Biofuel and Bioenergy Technology

Biofuel and
Bioenergy
Technology
Edited by
Wei-Hsin Chen, Keat Teong Lee and Hwai Chyuan Ong
Printed Edition of the Special Issue Published in Energies
www.mdpi.com/journal/energies
Biofuel and Bioenergy Technology
Special Issue Editors

Wei-Hsin Chen
Keat Teong Lee
Hwai Chyuan Ong
MDPI • Basel • Beijing • Wuhan • Barcelona • Belgrade

Special Issue Editors
Wei-Hsin Chen Keat Teong Lee Hwai Chyuan Ong
National Cheng Kung University Universiti Sains Malaysia Universiti Sains Malaysia
Taiwan Malaysia Malaysia
Editorial Office
MDPI
St. Alban-Anlage 66
4052 Basel, Switzerland
This is a reprint of articles from the Special Issue published online in the open access journal Energies
(ISSN 1996-1073) from 2017 to 2019 (available at: https://www.mdpi.com/journal/energies/special
issues/biofuel bioenergy)
For citation purposes, cite each article independently as indicated on the article page online and as
indicated below:
LastName, A.A.; LastName, B.B.; LastName, C.C. Article Title. Journal Name Year, Article Number,
Page Range.
ISBN 978-3-03897-596-0 (Pbk)

ISBN 978-3-03897-597-7 (PDF)

c 2019 by the authors. Articles in this book are Open Access and distributed under the Creative
Commons Attribution (CC BY) license, which allows users to download, copy and build upon
published articles, as long as the author and publisher are properly credited, which ensures maximum
dissemination and a wider impact of our publications.
The book as a whole is distributed by MDPI under the terms and conditions of the Creative Commons
license CC BY-NC-ND.
Contents
About the Special Issue Editors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Wei-Hsin Chen, Keat Teong Lee and Hwai Chyuan Ong

Reprinted from: Energies 2019, 12, 290, doi:10.3390/en12020290 . . . . . . . . . . . . . . . . . . . . 1
Qitai Eri, Wenzhen Wu and Xinjun Zhao

Numerical Investigation of the Air-Steam Biomass Gasification Process Based on
Thermodynamic Equilibrium Model
Reprinted from: Energies 2017, 10, 2163, doi:10.3390/en10122163 . . . . . . . . . . . . . . . . . . . 13
Jung-Chen Wu, Wei-Mon Yan, Chin-Tsan Wang, Chen-Hao Wang, Yi-Hao Pai,
Kai-Chin Wang, Yan-Ming Chen, Tzu-Hsuan Lan and Sangeetha Thangavel
Treatment of Oily Wastewater by the Optimization of Fe2O3 Calcination Temperatures in
Innovative Bio-Electron-Fenton Microbial Fuel Cells
Reprinted from: Energies 2018, 11, 565, doi:10.3390/en11030565 . . . . . . . . . . . . . . . . . . . . 32
Yan-Ming Chen, Chin-Tsan Wang and Yung-Chin Yang

Effect of Wall Boundary Layer Thickness on Power Performance of a Recirculation Microbial
Fuel Cell
I-Ching Kuan, Wei-Chen Kao, Chun-Ling Chen and Chi-Yang Yu

Microbial Biodiesel Production by Direct Transesterification of Rhodotorula glutinis Biomass
Rei-Yu Chein and Wen-Hwai Hsu

Analysis of Syngas Production from Biogas via the Tri-Reforming Process
Hoang Chinh Nguyen, Dinh Thi My Huong, Horng-Yi Juan, Chia-Hung Su and
Chien-Chung Chien
Liquid Lipase-Catalyzed Esterification of Oleic Acid with Methanol for Biodiesel
Production in the Presence of Superabsorbent Polymer: Optimization by Using Response
Surface Methodology
Jeeban Poudel, Sujeeta Karki and Sea Cheon Oh

Valorization of Waste Wood as a Solid Fuel by Torrefaction
Mohammed Ali Musa, Idrus Syazwani, Che Man Hasfalina and Nik Norsyahariati Nik Daud
Wastewater Treatment and Biogas Recovery Using Anaerobic Membrane Bioreactors
(AnMBRs): Strategies and Achievements
Hynek Roubı́k, Jana Mazancová, Phung Le Dinh, Dung Dinh Van and Jan Banout
Biogas Quality across Small-Scale Biogas Plants: A Case of Central Vietnam
v
Alvaro Fernandes, Joerg Brabandt, Oliver Posdziech, Ali Saadabadi, Mayra Recalde,
Liyuan Fan, Eva O.J. Promes, Ming Liu, Theo Woudstra,and P.V. Aravind
Design, Construction, and Testing of a Gasifier-Specific Solid Oxide Fuel Cell System
Bohwa Kim, Ramasamy Praveenkumar, Eunji Choi, Kyubock Lee, Sang Goo Jeon and
You-Kwan Oh
Prospecting for Oleaginous and Robust Chlorella spp. for Coal-Fired Flue-Gas-Mediated
Biodiesel Production
Aditi David, Tanvi Govil, Abhilash Kumar Tripathi, Julie McGeary, Kylie Farrar and
Rajesh Kumar Sani
Thermophilic Anaerobic Digestion: Enhanced and Sustainable Methane Production from
Co-Digestion of Food and Lignocellulosic Wastes
Reprinted from: Energies, 11, 2058, doi:10.3390/en11082058 . . . . . . . . . . . . . . . . . . . . . . 169
David Valero, Carlos Rico, Blondy Canto-Canché, Jorge Arturo Domı́nguez-Maldonado,

Raul Tapia-Tussell, Alberto Cortes-Velazquez and Liliana Alzate-Gaviria
Enhancing Biochemical Methane Potential and Enrichment of Specific Electroactive
Communities from Nixtamalization Wastewater using Granular Activated Carbon as a
Conductive Material
Urszula Dzieko ńska-Kubczak, Joanna Berłowska, Piotr Dziugan, Piotr Patelski,

Katarzyna Pielech-Przybylska and Maria Balcerek
Nitric Acid Pretreatment of Jerusalem Artichoke Stalks for Enzymatic Saccharification and
Bioethanol Production
Anna Brunerová, Hynek Roubı́k and Milan Brožek

Bamboo Fiber and Sugarcane Skin as a Bio-Briquette Fuel
José Marı́a Encinar, Ana Pardal, Nuria Sánchez and Sergio Nogales
Biodiesel by Transesterification of Rapeseed Oil Using Ultrasound: A Kinetic Study of
Base-Catalysed Reactions
Guan-Bang Chen, Jia-Wen Li, Hsien-Tsung Lin, Fang-Hsien Wu and Yei-Chin Chao
A Study of the Production and Combustion Characteristics of Pyrolytic Oil from Sewage Sludge
Using the Taguchi Method
Seong Ju Kim, Byung Hwan Um, Dong Joong Im, Jin Hyung Lee and Kyeong Keun Oh
Combined Ball Milling and Ethanol Organosolv Pretreatment to Improve the Enzymatic
Digestibility of Three Types of Herbaceous Biomass
Chia-Hung Su, Hoang Chinh Nguyen, Uyen Khanh Pham, My Linh Nguyen and
Horng-Yi Juan
Biodiesel Production from a Novel Nonedible Feedstock, Soursop (Annona muricata L.) Seed Oil
vi
Mohammad Anwar, Mohammad G. Rasul, Nanjappa Ashwath and Md Mofijur Rahman
Optimisation of Second-Generation Biodiesel Production from Australian Native Stone Fruit
Oil Using Response Surface Method
Tae Hoon Kim, Dongjoong Im, Kyeong Keun Oh and Tae Hyun Kim
Effects of Organosolv Pretreatment Using Temperature-Controlled Bench-Scale Ball Milling on
Enzymatic Saccharification of Miscanthus × giganteus
Ziaur Rahman, Javed Nawab, Bong Hyun Sung and Sun Chang Kim
A Critical Analysis of Bio-Hydrocarbon Production in Bacteria: Current Challenges and
Future Directions
Chen Li, Ashanti M. Sallee, Xiaoyu Zhang and Sandeep Kumar

Electrochemical Hydrogenation of Acetone to Produce Isopropanol Using a Polymer Electrolyte
Membrane Reactor
Mohammad Anwar, Mohammad G. Rasul and Nanjappa Ashwath

A Systematic Multivariate Analysis of Carica papaya Biodiesel Blends and Their Interactive Effect
on Performance
Mehdi Bidabadi, Peyman Ghashghaei Nejad, Hamed Rasam, Sadegh Sadeghi and
Bahman Shabani
Mathematical Modeling of Non-Premixed Laminar Flow Flames Fed with Biofuel
in Counter-Flow Arrangement Considering Porosity and Thermophoresis Effects:
An Asymptotic Approach
Reprinted from: Energies 2018, 11, 2945, doi:10.3390/en11112945. . . . . . . . . . . . . . . . . . . . 372
David Längauer, Yu-Ying Lin, Wei-Hsin Chen, Chao-Wen Wang, Michal Šafář and
Vladimı́r Čablı́k
Simultaneous Extraction and Emulsification of Food Waste Liquefaction Bio-Oil
Martin Černý, Monika Vı́tězová, Tomáš Vı́těz, Milan Bartoš and Ivan Kushkevych
Variation in the Distribution of Hydrogen Producers from the Clostridiales Order in Biogas
Reactors Depending on Different Input Substrates
vii
About the Special Issue Editors
Wei-Hsin Chen (Distinguished Professor) received a B.S. from the Department of Chemical
Engineering, Tunghai University in 1988, completed his Ph.D. at the Institute of Aeronautics and
Astronautics, National Cheng Kung University in 1993. After receiving his Ph.D., Dr. Chen worked
in an iron and steel corporation as a process engineer for one and a half years (1994–1995). He joined
the Department of Environmental Engineering and Science, Fooyin University in 1995 and was
promoted to a full professor in 2001. In 2005, he moved to the Department of Marine Engineering,
National Taiwan Ocean University. Two years later (2007), he moved to the Department of Greenergy,
National University of Tainan. Currently, he is a faculty member and distinguished professor at
the Department of Aeronautics and Astronautics, National Cheng Kung University. Professor Chen
visited Princeton University, USA, from 2004 to 2005, the University of New South Wales, Australia,
in 2007, the University of Edinburgh, UK, in 2009, the University of British Columbia, Canada,
from 2012 to 2013, and the University of Lorraine, France, in 2017 and 2019 as a visiting professor
and invited lecturer. His teaching courses at the National Cheng Kung University include Bioenergy,
Materials Engineering and Science, Energy Experiments, and Engineering Mathematics. His research
topics include bioenergy, hydrogen energy, clean energy, carbon capture, and aerosol physics. He
has published over 500 papers in international and domestic journals and conferences. He is the
editor, associate editor, guest editor, and editorial member of a number of international journals,
including Applied Energy, Energy Conversion and Management, International Journal of Hydrogen Energy,
International Journal of Energy Research, and Energies. He is also the author of several books concerning
energy science and air pollution. His important awards include the 2015 Outstanding Research
Award (Ministry of Science and Technology, Taiwan), the 2015 Highly Cited Paper Award (Applied
Energy, Elsevier), the 2017 Outstanding Engineering Professor Award (Chinese Institute of Engineers,
Taiwan) as well as the 2016, 2017, and 2018 Highly Cited Researcher Award (Web of Science).
Keat Teong Lee (Professor. Dr.) obtained his Ph.D in Chemical Engineering from Universiti Sains
Malaysia (USM) in 2004. He currently holds the position of Professor at the School of Chemical
Engineering, and he is also the Director of the Research Creativity & Management Office, Universiti
Sains Malaysia. Dr. Lee has co-authored two books, 10 book chapters, 30 review papers and more
than 100 research papers in peer reviewed international journals. Dr. Lee is currently the Co-Editor
for Energy Conversion and Management (Elsevier) and an Editorial Board Member for Bioresource
Technology (Elsevier) and Energy Science & Engineering (Wiley). He has also won numerous awards
including the Young Scientist Award 2011 from The International Forum on Industrial Bioprocess
and the 2012 Top Research Scientists Malaysia Award from the Academy Sciences of Malaysia. He
is one of the four Malaysian researchers to be recognized as one of the most cited researchers in
the latest Shanghai Academic Ranking of World Universities 2016 by Subjects (Energy Science &
Engineering). Dr. Lee is now working on the production of biofuels (biodiesel and bioethanol) from
biomass (including macro and microalgae) using various technologies. Apart from that, he also has
a special interest in the social and sustainability aspects of biofuels.
ix
Hwai Chyuan Ong (Dr.) obtained his B.Eng. (Hons.) in Mechanical Engineering from the Faculty
of Engineering, University of Malaya, with distinction. Then, he obtained his Ph.D. in Mechanical
Engineering from the same university in December 2012. His research interests are wide-ranging
under the general umbrella of renewable energy. In particular, these include biofuel and bioenergy,
solar thermal, and green technology and environmental science. He is currently appointed as a Senior
Lecturer at the Department of Mechanical Engineering, University of Malaya. He is also a Chartered
Engineer of Engineering Council (CEng) under the Institution of Mechanical Engineers (IMechE),
United Kingdom. He has published more than 100 high impact SCI journal papers with an H-index
of 25 (WOS). In 2017 and 2018, he received the Malaysia’s Research Star Award (frontier researcher)
and in 2016, he received the Malaysia’s Rising Star Award (young researcher) from the Ministry of
Higher Education and Clarivate Analytics. In 2018, he also received the Outstanding Research Award
and the most Highly Cited Paper Award at the University of Malaya Excellence Awards. Currently,
he is an Associate Editor of the Journal of Renewable and Sustainable Energy (IF: 1.135) and was a Guest
Editor in Energies (SCI, IF = 2.262) in the Special Issues “Biofuel and Bioenergy Technology” and
“Biomass Processing for Biofuels, Bioenergy and Chemicals”.
x
energies
Editorial
Wei-Hsin Chen 1, *, Keat Teong Lee 2, * and Hwai Chyuan Ong 3, *
1 Department of Aeronautics and Astronautics, National Cheng Kung University, Tainan 701, Taiwan
2 School of Chemical Engineering, Universiti Sains Malaysia, Nibong Tebal 14300, Pulau Pinang, Malaysia
3 Department of Mechanical Engineering, University of Malaya, Kuala Lumpur 50603, Malaysia
* Correspondence: [email protected] or [email protected] (W.-H.C.); [email protected] (K.T.L.);
[email protected] (H.C.O.)
Received: 24 December 2018; Accepted: 10 January 2019; Published: 18 January 2019
1. Introduction
Biomass is considered as a renewable resource because of its short life cycle, and biomass-derived
biofuels are potential substitutes to fossil fuels. When biomass grows, all carbon in biomass comes
from the atmosphere and is liberated into the environment when it is burned. Therefore, biomass is
thought of as a carbon-neutral fuel. For these reasons, the development of bioenergy is an effective
countermeasure to elongate fossil fuel reserves, lessen greenhouse gas (GHG) emissions, and mitigate
global warming and climate change. Biomass can be converted into biofuels through a variety of routes
such as physical, thermochemical, chemical, and biological methods. The common and important
biofuels for bioenergy include charcoal, biochar, biodiesel, bioethanol, biobutanol, pyrolysis and
liquefaction bio-oils, synthesis gas (syngas), biogas, and biohydrogen, etc. On account of the merit
of bioenergy for environmental sustainability, biofuel and bioenergy technology plays a crucial role
for renewable energy development. This Special Issue aims to publish high-quality review and
research papers, addressing recent advances in biofuel and bioenergy. State-of-the-art studies of
advanced techniques of biorefinery for biofuel production are also included. Research involving
experimental studies, recent developments, and novel and emerging technologies in this field are
covered. The particular topics of interest in the original call for papers included, but were not limited to:
• Novel and unexploited biomass resources for biofuel and bioenergy production
• New emerging technologies for biofuel and bioenergy production
• Development of thermochemical conversion routes for biofuel and bioenergy produciton
• Advanced biorefinery processes for biofuel and biochemicals production
• Bioreactors or microbial fuel cell for bioenergy and power production
• State-of-the-art review in the progress of biofuel and bioenergy technology
This Special Issue of Energies on the subject of “Biofuel and Bioenergy Technology” contains
the successful invited submissions [1–27]. A total of twenty-seven technical papers which cover
diversified biofuel and bioenergy technology related researches have shown critical results and
contributed significant findings in biomass processing [1,2], bio-oil and biodiesel [3–11], syngas [12–14],
biogas/methane [15–19], bioethanol and alcohol-based fuels [20–22], solid fuel [23–25] and also
microbial fuel cell [13,26,27] developments.
2. Statistics of the Special Issue

The response to our call had the following statistics:
• Submissions (46);
• Publications (27);
• Rejections (19);
Energies 2019, 12, 290; doi:10.3390/en12020290 1 www.mdpi.com/journal/energies

Energies 2019, 12, 290
• Article types: research articles (25); review articles (2).
The authors’ geographical distribution (published papers) is:
• Taiwan (8);
• Korea (4);
• Czech Republic (3)
• Australia (3);
• USA (2);
• China (1);
• Malaysia (1);
• Mexico (1);
• Pakistan (1);
• Poland (1);
• Spain (1);
• The Netherlands (1).
Published submissions are related to the most important techniques and analysis applied to the
biofuel and bioenergy technology. In summary, the edition and selections of papers for this special
issue are very inspiring and rewarding. We thank the editorial staff and reviewers for their efforts and
help during the process.
3. Brief Overview of the Contributions to This Special Issue

Table 1 provides some of the key information, including the research type, field of study,
final product as well as the key findings. As observed, a majority of the publications (twenty-three
papers) focus on experimental work to improve or explore novel technologies for energy-products
synthesis, while three papers focus on modelling studies and two papers focus on literature review
studies. The following discussion highlights and groups the research findings in accordance to the
corresponding research field or work.
As the initial step in most synthesis routes, biomass processing can enhance the substrate’s quality
for other synthesis processes. Thus, commonly, these are treated as pretreatment to enhance the
characteristics of the biomass. In two research works [1,2], the combination of physical treatment
(ball milling) and chemical treatment (ethanol organosolv) showed improved glucan digestibility.
Three different biomasses such as giant miscanthus, corn stover and wheat straw were pretreated
with ball milling and ethanol organosolv and the overall biomass size was reduced as a result of the
prolonged pre-treatment [1]. Due to the improved physicochemical characteristics resulting from the
pre-treatment, a maximum of 91% glucan digestibility could be achieved. A parametric study on
combined ball milling and organosolv was performed as well to optimize the glucan digestibility [2].
It was determined that at 170 ◦ C, with reaction time of 90 min and ethanol concentration of 40%
and liquid/solid ratio of 10, the pretreatment process achieved the best results. Thus, the biomass
processing method could be beneficial in generating desired products.
2
Table 1. Key Information of the Publications Submitted to Special Issue.
Research Research Type Technology/

Product Key Findings
Work Experimental Modelling Review Field of Work
Biodiesel • Reduction in brake power, torque and brake

Energies 2019, 12, 290
Anwar et al., blend of thermal efficiency.

x Blending
2018 [5] papaya seed • Significant effect on brake specific fuel consumption.
oil
Australian • Optimisation with response surface methodology.

Anwar et al., Alkali-catalysed native stone • Maximum biodiesel yield of 95.8%.
x • Met ASTM D6751 and EN14214 standards.
2018 [6] transesterification fruit
biodiesel • Potential second-generation biodiesel.
• Oxidizer and fuel Lewis number were between 0.4 and 1,

the maximum flame temperature was ~1860 K.
• Per unit of fuel Lewis number, the minimum
3
Mathematic thermophoretic force was −1.48 × 10−8 N.
Bidabadi et
x asymptotic - • Per unit of oxidizer Lewis number, the minimum
al., 2018 [25]
technique thermophoretic force was −1.53 × 10−8 N.
• Per unit of porosity factor, the minimum thermophoretic
force was −1.28 × 10−8 N.
• Low ash content for bamboo fibre (1.16%) and sugarcane

skin (8.62%).
Brunerová et High-Pressure Bio-Briquette • Satisfactory mechanical durability for bamboo fibre
x (97.80%) and sugarcane skin (97.70%).
al., 2018 [24] Densification Fuel
• These products can be used for bio-briquette
fuel production.
Table 1. Cont.

• Occurrence of potentially harmful microorganisms such as

Energies 2019, 12, 290
Biogas study
Černý et al., Biogas Clostridium novyi was detected at higher ratio (65.63%) in
x with DNA
2018 [15] (Hydrogen) the population of the bioreactor.
analysis
• First-Law Efficiency increased with increased reaction

temperature for higher hydrogen and carbon
Chein et al., Tri-Reforming Syngas monoxide yields.
x
2018 [12] Process (hydrogen) • Second-Law Efficiency decreased with increased reaction
temperature due to more complete chemical reaction.
• Optimisation with Taguchi Method.

Chen et al., • Maximum pyrolytic oil yield of 10.19%.
4
x Pyrolysis Pyrolytic Oil • Synthesis conditions: 450 ◦ C, 60 min, 10 ◦ C/min and
2018 [3]
nitrogen flow of 700 mL/min.
• Hydrodynamic boundary layer of 1.6 cm (thin layer)

Chen et al., Microbial showed maximum voltage of 22 mV and charged transfer
x -
2018 [26] Fuel Cell resistance of 39 Ω.
• Food wastes (corn stover, prairie cordgrass and unbleached

Thermophilic paper) undergone thermophilic anaerobic digestion.
David et al.,
x anaerobic Methane • Highest methane yield of 305.45 L/kg was achieved after
2018 [18]
digestion 30 days of incubation at 60 ◦ C at 100 rpm.
Table 1. Cont.

Nitric acid • Jerusalem artichoke stalks were converted into bioethanol

Energies 2019, 12, 290
Dziekońska- pretreatment for with nitric acid as catalyst.

Kubczak et x enzymatic Bioethanol • Nitric acid pretreated hydrolysates led to 30%
al., 2018 [20] hydrolysis and improvement in ethanol yield (77–82% of theoretical yield).
fermentation
Transesterification • Ultrasonic accelerated rate of biodiesel

Encinar et with base- transesterification reactions.
x Biodiesel
al., 2018 [7] catalysed • Reaction followed a pseudo-first order kinetic model.
reactions
Equilibrium • Simulations were performed with two different models

Eri et al., (with and without tar).
x constants -
2018 [14] • The simulations were validated by experimental data.
modelling
5
Fernedas Gasifier-Specific • Validation data showed good agreement between
et al., x x Solid Oxide Fuel - experimental and simulation data.
2018 [13] Cell System • System efficiencies were estimated to be 33.7–34.5%.
• M082 strain showed maximum lipid content (397 mg fatty

Photobioreactor acid methyl ester (FAME)/g cell) with good tolerance to
Kim et al., Microalgal
x with coal-fired high temperature.
2018 [10] biodiesel
flue-gas • FAME produced met the international standards.
Table 1. Cont.

• Combined pretreatment on giant miscanthus, corn stover

Energies 2019, 12, 290
and wheat straw show varied results (increased of glucan

Ball milling and content for giant miscanthus, removal of cellulose for corn
Kim et al.,
x ethanol - stover).
2018 [1]
organosolv • Enzymatic digestibility was improved with 91%
glucan digestibility.
• Pretreatment was performed using a 30 L bench-scale ball

mill reactor.
• Pretreatment conditions were varied: room temperature to
Ball milling and 170 ◦ C, time from 30 to 120 min, ethanol concentration from
Kim et al.,
x ethanol - 30% to 60%, liquid/solid ratio from 10 to 20.
2018 [2]
organosolv • Highest glucan digestibility was performed at 170 ◦ C,
6
reaction time to 90 min, 40% of ethanol concentration and
L/S = 10.
• Acid-catalysed synthesis by 0.6 M sulphuric acid at 70 ◦ C

for 20 h yielded 111% of FAME.
Kuan et al., • Base-catalysed synthesis by 1.0 g/L of sodium hydroxide at
x Transesterification Biodiesel 70 ◦ C for 10 h yielded 102% of FAME.
2018 [8]
• Direct transesterification shortened the reaction time and
improved FAME yield.
Table 1. Cont.

• Emulsified ratio (bio-oil to emulsifier, B/E ratio) at 1

Energies 2019, 12, 290
Simultaneous showed higher solubility of 66.48 wt %.

Längauer et Emulsified • At higher temperature, higher solubility was also observed.
x Extraction and
al., 2018 [4] bio-oil • Methanol as co-surfactant also improved better solubility
Emulsification
from 58.83 to 70.96 wt %.
Electrochemical • Polymer electrolyte membrane fuel cell was used to

Hydrogenation produce isopropanol as main product and diisopropyl ether
Li et al., using polymer as byproduct.
x Isopropanol • High selectivity and (>90%) and high current efficiency
2018 [21] electrolyte
membrane (59.7%) were observed at mild conditions of 65 ◦ C and at
reactor atmospheric pressure.
7
• Anaerobic digestion technologies were critically reviewed.
Anaerobic • Factors on membrane fouling, microbial environment
Musa et al., Membrane conditions as well as parameters on the operations of
x - AnMBRs were discussed.
2018 [19] Bioreactors
(AnMBRs) • Microfiltration as the mean to reduce energy and water
usage in the AnMBRs was suggested.
• Optimisation with Response Surface Methodology

• Superadsorbent polymer (SAP), as water removal agent,
Liquid Lipase was used in esterification.
Nguyen et
x Catalyzed Biodiesel • The polymer improved the conversion to 96.73% at 35.25
al., 2018 [11]
Esterification ◦ C, methanol to oleic acid molar ratio of 3.44:1, SAP
loading of 10.55% and enzyme loading of 11.98%.
Table 1. Cont.

• Wood waste was torrefied at 200–400 ◦ C and 0–50 min.

Energies 2019, 12, 290
Poudel et al., Torrefied • 300 ◦ C as the optimal temperature for torrefaction based on
x Torrefaction
2018 [23] Biomass Van Krevelen diagram.
• Bioenergy products (alcohols and n-alkene hydrocarbons

Bio-hydrocarbon (C2 to C18 ) as produced by engineered microorganisms
Rahman et showed promising energy potential.
x Production in -
al., 2018 [22] • The review discussed the complexity of metabolic networks
Bacteria
to obtain these bio-hydrocarbon products.
• Biogas composition was measured for 107 small-scale

biogas plants, respectively.
8
Roubík, et Biogas Plant Biogas • Mean compositions as follows: For plants younger than 5
x
al., 2018 [16] Study (methane) years, CH4 was 65.44% and CO2 was 29.31%; for plants
older than 5 years, CH4 was 64.57% and CO2 was 29.93%.
• Soursop (Annona muricata) seeds were used to produce

bio-oil (29.6% (w/w)).
• Bio-diesel with highest of 97.02% was produced under
Two-step acid-catalysed conditions of 65 ◦ C, 1% sulphuric acid,
Su et al., reaction time of 90 min and methanol: oil ratio of 10:1 and
x acid-catalysed Biodiesel
2018 [9] under base-catalysed conditions of 65 ◦ C, 0.6% NaOH,
transesterification
reaction time of 30 min and methanol: oil ratio of 8:1.
• Produced biodiesel met the EN14214 and
D6751 requirements.
Table 1. Cont.

• Biochemical Methane Potential (BMP) showed that the

Energies 2019, 12, 290
addition of granular activated carbon (GAC) improved the

Valero et al., methane yield by 34% for instance testing and 54% for 10
x - Biomethane days of GAC biofilm development.
2018 [17]
• Addition of GAC can improve digester’s anaerobic
digestion performance.
• Different calcination temperatures (500–900 ◦ C) of iron

oxide (Fe2 O3 ) were tested to investigate their photocatalytic
properties within the cathodic chambers.
• Calcinated Fe2 O3 improved the bio-electro-Fenton
Wu et al., Microbial Fuel microbial fuel cell (Bio-E-Fenton-MFC) on degrading
x -
2018 [27] Cell oily wastewater.
9
• Within one hour, oily water was best-degraded up to 99.3%
with electrode material synthesised at 500 ◦ C with
maximum power density of 52.5 mW/m2 .
Energies 2019, 12, 290
Bio-oils can be synthesized from sewage sludge by using pyrolysis techniques [3].
Taguchi optimization suggested the best pyrolysis was performed at 450 ◦ C, 60 min and 10 ◦ C/min,
which also showed consistency with other research work. Nonetheless, under most conditions,
pyrolytic oil/bio-oil requires further processing or upgrading for use as biodiesel. To maintain
the stability of bio-oil, blending with emulsifier resulted in high solubility (58.83–70.96 wt %) [4].
These findings suggest that simple blending could improve the properties of biodiesel or bio-oil
tremendously, which is worthy of further investigation. Aside from using bio-oil as a precursor for
biodiesel [3], biodiesel could be directly synthesized using other oil materials such as Australian native
stone fruit oil [6], rapeseed oil [7], Rhodotorula glutinis [8] and soursop seed oil [9] via transesterification
techniques. Transesterification of Australian native stone biodiesel showed a high yield of 95.8%
with the response surface methodology optimization and its quality fulfilled the ASTM D6751 and
EN14214 requirements [6]. Kuan et al. [8] investigate both direct acid and base-transesterification on
Rhodotorula glutinis biomass which gave 111% yield of FAME and 102% yield of FAME, respectively,
which were regarded as of good biodiesel quality. Another research work by Su et al. [9] used
soursop seed to produce bio-oil which was eventually upgraded to biodiesel using a two-step
acid catalyzed transesterification. The biodiesel produced met both EN14214 and D6751 standards.
Encinar et al. [7] performed rapeseed transesterification with KOH catalyst as well as with the aid
of ultrasound whereby the kinetic behavior obeyed a pseudo-first order trend. Liquid lipase-based
esterification was attempted and optimized using RSM to enhance the usage of water removal agent
in the system [11]. In Anwar et al’s, [5] work, it was found that by blending papaya oil biodiesel with
varying contents (5–20%) with diesel could improve the engine testing properties. Microalgal biodiesel
was generated using photobioreactor with coal-fired flue gas from three strains (M082, M134 and
KR-1) [10]. Among the strains, M082 generated high lipid value of 397 mg/g which was regarded to
be a suitable feedstock for biodiesel production.
In recent years, gasification also garners high interest due to its rapid processing step and high
yield of syngas which could be directly used for combustion. One of the main constitutes of syngas
is hydrogen which usually provides high calorific value. Based on Chein and Hsu’s [12] work,
the tri-reforming process could produce good quality syngas. In addition, it was also found that at
higher reaction temperatures, more hydrogen and carbon monoxide were produced. In pilot plant
study, the gasifier which was embedded with specific solid oxide fuel cell system in an industrial scale
was investigated in details [13]. To further understanding the gasification process, a thermodynamic
equilibrium constants derivation and modelling was performed for two cases, with and without tar.
The simulated data were validated with experimental data [14]. These findings could serve as good
guides for future development of gasification process.
Bio-digester or bioreactor could also be used for biogas production. Černý et al. [15] discovered
that microorganisms such as Clostridium novyi were detected at higher ratio (65.63%) in the population
of the bioreactor for the biogas production. Such detection could serve as an important reminder to
seek ways to inhibit these harmful microorganisms in the system. In an investigation and survey of 107
biogas plants, it was also found that the younger plant (<5 years) produced higher CH4 (65.44%) and
CO2 (29.31%) [16]. Addition of granular activated carbon (GAC) in the digestion system could also
directly improve the methane production by more than 34% [17]. Thermophilic anaerobic digestion is
another interesting field of research. David et al. [18] found that by co-digestion under such conditions,
high yield of methane (up to 305.45 kg/L) could be achieved. Musa’s work critically reviewed some of
the more critical findings on anaerobic membrane reactors for biogas recovery especially on membrane
fouling and parameters of operation [19].
The studies on alcohol-based biofuels are increasing due to its high energy-content and
suitability as fuel products. Dziekońska-Kubczak et al. [20] report that nitric acid as a form of
chemical pretreatment could enhance the bioethanol production up to 30%. A Polymer Electrolyte
Membrane Fuel Cell was used to produce isopropanol from acetone for use as a biofuel [21].
The hydrogenation process consumes less energy and less chemical wastes compared to other
10
Energies 2019, 12, 290
techniques [21]. Bio-hydrocarbons (alcohol and alkenes) produced from bacteria and their synthesis
mechanisms are reviewed by Rahman et al. [22], as well as future challenges and complexity.
High pressure densification and torrefaction are currently attracting attention among the research
community as these methods can produce potential solid-based fuels which require no further upgrade
and can be use directly. Both methods are usually applied in mild synthesis conditions which
differ from common thermochemical conversion techniques like slow pyrolysis and fast pyrolysis.
For example, biomass with low ash content (1.16–8.62%) and good mechanical durability (97%) such as
bamboo fibre and sugarcane skin could be directly densified as bio-briquette fuel without any energy
processing [24]. As for torrefaction, a form of mild pyrolysis, wood wastes could be converted into
torrefied biomass as low as 300 ◦ C [23]. As observed, both methods consume relatively lower energy
requirement and are simpler in term of synthesis process. In a modelling study, Lycopodium particles
were also simulated and modelled as biofuel and burned in air environment [25]. It was discovered
that the particles of Lycopodium were greatly influenced by thermophoretic force. Microbial fuel cell
studies were also being investigated thoroughly. The effect of the hydrodynamic layer thickness
was found to be significant on the voltage and charged transfer resistance [26]. In another study,
the calcination temperature on the cathodic chambers was studied and it was found that the electrode
synthesized at 500 ◦ C could degrade oily wastewater up to 99.3% [27]. Thus, the microbial fuel cell
shows tremendous potential to be developed for other applications.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Kim, S.; Um, B.; Im, D.; Lee, J.; Oh, K. Combined Ball Milling and Ethanol Organosolv Pretreatment
to Improve the Enzymatic Digestibility of Three Types of Herbaceous Biomass. Energies 2018, 11, 2457.
[CrossRef]
2. Kim, T.; Im, D.; Oh, K.; Kim, T. Effects of Organosolv Pretreatment Using Temperature-Controlled Bench-Scale
Ball Milling on Enzymatic Saccharification of Miscanthus × giganteus. Energies 2018, 11, 2657. [CrossRef]
3. Chen, G.-B.; Li, J.-W.; Lin, H.-T.; Wu, F.-H.; Chao, Y.-C. A Study of the Production and Combustion
Characteristics of Pyrolytic Oil from Sewage Sludge Using the Taguchi Method. Energies 2018, 11, 2260.
[CrossRef]
4. Längauer, D.; Lin, Y.-Y.; Chen, W.-H.; Wang, C.-W.; Šafář, M.; Čablík, V. Simultaneous Extraction and
Emulsification of Food Waste Liquefaction Bio-Oil. Energies 2018, 11, 3031. [CrossRef]
5. Anwar, M.; Rasul, M.; Ashwath, N. A Systematic Multivariate Analysis of Carica papaya Biodiesel Blends
and Their Interactive Effect on Performance. Energies 2018, 11, 2931. [CrossRef]
6. Anwar, M.; Rasul, M.; Ashwath, N.; Rahman, M. Optimisation of Second-Generation Biodiesel Production
from Australian Native Stone Fruit Oil Using Response Surface Method. Energies 2018, 11, 2566. [CrossRef]
7. Encinar, J.; Pardal, A.; Sánchez, N.; Nogales, S. Biodiesel by Transesterification of Rapeseed Oil Using
Ultrasound: A Kinetic Study of Base-Catalysed Reactions. Energies 2018, 11, 2229. [CrossRef]
8. Kuan, I.-C.; Kao, W.-C.; Chen, C.-L.; Yu, C.-Y. Microbial Biodiesel Production by Direct Transesterification of
Rhodotorula glutinis Biomass. Energies 2018, 11, 1036. [CrossRef]
9. Su, C.-H.; Nguyen, H.; Pham, U.; Nguyen, M.; Juan, H.-Y. Biodiesel Production from a Novel Nonedible
Feedstock, Soursop (Annona muricata L.) Seed Oil. Energies 2018, 11, 2562. [CrossRef]
10. Kim, B.; Praveenkumar, R.; Choi, E.; Lee, K.; Jeon, S.; Oh, Y.-K. Prospecting for Oleaginous and Robust
Chlorella spp. for Coal-Fired Flue-Gas-Mediated Biodiesel Production. Energies 2018, 11, 2026. [CrossRef]
11. Nguyen, H.; Huong, D.; Juan, H.-Y.; Su, C.-H.; Chien, C.-C. Liquid Lipase-Catalyzed Esterification of Oleic
Acid with Methanol for Biodiesel Production in the Presence of Superabsorbent Polymer: Optimization by
Using Response Surface Methodology. Energies 2018, 11, 1085. [CrossRef]
12. Chein, R.-Y.; Hsu, W.-H. Analysis of Syngas Production from Biogas via the Tri-Reforming Process. Energies
2018, 11, 1075. [CrossRef]
13. Fernandes, A.; Brabandt, J.; Posdziech, O.; Saadabadi, A.; Recalde, M.; Fan, L.; Promes, E.; Liu, M.;
Woudstra, T.; Aravind, P. Design, Construction, and Testing of a Gasifier-Specific Solid Oxide Fuel Cell
System. Energies 2018, 11, 1985. [CrossRef]
11
Energies 2019, 12, 290
14. Eri, Q.; Wu, W.; Zhao, X. Numerical Investigation of the Air-Steam Biomass Gasification Process Based on
Thermodynamic Equilibrium Model. Energies 2017, 10, 2163. [CrossRef]
15. Černý, M.; Vítězová, M.; Vítěz, T.; Bartoš, M.; Kushkevych, I. Variation in the Distribution of Hydrogen
Producers from the Clostridiales Order in Biogas Reactors Depending on Different Input Substrates. Energies
2018, 11, 3270. [CrossRef]
16. Roubík, H.; Mazancová, J.; Le Dinh, P.; Dinh Van, D.; Banout, J. Biogas Quality across Small-Scale Biogas
Plants: A Case of Central Vietnam. Energies 2018, 11, 1794. [CrossRef]
17. Valero, D.; Rico, C.; Canto-Canché, B.; Domínguez-Maldonado, J.; Tapia-Tussell, R.; Cortes-Velazquez, A.;
Alzate-Gaviria, L. Enhancing Biochemical Methane Potential and Enrichment of Specific Electroactive
Communities from Nixtamalization Wastewater using Granular Activated Carbon as a Conductive Material.
Energies 2018, 11, 2101. [CrossRef]
18. David, A.; Govil, T.; Tripathi, A.; McGeary, J.; Farrar, K.; Sani, R. Thermophilic Anaerobic Digestion:
Enhanced and Sustainable Methane Production from Co-Digestion of Food and Lignocellulosic Wastes.
19. Musa, M.; Idrus, S.; Che Man, H.; Nik Daud, N. Wastewater Treatment and Biogas Recovery Using Anaerobic
Membrane Bioreactors (AnMBRs): Strategies and Achievements. Energies 2018, 11, 1675. [CrossRef]
20. Dziekońska-Kubczak, U.; Berłowska, J.; Dziugan, P.; Patelski, P.; Pielech-Przybylska, K.; Balcerek, M. Nitric
Acid Pretreatment of Jerusalem Artichoke Stalks for Enzymatic Saccharification and Bioethanol Production.
21. Li, C.; Sallee, A.; Zhang, X.; Kumar, S. Electrochemical Hydrogenation of Acetone to Produce Isopropanol
Using a Polymer Electrolyte Membrane Reactor. Energies 2018, 11, 2691. [CrossRef]
22. Rahman, Z.; Nawab, J.; Sung, B.; Kim, S. A Critical Analysis of Bio-Hydrocarbon Production in Bacteria:
Current Challenges and Future Directions. Energies 2018, 11, 2663. [CrossRef]
23. Poudel, J.; Karki, S.; Oh, S. Valorization of Waste Wood as a Solid Fuel by Torrefaction. Energies 2018, 11,
1641. [CrossRef]
24. Brunerová, A.; Roubík, H.; Brožek, M. Bamboo Fiber and Sugarcane Skin as a Bio-Briquette Fuel. Energies
2018, 11, 2186. [CrossRef]
25. Bidabadi, M.; Ghashghaei Nejad, P.; Rasam, H.; Sadeghi, S.; Shabani, B. Mathematical Modeling of
Non-Premixed Laminar Flow Flames Fed with Biofuel in Counter-Flow Arrangement Considering Porosity
and Thermophoresis Effects: An Asymptotic Approach. Energies 2018, 11, 2945. [CrossRef]
26. Chen, Y.-M.; Wang, C.-T.; Yang, Y.-C. Effect of Wall Boundary Layer Thickness on Power Performance of a
Recirculation Microbial Fuel Cell. Energies 2018, 11, 1003. [CrossRef]
27. Wu, J.-C.; Yan, W.-M.; Wang, C.-T.; Wang, C.-H.; Pai, Y.-H.; Wang, K.-C.; Chen, Y.-M.; Lan, T.-H.; Thangavel, S.
Treatment of Oily Wastewater by the Optimization of Fe2O3 Calcination Temperatures in Innovative
Bio-Electron-Fenton Microbial Fuel Cells. Energies 2018, 11, 565. [CrossRef]
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
12
energies
Article
Numerical Investigation of the Air-Steam Biomass
Gasification Process Based on Thermodynamic
Equilibrium Model
Qitai Eri 1 , Wenzhen Wu 1 and Xinjun Zhao 1,2, *
1 School of Energy and Power Engineering, Beihang University, Beijing 100191, China;
[email protected] (Q.E.); [email protected] (W.W.)
2 The 41st Research Institute, the Sixth Academy of CASIC, Hohhot 010010, China
* Correspondence: [email protected]; Tel.: +86-010-8233-9866
Received: 20 November 2017; Accepted: 14 December 2017; Published: 18 December 2017
Abstract: In the present work, the air-steam biomass gasification model with tar has been developed
based on the equilibrium constants. The simulation results based on two different models (with and
without tar) have been validated by the experimental data. The model with tar can well predict the tar
content in gasification; meanwhile, the predicted gas yield (GY), based on the model with tar, is much
closer to the experimental data. The energy exchange between the gasifier and the surrounding has
been studied based on the dimensionless heat transfer ratio (DHTR), and the relationship between
DHTR and the process parameters is given by a formula. The influence of process parameters on
the syngas composition, tar content, GY, lower heating value (LHV), and exergy efficiency have
been researched.
Keywords: air-steam gasification; equilibrium model; tar; energy exchange; exergy efficiency
1. Introduction
Biomass gasification is a thermochemical conversion method [1,2], which converts biomass into
combustible gases through partial oxidation. Decomposition, partial oxidation, and various other
chemical reactions occur during gasification, involving heat and mass transfer processes. Biomass
gasification has gained attention worldwide due to its capacity to handle a wide range of biomass
feedstocks and its high efficiency of energy conversion [3–5].
Due to its inherent complexity, simulating the gasification process and predicting the process
performance is still at a preliminary stage [6]. In the simulation of gasification, thermodynamic
equilibrium modeling does not consider the mechanism of transformation; meanwhile, it is
independent of gasifier geometry and not limited to a specified range of operating conditions.
Equilibrium model represents the gasification process as a single reaction, so it is useful to predict
the exit gas composition and estimate the trend of the gasifier output with variations in process
parameters [7].
Equilibrium models can be categorized into stoichiometric models based on equilibrium constants
of the most important reactions [8] and non-stoichiometric models based on minimization of the
Gibbs free energy [9–13]. Some researchers [14–16] simulated the gasification process based on
non-stoichiometric models, whereas other researchers [17–19] paid more attention to stoichiometric
models. However, both the approaches are essentially equivalent [20]. Air, steam, and air-steam
are the generally used gasifying agents in gasification. Air is comprehensively used [21,22], since it
is cheap. Steam gasification was found to produce gas with a high heating value [23,24] compared
to air gasification due to the high content of H2 in the syngas. The pure steam gasification process
requires an indirect or external heat supply for endothermic gasification reaction, which increases the

Energies 2017, 10, 2163
cost of gasification. Alternatively, air-steam gasification [25,26] decreases the cost, since the partial
combustion reaction provides the heat. Meanwhile, the heating value of the syngas is still acceptable,
so this process has received a lot of attention.
In the air-steam gasification process, when the energy balance equation is introduced to the
equilibrium model, the value of energy exchange between the gasifier and the surrounding should
be given. If this value is inaccurate, the simulation results are very likely to be misleading. However,
this value is usually determined through experiments, which causes the data to be insufficient. If the
relationship between the energy exchange and the process parameters can be determined by a formula,
the value will be obtained accurately and conveniently. Tar is generally neglected in the simulation
study based on the stoichiometric models [18,27]. In fact, tar is invariably present in the gasification
process, especially at low gasification temperatures [28,29]. The components of tar are very complex,
and the primary tars are generally obtained at temperatures below 973 K. As temperature increases,
primary tars continue to decompose with the formation of secondary tars, comprising phenolics and
olefins [30]. In some simulations of gasification, phenol was used as the model compound for tars [31].
Hence, in this study, the air-steam biomass gasification model with tar has been developed based
on the equilibrium constants, and phenol is used as the model compound for tars. The effects of
process parameters on the syngas composition, tar content, gas yield, the lower heating value of the
syngas, and the exergy efficiency of gasification have been studied. Moreover, the relationship between
the energy exchange and the process parameters has also been researched.
2. Thermodynamic Equilibrium Model

For the present air-steam thermodynamic equilibrium model, some assumptions have been made
for the biomass gasification process: (1) temperature of the products is equal to the gasification
temperature; (2) oxygen is consumed completely; (3) the gas mixture is in equilibrium and is
homogeneous; (4) the residence time is long enough, and all the chemical reactions can reach an
equilibrium state; (5) nitrogen does not take part in the reactions; (6) the biomass is composed of C, H,
and O; (7) the solid products include only the char, which is composed of entirely of carbon; (8) ash is
not considered in the simulation; (9) phenol is regarded as tar; (10) atmospheric pressure is assumed in
the gasifier.
The overall gasification reaction with air and steam can be written as follows:
CHx Oy + m(O2 + 3.76N2 ) + nH2 O = x1 CO + x2 CO2 +

(1)
x3 H2 + x4 CH4 + x5 H2 O + 3.76mN2 + x6 C + x7 C6 H6 O
where x and y are the number of atoms of hydrogen and oxygen for each atom of carbon in the biomass,
respectively. m and n are the molar amounts of air and steam, respectively, fed per mole of biomass.
x1 , x2 , x3 , x4 , x5 , x6 , and x7 are the number of moles of the corresponding products, respectively.
The element balances of carbon, hydrogen, and oxygen elements are given below:
1 = x1 + x2 + x4 + x6 + 6x7 (2)
x + 2n = 2x3 + 4x4 + 2x5 + 6x7 (3)
y + 2m + n = x1 + 2x2 + x5 + x7 (4)
2.1. Equilibrium Constants and Non-Equilibrium Factor

When biomass is added to the gasifier, the pyrolysis reactions occur instantaneously. This is
followed by a number of competing intermediate reactions, including both heterogeneous reactions
and homogeneous gas phase reactions. Char, tar, and other gases (CO, CO2 , H2 O, H2 , and CH4 )
are produced during the gasification process. The three independent equilibrium reactions used for
equilibrium calculation are considered to be:
14
Energies 2017, 10, 2163
C + H2 O ↔ CO + H2 − 131 kJ/mol (5)
C + 2H2 ↔ CH4 + 75 kJ/mol (6)
CO + H2 O ↔ CO2 + H2 + 41 kJ/mol (7)
The above three reactions are water gas reaction, methane generation reaction, and water gas shift
reaction, respectively. The equilibrium constants for these three reactions can be shown as follows:
( PCO /Ptotal )( PH2 /Ptotal ) x x 1

K1 = = 1 3· (8)
( PH2 O /Ptotal ) x5 xtotal
( PCH4 /Ptotal ) x4
K2 = = · xtotal (9)
( PH2 /Ptotal )2 ( x3 )2
( PCO2 /Ptotal )( PH2 /Ptotal ) x ·x
K3 = = 2 3 (10)
( PCO /Ptotal )( PH2 O /Ptotal ) x1 · x5
where K and P are equilibrium constant and pressure, respectively. xtotal is the total number of moles
of syngas.
The equilibrium constant can be calculated based on the Gibbs free energy; meanwhile, the
standard Gibbs function of the reaction is dependent on the standard enthalpy change. Through
mathematical conversion, the equilibrium constant is associated with the standard enthalpy change
and a constant of integration, which can be found from the related reference [32]. From the data of
standard enthalpy of formation and standard Gibbs function of formation, also shown in that paper,
this constant of integration can be calculated. Thus, the correlations between the equilibrium constants
and the temperatures for the Reactions (5)–(7) are obtained shown below:
−15702.01 39900
ln K1 = + 1.384 ln T − 0.000621T + + 7.642 (11)
T T2
7082.848 35050
ln K2 = − 6.567 ln T + 0.003733T − 3.60667 × 10−7 T 2 + + 32.541 (12)
T T2
5872.461 58200
ln K3 = + 1.86 ln T − 0.000269T − − 18.014 (13)
T T2
In the actual gasification system, it is a fact that methane generation reaction is not at equilibrium
state. Some researchers [17,27] found that the methane yield was underestimated in the simulation
based on the conventional thermodynamic equilibrium model. Hence, it is necessary to consider the
non-equilibrium behavior. According to the linear regression analysis of many experimental data sets,
the dependency of CH4 with respect to the process parameters was obtained [17]. It was observed
that the methane yield had a significant relationship with ER, and that was consistent with the results
in the literature [33,34]. Meanwhile, the tar is formed in pyrolysis stage accompanied with H2 , CO,
CO2 , H2 O, and CH4 , and it can also convert to these gases according to the decomposition reactions.
So, it is reasonable to believe that the equilibrium relationship between these gases will not change,
and the effect of the tar on these gases is just revealed from Equation (9). Hence, the non-equilibrium
modification is needed to apply to K2 , which is shown below:
K2∗ = (38.75 − 30.7· ER)·K2 (14)
where ER is equivalence ratio, and it is defined as the ratio of the actual amount of air fed to the gasifier
to the stoichiometric amount of air necessary for the combustion of biomass.
Meanwhile, the Equations (8) and (9) can be combined to give the Equation (15) as shown below:
x1 · x4
K1 · K2 = (15)
x3 · x5
15
Energies 2017, 10, 2163
In the gasification process, the carbon in biomass gets partially converted into gas and the rest is
converted into char. Hence, the carbon conversion ratio is defined as the ratio of total number of moles
of carbon in syngas to the number of moles of carbon in the biomass feed. Many studies [34–36] have
found that carbon conversion ratio cannot be calculated directly, and the empirical formula obtained
based on experimental data sets will be suitable. Generally, the carbon conversion ratio can be fitted
as a function of process parameters, such as ER [34], or both ER and temperature [17,37]. It has been
found that the temperature has a small effect on carbon conversion ratio. An empirical formula was
proposed by Lim et al. [17] based on many experimental data points under different experimental
conditions, and the temperature influence had been considered. So, this formula can be considered as
a representative one, and it is shown as follows:
f = x1 + x2 + x4 + 6x7 = 0.901 + 0.439(1 − e(−ER+0.0003T ) ) (16)
2.2. Energy Balance in the Gasification Process

Based on the first law of thermodynamics, the equation for energy conservation of the reactants
and products of gasification can be written as follows:
T2
Q = ∑ (wout ·H0out ) + (∑ (wout ·Cp,out ( T )))dT
298.15 (17)
T1
−∑ (win ·H0in ) − (∑ (win ·Cp,in ( T )))dT
298.15
where Cp is the specific heat at constant pressure, and T1 and T2 are the inlet temperature of reactants
and the outlet temperature of products, respectively. H0 is the standard enthalpy of formation, and the
value for the gas can be found in the literature [8,32]. Q is the energy exchange between the gasifier
and the environment, which is positive for the endothermic reaction and negative for the exothermic
reaction. w is the molar amount of the corresponding material. The subscript “in” corresponds to the
reactants and the subscript “out” corresponds to the products.
The equation for the complete combustion reaction of biomass is shown below:
x y x
CHx Oy + (1 + − )O2 → CO2 + H2 O + LHVbiomass (18)
4 2 2
The standard enthalpy of formation for biomass can be calculated as follows:
x
H0f,CHx Oy = H0f,CO2 + H0f,H2 O + LHVbiomass (19)
2
where LHVbiomass is the lower heating value of the biomass in MJ/kg.
2.3. Exergy Efficiency of Gasification

Exergy is the maximum amount of theoretically available energy, which is used for evaluating the
quality of energy. The conversion of exergy before and after gasification is controlled by the second
law of thermodynamics, which considers the loss due to increase in entropy. The quality of energy
could not be considered in the first law of thermodynamics during the gasification process. Hence, it is
highly necessary to analyze the exergy efficiency.
Exergy includes two components [8]: chemical exergy and physical exergy.
The chemical exergy of the compound is calculated as follows [38]:
Echem,CHO = wC Echem,C + 0.5wH2 Echem,H2 + 0.5wO2 Echem,O2 (20)
The chemical exergies for pure components are given in the literature [39,40].
16
Energies 2017, 10, 2163
The physical exergy of a component is calculated as follows [41]:
T T
Cp
Ephy = Cp dT − T0 dT = (h − h0 ) − T0 (s − s0 ) (21)
T
T0 T0
where h and s are the enthalpy and entropy, respectively.

The exergy of biomass is calculated from its lower heating value as follows:
Ebiomass = β·LHVbiomass (22)
where β is a multiplication factor [8], which is expressed in terms of oxygen-carbon and

hydrogen-carbon ratios.
The exergy balance of the reaction process is as follows:
∑ Ereactants + W =∑ Eproducts +Eloss (23)
where the exergy of reactants, Ereactants , comprises biomass exergy and the exergy of the inlet gas
mixture. The exergy of products, Eproducts , comprises syngas exergy and unconverted carbon exergy.
Eloss is the exergy loss during the gasification process. W is the exergy accompanying the energy
exchange, which is positive for the endothermic reaction and negative for the exothermic reaction.
Exergy efficiency (η) corresponding to the endothermic reaction is defined as follows:
∑ Eproducts
η= (24)
∑ Ereactants + W
Exergy efficiency corresponding to the exothermic reaction is defined as follows:
∑ Eproducts − W
η= (25)
∑ Ereactants
2.4. Other Parameters

Steam to biomass ratio (SBR) is defined as the ratio of the steam mass flow rate to the biomass
mass flow rate on dry basis.
Gas yield (GY) is calculated by dividing the rate of production of dry syngas, excluding nitrogen,
by the biomass feeding rate, and its unit is Nm3 /kg.
When the combustible components of the syngas are CO, H2 , and CH4 , the lower heating value
(LHV) of the syngas is calculated as follows [42]:
LHV = 12.64φCO + 10.8φH2 + 35.8φCH4 (26)
where LHV is expressed in MJ/Nm3 . φ is the corresponding molar fraction of combustible gas species
in the dry syngas, which includes nitrogen.
3. Model Validation
In order to validate the equilibrium model, the experimental results reported by some researchers
have been used as comparison, and the operating conditions of the samples chosen to validate the
model are presented in Table 1. The validation results are shown through Figures 1–6.
17
Energies 2017, 10, 2163
Table 1. Operating conditions for the experimental runs of gasification.
Expt. Sample No. Author(s) Gasification Agent T (◦ C) SBR ER

1 Campoy et al. [43] Steam + Air 752 0.18 0.23
2 Campoy et al. Steam + Air 727 0.28 0.19
7 Jarungthammachote et al. [44] Steam + Air 909 0.17 0.34
8 Jarungthammachote et al. Steam + Air 903 0.21 0.34
13 Salami et al. [45] Steam + Air 770 0.23 0.29
14 Salami et al. Steam + Air 770 0.4 0.29
27 Cheng et al. [46] Steam 714 0.15 0
28 Cheng et al. Steam 788 0.15 0
29 Cheng et al. Steam 863 0.15 0
30 Cerone et al. [47] Steam + Air 702 0.233 0.19
31 Cerone et al. Steam + Oxygen 742 0.3 0.18
32 Ruoppolo et al. [48] Steam + Oxygen + Nitrogen 1 732 0.23 0.22
1 The proportion of nitrogen is not presented in the experiment, so nitrogen is not considered for simulation.
The comparisons between the experimental results and numerical results for syngas composition,
LHV, and GY are shown in Figures 1 and 2. The molar fraction of individual component of syngas is
defined as the number of moles of individual component divided by the total number of moles of dry
syngas, which includes nitrogen.
The error between experimental results and predicted results is estimated using the statistical
parameter of root mean square (RMS) error, and it is defined as follows:

N

∑ (Expi − Modi )2
i
RMS = (27)
N
where Exp and Mod are the experimental molar fractions and the predicted molar fractions of syngas
species, respectively. N is the number of data.
The RMS errors are calculated for each set of data point, and the maximum value of RMS error is
0.041 for sample No. 19, as shown in Figure 3.
18
Energies 2017, 10, 2163
Figure 1. Comparison between experimental data and predicted data for syngas composition.
Figure 2. Comparisons between experimental data and predicted data for lower heating value (LHV)
and gas yield (GY).
For the first 6 data points, the simulation values for syngas composition well fit the experimental
data. Only the molar fraction of H2 is slightly above the experimental data, so the predicted LHV
and GY are just above the experimental data, as shown in Figure 2. For the next 6 data points,
the predicted syngas composition also very well fits the experimental data. From sample No. 13 to 19,
it shows that there is a difference in composition between the model results and the experimental data,
as shown in Figures 1 and 3. However, at high gasification temperature, from sample No. 20 to 26,
the RMS error increases less obviously with increasing SBR, as shown in Figure 3. Meanwhile, it is also
observed that the predicted LHV is closer to the experimental data at high gasification temperature,
as shown in Figure 2. The effect of SBR on the accuracy of calculation depends on the gasification
temperature. When the same value of SBR is choosen, the difference between the simualtion result
and experimental result at low gasification temperature is greater than the value at high gasification
temperature. Through the present results, it is evident that the predicted molar fraction of CH4 well
fits with the experimental data at all the data set points.
19
Energies 2017, 10, 2163
Figure 3. RMS errors from comparisons between experimental data and predicted results.
Figure 4. Comparison between experimental data and predicted date for tar content.
The comparisons between the experimental results and the numerical results for tar content and
GY are shown in Figures 4 and 5. Effects of the models, with or without tar, on the simulation results
are also researched. The tar content in g/Nm3 is the weight of the tar divided by the volume of dry
syngas, which includes nitrogen. From Figure 4, the tar content can be predicted very well. Moreover,
the predicted GY based on the model with tar are much closer to the experimental data. For the model
without tar, all the tar gets converted into syngas; hence, the predicted GY is larger than the value
calculated from the model with tar.
Figure 5. Comparison between experimental data and predicted data based on different models for GY.
20
Energies 2017, 10, 2163
Figure 6. Comparison between experimental data and predicted data based on different models for
syngas composition (E-experimental data, M1-model with tar, M2-model without tar).
From Figure 6, the predicted total molar fraction of the combustible gases based on the model
without tar is larger than the predicted value based on the model with tar. From Figures 4–6,
the predicted values based on the model with tar are closer to the experimental data. This shows that
the gasification model with tar is reasonable and effective. At low gasification temperature with small
ER and SBR, the tar content is very high, as shown in Table 1 and Figure 4. In this case the simulation
result will be inaccurate if the tar is not considered in the gasification model.
Considering the effects of the process parameters on syngas composition, tar content, LHV,
and GY, the proposed equilibrium model with tar is suitable for the simulation of gasification at
relatively low gasification temperature with small ER and SBR.
4. Results and Discussion

In this study, gasification of woody biomass is studied with ER varying from 0.1 to 0.5, SBR from
0.15 to 0.6, and the gasification temperature from 900 K to 1100 K. The molecular formula for woody
biomass is CH1.4 O0.64 on the basis of an earlier reference [43]. The temperature of the inlet biomass is
298.15 K, and the temperatures of inlet air and steam are 673.15 K. The LHV of woody biomass from
the same reference work is 17.1 MJ/kg, and the biomass is completely dry in the simulation study.
4.1. The Effect of Gasification Temperature

The effects of gasification temperature on syngas composition and tar content are shown in
Figure 7. The tar content in g/kg is the weight of the tar divided by the weight of biomass. The result
from Figure 7 reveals the expected strong effect of temperature on the gasification products. The molar
fractions of CO and H2 increase significantly as gasification temperature increases, whereas the molar
fractions of CO2 and CH4 show an opposite trend, as mentioned in the literature [35,42]. Meanwhile,
with increase in temperature, the tar content decreases with the value changing from 137 g/kg at 900 K
to 0 g/kg at 1050 K. The value of 0 shows that the tar cannot be predicted based on the proposed
model in some calculation conditions, which does not agree with the fact. Hence, from this perspective,
the model has a certain application scope, which is consistent with the conclusion of previous section of
this article. As the tar content decreases and the syngas volume increases with increase in temperature,
GY increases with increase in temperature, as shown in Figure 8. Meanwhile, LHV also increases with
increase in temperature due to the formation of large molar fractions of CO and H2 .
21
Energies 2017, 10, 2163
Figure 7. Effects of gasification temperature on syngas composition and tar content at ER = 0.2 and
SBR = 0.3.
Figure 8. Effects of gasification temperature on LHV and GY at ER = 0.2 and SBR = 0.3.
At high ER, the influence of the gasification temperature on the syngas composition is shown
in Figure 9. The varying trend of syngas composition shown in Figure 9 is similar to the trend
shown in Figure 7, with increase in gasification temperature. However, the total molar fraction of the
combustible gases decreases, and this results in the small value of LHV when comparing 4.04 shown in
Figure 10 with 6.86 shown in Figure 8 at the same temperature of 1100 K. Meanwhile, tar is not formed
throughout the range of temperature, so it is advisable to decrease the tar content by increasing ER.
Figure 9. Effect of gasification temperature on syngas composition at ER = 0.4 and SBR = 0.5.
22
Energies 2017, 10, 2163
Figure 10. Effects of gasification temperature on LHV and GY at ER = 0.4 and SBR = 0.5.
4.2. The Effect of ER

At the gasification temperature of 950 K, the effects of ER on syngas composition and tar content
are shown in Figure 11. The molar fractions of the combustible gases decrease with increase in ER due
to increase in N2 content, so LHV decreases, as shown in Figure 12. With the conversion of tar and
participation of oxygen in the gasifier, GY increases with increase in ER.
Figure 11. Effects of ER on syngas composition and tar content at T = 950 K and SBR = 0.3.
Figure 12. Effects of ER on LHV and GY at T = 950 K and SBR = 0.3.
At the gasification temperature of 1100 K, the influence of ER on syngas composition is shown

in Figure 13. The percentage of CO2 is defined as the ratio of the molar quantity of CO2 to the total
molar quantity of CO, CO2 , H2 , and CH4 products. The varying trend of the syngas composition
shown in Figure 13 is similar to the trend shown in Figure 11. The percentage of the CO2 increases
sharply with increase in ER due to combustion of more carbon into CO2 , and this is consistent with
23
Energies 2017, 10, 2163
previous reports [17,34]. The molar fraction of CO2 does not differ much with increasing ER, as seen in
Figure 13, and in some studies this value increases slightly [9]. At high temperatures tar is not formed
in the entire range of ER. This shows that temperature has an important effect on tar conversion in the
gasification process. The effects of ER on LHV and GY shown in Figure 14 are similar to that shown in
Figure 12. The GY increases as ER increases, whereas LHV decreases with increase in ER.
Figure 13. Effect of ER on syngas composition at T = 1100 K and SBR = 0.5.
Figure 14. Effects of ER on LHV and GY at T = 1100 K and SBR = 0.5.
4.3. The Effect of SBR

Due to the significant influence of Reaction (7), the molar fractions of H2 and CO2 increase as SBR
increases, whereas the value of CO decreases at the same time, as shown in Figure 15. The tar content
decreases as SBR increases, which shows the same trend as that reported earlier [45]. Since the trend
for the decrease in molar fraction of CO is not evident when SBR is below 0.4, the LHV increases as
SBR increases over a maximum range of SBR, followed by a slight decrease, as shown in Figure 16.
The GY increases as SBR increases, since Reaction (7) is favored as also mentioned earlier [17].
Figure 15. Effects of SBR on syngas composition and tar content at T = 950 K and ER = 0.3.
24
Energies 2017, 10, 2163
Figure 16. Effects of SBR on LHV and GY at T = 950 K and ER = 0.3.
From Figure 17, the molar fraction of CH4 varies very less with increasing SBR, and the increase
in the molar fraction of H2 is not significant. However, at high gasification temperature the molar
fraction of CO decreases apparently as SBR increases, which is very different to the trend shown in
Figure 15. The result is consistent with earlier studies [9,43]. Meanwhile, in the formula for calculating
LHV (Equation (26)), the coefficient corresponding to the molar fraction of CO is larger than that of H2 .
Therefore, the LHV decreases as SBR increases, as shown in Figure 18 for wide range of SBR. This is
different from the trend at low temperature, as shown in Figure 16. It can also be seen that increase in
H2 is always accompanied with a decrease in CO and an increase of CO2 . Hence, the value of SBR
should not be too big, especially at high gasification temperature.
Figure 17. Effect of SBR on syngas composition at T = 1100 K and ER = 0.2.
Figure 18. Effects of SBR on LHV and GY at T = 1100 K and ER = 0.2.
25
Energies 2017, 10, 2163
4.4. The Effect of Process Parameters on Energy Exchange

In the air-steam gasification process, the dimensionless heat transfer ratio (DHTR), which can be
affected by the process parameters, is adopted to represent the value and direction of energy exchange
between the gasifier and the environment, and it is defined as follows:
Q
DHTR =
(28)

T1

∑ (win ·ΔHf,in
0 )+ (∑ (win ·Cp,in ( T )))dT
298.15
A DHTR value of zero corresponds to an auto-thermal reaction, which means there is no energy
exchange between the gasifier and the environment. The value of more than zero corresponds to an
endothermic reaction, whereas the value below zero corresponds to an exothermic reaction.
The distributions of DHTR are shown in Figures 19–22, and the air-steam biomass gasification
is not an auto-thermal reaction process for most of the process parameters that are shown in the
figures. At the same temperature and ER, the effect of SBR on DHTR is not evident, except when
the temperature is low and ER is high, as shown in Figure 19. The influence of ER on DHTR is
obvious, and the system changes from the endothermic state to the exothermic state as ER increases.
This is attributed to an increase in the released heat due to a more complete combustion reaction.
With increase in gasification temperature, the auto-thermal reaction zone moves from low ER area to
high ER area. ER corresponding to the auto-thermal reaction zone changes from about 0.15 at 900 K
temperature to about 0.3 at 1100 K gasification temperature.
Figure 19. Dimensionless heat transfer ratio (DHTR) distribution at T = 900 K.
Figure 20. DHTR distribution at T = 950 K.
26
Energies 2017, 10, 2163
Based on 450 numerical data of DHTR, the relationship between DHTR and the process
parameters, which is evaluated in terms of the correlation coefficient, is determined as follows:
DHTR = −0.904090 − 1.426036 × ER + 0.179806 × SBR + 0.001144 × T (29)
The correlation coefficient is 0.983, which is a very high value, so it clearly shows that the
formula is accurate. According to this formula, the value of energy exchange can be determined
conveniently. In some cases, the energy balance equation is introduced to the equilibrium model.
Based on Equation (29), the energy exchange can change with the changes of the process parameters,
so the predicted results obtained by solving the simultaneous equations will be more reasonable.
4.5. The Effect of Process Parameters on Exergy Efficiency

The distributions of exergy efficiency are shown in Figures 23–25. The exergy efficiency decreases
slightly as SBR increases when both gasification temperature and ER are small. From the figures,
the exergy efficiency is between 0.8 and 0.9 under all simulation conditions. The effect of ER on exergy
efficiency is obvious, and the exergy efficiency decreases as ER increases.
In exothermic state, based on Equation (25), the exergy efficiency (η) can also be expressed
as follows:
∑ Eproducts − W ∑ Eproducts −W
η= = + = ηp + ηe (30)
∑ Ereactants ∑ Ereactants ∑ Ereactants
27
Energies 2017, 10, 2163
where η p is the exergy efficiency of products, which is termed as the overall exergy efficiency in some
studies [9], and η e is the exergy efficiency of energy exchange.
Figure 23. Exergy efficiency distribution at T = 900 K.
28
Energies 2017, 10, 2163
For the exothermic reaction, the DHTR value decreases with increase in ER, as shown in
Figures 19–22. By comparing the formulas of these two parameters, DHTR and η e , the trends are just
inverse as ER increases. So, the exergy efficiency of energy exchange increases with increase in ER.
However, the total exergy efficiency decreases, as shown in Figures 23–25. So the exergy efficiency of
products decreases rapidly, and this is the reason for the obvious decrease of exergy efficiency as ER
increases in some studies [9].
5. Conclusions
In the present work, the air-steam biomass gasification model with tar has been developed based
on the equilibrium constants, and phenol is used as the model compound for the tar. The simulation
results based on two different models (with and without tar) have been validated by the experimental
data reported in the references, and the model with tar can well predict the tar content in the gasification.
Meanwhile, the predicted GY and total molar fraction of the combustible gases, based on the model
with tar, agree better with the experimental data. So, the simulation results are much less accurate if
the tar is not considered in the gasification model. Considering the effects of the process parameters
on syngas composition, tar content, LHV, and GY, the proposed equilibrium model with tar is suitable
for the simulation of the gasification process at a relatively low gasification temperature with a small
ER and SBR.
The energy exchange between the gasifier and its surroundings has been studied based on DHTR.
The influence of SBR on DHTR is not obvious, whereas the effects of ER and temperature on DHTR are
evident. Moreover, the relationship between DHTR and the process parameters is given by a formula.
The influence of process parameters on syngas composition, tar content, GY, LHV, and exergy
efficiency has also been parametrically researched. With increasing gasification temperature, the total
molar fraction of combustible gases, GY, and LHV all increase, whereas the tar content and exergy
efficiency all decrease. GY increases as ER increases, whereas tar content, total molar fraction of
combustible gases, LHV, and exergy efficiency all decrease. With increasing SBR, the total molar
fraction of combustible gases and GY increase, whereas tar content and exergy efficiency all decrease.
Author Contributions: Xinjun Zhao did the simulation and wrote the paper. Qitai Eri and Wenzhen Wu analyzed
the data.
References
1. Pereira, E.G.; da Silva, J.N.; de Oliveira, J.L.; Machado, C.S. Sustainable energy: A review of gasification
technologies. Renew. Sustain. Energy Rev. 2012, 16, 4753–4762. [CrossRef]
2. Jayathilake, R.; Rudra, S. Numerical and Experimental Investigation of Equivalence Ratio (ER) and Feedstock
Particle Size on Birchwood Gasification. Energies 2017, 10, 1232. [CrossRef]
3. Baruah, D.; Baruah, D.C. Modeling of biomass gasification: A review. Renew. Sustain. Energy Rev. 2014, 39,
806–815. [CrossRef]
4. Purohit, P. Economic potential of biomass gasification projects under clean development mechanism in India.
J. Clean. Prod. 2009, 17, 181–193. [CrossRef]
5. González-Vázquez, M.; García, R.; Pevida, C.; Rubiera, F. Optimization of a Bubbling Fluidized Bed Plant for
Low-Temperature Gasification of Biomass. Energies 2017, 10, 306. [CrossRef]
6. Ahmed, T.Y.; Ahmad, M.M.; Yusup, S.; Inayat, A.; Khan, Z. Mathematical and computational approaches
for design of biomass gasification for hydrogen production: A review. Renew. Sustain. Energy Rev. 2012, 16,
2304–2315. [CrossRef]
7. Sharma, S.; Sheth, P.N. Air–steam biomass gasification: Experiments, modeling and simulation.
Energy Convers. Manag. 2016, 110, 307–318. [CrossRef]
8. Loha, C.; Chattopadhyay, H.; Chatterjee, P.K. Thermodynamic analysis of hydrogen rich synthetic gas
generation from fluidized bed gasification of rice husk. Energy 2011, 36, 4063–4071. [CrossRef]
29
Energies 2017, 10, 2163
9. Tapasvi, D.; Kempegowda, R.S.; Tran, K.-Q.; Skreiberg, Ø.; Grønli, M. A simulation study on the torrefied
biomass gasification. Energy Convers. Manag. 2015, 90, 446–457. [CrossRef]
10. Jangsawang, W.; Klimanek, A.; Gupta, A.K. Enhanced yield of hydrogen from wastes using high temperature
steam gasification. J. Energy Resour. Technol. 2006, 128, 179–185. [CrossRef]
11. Rodríguez-Olalde, N.; Mendoza-Chávez, E.; Castro-Montoya, A.; Saucedo-Luna, J.; Maya-Yescas, R.;
Rutiaga-Quiñones, J.; Ortega, J. Simulation of Syngas Production from Lignin Using Guaiacol as a Model
Compound. Energies 2015, 8, 6705–6714. [CrossRef]
12. Suwatthikul, A.; Limprachaya, S.; Kittisupakorn, P.; Mujtaba, I. Simulation of Steam Gasification in a
Fluidized Bed Reactor with Energy Self-Sufficient Condition. Energies 2017, 10, 314. [CrossRef]
13. He, J.; Göransson, K.; Söderlind, U.; Zhang, W. Simulation of biomass gasification in a dual fluidized bed
gasifier. Biomass Convers. Biorefin. 2012, 2, 1–10. [CrossRef]
14. Rofouie, P.; Moshkelani, M.; Perrier, M.; Paris, J. A modified thermodynamic equilibrium model for woody
biomass gasification. Can. J. Chem. Eng. 2014, 92, 593–602. [CrossRef]
15. Meng, W.X.; Banerjee, S.; Zhang, X.; Agarwal, R.K. Process simulation of multi-stage chemical-looping
combustion using Aspen Plus. Energy 2015, 90, 1869–1877. [CrossRef]
16. Chutichai, B.; Patcharavorachot, Y.; Assabumrungrat, S.; Arpornwichanop, A. Parametric analysis of a
circulating fluidized bed biomass gasifier for hydrogen production. Energy 2015, 82, 406–413. [CrossRef]
17. Lim, Y.-i.; Lee, U.-D. Quasi-equilibrium thermodynamic model with empirical equations for air–steam
biomass gasification in fluidized-beds. Fuel Process. Technol. 2014, 128, 199–210. [CrossRef]
18. Karmakar, M.K.; Datta, A.B. Generation of hydrogen rich gas through fluidized bed gasification of biomass.
Bioresour. Technol. 2011, 102, 1907–1913. [CrossRef] [PubMed]
19. Mendiburu, A.Z.; Carvalho, J.A.; Coronado, C.J.R. Thermochemical equilibrium modeling of biomass
downdraft gasifier: Stoichiometric models. Energy 2014, 66, 189–201. [CrossRef]
20. Jarungthammachote, S.; Dutta, A. Thermodynamic equilibrium model and second law analysis of a
downdraft waste gasifier. Energy 2007, 32, 1660–1669. [CrossRef]
21. Kim, Y.D.; Yang, C.W.; Kim, B.J.; Kim, K.S.; Lee, J.W.; Moon, J.H.; Yang, W.; Yu, T.U.; Lee, U.D. Air-blown
gasification of woody biomass in a bubbling fluidized bed gasifier. Appl. Energy 2013, 112, 414–420.
[CrossRef]
22. Sahoo, A.; Ram, D.K. Gasifier performance and energy analysis for fluidized bed gasification of sugarcane
bagasse. Energy 2015, 90, 1420–1425. [CrossRef]
23. Umeki, K.; Yamamoto, K.; Namioka, T.; Yoshikawa, K. High temperature steam-only gasification of woody
biomass. Appl. Energy 2010, 87, 791–798. [CrossRef]
24. Fremaux, S.; Beheshti, S.-M.; Ghassemi, H.; Shahsavan-Markadeh, R. An experimental study on
hydrogen-rich gas production via steam gasification of biomass in a research-scale fluidized bed.
25. Gil-Lalaguna, N.; Sánchez, J.L.; Murillo, M.B.; Rodríguez, E.; Gea, G. Air–steam gasification of sewage sludge
in a fluidized bed. Influence of some operating conditions. Chem. Eng. J. 2014, 248, 373–382. [CrossRef]
26. Ruoppolo, G.; Ammendola, P.; Chirone, R.; Miccio, F. H2 -rich syngas production by fluidized bed gasification
of biomass and plastic fuel. Waste Manag. 2012, 32, 724–732. [CrossRef] [PubMed]
27. Huang, H.J.; Ramaswamy, S. Modeling biomass gasification using thermodynamic equilibrium approach.
Appl. Biochem. Biotechnol. 2009, 154, 14–25. [CrossRef] [PubMed]
28. Mayerhofer, M.; Fendt, S.; Spliethoff, H.; Gaderer, M. Fluidized bed gasification of biomass—In bed
investigation of gas and tar formation. Fuel 2014, 117, 1248–1255. [CrossRef]
29. Martínez-Lera, S.; Pallarés Ranz, J. On the development of a wood gasification modelling approach with
special emphasis on primary devolatilization and tar formation and destruction phenomena. Energy 2016,
113, 643–652. [CrossRef]
30. Morf, P.; Hasler, P.; Nussbaumer, T. Mechanisms and kinetics of homogeneous secondary reactions of tar
from continuous pyrolysis of wood chips. Fuel 2002, 81, 843–853. [CrossRef]
31. Ji, P.; Feng, W.; Chen, B. Production of ultrapure hydrogen from biomass gasification with air. Chem. Eng. Sci.
2009, 64, 582–592. [CrossRef]
32. Zainal, Z.A.; Ali, R.; Lean, C.H.; Seetharamu, K.N. Prediction of performance of a downdraft gasifier
using equilibrium modeling for different biomass materials. Energy Convers. Manag. 2001, 42, 1499–1515.
[CrossRef]
30
Energies 2017, 10, 2163
33. Gómez-Barea, A.; Leckner, B. Estimation of gas composition and char conversion in a fluidized bed biomass
gasifier. Fuel 2013, 107, 419–431. [CrossRef]
34. Li, X.T.; Grace, J.R.; Lim, C.J.; Watkinson, A.P.; Chen, H.P.; Kim, J.R. Biomass gasification in a circulating
fluidized bed. Biomass Bioenergy 2004, 26, 171–193. [CrossRef]
35. Meng, X.; de Jong, W.; Fu, N.; Verkooijen, A.H. M. Biomass gasification in a 100 kWth steam-oxygen blown
circulating fluidized bed gasifier: Effects of operational conditions on product gas distribution and tar
formation. Biomass Bioenergy 2011, 35, 2910–2924. [CrossRef]
36. Lv, P.M.; Xiong, Z.H.; Chang, J.; Wu, C.Z.; Chen, Y.; Zhu, J.X. An experimental study on biomass air-steam
gasification in a fluidized bed. Bioresour. Technol. 2004, 95, 95–101. [CrossRef] [PubMed]
37. Li, X.; Grace, J.R.; Watkinson, A.P.; Lim, C.J.; Ergudenler, A. Equilibrium modeling of gasification: A free
energy minimization approach and its application to a circulating fluidized bed coal gasifier. Fuel 2001, 80,
38. Srinivas, T.; Gupta, A.V.S.S.K.S.; Reddy, B.V. Thermodynamic equilibrium model and exergy analysis of a
biomass gasifier. J. Energy Resour. Technol. 2009, 131, 98–105. [CrossRef]
39. Demirel, Y. Thermodynamic analysis. Arab. J. Sci. Eng. 2013, 38, 221–249. [CrossRef]
40. Pellegrini, L.; Deoliveirajr, S. Exergy analysis of sugarcane bagasse gasification. Energy 2007, 32, 314–327.
[CrossRef]
41. Datta, A.; Ganguly, R.; Sarkar, L. Energy and exergy analyses of an externally fired gas turbine (EFGT) cycle
integrated with biomass gasifier for distributed power generation. Energy 2010, 35, 341–350. [CrossRef]
42. Wang, R.; Huang, Q.; Lu, P.; Li, W.; Wang, S.; Chi, Y.; Yan, J. Experimental study on air/steam gasification
of leather scraps using U-type catalytic gasification for producing hydrogen-enriched syngas. Int. J.
Hydrogen Energy 2015, 40, 8322–8329. [CrossRef]
43. Campoy, M.; Gómez-Barea, A.; Vidal, F.B.; Ollero, P. Air-steam gasification of biomass in a fluidized bed
under simulated autothermal and adiabatic conditions. Ind. Eng. Chem. Res. 2008, 47, 5957–5965. [CrossRef]
44. Jarungthammachote, S.; Dutta, A. Experimental investigation of a multi-stage air-steam gasification process
for hydrogen enriched gas production. Int. J. Energy Res. 2012, 36, 335–345. [CrossRef]
45. Salami, N.; Skála, Z. Use of the steam as gasifying agent in fluidized bed gasifier. Chem. Biochem. Eng. Q. J.
2015, 29, 13–18. [CrossRef]
46. Cheng, G.; Li, Q.; Qi, F.; Xiao, B.; Liu, S.; Hu, Z.; He, P. Allothermal gasification of biomass using micron size
biomass as external heat source. Bioresour. Technol. 2012, 107, 471–475. [CrossRef] [PubMed]
47. Cerone, N.; Zimbardi, F.; Villone, A.; Strjiugas, N.; Kiyikci, E.G. Gasification of wood and torrefied wood
with air, oxygen, and steam in a fixed-bed pilot plant. Energy Fuels 2016, 30, 4034–4043. [CrossRef]
48. Ruoppolo, G.; Miccio, F.; Brachi, P.; Picarelli, A.; Chirone, R. Fluidized bed gasification of biomass and
biomass/coal pellets in oxygen and steam atmosphere. Chem. Eng. Trans. 2013, 32, 595–600.
31
energies
Article
Treatment of Oily Wastewater by the Optimization
of Fe2O3 Calcination Temperatures in Innovative
Bio-Electron-Fenton Microbial Fuel Cells
Jung-Chen Wu 1 , Wei-Mon Yan 2 , Chin-Tsan Wang 3, *, Chen-Hao Wang 1, *, Yi-Hao Pai 4 ,
Kai-Chin Wang 1 , Yan-Ming Chen 5 , Tzu-Hsuan Lan 5 and Sangeetha Thangavel 2
1 Department of Materials Science and Engineering, National Taiwan University of Science and Technology,
No. 43, Keelung Rd., Sec. 4, Da’an Dist., Taipei 10607, Taiwan; [email protected] (J.-C.W.);
[email protected] (K.-C.W.)
2 Department of Energy and Refrigerating Air-Conditioning Engineering, National Taipei University of Technology,
1, Sec. 3, Zhongxiao E. Rd., Taipei 10608, Taiwan;
[email protected] (W.-M.Y.); [email protected] (S.T.)
3 Department of Mechanical and Electro-Mechanical Engineering, National Ilan University, No. 1, Sec. 1,
Shennong Rd., Yilan 26047, Taiwan
4 Department of Opto-Electronic Engineering, National Dong Hwa University, No. 1, Sec. 2, Da Hsueh Rd.,
Shoufeng, Hualien 97401, Taiwan; [email protected]
5 Department of Materials and Mineral Resources Engineering & Institute of Materials Science and Engineering
National Taipei University of Technology, No. 1, Sec. 3, Zhong-Xiao E. Rd., Taipei 10608, Taiwan;
[email protected] (Y.-M.C.); [email protected] (T.-H.L.)
* Correspondence: [email protected] (C.-T.W.); [email protected] (C.-H.W.); Tel.: +886-3-935-7400
(ext. 7459) (C.-T.W.); +886-2-2730-3715 (C.-H.W.); Fax: +886-3-931-1326 (C.-T.W.); +886-2-2737-6544 (C.-H.W.)
Received: 20 January 2018; Accepted: 2 March 2018; Published: 6 March 2018
Abstract: Due to the fact that Iron oxide (Fe2 O3 ) is known to have a good effect on the photochemical
reaction of catalysts, an investigation in this study into the enhancement of the degradation
performance of bio-electro-Fenton microbial fuel cells (Bio-E-Fenton MFCs) was carried out using
three photocatalytic cathodes. These cathodes were produced at different calcination temperatures
of Fe2 O3 ranging from 500 ◦ C to 900 ◦ C for realizing their performance as photo catalysts within
the cathodic chamber of an MFC, and they were compared for their ability to degrade oily wastewater.
Results show that a suitable temperature for the calcination of iron oxide would have a significantly
positive effect on the performance of Bio-E-Fenton MFCs. An optimal calcination temperature of
500 ◦ C for Fe2 O3 in the electrode material of the cathode was observed to produce a maximum
power density of 52.5 mW/m2 and a chemical oxygen demand (COD) degradation rate of oily
wastewater (catholyte) of 99.3% within one hour of operation. These novel findings will be useful
for the improvement of the performance and applications of Bio-E-Fenton MFCs and their future
applications in the field of wastewater treatment.
Keywords: bio-electro-Fenton microbial fuel cells (Bio-E-Fenton MFCs); wastewater; photo catalyst;
degradation; calcination; chemical oxygen demand (COD)
1. Introduction
The Bio-Electro-Fenton Microbial Fuel Cells (Bio-E-Fenton MFCs) is a new framework and has been
operated extensively because of its simultaneous wastewater treatment and power generation capability.
Significantly, the electro-Fenton reaction consists of iron ions (Fe2+ ) and hydrogen peroxide, and has
become a useful technology for treating organic pollutants in wastewater [1–10] as hydroxyl radicals (•OH)
are generated in abundance within the electro-Fenton reaction [11]. Considering the working principle of

Energies 2018, 11, 565
Bio-E-Fenton MFCs, electrons and protons are produced at the anode as a result of the microbial oxidation
of substrates. Then, electrons move towards the cathode through an external circuit pathway, whereas
the protons directly penetrate through a membrane and arrive at the cathode simultaneously. Finally,
a complete reaction of redox is attained with the aid of oxygen in the cathodic reduction reaction of
Bio-E-Fenton MFCs [12]. Although the electro-Fenton reaction is very feasible and effective in treating
organic pollutants in wastewater, the COD removal is incomplete [13,14]. A new electro-Fenton system
with an Fe@Fe2 O3 /CNT oxygen-fed gas diffusion cathode was able to degrade rhodamine B (RhB) reaching
91.5% within 120 min at neutral pH [11]. The anode was a Pt sheet with a surface area of 2.0 cm2 and it also
served as a pseudo reference electrode. Cyclic voltammetry experiments were performed and a couple
of well-defined peaks were observed at ∼0.62 V, which were ascribed to the overall redox potentials
of the Fe3+ /Fe2+ couple. However, 74.1% and 60.1% of RhB was degraded on the Fe@Fe2 O3 /ACF
and Fe@Fe2 O3 /graphite cathodes, respectively. Polyaniline (PANI)-Fe2 O3 nanocomposite was used to
successfully enhance the dye adsorption efficiency of acid violet 19 dye, as the dye concentration was
increased from 20–80 mg/lt, and the percentage removal of dye decreased from 98.5–90% for the adsorbent
at 10 gm/lt [15]. For another kind of photo catalytic material like MoS2 , Liu et al. [16] have indicated that
a system with MoS2 -RGO composites has a better photo catalytic performance than using pure MoS2 .
Here, the MoS2 -RGO composite with 0.5 wt % RGO achieved the highest methylene blue (MB) degradation
rate of 99% within 60 min [16]. Nan Xu et al. [17] reported that the anode was carbon felt assisted with
granular graphite to enhance the anodic power density. Anaerobic sludge collected from the Mangrove
Wetland in Shenzhen was used as a biocatalyst in the anodic chamber. The system with E2 (17β-estrodial)
in the catholyte enhanced the voltage output to 0.29 V and the E2 removal efficiency leveled off
within 6 h, reaching 85.5%, 96.4%, 88.6%, and 75.6% with an external load of 50, 150, 500, and 1000 Ω,
respectively, while the corresponding power density values were 3.47, 4.35, 3.01, and 2.41 W/m3 [17].
Nevertheless, facing the issue of the weak performance of MFCs, many kinds of electrode
materials such as carbon nanotube/polyaniline carbon paper, CNT/PANI carbon paper, have been
developed [18]. Results showed that a lower ohm loss and a better power performance with
a maximum power density of 1574 mW/m2 were executed by using the CNT/PANI electrode.
A voltage of 1.18 V and current of 12.8 mA in MFCs were achieved by using CNT/PANI carbon
paper [18]. But some research works regarding the development of non-precious metal catalysts
can be implemented for an effective oxygen reduction reaction (ORR) and the enhancement of
MFC performance. Vitamin B12 (py-B12/C) pyrolyzed and supported by carbon black can be
utilized as the cathode catalyst in the solid phase MFCs [19]. Therefore, an innovative endeavor
of employing calcination iron oxide (Fe2 O3 ) as a photo catalyst combined with the Bio-E-Fenton
system, while simultaneously generating power, has not been previously reported for the treatment
of organic wastewater. Fenton’s reagent could function well under acidic conditions and during
the anodic oxidation and cathode aeration resulted in a pH increase, inhibiting the generation of •OH
and decreasing the COD removal rate [20–22]. In the experiment, 10% diluted sulfuric acid was added
to the cathode solution to obtain a pH value of 3.0 [23]. The experimental mechanism of Fe2 O3 /carbon
felt (CF) photo catalyst bio-electro-Fenton degradation of oily wastewater is illustrated in Figure 1.
The first step 1 in the cathode reaction: 2H+ + 2e− + O2 → H2 O2 (1–1), resulted in H2 O2 accumulation
in the cathode chamber via the two-electron reduction of dissolved O2 in the Bio-E-Fenton MFCs and
then FeSO4 powder was also added as a source of Fe2+ : 2H+ + Fe → Fe2+ + H2 (1–2); the second step
2 Fe2 O3 photo catalyst was added in the cathode chamber and a 300 W halogen lamp was used for
the photo excitation of Fe2 O3 in order to produce more hydroxyl radicals; the third step 3 finally
showed the bio-electro-Fenton reaction at the cathode, where the hydroxyl radicals played a vital role
in the degradation of products according to the reaction: Fe2+ + H2 O2 → Fe3+ + •OH + OH− (1–3).
This was followed by the reduction of Fe3+ to Fe2+ . This was a chain reaction where Fe3+ got reduced
to Fe2+ and then Fe2+ was oxidized into Fe3+ . This was a completely physical reaction which took
place at the cathode and microbes had no role to play. In the conclusion, the Bio-E-Fenton MFCs was
reported to be an effective system to treat organic wastewater, as well as bio-refractory pollutants [13].
33
Energies 2018, 11, 565
Figure 1. Proposed mechanism of Fe2 O3 /CF catalytic Bio-E-Fenton degradation of oily wastewater.
This research study has been carried out for gaining an in-depth understanding of the effects
of calcination temperatures (500 ◦ C to 900 ◦ C) on the performance of Bio-E-Fenton MFCs. Based on
the calcination temperature of Fe2 O3 , the photo catalyst structure will change and can be applied in
this study for investigating the performance of MFCs. The results showed that the intensity of Fe2 O3
will not vary greatly due to high calcination temperatures, but the crystalline particle size and surface
morphology will be changed immensely [24]. This feature of Fe2 O3 will enhance the COD removal
efficiency of the MFC significantly [25–28]. Fe2 O3 /carbon felt (CF) has been selected to act as part of
the cathode in Bio-E-Fenton MFCs because the photo catalysis of Fe2 O3 can also produce hydroxyl
radicals [11]. It can further combine with FeSO4 to act as an iron source with H2 O2 for producing more
hydroxyl radicals (•OH), resulting in a higher degradation rate within a short time. In addition, a pH
of 3 was maintained at the cathode chamber to treat refractory organic wastewater and this condition
will not affect the pH of the anode chamber.
2. Experimental Design
2.1. Reactor Construction and Operation

An acrylic dual chambered tank with dimensions of 85 mm × 70 mm × 55 mm was selected
for the construction of the Bio-E-Fenton MFC, as shown in Figure 2. Each chamber had a total
volume of 200 mL with a proton exchange membrane Nafion-117 (80 mm × 70 mm) and carbon felt
(CF, 60 mm × 60 mm × 5 mm) as the anode and cathode electrodes, respectively. The preparation method
for the carbon felt is as follows: It was initially washed in hot H2 O2 (10%, 90 ◦ C) solution for 3 h to
develop local Quinone sites on the carbon surface for improvement of the anode biocompatibility and
increased quantity of anthraquinone [29]. In order to prevent the thermal heat transfer from the cathode
to the anode, a Bakelite plate with dimensions of 100 mm × 50 mm × 20 mm (Figure 2) was employed
in this study. Therefore, the power density would be influenced by the effect of the calcination temperature
of Fe2 O3 in the cathode chamber, but not by the temperature of the anode. In addition, a control reactor
without Fe2 O3 was not constructed and it was considered from the similar work of [30].
34
Energies 2018, 11, 565
Figure 2. The experimental set-up of the proposed reactor.
2.2. Anolyte/Catholyte and Fe2 O3 Catalyst Preparation

Kim et al. [31] proved that lactic acid wastewater can be employed to sustain the production
of electrical energy in MFCs. Therefore, dairy wastewater was used in this study as the anolyte.
During the early stage of fermentation of the dairy wastewater in the chamber, precipitate particulates
were formed. Three layers were naturally formed when the wastewater was kept stagnant and were
classified as top-level supernatant (clear fluid), a mid-level interface, and a bottom-level precipitate.
Prior to the study, a conductivity experiment confirmed that the top-level supernatant in the chamber
had the best conductivity of all, with an open-circuit voltage of 0.70 V, limiting current of 0.547 mA/m2 ,
and an achievable maximum power density of 101.4 mW/m2 . In addition, considering the fact that
viscosity would influence microbial activity and also affect the power generation of MFCs [32], samples
from the top-level supernatant were selected for further studies.
The water bodies [33] have long been contaminated by oil pollution [34–36], so artificially prepared
oily wastewater was used as the catholyte in this study. A total of 1 mL of diesel was added to 1 L
of water and heated in a magnetic heater and blended with a blender for 1 day at 50 ◦ C. As diesel
was immiscible in water, 10 g of emulsifying (Tween 80) agent was added to aid the dissolution of
the diesel in water. Molognoni et al. [37] assessed the bio-electrochemical treatability of industrial
(dairy) wastewater by MFCs, and an MFC was built and continuously operated for 72 days, during
which time the anodic chamber was fed with dairy wastewater and the cathodic chamber with
an aerated mineral solution. The study demonstrated that industrial effluents from agrifood facilities
can be treated by bio-electrochemical systems (BESs) with >85% (average) organic matter removal.
The study by Marashi and Kariminia [38] proved that the wastewater concentration had an influence on
the MFC performance and using the raw wastewater with the concentration of 8000 mg COD L−1 resulted
in the highest power density (65.6 mW m−2 ) production. Kim et al. [39] reported that the MFC–Anaerobic
fluidized membrane bioreactors (AFMBRs) achieved 89 ± 3% removal of the COD with an effluent of
36 ± 6 mg-COD/L over 112 days of operation.
A total of 270 mg of FeCl3 (98%, Acros, Taipei, Taiwan) was dissolved in 30 mL of deionized water,
and then 1 M of NaOH (98%, Fisher, Hampton, VA, USA) 1 mL and 0.5 M oxalic acid (98%, Acros) 750 μL
were added to the Teflon tank, and the microwave was set up at 160 ◦ C for 30 min. After that, the Fe2 O3
sample solution was filtered through a membrane filter with a pore size of 0.2 μm (Advantec, Toyo, Japan)
and mixed with the 80 ◦ C deionized water and Fe2 O3 powder (by freeze-drying at −50 ◦ C) under
a vacuum for 5 h. Calcination of Fe2 O3 was achieved by loading it into a fused aluminum oxide boat and
the temperatures were set to 500 ◦ C, 700 ◦ C, and 900 ◦ C for 30 min per/sample in a high temperature
furnace. Following the calcination, the furnace was cooled to room temperature by natural convection.
35
Energies 2018, 11, 565
2.3. Experimental Analysis

Electrochemical analysis was performed by the workstation (Jiehan ECW-5600, Taipei, Taiwan) to
measure the polarization performance of the Bio-E-Fenton MFCs. For COD analysis, an instrument
(SUNTEX-V2000 photometer, Taipei, Taiwan) was utilized with solutions diluted till 100× with deionized
water. The temperature for the calcination of Fe2 O3 (500 ◦ C–900 ◦ C) was applied using high temperature
furnaces (Riki JH-2, Taipei, Taiwan). The pH was measured in the anode/cathode chambers by using
a pH meter (SUNTEX SP-2300, Taipei, Taiwan) and dissolved oxygen (DO) by using the DO analyzer
(CLEAN DO-200, Taipei, Taiwan). For material analysis, instruments like an X-ray diffractometer
(Bruker D2 Phaser), field emission scanning electron microscope (FE-SEM) (JSM-6500F), and Field
Emission Gun Transmission Electron Microscopy, FEG-TEM (FEI Tecnai™ G2 F-20 S-TWIN), were used.
3. Results and Discussions
3.1. Performance of the Fe2 O3 -C500 ◦ C/CF in Bio-E-Fenton MFCs Power Generation
Polarization curves for the Bio-E-Fenton MFCs using Fe2 O3 calcination from 500 ◦ C–900 ◦ C
are portrayed in Figure 3. The conditions adopted for the polarization test were listed as follows;
two-electrode measurement, anode electrode connected (by copper wire) to the working electrode
(WE); reference electrode 2 (RE2) was in the dairy wastewater solution; the cathode electrode was
connected (by copper wire) to the counter electrode (CE) and reference electrode 1 (RE1) in oily
wastewater solution. Here, the carbon felt (60 mm × 60 mm × 5 mm) was selected as the electrode
and can be used to calculate the power density (power/working area of electrode) and current density
(current/working area of electrode). It was clear that Fe2 O3 -500 ◦ C /CF produced an open circuit
voltage of 0.55 V, a maximum current density of 349 mA/m2 , and a maximum power density of
52.5 mW/m2 , which was 2.6 fold higher compared to Fe2 O3 -900 ◦ C. A comparison between previous
studies of Bio-E-Fenton MFCs and this study has been made and the results are shown in Table 1.
In addition, an original report related to TiO2 has been addressed for showing its optoelectronic ability
to generate hydrogen in water [40]. Table 2 shows the list of related cases of photo catalysts.
Figure 3. Polarization and power density curves of BEF-MFC at different calcination temperatures of Fe2 O3 .
36
Energies 2018, 11, 565
Table 1. Research studies using Bio-E-Fenton systems.
Power Density (mW/m2 ) COD Removal (%/h) Ref.

430 78/1080 [13]
625 88/48 [14]
52.5 99.3/1 This study
Table 2. Different kinds of photo catalyst systems employed.
Material Wavelength (nm) COD Removal (%/min) pH Ref.

Fe@Fe2 O3 /CNT 555 91.5/120 8 [11]
PANI-Fe2 O3 545 98.5/90 8 [15]
MoS2 -RGO 450 99.0/60 7 [16]
Fe2 O3 -C500 ◦ C/CF 410 99.3/60 3 This study
In consideration of the COD degradation rate, the Fe2O3-500 ◦ C/CF performed better with 99.3%
removal in 1 h and with an effluent concentration of 152 ± 5 mg-COD/mL/1 day. Generally speaking,
a higher COD degradation rate was observed in this study compared to the other studies involving general
bio-electro-Fenton systems [41] for the treatment of oily wastewater, with an efficiency that was 1.4 times
higher (Table 3). The pH and DO at 700 ◦ C were lower than those at 500 ◦ C and 900 ◦ C because under
the constant resistance (1 KΩ) discharge at 700 ◦ C, the DO and pH values affected the voltage output
recorded from the long-term measurement. Nevertheless, the results of Table 3 indicated that a better power
performance of the system and COD removal rate in the cathode chamber were observed at a calcination
temperature of Fe2O3-500 ◦ C, as it had a suitable pH and DO for befitting the biocompatibility [42–44] and
electrical conductivity [45,46] in the system. The results showed that the maximum power density was
52.5 mW/m2 and it was influenced by the calcination temperature of Fe2O3, which obviously showed that
the power density decreased with increasing calcination temperature. In addition, the average temperature
of the anode chamber was controlled at about 35~45 ◦ C (by Bakelite plate) and the cathode temperature was
higher than 70 ◦ C (Shown in Table 3), which would be the effect of evaporation. Therefore, the measurement
time was one hour and the time interval was based on the experimental reliability and neglecting the effect
of evaporation in the cathode chamber.
Table 3. Performance of system @ calcination conditions of Fe2 O3 .
Parameter 500 ◦ C 700 ◦ C 900 ◦ C BF [30]

COD removal (%) @cathode 99.3 83.2 75.2 74.0
DO (ppm) @cathode 4.13 2.82 3.59 2.44
pHaverage 2.1 1.8 2.4 3.0
PDmax (mW/m2 ) 52.5 28.1 20.1 26.0
Temp. average /◦ C anode 34.6 43.7 42.7 22.6
Temp. average /◦ C cathode 71.6 74.9 72.7 24.1
BF: Bio-electro-Fenton system [30].
To further understand the impact of Fe2 O3 -C500-/CF on the degradation rate, the following three
reasons were formulated: (1) the in situ reaction between Fe2+ leached from Fe2 O3 /FeSO4 and H2 O2
generated from the CF was kinetically more favorable than the reaction between H2 O2 from the CF
and Fe2+ in the bulk solution; (2) the calcination of Fe2 O3 resulted in the formation of different crystal
structures that affected the COD removal efficiency [47]; and (3) the photo catalyst accelerated chemical
reactions [48] so that the strong oxidation of the hydroxyl radicals could quickly break the covalent
bonds of diesel and enhance the degradation of wastewater [49].
37
Energies 2018, 11, 565
3.2. Characterizations of the Calcinated Fe2 O3

Figure 4 displays XRD patterns of the calcinated Fe2 O3 . The characteristic peaks of 500 ◦ C Fe2 O3
were at 24.2◦ , 33.3◦ , 35.6◦ , 40.9◦ , 49.6◦ , 54.2◦ , 62.6◦ , and 64.2◦ , corresponding to facet indexes of (012),
ć
(104), (110), (113), (024), (116), (214), and (300), respectively. The diffraction peaks at a 2θ value of 35.6◦
were ascribed to the (110) reflection of Fe2 O3 [11]. Fe2 O3 crystallographic structures can be controlled
by varying their calcination temperatures. At a high calcination temperature, the intensity did not vary
greatly, but the crystalline particle size and surface morphology changed immensely [24].
ć
900ć
700ć
ć
Intensity (a.u)
500ć
PDF 89-8104 Fe2O3
20 30 40 50 60 70 80
2 theta (degree)
Figure 4. XRD patterns of Fe2 O3 at different temperatures.
3.3. The Morphologies of the Fe2 O3 at Different Calcination Temperatures

For analyzing the surface morphology of Fe2 O3 at different temperatures, field emission scanning
electron microscopy was used (Figure 5a–c). The Fe2 O3 had an average particle size of 581 nm at 500 ◦ C,
984 nm at 700 ◦ C, and 1255 nm at 900 ◦ C. Fe2 O3 at 500 ◦ C showed a layered stacked morphology and
the particles were uniformly distributed with a large surface area. This showed that Fe2 O3 -C at 500 ◦ C
was better in terms of surface modification [50–52], which resulted in a maximum power density of
52.5 mW/m2 . In addition, the surface modification of Fe2 O3 was indeed the main reason behind its
best performance compared to the other systems.
Figure 5. Cont.
38
Energies 2018, 11, 565
Figure 5. FE-SEM images of Fe2 O3 at different calcination temperatures (a) 500 ◦ C (b) 700 ◦ C (c) 900 ◦ C.
The TEM images depicted that highly-agglomerated Fe2 O3 with a uniform size of ~28 nm in
diameter affected the material structure of Fe2 O3 after the calcination processes, as shown in Figure 6a, b.
It should be noted that the structures after calcination at 500 ◦ C to 700 ◦ C led to a greater particle size
and the lattice plane spacing of 0.21 to 0.26 nm. As shown in Figure 6c, calcination at 900 ◦ C resulted in
sheet-like structures and lattice plane spacing of 0.27 nm. The selected area diffraction pattern (SADP)
showed excellent crystalline structure formations at 500 ◦ C, which improved the COD degradation
efficiency [25–28]. In addition, Fe2 O3 -C500 ◦ C/CF had a better degradation rate of 99.3% in 1 h. This was
a novel attempt where the Fe2 O3 -C500 ◦ C/CF was combined with the bio-electro-Fenton reagent in MFCs.
Figure 6. TEM images of Fe2 O3 at different calcination temperatures (a) 500 ◦ C (b) 700 ◦ C (c) 900 ◦ C.
4. Conclusions
In this study, the photo catalytic ability of iron oxide (Fe2 O3 ) was investigated at different
calcination temperatures ranging from 500 ◦ C to 900 ◦ C in a Bio-E-Fenton MFC for evaluating its
effect on the degradation of dairy and oily wastewaters and power generation. A series of studies
were executed and useful findings were addressed as follows: For the study of Fe2 O3 in Bio-E-Fenton
MFCs, firstly, the calcination at 500 ◦ C brought about a better output voltage than in the case of
the calcination at 900 ◦ C, and its power density was also 2.6 times higher than at 900 ◦ C. In addition,
the COD degradation efficiency showed that Fe2 O3 -C at 500 ◦ C/CF had a better performance of
99.3% removal in 1 h and with an effluent of (152 ± 5 mg-COD/mL/1 day) operation, which was
1.3 times higher than the Fe2 O3 -C at 900 ◦ C/CF. Further, these evidences proved that calcination
temperatures of 500 ◦ C with 52.5 mW/m2 would be optimal for Fe2 O3 in Bio-E-Fenton MFCs because
of its morphology with a uniform particle size and a larger surface area. In contrast, the degradation
performance of 80.1% in 1 h was obtained by combining Fe2 O3 -C500 ◦ C/CF with a bio-electro-Fenton
reagent under non-illuminated conditions. Last but not the least, this innovative process was very
promising for the treatment of organic wastewater treatment and the improved performance of related
bio-electrochemical systems in the future.
39
Energies 2018, 11, 565
Acknowledgments: The authors would like to acknowledge the generous funding support from MOST Taiwan under
contract #MOST-104-2622-E-197-003-CC3, MOST-103-2221-E-197-022-MY3 and MOST-104-2628-E-011-003-MY3.
Author Contributions: Jung-Chen Wu: Operation of experiments, analysis and writing of article; Wei-Mon Yan:
Support of experiments and integrated system; Chin-Tsan Wang: Guidance of experiments and modification of
manuscript and reply of comments; Chen-Hao Wang: Guidance of chemical experiments; Yi-Hao Pai: Support and
assistance of photo-catalyst; Kai-Chin Wang: Assistance of analysis in chemical part; Yan-Ming Chen: Assistance
of analysis in material part; Tzu-Hsuan Lan: Assistance of analysis in material part; Sangeetha Thangavel: English
modification of manuscript.
References
1. Khoufi, S.; Aloui, F.; Sayadi, S. Treatment of olive oil mill wastewater by combined process electro-Fenton
reaction and anaerobic digestion. Water Res. 2006, 40, 2007–2016. [CrossRef] [PubMed]
2. Brillas, E.; Casado, J. Aniline degradation by Electro-Fenton® and peroxi-coagulation processes using a flow
reactor for wastewater treatment. Chemosphere 2002, 47, 241–248. [CrossRef]
3. Liu, H.; Wang, C.; Li, X.; Xuan, X.; Jiang, C.; Cui, H.N. A novel electro-Fenton process for water treatment:
Reaction-controlled pH adjustment and performance assessment. Environ. Sci. Technol. 2007, 41, 2937–2942.
[CrossRef] [PubMed]
4. Panizza, M.; Oturan, M.A. Degradation of Alizarin Red by electro-Fenton process using a graphite-felt cathode.
Electrochim. Acta 2011, 56, 7084–7087. [CrossRef]
5. Wang, A.; Qu, J.; Ru, J.; Liu, H.; Ge, J. Mineralization of an azo dye Acid Red 14 by electro-Fenton’s reagent
using an activated carbon fiber cathode. Dyes Pigment. 2005, 65, 227–233. [CrossRef]
6. Luo, M.; Yuan, S.; Tong, M.; Liao, P.; Xie, W.; Xu, X. An integrated catalyst of Pd supported on magnetic
Fe3 O4 nanoparticles: Simultaneous production of H2 O2 and Fe2+ for efficient electro-Fenton degradation of
organic contaminants. Water Res. 2014, 48, 190–199. [CrossRef] [PubMed]
7. Nidheesh, P.; Gandhimathi, R. Trends in electro-Fenton process for water and wastewater treatment: An overview.
Desalination 2012, 299, 1–15. [CrossRef]
8. Nidheesh, P.; Gandhimathi, R.; Velmathi, S.; Sanjini, N. Magnetite as a heterogeneous electro Fenton catalyst
for the removal of Rhodamine B from aqueous solution. RSC Adv. 2014, 4, 5698–5708. [CrossRef]
9. Ghoneim, M.M.; El-Desoky, H.S.; Zidan, N.M. Electro-Fenton oxidation of Sunset Yellow FCF azo-dye in
aqueous solutions. Desalination 2011, 274, 22–30. [CrossRef]
10. Babuponnusami, A.; Muthukumar, K. A review on Fenton and improvements to the Fenton process for
wastewater treatment. J. Environ. Chem. Eng. 2014, 2, 557–572. [CrossRef]
11. Ai, Z.; Mei, T.; Liu, J.; Li, J.; Jia, F.; Zhang, L.; Qiu, J. Fe@Fe2 O3 core–shell nanowires as an iron reagent. 3. Their
combination with CNTs as an effective oxygen-fed gas diffusion electrode in a neutral electro-Fenton system.
J. Phys. Chem. C 2007, 111, 14799–14803. [CrossRef]
12. Zhuang, L.; Zhou, S.; Li, Y.; Liu, T.; Huang, D. In situ Fenton-enhanced cathodic reaction for sustainable
increased electricity generation in microbial fuel cells. J. Power Sources 2010, 195, 1379–1382. [CrossRef]
13. Li, Y.; Lu, A.; Ding, H.; Wang, X.; Wang, C.; Zeng, C.; Yan, Y. Microbial fuel cells using natural pyrrhotite as
the cathodic heterogeneous Fenton catalyst towards the degradation of biorefractory organics in landfill leachate.
Electrochem. Commun. 2010, 12, 944–947. [CrossRef]
14. Wang, Y.; Feng, C.; Li, Y.; Gao, J.; Yu, C.-P. Enhancement of emerging contaminants removal using Fenton reaction
driven by H2 O2 -producing microbial fuel cells. Chem. Eng. J. 2017, 307, 679–686. [CrossRef]
15. Patil, M.R.; Shrivastava, V.S. Adsorption removal of carcinogenic acid violet19 dye from aqueous solution by
polyaniline-Fe2 O3 magnetic nano-composite. J. Mater. Environ. Sci 2015, 6, 11–21.
16. Li, J.; Liu, X.; Pan, L.; Qin, W.; Chen, T.; Sun, Z. MoS2 –reduced graphene oxide composites synthesized via
a microwave-assisted method for visible-light photocatalytic degradation of methylene blue. RSC Adv. 2014,
4, 9647–9651. [CrossRef]
17. Xu, N.; Zeng, Y.; Li, J.; Zhang, Y.; Sun, W. Removal of 17β-estrodial in a bio-electro-Fenton system: Contribution of
oxidation and generation of hydroxyl radicals with the Fenton reaction and carbon felt cathode. RSC Adv. 2015,
5, 56832–56840. [CrossRef]
18. Wang, C.-T.; Huang, R.-Y.; Lee, Y.-C.; Zhang, C.-D. Electrode material of carbon nanotube/polyaniline carbon
paper applied in microbial fuel cells. J. Clean Energy Technol. 2013, 1, 206–210. [CrossRef]
40
Energies 2018, 11, 565
19. Wang, C.-H.; Wang, C.-T.; Huang, H.-C.; Chang, S.-T.; Liao, F.-Y. High stability pyrolyzed vitamin B12 as a
non-precious metal catalyst of oxygen reduction reaction in microbial fuel cells. RSC Adv. 2013, 3, 15375–15381.
[CrossRef]
20. Xu, N.; Zhou, S.; Yuan, Y.; Qin, H.; Zheng, Y.; Shu, C. Coupling of anodic biooxidation and cathodic
bioelectro-Fenton for enhanced swine wastewater treatment. Bioresour. Technol. 2011, 102, 7777–7783. [CrossRef]
[PubMed]
21. Fenton, H.J.H. LXXIII.—Oxidation of tartaric acid in presence of iron. J. Chem. Soc. Trans. 1894, 65, 899–910.
[CrossRef]
22. Anotai, J.; Chen, C.-M.; Bellotindos, L.M.; Lu, M.-C. Treatment of TFT-LCD wastewater containing ethanolamine
by fluidized-bed Fenton technology. Bioresour. Technol. 2012, 113, 272–275. [CrossRef] [PubMed]
23. Zhu, X.; Ni, J. Simultaneous processes of electricity generation and p-nitrophenol degradation in a microbial
fuel cell. Electrochem. Commun. 2009, 11, 274–277. [CrossRef]
24. Fay, S.; Kroll, U.; Bucher, C.; Vallat-Sauvain, E.; Shah, A. Low pressure chemical vapour deposition of ZnO
layers for thin-film solar cells: Temperature-induced morphological changes. Sol. Energy Mater. Sol. Cells 2005,
86, 385–397. [CrossRef]
25. Tian, C.; Zhang, Q.; Wu, A.; Jiang, M.; Liang, Z.; Jiang, B.; Fu, H. Cost-effective large-scale synthesis of ZnO
photocatalyst with excellent performance for dye photodegradation. Chem. Commun. 2012, 48, 2858–2860.
[CrossRef] [PubMed]
26. Yu, C.; Cao, F.; Li, X.; Li, G.; Xie, Y.; Jimmy, C.Y.; Shu, Q.; Fan, Q.; Chen, J. Hydrothermal synthesis and
characterization of novel PbWO4 microspheres with hierarchical nanostructures and enhanced photocatalytic
performance in dye degradation. Chem. Eng. J. 2013, 219, 86–95. [CrossRef]
27. Zhou, W.; Sun, F.; Pan, K.; Tian, G.; Jiang, B.; Ren, Z.; Tian, C.; Fu, H. Well-Ordered Large-Pore
Mesoporous Anatase TiO2 with Remarkably High Thermal Stability and Improved Crystallinity: Preparation,
Characterization, and Photocatalytic Performance. Adv. Funct. Mater. 2011, 21, 1922–1930. [CrossRef]
28. Xiong, J.; Cheng, G.; Li, G.; Qin, F.; Chen, R. Well-crystallized square-like 2D BiOCl nanoplates:
Mannitol-assisted hydrothermal synthesis and improved visible-light-driven photocatalytic performance.
RSC Adv. 2011, 1, 1542–1553. [CrossRef]
29. Feng, C.-H.; Li, F.-B.; Mai, H.-J.; Li, X.-Z. Bio-electro-Fenton process driven by microbial fuel cell for
wastewater treatment. Environ. Sci. Technol. 2010, 44, 1875–1880. [CrossRef] [PubMed]
30. Oonnittan, A.; Sillanpaa, M.E.T. Water treatment by electro-Fenton process. Curr. Org. Chem. 2012, 16, 2060–2072.
[CrossRef]
31. Kim, H.J.; Hyun, M.S.; Chang, I.S.; Kim, B.H. A microbial fuel cell type lactate biosensor using a metal-reducing
bacterium, Shewanella putrefaciens. J. Microbiol. Biotechnol. 1999, 9, 365–367.
32. Diels, A.M.; Callewaert, L.; Wuytack, E.Y.; Masschalck, B.; Michiels, C.W. Inactivation of Escherichia coli by
high-pressure homogenisation is influenced by fluid viscosity but not by water activity and product composition.
Int. J. Food Microbiol. 2005, 101, 281–291. [CrossRef] [PubMed]
33. Chen, H.S.; Zou, X.G. Treatment in Water Bodies Pollution by Ecological Floating Bed Technology.
China Water Kesources 2005, 5, 022.
34. Toyoda, M.; Inagaki, M. Heavy oil sorption using exfoliated graphite: New application of exfoliated graphite
to protect heavy oil pollution. Carbon 2000, 38, 199–210. [CrossRef]
35. Blumer, M. Oil pollution of the ocean. In Oil on the Sea; Springer: Berlin, Germany, 1969; pp. 5–13.
36. Nelson-Smith, A. The problem of oil pollution of the sea. In Advances in Marine Biology; Elsevier: Amsterdam,
The Netherlands, 1971; Volume 8, pp. 215–306.
37. Molognoni, D.; Chiarolla, S.; Cecconet, D.; Callegari, A.; Capodaglio, A.G. Industrial wastewater treatment with
a bioelectrochemical process: Assessment of depuration efficiency and energy production. Water Sci. Technol. 2018,
77, 134–144. [CrossRef] [PubMed]
38. Marashi, S.; Kariminia, H. Performance of a single chamber microbial fuel cell at different organic loads and pH
values using purified terephthalic acid wastewater. J. Environ. Health Sci. Eng. 2015, 13, 27. [CrossRef] [PubMed]
39. Kim, K.-Y.; Yang, W.; Ye, Y.; LaBarge, N.; Logan, B.E. Performance of anaerobic fluidized membrane bioreactors
using effluents of microbial fuel cells treating domestic wastewater. Bioresour. Technol. 2016, 208, 58–63. [CrossRef]
[PubMed]
40. Fujishima, A.; Honda, K. Electrochemical photolysis of water at a semiconductor electrode. Nature 1972, 238, 37–38.
[CrossRef] [PubMed]
41
Energies 2018, 11, 565
41. Xu, N.; Zhang, Y.; Tao, H.; Zhou, S.; Zeng, Y. Bio-electro-Fenton system for enhanced estrogens degradation.
42. Jadhav, G.; Ghangrekar, M. Performance of microbial fuel cell subjected to variation in pH, temperature,
external load and substrate concentration. Bioresour. Technol. 2009, 100, 717–723. [CrossRef] [PubMed]
43. Adani, F.; Baido, D.; Calcaterra, E.; Genevini, P. The influence of biomass temperature on biostabilization–biodrying
of municipal solid waste. Bioresour. Technol. 2002, 83, 173–179. [CrossRef]
44. Liang, C.; Das, K.; McClendon, R. The influence of temperature and moisture contents regimes on the aerobic
microbial activity of a biosolids composting blend. Bioresour. Technol. 2003, 86, 131–137. [CrossRef]
45. Wang, Y.; Li, B.; Zeng, L.; Cui, D.; Xiang, X.; Li, W. Polyaniline/mesoporous tungsten trioxide composite as
anode electrocatalyst for high-performance microbial fuel cells. Biosens. Bioelectron. 2013, 41, 582–588. [CrossRef]
[PubMed]
46. Wang, Z.; Zheng, Y.; Xiao, Y.; Wu, S.; Wu, Y.; Yang, Z.; Zhao, F. Analysis of oxygen reduction and microbial
community of air-diffusion biocathode in microbial fuel cells. Bioresour. Technol. 2013, 144, 74–79. [CrossRef]
[PubMed]
47. Chakrabarti, S.; Dutta, B.K. Photocatalytic degradation of model textile dyes in wastewater using ZnO as
semiconductor catalyst. J. Hazard. Mater. 2004, 112, 269–278. [CrossRef] [PubMed]
48. Smejkal, Q.; Linke, D.; Bentrup, U.; Pohl, M.-M.; Berndt, H.; Baerns, M.; Brückner, A. Combining accelerated
activity tests and catalyst characterization: A time-saving way to study the deactivation of vinylacetate catalysts.
Appl. Catal. A Gen. 2004, 268, 67–76. [CrossRef]
49. Pignatello, J.J.; Oliveros, E.; MacKay, A. Advanced oxidation processes for organic contaminant destruction
based on the Fenton reaction and related chemistry. Crit. Rev. Environ. Sci. Technol. 2006, 36, 1–84. [CrossRef]
50. Rezaei, F.; Richard, T.L.; Logan, B.E. Analysis of chitin particle size on maximum power generation, power
longevity, and Coulombic efficiency in solid–substrate microbial fuel cells. J. Power Sources 2009, 192, 304–309.
[CrossRef]
51. Xiao, L.; Damien, J.; Luo, J.; Jang, H.D.; Huang, J.; He, Z. Crumpled graphene particles for microbial fuel cell electrodes.
J. Power Sources 2012, 208, 187–192. [CrossRef]
52. Fan, Y.; Xu, S.; Schaller, R.; Jiao, J.; Chaplen, F.; Liu, H. Nanoparticle decorated anodes for enhanced current
generation in microbial electrochemical cells. Biosens. Bioelectron. 2011, 26, 1908–1912. [CrossRef] [PubMed]
42
energies
Article
Effect of Wall Boundary Layer Thickness on Power
Performance of a Recirculation Microbial Fuel Cell
Yan-Ming Chen 1 , Chin-Tsan Wang 2, * and Yung-Chin Yang 1, *
1 Institute of Materials Science and Engineering, National Taipei University of Technology, No.1, Sec. 3,
Zhongxiao E. Rd., 106 Taipei, Taiwan; [email protected]
2 Department of Mechanical and Electro-Mechanical Engineering, National I-Lan University, No.1, Sec. 1,
Shennong Rd., 26047 I Lan, Taiwan
* Correspondence: [email protected] (C.-T.W.); [email protected] (Y.-C.Y.);
Tel.: +886-3-935-7400 (ext. 7459) (C.-T.W.); +886-2-2771-2171 (ext. 2762) (Y.-C.Y.)
Received: 23 March 2018; Accepted: 17 April 2018; Published: 20 April 2018
Abstract: Hydrodynamic boundary layer is a significant phenomenon occurring in a flow through

a bluff body, and this includes the flow motion and mass transfer. Thus, it could affect the biofilm
formation and the mass transfer of substrates in microbial fuel cells (MFCs). Therefore, understanding
the role of hydrodynamic boundary layer thicknesses in MFCs is truly important. In this study, three
hydrodynamic boundary layers of thickness 1.6, 4.1, and 5 cm were applied to the recirculation mode
membrane-less MFC to investigate the electricity production performance. The results showed that
the thin hydrodynamic boundary could enhance the voltage output of MFC due to the strong shear
rate effect. Thus, a maximum voltage of 21 mV was obtained in the MFC with a hydrodynamic
boundary layer thickness of 1.6 cm, and this voltage output obtained was 15 times higher than that of
MFC with 5 cm hydrodynamic boundary layer thickness. Moreover, the charge transfer resistance
of anode decreased with decreasing hydrodynamic boundary layer thickness. The charge transfer
resistance of MFC with hydrodynamic boundary layer of thickness 1.6 cm was 39 Ω, which was
0.79 times lesser than that of MFC with 5 cm thickness. These observations would be useful for
enhancing the performance of recirculation mode MFCs.
Keywords: MFC; hydrodynamic boundary layer; recirculation mode; shear rate; voltage; charge
transfer resistance
1. Introduction
In recent years, the development of renewable energy (solar energy, wind power, bio-energy, and
fuel cell) and electric transportation/storage technologies (supercapacitors) has begun to flourish due
to the consumption and pollution that is caused by fossil fuels [1–3]. Microbial fuel cells (MFCs) are
renewable energy technology transducers [4], which has the potential for application in wastewater
treatment plants [5,6], electrical devices [7,8], and biosensor [9,10], because it can use bacteria for
generating electricity from waste water. However, the power densities of MFCs are still low due to
their high internal resistances [11–13]. Thus, at present, there are many research works that are focused
on MFCs. In MFCs, microorganisms function as a biocatalyst, which can convert organic matter into
electricity by using three extracellular electron transfer (EET) pathways: (1) direct electron transfer
(c-type cytochromes); (2) electron mediators; and, (3) solid conductive matrix (electrical conductive pili
or nanowires) [14]. Nevertheless, the microbes using electron mediators are not suitable for commercial
application as they are planktonic and could be easily washed out along with the effluent [15], and
also has higher energy losses than other EET pathways [14]. Hence, it obvious that anode biofilm plays
a key role in the electricity production of MFCs [16,17], as the bacteria that use direct electron transfer
and solid conductive matrix pathway will be attached on the anode electrode.

Energies 2018, 11, 1003
Recently, there are different methods to enhance the anode biofilm formation. Li et al. demonstrated
that the gravity effect being applied to the MFC not only facilitated easy attachment of microbes on
the electrode, but also increased the biofilm thickness. Thus, the startup time was shorter than the
control by 12%, and the maximum current density reached 8.41 ± 0.13 A/m2 , which was 29% higher
than without the gravitational effect [18]. He et al. indicated that if the fiber diameter of anode fiber
was lesser than the bacteria size, the thickness of biofilm will be less, which may result in less current
density due to the fact that the fibers will form small pores and hinder the bacterial penetration [19].
Ren et al. found that the CNT-based electrode materials: Spin-Spray Layer-by-Layer (SSLbL) CNT
can guarantee a thicker biofilm than other CNT materials. Therefore, it can result in a high current
density of 2.59 A/m2 [20]. Although, these studies stated that enhanced biofilm thickness lead to the
enhanced electricity production of MFC, but most of the biofilm formation is dependent on natural
adsorption [21]. In addition, very thick biofilms will not only block the electron transfer of MFCs [21],
but also will become a diffusion barrier on the physical and chemical exchanges between the bacteria
and substrate during the mass transfer processes [22].
In MFCs, the hydrodynamic effect is very significant because the major composition of anolyte
is water and the physical effect of fluid motion can influence the MFC performance and behavior
of biofilm microbial communities [23]. Nowadays, flow rate and recirculation rate are vital flow
parameters in continuous and recirculation mode MFCs due to two reasons, first, the flow rate and
recirculation rate can be easily set by peristaltic pump; second, the effect of these parameters is equally
important as that of the effects of shear stress [23–26]. Increasing the flow/recirculation rate could
enhance the shear rate [27,28]. Thus, resulting in a denser and thicker biofilm. Moreover, high shear
rates can facilitate faster bacterial attachment on the electrode and shorten the startup time of MFC.
Pham et al. indicated that the current and power output of MFC under the shear rate of about
120/s which was higher than that of low shear rate (∼ =0.3/s) by 2 to 3 times [27]. On the other hand,
increasing the shear stress also can enhance the mass transfer rate between the substrate and the anode
biofilm [28,29], and this will efficiently limit the mass transfer losses of MFCs [30,31]. Wang et al.
reported that a high recirculation rate of 40 mL/min increased the mass transfer rate of MFC due to high
shear stress. Therefore, warburg diffusion resistance of MFC was only 265.4 Ω, which is 73% less than
the recirculation rate of 4 mL/min, than, the maximum current density of 50.19 mA/m2 [31]. Therefore,
the shear stress effect is worthy for investigation. Furthermore, the employment of honeycomb and
avoiding the membrane in MFC was also important in this study. Because many researches indicated
that the use of membranes would not only limit the MFC application [32], but also decrease system
performance [33] and enhance the cost [32,34]. Adding a honeycomb pattern could maintain the liquid
flow inside a reactor in a homogeneous, symmetrical, and uniform manner, which can enhance the
mass transfer of substrates [31].
However, when anolyte with a value of flow/recirculation rate flows through the MFC anode
chamber, viscous shearing forces occur at the chamber wall and the hydrodynamic boundary layer is
formed on the chamber wall. The thickness of the hydrodynamic boundary layer will increase with
an increase in distance from the leading edges of the chamber wall. The flow velocity gradient will
result in a strong shear stress within the hydrodynamic boundary layer [35]. Therefore, the effect of the
hydrodynamic boundary layer would not only affect the biofilm structure, but it would also influence
the mass transfer [36,37]. So, an in-depth understanding of the effects of hydrodynamic boundary layer
thickness is necessary. But, the effects of hydrodynamic boundary layer on MFC electricity production
has not been well discussed yet. Thus, in this study, three different hydrodynamic boundary layer
thicknesses (1.6, 4.1, and 5 cm) were applied to recirculation mode membrane-less MFCs with a
honeycomb pattern for investigating the hydrodynamic boundary layer effect on internal resistance.
44
Energies 2018, 11, 1003
2.1. Voltage Output of MFCs

In order to investigate the effect of hydrodynamic boundary layer on MFCs, three different
hydrodynamic boundary thickness (1.6, 4.1, and 5 cm) were applied to MFCs, and the voltage output
was shown in Figure 1. All three MFCs obtained the similar output voltage at the beginning of the
operation. However, it could be observed that the decrease in thickness of the hydrodynamic boundary
layer shortened the startup time of MFCs. The voltage of MFC-1 started to rise up by the day 3.5,
but MFC-2 and MFC-3 started voltage production 1.5 days later when compared to MFC-1. At day 5,
the voltage of MFC-1 increased rapidly, and reached 11 mV at day 6. Although, the voltage of MFC-2
had the similar trend with MFC-1 at day 5, but it did not increase more than MFC-1. The voltage of
MFC-2 at day 6 reached ∼ =4 mV, which was just 0.36 times that of MFC-1. Nevertheless, the voltage of
MFC-2 was still higher than MFC-3 and the voltage of MFC-3 was still very similar to the voltage at
the beginning of the operation. The reason why less hydrodynamic boundary thickness had faster
and higher voltage production was due to the shear rate effect [27]. In a pipe flow, the shear rate is
high at the inlet, but it will decrease gradually to the fully developed value [38], which means that
a thin hydrodynamic boundary layer could result in a higher shear rate than that of thick boundary
layer. The high shear rate could enhance the bacteria attachment to the electrode [27]. In this study, the
MFC-1 had the highest shear rate of 0.047/s and MFC-3 has the lowest shear rate of 0.016/s (Table 1).
Therefore, the MFC-1 had the shortest startup time, than MFC-2 and MFC-3. In addition, the shear rate
of three MFCs were less than 120/s. Therefore, the bacteria were not detached from the electrode [27].
After day 6, the voltage of three MFCs started to decrease. Hence, the anolyte of three MFCs were
replaced by fresh medium. However, from day 6 to 10, the voltage of three MFCs failed to rise up
to the same level as day 6, and the voltage of MFC-1 even started decreasing. This might be due to
the fact that a high concentration of glucose may have a toxic effect on bacteria [39]. Nonetheless,
the voltage of MFC-1, MFC-2, and MFC-3 started to increase after day 11.7 and the MFC-1 reached a
maximum voltage of 21 mV by day 16, which was 15 times that of MFC-3.
Figure 1. The voltage output versus time of microbial fuel cells (MFCs) with different hydrodynamic
boundary layer thicknesses.
Table 1. Internal resistance of MFCs with respect to hydrodynamic boundary layer thickness.
Kinds δ (cm) G (1/S) R1 (Ω) Ra,g (Ω) Ca,g (C) Ra,t (Ω) Ca,t (C) Rc (Ω) Cc (C) Wc (C)
MFC-1 1.6 0.047 47.4 39 0.0015 12.6 1.5 × 10−6 15.8 1.6 × 10−6 0.00086
MFC-2 4.1 0.02 45 45 0.0013 12 0.0025 3.85 × 10−6 0.00098 0.00016
MFC-3 5 0.016 40.5 49.1 0.0018 11 1.7 × 10−6 13.5 0.0008 0.00085
45
Energies 2018, 11, 1003
2.2. Internal Resistance of MFCs

In order to investigate the hydrodynamic boundary layer thickness effect on the internal resistance
of the reactor, the equivalent circuit and internal resistance of MFCs were analyzed, as shown in Figure 2
and Table 1. According to Figure 2, it could be observed that the internal resistance of the MFCs
was composed of ohmic losses, charge transfer losses of anode/cathode, and concentration losses.
R1 represented the ohmic losses in the MFCs [40] and the value of ohmic losses can be related to the
conductivity of anolyte or the conductivity of electrode material. Ra , which represented the charge
transfer losses of anode, which can indicate the activity of the biofilm [41]. Thus, two kind of charge
transfer resistances (Ra,g : graphite felt sheet, Ra,t : titanium wire) occurred at anode. Capacitance Ca,g
and Ca,t indicated the electrical double layer phenomenon [42] of graphite felt sheet and titanium wire,
and this phenomenon occurred at the interface between the electrode and anolyte. Rc and Cc denoted
the charge transfer losses and the electrical double layer of cathode. In addition, the Wc represented
the concentration losses of MFCs [11], and the concentration losses could describe the diffusion of
substrate, oxygen, or other reactants in MFCs.
Figure 2. The equivalent circuit of MFC.
In Table 1, the ohmic losses were 47.4, 45, and 40.5 Ω for hydrodynamic boundary layers with
thicknesses of 1.6, 4, and 5 cm, respectively. These results showed that the ohmic losses were similar
due to the same conductivity of the Phosphate buffered saline (PBS) [31]. On the other hand, it can
be clearly seen that the value of Ra,g decreased when the hydrodynamic boundary layer thickness
decreased. The charge transfer loss was 39 Ω with a hydrodynamic boundary layer thickness of 1.6 cm
and increased to 49.1 Ω when the thickness was 5 cm, and these results were in accordance with the
voltage output of MFCs (Figure 1). Due to the strong shear rate effect within the thin hydrodynamic
boundary layer thickness [35,38], the physical properties of biofilms, such as density and strength,
were influenced by fluid shear. Also, the easy attachment of bacteria on the electrode can result in a
thicker and denser biofilm structure [27,28,43]. This biofilm can produce higher biomass than that of
biofilm in lower shear rate environment [23]. Hence, the electron transfer rate could be enhanced and
a lower charge transfer resistance could be obtained. Whereas, the effect of boundary layer thickness
did not significantly influence the charge transfer losses of titanium wire and the values of Ra,t was
very similar. This may be due to the smooth surface of titanium wire, which cannot facilitate easy
bacterial attachment [40]. The capacitance of MFC was related to the electron or ion accumulation
phenomena in the biofilm or the substrate [36]. In this study, the capacitances (Ca,g ) were 0.0015,
0.0013, 0.0018 C for the hydrodynamic boundary layers of thickness 1.6, 4.1, and 5 cm, respectively.
These values were very similar to each other and it can be observed that the thickness of hydrodynamic
boundary layer had no effect on the values of Ca,g and these results were similar with the research
of Ter Heijne et al. [36]. Moreover, the capacitances of Ca,t also reported the similar values (around
1.5 × 10−6 ), except for the hydrodynamic boundary layer thickness of 4.1 cm. However, this study
suggests that electrochemical impedance spectroscopy (EIS) should be analyzed using three-electrode
mode for the half-potential for data accuracy. Regarding the internal resistance of the cathode, the
cathode of all three MFCs was installed at the top of the reactor, and, the oxygen was directed to the
cathode and it did not pass through the hydrodynamic boundary layer. Thus, the thickness of the
hydrodynamic boundary layer did not influence the diffusion of oxygen. Hence, the concentration
losses (Wc ) of three MFCs were similar.
46
Energies 2018, 11, 1003
2.3. The Mechanism of Hydrodynamic Boundary Layer Effects in Recirculation Mode MFCs
In recirculation mode MFCs, the hydrodynamics is significantly important as to the microbe,
electrode biofilm, and the diffusion of reactant being affected by the flow field [44,45]. Thus, the
effect of the flow field in the MFC chamber is necessary to be described. According to the results
of this study and other references, the working mechanism model of the hydrodynamic boundary
layer effect was illustrated in Figure 3. At first, when the anolyte flowed through the surge tank and
passed through the honeycomb into the MFC chamber, the fluid particles of the anolyte in the layer,
which were in contact with the chamber wall of MFC had completely stopped due to the no-slip
condition. Particles of anolyte also slowed down gradually because of the friction in adjacent layers.
Then, in order to compensate the decrease of mass flow rate near the MFC chamber wall, the velocity
of the fluid at the midsection of the MFC chamber needed to be increased. Due to the velocity gradient
(shear rate) that was developed in the MFC chamber, the velocity (hydrodynamic) boundary layer was
formed on the MFC chamber wall [38]. The velocity boundary layer thickness increased along with
the flow direction due to the viscosity of liquid [35,46]. Thus, the MFC-1 had the much thinner velocity
boundary layer, and the MFC-3 had the thickest velocity boundary layer (see Figure 3).
Figure 3. The mechanism of hydrodynamic boundary layer.
However, due to the shear rate effect, the efficiency of bacteria that were attached to the electrode
was enhanced [27]. The anaerobic biofilm was formed at the anode electrode because the environment
was more anaerobic than cathode, and the aerobic microbes moved to the cathode due to the presence
of oxygen. Thin velocity boundary layer could result in a much thicker and denser biofilm than that of
the thick velocity boundary layer due to the strong shear rate effect. This could enhance the current
and power output of MFC [27]. The biofilm thickness reached maximum within the transition zone
between laminar and turbulent because of two reasons, first, the diffusion of substrate was limited in
laminar zone, second, the bacteria was detached from the biofilm in turbulent flow [47]. In this study,
47
Energies 2018, 11, 1003
the inlet Reynolds number was Re = 89.6, which meant that the inertial forces were dominant in MFC,
and that the velocity of boundary layer was laminar. Thus, the mass transfer of MFCs needed to be
enhanced by increasing the Reynolds number or the flow convective [45]. Although the thickness
of the biofilm would become a diffusion barrier [22] the thickness of the biofilm (about 40 μm) was
smaller than the hydrodynamic boundary layer [37]. Therefore, the mass transfer of recirculation
mode MFC was mainly determined by the thickness of hydrodynamic boundary layer [35,36]. Inside
the velocity boundary layer, the concentration boundary layer occurred due to the concentration
gradient between the bulk liquid and the chamber wall. The thickness of the concentration boundary
layer increased with the flow direction until it reached the velocity boundary layer [48]. Therefore,
the thickest concentration boundary layer occurred in MFC-3. In the concentration boundary layer,
the mass transfer of reactant relied on diffusion, which meant that a higher mass transfer could be
obtained in a thin concentration boundary layer due to a high concentration gradient. In recirculation
mode MFCs, due to the moving fluid, the mass transfer not only depended on diffusion, but also on
bulk fluid motion. The fluid motion enhanced the mass transfer as the reactant in the fluid can move
from high a concentration area to a low concentration area [48]. Hence, the strong shear rate effect in
thin velocity boundary layer not only reduced the charge transfer losses, but also the concentration
losses of MFCs [31,35,36]. It is worth mentioning that the boundary layer had no effect on the oxygen
diffusion as the oxygen was in direct contact with the cathode in this study. Moreover, research works
had indicated that the viscous sublayer was also important [49–51] due to the fact that it could affect
the biofilm accumulation and this could be suggested as an investigation for prominent future research.
2.4. Applications
Microbial fuel cells can be used in wastewater treatment plants or microfluidic devices in the future.
However, the wastewater treatment plants and microfluidic devices are the continuous/recirculation
flow mode systems. Thus, the effect of hydrodynamic boundary layer can not be ignored, and it
must be considered. The thin hydrodynamic boundary layer could significantly enhance the biofilm
formation and substrate mass transfer, and the maintainence of a thin hydrodynamic boundary layer
is a challenge in MFC performance and commercialization. There are several methods to maintain
the anode/cathode electrode in a thin hydrodynamic boundary layer. Firstly, the electrode placement
should be place near the flow entrance, this suggestion could be indirect proof by other research.
Ye et al. showed that the biofilm were thick at the beginning of the microchannel, and thin at the end
of the microchannel [52]; Second, if the anolyte flows past the porous electrode (graphite/carbon
felt) a thin porous electrode of 0.1 mm porosity should be chosen as the substrate transfer and the
diffusion of protons from the biofilm will be limited if the flow passes through the thick felt [53,54];
Third, using the flow control method to inhibit the hydrodynamic boundary layer developing, such
as increasing the number of anolyte inlet along the MFC flow channel. This method could let the
new hydrodynamic boundary layer replace the original hydrodynamic boundary layer, and keep the
hydrodynamic boundary layer thickness thin [35]. The results of this research significantly have great
contribution on the design of any continuous/recirculation flow mode MFC systems in the future.
3. Materials and Methods
3.1. Reactor Construction

In order to investigate the effects of hydrodynamic boundary layer on the electricity production of
MFC, three single chambered reactors were fabricated with Polymethylmethacrylate (PMMA) sheets
(Figure 4A) [30]. The reactors were of symmetrical structure, and consisted of one MFC chamber,
two honeycombs, and two surge tanks. No membranes were used, as it was a membrane-less reactor.
The inner space dimension of MFC chamber was 10 cm × 10 cm × 10 cm. Honeycomb structure
was made of plastic straws, and each one had an inner space dimension of 10 cm × 10 cm × 10 cm.
The surge tanks had an inner space dimension of 10 cm × 17 cm × 17.5 cm. Two graphite felt sheets
48
Energies 2018, 11, 1003
(each with L × W × D: 0.5 cm × 2.8 cm × 0.5 cm) were used as the anode and cathode electrode,
and they were embedded at the bottom and top of the MFC chambers. The anode electrode was
connected using a titanium wire to an external resistance (1000 Ω), which was further connected to the
cathode electrode. In order to generate three different hydrodynamic boundary layer thicknesses in
the MFC chamber, the anode and cathode electrodes were placed at the distances of 1 cm (MFC-1),
6 cm (MFC-2), and 9 cm (MFC-3), which were measured from the leading edges of the chamber wall.
3.2. Inoculation and Operational Conditions

The MFCs (MFC-1, MFC-2, MFC-3) were inoculated with 1:1 ratio of wastewater from the
Department of Biotechnology and Animal Science, National Ilan University, Yilan, Taiwan [11] and
50 mM PBS [31]. 50 mM PBS contained Na2 HPO4 , 4.58 g/L; NaH2 PHO4 ·H2 O, 2.45 g/L; NH4 Cl,
0.31 g/L; KCl, 0.13 g/L; and, 0.2 g/L glucose was used as the substrate. All of the MFCs were
operated at the same flow rate of 480 mL/min from the beginning till the end of the experiment and
the experimental setup was shown in Figure 4B.
Figure 4. The (A) clear side view and (B) experimental setup of recirculation mode MFCs.
3.3. Electrochemical Analysis

Three MFCs (MFC-1, MFC-2, and MFC-3) were connected with an external resistance of 1000 Ω.
The voltage of MFCs was measured by an automatic data acquisition system (Jiehan 5020, Jiehan
Technology Corporation, Taiwan, Republic of China) with a sample rate of 1 point/min [55]. The EIS
analysis of MFCs was measured by an impedance analyzer (HIOKY 3522-50 LCR HiTESTER, Hioki
E.E. Corporation, Nagano, Japan), and the scan frequency was set as 100 KHz to 0.1 Hz with an AC
amplitude of 10 mV [56]. The value of equivalent circuit of MFCs was analyzed by Zware.
3.4. Reynolds Number (Re)

Reynolds number is a dimensionless quantity that can be used to investigate the different fluid
flow situations in fluid mechanics. Reynolds number (Re) is a ratio of inertia force to viscous forces,
and is defined as shown in Equation (1) [57]:

Re = ρVDT μ (1)
where ρ represents the density of water (997 kg/m3 ); V indicate the inlet flow velocity (80 × 10−5 m/s,
corresponding to the flow rate of 480 mL/min); DT represents the hydraulic diameter (10 cm); and,
μ is viscosity of water (8.9 × 10−4 kg/ms).
49
Energies 2018, 11, 1003
3.5. Flat Plate Boundary Layer Thickness (δ) and Shear Rate
The hydrodynamic boundary layer was formed by liquid flow through a flat plate due to the
viscosity of liquid particle, and the thickness of flat plate boundary layer thickness, which can be
defined as Equation (2) [46]:
δ = 5 μx ρV (2)
where x is the electrode position of 1 cm (MFC-1), 6 cm (MFC-2), and 9 cm (MFC-3). (Corresponding to

the hydrodynamic boundary layer thickness of 1.6, 4.1, and 5 cm [58]), and the schematic diagram of
hydrodynamic boundary layer that was applied to MFCs were showed in Figure 5.
The hydrodynamic boundary layer thicknesses were chosen based on different flow field.
hydrodynamic boundary of thickness 5 cm occurred at the hydrodynamically fully developed region.
In this region, the velocity profile was fully developed and remained unchanged. The hydrodynamic
boundary of thickness 1.6 cm occurred near the entrance of the MFC chamber. The velocity profile on
this region was similar to the inlet flow.
Shear rate can influence the mass transfer and biofilm formation, and the shear rate is defined as
shown in Equation (3) [27]:
G = V/d (3)
where V is the flow velocity at hydrodynamic boundary layer (0.99 times less than inlet flow velocity)
(m/s), and d represents the thickness of the hydrodynamic boundary layer (m).
Figure 5. The schematic diagram of hydrodynamic boundary layer in the MFCs.
4. Conclusions
Hydrodynamic boundary layers of thickness 1.6, 4.1, and 5 cm were applied to the recirculation
mode MFCs to investigate the electricity production in this study. It was observed that the voltage of
MFCs increased with decreasing hydrodynamic boundary layer thicknesses. Due to the high shear
rate of 0.047/s that occurred in the hydrodynamic boundary layer of 1.6 cm, the maximum voltage
50
Energies 2018, 11, 1003
of 21 mV was observed in MFC-1, which was 15 times higher than that of MFC-3. These results
were further confirmed by EIS analysis. The charge transfer losses of anode decreased with the
decreasing hydrodynamic boundary layer thicknesses. The charge transfer resistance (Ra,g ) of MFC-1
was only 39 Ω, which was only 0.79 times lesser than that of MFC-3. These findings were useful for
the improvement in the performance of the continuous/recirculation flow mode MFCs.
Acknowledgments: The authors would like to acknowledge the generous funding support from MOST Taiwan
under contract No. MOST 103-2221-E-197-022-MY3, 106-2923-E-197-001-MY3, 106-2622-E-197-007-CC3 and
106-2221-E-197-019, and thank Sangeetha Thangavel (Department of Energy and Refrigerating Air-Conditioning
Engineering, National Taipei University of Technology, Taipei, Taiwan) for correcting the English grammar mistakes.
Author Contributions: Yan-Ming Chen: Operation of experiments, analysis and writing of article; Yung-Chin Yang:
Guidance of material experiments; Chin-Tsan Wang: Guidance of experiments and modification of manuscript and
reply of comments.
References
1. Repp, S.; Harputlu, E.; Gurgen, S.; Castellano, M.; Kremer, N.; Pompe, N.; Wörner, J.; Hoffmann, A.;
Thomann, R.; Emen, F.M. Synergetic effects of Fe3+ doped spinel Li4 Ti5 O12 nanoparticles on reduced graphene
oxide for high surface electrode hybrid supercapacitors. Nanoscale 2018, 10, 1877–1884. [CrossRef] [PubMed]
2. Genc, R.; Alas, M.O.; Harputlu, E.; Repp, S.; Kremer, N.; Castellano, M.; Colak, S.G.; Ocakoglu, K.; Erdem, E.
High-capacitance hybrid supercapacitor based on multi-colored fluorescent carbon-dots. Sci. Rep. 2017, 7,
11222. [CrossRef] [PubMed]
3. Wang, C.-T.; Chen, W.-J.; Huang, R.-Y. Influence of growth curve phase on electricity performance of
microbial fuel cell by escherichia coli. Int. J. Hydrogen Energy 2010, 35, 7217–7223. [CrossRef]
4. Khan, M.D.; Khan, N.; Sultana, S.; Joshi, R.; Ahmed, S.; Yu, E.; Scott, K.; Ahmad, A.; Khan, M.Z.
Bioelectrochemical conversion of waste to energy using microbial fuel cell technology. Process Biochem.
2017, 57, 141–158. [CrossRef]
5. Singh, H.M.; Pathak, A.K.; Chopra, K.; Tyagi, V.; Anand, S.; Kothari, R. Microbial fuel cells: A sustainable
solution for bioelectricity generation and wastewater treatment. Biofuels 2018, 1–21. [CrossRef]
6. Wu, Y.-C.; Wang, Z.-J.; Zheng, Y.; Xiao, Y.; Yang, Z.-H.; Zhao, F. Light intensity affects the performance of
photo microbial fuel cells with Desmodesmus sp. A8 as cathodic microorganism. Appl. Energy 2014, 116,
86–90. [CrossRef]
7. Wang, C.-T.; Chen, Y.-M.; Qi, Z.-Q.; Wang, Y.-T.; Yang, Y.-C. Types of simplified flow channels without flow
obstacles in microbial fuel cells. Int. J. Hydrogen Energy 2014, 39, 14306–14311. [CrossRef]
8. Yong, X.-Y.; Yan, Z.-Y.; Shen, H.-B.; Zhou, J.; Wu, X.-Y.; Zhang, L.-J.; Zheng, T.; Jiang, M.; Wei, P.; Jia, H.-H.
An integrated aerobic-anaerobic strategy for performance enhancement of pseudomonas aeruginosa-inoculated
microbial fuel cell. Bioresour. Technol. 2017, 241, 1191–1196. [CrossRef] [PubMed]
9. Chouler, J.; Cruz-Izquierdo, Á.; Rengaraj, S.; Scott, J.L.; Di Lorenzo, M. A screen-printed paper microbial fuel cell
biosensor for detection of toxic compounds in water. Biosens. Bioelectron. 2018, 102, 49–56. [CrossRef] [PubMed]
10. Sun, J.-Z.; Kingori, G.P.; Si, R.-W.; Zhai, D.-D.; Liao, Z.-H.; Sun, D.-Z.; Zheng, T.; Yong, Y.-C. Microbial fuel
cell-based biosensors for environmental monitoring: A review. Water Sci. Technol. 2015, 71, 801–809. [CrossRef]
[PubMed]
11. Wang, C.-T.; Lee, Y.-C.; Ou, Y.-T.; Yang, Y.-C.; Chong, W.-T.; Sangeetha, T.; Yan, W.-M. Exposing effect of
comb-type cathode electrode on the performance of sediment microbial fuel cells. Appl. Energy 2017, 204,
12. Liu, X.-W.; Huang, Y.-X.; Sun, X.-F.; Sheng, G.-P.; Zhao, F.; Wang, S.-G.; Yu, H.-Q. Conductive carbon
nanotube hydrogel as a bioanode for enhanced microbial electrocatalysis. ACS Appl. Mater. Interfaces 2014, 6,
8158–8164. [CrossRef] [PubMed]
13. Wang, Q.-Q.; Wu, X.-Y.; Yu, Y.-Y.; Sun, D.-Z.; Jia, H.-H.; Yong, Y.-C. Facile in-situ fabrication of
graphene/riboflavin electrode for microbial fuel cells. Electrochim. Acta 2017, 232, 439–444. [CrossRef]
14. Torres, C.I.; Marcus, A.K.; Lee, H.-S.; Parameswaran, P.; Krajmalnik-Brown, R.; Rittmann, B.E. A kinetic
perspective on extracellular electron transfer by anode-respiring bacteria. FEMS Microbiol. Rev. 2009, 34,
51
Energies 2018, 11, 1003
15. Du, Z.; Li, Q.; Tong, M.; Li, S.; Li, H. Electricity generation using membrane-less microbial fuel cell during
wastewater treatment. Chin. J. Chem. Eng. 2008, 16, 772–777. [CrossRef]
16. Feng, C.; Liu, Y.; Li, Q.; Che, Y.; Li, N.; Wang, X. Quaternary ammonium compound in anolyte
without functionalization accelerates the startup of bioelectrochemical systems using real wastewater.
17. Wang, Z.; Mahadevan, G.D.; Wu, Y.; Zhao, F. Progress of air-breathing cathode in microbial fuel cells. J. Power
Sources 2017, 356, 245–255. [CrossRef]
18. Li, T.; Zhou, L.; Qian, Y.; Wan, L.; Du, Q.; Li, N.; Wang, X. Gravity settling of planktonic bacteria to anodes
enhances current production of microbial fuel cells. Appl. Energy 2017, 198, 261–266. [CrossRef]
19. He, G.; Gu, Y.; He, S.; Schröder, U.; Chen, S.; Hou, H. Effect of fiber diameter on the behavior of biofilm
and anodic performance of fiber electrodes in microbial fuel cells. Bioresour. Technol. 2011, 102, 10763–10766.
[CrossRef] [PubMed]
20. Ren, H.; Pyo, S.; Lee, J.-I.; Park, T.-J.; Gittleson, F.S.; Leung, F.C.; Kim, J.; Taylor, A.D.; Lee, H.-S.; Chae, J.
A high power density miniaturized microbial fuel cell having carbon nanotube anodes. J. Power Sources 2015,
273, 823–830. [CrossRef]
21. Wang, X.; Lin, H.; Wang, J.; Xie, B.; Huang, W. Influence of the biofilm formation process on the properties of
biofilm electrode material. Mater. Lett. 2012, 78, 174–176. [CrossRef]
22. Cabije, A.H.; Agapay, R.C.; Tampus, M.V. Carbon-nitrogen-phosphorus removal and biofilm growth
characteristics in an integrated wastewater treatment system involving a rotating biological contactor.
Asia-Pac. J. Chem. Eng. 2009, 4, 735–743. [CrossRef]
23. Oliveira, V.; Simões, M.; Melo, L.; Pinto, A. Overview on the developments of microbial fuel cells.
Biochem. Eng. J. 2013, 73, 53–64. [CrossRef]
24. Ieropoulos, I.; Winfield, J.; Greenman, J. Effects of flow-rate, inoculum and time on the internal resistance of
microbial fuel cells. Bioresour. Technol. 2010, 101, 3520–3525. [CrossRef] [PubMed]
25. Zhang, F.; Jacobson, K.S.; Torres, P.; He, Z. Effects of anolyte recirculation rates and catholytes on electricity
generation in a litre-scale upflow microbial fuel cell. Energy Environ. Sci. 2010, 3, 1347–1352. [CrossRef]
26. Zhang, L.; Zhu, X.; Li, J.; Kashima, H.; Liao, Q.; Regan, J.M. Step-feed strategy enhances performance of
unbuffered air-cathode microbial fuel cells. RSC Adv. 2017, 7, 33961–33966. [CrossRef]
27. Pham, H.T.; Boon, N.; Aelterman, P.; Clauwaert, P.; De Schamphelaire, L.; Van Oostveldt, P.; Verbeken, K.;
Rabaey, K.; Verstraete, W. High shear enrichment improves the performance of the anodophilic microbial
consortium in a microbial fuel cell. Microb. Biotechnol. 2008, 1, 487–496. [CrossRef] [PubMed]
28. Shen, Y.; Wang, M.; Chang, I.S.; Ng, H.Y. Effect of shear rate on the response of microbial fuel cell toxicity
sensor to Cu (II). Bioresour. Technol. 2013, 136, 707–710. [CrossRef] [PubMed]
29. Celmer, D.; Oleszkiewicz, J.; Cicek, N. Impact of shear force on the biofilm structure and performance of a
membrane biofilm reactor for tertiary hydrogen-driven denitrification of municipal wastewater. Water Res.
2008, 42, 3057–3065. [CrossRef] [PubMed]
30. Wang, C.-T.; Huang, Y.-S.; Sangeetha, T.; Chen, Y.-M.; Chong, W.-T.; Ong, H.-C.; Zhao, F.; Yan, W.-M.
Novel bufferless photosynthetic microbial fuel cell (PMFCs) for enhanced electrochemical performance.
31. Wang, C.-T.; Huang, Y.-S.; Sangeetha, T.; Yan, W.-M. Assessment of recirculation batch mode operation in
bufferless bio-cathode microbial fuel cells (MFCs). Appl. Energy 2018, 209, 120–126. [CrossRef]
32. Jang, J.K.; Chang, I.S.; Kang, K.H.; Moon, H.; Cho, K.S.; Kim, B.H. Construction and operation of a novel
mediator-and membrane-less microbial fuel cell. Process Biochem. 2004, 39, 1007–1012. [CrossRef]
33. Logan, B.E. Microbial Fuel Cells; John Wiley & Sons: Hoboken, NJ, USA, 2008.
34. Ghangrekar, M.; Shinde, V. Performance of membrane-less microbial fuel cell treating wastewater and effect
of electrode distance and area on electricity production. Bioresour. Technol. 2007, 98, 2879–2885. [CrossRef]
[PubMed]
35. Yang, Y.; Ye, D.; Liao, Q.; Zhang, P.; Zhu, X.; Li, J.; Fu, Q. Enhanced biofilm distribution and cell performance of
microfluidic microbial fuel cells with multiple anolyte inlets. Biosens. Bioelectron. 2016, 79, 406–410. [CrossRef]
[PubMed]
36. Ter Heijne, A.; Schaetzle, O.; Gimenez, S.; Fabregat-Santiago, F.; Bisquert, J.; Strik, D.P.; Barriere, F.;
Buisman, C.J.; Hamelers, H.V. Identifying charge and mass transfer resistances of an oxygen reducing
biocathode. Energy Environ. Sci. 2011, 4, 5035–5043. [CrossRef]
52
Energies 2018, 11, 1003
37. Ter Heijne, A.; Hamelers, H.V.; Saakes, M.; Buisman, C.J. Performance of non-porous graphite and
titanium-based anodes in microbial fuel cells. Electrochim. Acta 2008, 53, 5697–5703. [CrossRef]
38. Yunus, A.C.; Cimbala, J.M. Fluid Mechanics Fundamentals and Applications, International ed.; McGraw Hill
Publication: New York, NY, USA, 2006; pp. 325–326.
39. D’souza Rohan, V.D.; Rohan, G.; Satish, B. Bioelectricity production from microbial fuel using escherichia
coli (glucose and brewery waste). Int. Res. J. Biol. Sci. 2013, 2, 50–54.
40. Chen, Y.-M.; Wang, C.-T.; Yang, Y.-C.; Chen, W.-J. Application of aluminum-alloy mesh composite carbon
cloth for the design of anode/cathode electrodes in escherichia coli microbial fuel cell. Int. J. Hydrogen Energy
2013, 38, 11131–11137. [CrossRef]
41. Liu, P.; Liang, P.; Jiang, Y.; Hao, W.; Miao, B.; Wang, D.; Huang, X. Stimulated electron transfer inside
electroactive biofilm by magnetite for increased performance microbial fuel cell. Appl. Energy 2018, 216,
42. Zhang, F.; Merrill, M.D.; Tokash, J.C.; Saito, T.; Cheng, S.; Hickner, M.A.; Logan, B.E. Mesh optimization for
microbial fuel cell cathodes constructed around stainless steel mesh current collectors. J. Power Sources 2011,
196, 1097–1102. [CrossRef]
43. Stoodley, P.; Sauer, K.; Davies, D.G.; Costerton, J.W. Biofilms as complex differentiated communities.
Annu. Rev. Microbiol. 2002, 56, 187–209. [CrossRef] [PubMed]
44. Wang, C.-T. Flow control in microbial fuel cells. In Technology and Application of Microbial Fuel Cells; InTech:
Rijeka, Croatia, 2014.
45. Costerton, J.W.; Lewandowski, Z.; Caldwell, D.E.; Korber, D.R.; Lappin-Scott, H.M. Microbial biofilms.
Annu. Rev.‘Microbiol. 1995, 49, 711–745. [CrossRef] [PubMed]
46. Schlichting, H.; Gersten, K. Boundary-Layer Theory, 7th ed.; McGraw-Hill Book Company: Columbus, OH,
USA, 1989; p. 140.
47. Lewandowski, Z.; Walser, G. Influence of Hydrodynamics on Biofilm Accumulation; Environmental Engineering;
ASCE: Reston, VA, USA, 1991; pp. 619–624.
48. Cengel, A. Heat and Mass Transfer A Practical Approach, 3rd ed.; McGraw-Hill Book Company: Columbus,
OH, USA, 2006; p. 810.
49. Bryers, J.D.; Characklis, W.G. Processes governing primary biofilm formation. Biotechnol. Bioeng. 1982, 24,
50. Kim, T.W.; Micheli, F. Decreased solar radiation and increased temperature combine to facilitate fouling by
marine non-indigenous species. Biofouling 2013, 29, 501–512. [CrossRef] [PubMed]
51. De Beer, D.; Stoodley, P.; Lewandowski, Z. Liquid flow in heterogeneous biofilms. Biotechnol. Bioeng. 1994,
52. Ye, D.; Yang, Y.; Li, J.; Zhu, X.; Liao, Q.; Deng, B.; Chen, R. Performance of a microfluidic microbial fuel cell
based on graphite electrodes. Int. J. Hydrogen Energy 2013, 38, 15710–15715. [CrossRef]
53. Sleutels, T.H.; Lodder, R.; Hamelers, H.V.; Buisman, C.J. Improved performance of porous bio-anodes
in microbial electrolysis cells by enhancing mass and charge transport. Int. J. Hydrogen Energy 2009, 34,
9655–9661. [CrossRef]
54. Sleutels, T.H.; Hamelers, H.V.; Buisman, C.J. Effect of mass and charge transport speed and direction in porous
anodes on microbial electrolysis cell performance. Bioresour. Technol. 2011, 102, 399–403. [CrossRef] [PubMed]
55. Wang, C.-T.; Chen, Y.-M.; Hu, Z.-Y.; Chong, W.-T. Dynamic power response of microbial fuel cells under
external electrical exciting. Int. J. Hydrogen Energy 2017, 42, 22208–22213. [CrossRef]
56. Hutchinson, A.J.; Tokash, J.C.; Logan, B.E. Analysis of carbon fiber brush loading in anodes on startup and
performance of microbial fuel cells. J. Power Sources 2011, 196, 9213–9219. [CrossRef]
57. Van Dyke, M.; Van Dyke, M. An Album of Fluid Motion; Parabolic Press: Stanford, CA, USA, 1982.
58. Chen, Y.-M.; Wang, C.-T.; Yang, Y.-C.; Wang, Y.-T. Effect of Boudary Layer Thickness on the Performance of
Microbial Fuel Cell. In Proceedings of the EMChIE 2015, Tarragona, Spain, 10–12 June 2015.
53
energies
Article
Microbial Biodiesel Production by Direct
Transesterification of Rhodotorula glutinis Biomass
I-Ching Kuan, Wei-Chen Kao, Chun-Ling Chen and Chi-Yang Yu *
Department of Bioengineering, Tatung University, Taipei 10452, Taiwan; [email protected] (I.-C.K.);
[email protected] (W.-C.K.); [email protected] (C.-L.C.)
* Correspondence: [email protected]; Tel.: +886-2-2182-2928 (ext. 6330)
Abstract: (1) Background: Lipids derived from oleaginous microbes have become promising
alternative feedstocks for biodiesel. This is mainly because the lipid production rate from microbes
is one to two orders of magnitude higher than those of energy crops. However, the conventional
process for converting these lipids to biodiesel still requires a large amount of energy and organic
solvents; (2) Methods: In this study, an oleaginous yeast, Rhodotorula glutinis, was used for direct
transesterification without lipid pre-extraction to produce biodiesel, using sulfuric acid or sodium
hydroxide as a catalyst. Such processes decreased the amount of energy and organic solvents
required simultaneously; (3) Results: When 1 g of dry R. glutinis biomass was subject to direct
transesterification in 20 mL of methanol catalyzed by 0.6 M H2 SO4 at 70 ◦ C for 20 h, the fatty acid
methyl ester (FAME) yield reached 111%. Using the same amount of biomass and methanol loading
but catalyzed by 1 g/L NaOH at 70 ◦ C for 10 h, the FAME yield reached 102%. The acid-catalyzed
process showed a superior moisture tolerance; when the biomass contained 70% moisture, the FAME
yield was 43% as opposed to 34% of the base-catalyzed counterpart; (4) Conclusions: Compared to
conventional transesterification, which requires lipid pre-extraction, direct transesterification not
only simplifies the process and shortens the reaction time, but also improves the FAME yield.
Keywords: biodiesel; direct transesterification; Rhodotorula glutinis; single cell oil
1. Introduction
Biodiesel is one of the most promising renewable fuels in transportation. It can be used as
a drop-in replacement fuel for existing diesel vehicles and boiler engines without major modifications.
It is also compatible with current fuel infrastructure [1]. Biodiesel is defined as the fatty acid, alkyl
monoesters, derived from renewable feedstocks such as vegetable oils, animal fats, and waste cooking
oil. Compared to conventional petrodiesel, biodiesel is highly degradable, non-toxic, and cleaner in
exhaust emissions, with the exception of NOx [2]. The combustion properties of biodiesel are similar
to those of petroleum diesel.
Currently, commercial biodiesel is produced by the transesterification of plant oils with short-chain
alcohols, using alkaline catalysts such as NaOH and KOH; most of these plant oils are edible, including
rapeseed, sunflower, palm, and soybean oil [3]. The biodiesel derived from these oils (first generation
biodiesel) has two major drawbacks: (1) the high cost of feedstocks renders biodiesel unable to
compete commercially with petrodiesel [4]; (2) the utilization of these edible plants for fuel production
may endanger the world’s food supply [5]. These problems led to the development of the second
generation which mainly utilizes non-edible plant oils such as jatropha, jojoba, and waste cooking
oil [6]. However, the supply of these non-edible oils is not likely to meet the global demand for
biodiesel. Recently, the use of oil-accumulating microbes as feedstocks for biodiesel production has
drawn a lot of attention [7]; the biodiesel derived from these oleaginous microalgae, bacteria, yeasts,
and fungi is commonly referred to as the third-generation.

Energies 2018, 11, 1036
These microbial lipids, also known as single cell oils (SCO), have certain advantages over oils
derived from plants. Firstly, these oleaginous microbes can be cultivated all year round with suitable
substrates on non-arable lands; secondly, the oil production rate is one to two orders of magnitude
higher than those of conventional energy crops [8]. In addition, the lipid compositions of many
oleaginous microbes are not too different from those of vegetable oils [9]. This suggests that the
fuel properties of biodiesel derived from microbial origins should be similar to those derived from
vegetables oils. Among the oleaginous microbes, microalgae have been extensively studied for their
potential as alternative feedstocks for biodiesel. Microalgae are mostly phototrophic and capable of
accumulating a large amount of biomass and lipids rapidly [10]. Although microalgae are very effective
in producing SCO, if grown photo synthetically, they require a large area of land to be cultivated,
and are subject to daily and seasonal variations [9,10]. Compared to microalgae, oleaginous yeasts
could be a better alternative for the production of SCO. The growth rates of these yeasts are faster
than those of microalgae; the duplication time could be less than 1 h in certain cases [11]. Unlike
phototrophic microalgae, the growth of oleaginous yeasts is not affected by the variation in weather
and sunlight and is less prone to contamination by other microorganisms. Several yeasts are known
for their ability to accumulate SCO, such as Cryptococcus curvatus, Lipomyces starkeyi, Rhodosporidium
toruloides, and Rhodotorula glutinis [12]. R. glutinis caught our attention as an ideal SCO-producing
strain because it could be cultivated with a variety of low cost carbon sources such as crude glycerol [13]
and lignocellulosic biomass hydrolysate [14]. R. glutinis is also very effective in accumulating SCO
with oil content of up to 72% [9]. R. glutinis is also capable of synthesizing carotenoids including
β-carotene, torulene, and torularhodin [15]; the composition of these carotenoids depends on the
cultivation conditions.
The conventional process for converting yeast SCO to biodiesel includes the following steps in
this order: cell disruption, oil extraction, separation and transesterification. In order to decrease the
energy expenditure and the amount of solvents used in these steps, many researchers combine
the aforementioned steps into one, which is usually described as direct transesterification [16].
The simplified process reduces the overall high cost of biodiesel derived from microbial origins.
A fatty acid methyl esters (FAME) yield of up to 98% was reached by direct transesterification of dry R.
toruloides biomass catalyzed by either H2 SO4 or HCl [17]. Using the same dry biomass, a similar yield
was also obtained by direct transesterification using NaOH [18].
In this work, dry R. glutinis biomass was converted to FAME by direct transesterification with
either acidic or basic catalysts. The effects of catalyst amount, reaction temperature, incubation time,
and methanol loading on FAME yield were studied. In order to evaluate the feasibility of using wet
biomass directly, the influences of moisture content on the FAME yield were also examined. Finally,
under optimized reaction conditions, the yields and FAME compositions from direct transesterification
were compared with those from the conventional processes.
2.1. Effect of Catalyst Concentration on Transesterification

The average biomass concentration of R. glutinis from fermentation was 16 ± 5 g/L, with a lipid
content of 39± 6%. Simple calculation revealed that 1 g of FAME is expected for 1 g of microbial lipid
under complete transesterification. Thus the highest theoretical yield is 100%.When 0.1 and 0.2 g/L
NaOH were used as catalysts, no FAME product was detected (Figure 1a). The added NaOH could be
consumed by the saponification side reaction because there was an excess of yeast lipids. The FAME
yield increased to 18% when the catalyst concentration was 0.5 g/L; the highest FAME yield of 94%
was observed with 1 g/L NaOH. However, the FAME yield started to decrease rapidly as the catalyst
concentration increased further. No FAME was observed with 4 g/L NaOH; a high alkaline catalyst
concentration may facilitate saponification, thus leading to this low FAME yield [19]. Our results
indicated that only a narrow range of NaOH concentration was suitable for FAME production. Similar
55
Energies 2018, 11, 1036
results were observed by others using dry R. toruloides biomass as a feedstock [18]. The NaOH
concentration of 1 g/L was selected for later experiments.
When 0.05 M H2 SO4 was used as a catalyst, the FAME yield was only 46% (Figure 1b). The FAME
yield increased with the concentration of H2 SO4 , and the highest FAME yield of 103% was observed
with 0.6 M H2 SO4 . However, a further increase in the H2 SO4 concentration to 0.8 M led to a decrease in
FAME yield; a similar observation was made by others using a dry R. toruloides biomass as a feedstock,
and the decrease in FAME yield was explained as a result of side reactions such as polymerization
under harsh conditions [17]. The transesterification reaction mixture appeared orange when H2 SO4
concentration was below 0.1 M, but the color changed to light yellow as the catalyst concentration
increased further; the color change may be due to the oxidation of the caroteinoids present in the
biomass [20]. The H2 SO4 concentration of 0.6 M was selected for later experiments.

<LHOG
<LHOG

1D2+J/ +620
(a) (b)
Figure 1. The effect of catalyst concentration on fatty acid methyl ester (FAME) yield. (a) NaOH;
(b) H2 SO4 . For both catalysts, direct transesterification on 1 g of dry R. glutinis biomass was carried out
with a methanol loading of 1:20 (biomass: methanol, w/v) at 70 ◦ C for 15 h. Data were obtained by
averaging three individual measurements and the standard deviations were indicated with error bars.
2.2. Effect of Reaction Time and Temperature on Transesterification

The FAME yield obtained by NaOH catalysis was only 49% after 2 h of incubation, while increasing
dramatically up to a yield of 102% after 10 h of incubation (Figure 2). Further extension of the reaction
time did not improve the yield, indicating that the reaction reached completion after 10 h. When H2 SO4
was used as a catalyst, the FAME yield increased with the reaction time of up to 20 h with a yield
of 111%. More FAME was formed than the available lipid which could be explained by the following
reasons: (1) The lipid content within the biomass was underestimated; (2) some phopholipids were
also converted to FAME [21,22]. Further increase in the reaction time beyond 20 h did not improve the
FAME yield. For later experiments, the reaction times of 10 and 20 h were used for NaOH and H2 SO4
catalysts, respectively.
With a 10 h reaction time and NaOH as a catalyst; the FAME yields at 50 and 60 ◦ C were only
61% and 94%, respectively; even increasing the reaction duration up to 20 h did not enhance the FAME
yield. Thus, 70 ◦ C was optimal for FAME production catalyzed by NaOH. A similar observation was
also obtained with catalysis by H2 SO4 . A previous study reported an optimal temperature of 50 ◦ C
for converting the dry biomass of oleaginous R. toruloides directly to biodiesel within a lidded test
tube [18]. The difference in the optimal reaction temperature might be attributed to a different reaction
setup. In later experiments, for both NaOH and H2 SO4 , a reaction temperature of 70 ◦ C was used.
56
Energies 2018, 11, 1036

<LHOG

1D2+
+62

7LPHK
Figure 2. The effect of reaction time on FAME yield. Direct transesterification on 1 g of dry R. glutinis
biomass was carried out with a methanol loading of 1:20 (biomass: methanol, w/v) at 70◦ C using 1 g/L
NaOH or 0.6 M H2 SO4 as a catalyst. Data were obtained by averaging three individual measurements
and the standard deviations were indicated with error bars.
2.3. Effect of Methanol Loading on Transesterification

Methanol serves as one of the reactants and also acts as a solvent to weaken and disrupt the cell
walls [23]. The effect of methanol loading on transesterification is shown in Figure 3. When NaOH
was used as a catalyst, almost no FAME was formed with a biomass to methanol ratio (w/v) of 1:10;
this could be because themethanol loading is insufficient to disrupt the cell walls effectively. When the
biomass to methanol ratio was increased to 1:20, the highest yield of 102% was obtained. Nevertheless,
the yield started to decrease as the ratio was further increased; similar results were also observed by
others [24]. One possible explanation for the decreased yield at high methanol loading is that excess
methanol dilutes the biomass and catalyst concentrations, thus giving rise to a lower conversion [18,24].
On the other hand, when H2 SO4 was employed as a catalyst, the methanol loading had little influence
on the FAME yield within the range we studied. For both catalysts, a biomass to methanol ratio (w/v)
of 1:20 was used for later experiments.

<LHOG

1D2+
+62

%LRPDVVPHWKDQROZY
Figure 3. The effect of methanol loading on transesterification. Direct transesterification on 1 g of

dry R. glutinis biomass was carried out at 70 ◦ C with 1 g/L NaOH and 0.6 M H2 SO4 for 10 and
20 h, respectively. Data were obtained by averaging three individual measurements and the standard
deviations were indicated with error bars.
57
Energies 2018, 11, 1036
2.4. Effect of Moisture on Transesterification

In an esterification reaction, the moisture content should be minimized because the presence of
water favors the generation of free fatty acids from the hydrolysis of triglycerides, thus enhancing the
degree of saponification and lowering the FAME yield. Previous studies on the transesterification of
vegetable oils have shown that the presence of water even at very low concentrations has a detrimental
effect on the FAME yield [25]. To determine the effect of moisture content on the FAME yield, various
amounts of water were added back to 1 g of dry biomass; the moisture content was defined by the
following equation:
moisture content (%) = amount of added water (g)/(amount of added water (g) +
(1)
1 g of biomass)
Using NaOH as a catalyst, the FAME yield was decreased with an increase in the moisture content
(Figure 4); the FAME yield was 34% when the biomass contained 70% moisture. Our results were
quite different from those reported for R. toruloides (also using NaOH as a catalyst), which showed
that there was almost no formation of FAME with moisture content above 10% [18]. In the case of
H2 SO4 , the process tolerated moisture fairly well up to 50%. However, the FAME yield decreased
rapidly to 85% and 43% when the biomass contained 60% and 70% moisture, respectively. Similar
results were obtained by others using dry microalgae biomass, Chaetoceros gracilis, as a feed stock
with a higher methanol loading of 1:40 [21]. We have also examined the feasibility of using a wet R.
glutinis biomass collected from centrifugation without drying (the moisture content was 72%) for direct
transesterification. However, this included a heating step at 80◦ C for 10 min before centrifugation.
Consequently, the FAME yield could reach 73%. Yellapu et al. reported that the yield could be further
improved by pre-treating the wet biomass with surfactants such as N-lauroyl sarcosine, and the
reaction time may also be shortened by carrying out the reaction under sonication [26].

<LHOG

1D2+
+62

0RLVWXUH
Figure 4. The effect of moisture on FAME yield. Direct transesterification on 1 g of dry R. glutinis
biomass was carried out at 70 ◦ C with 1 g/L NaOH and 0.6 M H2 SO4 for 10 and 20 h, respectively.
A methanol loading of 1:20 (biomass: methanol, w/v) was used for both catalysts. Data were obtained by
averaging three individual measurements and the standard deviations were indicated with error bars.
2.5. Comparison of Yields and Compositions of FAME Derived from Different Transesterification Methods
The yields and compositions of FAME produced by different transesterification methods are
listed in Table 1. The conventional methods resulted in lower FAME yields of 78% and 85% using
NaOH and H2 SO4 as catalysts, respectively; the acidic method had a yield similar to that of the
previous report (81%), using the same biomass [27]. Direct transesterification, regardless of the
catalyst used, resulted in significantly higher yields in a shorter time (approximately 4 h less) than
conventional methods, which involved cell disruption, lipid extraction, and transesterification in that
58
Energies 2018, 11, 1036
order. For direct transesterification, the compositions of FAME obtained with either catalyst were
almost the same; the oleic acid methyl ester was the dominant FAME, followed by linoleic, and then
palmitic acid methyl ester. However, a previous report showed that the FAME composition was
catalyst dependent, especially for oleic and linolenic acid methyl esters [18]. The content of linoleic
acid methyl ester from direct transesterification was about 20% higher than the linoleic acid content
of R. glutinis cultured with glycerol as a carbon source [13,14]. The higher amount of linoleic acid
methyl ester was explained by the fact that some FAME was derived from the phospholipids in the
cell membrane [22,23]. When conventional methods were used, the linoleic acid methyl ester was the
dominant FAME, followed by oleic acid methyl ester instead.
Table 1. Yields and compositions of FAME derived from R. glutinis using different transesterification methods.
FAME Yield FAME Composition (%)

Method
(%) a C16:0 C16:1 C18:0 C18:1 C18:2 C18:3
Direct acid-catalyzed 111.0 ± 0.1 16.8 0.7 4.1 42.2 33.8 2.5
Direct base-catalyzed 102.0 ± 0.8 18.8 0.7 4 41.1 32.9 2.6
Conventional acid-catalyzed b 85.2 ± 0.9 18.1 1.6 3 33.3 41.1 4.1
Conventional base-catalyzed c 77.9 ± 0.7 21 1 2.5 30.5 41.2 3.7
a Data were obtained by averaging three individual measurements. b The lipid was first extracted as described
in Methods (Total lipid analysis) except that 1 g of dry biomass was used and reagents required were scaled-up
accordingly. Transesterification of extracted lipid was carried out with 0.6 M H2 SO4 for 20 h at 70 ◦ C, using
a methanol loading of 1:20 (biomass: methanol, w/v). c The lipid was first extracted as described in Methods
(Total lipid analysis) except that 1 g of dry biomass was used and the reagents required were scaled-up accordingly.
The transesterification of extracted lipid was carried out with 1 g/L NaOH for 10 h at 70 ◦ C, using a methanol
loading of 1:20 (biomass: methanol, w/v).
3.1. Microorganism and Medium

The seed medium for culturing R. glutinis BCRC (Bioresource Collection and Research Center,
Hsinchu, Taiwan) 22,360 contained 3 g of yeast extract, 3 g of malt extract, 5 g of peptone, and 10 g of
dextrose per liter. The fermentation medium contained 80 g of crude glycerol (Yu-Hwa Biodiesel
Company, Changhua, Taiwan), 2 g of yeast extract, 2 g of (NH4 )2 SO4 , 1 g of KH2 PO4 , 0.5 g of
MgSO4 ·7H2 O, 0.1 g of CaCl2 , and 0.1 g of NaCl per liter. The crude glycerol, dark brown in color and
slightly viscous, was used directly without any pretreatment or purification.
3.2. Production of Biomass

Under aerobic conditions, 300 mL of seed culture was prepared in a seed medium after a
24-h incubation at 24 ◦ C with orbital shaking at 150 rpm. The seed culture was inoculated into
a 5-L stirred-tank fermentor (BIOSTAT® A plus, Sartorius, Gottingen, Germany) containing 2.7 L of
fermentation medium (3 L working volume). The pH was maintained at 5.5 by automatically feeding
1.0 N NaOH or 1.0 N HCl solution into the medium. The fermentor was operated at 24 ◦ C with
dissolved oxygen controlled at 30, ±2% of the saturation level. The agitation during the process was
limited to a range from 200 to 400 rpm to avoid potential damage resulting from high shear force.
The biomass was harvested after 72 h of incubation by centrifugation at 12,021× g, and then the pellets
were washed with deionized water to remove any residual fermentation medium. The biomass was
stored at −80 ◦ C for 24 h before lyophilized in a Freezemobile 12 XL freeze dryer (Virtis). The dried
biomass was stored in a refrigerator for later use.
3.3. Total Lipid Analysis

The extraction of lipids from a biomass was modified from the procedure reported by Bligh and
Dyer [28]. Fifty milliliters of chloroform/methanol mixture (2:1, v/v) was mixed with 0.5 g of finely
59
Energies 2018, 11, 1036
ground powder of dry biomass. The suspension was then subject to ultrasonication with a Misonix
XL2020 sonicator for 6 min on an ice bath; the ultrasonic horn was directly immersed in the suspension
and the on/off cycle was set to 60/60 s in order to minimize the heat generated. The suspension was
incubated at room temperature for 1 h on a rocking mixer at 100 rpm. The solvent phase was collected
by centrifugation at 12,100× g for 10 min; the biomass was extracted with an additional 30 mL of
chloroform/methanol mixture for 1 h. The combined solvent phase was evaporated at 65 ◦ C under
vacuum (337 mbar), and then the remaining lipid was weighted. The lipid content was defined as below:
lipid content (%) = weight of lipid (g)/weight of biomass (g) × 100 (2)
3.4. Direct Transesterification

One gram of dry biomass and various amounts of methanol (biomass to methanol ratio in the
range of 1:10 to 1:60, w/v) were added to a 100 mL round-bottom flask fitted with a condenser. NaOH
(from 0.1 to 4 g/L) or H2 SO4 (from 0.05 to 0.8 M) dissolved in methanol was used as a catalyst.
The suspension was heated at 40 to 70 ◦ C in a water bath under atmospheric pressure with vigorous
mixing for 2 to 30 h, followed by a centrifugation at 12,100× g for 10 min and a collection of the
supernatant. The residual biomass was then washed with 10 mL of n-hexane and the washing fraction
was collected after centrifugation as described above. The supernatant and the washing fraction were
combined and extracted with equal volume of n-hexane. The n-hexane layer was collected and the
solvent was evaporated at 65 ◦ C under vacuum (337 mbar). The biodiesel product was collected and
weighted. The experiments were performed in triplicate; the average from triplicate was plotted in the
figures and the standard deviation was indicated with an error bar.
3.5. Analysis of Fatty Acid Methyl Esters

The content of FAME in the transesterification product was measured according to the standard
of Taiwan, CNS-15051. The biodiesel product was first treated with sodium sulfate followed by
centrifugation at 2040× g for 5 min. Fifty microliters of the treated sample was mixed with 1 mL
of 10 mg/mL methyl heptadecanoate in hexane as an internal standard. One microliter of the
sample was injected into a gas chromatograph (GC-2014, Shimadzu, Kyoto, Japan) equipped with
a flame-ionization detector (FID). A BPX70 capillary column (30 m × 0.25 mm i.d.; SGE Analytical
Science, Ringwood, Australia) with nitrogen as carrier gas was used. The injector and FID temperatures
were both set at 250 ◦ C. The oven temperature was initially held at 150 ◦ C for 30 s and then increased
to 180 ◦ C at 10 ◦ C/min, finally to 198 ◦ C at 1.5 ◦ C/min. The content of FAME was calculated using the
following equation:
∑ A − AEI C × VEI
FAME content (%) = × EI × 100 (3)
AEI m
where ΣA is the summation of peak area of all the FAME peaks (from C14:0 to C24:1); AEI is the peak
area of internal standard, methyl heptadecanoate; CEI is the concentration of methyl heptadecanoate,
VEI is the volume of methyl heptadecanoate, and m is the mass of the biodiesel sample. The FAME
yield was calculated using the following equation:
FAME content (%) × weight of biodiesel (g)

FAME yield (%) = × 100 (4)
lipid content (%) × weight of biomass (g)
4. Conclusions
In this study, the oleaginousyeast, R. glutinis, was used for the direct transesterification without
lipid pre-extraction to produce FAME. The use of chloroform, a highly toxic organic solvent often
applied in lipid extraction, was avoided. The highest FAME yield of 111% was obtained with 0.6 M
H2 SO4 as a catalyst and a methanol loading of 1:20 (biomass to methanol, w/v), the reaction was
carried out at 70 ◦ C for 20 h. Using the same methanol loading and temperature, a slightly lower yield
60
Energies 2018, 11, 1036
of 102% was obtained with 1 g/L NaOH after 10 h of incubation. Although basic catalysts are often
avoided because of the potential saponification caused by the free fatty acids present in SCO, the NaOH
catalyst was still able to achieve similar FAME yield in half the reaction time required by the H2 SO4
catalyst. However, the acid-catalyzed direct transesterification tolerates moisture content much better
than the base-catalyzed process; thus, the H2 SO4 catalyst is more suitable for wet biomass feedstocks.
Compared to conventional transesterification, the direct process is not only simpler, but also improves
the FAME yield by 24–29% with a shorter reaction time and a lower energy consumption. The reaction
processes reported in this work simplify for the production of SCO-derived biodiesel and reduce its
cost. Our preliminary work also indicates that wet biomass can be used directly as a feedstock with
some minor adjustments in reaction conditions, suggesting that further process simplification and cost
reduction are possible.
Author Contributions: Chi-Yang Yu and I-Ching Kuan designed the experiments and analyzed the data.
Wei-Chen Kao and Chun-Ling Chen performed the experiments using the basic and acidic catalyst, respectively.
Chi-Yang Yu and I-Ching Kuan prepared the manuscript. All authors read and approved the final manuscript.
All authors contributed equally to this work.
Acknowledgments: We thank Hong-Wei Yen (Department of Chemical and Materials Engineering, Tunghai
University, Taiwan) and Yaw-Nan Chang (Department of Biotechnology, National Formosa University, Taiwan)
for providing R. glutinis BCRC 22360.This work was supported by the Ministry of Science and Technology (MOST
105-2621-M-036-001-MY2) in Taiwan.
References
1. Sitepu, I.R.; Garay, L.A.; Sestric, R.; Levin, D.; Block, D.E.; German, J.B.; Boundy-Mills, K.L. Oleaginous yeasts
for biodiesel: Current and future trends in biology and production. Biotechnol. Adv. 2014, 32, 1336–1360.
[CrossRef] [PubMed]
2. Canakci, M.; Sanli, H. Biodiesel production from various feedstocks and their effects on the fuel properties.
J. Ind. Microbiol. Biotechnol. 2008, 35, 431–441. [CrossRef] [PubMed]
3. Fukuda, H.; Kondo, A.; Noda, H. Biodiesel fuel production by transesterification of oils. J. Biosci. Bioeng.
2001, 92, 405–416. [CrossRef]
4. Demirbas, A. Importance of biodiesel as transportation fuel. Energy Policy 2007, 35, 4661–4670. [CrossRef]
5. Escobar, J.C.; Lora, E.S.; Venturini, O.J.; Yáñez, E.E.; Castillo, E.F.; Almazan, O. Biofuels: Environment,
technology and food security. Renew. Sustain. Energy Rev. 2009, 13, 1275–1287. [CrossRef]
6. Kulkarni, M.G.; Dalai, A.K. Waste cooking oil-an economical source for biodiesel: A review. Ind. Eng. Chem. Res.
2006, 45, 2901–2913. [CrossRef]
7. Huang, C.; Chen, X.-F.; Xiong, L.; Chen, X.-D.; Ma, L.-L.; Chen, Y. Single cell oil production from low-cost
substrates: The possibility and potential of its industrialization. Biotechnol. Adv. 2013, 31, 129–139. [CrossRef]
[PubMed]
8. Chen, C.-Y.; Yeh, K.-L.; Aisyah, R.; Lee, D.-J.; Chang, J.-S. Cultivation, photobioreactor design and harvesting
of microalgae for biodiesel production: A critical review. Bioresour. Technol. 2011, 102, 71–81. [CrossRef]
[PubMed]
9. Meng, X.; Yang, J.; Xu, X.; Zhang, L.; Nie, Q.; Xian, M. Biodiesel production from oleaginous microorganisms.
Renew. Energy 2009, 34, 1–5. [CrossRef]
10. Chisti, Y. Biodiesel from microalgae. Biotechnol. Adv. 2007, 25, 294–306. [CrossRef] [PubMed]
11. Yousuf, A.; Khan, M.R.; Islam, M.A.; Wahid, Z.A.; Pirozzi, D. Technical difficulties and solutions of direct
transesterification process of microbial oil for biodiesel synthesis. Biotechnol. Lett. 2017, 39, 13–23. [CrossRef]
[PubMed]
12. Ratledge, C.; Cohen, Z. Microbial and algal oils: Do they have a future for biodiesel or as commodity oils?
Lipid Technol. 2008, 20, 155–160. [CrossRef]
13. Yen, H.-W.; Zhang, Z. Effects of dissolved oxygen level on cell growth and total lipid accumulation in the
cultivation of Rhodotorula glutinis. J. Biosci. Bioeng. 2011, 112, 71–74. [CrossRef] [PubMed]
61
Energies 2018, 11, 1036
14. Yen, H.-W.; Chang, J.-T. Growth of oleaginous Rhodotorula glutinis in an internal-loop airlift bioreactor
by using lignocellulosic biomass hydrolysate as the carbon source. J. Biosci. Bioeng. 2015, 119, 580–584.
[CrossRef] [PubMed]
15. Kot, A.M.; Błażejak, S.; Kurcz, A.; Gientka, I.; Kieliszek, M. Rhodotorula glutinis—Potential source of lipids,
carotenoids, and enzymes for use in industries. Appl. Microbiol. Biotechnol. 2016, 100, 6103–6117. [CrossRef]
[PubMed]
16. Cheirsilp, B.; Louhasakul, Y. Industrial wastes as a promising renewable source for production of microbial
lipid and direct transesterification of the lipid into biodiesel. Bioresour. Technol. 2013, 142, 329–337. [CrossRef]
[PubMed]
17. Liu, B.; Zhao, Z. Biodiesel production by direct methanolysis of oleaginous microbial biomass.
J. Chem. Technol. Biotechnol. 2007, 82, 775–780. [CrossRef]
18. Thliveros, P.; Uçkun Kiran, E.; Webb, C. Microbial biodiesel production by direct methanolysis of oleaginous
biomass. Bioresour. Technol. 2014, 157, 181–187. [CrossRef] [PubMed]
19. Leung, D.Y.C.; Guo, Y. Transesterification of neat and used frying oil: Optimization for biodiesel production.
Fuel Process. Technol. 2006, 87, 883–890. [CrossRef]
20. Boon, C.S.; McClements, D.J.; Weiss, J.; Decker, E.A. Factors influencing the chemical stability of carotenoids
in foods. Crit. Rev. Food Sci. Nutr. 2010, 50, 515–532. [CrossRef] [PubMed]
21. Wahlen, B.D.; Willis, R.M.; Seefeldt, L.C. Biodiesel production by simultaneous extraction and conversion of
total lipids from microalgae, cyanobacteria, and wild mixed-cultures. Bioresour. Technol. 2011, 102, 2724–2730.
[CrossRef] [PubMed]
22. Vicente, G.; Bautista, L.F.; Rodríguez, R.; Gutiérrez, F.J.; Sádaba, I.; Ruiz-Vázquez, R.M.; Torres-Martínez, S.;
Garre, V. Biodiesel production from biomass of an oleaginous fungus. Biochem. Eng. J. 2009, 48, 22–27.
[CrossRef]
23. Zhang, X.; Yan, S.; Tyagi, R.D.; Surampalli, R.Y.; Valéro, J.R. Ultrasonication aided in-situ transesterification
of microbial lipids to biodiesel. Bioresour. Technol. 2014, 169, 175–180. [CrossRef] [PubMed]
24. Patil, P.D.; Gude, V.G.; Mannarswamy, A.; Cooke, P.; Munson-McGee, S.; Nirmalakhandan, N.; Lammers, P.;
Deng, S. Optimization of microwave-assisted transesterification of dry algal biomass using response surface
methodology. Bioresour. Technol. 2011, 102, 1399–1405. [CrossRef] [PubMed]
25. Canakci, M.; Van Gerpen, J. Biodiesel production via acid catalysis. Trans. ASAE 1999, 42, 1203–1210.
[CrossRef]
26. Yellapu, S.K.; Kaur, R.; Tyagi, R.D. Detergent assisted ultrasonication aided in situ transesterification for
biodiesel production from oleaginous yeast wet biomass. Bioresour. Technol. 2017, 224, 365–372. [CrossRef]
[PubMed]
27. Dai, C.-C.; Tao, J.; Xie, F.; Dai, Y.-J.; Zhao, M. Biodiesel generation from oleaginous yeast Rhodotorula glutinis
with xylose assimilating capacity. Afr. J. Biotechnol. 2007, 6, 2130–2134.
28. Bligh, E.G.; Dyer, W.J. A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 1959,
62
energies
Article
Analysis of Syngas Production from Biogas via the
Tri-Reforming Process
Rei-Yu Chein * and Wen-Hwai Hsu
Department of Mechanical Engineering, National Chung Hsing University, Taichung City 40227, Taiwan;
[email protected]
Abstract: The tri-reforming process was employed for syngas production from biogas at elevated
pressures in this study. In the tri-reforming process, air and water were added simultaneously as
reactants in addition to the main biogas components. The effects of various operating parameters
such as pressure, temperature and reactant composition on the reaction performance were studied
numerically. From the simulated results, it was found that methane and carbon dioxide conversions
can be enhanced and a higher hydrogen/carbon monoxide ratio can be obtained by increasing the
amount of air. However, a decreased hydrogen yield could result due to the reverse water–gas shift
reaction. A higher level of methane conversion and hydrogen/carbon monoxide ratio can be obtained
with increased water addition. However, negative carbon dioxide conversion could result due to
the water–gas shift and reverse carbon dioxide methanation reactions. The dry reforming reaction
resulting in positive carbon dioxide conversion can only be found at a high reaction temperature.
For all cases studied, low or negative carbon dioxide conversion was found because of carbon
dioxide production from methane oxidation, water–gas shift, and reverse carbon dioxide methanation
reactions. It was found that carbon dioxide conversion can be enhanced in the tri-reforming process
by a small amount of added water. It was also found that first-law efficiency increased with increased
reaction temperature because of higher hydrogen and carbon monoxide yields. Second-law efficiency
was found to decrease with increased temperature because of higher exergy destruction due to a
more complete chemical reaction at high temperatures.
Keywords: biogas; tri-reforming process; syngas; methane and carbon dioxide conversion;
hydrogen/carbon monoxide ratio; first-law/second-law efficiency
1. Introduction
The efficient production of syngas (a mixture of hydrogen and carbon monoxide) is gaining
significant attention worldwide as it is a versatile feedstock that can be used to produce a variety of
fuels and chemicals, such as methanol, Fischer–Tropsch fuels, H2 , and dimethyl ether (DME) [1]. Using
CH4 as the primary material, syngas can be produced from steam reforming (SR), partial oxidation
(POX), autothermal reforming (ATR), and dry reforming (DR). The tri-reforming (TR) process for
syngas production from CH4 has received growing attention because of its technical simplicity and
flexible operation [2–5]. In the TR process, the syngas is produced by combining SR, DR, and POX
in a single step. The TR process was proposed originally for syngas production from power plant
flue gas [6,7]. There are several advantages for syngas production from the TR process. As CO2 is
one of the reactants, there is no need for CO2 separation from the flue gas [6,7]. The H2 /CO ratio in
syngas can be altered by adjusting the relative amounts of the reactants. In addition, the presence of
H2 O and O2 in the feedstock helps to mitigate carbon deposition, and catalyst deactivation can be
prevented [8,9].

Energies 2018, 11, 1075
As fossil energy resources reduce sharply and environmental pollution becomes more serious,
searching for new materials for syngas production plays an important role in future energy
development [10–12]. Biogas is receiving much attention because of its considerable economic and
environmental benefits [13]. The biogas composition is related to the starting substrate, but is basically
composed of CH4 and CO2 with the volume ratio of 2 [14–16]. Both CH4 and CO2 are regarded as
major greenhouse gases (GHGs), which pose a serious threat to the global climate and environment.
Using biogas for syngas production, both CH4 and CO2 emissions into the atmosphere can be reduced.
Because of its potential for reducing global warming, further understanding of syngas production
from biogas is essential. Moreover, syngas can also be used for H2 production. In this case, H2 can
be enriched via the water–gas shift reaction using syngas and H2 O as feedstock. Among the various
alternative energy forms, hydrogen is considered an important energy carrier in the future [17]. It is
also an important raw material in the chemical industry and can be used as a fuel in fuel cells to
produce electrical energy. For reasons of sustainability, the use of renewable fuel sources such as biogas
or biomass for hydrogen production has received considerable attention [18–20].
Several studies have reported on syngas production from biogas via the TR process experimentally.
In the study of Vita et al. [21], tri-reforming simulated biogas over a Ni/ceria based catalysts was
carried out and the H2 O/CH4 and O2 /CH4 molar ratios, reaction temperature, and nickel content
effects on the catalyst’s performance were studied. They found that the H2 /CO ratio could be flexibly
adjusted using added amounts of oxygen and steam in order to meet the requirements of downstream
processes. In the study of Lau et al. [22], biogas was used as the fuel source in dry reforming and
combined dry/oxidative reforming reactions. The gas stream temperature and reactor space velocity
effects were examined experimentally. Their results indicated that an increase in the O2 /CH4 ratio at
low temperature promotes hydrogen production. In dry/oxidative reforming, they found that biogas
dry reforming is dominant and the overall reaction is net endothermic when the reaction temperature
is higher than 600 ◦ C. In the study of Zhu et al. [23], biogas reforming with added O2 through a
spark-shade plasma was conducted under an O2 / CH4 ratio of 0.60 and CO2 /CH4 ratios ranging from
0.17 to 1.00. Their results indicated that O2 and CH4 conversions decreased when the CO2 /CH4 ratio
was increased. They also reported that the partial oxidation of methane contributed mostly to CH4
conversion and the reverse water–gas shift (WGS) reaction dominated in CO2 conversion.
In addition to experimental work, several numerical TR process models using biogas as the
feedstock have also been reported in the literature. In the study of Corigliano and Fragiacomo [24],
biogas dry reforming analysis under various operating conditions was carried out using a numerical
model. The CO2 /CH4 ratio, pressure and temperature effects on reaction performance were reported.
In the study of Hernández and Martín [25], a process based on mass and energy balances, chemical
and phase equilibria, and rules of thumb was developed to optimize the production of methanol
using biogas as the raw material. Based on the production cost and carbon footprint, the optimized
CH4 /CO2 ratio contained in the biogas was found. In the study of Hajjaji et al. [26], a H2 production
system via biogas reforming was investigated using life-cycle assessment (LCA). They found that
the total GHG emissions from the system were about half of the life-cycle GHG of conventional H2
production systems via steam methane reforming. In the study of Zhang et al. [27], the effects of
various factors including reaction temperature, reactor pressure and CH4 flow rate on the syngas
compositions obtained from the TR process were investigated numerically. An optimum operating
condition for syngas production with a target ratio and maximized CO2 conversion were obtained.
In this work, the TR process is employed for syngas production using biogas as the feedstock. The
effects of various operating conditions such as pressure, temperature, biogas composition, air addition,
and H2 O additions are investigated. The novelty of this paper is the focus on CO2 conversion in the
TR process, which is seldom reported in the literature. Air is used as the added reactant in this study
instead of pure oxygen in the conventional TR process.
64
Energies 2018, 11, 1075
2. Modeling
2.1. Chemical Reaction

The following reactions are coupled and carried out in a single reactor in the TR porcess:
Steam reforming (SR):
CH4 + H2 O ↔ CO + 3H2 , ΔH0298K = +206 kJ/mole (1)
Dry reforming (DR):
CH4 + CO2 ↔ 2CO + 2H2 , ΔH0298K = +247 kJ/mole (2)
Partial oxidation (POX):
CH4 + 0.5O2 ↔ CO + 2H2 , ΔH0298K = −36 kJ/mole (3)
As shown in Equations (1)–(3), the TR process combines the endothermic SR and DR reactions
and the exothermic POX reaction. The heat released from POX is used as the heat supply for SR
and DR and makes the TR process energy efficient [28]. As noted by Cho et al. [29], the chemical
reactions involved in the TR process can be alternatively described using Equation (1) along with the
following reactions:
Reverse CO2 methanation (RCM):
CH4 + 2H2 O ↔ CO2 + 4H2 , ΔH0298K = +165 kJ/mole (4)
Water-gas shift (WGS):
CO + H2 O ↔ CO2 + H2 , ΔH0298K = −41 kJ/mole (5)
Complete oxidation of methane (COM):
CH4 + 2O2 ↔ CO2 + 2H2 O, ΔH0298K = −803 kJ/mole (6)
Note that with the chemical reactions described in Equations (1) and (4)–(6), the TR process
becomes the well-known catalytic partial oxidation of methane (CPOM). In the literature, there are
many studies devoted to the analysis of kinetic mechanisms for CPOM [30,31]. Similar to other
reforming process of CH4 , many reactions are likely to occur in the TR process. In addition to the study
of Cho et al. [29], studies of De Groote and Froment [32], Scognamiglio et al. [33], Chan and Wang [34],
and Izquierdo et al. [35] also reported that the reaction mechanism of CPOM is indirect in which the
process can be described by combining reactions of methane oxidation, methane–steam reforming,
and water–gas shift. According to these studies, reactions such as CO oxidation, H2 oxidation, and the
Boudouard reaction were not included.
Equations (1), (4) and (5) are the reactions involved in the conventional SR reaction. In this study,
syngas under high pressure is of interest for further fuel synthesis. The kinetic model for the SR
reaction over a nickel catalyst given by Xu and Froment [36] is adopted,
SR:

k1 p3H2 pCO
r1 = p p
CH4 H2 O − /DEN2 (7)
p2.5
H2
Keq,1
WGS:
k2 pH2 pCO2
r2 = pCO pH2 O − /DEN2 (8)
pH2 Keq,2
65
Energies 2018, 11, 1075
RCM:

k3 p4H2 pCO2
r3 = p p2
CH4 H2 O − /DEN2 (9)
p3.5
H2
Keq,3
KH2 O pH2 O
DEN = 1 + KCH4 pCH4 + KCO pCO + KH2 pH2 + (10)
pH2
For COM, the kinetic model of Trimm and Lam [37] is adopted in this study,
COM:
k4a pCH4 pO2 k4b pCH4 pO2
r4 = + (11)
(1 + KC + KC
2
(1 + KC
CH4 pCH4 + KO2 pO2 )
C
CH4 pCH4 O2 pO2 )
Equation (11) was derived over Pt-based catalyst support, while the model adsorption parameters
are adjusted for a Ni-based catalyst [38]. In Equations (7)–(11), ri is the reaction rate for SR (i = 1),
WGS (i = 2), RCM (i = 3), and COM (i = 4); Keq,i and ki are the chemical equilibrium constant and
rate constant for reaction i (i = 1,2,3,4); pj (j = CH4 , CO2 , H2 O, H2 , and CO) is the partial pressure
of species j; and Kj and KC j are the adsorption constants of species j. All of these kinetic parameters
are given in the Arrhenius function type and are functions of temperature, and can be found in the
literature [36,37]. It is noted that catalyst deactivation due to the thermal effect and carbon deposition
is neglected in this study [39]. For a reforming reaction involving CH4 , carbon formation is inevitable.
The carbon deposition on the catalyst surface is one of the reasons that causes catalyst deactivation.
In the tri-reforming process, the appearances of O2 and H2 O may suppress carbon formation [40,41].
Therefore, catalyst deactivation due to carbon deposition on the catalyst surface is neglected in
this study.
2.2. Process Simulation

In this study, Aspen Plus (v.10) is employed to carry out the TR process using biogas as the
feedstock. The flow process is depicted in Figure 1. The simulation is performed for a steady state.
The biogas stream is assumed to be purely composed of CH4 and CO2 with the designated molar
ratio. The air stream is composed of 21% O2 , 78% N2 , and 1% H2 . The purpose of H2 addition is to
avoid the singularity in chemical reaction rate computation. A 1% H2 addition is determined through
sensitivity analysis [42,43]. In order to produce high-pressure syngas for future use in fuel synthesis,
two compressors (COM-1 and COM-2) are used to increase the biogas and air pressures. In the H2 O
stream, a pump is used to increase the water pressure and it is then superheated in a boiler with heat
supplied from the high-temperature product stream. After mixing in a mixer, the reactant mixture
(TRI) is heated to a certain temperature before entering the insulated Rplug reactor (TR). The gas
mixture from the reactor (TRO) is sent to the boiler where the heat is recovered for superheating the
water. The TR process performance is characterized using the following dimensionless groups,
CH4 and CO2 conversions:
ni,in − ni,out
Xi = × 100%, i = CH4 , CO2 (12)
ni,in
H2 yield:
nH2 ,out − nH2 ,in
YH2 = (13)
nCH4 ,in
CO yield:
nCO,out − nCO,in
YCO = (14)
nCH4 , in
H2 /CO ratio:
YH2
H2 /CO = (15)
YCO
66
Energies 2018, 11, 1075
where ni,in is the molar flow rate of the i-th species supplied to the process; and ni,out is the molar
flow rate of the i-th species at reactor outlet. Based on these definitions, CH4 conversion is the ratio of
the CH4 consumption rate to the fed CH4 flow rate at the reactor inlet. Similarly, CO2 conversion is
the ratio of the CO2 consumption rate to the fed CO2 flow rate at the reactor inlet. The H2 and CO
yields are defined as the net increased amounts of H2 and CO from the reaction per fed CH4 flow
rate. The H2 /CO ratio is defined as the ratio of H2 yield to CO yield. Note that all these variables
are dimensionless.
BIOGAS
COM-1
TR
S1
MIXER
S5 TRI
AIR HEATER
COM-2
S2
S4
BOILER TRO
PUMP
S3
H2O
PRODUCT
Figure 1. Tri-reform process for syngas production using biogas as feedstock.
In addition to the reactant conversion and product yield, energy and exergy analyses based on the
resulting product stream flowing out of the TR reactor are also carried out. For the chemical reaction,
there are several ways to define the first- and second-law efficiencies [44,45]. Since the objective of the
TR process is to convert biogas into syngas, and noting that CO2 has zero low heating value (LHV),
the first-law efficiency is then defined as,
nCO,out LHVCO + nH2 ,out LHVH2

ηI = × 100% (16)
nCH4 ,in LHVCH4 + Wcomp + Wpump + Qheat
where Wi (i = comp, pump) and Qheat are the input work and heat input, respectively. The main heat
input occurs at the heater at which the mixed reactant is heated to a certain inlet temperature. The
exergetic analysis is carried out by considering three exergy transfers:
Exergy due to work:
ExW = Wcomp + Wpump (17)
Exergy due to heat transfer:

T0
ExQ = Qheat (1 − ) (18)
T
Exergy due to mass flow:
Exf,i = Ni {[(h − h0 ) − T0 (s − s0 )] + ∑ xk eCH

k + RT0 ∑ xk lnxk } (19)
In these equations, the subscript 0 denotes the reference state (25 ◦ C and 1 atm). The exergy due
to mass flow is contributed by physical exergy, chemical exergy and mixing exergy as shown on the
67
Energies 2018, 11, 1075
right-hand side of Equation (19). For second-law efficiency, this is generally defined as the ratio of
exergy recovered to the exergy supplied,
Exout
ηII = × 100% (20)
Exin
where Exin and Exout are the exergies supplied to and recovered from the system, respectively. Based
on Figure 1, Exin and Exout are expressed as,
Exin = Exbiogas + Exair + ExH2 O + ExQ + ExW , Exout = Exproduct (21)

The TR process using biogas as the feedstock is similar to the tri-reforming of methane (TRM).
The only difference is the CH4 and CO2 composition. We developed this work from our previous
study [46] and focused on using biogas as the feedstock. To verify the correctness of the model built
in Aspen Plus, the TRM using the reactor geometry and reactant composition reported in the studies
of Chein et al. [46] and Arab Aboosadi et al. [47] was carried out using the built model in Aspen
Plus. Figure 2 shows the comparison between the temperature and gas species distributions predicted
from a two-dimensional model [46] and from a model built in Aspen Plus. As shown in Figure 2, the
agreements for both temperature distribution shown in Figure 2a and species mole fractions shown in
Figure 2b are quite good at the reactor downstream. The discrepancies in the region near the reactor
inlet zone is believed due to the difference between one- and two-dimensional modeling. Since the
TR process performance is evaluated using results at the reactor outlet, good agreement between one-
and two-dimensional results is expected. In addition to the comparisons between numerical models,
experimental verification of the numerical model was given in our previous study [46]. Also note
that O2 is consumed rapidly as it enters the reactor shown in Figure 2b. That is, there will be no O2
available for oxidation of CO or H2 in the downstream of the reactor.
1700 0.5
(a) (b) ASPEN PLUS
2-D model [46]
ASPEN PLUS
1600 2-D model [46] H 2O
0.4
1500
H2
molr fraction
0.3
T (K)
1400
CO2
0.2
1300
0.1 CO
1200
CH4 O2
1100 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
reactor length (m) reactor length (m)
Figure 2. Comparison of results predicted from Aspen Plus and two-dimensional model [46] using the
optimized reactant composition reported by Arab Aboosadi et al. [47]. Tin = 1100 K, p = 20 atm, and
reactant composition CH4 /CO2 /H2 O/O2 = 1/1.3/2.46/0.47. (a) Temperature and (b) species mole
fraction variations along the reactor center line.
Based on the comparisons discussed above, the model built in Aspen Plus can be correctly
extended to the TR process using biogas as the feedstock. The base operating conditions are listed
in Table 1. The parameters listed in Table 1 are adopted from our previous study except for the
68
Energies 2018, 11, 1075
feedstock composition and total volumetric flow rate [46]. The molar ratio of the reactants are chosen
as CH4 :CO2 :Air:H2 O = 1:0.5:2:1 and the volume flow rate is fixed as 0.0723 L/min. As compared with
the previous study [46], a higher volume flow is used in this study because of the presence of N2 in the
air. For economy, air is added instead of pure oxygen. The advantage of this is to avoid the cost of
oxygen separation from air, but this obviously results in increased reactor volume. In the following,
the TR process performance is examined using reactant inlet temperature Tin as the primary parameter.
The effects of various pressures, catalyst weight/volume flow rate (W/F) ratios, CO2 /CH4 ratio in
biogas, amounts of air and H2 O on TR process performance are discussed.
Table 1. Reactor geometry and base operation conditions [46,47].
Parameter Value
Reactor length, L 2m
Reactor diameter, d(=2Rb ) 10 mm
Inlet pressure, pin 20 atm
Inlet temperature, Tin 300~1000 ◦ C
Reactant flow rate, F 0.0723 L min−1
Molar ratio of biogas CH4 :CO2 :Air:H2 O 1:0.5:2:1
Catalyst Ni/Al2 O3
Catalyst size, dp 0.42 mm
Catalyst weight, W 0.25 g
W/F ratio 0.0576 ghL−1
Heat-transfer condition Adiabatic
Figure 3 shows the TR process performance using the base operations listed in Table 1. In Figure 3a
the temperature variation along the reactor length for Tin = 900 ◦ C is shown. Due to the methane
oxidation reaction, the maximum temperature occurs in the near entrance region. The energy produced
from methane oxidation is used for steam reforming and dry reforming in the reactor downstream.
This causes the temperature to decrease along the reactor length. The CH4 and CO2 conversions are
shown in Figure 3b. The abrupt increase in CH4 conversion occurs at Tin = 550 ◦ C. This indicates that
Tin should be higher than 550 ◦ C in order to activate the catalyst. With temperature higher than 550 ◦ C,
CH4 conversion increases gradually with increased Tin . In Figure 3b, negative CO2 conversion results
for the low Tin regime. From the TR process chemical reactions, CO2 is produced by the methane
oxidation and WGS reactions and consumed by the dry reforming reaction. For low Tin , the WGS
reaction is dominated and CO2 consumption by DR is low. This results in negative CO2 conversion.
However, positive CO2 conversion can result when Tin becomes higher than 700 ◦ C, indicating that
DR is active. DR contributes to increase the H2 and CO yield in the high Tin regime. Figure 3b also
indicates that a complex interaction between CO2 consumption and production reactions occurs for Tin
in the 500~600 ◦ C range. The conversions of CH4 and CO2 from an equilibrium TR process obtained
from an Aspen Plus simulation are also shown in Figure 3b using the parameters listed in Table 1.
Since the results from the equilibrium process can be regarded as the theoretical limit of the reaction, it
can be seen that CH4 conversion from the catalytic reaction is lower than that from the equilibrium
reaction. Due to more CO2 production, lower CO2 conversion results from the equilibrium reaction.
For Tin higher than 800 ◦ C, CO2 conversion from the equilibrium reaction is higher than that from
the catalytic reaction. From Figure 3c, the H2 yield, CO yield and H2 /CO ratio are shown. It can be
seen that when Tin is lower than 500 ◦ C, the H2 and CO yields are very low due to inactive catalytic
reactions at low temperatures. In this low Tin regime, CO yield is much lower than H2 yield and
results in a high H2 /CO ratio. As Tin is higher than 550 ◦ C, the H2 /CO ratio decreases with Tin slowly
with a value close to 2. The decrease in H2 /CO with Tin is due to increased CO production from
the DR reaction while H2 decreases due to the reverse WGS reaction. In Figure 3d, the H2 yield, CO
yield, and H2 /CO ratio from the equilibrium TR process are also shown. As with conversions of CH4
and CO2 shown in Figure 3b, both yields of H2 and CO from the catalytic reaction are lower than
69
Energies 2018, 11, 1075
the equilibrium reaction. At high temperature, the H2 /CO ratio from both equilibrium and catalytic
reactions is about the same. In Figure 3d, the first- and second-law efficiencies are shown. Based on
Equation (16), the first-law efficiency depends on the H2 and CO yields. Because of higher H2 and CO
yields at higher Tin , ηI increases with increased Tin . However, the variation in ηII is opposite that of ηI.
Increased Tin implies that the chemical reaction is more complete towards the product side. Since the
chemical reaction is a highly irreversible process, high exergy destruction due to the chemical reaction
is expected. This results in decreased ηII as Tin increases. For the low Tin regime, exergy destruction
due to the chemical reaction is low because of low catalytic activity. Moreover, the contributions of
exergy destruction from compressors, pump, heaters and mixers are small. This leads to high ηII in the
low Tin regime.
1500 5
(a) (c) catalytic reaction
equilibrium reaction
4 H2/CO
1250
Yield, H2/CO
3
T( C)
1000
o
2
H2
CO
750
1
500 0
0 0.5 1 1.5 2 300 400 500 600 700 800 900 1000
reactor length (m) Tin(oC)
100 100
(b) (d)
CH4 90 K
75
80
50 70
conversion (%)
efficiency (%)
60
25 CO2
50
0 40
30 K
-25
20
-50
catalytic reaction 10
equilibrium reaction
0
-75 300 400 500 600 700 800 900 1000
300 400 500 600 700 800 900 1000
Tin(oC) Tin (oC)
Figure 3. Performance of tri-reforming (TR) process obtained using the base operation conditions listed
in Table 1. (a) Temperature variation along reactor with Tin = 900 ◦ C; (b) CH4 and CO2 conversions;
(c) H2 yield, CO yield, and H2 /CO ratio; and (d) First- and second-law efficiencies.
In Figure 4, the variation of species mole fraction of the TR process using the base operation
conditions listed in Table 1 is shown. It can be seen that significant mole fraction variation can be
found when Tin is higher than 500 ◦ C. The mole fractions of reactants (CH4 , CO2 , H2 O, O2 , and N2 )
decrease while the mole fractions of products (CO and H2 ) increase as Tin increases. Due to a highly
active methane oxidation reaction, O2 is consumed completely when Tin is greater than 550 ◦ C. Also
note that the variation trend of mole fractions of CO and H2 are similar to the yields of CO and H2
presented in Figure 3c. The yields of H2 and CO are used to characterize the TR process performance
in this study.
70
Energies 2018, 11, 1075
0.35
H2
0.3
N2
0.25
mole fraction 0.2

H2O
0.15 CO
0.1
CO2
0.05
CH4
O2
0
300 400 500 600 700 800 900 1000
O
T ( C)
Figure 4. Variations of species mole fraction of the TR process using the base operation conditions
listed in Table 1.
In the following, parametric studies based on the base operation conditions listed in Table 1 are
carried out. As listed in Table 1, the inlet temperature of the reactant is the primary parameter and
the amount of CH4 fed is used as the reference for the species contained in the reactant and product.
The catalyst weight (W) is varied from 0.025 g to 2.5 g; the operation pressure (P) is varied from 10
to 30 atm; the air/CH4 ratio is varied from 1 to 3; H2 O/CH4 is varied from 1 to 3; and CO2 /CH4 is
varied from 0.25 to 0.75.
In Figure 5 the effect of W/F ratios on the TR process is examined. The results shown in Figure 4
were obtained by varying the catalyst weight, while other parameters listed in Table 1 were kept fixed.
That is, higher W/F ratio results when the catalyst weight is increased. As shown in Figure 5a, a higher
temperature along the reactor length is obtained for the W/F = 0.00576 ghL−1 case. This indicates
that a smaller amount of energy released from methane oxidation reaction is used for endothermic SR
and DR reactions. For W/F = 0.0576 and 0.576 ghL−1 cases, temperature variations are identical at
the reactor downstream. That is, there is a limiting W/F ratio for the reaction. Increasing the W/F
ratio (either increasing catalyst weight or decreasing reactant volumetric flow rate) may not lead to
further improved reaction performance. In Figure 5b, CH4 and CO2 conversions are shown. Due
to low catalyst activity, CH4 conversion is low when Tin is low. It can be seen that the Tin at which
CH4 conversion abruptly increases can be decreased by increasing the W/F ratio. That is, the catalyst
activation temperature can be lowered with increased W/F ratio. As shown in Figure 5b, the Tin at
which CH4 conversion increases abruptly are 700 ◦ C, 500 ◦ C and 400 ◦ C for W/F = 0.00576, 0.0576,
and 0.576 gLh−1 , respectively. CH4 conversions for the W/F = 0.0576 and 0.576 gLh−1 cases become
identical when Tin is higher than 550 ◦ C. As discussed above, limited performance results when the
W/F ratio is increased. From Figure 5b, CO2 conversion has a negative value except in the high Tin
regime. This is due to CO2 formation in the methane oxidation and WGS reactions while DR is less
active. At high temperatures, CO2 is consumed via the dry reforming reaction, leading to positive
CO2 conversion. In Figure 5c, the H2 /CO ratios for various W/F ratios are shown. It can be seen that
H2 /CO ratio is about the same for the three W/F ratios studied when Tin is high. The H2 /CO ratio
close to a value of 2 can be obtained for the W/F range studied. In Figure 5d, variations in ηI and ηII
are shown. It can be seen that ηI can be enhanced by increasing the W/F ratio. However, ηII decreases
when the W/F ratio is increased because of a more complete chemical reaction.
71
Energies 2018, 11, 1075
1500 5
(a) (c)
4
W/F=0.00756 ghL-1
1250
W/F=0.00576 ghL-1 0.0576
0.576
3
H2/CO
T ( C)
1000 0.0576
o
0.576 2
750
1
500 0
0 0.5 1 1.5 2 300 400 500 600 700 800 900 1000
o
reactor length (m) Tin ( C)
100 100
(b) (d)
90 0.576 0.0576 0.00576
K,,
75
CH4 80
-1
70
50 0.0576 W/F=0.00576 ghL
0.576
conversion (%)
efficiency (%)
60
25 CO2 50
40
0.0576
0 0.576
30 W/F=0.00576 ghL-1
20
-25 K,
0.00576
0.0576 10
0.576
-50 0
300 400 500 600 700 800 900 1000 300 400 500 600 700 800 900 1000
Tin (oC) o
Tin ( C)
Figure 5. Effect of catalyst weight/volume flow rate (W/F) ratios on the TR process. Catalyst weight
is varied from 0.025 to 2.5 g while the other parameters were kept unchanged, as listed in Table 1.
(a) Temperature variation along reactor with Tin = 900 ◦ C; (b) CH4 and CO2 conversions; (c) H2 /CO
ratio; and (d) first- and second-law efficiencies.
In Figure 6, the reactor operating pressure effect on TR process performance is examined. From
Figure 6a the highest temperature increases with increased operating pressure. As the high temperature
in the near-entrance region of the reactor is due to the methane oxidation reaction, this implies that
methane oxidation can be enhanced by increasing the operating pressure. Due to the enhanced
methane oxidation reaction, the Tin at which an abrupt increase in CH4 conversion occurs can be
decreased by increasing the pressure, as shown in Figure 6b. Figure 6b also shows that CH4 conversion
can be increased in the low Tin regime when the pressure is increased. That is, increased operating
pressure can enhance catalyst activity at lower temperatures. In the high Tin regime, CH4 conversion
is slightly decreased as the pressure is increased. Although higher CH4 conversion can be obtained
from lower pressure operations, the resulting syngas may not be suitable for further use because
most applications involve high-pressure synthetic processes. A H2 /CO ratio with a value close to 2 is
obtained for all the pressures studied when Tin is high, as shown in Figure 6c. Because of the reduced
CH4 conversion at a high Tin regime, it can be seen that ηI decreases with increased Tin and pressure,
as shown in Figure 6d. However, Figure 6d shows that ηII increases with decreasing pressure because
of less exergy destruction by the chemical reaction.
72
Energies 2018, 11, 1075
1500 5
(a) (c)
4
1250
p=10 atm
3
p=10 atm 20 atm
H2/CO
T ( C)
1000 20 atm
o
30 atm 30 atm
2
750
1
500 0
0.5 1 1.5 2 600 700 800 900 1000
o
100 100
(b) (d) 10 atm
30 atm K,,
20 atm
75
75
30 atm
conversion (%)
efficiency (%)
50
20 atm
p=10 atm
50
25 10 atm 30 atm
20 atm 20 atm
p=10 atm
30 atm
25
0
K,
-25 0
300 400 500 600 700 800 900 1000 300 400 500 600 700 800 900 1000
o o
Tin ( C) Tin ( C)
Figure 6. Effect of pressure on the TR process. Pressure is varied from 10 to 30 atm while the other
parameters were kept unchanged, as listed in Table 1. (a) Temperature variation along reactor with
Tin = 900 ◦ C; (b) CH4 and CO2 conversions; (c) H2 /CO ratio; and (d) first- and second-law efficiencies.
The variation in reactant composition effect on TR process performance is examined in the

following. Figure 7 shows the air amount effect. Figure 7a shows that temperature can be increased
using more air as the reactant. That is, a more complete methane oxidation reaction is achieved when
the air supply is increased. With the increase in air amount, both CH4 and CO2 conversions can be
enhanced, as shown in Figure 7b. For the Air/CH4 = 3 case, 100% CH4 conversion can be reached
for Tin higher than 550 ◦ C. Due to the increased energy supply, dry reforming can occur in the lower
Tin regime resulting in increased CO2 conversion. However, negative CO2 conversion is still found
when Tin is low. Although more N2 is also introduced, increasing the volumetric flow rate of the entire
reactant, it does not affect CH4 and CO2 conversions. As shown in Figure 7c, a H2 /CO ratio with a
value higher than 2 can be obtained for the Air/CH4 = 3 case because DR is more active when the
temperature is high. For the Air/CH4 = 1 case, the H2 /CO value is lower than 2. This is due to the
reverse WGS reaction at high temperatures, reducing the H2 amount. Because of decreased H2 yield,
lower ηI in the higher Tin regime is obtained, as shown in Figure 7d. The reverse WGS reaction also
causes ηII to increase with Tin in the high Tin regime.
73
Energies 2018, 11, 1075
2000 5
(a) (c)
1750
4
3
1500
3
H2/CO
T(oC)
1250 2
3
2 2
1000
Air/CH4=1
Air/CH4=1 1
750
500 0
0 0.5 1 1.5 2 600 700 800 900 1000
o
100 (b) 100

3 (d)
2 90
1
75 K 3 2
80
Air/CH4=1
50 70
3 2
conversion (%)
CH4
efficiency (%)
3 60
25
2 Air/CH4=1
50
1
0
40
-25 CO2 30 K
20
-50
10
-75 0
300 400 500 600 700 800 900 1000 300 400 500 600 700 800 900 1000
o o
Tin ( C) Tin ( C)
Figure 7. Air/CH4 ratio effect on the TR process. The Air/CH4 ratio is varied from 1 to 3 while the other
parameters are kept unchanged, as listed in Table 1. (a) Temperature variation along the reactor with
Figure 8 shows the H2 O amount effect on TR process performance. With increased H2 O in the
reaction, lower temperature results at the reactor entrance region, as shown in Figure 8a, because of an
increased reactant volumetric flow rate and endothermic SR reaction. The increased H2 O amount does
not affect CH4 conversion, as shown in Figure 8b. However, more negative CO2 conversion results. In
addition to CO2 produced from methane oxidation, CO2 may also be produced from WGS and RCM
reactions, as indicated in Equations (4) and (5) when H2 O is increased. As shown in Figure 8c, a higher
H2 /CO ratio is obtained when H2 O is increased because of increased H2 yield. Figure 8d shows lower
ηI results when the H2 O amount is increased. This is because higher heating to the reactant is required
when the H2 O amount is increased. ηII increases with increased H2 O amount, indicating that less
exergy destruction results as H2 O is increased.
Figure 9 shows the amount of CO2 contained in the biogas effect on the TR process. As shown in
Figure 9a, the amount of CO2 does not affect the reaction temperature to a large extent. The temperature
increases slightly as the CO2 amount is decreased. As shown in Figure 9b, CH4 conversion is affected
insignificantly by the CO2 amount. However, CO2 conversion is always negative for the CO2 /CH4 =
0.25 case. That is, more CO2 is produced as a result of SR and WGS reactions than that consumed by
DR and reverse WGS reactions. In Figure 9c, higher H2 /CO results when CO2 is decreased. This may
be due to less CO formed from CO2 conversion. As with CH4 conversion, the CO2 amount effect on
first- and second-law efficiencies is not significant, as shown in Figure 9d.
74
Energies 2018, 11, 1075
1500 5
(a) (c)
4
1250
3
3
2 3 2
H2O/CH4=1
H2/CO
T (oC)
1000
H2O/CH4=1
2
750
1
500 0
0 0.5 1 1.5 2 600 700 800 900 1000
o
100 100
(b)
(d)
90
K,,
75 3
80
1 2
50 2 3 70 1
conversion (%)
efficiency (%)
CH4 60
25 K,
H2O/CH4=1
50 H2O/CH4=1
0 2
40
2 3
-25 30
CO2
3
20
-50
10
-75 0
300 400 500 600 700 800 900 1000 300 400 500 600 700 800 900 1000
o o
Tin ( C) Tin ( C)
Figure 8. Effect of H2 O/CH4 ratios on the TR process. H2 O/CH4 is varied from 1 to 3 while the other
parameters were kept unchanged, as listed in Table 1. (a) Temperature variation along reactor with
1500 5
(a) (c)
4
1250
3
CO2/CH4=0.25 0.5 0.75
H2/CO
T (oC)
1000 CO2/CH4=0.25
2 0.5
0.75
750
1
500 0
0 0.5 1 1.5 2 600 700 800 900 1000
o
100 100
(b) (d) 0.5 0.75
K,,
75
CO2/CH4=0.25 CH4
80
0.5
50 CO2/CH4=0.25
0.75
conversion (%)
efficiency (%)
0.75 60
25
0.5
0
40
0.25
-25 K,,
CO2
20
-50
-75 0
300 400 500 600 700 800 900 1000 300 400 500 600 700 800 900 1000
o o
Tin ( C) Tin ( C)
Figure 9. Effect of CO2 /CH4 ratios on the TR process. CO2 /CH4 is varied from 0.25 to 0.75 while the
other parameters were kept unchanged, as listed in Table 1. (a) Temperature variation along reactor with
75
Energies 2018, 11, 1075
Typical H2 and CO yield results are shown in Figure 10 for various air and H2 O amounts.
In Figure 10a, the H2 and CO yields increase with the increased air added in the reactant. As a high
temperature results in the Air/CH4 = 3 case, the H2 yield decreases with increased Tin due to the
reverse WGS reaction. Figure 10b shows that H2 yield can be enhanced by increased H2 O addition.
CO yield also decreases with H2 O addition because of inactive DR and reverse WGS reactions. As a
result, a higher H2 /CO ratio is obtained, as shown in Figure 8b.
2.5 (a) 2.5 (b)
3
2
2 2 2
1
3
1.5 1.5
1
Yield
Yield
3 H2O/CH4=1
1 1
2 2 3
H2
Air/CH4=1 CO
0.5 0.5
H2
CO
0 0
300 400 500 600 700 800 900 1000 300 400 500 600 700 800 900 1000
o o
Tin ( C) Tin ( C)
Figure 10. H2 and CO yields for various (a) air and (b) H2 O amounts added.
From the results shown above, the CO2 conversion is low or negative (CO2 production). Positive
CO2 conversion only occurs in the high Tin regime. It is, then, desirable to determine the way to
enhance CO2 conversion in the TR process. After several sets of numerical experiments, it was found
that high CO2 conversion can be obtained when H2 O is low. In this case, the TR process approaches the
dry reforming of methane (DRM). Figure 11 shows TR process performance with H2 O/CH4 = 0.001.
As shown in Figure 11a, the temperature drop occurs in the region very near the entrance because
DRM is a highly endothermic reaction. Large amounts of required heat leads to this temperature
drop. When the methane oxidation becomes active, energy release causes a temperature increase in
the reactor downstream. In Figure 11b, CH4 and CO2 conversions are shown for the H2 O/CH4 =
0.001 case. CO2 conversion is always positive and increases with increased Tin . The CO2 conversion is
lower than that of CH4 because of a low CO2 /CH4 ratio in the biogas. Because of small amounts of
H2 O, the H2 /CO ratio is close to unity, which is the stoichiometric H2 /CO ratio of DRM, as shown
in Figure 11c. Figure 11d shows that ηI increases with increased Tin because of higher H2 and CO
yield. The second-law efficiency states that ηII decreases with increased Tin because of higher exergy
destruction when the chemical reaction is more complete.
76
Energies 2018, 11, 1075
1250 5
(a) (c)
1000
3
H2/CO
T (oC)
2
750
500 0
0 0.5 1 1.5 2 600 700 800 900 1000
o
100 100
(b) (d)
90
CH4 K,,
80 80
70
K,
conversion (%)
efficiency (%)
60 CO2 60
50
40 40
30
20 20
10
0 0
300 400 500 600 700 800 900 1000 300 400 500 600 700 800 900 1000
o o
Tin ( C) Tin ( C)
Figure 11. TR process with H2 O/CH4 = 0.001 while the other parameters were kept unchanged,
as listed in Table 1. (a) Temperature variation along reactor with Tin = 900 ◦ C; (b) CH4 and CO2
conversions; (c) H2 /CO ratio; and (d) first- and second-law efficiencies.
4. Conclusions
The tri-reforming process was used in this study for syngas production from biogas. The effects
of various operating parameters such as pressure, temperature and reactant composition were studied
based on a model built in Aspen Plus. Based on the results obtained, the following conclusions can
be drawn:
(1) There appears to be a limiting space velocity for the reaction. Beyond this limiting value, the
reaction approaches the same performance. Lowering the reaction pressure could lead to higher
CH4 conversion, but the syngas produced may not be suitable for further applications.
(2) CH4 and CO2 conversions can be enhanced by increasing the amount of air in the reactant. Higher
amounts of air could result in decreased H2 yield due to the reverse water–gas shift reaction,
which is favorable at high reaction temperatures.
(3) A higher H2 /CO ratio can be obtained by increasing H2 O addition. However, the dry reforming
reaction is suppressed, leading to low CO2 or negative conversion.
(4) Dry reforming of CO2 can only be found when the reaction temperature is high. This results in
positive CO2 conversion and contributes to increased H2 and CO yields.
(5) Higher CO2 conversion can be obtained for the low H2 O addition case. However, low H2 /CO
with a value close to unity results.
(6) The first-law efficiency increases with the increased reaction temperature because of higher H2
and CO yields. The second-law efficiency decreases with the increased temperature because of
higher exergy destruction due to a more complete chemical reaction at high temperatures.
77
Energies 2018, 11, 1075
Author Contributions: This paper is a result of the collaboration between all co-authors. Rei-Yu Chein conceived
the idea, designed the study, and wrote the paper. Wen-Hwai Hsu established the simulation model and
data analysis.
Acknowledgments: Financial support from the Ministry of Science and Technology of Taiwan (MOST
106-2221-E-005-073-MY3) is acknowledged.
Abbreviation
Nomenclature
dp catalyst particle diameter, m
Ex exergy, kJ
eCH chemical exergy, kJ mol−1
F reactant volumetric flow rate, m3 s−1
Kj surface adsorption equilibrium constant of species j, Pa−1
KCj surface adsorption equilibrium constant of species j in combustion reaction, Pa−1
Keq,i equilibrium constant of reaction i
ki rate constant of reaction i, mol Pa0.5 kgcat s−1 , or mol Pa kgcat s−1
L length of reactor, m
LHV lower heating value, kJ mol−1
Ni total molar flow rate of a stream i, mole s−1
nj molar flow rate of species j, mole s−1
p pressure, Pa
Q heat transfer, W
R universal gas constant, 8.314 J mol−1 K−1
Rb reactor radius, m
ri kinetic rate of reaction i, mol kgcat s−1
s entropy, kJ mol−1 K−1
T temperature, K
W catalyst weight, g
Wcomp compressor work, W
Wpump pump work, W
X species conversion
x mole fraction
Y species yield
Subscript
in inlet
out outlet
0 reference state
Greek symbols
ΔH heat of reaction, kJ/mol
η efficiency
References
1. Raju, A.S.K.; Park, C.S.; Norbeck, J.M. Synthesis gas production using steam hydrogasification and steam
reforming. Fuel Process. Technol. 2009, 90, 330–336. [CrossRef]
2. Rathod, V.P.; Shete, J.; Bhale, P.V. Experimental investigation on biogas reforming to hydrogen rich syngas
production using solar energy. Int. J. Hydrog. Energy 2016, 41, 132–138. [CrossRef]
3. Su, B.; Han, W.; Jin, H. Proposal and assessment of a novel integrated CCHP system with biogas steam
reforming using solar energy. Appl. Energy 2017, 206, 1–11. [CrossRef]
4. Damanabi, A.T.; Bahadori, F. Improving GTL process by CO2 utilization in tri-reforming reactor and
application of membranes in Fischer-Tropsch reactor. J. CO2 Util. 2017, 21, 227–237. [CrossRef]
5. Dwivedi, A.; Gudi, R.; Biswas, P. An improved tri-reforming based methanol production process for
enhanced CO2 valorization. Int. J. Hydrog. Energy 2017, 42, 23227–23241. [CrossRef]
78
Energies 2018, 11, 1075
6. Song, C.; Pan, W. Tri-reforming of methane: A novel concept for catalytic production of industrially useful
synthesis gas with desired H2 /CO ratios. Catal. Today 2004, 98, 463–484. [CrossRef]
7. Majewski, A.J.; Wood, J. Tri-reforming of methane over Ni@SiO2 catalyst. Int. J. Hydrog. Energy 2014, 39,
12578–12585. [CrossRef]
8. Kumar, N.; Shojaee, M.; Spivey, J.J. Catalytic bi-reforming of methane: From greenhouse gases to syngas.
Curr. Opin. Chem. Eng. 2015, 9, 8–15. [CrossRef]
9. Choudhary, V.R.; Mondal, K.C.; Choudhary, T.V. Oxy-CO2 reforming of methane to syngas over
CoOx/MgO/SA-5205 catalyst. Fuel 2006, 85, 2484–2488. [CrossRef]
10. Weiland, P. Biogas production: Current state and perspectives. Appl. Microbiol. Biotechnol. 2010, 85, 849–860.
[CrossRef] [PubMed]
11. Nicoletti, G.; Arcuri, N.; Nicoletti, G.; Bruno, R. A technical and environmental comparison between
hydrogen and some fossil fuels. Energy Convers. Manag. 2015, 89, 205–213. [CrossRef]
12. Zhao, X.; Zhou, H.; Sikarwar, V.S.; Zhao, M.; Park, A.A.; Fennell, P.S.; Shen, L.; Fan, L. Biomass-based
chemical looping technologies: The good, the bad and the future. Energy Environ. Sci. 2017, 10, 1885–1910.
[CrossRef]
13. Samson, R.; LeDuy, A. Biogas production from anaerobic digestion of spirulina maxima algal biomass.
Biotechnol. Bioeng. 2012, 24, 1919–1924. [CrossRef] [PubMed]
14. Rasi, S.; Veijanen, A.; Rintala, J. Trace compounds of biogas from different biogas production plants. Energy
2017, 32, 1370–1380. [CrossRef]
15. Hagman, L.; Blumenthal, A.; Eklund, M.; Svensson, N. The role of biogas solutions in sustainable biorefineries.
J. Clean. Prod. 2018, 172, 3982–3989. [CrossRef]
16. Meyer, A.K.P.; Ehimen, E.A.; Holm-Nielsen, J.B. Future European biogas: Animal manure, straw and grass
potentials for a sustainable European biogas production. Biomass Bioenergy 2018, 111, 154–164. [CrossRef]
17. Molino, A.; Larocca, V.; Chianese, S.; Musmarra, D. Biofuels production by biomass gasification: A review.
18. Chianese, S.; Loipersböck, J.; Malits, M.; Rauch, R.; Hofbauer, H.; Molino, A.; Musmarra, D. Hydrogen from
the high temperature water gas shift reaction with an industrial Fe/Cr catalyst using biomass gasification
tar rich synthesis gas. Fuel Process. Technol. 2015, 132, 39–48. [CrossRef]
19. Molino, A.; Migliori, M.; Blasi, A.; Davoli, M.; Marino, T.; Chianese, S.; Catizzone, E.; Giordano, G. Municipal
waste leachate conversion via catalytic supercritical water gasification process. Fuel 2017, 206, 155–161.
[CrossRef]
20. Chianese, S.; Fail, S.; Binder, M.; Rauch, R.; Hofbauer, H.; Molino, A.; Blasi, A.; Musmarra, D. Experimental
investigations of hydrogen production from CO catalytic conversion of tar rich syngas by biomass
gasification. Catal. Today 2016, 277, 182–191. [CrossRef]
21. Vita, A.; Pino, L.; Cipitì, F.; Laganà, M.; Recupero, V. Biogas as renewable rawmaterial for syngas production
by tri-reforming process over NiCeO2 catalysts: Optimal operative condition and effect of nickel content.
22. Lau, C.S.; Tsolakis, A.; Wyszynski, M.L. Biogas upgrade to syn-gas (H2 -CO) via dry and oxidative reforming.
Int. J. Hydrog. Energy 2011, 36, 397–404. [CrossRef]
23. Zhu, X.; Li, K.; Liu, J.; Li, X.; Zhu, A. Effect of CO2 /CH4 ratio on biogas reforming with added O2 through
an unique spark-shade plasma. Int. J. Hydrog. Energy 2014, 39, 13902–13908. [CrossRef]
24. Corigliano, O.; Fragiacomo, P. Technical analysis of hydrogen-rich stream generation through CO2 reforming
of biogas by using numerical modeling. Fuel 2015, 158, 538–548. [CrossRef]
25. Hernández, B.; Martín, M. Optimal process operation for biogas reforming to methanol: Effects of dry
reforming and biogas composition. Ind. Eng. Chem. Res. 2016, 55, 6677–6685. [CrossRef]
26. Hajjaji, N.; Martinez, S.; Trably, E.; Steyer, J.; Helias, A. Life cycle assessment of hydrogen production from
biogas reforming. Int. J. Hydrog. Energy 2016, 41, 6064–6075. [CrossRef]
27. Zhang, Y.; Zhang, S.; Benson, T. A conceptual design by integrating dimethyl ether (DME) production with
tri-reforming process for CO2 emission reduction. Fuel Process. Technol. 2015, 131, 7–13. [CrossRef]
28. Solovev, S.A.; Kurilets, Y.; Orlik, S.N. Tri-reforming of methane on structured Ni-containing catalysts.
Theor. Exp. Chem. 2012, 48, 199–205. [CrossRef]
29. Cho, W.; Song, T.; Mitso, A.; McKinnon, T.J.; Ko, G.H.; Tolsma, J.E. Optimal design and operation of a natural
gas tri-reforming reactor for DME synthesis. Catal. Today 2009, 139, 261–267. [CrossRef]
79
Energies 2018, 11, 1075
30. York, A.P.E.; Xiao, T.; Green, M.L.H. Brief overview of the partial oxidation of methane to synthesis gas.
Top. Catal. 2003, 22, 345–358. [CrossRef]
31. Horn, R.; Williams, K.A.; Degenstein, N.J.; Bitsch-Larsen, A.; Dalle Nogare, D.; Tupy, S.A.; Schmidt, L.D.
Methane catalytic partial oxidation on autothermal Rh and Pt foam catalysts: Oxidation and reforming
zones, transport effects, and approach to thermodynamic equilibrium. J. Catal. 2007, 249, 380–393. [CrossRef]
32. De Groote, A.M.; Froment, D. Simulation of the catalytic partial oxidation of methane to syngas. Appl. Catal.
A 1996, 138, 245–264. [CrossRef]
33. Scognamiglio, D.; Russo, L.; Maffettone, P.L.; Salemme, L.; Simeone, M.; Crescitelli, S. Modeling temperature
profiles of a catalytic autothermal methane reformer with nickel catalyst. Ind. Eng. Chem. Res. 2009, 48,
1804–1815. [CrossRef]
34. Chan, S.H.; Wang, M.H. Thermodynamic analysis of natural gas fuel processing for fuel cell applications.
Int. J. Hydrog. Energy 2000, 25, 441–449. [CrossRef]
35. Izquierdo, U.; García-García, I.; Gutierrez, Á.M.; Arraibi, J.R.; Barrio, V.L.; Cambra, J.F.; Arias, P.L. Catalyst
deactivation and regeneration processes in biogas tri-reforming process. The effect of hydrogen sulfide
addition. Catalysts 2018, 8, 12. [CrossRef]
36. Xu, J.; Froment, G.F. Methane steam reforming, methanation and water-gas shift: I. Intrinsic kinetics. AIChE J.
1989, 35, 88–96. [CrossRef]
37. Trimm, D.L.; Lam, C.W. The combustion of methane on platinum-alumina fibre catalysts. I. Kinetics and
mechanism. Chem. Eng. Sci. 1908, 35, 1405–1413. [CrossRef]
38. De Smet, C.R.H.; de Croon, M.H.J.M.; Berger, R.J.; Marin, G.B.; Schouten, J.C. Design of adiabatic fixed-bed
reactors for the partial oxidation of methane to synthesis gas. Chem. Eng. Sci. 2001, 56, 4849–4861. [CrossRef]
39. Bartholomew, C.H. Mechanisms of catalyst deactivation. Appl. Catal. A Gen. 2001, 212, 17–60. [CrossRef]
40. Velasco, J.A.; Fernandez, C.; Lopez, L.; Cabrera, S.; Boutonnet, M.; Jaras, S. Catalytic partial oxidation of
methane over nickel and ruthenium based catalysts under low O2 /CH4 ratios and with addition of steam.
Fuel 2015, 153, 192–201. [CrossRef]
41. Ahmed, S.; Lee, S.H.E.; Ferrandon, M.S. Catalytic steam reforming of biogas e Effects of feed composition
and operating conditions. Int. J. Hydrog. Energy 2015, 40, 1005–1015. [CrossRef]
42. Barbieri, G.; Di Maio, F.P. Simulation of the methane steam re-forming process in a catalytic Pd-membrane
reactor. Ind. Eng. Chem. Res. 1997, 36, 2121–2127. [CrossRef]
43. Kim, J.H.; Choi, B.H.; Yi, J. Modified simulation of methane steam reforming in Pd-membrane/pack-bed
type reactor. J. Chem. Eng. Jpn. 1999, 32, 760–769. [CrossRef]
44. Lutz, A.E.; Bradshaw, R.W.; Keller, J.O.; Witmer, D.E. Thermodynamic analysis of hydrogen production by
steam reforming. Int. J. Hydrog. Energy 2003, 28, 159–167. [CrossRef]
45. Simpson, A.P.; Lutz, A.E. Exergy analysis of hydrogen production via steam methane reforming. Int. J.
Hydrog. Energy 2007, 32, 4811–4820. [CrossRef]
46. Chein, R.; Wang, C.; Yu, C. Parametric study on catalytic tri-reforming of methane for syngas production.
Energy 2017, 118, 1–17. [CrossRef]
47. Arab Aboosadi, Z.; Jahanmiri, A.H.; Rahimpour, M.R. Optimization of tri-reformer reactor to produce
synthesis gas for methanol production using differential evolution (DE) method. Appl. Energy 2011, 88,
2691–2701. [CrossRef]
80
energies
Article
Liquid Lipase-Catalyzed Esterification of Oleic Acid
with Methanol for Biodiesel Production in the
Presence of Superabsorbent Polymer: Optimization
by Using Response Surface Methodology
Hoang Chinh Nguyen 1,† , Dinh Thi My Huong 2,3,† , Horng-Yi Juan 3,† , Chia-Hung Su 3, *
and Chien-Chung Chien 3
1 Faculty of Applied Sciences, Ton Duc Thang University, Ho Chi Minh City 700000, Vietnam;
[email protected]
2 Faculty of Chemical Engineering, University of Technology and Education—The University of Danang,
Danang City 550000, Vietnam; [email protected]
3 Department of Chemical Engineering, Ming Chi University of Technology, New Taipei City 24301, Taiwan;
[email protected] (H.-Y.J.); [email protected] (C.-C.C.)
† These authors contributed equally to this work.
Abstract: Liquid lipase-catalyzed esterification of fatty acids with methanol is a promising process
for biodiesel production. However, water by-product from this process favors the reverse reaction,
thus reducing the reaction yield. To address this, superabsorbent polymer (SAP) was used as a
water-removal agent in the esterification in this study. SAP significantly enhanced the conversion
yield compared with the reaction without SAP. The lipase-catalyzed esterification in the presence
of SAP was then optimized by response surface methodology to maximize the reaction conversion.
A maximum conversion of 96.73% was obtained at a temperature of 35.25 ◦ C, methanol to oleic
acid molar ratio of 3.44:1, SAP loading of 10.55%, and enzyme loading of 11.98%. Under these
conditions, the Eversa Transform lipase could only be reused once. This study suggests that the
liquid lipase-catalyzed esterification of fatty acids using SAP as a water-removal agent is an efficient
process for producing biodiesel.
Keywords: biodiesel; esterification; liquid lipase; superabsorbent polymer; response surface methodology
1. Introduction
Extensive energy consumption and environmental pollution have stimulated the development
of renewable energy sources. Biodiesel, a renewable fuel derived from vegetable oil, is increasingly
considered a promising alternative to petrodiesel because of its superior combustion properties,
compatibility with diesel engines, and environmental benefits [1–4]. Therefore, biodiesel is being
produced globally to reduce the consumption of petrodiesel.
Biodiesel is commonly produced from edible feedstocks such as soybean, sunflower, and rapeseed
oils [5–8]; however, the use of these feedstocks for biodiesel production is restricted because of their
high cost (which accounts for 75% of the production cost) and competition with demand for the
food supply [9–11]. Therefore, inedible and waste oils have been developed as potential feedstocks
for biodiesel production [12–15]. These inedible and waste oils usually contain a high level of free
fatty acids, which must be esterified into biodiesel before the transesterification [16,17]. In recent
years, the esterification of fatty acids for biodiesel production has been widely investigated [18–20].
The common method is acid-catalyzed esterification [16,17,21]. Although biodiesel is successfully

Energies 2018, 11, 1085
produced from fatty acids through acid catalysis, this process retains several drawbacks such as
equipment corrosion and negative environmental effects [10,22,23]. Esterification using lipase as a
biocatalyst is considered a promising alternative for biodiesel production, because this method is
ecofriendly and proceeds at mild reaction conditions, thus reducing the energy consumption and
adverse environmental effects [10,24,25]. To improve the stability and reusability of lipase, immobilized
lipases have been developed and used for the reaction [10,14,24]. However, the rate of the reaction
catalyzed by immobilized lipase is relatively low because of the mass transfer limitation between the
enzyme and substrate [26,27]. Moreover, the high cost of immobilized lipase is the main drawback
that limits its industrial application [28].
Liquid lipase formulations have increasingly attracted attention as a promising alternative to
immobilized lipase for industrial applications because of their low cost (30- to 50-fold lower than
that of immobilized lipase) and high catalytic activity [29–31]. Recent studies have shown that liquid
lipases can be used for biodiesel production with high yield [26,29,32,33]. However, high water content
(from the feedstock and produced from the esterification of fatty acid and methanol) favors the reverse
reaction, thus lowering the reaction rate and production yield [32,34]. Efforts have been made to
remove the water from the reaction mixture, including the use of a molecular sieve, alumina, or silica
gel as adsorbents [35–37]. Although these adsorbents efficiently remove water from the reaction
solution, they cannot prevent the inactivation of lipase caused by water [37]. Superabsorbent polymer
(SAP) has been widely used for soil water conservation, sewage treatment, mineral dewatering, and
drug drying [37,38]. SAP demonstrates rapid water absorption and high water retention capacity [37]
and has been employed to remove water formed during the transesterification of corn oil and dimethyl
carbonate [37]. However, no report has mentioned the use of SAP as a water-removal agent for the
esterification process; this is an attractive research direction.
This study examined the potential use of SAP as a water-removal agent in the esterification of
fatty acid with methanol when using liquid lipase for biodiesel production. Oleic acid was used as a
model substrate, because it is one of the most common fatty acids in plant oils and animal fats [34].
Response surface methodology (RSM) was employed to analyze the effects of reaction conditions
(temperature, reaction time, SAP loading, and enzyme loading) on the reaction conversion. Liquid
lipase was also studied for its reusability.
2.1. Materials
Eversa Transform lipase (liquid lipase produced by Thermomyces lanuginosus) with activity of
100,000 PLU/g was obtained from Novozymes A/S (Bagsvaerd, Denmark). The SAP was provided
by Formosa Plastic Corp. (Kaohsiung, Taiwan). The SAP used in this study is mainly produced from
sodium polyacrylate and its properties are absorption capacity (0.9% NaCl) of 60 g/g, centrifuge
retention capacity (0.9% NaCl) of 38 g/g, and particle size distribution of 470 μm. Oleic acid (99%)
was provided by Showa Chemical Industry Co., Ltd. (Tokyo, Japan). Methanol, ethanol, and other
reagents were analytical grade and obtained from Echo Chemical Co. Ltd. (Miaoli, Taiwan).
2.2. Effect of SAP on the Esterification

A comparative study conducted lipase-catalyzed esterification of oleic acid with methanol with
and without the presence of SAP (5%, w/w) to investigate the effects of SAP on reaction conversion.
The reaction was initiated by adding 10% Eversa Transform lipase into reaction mixtures containing
methanol and oleic acid at a molar ratio of 3:1 and various amounts of water (0–30%, w/w). The reaction
was subsequently kept at 35 ◦ C with stirring for 150 min. The sample was regularly withdrawn for
determination of the reaction conversion.
The amount of oleic acid during the esterification was determined using a previously reported
procedure [39]. A sample was withdrawn from the reaction mixture, weighed, and dissolved in a
82
Energies 2018, 11, 1085
20-mL ethanol–diethyl ether solution (1:1, v/v). The sample was subsequently titrated against 0.1 M
KOH using phenolphthalein as the indicator to determine the acid value (AV). The reaction conversion
was then calculated as follows [40]:
AV1 − AV2
Reaction conversion (%) = × 100 (1)
AV1
where AV1 is the initial acid value, and AV2 is the acid value after esterification.
2.3. Optimization of Esterification Using RSM

A four-level and four-factorial central composite design was used to study the effects of reaction
factors on the reaction conversion. Esterifications with different reaction temperatures (30–50 ◦ C),
methanol:oleic acid molar ratios (1:1–9:1), SAP loadings (5–15%), and enzyme loadings (5–15%) were
performed in 100-mL screw-cap glass bottles with stirring for 150 min. After the reaction, a sample was
withdrawn from the reaction mixture to determine the reaction conversion. A quadratic equation was
then used to establish the relationship between the determined reaction conversion and reaction factors:
Y = β 0 + β 1 X1 + β 2 X2 + β 3 X3 + β 4 X4 + β 11 X12 + β 22 X22 + β 33 X32 + β 44 X42 + β 12 X1 X2

+ β 13 X1 X3 + β 14 X1 X4 + β 23 X2 X3 + β 24 X2 X4 (2)
+ β 34 X3 X4
where Y is the reaction conversion; X1 is the reaction temperature; X2 is the molar ratio of methanol
to oleic acid; X3 is the SAP loading; X4 is the enzyme loading; β0 is the regression coefficient for the
intercept term; β1 –β4 are linear parameters; β12 , β13 , β14 , β23 , β24 , and β34 are interaction parameters;
and β11 , β22 , β33 , and β44 are quadratic parameters. These parameters were determined using
the least-squares method [41], and an empirical model was subsequently employed to determine
the optimal reaction conditions for maximizing reaction conversion [41]. Minitab 16 (Minitab Inc.,
State College, PA, USA) was employed to develop the empirical model, perform analysis of variance
(ANOVA), and determine the optimal reaction conditions.
2.4. Enzyme Reuse

Eversa Transform lipase was reused in the esterification in the presence of SAP. The reaction was
conducted under the optimal conditions determined through RSM. After the reaction was completed,
the reaction mixture was centrifuged for phase separation. The oil phase was collected for the reaction
conversion determination, and the water phase containing liquid lipase was subsequently remixed
with fresh reactants and SAP to start a new reaction.
3.1. Effect of SAP on the Reaction Conversion

This study compared liquid lipase-catalyzed esterification of oleic acid with methanol in various
water contents with and without the presence of SAP. As shown in Figure 1, the conversion of the
reaction without SAP increased when the amount of water increased. This indicated that a certain
amount of water is required for the activity of liquid lipase. Nevertheless, a further increase in water
content caused a significant decrease in the conversion of the reaction without SAP. This is attributed
to high water content driving the reaction equilibrium to the reverse reaction [35,36,42] and decreasing
the activity of liquid lipase [32,35,37]. This result agrees with those of other studies [29,32,37]. Studies
have reported that a minimal amount of water is required for optimal enzyme activity, but excess
water adversely affects enzyme activity and stability, thus reducing the reaction conversion [32,35,37].
To overcome this obstacle, SAP was used as a water-removal agent for the reaction. The results showed
that the conversion of the reaction with SAP increased and reached the highest conversion when
83
Energies 2018, 11, 1085
increasing the water content from 0% to 5%. Remarkably, higher water contents resulted in no loss
in reaction conversion, indicating that excess water had no negative effect on the conversion of the
reaction with SAP. This is attributed to the efficient water absorption of the SAP that reduced the
negative effects caused by water [37]. In biodiesel production, the feedstock always contains various
amount of water; the presence of water in feedstock is the main concern, because it can reduce the
conversion yield [32,43,44]. The feedstock is thus treated to remove water before being used for the
reaction [44–46]. However, based on the results of this study, the water-removal can be eliminated by
adding SAP directly into the reaction solution. Therefore, the liquid lipase-catalyzed esterification
using SAP as a water-removal agent is a promising process for biodiesel synthesis from feedstock
containing high water content.
100
90
80
5HDFWLRQFRQYHUVLRQ
70
60
50
40
30
20
0 5 10 15 20 25 30
:DWHUDPRXQW
Figure 1. Effects of SAP on the esterification with varied water content. The reaction was conducted
under the following conditions: molar ratio of methanol to oleic acid of 3:1, a temperature of 35 ◦ C,
enzyme loading of 10%, in the presence of 5% SAP (•), or without SAP ().
3.2. The RSM Model Development

A central composite RSM model was employed to establish the relationships between reaction
conversion (measured response) and reaction factors (input variables)—temperature, molar ratio
of methanol to oleic acid, SAP loading, and enzyme loading. Table 1 shows the input variables
with their coded and un-coded values. Table 2 illustrates the experimental design for obtaining the
optimal reaction conditions. Based on the results shown in Table 2, the measured responses and
input variables in term of code values were input into an empirical model as the following quadratic
polynomial equation:
Y = 89.5 − 6.6X1 − 6.39X2 + 1.26X3 + 5.9X4 − 3.77X12 − 4.53X22 − 3.45X32 − 3.75X42

+ 0.86X1 X2 + 1.21X1 X3 − 0.26X1 X4 − 1.99X2 X3 + 0.15X2 X4 (3)
− 0.21X3 X4
where X3 , X4 , X1 X2 , X1 X3 , and X2 X4, have positive effects on the response, and the other parameters
have adverse effects.
84
Energies 2018, 11, 1085
Table 1. Coded values of the input variables for the central composite RSM design.
Variable Levels
Variables Symbols
−2 −1 0 1 2
Temperature (◦ C) X1 30 35 40 45 50
Methanol:oleic acid molar ratio X2 1 3 5 7 9
SAP loading (%) X3 5 7.5 10 12.5 15
Enzyme loading (%) X4 5 7.5 10 12.5 15
Table 2. Experimental design for the influences of the four independent variables on the reaction
conversion in coded values and experimental results.
Variable
Run Response, Y
X1 X2 X3 X4
1 1 1 1 1 66.87
2 1 −1 1 1 88.13
3 −2 0 0 0 85.83
4 1 1 −1 1 68.91
5 0 0 2 0 77.68
6 0 2 0 0 60.76
7 −1 −1 1 1 96.20
8 1 −1 −1 1 74.42
9 −1 −1 −1 1 95.88
10 0 0 0 2 81.21
11 1 1 1 −1 55.43
12 1 −1 1 −1 70.65
13 −1 1 −1 −1 65.83
14 −1 −1 −1 −1 80.98
15 −1 −1 1 −1 85.09
16 1 1 −1 −1 55.23
17 1 −1 −1 −1 64.56
18 0 −2 0 0 79.16
19 2 0 0 0 60.15
20 −1 1 1 1 78.20
21 0 0 0 −2 64.98
22 0 0 −2 0 70.88
23 −1 1 −1 1 83.59
24 −1 1 1 −1 65.38
25 0 0 0 0 90.92
26 0 0 0 0 89.90
27 0 0 0 0 88.79
28 0 0 0 0 90.92
29 0 0 0 0 87.92
30 0 0 0 0 88.36
31 0 0 0 0 89.72
Repeated experiments based on the central runs (25–31) showed a low coefficient of variance
(1.33%), indicating the high reproducibility and precision of the experiments. The model was evaluated
for statistical significance using the F test for ANOVA (Table 3). Results showed a very low p value
(<0.0001) of the model in the F test, confirming that the regression was statistically significant at the
95% confidence level. The coefficient of determination (R2 ) was determined to evaluate the quality
of the developed model. Result showed that a high R2 value (0.97) was achieved, signifying high
reliability of the model for predicting reaction conversion. As shown in Figure 2, the model predictions
were in good agreement with experimental values, indicating that the established model provided
satisfactory and accurate results. Table 4 presents the overall effects of the input variables on the
reaction conversion, which were examined using t tests. Low p values (<0.05) of the intercept term,
85
Energies 2018, 11, 1085
three linear terms (X1 , X2 , and X4 ), all quadratic terms, and an interaction term (X2 X3 ) indicated their
significant effects on the reaction. The developed model can therefore be used to forecast the optimal
reaction conditions for obtaining maximal responses.
Table 3. Analysis of variance for the empirical model.
Source DF b SS b MS b F Value Probability (P) > F

Model a 14 4319.58 308.54 32.91 <0.0001
Residual (error) 16 150.01 9.38 - -
Total 30 4469.59 - - -
aCoefficient of determination (R2 ) = 0.97; adjusted R2 = 0.94. b DF, degree of freedom; SS, sum of squares; MS,
mean square.
100
3UHGLFWHGUHDFWLRQFRQYHUVLRQ
90
80
70
60
50
50 60 70 80 90 100
$FWXDOUHDFWLRQFRQYHUVLRQ
Figure 2. Correlation between experimental and fitted conversions of the reaction.
Table 4. Significance of the coefficients in the empirical model.
Model Term Parameter Estimate Standard Error t Value a p Value

β0 89.50 1.16 77.34 0.000 b
β1 −6.60 0.63 −10.55 0.000 b
β2 −6.39 0.63 −10.22 0.000 b
β3 1.26 0.63 2.01 0.062
β4 5.90 0.63 9.43 0.000 b
β11 −3.77 0.57 −6.59 0.000 b
β22 −4.53 0.57 −7.91 0.000 b
β33 −3.45 0.57 −6.03 0.000 b
β44 −3.75 0.57 −6.54 0.000 b
β12 0.86 0.77 1.13 0.275
β13 1.21 0.77 1.58 0.133
β14 −0.26 0.77 −0.34 0.740
β23 −1.99 0.77 −2.61 0.019 b
β24 0.15 0.77 0.19 0.850
β34 −0.21 0.77 −0.27 0.788
a tα/2,n-p = t0.025,19 = 2.093. b p < 0.05 indicates that the model terms are significant.
86
Energies 2018, 11, 1085
3.3. Effect of Reaction Factors on Reaction Conversion

Figure 3 shows the effects of the methanol:oleic acid molar ratio and SAP loading on the reaction
conversion while maintaining temperature and enzyme loading at their central levels. Results showed
a significant interaction between the methanol:oleic acid molar ratio and SAP loading. At high
methanol:oleic acid molar ratios, SAP loading slightly affected the reaction conversion. However, at a
low methanol:oleic acid molar ratio, increasing the SAP loading significantly enhanced the reaction
conversion. This is attributed to the SAP absorbing water from the reaction mixture and enhancing the
enzyme activity to create a suitable microenvironment for an efficient reaction [37]. However, a higher
SAP loading resulted in a significant decrease in reaction conversion. Because liquid lipase requires
a minimal amount of water for optimal activity, high SAP loading caused less water content to be
present the reaction solution [35,37]. Consequently, the enzyme activity was reduced, leading to a
reduction in reaction conversion.
90
RQ
80
5HDFWLRQFRQYHUVL
70
60
50
40 14
12

1 2 10
J
3 4 8
LQ
0ROD 5
DG
UUDWL 6 6
RRI
OR
UHDFW 7
8
3
DQWV 9
6$
Figure 3. Response surface plot of combined effects of the reactant molar ratio and SAP loading on the
conversion of the reaction at a constant temperature (40 ◦ C) and enzyme loading (10%).
Figure 4 presents a response surface curve obtained by plotting the reaction conversion against
temperature and molar ratio of methanol to oleic acid while maintaining the other factors at their
central levels. At any temperature, the reaction conversion significantly increased when increasing
the methanol:oleic acid molar ratio. This was because a high amount of methanol is required for
esterification [20,47,48]. However, the reaction conversion decreased with a further increase in the
methanol:oleic acid molar ratio. This result is similar to that of other studies [10,14]. Studies have
reported that high methanol content in the reaction mixture can inhibit the activity of lipase [10,14].
In this study, the maximal reaction conversion was obtained at a methanol:oleic acid molar ratio of 3.34:1.
Figure 5 presents the effects of temperature and enzyme loading on the reaction conversion while
maintaining methanol:oleic acid molar ratio and SAP loading at their central levels. At any enzyme
loading level, the reaction conversion increased when increasing the temperature. Nevertheless, a high
temperature resulted in a decrease in the reaction conversion because the enzyme becomes inactive
at high temperatures [14,32,49]. This result agrees with that of other studies [32,33]. Studies have
demonstrated that lipases are sensitive to temperature, and therefore low or elevated temperatures
caused a dramatic decrease in their activity [14,33,49]. Similar to temperature, the enzyme loading
also significantly affected the reaction conversion. At a given temperature, increasing enzyme loading
led to a significant increase in reaction conversion. Studies have reported that an increase in the
amount of the enzyme increased contact between the enzyme active surface area and the reactants,
87
Energies 2018, 11, 1085
thus enhancing the reaction [14,50]. However, a further increase in enzyme loading resulted in a slight
decrease in reaction conversion. Excess enzyme possibly caused enzyme aggregation, which limited
the enzyme flexibility to react with the oleic acid–methanol interface, thus lowering the conversion of
the reaction [14,50].
90
RQ
80
70
5HDFWLRQFRQYHUVL
60
50
WV
40
WDQ
6
DF
30
UH
35 4
RI
40
WLR
7HP 2
SHUD 45
UD
WXUH
R&
DU
50 RO
0
Figure 4. Response surface plot of the combined effects of temperature and reactant molar ratio on the
conversion of the reaction at a constant SAP loading (10%) and enzyme loading (10%).
90
RQ
80
70
5HDFWLRQFRQYHUVL
60
50
40 14
12

30 10
QJ
35 8
GL
40
RD
7HP 6
HO
SHUD 45
WXUH
P
R& 50
]\
(Q
Figure 5. Response surface plot of the combined effects of temperature and enzyme loading on the
conversion of the reaction at a constant reactant molar ratio (5:1) and SAP loading (10%).
3.4. Obtaining Optimal Reaction Conditions

Based on the empirical model [Equation (3)], the canonical method was used to forecast the
optimal reaction conditions for maximizing reaction conversion. The highest conversion was predicted
to be 97.83% at 35.25 ◦ C, methanol:oleic acid molar ratio of 3.34:1, SAP loading of 10.55%, and enzyme
loading of 11.98%. An experiment was then carried out under optimal conditions to verify the
prediction. A conversion of 96.73% ± 0.15% was obtained, indicating consistent with the empirical
88
Energies 2018, 11, 1085
model prediction. The developed RSM model can be therefore used to describe the relationships
between the response and the variables in the liquid lipase-catalyzed esterification of oleic acid with
methanol. Furthermore, the reaction conversion was comparable with other process but exhibited a
shorter reaction time [25,29]. In this study, high reaction conversion (96.73%) was obtained under the
reaction time of 2.5 h whereas 24 h was required to yield conversion of 90.8% in the esterification of
oleic acid with methanol using lipase without SAP reported by Rosset et al. [25]. This suggests that the
liquid lipase-catalyzed esterification of fatty acids with methanol using SAP as a water-removal agent
is a promising process for producing biodiesel.
3.5. Reusability of Liquid Lipase

Although immobilized lipase is being used for its stability and reusability, its high cost restricts
its industrial application. To solve this concern, liquid lipase was developed as an alternative for
the reaction. Liquid lipase can be reused several times without significant loss in activity [29]. This
study investigated the reusability of Eversa Transform lipase through the esterification of oleic acid
and methanol under the optimal reaction conditions. As indicated in Figure 6, the Eversa Transform
lipase could only be reused once to drive the reaction to high conversion. After one cycle, the reaction
conversion decreased sharply. This was attributed to the inactivation effect of methanol on the
enzyme [10,14] and the low stability of liquid enzyme [51]. Further investigation is required to address
this limitation. Although liquid lipase demonstrated low reusability in this study, the enzyme remains
a promising alternative to immobilized lipase for industrial application because of its low cost [33,52].
Studies have demonstrated that the cost of liquid lipase is 30- to 50-fold lower than that of immobilized
lipase [30]. Additionally, the preparation of the liquid lipase is much simpler. Because of these merits,
liquid lipase is suggested as a potential alternative for the reaction to improve economic viability.
100
90
80
5HDFWLRQFRQYHUVLRQ
70
60
50
40
30
20
0 1 2 3 4 5
&\FOHQXPEHU
Figure 6. Reusability of liquid lipase for esterification of oleic acid and methanol.
4. Conclusions
This paper reports the liquid lipase-catalyzed esterification of oleic acid with methanol using SAP
as a water-removal agent for biodiesel production. The use of SAP significantly enhanced the reaction
conversion by suppressing the reverse reaction. The reaction conditions were then optimized to obtain
maximal conversion yield using RSM. A maximal conversion of 96.73% was achieved and verified
the optimization calculations. Although Eversa Transform lipase was ineffectively reused due to the
deactivation of enzyme caused by methanol, liquid lipase-catalyzed esterification in the presence of
SAP is promising for biodiesel production from feedstocks containing high water content.
89
Energies 2018, 11, 1085
Author Contributions: Hoang Chinh Nguyen and Chia-Hung Su conceived and designed the experiments;
Dinh Thi My Huong and Horng-Yi Juan performed the experiments; Hoang Chinh Nguyen and Chien-Chung Chien
analyzed the data; Chia-Hung Su contributed reagents and materials; Hoang Chinh Nguyen and Chia-Hung Su
wrote the paper.
Acknowledgments: The authors thank for the financial support from the Ministry of Science and Technology
(MOST) of Taiwan, R.O.C, under grant number 106-2221-E-131-028.
References
1. Da Silva Araújo, F.D.; do Nascimento Cavalcante, A.; Sousa, M.D.D.B.; de Moura, C.V.R.; Chaves, M.H.;
Aued-Pimentel, S.; Fernandes Caruso, M.S.; Tozetto, L.J.; Kaline Morais Chaves, S. Biodiesel production
from Bombacopsis glabra oil by methyl transesterification method. Energies 2017, 10, 1360. [CrossRef]
2. Kakati, J.; Gogoi, T. Biodiesel production from Kutkura (Meyna spinosa Roxb. Ex.) fruit seed oil: Its
characterization and engine performance evaluation with 10% and 20% blends. Energy Convers. Manag. 2016,
121, 152–161. [CrossRef]
3. Anwar, M.; Rasul, M.G.; Ashwath, N. Production optimization and quality assessment of papaya
(Carica papaya) biodiesel with response surface methodology. Energy Convers. Manag. 2018, 156, 103–112.
[CrossRef]
4. Mosarof, M.; Kalam, M.; Masjuki, H.; Alabdulkarem, A.; Ashraful, A.; Arslan, A.; Rashedul, H.; Monirul, I.
Optimization of performance, emission, friction and wear characteristics of palm and Calophyllum inophyllum
biodiesel blends. Energy Convers. Manag. 2016, 118, 119–134. [CrossRef]
5. Kim, K.H.; Lee, E.Y. Simultaneous production of transformer insulating oil and value-added glycerol
carbonates from soybean oil by lipase-catalyzed transesterification in dimethyl carbonate. Energies 2017, 11,
82. [CrossRef]
6. Jin, H.; Kolar, P.; Peretti, S.W.; Osborne, J.A.; Cheng, J.J. Kinetics and mechanism of NaOH-impregnated
calcined oyster shell-catalyzed transesterification of soybean oil. Energies 2017, 10, 1920. [CrossRef]
7. Vahid, B.R.; Haghighi, M. Biodiesel production from sunflower oil over MgO/MgAl2 O4 nanocatalyst:
Effect of fuel type on catalyst nanostructure and performance. Energy Convers. Manag. 2017, 134, 290–300.
[CrossRef]
8. Saka, S.; Kusdiana, D. Biodiesel fuel from rapeseed oil as prepared in supercritical methanol. Fuel 2001, 80,
9. Poudel, J.; Shah, M.; Karki, S.; Oh, S.C. Qualitative analysis of transesterification of waste pig fat in
supercritical alcohols. Energies 2017, 10, 265. [CrossRef]
10. Nguyen, H.C.; Liang, S.H.; Doan, T.T.; Su, C.H.; Yang, P.C. Lipase-catalyzed synthesis of biodiesel from black
soldier fly (Hermetica illucens): Optimization by using response surface methodology. Energy Convers. Manag.
2017, 145, 335–342. [CrossRef]
11. Martindale, W.; Trewavas, A. Fuelling the 9 billion. Nat. Biotechnol. 2008, 26, 1068–1070. [CrossRef] [PubMed]
12. Branco-Vieira, M.; San Martin, S.; Agurto, C.; Santos, M.A.D.; Freitas, M.A.; Mata, T.M.; Martins, A.A.;
Caetano, N.S. Potential of Phaeodactylum tricornutum for biodiesel production under natural conditions in
Chile. Energies 2017, 11, 54. [CrossRef]
13. Poudel, J.; Karki, S.; Sanjel, N.; Shah, M.; Oh, S.C. Comparison of biodiesel obtained from virgin cooking oil
and waste cooking oil using supercritical and catalytic transesterification. Energies 2017, 10, 546. [CrossRef]
14. Nguyen, H.C.; Liang, S.H.; Chen, S.S.; Su, C.H.; Lin, J.H.; Chien, C.C. Enzymatic production of biodiesel from
insect fat using methyl acetate as an acyl acceptor: Optimization by using response surface methodology.
15. Nguyen, H.C.; Liang, S.H.; Li, S.Y.; Su, C.H.; Chien, C.C.; Chen, Y.J.; Huong, D.T.M. Direct transesterification
of black soldier fly larvae (Hermetia illucens) for biodiesel production. J. Taiwan Inst. Chem. Eng. 2018, 85,
16. Yang, S.; Li, Q.; Gao, Y.; Zheng, L.; Liu, Z. Biodiesel production from swine manure via housefly larvae
(Musca domestica L.). Renew. Energy 2014, 66, 222–227. [CrossRef]
17. Berchmans, H.J.; Hirata, S. Biodiesel production from crude Jatropha curcas L. seed oil with a high content of
free fatty acids. Bioresour. Technol. 2008, 99, 1716–1721. [CrossRef] [PubMed]
90
Energies 2018, 11, 1085
18. Anguebes-Franseschi, F.; Abatal, M.; Bassam, A.; Escalante Soberanis, M.A.; May Tzuc, O.; Bucio-Galindo, L.;
Cordova Quiroz, A.V.; Aguilar Ucan, C.A.; Ramirez-Elias, M.A. Esterification optimization of crude African
palm olein using response surface methodology and heterogeneous acid catalysis. Energies 2018, 11, 157.
[CrossRef]
19. Bhuyan, M.S.U.S.; Alam, A.H.M.A.; Chu, Y.; Seo, Y.C. Biodiesel production potential from littered edible
oil fraction using directly synthesized S-TiO2 /MCM-41 catalyst in esterification process via non-catalytic
subcritical hydrolysis. Energies 2017, 10, 1290. [CrossRef]
20. Chongkhong, S.; Tongurai, C.; Chetpattananondh, P.; Bunyakan, C. Biodiesel production by esterification of
palm fatty acid distillate. Biomass Bioenergy 2007, 31, 563–568. [CrossRef]
21. Hayyan, A.; Alam, M.Z.; Mirghani, M.E.; Kabbashi, N.A.; Hakimi, N.I.N.M.; Siran, Y.M.; Tahiruddin, S.
Reduction of high content of free fatty acid in sludge palm oil via acid catalyst for biodiesel production.
22. Campanelli, P.; Banchero, M.; Manna, L. Synthesis of biodiesel from edible, non-edible and waste cooking
oils via supercritical methyl acetate transesterification. Fuel 2010, 89, 3675–3682. [CrossRef]
23. Minami, E.; Saka, S. Kinetics of hydrolysis and methyl esterification for biodiesel production in two-step
supercritical methanol process. Fuel 2006, 85, 2479–2483. [CrossRef]
24. Voulgaris, S.; Papadopoulou, A.A.; Alevizou, E.; Stamatis, H.; Voutsas, E. Measurement and prediction of
solvent effect on enzymatic esterification reactions. Fluid Phase Equilib. 2015, 398, 51–62. [CrossRef]
25. Rosset, I.G.; Cavalheiro, M.C.H.; Assaf, E.M.; Porto, A.L. Enzymatic esterification of oleic acid with aliphatic
alcohols for the biodiesel production by Candida antarctica lipase. Catal. Lett. 2013, 143, 863–872. [CrossRef]
26. Andrade, T.A.; Errico, M.; Christensen, K.V. Evaluation of reaction mechanisms and kinetic parameters for
the transesterification of castor oil by liquid enzymes. Ind. Eng. Chem. Res. 2017, 56, 9478–9488. [CrossRef]
27. He, Y.; Li, J.; Kodali, S.; Balle, T.; Chen, B.; Guo, Z. Liquid lipases for enzymatic concentration of n-3
polyunsaturated fatty acids in monoacylglycerols via ethanolysis: Catalytic specificity and parameterization.
28. Alves, J.S.; Vieira, N.S.; Cunha, A.S.; Silva, A.M.; Ayub, M.A.Z.; Fernandez-Lafuente, R.; Rodrigues, R.C.
Combi-lipase for heterogeneous substrates: A new approach for hydrolysis of soybean oil using mixtures of
biocatalysts. RSC Adv. 2014, 4, 6863–6868. [CrossRef]
29. Ren, H.; Li, Y.; Du, W.; Liu, D. Free lipase-catalyzed esterification of oleic acid for fatty acid ethyl ester
preparation with response surface optimization. J. Am. Oil Chem. Soc. 2013, 90, 73–79. [CrossRef]
30. Cesarini, S.; Diaz, P.; Nielsen, P.M. Exploring a new, soluble lipase for FAMEs production in water-containing
systems using crude soybean oil as a feedstock. Process Biochem. 2013, 48, 484–487. [CrossRef]
31. Lv, L.; Dai, L.; Du, W.; Liu, D. Effect of water on lipase NS81006-catalyzed alcoholysis for biodiesel production.
Process Biochem. 2017, 58, 239–244. [CrossRef]
32. Andrade, T.A.; Errico, M.; Christensen, K.V. Influence of the reaction conditions on the enzyme catalyzed
transesterification of castor oil: A possible step in biodiesel production. Bioresour. Technol. 2017, 243, 366–374.
[CrossRef] [PubMed]
33. Adewale, P.; Vithanage, L.N.; Christopher, L. Optimization of enzyme-catalyzed biodiesel production from
crude tall oil using Taguchi method. Energy Convers. Manag. 2017, 154, 81–91. [CrossRef]
34. Pan, Y.; Alam, M.A.; Wang, Z.; Wu, J.; Zhang, Y.; Yuan, Z. Enhanced esterification of oleic acid and methanol
by deep eutectic solvent assisted Amberlyst heterogeneous catalyst. Bioresour. Technol. 2016, 220, 543–548.
[CrossRef] [PubMed]
35. Giacometti, J.; Giacometti, F.; Milin, Č.; Vasić-Rački, Đ.A. Kinetic characterisation of enzymatic esterification
in a solvent system: Adsorptive control of water with molecular sieves. J. Mol. Catal. B Enzym. 2001, 11,
36. Duan, Y.; Du, Z.; Yao, Y.; Li, R.; Wu, D. Effect of molecular sieves on lipase-catalyzed esterification of rutin
with stearic acid. J. Agric. Food Chem. 2006, 54, 6219–6225. [CrossRef] [PubMed]
37. Gu, J.; Xin, Z.; Meng, X.; Sun, S.; Qiao, Q.; Deng, H. Studies on biodiesel production from DDGS-extracted
corn oil at the catalysis of Novozym 435/super absorbent polymer. Fuel 2015, 146, 33–40. [CrossRef]
38. Ma, Z.; Li, Q.; Yue, Q.; Gao, B.; Xu, X.; Zhong, Q. Synthesis and characterization of a novel super-absorbent
based on wheat straw. Bioresour. Technol. 2011, 102, 2853–2858. [CrossRef] [PubMed]
39. Su, C.H. Recoverable and reusable hydrochloric acid used as a homogeneous catalyst for biodiesel production.
Appl. Energy 2013, 104, 503–509. [CrossRef]
91
Energies 2018, 11, 1085
40. Li, Q.; Zheng, L.; Cai, H.; Garza, E.; Yu, Z.; Zhou, S. From organic waste to biodiesel: Black soldier fly,
Hermetia illucens, makes it feasible. Fuel 2011, 90, 1545–1548. [CrossRef]
41. Arteaga, G.; Li-Chan, E.; Vazquez-Arteaga, M.; Nakai, S. Systematic experimental designs for product
formula optimization. Trends Food Sci. Technol. 1994, 5, 243–254. [CrossRef]
42. Li, S.F.; Wu, W.T. Lipase-immobilized electrospun PAN nanofibrous membranes for soybean oil hydrolysis.
Biochem. Eng. J. 2009, 45, 48–53. [CrossRef]
43. Cheng, J.; Yu, T.; Li, T.; Zhou, J.; Cen, K. Using wet microalgae for direct biodiesel production via microwave
irradiation. Bioresour. Technol. 2013, 131, 531–535. [CrossRef] [PubMed]
44. Kusdiana, D.; Saka, S. Effects of water on biodiesel fuel production by supercritical methanol treatment.
Bioresour. Technol. 2004, 91, 289–295. [CrossRef]
45. Wahidin, S.; Idris, A.; Shaleh, S.R.M. Rapid biodiesel production using wet microalgae via microwave
irradiation. Energy Convers. Manag. 2014, 84, 227–233. [CrossRef]
46. Halim, R.; Danquah, M.K.; Webley, P.A. Extraction of oil from microalgae for biodiesel production: A review.
Biotechnol. Adv. 2012, 30, 709–732. [CrossRef] [PubMed]
47. Veillette, M.; Giroir-Fendler, A.; Faucheux, N.; Heitz, M. Esterification of free fatty acids with methanol to
biodiesel using heterogeneous catalysts: From model acid oil to microalgae lipids. Chem. Eng. J. 2017, 308,
48. Berrios, M.; Siles, J.; Martin, M.; Martin, A. A kinetic study of the esterification of free fatty acids (FFA) in
sunflower oil. Fuel 2007, 86, 2383–2388. [CrossRef]
49. Köse, Ö.; Tüter, M.; Aksoy, H.A. Immobilized Candida antarctica lipase-catalyzed alcoholysis of cotton seed
oil in a solvent-free medium. Bioresour. Technol. 2002, 83, 125–129. [CrossRef]
50. Waghmare, G.V.; Rathod, V.K. Ultrasound assisted enzyme catalyzed hydrolysis of waste cooking oil under
solvent free condition. Ultrason. Sonochem. 2016, 32, 60–67. [CrossRef] [PubMed]
51. Rotková, J.; Šuláková, R.; Korecká, L.; Zdražilová, P.; Jandová, M.; Lenfeld, J.; Horák, D.; Bílková, Z. Laccase
immobilized on magnetic carriers for biotechnology applications. J. Magn. Magn. Mater. 2009, 321, 1335–1340.
[CrossRef]
52. Nielsen, P.; Rancke-Madsen, A.; Holm, H.; Burton, R. Production of biodiesel using liquid lipase formulations.
J. Am. Oil Chem. Soc. 2016, 93, 905–910. [CrossRef]
92
energies
Article
Valorization of Waste Wood as a Solid Fuel
by Torrefaction
Jeeban Poudel 1 , Sujeeta Karki 2 and Sea Cheon Oh 2, *
1 Waste & Biomass Energy Technology Center, Kongju National University, 1223-24 Cheonan-Daero, Seobuk,
Chungnam 330-717, Korea; [email protected]
2 Department of Environmental Engineering, Kongju National University, 1223-24 Cheonan-Daero, Seobuk,
Chungnam 330-717, Korea; [email protected]
Received: 8 May 2018; Accepted: 15 June 2018; Published: 23 June 2018
Abstract: The aim of this study was to investigate the optimal temperature range for waste wood
and the effect torrefaction residence time had on torrefied biomass feedstock. Temperature range
of 200–400 ◦ C and residence time of 0–50 min were considered. In order to investigate the effect of
temperature and residence time, torrefaction parameters, such as mass yield, energy yield, volatile
matter, ash content and calorific value were calculated. The Van Krevelen diagram was also used for
clarification, along with the CHO index based on molecular C, H, and O data. Torrefaction parameters,
such as net/gross calorific value and CHO increased with an increase in torrefaction temperature,
while a reduction in energy yield, mass yield, and volatile content were observed. Likewise, elevated
ash content was observed with higher torrefaction temperature. From the Van Krevelen diagram,
it was observed that at 300 ◦ C the torrefied feedstock came in the range of lignite. With better gross
calorific value and CHO index, less ash content and nominal mass loss, 300 ◦ C was found to be the
optimal torrefaction temperature for waste wood.
Keywords: waste wood; torrefaction; energy yield; mass yield; CHO index; gross calorific value;
Van Krevelen diagram
1. Introduction
In recent years, biomass has obtained remarkable attention because of the potential it holds to
replace the energy derived from fossil fuel. Biomass is considered to be an important renewable fuel
and the most widespread technology, and can be grouped into thermochemical (torrefaction, pyrolysis,
combustion, etc.), chemical (alkaline hydrolysis, etc.) and biochemical (fermentation, anaerobic
digestion, etc.) categories [1]. Biomass can be considered as a flexible source of energy as it can be
transformed into numerous energy products, for instance, bio-oil, syngas, and so forth. However,
biomass has numerous challenges, such as but not limited to, high moisture content, poor grindability,
hydroscopicity, low heating value, fibrous in nature, and so forth. [2,3]. These challenges confine the
combustion performance and escalate the handling and transportation cost of biomass. Therefore,
in order to overcome these challenges, several pretreatment approaches have been suggested. Thermal
degradation of woody biomass is a complex topic in itself and comprises of numerous fractions with
various thermal behaviors [4].

Energies 2018, 11, 1641
One of the promising approaches is to pretreat the biomass in the temperature range from
200–300 ◦ C in an inert environment, described as torrefaction. Torrefaction improves the dimensional
stability and durability of the wood [5]. This advancement can be curbed by the wood’s moisture;
therefore, the reduction of the wood’s moisture content is a must. In addition, the grindability of
the wood increases with torrefaction and reduces the energy required [6]. This reduction in wood
commination serves as major advantages for numerous energy derived applications, such as co-firing,
cement kilns, steel industry, coke, and so forth. [7]. Moreover, torrefied biomasses are more suitable
for further energy conversion processes, such as fast pyrolysis [8], gasification [9], and in the steel
industry, as a substitute for coke [10]. However, Koppenjan et al. [11] highlights the need to address
numerous issues, such as energy integration, volatile gases handling and the applicability of feedstock
for the advancement of torrefaction technologies. In addition, in spite of these advantages, some
of the detrimental effects are related to torrefaction degradation of strength and toughness [12,13].
The reduction in the cohesive nature of wood is associated with the disintegration of the hemicellulose
matrix and depolymerization of cellulose during torrefaction [14]. Although not spread in a large
commercial scale, there has been an enormous amount of studies for torrefaction of biomass in the
last decade. For instance, a comprehensive study by Rodriques et al. [15] evaluates the potential
torrefaction enhancement for production of bioenergy in Portugal. Similarly, Li et al. [16] highlights
the advantages of carbon dioxide torrefaction, such as elevated thermal stability and combustion
reactivity. In addition, Benavente and Fullana [17] utilized two-phase olive mill waste (TPOMW) to
determine the optimum torrefaction temperature range, and concluded that torrefaction of TPOMW
around 200 ◦ C min−1 is best for energy utilization. The assessment of the overall product yield, right
from the production of the raw materials to the utilization of the final energy, can be carried out by
conducting Life Cycle Assessment scenarios [18].
Decisive parameters like temperature, residence time, or properties of torrefaction products, have
crucial impact on torrefaction behavior, properties of the torrefied product, and comprehensive mass
and energy conversion efficacy [19]. However, the characteristic feature of the biomass material used
for torrefaction cannot be overlooked. The main objective of the present work is to study the effect
torrefaction has on the thermochemical properties of wood. The torrefied wood properties are based
on the Gross Calorific Value (GCV) (MJ/kg), elemental composition, proximate and gas analysis,
and energy yield (Yenergy (%)) and mass yield (Ymass (%)). The result of this work is to define the
optimal torrefaction temperature range for waste wood suitable for other energy conversion processes
or energy production use.

The following sections elaborate the reactor, equations and experimental setup of the process for
this study.
2.1. Materials
In this study, waste wood was investigated. The mixed waste wood collected from a waste wood
collection center in Cheonan, South Korea was first screened, and all the non-woody and coarse parts
were removed. The waste wood was homogeneously mixed and dried at 105 ◦ C for 24 h for the
torrefaction experiment. In Table 1, the properties of waste wood indicates about 13.5% moisture
content, 1.5% ash content and 20.6 MJ/kg GCV. The inorganic components of the raw waste wood are
referred as “other” in the table.
94
Energies 2018, 11, 1641
Table 1. Properties of waste wood (raw).
C 46.96
H 5.90
N 3.07
Elemental Analysis (wt %) a O 42.30
S 0.07
Cl 0.57
Other 1.13
Cd 20.02
Pb 16.02
Heavy Metal Analysis (mg/kg) Cr 4.61
Hg ND
As ND
Moisture (%) b 13.5
Volatile fraction (%) b 85.0
Ash (%) b 1.5
GCV (MJ/kg) a 20.6
LHV (MJ/kg) b 17.81
a Dry Basis; b Wet Basis; LHV (Lower Heating Value).
2.2. Torrefaction Experiment Setup and Procedure

The schematic diagram of the horizontal tubular reactor used for this study is illustrated in
Figure 1. The internal diameter and length of the reactor were 150 mm and 60 mm, respectively.
For each experimental run, 20 g of sample were weighed and flushed with nitrogen (21 min−1 ) until
the level of oxygen was below 1%. Torrefaction was then carried out for the temperature range
of 200–400 ◦ C for a residence time of 0–50 min. After completion of each experiment, the sample
was removed from the reactor and was weighed. The initial and final weights of the samples were
measured to determine mass yield (Ymass (%)), which was calculated with Equation (1):
mass after torrefaction

Mass Yield (Ymass (%)) = × 100% (1)
mass of raw sample
Figure 1. Schematic diagram of the horizontal tubular reactor used in this study.
95
Energies 2018, 11, 1641
In addition, the GCV of the raw and torrefied biomass was measured using bomb calorimeter
(Parr Instrument Co., Model 1672, Moline, IL, USA) and the energy yield (Yenergy (%)) was calculated
using Equation (2):
HHV (torrefied sample)

Energy Yield Yenergy (%) = Ymass × × 100% (2)
HHV(raw sample)
Thermogravimetric analysis (TGA) was carried out using a thermogravimetric analyzer

(TA Instruments, Q50, New Castle, DE, USA), 5 mg of sample was placed in a crucible and the
experiment was conducted in an inert environment with a nitrogen flow rate of 60 mL·min−1 .
The temperature range was allowed to rise up to 600 ◦ C during various heating rates, ranging from
10–30 ◦ C·min−1 . The changes in the atomic composition of waste wood were investigated by using the
Van Krevelen diagram. It is constructed using the molar ratio of hydrogen:carbon (hydrogen index) as
the ordinate to the molar ratio of oxygen:carbon (oxygen index) in the abscissa. The raw and torrefied
waste wood were plotted, and the specific location of these plots helped to identify the alterations in
the atomic composition of torrefied waste wood.

In this study, parameters such as thermogravimetric analysis, effect of torrefaction temperature
and residence time on torrefied product, and heavy metal analysis were used for the determination of
optimal temperature range for torrefaction of waste wood.
3.1. Thermogravimetric Analysis

The weight loss characteristics of waste wood as a function of torrefaction temperature are
illustrated in Figure 2. It can be seen from the figure, that until 250 ◦ C, there was no significant
weight loss for all three heating rates. However, the bulk weight loss of the waste wood is in the
temperature range of 250–370 ◦ C. There was an insignificant difference in the weight loss with respect
to heating rate, i.e., although the weight loss for the heating rate of 30 ◦ C·min−1 was highest followed
by 20 ◦ C·min−1 and the lowest was for 10 ◦ C and the differences were minimal. After the main
decomposition, i.e., after 370 ◦ C, weight loss was relatively slow.
R&PLQ
R&PLQ
R&PLQ

:HLJKWIUDFWLRQD

7RUUHIDFWLRQWHPSHUDWXUHR&
Figure 2. TGA analysis of waste wood with respect to torrefaction temperature.
96
Energies 2018, 11, 1641
3.2. Effect of Temperature on Torrefaction Parameters

The effect of temperature on torrefaction parameters, such as Ymass %, Yenergy (%), volatile matter
(VMc (%)), GCV (MJ/kg), and ash content (Ac (%)), are illustrated in Figure 3. It can be seen from
the graph that overall, there is a decrease in Yenergy (%), Ymass (%), and VMc (%), with an increase in
torrefaction temperature. Conversely, there is an increase in GCV (MJ/kg), and Ac (%), with increase
in torrefaction temperature.
As from the Figure 3, it can be observed that until 300 ◦ C, the mass loss was not significant after
which there was a loss of 42.77% at 350 ◦ C. This drastic alteration of mass loss can be attributed to the
drying process and thermal degradation of stable wood constituents [7,20,21]. Similar changes can be
observed for Yenergy (%), where a significant reduction of 27.80% was observed at 350 ◦ C. In contrast,
an increment in the GCV (MJ/kg) is observed with increase in torrefaction temperature. At elevated
torrefaction temperature, VMc (%) decreases while the Ac (%) increases, the results obtained in this
study is well supported by the results obtained by Pelaez-Samaniego et al. [22] for ponderosa pine
wood species.

*URVV&DORULILF9DOXH0-NJ

3HUFHQWDJH&RQWHQWV

7RUUHIDFWLRQWHPSHUDWXUHR&

Figure 3. Effect of torrefaction temperature on Energy Yield (Yenergy (%)), Mass Yield (Ymass (%)),
Volatile Matter (VMc (%)), Ash Content (Ac (%)) and Gross Calorific Value (GCV (MJ/kg)).
The GCV (MJ/kg) is one of the most important characteristics of a solid fuel. Although with
increasing torrefaction temperature, the GCV (MJ/kg) of the biomass feedstock keeps increasing;
it must be noted that Ymass (%) decreases with elevated torrefaction temperature. The loss of mass can
reach a point where the torrefaction process is regarded as incomplete. Therefore, this study aims to
provide the optimal range of temperatures that totally benefit torrefaction. Therefore, the temperature
range of 250–350 ◦ C was taken into account to further investigating the effect of torrefaction residence
time on torrefied waste wood for this study.
Coal contains higher percentage of carbon than biomass due to which the GCV (MJ/kg) of coal is
much higher than that of biomass. The importance of the hydrogen and oxygen index on the GCV
(MJ/kg) of waste wood is illustrated in Figure 4, presenting the Van Krevelen diagram, a graphical
97
Energies 2018, 11, 1641
illustration which portrays the decomposition of waste wood during torrefaction. During the
pretreatment, waste wood undergoes physiochemical changes that involve the loss or gain of integral
amounts of elemental composition, such as C, H, N, and O, leading to specific changes in the
Van Krevelen diagram. From the Van Krevelen diagram, it can be seen that as the torrefaction
temperature increases the molar ratio, i.e., H/C and O/C ratio decreases, bringing the waste wood
closer towards lignite and coal. Figure 4 also clearly indicates, that as the temperature increases
to 300 ◦ C, the waste wood appears closer to lignite. Torrefaction temperatures above 250 ◦ C,
the Van Krevelen plot demonstrates that the elemental composition of waste wood is shifted towards
that of coal and lignite. During torrefaction, carbon dioxide and water is released making the
torrefied product more valuable for thermal treatment processes, such as gasification and combustion.
Similar observations have been made by Granados et al. [23], and have reasoned it due to the large
decomposition of cellulose and hemicellulose leaving a lignin-rich material. In addition, not many
changes are observed around 200–250 ◦ C, and the product is still in the range of biomass. Thus, from
Figures 3 and 4, the optimal torrefaction temperature for waste wood is 300 ◦ C, with elevated GCV
(MJ/kg), nominal mass loss, and less Ac (%).
Figure 4. Van Krevelen diagram of waste wood with respect to torrefaction temperature.
To describe the oxidation state of organic compound present in organic materials, Mann et al. [24]
suggested the use of the CHO index, defined as:
2[O] − [H]
CHO = (3)
[C]
where, [O], [C] and [H] are mole fraction of oxygen, carbon, and hydrogen, respectively. Figure 5a
shows the CHO index values of waste wood (current study) and for different lignocellulosic biomass
obtained from [25], and Figure 5b denotes the CHO index obtained in this study with respect to
torrefaction temperature. The higher the CHO index, the greater the number of oxygenated compounds.
Whereas, a lower CHO index refers to the lowered amount of oxygenated compounds, denoting a
relatively lower amount of oxygen and higher amount of relative hydrogen content. The increase
in torrefaction temperature leads to a decrease the amount of oxygen and hydrogen, which in turn
increases the relative amount of carbon content, which provides a good CHO index. This value can be
correlated to the Van Krevelen diagram and an increase in GCV (MJ/kg) as well.
98
Energies 2018, 11, 1641

0.2
&XUUHQW6WXG\
+\EULG3RSODU 0.1
$OPRQG6KHOO
'HPROLVKHG:RRGV
2OLYH3LWWV
0.0
CHO Index

&+2,QGH[
-0.1

-0.2

-0.3
200 250 300 350

Torrefaction temperature (oC)
(a) (b)
Figure 5. CHO index values of (a) biomass feedstocks (data adapted from Reference [21]) (b) waste
wood with respect to torrefaction temperature.
3.3. Effect of Residence Time on Torrefaction Parameters

The effect of torrefaction residence time on Yenergy (%), Ymass (%), VMc (%), Ac (%) and GCV
(MJ/kg) is shown in Table 2. Holding time or residence time can be described as the total amount of
time the biomass feedstock is inside the torrefaction reactor [26]. The residence time had a significant
effect on all of the parameters except for VMc (%), where the overall decrease at 300 ◦ C was 1.6% at
50 min residence time. Additionally, at 50 min residence time, the Ymass (%) decreased by 59% at 350 ◦ C
for 50 min and Ac (%) increased by 82% at 300 ◦ C for 50 min residence time. In addition, the GCV
(MJ/kg) showed an increment of 24.75% at 250 ◦ C for 50 min residence time.
Table 2. Effect of torrefaction residence time on Energy Yield (Yenergy (%)), Mass Yield (Ymass (%)),
Volatile Matter (VMc (%)), Ash Content (Ac (%)) and Gross Calorific Value (GCV (MJ/kg)).
0 min Residence Time

250 ◦ C 300 ◦ C 350 ◦ C
Yenergy (%) 99.58 98.10 71.85
Ymass (%) 99.88 91.70 57.16
VMc (%) 98.34 98.03 96.52
Ac (%) 1.65 1.97 3.47
GCV (MJ/kg) 21.08 22.77 26.76
250 ◦ C 300 ◦ C 350 ◦ C
Yenergy (%) 87.10 59.15 49.87
Ymass (%) 79.84 53.06 39.68
VMc (%) 97.89 97.38 96.21
Ac (%) 2.11 2.61 3.79
GCV (MJ/kg) 23.09 23.73 26.75
250 ◦ C 300 ◦ C 350 ◦ C
Yenergy (%) 70.27 56.44 43.41
Ymass (%) 62.55 47.96 33.11
VMc (%) 97.55 96.33 95.72
Ac (%) 2.45 3.67 4.27
GCV (MJ/kg) 23.77 25.06 27.92
99
Energies 2018, 11, 1641
Table 2. Cont.

250 ◦ C 300 ◦ C 350 ◦ C
Yenergy (%) 74.23 53.45 52.66
Ymass (%) 65.45 43.87 38.33
VMc (%) 96.39 95.87 95.61
Ac (%) 3.60 4.13 4.39
GCV (MJ/kg) 24.00 25.94 29.25
250 ◦ C 300 ◦ C 350 ◦ C
Yenergy (%) 82.81 57.31 54.74
Ymass (%) 74.26 45.12 39.61
VMc (%) 96.65 96.25 95.31
Ac (%) 3.35 3.75 4.68
GCV (MJ/kg) 23.59 27.04 29.42
250 ◦ C 300 ◦ C 350 ◦ C
Yenergy (%) 89.24 59.00 53.00
Ymass (%) 71.79 44.42 39.57
VMc (%) 97.41 96.39 95.74
Ac (%) 2.58 3.60 4.26
GCV (MJ/kg) 26.31 28.28 28.52
The Yenergy (%) and Ymass (%) demonstrates significant alterations at elevated residence time.
However, this conversion rate may vary depending on the type of biomass used, for instance,
agricultural residue demonstrates higher conversion rate compared to woody biomass because of the
higher hemicellulose content [27]. In addition, a study by Chen et al. [20] reflects on how the polymeric
structure of feedstock influences the reactivity of the torrefaction reaction.
Various works have shown that the effect of torrefaction temperature has a more pronounced
effect than residence time, for instance, Chin et al. [28] concluded that although the GCV (MJ/kg)
increased with both residence time and temperature, the value was more influenced by torrefaction
temperature. In contrast, the present study has found that the overall increase in the GCV (MJ/kg) is
comparable for both temperature and residence time.
3.4. Heavy Metal Analysis of Waste Wood

The heavy metal analysis of the waste wood and the torrefied products for temperature ranging
from 250 ◦ C to 400 ◦ C (Table 3). The presence of heavy metals has a potentially damaging effect
on human physiology and other biological systems, when the tolerance levels are exceeded. It is
important to state that the cadmium problem in biomass utilization is caused by the technology
itself. It is imported via deposition onto the forest. The source of this deposition is the dissipation
of cadmium by anthropogenic processes, most notably fossil energy use [29]. The Cd, Pb and Cr
content of the raw WW were 20.02, 16.02 and 4.61 mg/kg, respectively (Table 2). Hg and As were
not detected. From Table 2, it was found that the concentrations of heavy metals increased following
torrefaction. This is due to the fact that the heavy metals were relatively concentrated by reducing of
volatile components in the biomass.
100
Energies 2018, 11, 1641
Table 3. Heavy metal concentration of WW as a function of the torrefaction temperature.
Raw 250 ◦ C 300 ◦ C 350 ◦ C

Cd 20.02 24.51 94.25 103.66
Heavy Metal (mg/kg) Pb 16.02 22.17 62.14 88.52
Cr 4.61 7.34 25.65 29.44
Hg ND ND ND ND
As ND ND ND ND
ND = Not Detected.
4. Conclusions
In the present work, waste wood under the influence of torrefaction residence time and
temperature was investigated. The results showed that there was an increase in the GCV (MJ/kg) of
waste wood with the increase in torrefaction temperature and residence time. However, after 350 ◦ C,
there was a negligible increase in the GCV (MJ/kg), but the mass loss was maximum; therefore,
for 0–50 min residence time, the temperature range of 250–350 ◦ C was taken into consideration.
Comparing the present data with the Van Krevelen diagram, we found that with an increase in
torrefaction severity the torrefied biomass behaved more like lignite and coal. Also, the CHO index
showed a positive influence with an increase in torrefaction temperature, i.e., the amount of oxygenated
compound decreased. Unlike other works, the effect of both residence time and temperature were
significant on Yenergy (%), Ymass (%), VMc (%), Ac (%) and GCV (MJ/kg) of the waste wood. However,
a decisive conclusion cannot be made and more investigation considering other torrefaction parameters,
such as grindability and heat-induced variations must be studied in detail. Calculating all of the
parameters, 300 ◦ C was considered as an optimal torrefaction temperature for waste wood, i.e.,
higher GCV (MJ/kg), lesser mass loss, comparatively low Ac (%) and the properties resembling more
towards lignite.
Author Contributions: J.P. and S.K. were responsible for the overall experiment, data analysis and arrangement
along with preparation of the manuscript. S.C.O. was the supervisor of the project and was mainly responsible for
the data and manuscript confirmation. All the authors were equally responsible for finalizing the manuscript
and submission.
Funding: This work was supported by a Grant from the Human Resources Development Program
(No. 20154030200940) of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) funded by
the Ministry of Trade, Industry and Energy of the Korean government.
References
1. Biochemical Conversion of Biomass in Liquid Transportation Fuels from Coal and Biomass: Technological Status, Costs,
and Environmental Impacts; The National Academies Press: Washington, DC, USA, 2009; pp. 117–162.
2. Doddapaneni, T.R.K.C.; Konttinen, J.; Hukka, T.I.; Moilanen, A. Influence of torrefaction pretreatment on the
pyrolysis of Eucalyptus clone: A study on kinetics, reaction mechanism and heat flow. Ind. Crops Prod. 2016,
92, 244–254. [CrossRef]
3. Van der Stelt, M.; Gerhauser, H.; Kiel, J.; Ptasinski, K. Biomass upgrading by torrefaction for the production
of biofuels: A review. Biomass Bioenergy 2011, 35, 3748–3762. [CrossRef]
4. Peng, J.H.; Bi, H.T.; Lim, C.J.; Sokhansanj, S.A. Study on density, hardness, and moisture uptake of torrefied
wood pellets. Energy Fuels 2013, 27, 967–974. [CrossRef]
5. Kamdem, D.; Pizzi, A.; Jermannaud, A. Durability of heat-treated wood. Holz als Roh-und Werkstoff 2002,
60, 1–6. [CrossRef]
6. Bridgeman, T.; Jones, J.; Williams, A.; Waldron, D. An investigation of the grindability of two torrefied energy
crops. Fuel 2010, 89, 3911–3918. [CrossRef]
7. Phanphanich, M.; Mani, S. Impact of torrefaction on the grindability and fuel characteristics of forest biomass.
101
Energies 2018, 11, 1641
8. Boateng, A.; Mullen, C. Fast pyrolysis of biomass thermally pretreated by torrefaction. J. Anal. Appl. Pyrolysis
2013, 100, 95–102. [CrossRef]
9. Svoboda, K.; Pohořelý, M.; Hartman, M.; Martinec, J. Pretreatment and feeding of biomass for pressurized
entrained flow gasification. Fuel Process. Technol. 2009, 90, 629–635. [CrossRef]
10. Thrän, D.; Witt, J.; Schaubach, K.; Kiel, J.; Carbo, M.; Maier, J.; Ndibe, C.; Koppejan, J.; Alakangas, E.; Majer, S.
Moving torrefaction towards market introduction–Technical improvements and economic-environmental
assessment along the overall torrefaction supply chain through the SECTOR project. Biomass Bioenergy 2016,
89, 184–200. [CrossRef]
11. Koppejan, J.; Sokhansanj, S.; Melin, S.; Madrali, S. Status overview of torrefaction technologies.
IEA Bioenergy Task 2012, 32, 1–54.
12. Stamm, A.J. Thermal degradation of wood and cellulose. Ind. Eng. Chem. 1956, 48, 413–417. [CrossRef]
13. Stanzl-Tschegg, S.; Beikircher, W.; Loidl, D. Comparison of mechanical properties of thermally modified
wood at growth ring and cell wall level by means of instrumented indentation tests. Holzforschung 2009, 63,
14. Bergman, P.C.; Boersma, A.; Zwart, R.; Kiel, J. Torrefaction for Biomass Co-Firing in Existing Coal-Fired Power
Stations “Biocoal”; Report No.ECN-C-05-013; Energy Centre of Netherlands: Petten, The Netherlands, 2005.
15. Rodrigues, A.; Loureiro, L.; Nunes, L.J.R. Torrefaction of woody biomasses from poplar SRC and Portuguese
roundwood: Properties of torrefied products. Biomass Bioenergy 2018, 107, 55–65. [CrossRef]
16. Li, S.-X.; Zou, J.-Y.; Li, M.-F.; Wu, X.-F.; Bian, J.; Xue, Z.-M. Structural and thermal properties of Populus
tomentosa during carbon dioxide torrefaction. Energy 2017, 124, 321–329. [CrossRef]
17. Benavente, V.; Fullana, A. Torrefaction of olive mill waste. Biomass Bioenergy 2015, 73, 186–194. [CrossRef]
18. Christoforou, E.A.; Fokaides, P.A. Life cycle assessment (LCA) of olive husk torrefaction. Renew Energy 2016,
90, 257–266. [CrossRef]
19. Prins Mark, J.; Ptasinski, K.; Janssen, F. Torrefaction of wood Part 2. Analysis of products J. Anal. Appl. Pyrolysis
2006, 77, 35–40. [CrossRef]
20. Chen, W.H.; Kuo, P.C. A study on torrefaction of various biomass materials and its impact on lignocellulosic
structure simulated by a thermogravimetry. Energy 2010, 35, 2580–2586. [CrossRef]
21. Hill, S.J.; Grigsby, W.J.; Hall, P.W. Chemical and cellulose crystallite changes in Pinus radiata during
torrefaction. Biomass Bioenergy 2013, 56, 92–98. [CrossRef]
22. Pelaez-Samaniego, M.R.; Yadama, V.; Garcia-Perez, M.; Lowell, E.; McDonald, A.G. Effect of temperature
during wood torrefaction on the formation of lignin liquid intermediates. J. Anal. Appl. Pyrolysis 2014,
109, 222–233. [CrossRef]
23. Granados, D.; Basu, P.; Chejne, F.; Nhuchhen, D. Detailed investigation into torrefaction of wood in a
two-stage inclined rotary torrefier. Energy Fuels 2016, 31, 647–658. [CrossRef]
24. Mann, B.F.; Chen, H.; Herndon, E.M.; Chu, R.K.; Tolic, N.; Portier, E.F.; Chowdhury, T.R.; Robinson, E.W.;
Callister, S.J.; Wullschleger, S.D. Indexing permafrost soil organic matter degradation using high-resolution
mass spectrometry. PLoS ONE 2015, 10. [CrossRef] [PubMed]
25. Jenkins, B.; Baxter, L.; Miles, T., Jr.; Miles, T. Combustion properties of biomass. Fuel Process. Technol. 1998,
54, 17–46. [CrossRef]
26. Strandberg, M.; Olofsson, I.; Pommer, L.; Wiklund-Lindström, S.; Åberg, K.; Nordin, A. Effects of temperature
and residence time on continuous torrefaction of spruce wood. Fuel Process. Technol. 2015, 134, 387–398.
[CrossRef]
27. Bridgeman, T.G.; Jones, J.M.; Shield, I.; Williams, P.T. Torrefaction of reed canary grass, wheat straw and
willow to enhance solid fuel qualities and combustion properties. Fuel 2008, 87, 844–856. [CrossRef]
28. Chin, K.; H’ng, P.; Go, W.; Wong, W.; Lim, T.; Maminski, M.; Paridah, M.; Luqman, A. Optimization of
torrefaction conditions for high energy density solid biofuel from oil palm biomass and fast growing species
available in Malaysia. Ind. Crops Prod. 2013, 49, 768–774. [CrossRef]
29. Narodoslawsky, M.; Obernberger, I. From waste to raw material—The route from biomass to wood ash for
cadmium and other heavy metals. J. Hazard. Mater. 1996, 50, 157–168. [CrossRef]
102
energies
Review
Wastewater Treatment and Biogas Recovery Using
Anaerobic Membrane Bioreactors (AnMBRs):
Strategies and Achievements
Mohammed Ali Musa 1,2 , Syazwani Idrus 1, *, Hasfalina Che Man 1 and
Nik Norsyahariati Nik Daud 1
1 Department of Civil Engineering, Faculty of Engineering, University Putra Malaysia, Serdang 43400,
Selangor, Malaysia; [email protected] (M.A.M.); [email protected] (H.C.M.);
[email protected] (N.N.N.D.)
2 Department of Civil and Water Resources Engineering, University of Maiduguri, P.M.B. 1069,
Maiduguri 600230, Borno State, Nigeria
* Correspondence: [email protected], Tel.:+60-13-692-2301
Received: 17 May 2018; Accepted: 19 June 2018; Published: 27 June 2018
Abstract: Anaerobic digestion is one of the most essential treatment technologies applied to industrial
and municipal wastewater treatment. Membrane-coupled anaerobic bioreactors have been used
as one alternative to the conventional anaerobic digestion process. They are presumed to offer
the advantage of completely reducing or minimizing the volume of sludge and increasing biogas
production. However, researchers have consistently reported different kinds of fouling that resulted
in the reduction of membrane life span. Depending on the strength of the effluent, factors such as
high suspended and dissolved solids, fats, oil and grease, transmembrane pressure (TMP) and flux
were reported as major contributors to the membrane fouling. Moreover, extracellular polymeric
substances (EPSs) are an important biological substance that defines the properties of sludge flocs,
including adhesion, hydrophobicity and settling and have been found to accelerate membrane fouling
as well. Extensive studies of AnMBR have been done at laboratory while little is reported at the
pilot scale. The significance of factors such as organic loading rates (OLRs), hydraulic retention
time (HRT), pH and temperature on the operations of AnMBRs have been discussed. Microbial
environmental conditions also played the most important role in the production of biogas and the
chemical oxygen demand (COD) removal, but adverse effects of volatile fatty acids formation were
reported as the main inhibitory effect. Generally, evaluating the potential parameters and most
cost effective technology involved in the production of biogas and its inhibitory effects as well
as the effluent quality after treatment is technically challenging, thus future research perspectives
relating to food to microorganism F/M ratio interaction, sufficient biofilm within the reactor for
microbial attachment was recommended. For the purpose of energy savings and meeting water
quality discharge limit, the use of micro filtration was also proposed.
Keywords: anaerobic digestion; biogas production; wastewater treatment; membrane bioreactors
1. Introduction
Over the past century, there has been quite a number of studies on the most economic, efficient
and environmental friendly wastewater treatment technologies. Conventional aerobic methods have
existed for over a century now, but they have major drawbacks that include sludge production,
high energy use for aeration, large operating space and a higher maintenance cost. Moreover,
the systems are characterized by uncontrolled release of potential atmospheric greenhouse gases
such as methane (CH4 ), carbon dioxide (CO2 ), and nitrogen oxide (N2 O) that contribute immensely to

Energies 2018, 11, 1675
deterioration of the environment [1], although in the aerobic process quality effluents are produced
which comply with standards limits set by the different regulatory bodies, especially in developed
countries. However, one obvious disadvantage is the lack of material and energy recovery [2,3]. Hence,
interest continues to grow in finding the best alternative.
The advent of anaerobic digestion in the field of wastewater treatment marks the beginning of
economic and efficient technology [4]. This is seen in the quality of the effluents discharged, material
recovered and energy generated as well as the mode of sludge production, handling and processing [5].
One of the early constraints detected from the onset of this technology was the long hydraulic retention
time coupled with the slow growing methanogenic bacteria [6]. However, towards attaining higher
efficiency, research focus was moved towards coupling anaerobic bioreactors with membrane filtration
units. These systems were seen to offer a more unique and prudent technique over conventional
anaerobic method [7]. It simply combines the anaerobic process and membrane technology operating
simultaneously. Table 1 presents a comparison of conventional anaerobic treatment and anaerobic
membrane bioreactor (AnMBR).
Table 1. Comparison of conventional anaerobic treatment and anaerobic membrane bioreactor

(AnMBR) [8].
Feature Conventional Anaerobic Treatment AnMBR

Discharge quality Moderate-Poor High
Sludge volume low low
substrate loading concentration High High
Removal efficiency (Effluent) High High
Biomass retention Low Complete
Footprint High-Moderate Low
Alkalinity Requirement Depends on microbial activity Depends on microbial activity
Nutrient requirement Low Low
Startup period 2–4 Months Less than 2 Weeks
Temperature requirement Low-Moderate Low-Moderate
Energy input Low Low
Pre-treatment requirement Not necessary Mostly for high solid substrates
Biogas recovery Yes Yes
The combination of anaerobic bioreactors with membranes is presumed to reduce the overall
energy demand and to facilitate the retention of microorganisms so as to operate with high biomass
concentration [9]. Depending on the intended use of the treated effluent, membranes such as
microfiltration, ultrafiltration, nano-filtration and reverse osmosis are usually coupled to anaerobic
reactors in both industrial and municipal wastewater treatment systems [10]. Buntner et al. [11]
studied the combination of an upflow anaerobic sludge blanket (UASB) reactor with ultrafiltration
membranes for dairy wastewater treatment at ambient temperature. The intent of this combination
was to decrease the COD of the dairy wastewater whilst producing biogas rich in methane and
diminish the overall sludge production as well as obtaining high quality effluent. About 150 L/kg
COD of biogas was achieved out of which 73% was methane gas. The organic loading rate was up to
4.85 kg COD/m3 ·day and 95% COD removal was realized, reaching 99% during its stable operation
period. However, this is lower compared to the work of Deowan et al. [12]. Both studies were
reported at pilot scales, but one configured the membrane externally while the other was submerged
within the bioreactor. The aim of this review paper is to gain insight knowledge on the strategies
and achievements recorded in the recent past on the use of AnMBR technology, since the idea was
presented as a possible solution or substitute that might overcome the disadvantages of conventional
wastewater treatment in terms of space utilization, superior effluent quality, energy generation and the
overall operation and maintenance cost.
104
Energies 2018, 11, 1675
2. Fundamentals of AnMBR
An anaerobic reactor coupled with a membrane was first commercialized by Dorr-Oliver in the
early 1980s [13]. It was constructed to treat high strength wastewater (whey) and was named anaerobic
membrane bioreactor system. Since then, research on the feasibility and treatment efficiencies of
AnMBRs continue, most especially on the type of material and configurations for treating low, medium,
and high strength wastewater [14–17]. The two forms of membrane configuration includes internal
and external, as shown in Figure 1, but the external configuration is the most widely reported.
(a) (b)
Figure 1. A schematic of AnMBR configurations—(a) Side or external membrane (b) submerged membrane.
Membrane Configuration Performance

A pilot scale study of municipal wastewater treatment with an external membrane configuration
was reported by Huang et al. [14]. The system achieved COD removal efficiency close to 90%, but a
slow and linear increase in the filtration resistance was observed under critical flux conditions and
subsequently resulted in fouling due to solid accumulation on the surface of the membrane. However,
it was observed that gas sparging and additional shear were needed to control membrane fouling.
According to Chu et al. [10], high shear stress may interfere with the biological activities during
anaerobic digestion of biomass. Furthermore, studies of Zhao et al. [15] showed that the external
membrane configuration system is easier to maintain and monitor. However, these require more energy
with high hydraulic shear force which might also disrupt anaerobic bio-solids favoring substrates
with smaller particle sizes and this in turn causes membrane fouling. In another development,
thorough investigations of submersible membrane module coupled to anaerobic bioreactors were
reported in [16–19] and it is believed to overcome the numerous drawbacks of conventional methods
of wastewater treatment. However, it also appears that the technology is more suitable for low strength
organic loads, especially municipal wastewater as examined in the previous research of [20–22].
Martinez-Sosa et al. [3] also examined a submerged anaerobic membrane bioreactor and found
97% COD removal. The final effluent COD was less than 20 mg/L at volumetric organic loading
rate of 0.5 to 12.5 kg/m3 ·day. The system could be termed efficient even though, the temperature
fluctuates between 12 ◦ C and 26 ◦ C. Likewise the membrane permeability was not outstanding due to
intermittent suction mode and membrane flux that lead to frequent membrane cleaning. Furthermore,
investigation of Deowan et al. [12] on a submerged anaerobic membrane bioreactor (SAMBR) treating
textile wastewater revealed COD removal efficiency around 90% with negligible fluctuations in some
phases. However, 60% color removal was reported in the first phase and subsequently dropped to
between 20–50% after the system stabilized. This phenomenon was attributed to color adsorption
on the surface of the membrane. In a similar configuration manner, the defects in COD removal
were ascribed to the presence of high total nitrogen (TN) and total phosphorus (TP) [23]. In view
of that, application of anoxic conditions or an aerobic process or a combination of the two could
enhance removal of both TN and TP. Katayon et al. [24], experimented on the effect of vertical and
105
Energies 2018, 11, 1675
horizontal membrane configurations, in which two procedures consisting of low and high mixed liquor
suspended solids (MLSS) concentrations were considered. The results showed that horizontally placed
membrane was able to removed 99.2% total solids and 99.73% turbidity at lower MLSS concentration
at a mean flux value of 5.03 L/m2 h. This is quite higher than those found in vertical modules with high
MLSS concentrations and mean flux value of 2.27 L/m2 h. Hence, for higher efficiency, maintaining
high biomass concentration and sufficient microbial activity with minimal energy utilization would
enhance the process of anaerobic membrane bioreactor operation processes.
3. Membrane Performance of Various Wastewater Treatments
3.1. Industrial Wastewater

Industrial regulatory issues towards meeting the stringent water quality discharge permissible
limits motivate researchers to study more on finding lasting solutions to the problems associated
with the effluent released to the environment. Most industrial wastewaters are regarded as high
strength wastewater, because they contain large amounts of settleable, dissolved and suspended
solids or other elements such as heavy metals in greater proportion [25,26]. However, the strength of
the wastewater could differ from one industry to another owing to the different types of operations.
Generally, industrial wastewater comes from streams such as production lines, cooling towers or boiler
and cleaning processes that contain diverse substances. These may include organic and inorganic
compounds, viruses, bacteria and toxic compounds. Therefore, applying less energy and achieving
high COD removal with minimum sludge production is of utmost priority to the water and wastewater
treatment industry. For instance, a pilot scale study of SAMBR treating high strength wastewater
(raw tannery wastewater) achieved higher COD removal efficiency up to 90% at organic loading
rate OLR of 6 g/L·day and biogas yield (0.160 L/g COD removed) [27]. The system performed
efficiently, but was strongly characterized by a high hydraulic retention time (HRT) (40 h) and as
such, high energy was expended, although the permeate flux remained at (6.8 LMH) which was
well below the critical flux (17.5 LMH) as determine in the earlier work of Hu et al. [28]. Similarly,
Fush et al. [29] reported treatment of three different high strength wastewater effluents (artificial
wastewater, animal slaughterhouse, and sauerkraut brine) using continuous stirrer tank reactor CSTR
coupled with membrane filtration units. The COD removal in all the reactors were >90% at OLR of 20 g
COD/L·day, 8 g COD/L·day and 6–8 g COD/L·day respectively. The methane yields were in the range
of 0.17–0.30, 0.20–0.34, and 0.12–0.32 Ln/g·COD fed. On the other hand, Saddoud et al. [30] showed
that an anaerobic cross-flow ultrafiltration membrane bioreactor exhibited high efficiency removal
of suspended solid (SS), biochemical oxygen demand BOD5 , COD and microorganisms. It reached
>100%, 90%, 88%, and 100% respectively. Biogas production got to 30 liters per day with average of
0.27 L CH4/g·COD yield. Interestingly, this high yield of biogas was achieved at low organic loading
rate (OLR) of 2 g COD/L·day.
Most recently, the filtration performance of an AnMBR treating high strength lipid-rich wastewater
and corn-to-ethanol thin stillage was conducted by Dereli et al. [31]. The reactors delivered a high
COD removal efficiency of up to 99% under stable operating conditions with an average OLR of 8.3,
7.8, and 6.1 kg COD/m3 ·day. However, the permeate quality turned out to be inferior in quality with
increased in solid retention time (SRT). Table 2 present some examples of the biological and membrane
performance of different AnMBR applications for treatment of industrial wastewater. Unfortunately,
information regarding the membrane performance with respect to biogas production is quite limited
in most of the studies.
3.2. Municipal Wastewater

Effluents characterized by low organic strength and high particulate organic matter are mostly
reported as municipal wastewater [32]. The ability of AnMBRs in retaining biomass within the reactor
has made it a subject of research for municipal wastewater treatment as a possible alternative to the
106
Energies 2018, 11, 1675
conventional anaerobic treatment process. According to Ho et al. [33], Kocadagistan and Nazmi [34],
bioreactors coupled with membranes used for the treatment of municipal wastewater (MWW) have
shown excellent effluent quality that meets the stringent discharge standards in terms of COD,
suspended solids and pathogen counts removals when compared to conventional anaerobic methods.
Moreover, the use membrane bioreactor (MBR), for anaerobic treatment of municipal wastewater in
high temperate climate is still a challenge. This is because, domestic wastewater is usually complex
in nature and characterized by high fraction of particulate organic matter. It could be in the form of
proteins, suspended solids of different origin, and fatty acids with large to moderate biodegradability
portions [35]. Based on these characteristics, operation of MWW treatment reactor under psychrophilic
temperature (<20 ◦ C) might be some how difficult. Smith et al. [36] compared simulated and actual
domestic wastewater (DWW) at the bench-scale using submerged flat-sheet microfiltration membranes.
A psychrophilic temperature of 15 ◦ C was set. An average removal efficiency of 92 ± 5% COD
was achieved, which corresponds to an average permeate COD of 36 ± 21 mg/L in simulated
DWW, while 69 ± 10% was realized during actual DWW treatment. However, it is obvious the
membrane in this study utilizes a lot of energy with relatively high concentration of dissolved methane.
Although previous references [37–41] have reported the performance of simulated and actual DWW
treatments, only a few of them were performed at psychrophilic temperature (15 ◦ C and below).
Table 3 summarizes some studies on the use of AnMBRs for domestic wastewater treatment.
Another investigation of Smith et al. [52] showed a simulated domestic wastewater using
AnMBR at psychrophilic temperatures of 15, 12, 9, 6, and 3 ◦ C. Remarkably, a total reduction of
COD > 95% was realized at a temperature of 6 ◦ C. The success was attributed to viable microbial
activity in the membrane biofilm but subsequently give rise to high dissolved methane oversaturation
in permeate and consequently fell to 86% at 3 ◦ C. However, decreasing temperature resulted in
increased soluble COD in the bioreactor. Thus, this signifies a reduction in suspended biomass
activity. Dolejs et al. [53] conducted a research on the effect of psychrophilic temperature shocks on a
gas-lift anaerobic membrane bioreactor (Gl-AnMBR) used for treating synthetic domestic wastewater.
The stability of the system was measured by transiting between mesophilic and psychrophilic stages
having several psychrophilic shocks (12–48 h). The result showed an average COD removal of 94 ± 2%
at mesophilic with an average methane yield of 0.19 L CH4 /g COD removed (including psychrophilic
shocks). More than 80% of the influent COD accumulated in the reactor under psychrophilic in
comparison to 39% under mesophilic conditions.
A pilot scale study of anaerobic urban wastewater treatment in a submerged hollow fiber
membrane bioreactor was reported by Gimenez et al. [20]. They assessed the effect of a number
of operational variables on both biological and physical separation process performance. Mesophilic
temperature of 33 ◦ C, at 70 days SRT, and HRT ranging from 20 h down to 6 h. COD removal stood
almost at 90%, with no trace of irrevocable fouling observed but yet the methane yield was very
low and this was mainly attributed to influent COD/SO4 –S ratio. The work of Gouveia et al. [54]
on a pilot scale AnMBR coupled to an external ultrafiltration treating MWW also attained COD
removal efficiency of 87 ± 1% with specific methane yield of 0.18 and 0.23 Nm3 CH4 /kg COD
removed at a lower temperature of 18 ± 2 ◦ C. More recently, a toxicity reduction in wastewater
(synthetic wastewater) aiming at reuse possibilities was performed at relatively psychrophilic
temperature (25 ◦ C) using submerged anaerobic membrane bioreactor with forward osmosis membrane
(FO-AnMBR) [55]. The FO-AnMBR process exhibited >96% removal of organic carbon, nearly 100% of
total phosphorus and 62% of ammonia-nitrogen respectively. This suggests better removal efficiency
than the conventional AnMBR. The average COD removal was 96.7% corresponding to the influent
COD concentration of 460 mg/L and a methane production of 0.21 L CH4/g COD. Thus, the system
demonstrates high feasibility of energy recovery.
107
Table 2. References on anaerobic membrane bioreactors (AnMBRs) for industrial wastewater.
Reactor Maximum
Type of System/Module Biogas Production
Membrane Wastewater Working Influent Effluent COD COD
Configuration/Membrane Operating Condition (L CH4 /g COD Reference
Type/Material/Characteristic Treated Volume + COD mg/L Mg/L Removal
Configuration removed
Scale (%)
Flux = 8.4 L/m2 ·h,
-/ceramic (ZrO2 –TiO2 )/ HRT = 1.7–5, 3.5, 3–3.5,
AnMBR/External cross flow Industrial V = 50 L
Energies 2018, 11, 1675
Pore size = 0.2 μm SRT = 120–450 day, 4300 830 78 0.5 and 0.6 [42]
membrane/Tubular wastewater S=P
Area = 0.25 m2 Temp = 35–37 ◦ C,
MLSS = OLR = 2.5 g, COD L/day
MF (Polipropilen-PP)/Chlorinated Flux = 3–0.9 L/m2 ·h,
Synthetic
SAMBR/Submerged/Flat sheet and polyethylene/ HRT = 390, 167, 168, V=4L
industrial 20,000–23,000 15.2 85–90 - [43]
hollow fiber Pore size = 0.05 μm SRT = ∞, Temp = 35 ◦ C, S=L
wastewater
Area = 0.66 m2 MLSS = OLR = 0.3–0.54 g, COD/L·day
Submerged anaerobic Flux = 6 L/m2 ·h,
MF/Curtain-type/ Synthetic
membrane bioreactor HRT = 2.2 h, SRT = -, Temp = 35 ◦ C, V = 25 L
Pore size = 0.2 μm industrial 223 ± 111 50 ± 22 87 0.12 [44]
(SAnMBR)/Submerged/ MLSS =10.9 g/L, S=P
Area 5.4 m2 wastewater
Hollow fiber OLR = 3.0 kg COD/m3 ·day
Flux = 7.2 L/m2 ·h,
MF/Curtain-type/ HRT = 35 h, SRT = 40 day,
SAnMBR/Submerged/ Paper mill V = 10 L
Pore size = 0.4 μm Temp = 21 ◦ C, 11,415 ± 15 228.3 ± 5 98 - [45]
Hollow fiber wastewater S=L
Area = 0.040 m2 MLSS = 12.90 g/L,
OLR = 7.0 kg COD/m3 ·day
UF/Hollow Flux = 20 L/m2 ·h, 8056.5
108
fiber/Polyvinylidene fluoride Synthetic HRT = 48, 36, 24, and 18 h, 1218.2
V = 4.4 m3 93.6 and
AnMBR (PVDF) anti-biotic Temp = 35 ± 1 ◦ C, 7892–21,986 1523.5 94 and 130.7 [46]
S=P 98.7
Pore size = - solvent MLSS = 16.5–12.4 g/L, 1828.3
Area = 20 m2 OLR = 3.9–12.7 kg, COD/m3 ·day 2207.7
Flux = 20 L/m2 ·h,
UF/Hollow fiber/
Synthetic HRT = 48 h,
Polyvinylidene fluoride (PVDF) V = 4.4 m3
AnMBR anti-biotic Temp = 37 ◦ C, 15,000–25,000 4000 96.5 - [47]
Pore size = - S=P
solvent MLSS = 52.2 g/L,
Area = 20 m2
OLR = 3.79 kg , COD/m3 ·day
Flux = 20 L/m2 ·h,
UF/Hollow fiber/ HRT = 48–24 h,
Synthetic
Polyvinylidene fluoride (PVDF) Temp = (35 ± 3 ◦ C, 25 ± 3 ◦ C, 15 ± 3 ◦ C, 25 V = 4.4 m3
AnMBR anti-biotic 1000–25,000 - 95 - [48]
Pore size = - ± 3 ◦ C), S=P
solvent
Area = 20 m2 MLSS = 16.5 g/L,
OLR = 10.0 kg, COD/m3 ·day
Flux = 8.64 ± 0.69, L/m2 ·h,
UF/Hollow fiber/ HRT = 44 h,
Brewery V = 15 L
AnMBR Pore size = 0.04 μm Temp = (35 ◦ C), 19,100 171 99 0.53 ± 0.015 [49]
wastewater S=L
Area = 0.047 m2 MLSS = 2.8 g/L,
OLR = 3.5–11.5 g , COD/L·d
Table 2. Cont.
Reactor Maximum
Type/Material/Characteristic Treated Volume + COD mg/L Mg/L Removal
Configuration removed
Scale (%)
Flux = 33.4–16.2 L/m2 ·h,
UF/Hollow fiber/
HRT = 3 d, 278.9 ± 4.2 mg/L 94.5 ± 2.9
Polyvinylidene fluoride (PVDF) Bamboo V = 15 L 17,160 ± 13.2 ± 1.2,
Energies 2018, 11, 1675
AnMBR and B-AnMBR Temp = 32 ± 2 ◦ C, and and 89.1 ± [50]

Pore size = 0.02 μm wastewater S=L 814 10.3 ± 0.8
MLSS = 16 g/L, 125.3 ± 3.2 mg/L 3.1
Area = 0.07 m
OLR = 6 kg , COD/m3 ·day
Flux = 6 L/m2 ·h,
UF/Hollow fiber/
HRT = 30.6 h, 46.1 ± 2.9
Polyvinylidene fluoride (PVDF) Pharmaceutical V = 10 L 16,249 ± 8723 ± 593 and 24.5 ± 2.1
C-AnMBR and B-AnMBR Temp = 27 ± 1.0 ◦ C, and 60.3 ± [51]
Pore size = 0.04 μm wastewater S=L 714 6432 ± 445 And 17.6 ± 1.5
MLSS = 9500–10,200 and 10,000 mg/L, 2.8
Area = 0.047 m2
OLR = 13.0 ± 0.6 kg, COD/m3 ·day
AnMBR = anaerobic membrane bioreactor, COD = chemical oxygen demand, HRT = hydraulic retention time, MLSS = Mixed liquor suspended solids, OLR = organic loading rate,
P = pilot scale, - = not reported.
Table 3. References on anaerobic membrane bioreactors (AnMBRs) for municipal wastewater treatment.
Reactor Maximum
Type/Material/Characteristic Treated Volume + COD mg/L mg/L Removal
Configuration removed)
109
Scale (%)
-/Non-woven
Raw Flux = 30 L/m2 ·h, HRT = 4 h, SRT = ∞,
Fibrous (chlorinated polyethylene)/ V = 2.7 L
AFMBR/Submerged/ Hollow fiber municipal Temp = 23 ± 1 ◦ C, MLSS = -, - 23.5 97.07 2.11 [56]
Pore size = 0.1 μm S=L
wastewater OLR = 3–6
Area 0.022 m2
Municipal
(wastewater Flux = 9.72 L/m2 ·h,
MF/Polyvinylidene fluoride/ from the HRT 110 h, SRT = -,
SAMBRs/Submerged/ V = 0.628 L 15,000 ±
Pore size = 0.02 μm ethanol Temp = 23–25 ◦ C, 202 ± 23 98.2 - [57]
Hollow fiber S=L 1000
Area = - fermentation MLSS = 8–12 g/L,
of food OLR =2.3–3.6 g/L·day
waste)
Flux = 2.5 L/m2 ·h,
UF/Polyvinylidene fluoride/ Raw
HRT = 8 h, V = 0.7 m3 525 ± 174, 222 ± 61,
UASB/Submerged/Tubular Pore size = - municipal 68.6 - [58]
Temp = 18–21 ◦ C, S=P 657 ± 235 130 ± 55
Area = 0.2375 m2 wastewater
MLSS = -, OLR = -
Flux = 5.3 L/m2 ·h, HRT = 12 h, SRT = 25–30
CG-AnMBR and -/A polyvinylidence (PVDF)/ Synthetic day,
V=3L
SG-AnMBR/Submerged/ Pore size = 0.22 μm domestic Temp = 20 ◦ C, 330–370 - 90 156.3 ± 5.8 [59]
S=L
Hollow fiber Area = 0.06 m2 wastewater MLSS = 20.50 ± 1.53 g/L,
OLR = -
Flux = 7.8 L/m2 ·h
MF/-/ Real HRT = 35 h, SRT = 25–30 day
V = 3 m3 896 GC
MBR/Submerged/Flat sheet Pore size = 0.4 μm municipal Temp = 6.5–21 ◦ C/7.0–20 ◦ C 18.7 ± 2.9 93 - [60]
S=P mL−1
Area = 16 m2 wastewater MLSS = 5300–9800 mg/L
OLR = -
Table 3. Cont.
Reactor Maximum
Configuration removed)
Scale (%)
-/Ceramic membrane/ Flux = 8 L/m2 ·h, HRT = 5.8 h, SRT = 60 day,
Domestic V = 3.6 L
AnCMBRs/Submerged/ Flat sheet Pore size = 80 nm Temp = 25 ◦ C, MLSS = -, 417 ± 61 - 87 - [61]
wastewater S=L
Energies 2018, 11, 1675
Area = 0.08 m2 OLR =10.0 kg COD/m3 ·day

Synthetic
-/Polyethylene terephthalate/ municipal Flux = -, HRT = 42–12 h
V=6L
SAnMBR/Submerged/Flat-sheet Pore size = 0.2 μm wastewater SRT = -, Temp = 25 ± 1 ◦ C - 17.1 95.5–98.8 2.30–4.25 [62]
S=L
Area = 0.116 m2 (alcohol MLSS = -, OLR = 3.0–6.0 kg COD/m3 ·day
ethoxylates)
Flux = 8.3–9.5 L/m2 ·h and 6.0–6.7 L/m2 ·h,
UF/Polyvinylidene fluoride/
Domestic HRT = 5.8–4.8 and 8.0 –7.1 h, SRT = 50 day, V = 15 L
SAnMBR/Submerged/Flat-sheet Pore size =0.2 μm 400 - 90 276 ± 13 [63]
wastewater Temp = 35 ◦ C and 25 ◦ C, S=L
Area 0.735 m2
MLSS = -, OLR = 0.43–0.90 kg COD/m3 ·day
Flux 12.3 L/m2 ·h,
UF/Polyethersulfone/ Synthetic
HRT = 6 h, SRT = 126 day, V=7L
AnMBR/External/Tubular Pore size = 30 μm municipal 530 ± 30 42 and 52 92 and 90 - [64]
Temp = 25 ◦ C and 15 ◦ C, S=L
MLSS = -, OLR = 2 kg/m3 ·day
Flux 12.3 L/m2 ·h,
HRT = 6 h,
UF/Polyethersulfone/ Synthetic
SRT = 126 day, V=7L 149 ± 5.9 to
AnMBR/External/Tubular Pore size = 30 μm municipal 530 ± 30 92 - [65]
Temp = 25–15 ◦ C, S=L
110
42 ± 4.4,
MLSS = -,
OLR = 2 kg/m3 ·day
AnMBR = anaerobic membrane bioreactor, COD = chemical oxygen demand, HRT = hydraulic retention time, MLSS = Mixed liquor suspended solids, OLR = organic loading rate,
P = pilot scale, - = not reported.
Energies 2018, 11, 1675
3.3. Synthetic Wastewater

Compounds such as starch, glucose, molasses, peptone, yeast, and cellulose are usually used as
synthetic substrates to test new concept of AnMBR. The results of a number of studies are summarized
in Table 4. The COD removal efficiencies of those investigations were generally >90%, with OLR less
than 10 kg COD/m3 ·day. Effect of HRT and SRT on treatment performance of submerged AnMBR for
synthetic low-strength wastewater reveals a total COD removal efficiencies higher than 97% at all the
operating conditions with a maximum biogas production in terms of mixed liquor volatile suspended
solid (MLVSS) removals (0.056 L CH4/g MLVSS) at infinite SRT [66]. Though increasing in OLR with
short hydraulic time HRT and long SRT boosted the methanogenic environment, but membrane fouling
was worsened due to a decrease in HRT which heightened the growth of biomass and accumulation
of soluble microbial products (SMP). The experiment performed by Jeison et al. [67] on synthetic
wastewater using submerged anaerobic membrane bioreactor showed a reversible cake layer formation
on short term basis. Additionally, cake consolidation was detected during a long-term operation at a
flux close to the critical point. Remarkably, increasing OLR from 50–60 g COD/L·day towards the end
of the operation presented high COD removal efficiency greater than 90%.
Fallah et al. [68] reported a significant COD removal of (<99%). In their studies, they considered
two hydraulic retention times (24 h and 18 h) using a MBR to remove styrene from a synthetic
wastewater having a chemical oxygen demand and styrene concentration of 1500 mg/L and 50 mg/L.
nonetheless, reduction in HRT to 18 h caused a release of extracellular polymeric substance (EPS) from
the bacterial cells that led to the rise in soluble microbial product (SMP) and sludge deflocculation.
Moreover, the dramatic rise in transmembrane pressure TMP which was operating fairly low and
constant for a number of days give rise to severe membrane fouling. This trend was attributed to
the rise in SMP concentrations and the decrease in mean floc size. Similar research was reported in
Ho et al. [69]. The HRT were varied during treatment of synthetic municipal wastewater. The permeate
quality was outstanding, irrespective of HRT differences with over 90% COD removal at HRT of
6 h. Methane produce was 0.21 to 0.22 CH4 /g COD removed. Conversely, the fraction of methane
recovered from the synthetic municipal wastewater declined from 48 to 35% with the reduction of
HRT from 12 to 6 h. Subsequently, the result of the increased mixed-liquor soluble COD which was
precluded and accumulated in the AnMBR drastically affects the performance.
111
Table 4. References on anaerobic membrane bioreactors for synthetic wastewater treatment.
Reactor Maximum
Configuration/Membrane Operating Condition (L CH4 /g Reference
Configuration CODremoved)
Scale (%)
Synthetic
municipal
Energies 2018, 11, 1675
-/Non-woven wastewater
SAMBR/ Flux = -, HRT = 24–12 h,
fibrous (chlorinated polyethylene)/ (linear alkyl V=6L
Submerged/ SRT = Temp = 25 ± 1 ◦ C - 23.5 97.07 2.11 [74]
Pore size = 0.2 μm benzene S=L
Flat sheet MLSS = -, OLR = 3–6 kg COD/m3 ·day
Area = 0.116 m2 sulfonate
concentration
in sewage)
Flux = 15 L/m2 ·h, 252 ± 27,
SAMBRs/ MF/Polymethyl methacrylate/ HRT = 12, 8, 6, 4 and 2 h, SRT = 200 day, 236 ± 27,
Synthetic V=3L
Submerged/ Pore size = 0.2 μm Temp = 35 ± 1 ◦ C, 544 ± 22 14 ± 2 >97 249 ± 29, [75]
sewage S=L
Flat sheet Area = 0.116 m2 MLSS = 6000 mg/L and 7000 mg/L, 264 ± 46,
OLR = - 134 ± 23
Flux = 10.64L/m2 ·h,
MF/Polymethyl methacrylate/
Synthetic HRT = 8 h, SRT = 140 day, V=4L
SMBR/-/- Pore size = 0.04 μm - - 91 and 98 - [76]
wastewater Temp = 18 ± 15 ◦ C, MLSS = 6000 mg/L, S=L
Area = 0.047 m2
OLR = 1.77 ± 0.03 g COD/L·day
Flux = 3.532 LMH
Cathode/Stainless steel mesh/ HRT = -, SRT = -
Synthetic V = 6.8 L
AnOMBR/submerged/- Pore size = - Temp = 35 ± 1 ◦ C 2000 - 71.1 0.254 [77]
wastewater S=P
112
Area 1.5 m2 MLSS = 16.31 g/L
OLR = -
MF/Non-woven Flux = 5 L/m2 ·h,
fibrous (chlorinated polyethylene)/ Synthetic HRT = 12–6 h, V = 3.2 L
SAMBRs/Submerged/Flat sheet - 29 ± 8 93.2 ± 6.6 1.7 ± 0.3 [78]
Pore size = 0.2 μm sewage SRT = ,Temp = 35 ◦ C, S=L
Area = 0.116 m2 MLSS = -,OLR = 3–6 kg COD/m3 ·day
-/Hollow fiber/ Synthetic Flux = 11.3 L/m2 ·h, 0.31 ± 0.02,
Integrated anaerobic fluidized bed V = 6.1 L
Pore size = 0.4 μm benzothiazole HRT = 24 h, Temp = 35 ◦ C, MLSS = -, - 230 96 0.32 ± 0.03 and [79]
membrane (IAFMBR)/-/- S=L
Area = 0.21 m2 wastewater OLR = - 0.31 ± 0.01
Flux = -,
MF/Chlorinated polyethylene/
Synthetic HRT = 48, 24, 12, 6 h, V=6L
SAMBRs/Submerged/Flat sheet Pore size = 0.2 μm - 134 94 - [80]
sewage Temp = 25, 15, 10 ◦ C, S=L
Area = 0.116 m2
MLSS = -, OLR = -
-/Polyvinylidence (PVDF) hollow Flux = 20.2 ± 8.5 L/m2 ·h, HRT = 3.5 ± 1.1,
0.095 ± 0.15 0.14
fiber/ Synthetic 2.2 ± 0.5, 2.3 ± 0.3 day, Temp = 35 ◦ C, V = 3.5 L
AnMBRs/External/- 6752 ± 663 ± 0.23 0.28 ± 96.7 ± 2.7 0.50 ± 0.17 [81]
Pore size = 0.22 μm sewage MLSS = 4.78 ± 1.9 g L−1 , S=L
0.35
Area = 0.06 m2 OLR = 3.4 kg COD/m3 ·day
-/Polyvinylidence (PVDF) hollow Flux = 7 L/m2 ·h,
fiber/ Synthetic HRT = 12 h, Temp = 20 ◦ C, V=4L
EG-AnMBR and SG-AnMBR - - <90% 160 [82]
Pore size = 0.22 μm sewage MLSS = 22.34 ± 0.41 g/L, S=L
Area = 0.06 m2 OLR = 0.53–0.59 kg COD/m3
AnMBR = anaerobic membrane bioreactor, COD = chemical oxygen demand, HRT = hydraulic retention time, MLSS = Mixed liquor suspended solids, OLR = organic loading rate, P =
pilot scale, - = not reported.
Energies 2018, 11, 1675
4. Effect of Microbial Activity on Anaerobic Membrane Performance

During microbial activity, the concentration of volatile fatty acids (VFA) usually reflects the state of
anaerobic digestion performance, specifically in the acetogenic and methanogenic phase. According to
Wijekoon et al. [70], 85–96% of high strength molasses-based synthetic wastewater was removed as total
chemical oxygen demand (COD) at optimum organic loading rate of 8 ± 0.3 kg COD/m3 ·day. Though
it was at higher temperature, but biogas production of 15, 20 and 35 L/day at OLR 5.1 ± 0.1, 8.1 ± 0.3
and 12.0 ± 0.2 kg COD/m3 ·day respectively was achieved. It was seen from the results of the treatment,
increasing loading rate amounts to increase in hydrolytic and methanogenic activities. However, it was
also observed that the process performance reached its maximum level with continuing increase in
loading rate which subsequently reduce the biological activities of the system.
In anaerobic digestion, rapid methane formation is mostly attributed to the presence of acetic acid
and butyric acid. However, acetic acid contributes 70% of the total acidic portion. This might be due
to the fact that, all the volatile acids are converted into acetate during metabolism. The presence of
propionic acid tends to upset anaerobic digestion processes due to its toxicity among the VFAs, but this
effect could be counteracted under thermophilic conditions. The experiments of Speece et al. [71]
showed that propionic acid degradation rate to other intermediate was very poor, largely because of
lower partial pressure (H2 ) demand, but it also supports process start up and stabilization under strict
anaerobic conditions.
Moreover, methanogenic activity might be inhibited with propionic acid concentrations greater
than 1–2 g/L. It could also withstand acetic and butyric acid concentrations up to 10 g/L.
Ahmed et al. [72] have found that, microbial respiratory quinones are essential parts of the bacterial
respiratory chain that perform a vital role in electron transfer during the period of microbial respiration.
However, an investigation of Hiraishi et al. [73] demonstrated that alterations in microbial community
structure in a mixed culture of microbes could effectively be quantified using quinone profiles.
Recent research on the effects of temperature and temperature shock on the performance and
microbial community structure of a submerged anaerobic membrane bioreactor (SAnMBR) was
demonstrated by Gao et al. [21]. The result obtained indicates that not only the diversity, but also
the species richness of microbial populations are affected by temperature variation. It further proves
that submerged AnMBR performance under temperature shock conditions have no effect on the
COD removal ability of the reactor. The biogas production rates were 0.21 ± 0.03, 0.20 ± 0.03
and 0.21 ± 0.02 L/g COD removed. Moreover, no major change is observed in the production and
composition of biogas. Nevertheless, temporary production of biogas occurred at temperature shock
which affected the abundance and diversity of microbial populations.
According to Iranpour et al. [83], the similarities that exist among the microbial community
working under variable temperatures may be connected to the development of thermomesophiles
rather than thermophiles. During microbial activity in anaerobic digestion, the release of soluble
organic compounds called SMP in normal biomass metabolism is of great interest, not only in
terms of achieving discharge standard limits, but also in setting the lower limit for treatments [84].
The significance of SMP in all kinds of wastewater treatment is at the moment objectively well
recognized, but complications still come about in trying to measure SMP and draw conclusions when
they are present in effluents from plants treating highly complex feeds. However, researchers like
Sciener et al. [85] and Chidoba et al. [86] have studied this and conclude that large portions of the
soluble organic matter in the effluent from the biological treatment processes are actually SMP.
The work of Judd [87] and Li et al. [88] reveals that microbial activity decreases under prolonged
SRT in AnMBR. This is because microbes take a long time to degrade the inorganic matter and require
a high concentration of biomass to ensure all the organics are totally degraded. Still, the application of
high SRT is advantageous, since it favors biomass growth which is responsible for biodegradability
of organic pollutants. It also provides room for higher MLSS in AnMBR that establish starvation
conditions to achieve good quality effluent and create a low F/M ratio [89,90].
113
Energies 2018, 11, 1675
5. Operational and Performance Parameters
5.1. Temperature
In anaerobic digestion (AD) processes, temperature is a major factor that plays an important role in
the stabilization and performance of the whole system [91,92]. Under strict anaerobic conditions, some
bacteria thrive well under psychrophilic (<25 ◦ C), mesophilic (25–40 ◦ C) and thermophilic (>45 ◦ C)
conditions [93]. According to El-Mashad et al. [92] and Duran et al. [94], thermophilic conditions offer
more advantage in terms of specific growth and metabolic rates with comparatively less ammonia
inhibition than mesophiles. However, it could also cause higher microbial death rates, poor supernatant
quality and reduced process stability due to chronically high propionate concentrations than mesophilic
bacteria. Studies on biogas production and wastewater treatment at varying temperature such as
psychrophilic (0–20 ◦ C), mesophilic (20–42 ◦ C), and Thermophilic (42–75 ◦ C) were well investigated
in [10,12,14,94].
The diversity and activities of microbial communities coupled with thermodynamic equilibrium
of the biochemical reactions are adversely affected by temperature [95]. It usually shifts the abundance
and activities of specific microbial populations and determines the roles of specific taxa in the AD
food chain. It all begins with substrate hydrolysis, followed by acidogenesis, then acetogenesis,
and ends with methanogenesis [96]. Study on temperature adaptation was examined by Chidoba,
1985 [86]. It was initially set at mesophilic condition (37 ◦ C) and gradually switched to thermophilic
(55 ◦ C). A drastic reduction in biogas production by 15% from the original state of mesophilic was
observed. This was attributed to the change in temperature along with increase in VFA to 3000 mg/L.
More detailed studies were reported by Song et al. [97] on thermophilic and mesophilic temperature
co-phase anaerobic digestions. Their examination exclusively focused on the sewage sludge using
the exchange process of digesting sludge between spatially separated mesophilic and thermophilic
digesters and compared with single-stage mesophilic and thermophilic anaerobic digestions. It was
confirmed that the system stability, effluent quality, specific methane production during single-stage
operation was greater in mesophilic than thermophilic., but volatile solids (VS) reduction and total
coliform destruction were much higher in single-stage thermophilic than mesophilic digestion.
In the past, study on the effect of temperature range between 40–64 ◦ C using cow manure as the
main substrate was well studied by Angelidaki, 1994 [98]. Two different ammonia concentrations
(2.5 and 6.0 g N/L) were continuously fed to the lab-scale reactor. It was observed at some stage,
precisely (HRT 15 days); the high temperature and ammonia loading resulted in poor process
performance. Consequently, a significant change in the amount of biogas production and process
stability was seen when the ammonia load was high and temperature reduces to below 55 ◦ C.
5.2. pH
Volatile fatty acid (VFA) concentration usually determines the pH of the effluent, which is
one of the influential factor in anaerobic digestion (AD) processes [99]. Different range of pH is
required for bacterial growth in AD, comprehensively between 4.0 to 8.5 [100]. However, a pH
level of 6.8 to 7.2 suitably favors methanogenic bacteria [101,102]. On the other hand, hydrolysis
and acidogenesis thrive well in the pH range between 5.5 and 6.5 [103,104]. It was also shown that,
excessively alkaline pH may result in microbial granules disintegration and subsequent failure [105].
The pH adjustment studies in [101] favorably increased methane yield. The maximum cumulative
biogas production reached 16,607 mL at pH 7.0 (0.4535 L methane/g VS). However, the yield
decreased to 6916 and 9739 mL at pH 6.0 and 8.0, equivalent to 0.1889 L methane/g VS and
0.2659 L methane/g VS, respectively.
Recent research on novel biogas-pH automation control strategy using a combined gas-liquor
phase monitoring was developed by Yu et al. [106] using an AnMBR treating high starch wastewater
COD (27.53 g/L). The biogas pH progressed with threshold between biogas production rate
>98 NmL·h−1 preventing overload and pH > 7.4 preventing under load. The OLR and the effluent COD
114
Energies 2018, 11, 1675
was doubled to 11.81 kg COD/m3 ·day and halved as 253.4 mg/L, respectively. In another development,
a ternary contour was performed by Mao et al. [107] to picture the pH dissimilarities in ternary buffer
system. The variation was controlled by a system composing of VFAs, ammonia and carbonate.
However, the accurate simulation of pH variation in AnMBR during methanogenesis is extremely
challenging due to enormous magnitude of electrolytes, and as such the ternary macro-quantity
buffer salts were chosen for the visualization which denotes the critical pH buffer capacity to during
methanogenic stage.
Kim et al. [104] demonstrated that one major problem faced with COD removal especially for
starch wastewater is the provision of sufficient carbonate alkalinity. The effect of high pH shocks
(pH 8.0, 9.1 and 10.0) on the performance of submerged AnMBR was well reported in [105]. It was
found that; pH 8.0 had slight influence, while pH 9.1 and 10.0 shocks have put forth vigorous impact
on biogas production, COD removal, and membrane filtration performance. In addition, colloids
and solutes accumulation in the sludge suspension further accelerates deterioration of membrane
performance. Interestingly, when neutral pH (7) was taken up again, it cost the reactor approximately
1, 6, and 30 days to recover from the previous shocks.
An experiment on the effect of mixed liquor (pH 5 and 9) in the removal of trace organics in both
acidic and basic environment was also reported in [106]. A reduction in total organic carbon TOC
and total nitrogen TN removal efficiencies was detected with ionisable trace organic contaminants.
Hence, the removal efficiencies were extremely pH dependent. Conversely, the biological performance
was favorable at near optimum pH (i.e., approximately pH 6–7) with respect to TOC and TN removal
efficiencies. As result of the differences in attainment of optimum pH by acidogenic bacteria (5.5–6.5)
and methanogens (6.5–8.2), several studies have been conducted on phase separation of acidogenesis
and methanogenesis using AnMBRs. For example, Mao et al. [107] demonstrated a two-stage reactor
configuration and optimizes each one phase as an entity. VFA accumulation was drastically reduced
and the system stability improves to the extent of forbearing greater loading rate and toxicity. Therefore,
maintenance and adherence to these features would positively escalate the chance of achieving a high
methane yield.
5.3. Effect of OLR

The amount of volatile solids VS fed to a bioreactor at an interval of time under continuous
operation is termed as the OLR. Under normal operating conditions, it is expected that an increase in
organic loading rate would also increase the quantity of biogas. AnMBR processes have the advantage
of tolerating changes in organic loading similar to tolerance to fluctuations in temperature. Organic
loadings ranging from 0.5 to 12.5 kg/m3 ·day was applied to AnMBR for the treatment of domestic
wastewater. The system achieved 97% (COD) removal with less than 20 mg/L effluent COD [108]. In a
similar experiment by Vincent et al. [109], less than 50 mg/L soluble COD in the effluent was obtained
at organic loading rate 0.25 kg/m3 ·day and 0.7 kg/m3 ·day using AnMBR. This is not the case in the
work of Qiao et al. [110]. The system exhibited very poor effluent quality after a long HRT. Study of the
performance of AnMBR by Yingyu et al. [111] showed that biogas yield rose linearly with increasing
organic loading. A similar trend was observed by Wijekoon et al. [70] using a two-stage thermophilic
AnMBR with continuous increasing loading rate from 5 to 12 kg COD/m3 ·day. Bornare et al. [112]
achieved biogas production increased from 159 to 289 L/day, but they observed a decreased in yield
from 0.48 to 0.42 L biogas/g CODremoved when OLR was increased from 0.62 to 1.32 kg COD/m3 ·day.
However, Dereli et al. [113] indicated that OLR could not be an independent parameter and therefore
should be assessed along with SRT.
High performance integrated anaerobic–aerobic fixed-film pilot-scale reactor with an arranged
media for treating slaughter house wastewater showed a removal efficiency of 93% at OLR of 0.77 kg
COD with methane yield of 0.38 m3 CH4 /kg COD [114]. High mixing as a result of system integration
causes low extension of the anaerobic process and consequently affects the methanogenic activities.
A fact known to most researchers in the field of anaerobic digestion is that an increase in OLR
115
Energies 2018, 11, 1675
could result in excessive VFA formation which may inhibit microbial activities and deteriorate the
system. For example, the research of Saddoud et al. [115] confirms that accumulation of VFA was the
main reason for methanogenic inhibition, thus, resulting in a decrease in methane yield at OLR of
16.3 kg COD/m3 in one-phase AnMBR. To counteract this effect, they proposed the use of a two-stage
AnMBR coupled with an anaerobic filter as acidogenic reactor while a jet flow AnMBR was used as
methanogenic reactor at high OLR and realized a substantial improvement in biogas production in
the subsequent stage. Jeison et al. [116] presented removal efficiency below 50% using continuous
stirrer tank reactor CSTR fed with acidified and partially acidified substrate. Even though the organic
loading rates reached 10–17 kg COD/m3 ·day, the low removal efficiency might be connected to
soluble microbial product level within the AnMBR and these could cause irreversible pore fouling in
the membranes.
5.4. Effect of HRT and SRT

Operational parameters such as hydraulic retention time (HRT) and solid retention time (SRT) are
two factors that play a vital role in the treatment performance of AnMBRs. A study by Mei et al. [61]
demonstrated using submerged anaerobic membrane bioreactors (SAnMBRs). SRTs of 30, 60 and
infinite and HRT of 12, 10 and 8 h were used. A total COD removal efficiency >97% was observed
at all operating conditions. Biogas production rate reached 0.056 L CH4/g MLVSS day at an infinite
SRT. However membrane fouling occurred as a result of shorter HRT and infinite SRT. It was also seen
that, longer SRT was the main cause of higher SMP production. The presence of SMP subsequently
introduces more nutrients onto the membrane surface that caused the blockage of the pores and
enhanced biocake formation. Dong et al. [117] reported the influence of SRT and HRT on bioprocess
performance of pilot and bench scale AnMBRs using municipal wastewater as substrate. The SRT
and HRT applied were 40–10 day and 2.5 to 8.5 h. A good permeate quality with COD concentration
(40 mg/L) and BOD5 (10 mg/L) was observed in all conditions. The range of values tested for SRT and
HRT have not considerably interfered with COD and BOD5 removal efficiencies. Moreover prolonged
SRTs caused a reduced sludge production and increase methane yield.
Similarly, in the research of Ozgun et al. [65], reduction in permeate COD concentrations
from 16.5 to 5 mg/L resulted in an increased methane yield from 0.12 to 0.25 L CH4 /g COD.
The improvement in the quality of effluent and the methane was attributed to prolonged SRT from
30 days to infinite. Salazar et al. [39] and Hu et al. [28] revealed a permeate COD concentration increase
from 10 to 50 mg/L with decreasing HRT. However, Liao et al. [35] strongly suggested that, lower HRTs
give room for shorter contact time between microorganisms and substrate and therefore might pave
way for a part of influent COD leaving the reactor without proper treatment. One very important
aspect of AnMBRs is the enabling environment that allows SRT to be completely independent from
HRT irrespective of the sludge properties. Conversely, experience and frequent practice shows that
longer SRTs operation yield more quantities of biogas. This is because any reduction in the SRT may
decrease the extent of reactions required for stable digestion. For instance Huang et al. [66] clearly
reported a methane yield of (0.670 ± 0.203 L CH4 /day), (0.906 ± 0.357 L CH4 /day), (1.290 ± 0.267 L
CH4 /day) at longer SRT of 30, 60, and infinite days respectively. Therefore, it is apparent that,
longer SRT in AnMBRs operations give room for minimal sludge production, and hence cuts disposal
cost significantly. Based on the studies on the effects of HRT and SRT conducted so far, it could be seen
that prolonging HRT may result in inadequate utilization of AnMBRs’ volume and its reduction may
lead to rapid VFA accumulation which may hinder methanogenic activities. Prolonged SRT might also
result to membrane fouling. It could also encourage the release of soluble microbial products (SMP) as
well as rapid cake formation and excessive decline in flux.
6. Inhibitors
Inhibitory substances are the primary cause of disparity in anaerobic reactors. This might also lead
to entire failure of the digestion processes when the concentration is beyond tolerable limits. Anaerobic
116
Energies 2018, 11, 1675
digestion usually presents different variations in the level of inhibition and toxicity. Mechanisms
such as synergism, acclimation, and complexity substrate could considerably upset the phenomenon
of inhibition. Nevertheless, bioreactor failures are regularly reported as the result of high ammonia
inhibition which directly affects the microbial activity [118].
Kayhanian et al. [119], reported ammonia as one of the inhibitory substances found in anaerobic
digestion. It is mainly as a result of biological degradation of nitrogenous matter in the form of
proteins and urea. Inorganic ammonia nitrogen like ammonium ion NH4 + and free ammonia (FA) are
usually found in aqueous form. FA is considered as the primary cause of inhibition since it is freely
membrane-permeable [120,121]. Calli et al. [122] reveal an apparent COD removal of 78–96% in a
study of the effect of high free ammonia concentrations in synthetic wastewater using UASB reactor.
The reactor was fed at OLR of 1.2 kg COD m3 /day with total ammonia nitrogen concentration total
ammonia nitrogen (TAN) increasing from 1000 to 6000 mg/L. However, Tchobanoglous et al. [123]
showed that, the accumulation of propionic acid along with reduction in eubacterial suggest a
sensitivity of propionate degrading acetogenic bacteria to free ammonia than methanogenic archaea.
A theoretical stoichiometric relationship of estimating the quantity of ammonia that could be generated
from organic substrate in anaerobic biodegradation is shown in Equation (1):
CaHbOcNd + ((4a − b − 2c + 3d)/4)H2 O → ((4a + b − 2c − 3d)/8)CH4 + ((4a − b + 2c + 3d)/8)CO2 + dNH3 (1)
According to Kayhanian [124], methanogens are the most sensitive to ammonia inhibition among
the four bacteria types that exist in anaerobic digestion processes. The effect might even cause the
bacteria to cease to grow. Study on the effects of ammonia on propionate degradation and microbial
community in bioreactors showed that, using propionate as a sole carbon source result to reactor failure
after four hydraulic retention times [125]. A total ammonia nitrogen (TAN) concentration of 2.5 g NL−1
at OLR of 0.8 g propionic acid (HPr)/L·day were observed and 95% of the degraded HPr was converted
to methane. On the average, the degradation rate of HPr is below 53%, likewise an average of 74%
and 99% HPr degradation and methane recovery rates was also recorded during the last HRT. Thus,
these behaviors demonstrate an alteration of the microbial community. Frequent maintenance, change
in intracellular pH of methanogens and inhibition of a specific enzyme reactions are some of the
numerous suggested pathways for overcoming ammonia inhibition. Thorough understanding of
the ammonia toxicity occurrence which ammonia may affect methanogenic bacteria is not readily
available but, is the few available studies with unadulterated cultures this was revealed to influence
the treatment in two ways: (i) direct inhibition of methane- producing enzymes by ammonium ion
and/or (ii) the hydrophobic nature of ammonia molecules which may diffuse passively into bacterial
cells causing proton imbalance [126].
Another important anaerobic digestion inhibiting parameter is sulfate. Sulfate is a common
constituent of many industrial wastewaters that is converted in to sulfide by the sulfate-reducing
bacteria (SRB) [127]. Major SRB includes complete and incomplete oxidizers. Complete oxidizers
convert acetate to CO2 and HCO3 − completely. Compounds such as lactate are reduced to acetate
and CO2 by incomplete oxidizers. Furthermore, primary and secondary inhibitions are the two stages
that exist during sulfate reduction [128]. Methane production is suppressed by primary inhibitors as a
result of competition for common organic and inorganic substrate by the SBR, but secondary inhibition
is caused by various bacterial groups due to the toxicity of sulfide [129,130].
7. Membrane Fouling
Membrane fouling is the result of gradual accumulation of suspended solids (SS) and dissolved
solids (DS) mostly in the form of fats, oil and grease onto the surface of the membrane [131].
Mechanisms that propel fouling occur via: (1) deposition of sludge flocs onto the surface, (2) adsorption
of solutes within the membrane, (3) surface cake layer formation, and (4) the shear force that exists
between membrane surface, soluble microbial products (SMPs) and extracellular polymeric substances
117
Energies 2018, 11, 1675
(EPS) [132]. The surface blockage or clogging phenomenon is always attributed to the type of
membrane itself, biomass and operating conditions. Fouling due to the membrane itself could
be attributed to the nature of the material, configuration, hydrophobicity, porosity and pore size
of the membrane [8]. Moreover, biomass characteristic such as MLSS, EPS or SMP, flock structure
and dissolved matter contribute immensely to causing fouling [4]. Conditions at which membrane
bioreactor operates might also increase fouling if not properly monitored. Other factors like cross
flow velocity, HRT or SRT, aeration and transmembrane pressure could also contribute significantly to
fouling [133].
Zhang et al. [134] further revealed that, the use of adsorbent/flocculants in (AnMBR) could
rarely overcome the effect of fouling completely. However, their research revealed that, among the
combination of eight additives (three powdered activated carbon PACs, two granular activated carbons,
one cationic polymer, and two metal salts), 400 mg/L of PAC was able to reduce transmembrane
pressure rise from 0.94 to 0.06 kPa/h. This outcome signified an outstanding fouling reduction
technique. Still, the effect of increasing in OLR and reduction of HRT (24–18 h) is clearly confirmed
as the main factor responsible for severe membrane fouling, but research by Jeison et al. [67] showed
that, extracellular polymeric compounds (EPS) from microbial cells was responsible for the release
of SMP and a drastic increase in TMP during the long term operation that led to the occurrence of
fouling. Filtration performance of membrane bioreactors is largely dependent on the mixed liquor
suspended solids. In the experiments performed by Meng et al. [135], filtration resistance including
membrane resistance (12%), cake resistance (80%), blocking and irremovable fouling resistance (8%),
depicts that the formation of cake layer is the main cause of membrane fouling. Lee et al. [136] pointed
out that cake layers are not uniformly distributed on the entire surface of all of the membrane fibers.
It could be wholly covered by a static sludge cake that could not be removed by the sheer force due
to aeration, and partially by a thin sludge film that might frequently washed away due to aeration
turbulence. A study of Jeison [67] concluded that cake layer formation could be removed despite
longer operation period (>200 days). However, during these periods, cake consolidation was seen
close to the critical flux. A normal backwashing cycle was unable to remove the consolidated cake and
as such physical external treatment was applied. Similarly, Cho et al. [131] and Di-Bella et al. [137]
showed that cake layer formations on AnMBRs coupled with external aeration zone could easily be
removed. This study suggest that AnMBRs operating with aerobic zones external to the membrane
module could achieve higher cake layer removal as compared to cake layer formed on membrane
embedded within an anaerobic bioreactor.
8. Conclusions
They has been a quite significant development of technologies towards meeting the stringent
environmental regulatory discharge requirements and biogas production. These can be seen in
the manner, in which MBRs are configured, variations of operational parameters, and the different
techniques adopted towards provision of the basic living conditions for microorganisms to strive.
However, membrane lifespan is still the main concern of stakeholders in the water and wastewater
treatment industries. The efficiency of physical, chemical and biological methods of reversing fouling
on membrane surface is being exploited. Though physical and chemical methods have been sufficient,
the disadvantages are huge. A lot of energy is consumed during aeration and considerable amount
of chemicals are utilized which does not favor the players in this field with regards to cost and
environmentally-wise. Research is still ongoing on the most economical biological methods of
producing quality effluent and biogas production using AnMBRs, and future studies should take in to
account the following factors:
1. Food to microorganism F/M ratio interaction and HRT of bioreactors operating at thermophilic,
mesophilic or psychrophilic temperature. These will offer more information on biogas production
and effluent quality.
118
Energies 2018, 11, 1675
2. Provision of lining or rough surfaces within the bioreactor (biofilm) would facilitate the retention
of large microbial populations preventing them from exiting the reactor along with the effluent if
shorter HRT is to be used.
3. If the purpose of the treatment is to meet discharge standard limits, and subsequently discharge
the effluent to water bodies, then the use of microfiltration should be encouraged rather than
ultrafiltration. This is because; smaller size membrane surface area would require more energy
for water to pass through and as such applying ultrafiltration for such treatment would be costly.
Author Contributions: M.A.M. and S.I. conceived the concept for the review; M.A.M. and S.I. analyzed the
existing literature; H.C.M. and N.N.N.D. wrote and edited the paper. All authors have read and approved
this review.
Acknowledgments: The work required for the preparation and writing of this review was funded by Universiti
Putra Malaysia (UPM) baseline funding GP-IPS/9633400 awarded to Mohammed Ali Musa.
Conflicts of Interest: The authors declare no conflicts of interest.
References
1. Tock, J.Y.; Lai, C.L.; Lee, K.T.; Tan, K.T.; Bhatia, S. Banana Biomass as Potential Renewable Energy Resource:
A Malaysian Case Study. Renew. Sustain. Energy Rev. 2010, 14, 798–805. [CrossRef]
2. Verstraete, W.; van de Caveye, P.; Diamantis, V. Maximum use of resources present in domestic used water.
3. Martinez-Sosa, D.; Helmreich, B.; Netter, T.; Paris, S.; Bischof, F.; Horn, H. Anaerobic Submerged Membrane
Bioreactor (AnSMBR) for Municipal Wastewater Treatment under Mesophilic and Psychrophilic Temperature
Conditions. Bioresour. Technol. 2011, 102, 10377–10385. [CrossRef] [PubMed]
4. Yoon, Y.; Kim, S.; Oh, S.; Kim, C. Potential of Anaerobic Digestion for Material Recovery and Energy
Production in Waste Biomass from a Poultry Slaughterhouse. Waste Manag. 2014, 34, 204–209. [CrossRef]
[PubMed]
5. Kim, K.Y.; Yang, W.; Ye, Y.; LaBarge, N.; Logan, B.E. Performance of Anaerobic Fluidized Membrane
Bioreactors Using Effluents of Microbial Fuel Cells Treating Domestic Wastewater. Bioresour. Technol. 2016,
6. Kanai, M.; Ferre, V.; Wakahara, S.; Yamamoto, T.; Moro, M. A Novel Combination of Methane Fermentation
and MBR—Kubota Submerged Anaerobic Membrane Bioreactor Process. Desalination 2010, 250, 964–967.
[CrossRef]
7. Harb, M.; Hong, P. Anaerobic Membrane Bioreactor Effluent Reuse: A Review of Microbial Safety Concerns.
Fermentation 2017, 3, 39. [CrossRef]
8. Lin, H.; Peng, W.; Zhang, M.; Chen, J.; Hong, H.; Zhang, Y. A Review on Anaerobic Membrane Bioreactors.
Applications, Membrane Fouling and Future Perspectives. Desalination 2013, 314, 169–188. [CrossRef]
9. Padmasiri, S.I.; Zhang, J.; Fitch, M.; Norddahl, B.; Morgenroth, E.; Raskin, L. Methanogenic Population
Dynamics and Performance of an Anaerobic Membrane Bioreactor (AnMBR) Treating Swine Manure under
High Shear Conditions. Water Res. 2007, 41, 134–144. [CrossRef] [PubMed]
10. Chu, L.B.; Yang, F.L.; Zhang, X.W. Anaerobic Treatment of Domestic Wastewater in a Membrane-Coupled
Expended Granular Sludge Bed (EGSB) Reactor under Moderate to Low Temperature. Process Biochem. 2005,
40, 1063–1070. [CrossRef]
11. Buntner, D.; Sanchez, A.; Garrido, J.M. Feasibility of Combined UASB and MBR System in Dairy Wastewater
Treatment at Ambient Temperatures. Chem. Eng. J. 2013, 230, 475–481. [CrossRef]
12. Deowan, S.A.; Galiano, F.; Hoinkis, J.; Figoli, A.; Drioli, E. Submerged Membrane Bioreactor (SMBR) for
Treatment of Textile Dye Wastewatertowards Developing Novel MBR Process. APCBEE Procedia 2013, 5,
13. Skouteris, G.; Hermosilla, D.; López, P.; Negro, C.; Blanco, Á. Anaerobic Membrane Bioreactors for
Wastewater Treatment. A Review. Chem. Eng. 2012, 198–199, 138–148. [CrossRef]
14. Huang, Z.; Ong, S.L.; Ng, H.Y. Feasibility of Submerged Anaerobic Membrane Bioreactor (SAMBR) for
Treatment of Low-Strength Wastewater. Water Sci. Technol. 2008, 58, 1925–1931. [CrossRef] [PubMed]
119
Energies 2018, 11, 1675
15. Zhao, L.; Liu, D.; Wang, X. Effect of Several Factors on Peracetic Acid Pretreatment of Sugarcane Bagasse for
Enzymatic Hydrolysis. J. Chem. Technol. Biotechnol. 2007, 82, 115–121. [CrossRef]
16. Van Zyl, P.J.; Wentzel, M.C.; Ekama, G.A.; Riedel, K.J. Design and Start-up of a High Rate Anaerobic
Membrane Bioreactor for the Treatment of a Low pH, High Strength, Dissolved Organic Waste Water.
Water Sci. Technol. 2008, 57, 291–295. [CrossRef] [PubMed]
17. Akram, A.; Stuckey, D.C.; Kram, A.A.; Tuckey, D.C.S. Biomass Acclimatisation and Adaptation during
Start-Up of a Submerged Anaerobic Membrane Bioreactor (SAMBR) Biomass Acclimatisation and Adaptation
during Start-Up of a Submerged Anaerobic Membrane Bioreactor (Sambr). Environ. Technol. 2014, 29,
18. Vyrides, I.; Stuckey, D.C. Saline Sewage Treatment Using a Submerged Anaerobic Membrane Reactor
(SAMBR): Effects of Activated Carbon Addition and Biogas-Sparging Time. Water Res. 2009, 43, 933–942.
[CrossRef] [PubMed]
19. Lin, H.; Liao, B.; Chen, J.; Gao, W.; Wang, L.; Wang, F.; Lu, X. New Insights into Membrane Fouling in a
Submerged Anaerobic Membrane Bioreactor Based on Characterization of Cake Sludge and Bulk Sludge.
20. Giménez, J.B.; Robles, A.; Carretero, L.; Durán, F.; Ruano, M.V.; Gatti, M.N.; Ribes, J.; Ferrer, J.; Seco, A.
Experimental Study of the Anaerobic Urban Wastewater Treatment in a Submerged Hollow-Fibre Membrane
Bioreactor at Pilot Scale. Bioresour. Technol. 2011, 102, 8799–8806. [CrossRef] [PubMed]
21. Gao, W.J.; Leung, K.T.; Qin, W.S.; Liao, B.Q. Effects of temperature and temperature Shock on the Performance
and Microbial Community Structure of a Submerged Anaerobic Membrane Bioreactor. Bioresour. Technol.
2011, 102, 8733–8740. [CrossRef] [PubMed]
22. Zhang, X.; Wang, Z.; Wu, Z.; Wei, T.; Lu, F.; Tong, J.; Mai, S. Membrane Fouling in an Anaerobic Dynamic
Membrane Bioreactor (AnDMBR) for Municipal Wastewater Treatment: Characteristics of Membrane
Foulants and Bulk Sludge. Process Biochem. 2011, 46, 1538–1544. [CrossRef]
23. Lin, H.; Chen, J.; Wang, F.; Ding, L.; Hong, H. Feasibility evaluation of submerged anaerobic membrane
bioreactor for municipal secondary wastewater treatment. Desalination 2011, 280, 120–126. [CrossRef]
24. Katayon, S.; Noor, M.J.M.M.; Ahmad, J.; Ghani, L.A.A.; Nagaoka, H.; Aya, H. Effects of Mixed Liquor
Suspended Solid Concentrations on Membrane Bioreactor Efficiency for Treatment of Food Industry
Wastewater. Desalination 2004, 167, 153–158. [CrossRef]
25. Hogye, S.P. National Small Flows Clearinghouse at West Virginia University, Morgantown; National
Environmental Services Center (NESC): Morgantown, WV, USA, 2003.
26. Robinson, A.H. New Developments in the Application of Membrane Bioreactors (MBR) for Industrial Wastewater
Treatment; Wehrle Environmental: Witney, UK, 2001.
27. Umaiyakunjaram, R.; Shanmugam, P. Study on Submerged Anaerobic Membrane Bioreactor (SAMBR)
Treating High Suspended Solids Raw Tannery Wastewater for Biogas Production. Bioresour. Technol. 2016,
28. Hu, A.Y.; Stuckey, D.C. Treatment of dilute wastewaters using a Novel Submerged Anaerobic Membrane
Bioreactor. J. Environ. Eng. 2006, 132, 190–198. [CrossRef]
29. Fuchs, W.; Binder, H.; Mavrias, G.; Braun, R. Anaerobic Treatment of Wastewater with High Organic Content
Using a Stirred Tank Reactor Coupled with a Membrane Filtration Unit. Water Res. 2003, 37, 902–908.
[CrossRef]
30. Saddoud, A.; Ellouze, M.; Dhouib, A.; Sayadi, S. Anaerobic Membrane Bioreactor Treatment of Domestic
Wastewater in Tunisia. Desalination 2007, 207, 205–215. [CrossRef]
31. Dereli, R.K.; Heffernan, B.; Grelot, A.; van der Zee, F.P.; van Lier, J.B. Influence of High Lipid Containing
Wastewater on Filtration Performance and Fouling in AnMBRs Operated at Different Solids Retention Times.
Sep. Purif. Technol. 2015, 139, 43–52. [CrossRef]
32. Van Lier, J.B. High-Rate Anaerobic Wastewater Treatment: Diversifying from End-of-the-Pipe Treatment to
Resource-Oriented Conversion Techniques. Water Sci. Technol. 2008, 57, 1137–1148. [CrossRef] [PubMed]
33. Ho, J.; Sung, S. Methanogenic Activities in Anaerobic Membrane Bioreactors (AnMBR) Treating Synthetic
Municipal Wastewater. Bioresour. Technol. 2010, 101, 2191–2196. [CrossRef] [PubMed]
34. Kocadagistan, E.; Topcu, N. Treatment Investigation of the Erzurum City Municipal Wastewaters with
Anaerobic Membrane Bioreactors. Desalination 2007, 216, 367–376. [CrossRef]
120
Energies 2018, 11, 1675
35. Liao, B.; Kraemer, J.T.; Bagley, D.M. Anaerobic Membrane Bioreactors. Applications and Research Directions
Anaerobic Membrane Bioreactors. Applications and Research Directions. Crit. Rev. Environ. Sci. Technol.
2006, 36, 489–530. [CrossRef]
36. Smith, A.L.; Skerlos, S.J.; Raskin, L. Psychrophilic Anaerobic Membrane Bioreactor Treatment of Domestic
Wastewater. Water Res. 2013, 47, 1655–1665. [CrossRef] [PubMed]
37. Lew, B.; Tarre, S.; Beliavski, M.; Dosoretz, C.; Green, M. Anaerobic Membrane Bioreactor (AnMBR) for
Domestic Wastewater Treatment. Desalination 2009, 243, 251–257. [CrossRef]
38. Martinez-Sosa, D.; Helmreich, B.; Horn, H. Anaerobic Submerged Membrane Bioreactor (AnSMBR) Treating
Low-Strength Wastewater under Psychrophilic Temperature Conditions. Process Biochem. 2012, 47, 792–798.
[CrossRef]
39. Salazar-Peláez, M.L.; Morgan-Sagastume, J.M.; Noyola, A. Influence of Hydraulic Retention Time on Fouling
in a UASB Coupled with an External Ultrafiltration Membrane Treating Synthetic Municipal Wastewater.
Desalination 2011, 277, 164–170. [CrossRef]
40. Gao, D.W.; Zhang, T.; Tang, C.Y.Y.; Wu, W.M.; Wong, C.Y.; Lee, Y.H.; Yeh, D.H.; Criddle, C.S. Membrane
Fouling in an Anaerobic Membrane Bioreactor: Differences in Relative Abundance of Bacterial Species in the
Membrane Foulant Layer and in Suspension. J. Membr. Sci. 2010, 364, 331–338. [CrossRef]
41. Giménez, J.B.; Martí, N.; Ferrer, J.; Seco, A. Methane Recovery Efficiency in a Submerged Anaerobic
Membrane Bioreactor (SAnMBR) Treating Sulphate-Rich Urban Wastewater: Evaluation of Methane Losses
with the Effluent. Bioresour. Technol. 2012, 118, 67–72. [CrossRef] [PubMed]
42. Svojitka, J.; Dvorak, L.; Studer, M.; Straub, J.O.; Fromelt, H.; Wintgens, T. Performance of an Anaerobic
Membrane Bioreactor for Pharmaceutical Wastewater Treatment. Bioresour. Technol. 2017, 229, 180–189.
[CrossRef] [PubMed]
43. Kaya, Y.; Bacaksiz, A.M.; Bayrak, H.; Gnder, Z.B.; Vergili, I.; Hasar, H.; Yilmaz, G. Treatment of Chemical
Synthesis-Based Pharmaceutical Wastewater in an Ozonation-Anaerobic Membrane Bioreactor (AnMBR)
System. Chem. Eng. J. 2017, 322, 293–301. [CrossRef]
44. Mei, X.; Wang, Z.; Miao, Y.; Wu, Z. A Pilot-Scale Anaerobic Membrane Bioreactor under Short Hydraulic
Retention Time for Municipal Wastewater Treatment: Performance and Microbial Community Identification.
J. Water Reuse Desalin. 2017. [CrossRef]
45. Erkan, H.S.; Engin, G.O. The Investigation of Paper Mill Industry Wastewater Treatment and Activated
Sludge Properties in a Submerged Membrane Bioreactor. Water Sci. Technol. 2017. [CrossRef] [PubMed]
46. Hu, D.; Su, H.; Chen, Z.; Cui, Y.; Ran, C.; Xu, J.; Xiao, T.; Li, X.; Wang, H.; Tian, Y.; et al. Performance
Evaluation and Microbial Community Dynamics in a Novel AnMBR for Treating Antibiotic Solvent
Wastewater. Bioresour. Technol. 2017, 243, 218–227. [CrossRef] [PubMed]
47. Hu, D.; Xiao, T.; Chen, Z.; Wang, H.; Xu, J.; Li, X.; Su, H.; Zhang, Y. Effect of the High Cross Flow Velocity on
Performance of a Pilot-Scale Anaerobic Membrane Bioreactor for Treating Antibiotic Solvent Wastewater.
48. Hu, D.; Xu, J.; Chen, Z.; Wu, P.; Wang, Z.; Wang, P.; Xiao, T.; Su, H.; Li, X.; Wang, H.; et al. Performance of a
Pilot Split-Type Anaerobic Membrane Bioreactor (AnMBR) Treating Antibiotics Solvent Wastewater at Low
Temperatures. Chem. Eng. J. 2017, 325, 502–512. [CrossRef]
49. Chen, H.; Chang, S.; Guo, Q.; Hong, Y.; Wu, P. Brewery Wastewater Treatment Using an Anaerobic Membrane
Bioreactor. Biochem. Eng. J. 2016, 105, 321–331. [CrossRef]
50. Xia, T.; Gao, X.; Wang, C.; Xu, X.; Zhu, L. An Enhanced Anaerobic Membrane Bioreactor Treating Bamboo
Industry Wastewater by Bamboo Charcoal Addition: Performance and Microbial Community Analysis.
51. Ng, K.K.; Shi, X.; Ng, H.Y. Evaluation of System Performance and Microbial Communities of Abioaugmented
Anaerobic Membrane Bioreactor Treating Pharmaceutical Wastewater. Water Res. 2015, 81, 311–324.
[CrossRef] [PubMed]
52. Smith, A.L.; Skerlos, S.J.; Raskin, L. Anaerobic Membrane Bioreactor Treatment of Domestic Wastewater
at Psychrophilic Temperatures Ranging from 15 ◦ C to 3 ◦ C. Environ. Sci. Water Res. Technol. 2015, 1, 56–64.
[CrossRef]
53. Dolejs, P.; Ozcan, O.; Bair, R.; Ariunbaatar, J.; Bartacek, J.; Lens, P.N.L.; Yeh, D.H. Effect of Psychrophilic
Temperature Shocks on a Gas-Lift Anaerobic Membrane Bioreactor (Gl-AnMBR) Treating Synthetic Domestic
Wastewater. J. Water Process Eng. 2017, 16, 108–114. [CrossRef]
121
Energies 2018, 11, 1675
54. Gouveia, J.; Plaza, F.; Garralon, G.; Fdz-polanco, F.; Peña, M. Long-Term Operation of a Pilot Scale Anaerobic
Membrane Bioreactor (AnMBR) for the Treatment of Municipal Wastewater under Psychrophilic Conditions.
55. Chen, L.; Gu, Y.; Cao, C.; Zhang, J.; Ng, J.W.; Tang, C. Performance of a Submerged Anaerobic Membrane
Bioreactor with Forward Osmosis Membrane for Low-Strength Wastewater Treatment. Water Res. 2014, 50,
56. Wu, B.; Li, Y.; Lim, W.; Lee, S.L.; Guo, Q.; Fane, A.G.; Liu, Y. Single-Stage versus Two-Stage Anaerobic
Fluidized Bed Bioreactors in Treating Municipal Wastewater: Performance, Foulant Characteristics,
and Microbial Community. Chemosphere 2017, 171, 158–167. [CrossRef] [PubMed]
57. Zhu, X.; Li, M.; Zheng, W.; Liu, R.; Chen, L. Performance and Microbial Community of a Membrane
Bioreactor System-Treating Wastewater from Ethanol Fermentation of Food Waste. J. Environ. Sci. 2017, 53,
58. Cerón-Vivas, A.; Noyola, A. Fouling Membrane in an Anaerobic Membrane Bioreactor Treating Municipal
Wastewater. Water Pract. Technol. 2017, 12, 314–321. [CrossRef]
59. Chen, C.; Guo, W.S.; Ngo, H.H.; Liu, Y.; Du, B.; Wei, Q.; Wei, D.; Nguyen, D.D.; Chang, W.S. Evaluation
of a Sponge Assisted-Granular Anaerobic Membrane Bioreactor (SG-AnMBR) for Municipal Wastewater
Treatment. Renew. Energy 2017, 111, 620–627. [CrossRef]
60. Gurung, K.; Ncibi, M.C.; Sillanpää, M. Assessing Membrane Fouling and the Performance of Pilot-Scale
Membrane Bioreactor (MBR) to Treat Real Municipal Wastewater during Winter Season in Nordic Regions.
Sci. Total Environ. 2017, 579, 1289–1297. [CrossRef] [PubMed]
61. Mei, X.; Quek, P.J.; Wang, Z.; Ng, H.Y. Alkali-Assisted Membrane Cleaning for Fouling Control of Anaerobic
Ceramic Membrane Bioreactor. Bioresour. Technol. 2017, 240, 25–32. [CrossRef] [PubMed]
62. Nie, Y.; Tian, X.; Zhou, Z.; Li, Y. Impact of Food to Microorganism Ratio and Alcohol Ethoxylate Dosage
on Methane Production in Treatment of Low-Strength Wastewater by a Submerged Anaerobic Membrane
Bioreactor. Front. Environ. Sci. Eng. 2017, 11, 6. [CrossRef]
63. Mei, X.; Wang, Z.; Miao, Y.; Wu, Z. Recover Energy from Domestic Wastewater Using Anaerobic Membrane
Bioreactor: Operating Parameters Optimization and Energy Balance Analysis. Energy 2016, 98, 146–154.
[CrossRef]
64. Ozgun, H.; Tao, Y.; Ersahin, M.E.; Zhou, Z.; Gimenez, J.B.; Spanjers, H.; van Lier, J.B. Impact of Temperature
on Feed-Flow Characteristics and Filtration Performance of an Upflow Anaerobic Sludge Blanket Coupled
Ultrafiltration Membrane Treating Municipal Wastewater. Water Res. 2015, 83, 71–83. [CrossRef] [PubMed]
65. Ozgun, H.; Gimenez, J.B.; Ersahin, M.E.; Tao, Y.; Spanjers, H.; van Lier, J.B. Impact of Membrane Addition
for Effluent Extraction on the Performance and Sludge Characteristics of Upflow Anaerobic Sludge Blanket
Reactors Treating Municipal Wastewater. J. Membr. Sci. 2015, 479, 95–104. [CrossRef]
66. Huang, Z.; Ong, S.L.; Ng, H.Y. Submerged Anaerobic Membrane Bioreactor for Low-Strength Wastewater
Treatment: Effect of HRT and SRT on Treatment Performance and Membrane Fouling. Water Res. 2011, 45,
67. Jeison, D.; van Lier, J.B. Cake Formation and Consolidation: Main Factors Governing the Applicable Flux in
Anaerobic Submerged Membrane Bioreactors (AnSMBR) Treating Acidified Wastewaters. Sep. Purif. Technol.
2007, 56, 71–78. [CrossRef]
68. Fallah, N.; Bonakdarpour, B.; Nasernejad, B.; Moghadam, M.R.A. Long-Term Operation of Submerged
Membrane Bioreactor (MBR) for the Treatment of Synthetic Wastewater Containing Styrene as Volatile
Organic Compound (VOC): Effect of Hydraulic Retention Time (HRT). J. Hazard. Mater. 2010, 178, 718–724.
[CrossRef] [PubMed]
69. Ho, J.; Sung, S. Anaerobic Membrane Bioreactor Treatment of Synthetic Municipal Wastewater at Ambient
Anaerobic Membrane Bioreactor Treatment of Synthetic Municipal Wastewater at Ambient Temperature.
Water Environ. Res. 2009, 81, 922–928. [CrossRef] [PubMed]
70. Wijekoon, K.C.; Visvanathan, C.; Abeynayaka, A. Effect of Organic Loading Rate on VFA Production, Organic
Matter Removal and Microbial Activity of a Two-Stage Thermophilic Anaerobic Membrane Bioreactor.
122
Energies 2018, 11, 1675
71. Speece, R.E.; Boonyakitsombut, S.; Kim, M.; Azbar, N.; Ursillo, P. Overview of Anaerobic Treatment:
Thermophilic and Propionate Implications—Keynote Address, Association of Environmental Engineering
and Science Professors, 78th Annual Water Environment Federation Technical Exposition and Conference,
Washington, DC, USA, 27 October–2 November 2005. Water Environ. Res. 2006, 78, 460–473. [PubMed]
72. Ahmed, Z.; Cho, J.; Lim, B.-R.; Song, K.-G.; Ahn, K.-H. Effects of Sludge Retention Time on Membrane
Fouling and Microbial Community Structure in a Membrane Bioreactor. J. Membr. Sci. 2007, 287, 211–218.
[CrossRef]
73. Hiraishi, A.; Ueda, Y.; Ishihara, J. Quinone Profiling of Bacterial Removal Quinone Profiling of Bacterial
Communities in Natural and Synthetic Sewage Activated Sludge for Enhanced Phosphate Removal.
Appl. Environ. Microbiol. 1998, 64, 992–998. [PubMed]
74. Nie, Y.; Chen, R.; Tian, X.; Li, Y.-Y. Impact of Water Characteristics on the Bioenergy Recovery from Sewage
Treatment by Anaerobic Membrane Bioreactor via a Comprehensive Study on the Response of Microbial
Community and Methanogenic Activity. Energy 2017, 139, 459–467. [CrossRef]
75. Kunacheva, C.; Soh, Y.N.A.; Trzcinski, A.P.; Stuckey, D.C. Soluble Microbial Products (SMPs) in the Effluent
from a Submerged Anaerobic Membrane Bioreactor (SAMBR) under Different HRTs and Transient Loading
Conditions. Chem. Eng. J. 2017, 311, 72–81. [CrossRef]
76. Zolfaghari, M.; Drogui, P.; Seyhi, B.; Brar, S.K.; Buelna, G.; Dubé, R.; Klai, N. Investigation on Removal
Pathways of Di 2-Ethyl Hexyl Phthalate from Synthetic Municipal Wastewater Using a Submerged Membrane
Bioreactor. J. Environ. Sci. 2015, 37, 37–50. [CrossRef] [PubMed]
77. Zhang, H.; Jiang, W.; Cui, H. Performance of Anaerobic Forward Osmosis Membrane Bioreactor Coupled
with Microbial Electrolysis Cell (AnOMEBR) for Energy Recovery and Membrane Fouling Alleviation.
Chem. Eng. J. 2017, 321, 375–383. [CrossRef]
78. Xiao, Y.; Yaohari, H.; de Araujo, C.; Sze, C.C.; Stuckey, D.C. Removal of Selected Pharmaceuticals in an
Anaerobic Membrane Bioreactor (AnMBR) with/without Powdered Activated Carbon (PAC). Chem. Eng. J.
2017, 321, 335–345. [CrossRef]
79. Li, Y.; Hu, Q.; Chen, C.H.; Wang, X.L.; Gao, D.W. Performance and Microbial Community Structure in an
Integrated Anaerobic Fluidized-Bed Membrane Bioreactor Treating Synthetic Benzothiazole Contaminated
Wastewater. Bioresour. Technol. 2017, 236, 1–10. [CrossRef] [PubMed]
80. Watanabe, R.; Nie, Y.; Wakahara, S.; Komori, D.; Li, Y.Y. Investigation on the Response of Anaerobic
Membrane Bioreactor to Temperature Decrease from 25 ◦ C to 10 ◦ C in Sewage Treatment. Bioresour. Technol.
2017, 243, 747–754. [CrossRef] [PubMed]
81. Basset, N.; Santos, E.; Dosta, J.; Matalvarez, J. Start-up and Operation of an AnMBR for Winery Wastewater
Treatment. Ecol. Eng. 2016, 86, 279–289. [CrossRef]
82. Chen, C.; Guo, W.; Ngo, H.H.; Chang, S.W.; Nguyen, D.D.; Nguyen, P.D.; Bui, X.T.; Wu, Y. Impact of Reactor
Configurations on the Performance of a Granular Anaerobic Membrane Bioreactor for Municipal Wastewater
Treatment. Int. Biodeterior. Biodegrad. 2017, 121, 131–138. [CrossRef]
83. Iranpour, R.; Cox, H.H.J.; Shao, Y.J.; Moghaddam, O.; Kearney, R.J.; Deshusses, M.A.; Stenstrom, M.K.;
Ahring, B.K. Changing Mesophilic Wastewater Sludge Digestion into Thermophilic Operation at Terminal
Island Treatment Plant. Water Environ. Res. 2002, 74, 494–507. [CrossRef] [PubMed]
84. Barker, D.J.; Stuckey, D.C. A Review of Soluble Microbial Products (SMP) in Wastewater Treatment Systems.
Water Res. 1999, 33, 3063–3082. [CrossRef]
85. Sciener, P.; Nachaiyasit, S.; Stuckey, D.C. Production of Soluble Microbial Products (SMP) in an Anaerobic
Baffled Reactor: Composition, Biodegradability and the Effect of Process Parameters. Environ. Technol. 1998,
4, 391–400. [CrossRef]
86. Chudoba, J.A.N. Inhibitory effect of refractory organic compounds produced by activated sludge
micro-organisms on microbial activity and flocculation. Water Res. 1985, 19, 197–200. [CrossRef]
87. Judd, S. Principles and Applications of Membrane Bioreactors in Water and Wastewater Treatment; Butterworth-
Heinemann: Oxford, UK, 2006.
88. Li, H.; Yang, M.; Zhang, Y.; Yu, T.; Kamagata, Y. Nitrification Performance and Microbial Community
Dynamics in a Submerged Membrane Bioreactor with Complete Sludge Retention. J. Biotechnol. 2006, 123,
123
Energies 2018, 11, 1675
89. Van Den Broeck, R.; van Dierdonck, J.; Nijskens, P.; Dotremont, C.; Krzeminski, P. The Influence of Solids
Retention Time on Activated Sludge Bioflocculation and Membrane Fouling in a Membrane Bioreactor
(MBR). J. Membr. Sci. 2012, 401–402, 48–55. [CrossRef]
90. De Vrieze, J.; Saunders, A.M.; He, Y.; Fang, J.; Nielsen, P.H.; Verstraete, W.; Boon, N. Ammonia and
Temperature Determine Potential Clustering in the Anaerobic Digestion Microbiome. Water Res. 2015, 75,
91. Pap, B.; Györkei, Á.; Boboescu, I.Z.; Nagy, I.K.; Bíró, T.; Kondorosi, É.; Maróti, G. Temperature-Dependent
Transformation of Biogas-Producing Microbial Communities Points to the Increased Importance of
Hydrogenotrophic Methanogenesis under Thermophilic Operation. Bioresour. Technol. 2015, 177, 375–380.
[CrossRef] [PubMed]
92. El-Mashad, H.M.; Zeeman, G.; van Loon, W.K.P.; Bot, G.P.A.; Lettinga, G. Effect of Temperature and
Temperature Fluctuation on Thermophilic Anaerobic Digestion of Cattle Manure. Bioresour. Technol. 2004, 95,
93. Van Lier, J.B. Thermophilic Anaerobic Wastewater Treatment; Temperature Aspects and Process Stability.
Ph.D. Thesis, Wageningen University & Research, Wageningen, The Netherlands, 1995.
94. Duran, M.; Speece, R.E. Temperature-Staged Anaerobic Processes. Environ. Technol. 1997, 18, 747–753.
[CrossRef]
95. Wilson, C.A.; Murthy, S.M.; Fang, Y.; Novak, J.T. The Effect of Temperature on the Performance and Stability
of Thermophilic Anaerobic Digestion. Water Sci. Technol. 2008, 57, 297–304. [CrossRef] [PubMed]
96. Lin, Q.; de Vrieze, J.; Li, J.; Li, X. Temperature Affects Microbial Abundance, Activity and Interactions in
Anaerobic Digestion. Bioresour. Technol. 2016, 209, 228–236. [CrossRef] [PubMed]
97. Song, Y.C.; Kwon, S.J.; Woo, J.H. Mesophilic and Thermophilic Temperature Co-Phase Anaerobic Digestion
Compared with Single-Stage Mesophilic- and Thermophilic Digestion of Sewage Sludge. Water Res. 2004, 38,
98. Angelidaki, I. Anaerobic thermophilic digestion of manure at different ammonialoads: Effect of temperature.
Water Res. 1994, 28, 727–731. [CrossRef]
99. Chandra, R.; Vijay, V.K.; Subbarao, P.M.V.; Khura, T.K. Production of Methane from Anaerobic Digestion of
Jatropha and Pongamia Oil Cakes. Appl. Energy 2012, 93, 148–159. [CrossRef]
100. Appels, L.; van Assche, A.; Willems, K.; Degrève, J.; van Impe, J.; Dewil, R. Peracetic Acid Oxidation as an
Alternative Pre-Treatment for the Anaerobic Digestion of Waste Activated Sludge. Bioresour. Technol. 2011,
101. Zhou, J.; Zhang, R.; Liu, F.; Yong, X.; Wu, X.; Zheng, T.; Jiang, M.; Jia, H. Biogas Production and
Microbial Community Shift through Neutral pH Control during the Anaerobic Digestion of Pig Manure.
102. Wang, S.; Nomura, N.; Nakajima, T.; Uchiyama, H. Case Study of the Relationship between Fungi
and Bacteria Associated with High-Molecular-Weight Polycyclic Aromatic Hydrocarbon Degradation.
J. Biosci. Bioeng. 2012, 113, 624–630. [CrossRef] [PubMed]
103. Kim, J.; Park, C.; Kim, T.-H.; Lee, M.; Kim, S.; Kim, S.-W.; Lee, J. Effects of Various Pretreatments for Enhanced
Anaerobic Digestion with Waste Activated Sludge. J. Biosci. Bioeng. 2003, 95, 271–275. [CrossRef]
104. Kim, M.; Gomec, C.Y.; Ahn, Y.; Speece, R.E. Hydrolysis and Acidogenesis of Particulate Organic Material
in Mesophilic and Thermophilic Anaerobic Digestion. Environ. Technol. 2003, 24, 1183–1190. [CrossRef]
[PubMed]
105. Sandberg, M.; Ahring, B.K. Anaerobic Treatment of Fish Meal Process Waste Water in a UASB Reactor at
High pH. Appl. Microbiol. 1992, 36, 800–804. [CrossRef]
106. Yu, D.; Liu, J.; Sui, Q.; Wei, Y. Biogas-pH Automation Control Strategy for Optimizing Organic Loading Rate
of Anaerobic Membrane Bioreactor Treating High COD Wastewater. Bioresour. Technol. 2016, 203, 62–70.
[CrossRef] [PubMed]
107. Mao, C.; Feng, Y.; Wang, X.; Ren, G. Review on Research Achievements of Biogas from Anaerobic Digestion.
Renew. Sustain. Energy Rev. 2015, 45, 540–555. [CrossRef]
108. Wen, C.; Huang, X.; Qian, Y. Domestic Wastewater Treatment Using an Anaerobic Bioreactor Coupled with
Membrane Filtration. Process Biochem. 1999, 35, 335–340. [CrossRef]
109. Vincent, N.M.; Tong, J.; Yu, D.; Zhang, J.; Wei, Y. Membrane Fouling Characteristics of a Side-Stream Tubular
Anaerobic Membrane Bioreactor (AnMBR) Treating Domestic Wastewater. Processes 2018, 6, 50. [CrossRef]
124
Energies 2018, 11, 1675
110. Qiao, W.; Takayanagi, K.; Niu, Q.; Shofie, M.; Li, Y.Y. Long-Term Stability of Thermophilic Co-Digestion
Submerged Anaerobic Membrane Reactor Encountering High Organic Loading Rate, Persistent Propionate
and Detectable Hydrogen in Biogas. Bioresour. Technol. 2013, 149, 92–102. [CrossRef] [PubMed]
111. An, Y.Y.; Yang, F.L.; Bucciali, B.; Wong, F.S. Municipal Wastewater Treatment Using a UASB Coupled with
Cross-Flow Membrane Filtration. J. Environ. Eng. 2009, 135, 86–91. [CrossRef]
112. Bornare, J.; Kalyyanraman, V.; Sonde, R.R. Application of Anaerobic Membrane Bioreactor (AnMBR) for
Low-Strength Wastewater Treatment and Energy Generation. In Industrial Wastewater Treatment, Recycle and
Reuse; Butterworth-Heinemann: Oxford, UK, 2014; pp. 399–434.
113. Dereli, R.K.; Ersahin, M.E.; Ozgun, H.; Ozturk, I.; Jeison, D.; van der Zee, F.; van Lier, J.B. Potentials of
Anaerobic Membrane Bioreactors to Overcome Treatment Limitations Induced by Industrial Wastewaters.
114. Del Pozo, R.; Diez, V. Integrated Anaerobic-Aerobic Fixed-Film Reactor for Slaughterhouse Wastewater
Treatment. Water Res. 2005, 39, 1114–1122. [CrossRef] [PubMed]
115. Saddoud, A.; Sayadi, S. Application of Acidogenic Fixed-Bed Reactor prior to Anaerobic Membrane
Bioreactor for Sustainable Slaughterhouse Wastewater Treatment. J. Hazard. Mater. 2007, 149, 700–706.
[CrossRef] [PubMed]
116. Jeison, D.; Plugge, C.M.; Pereira, A.; van Lier, J.B. Effects of the Acidogenic Biomass on the Performance
of an Anaerobic Membrane Bioreactor for Wastewater Treatment. Bioresour. Technol. 2009, 100, 1951–1956.
[CrossRef] [PubMed]
117. Dong, Q.; Parker, W.; Dagnew, M. Influence of SRT and HRT on Bioprocess Performance in Anaerobic
Membrane Bioreactors Treating Municipal Wastewater. Water Environ. Res. 2016, 88, 158–167. [CrossRef]
[PubMed]
118. Hejnfelt, A.; Angelidaki, I. Anaerobic Digestion of Slaughterhouse by-Products. Biomass Bioenergy 2009, 33,
1046–1054. [CrossRef]
119. Kayhanian, M. Ammonia Inhibition in High-Solids Biogasification: An Overview and Practical Solutions.
Environ. Technol. 1999, 20, 355–365. [CrossRef]
120. De Baere, L.A.; Devocht, M.; van Assche, P.; Verstraete, W. Influence of High NaCl and NH4 Cl salt levels on
methanogenic associations. Water Res. 1984, 18, 543–548. [CrossRef]
121. Kroeker, E.J.; Shulte, D.D.; Sparling, A.B.; Lapp, H.M. Anaerobic Process Treatment Stability. J. WPCF 1979,
51, 718–727.
122. Calli, B.; Mertoglu, B.; Inanc, B.; Yenigun, O. Effects of High Free Ammonia Concentrations on the
Performances of Anaerobic Bioreactors. Process Biochem. 2005, 40, 1285–1292. [CrossRef]
123. Tchobanoglous, G.; Theisen, H.; Vigil, S. Integrated Waste Management: Engineering Principles and Management
Issues; McGraw-Hill: New York, NY, USA, 1993.
124. Kayhanian, M. Performance of a High-Soilds Anaerobic Digestion Process under Various Ammonia
Concentrations. J. Chem. Technol. Biotechnol. 1994, 59, 349–352. [CrossRef]
125. Li, Y.; Zhang, Y.; Kong, X.; Li, L.; Yuan, Z.; Dong, R.; Sun, Y. Effects of Ammonia on Propionate Degradation and
Microbial Community in Digesters Using Propionate as a Sole Carbon Source. J. Chem. Technol. Biotechnol. 2017.
[CrossRef]
126. Gallert, C.; Bauer, S.; Winter, J. Effect of Ammonia on the Anaerobic Degradation of Protein by a Mesophilic
and Thermophilic Biowaste Population. Appl. Microbiol. Biotechnol. 1998, 50, 495–501. [CrossRef] [PubMed]
127. Koster, I.W.; Rinzema, A.; de Vegt, A.L.; Lettinga, G. Sulfide Inhibition of the Methanogenic activity of
granular sludge at various pH-levels. Water Res. 1986, 20, 1561–1567. [CrossRef]
128. Harada, H.; Uemura, S.; Momonoi, K. Interaction Between Sulfate Reducing Bacteria and Methane Producing
Bacteria In Uasb Reactors Fed With Low Strength Wastes Containing Different Levels of Sulfate. Water Res.
1994, 28, 355–367. [CrossRef]
129. Stefanie, J.W.H.; Elferink, O.; Visser, A.; Pol, L.W.H.; Stams, A.J.M. Sulfate Reduction in Methanogenic
Bioreactors. FEMS Microbiol. Rev. 1994, 15, 119–136. [CrossRef]
130. Colleran, E.; Pender, S.; Philpott, U.; O’Flaherty, V.; Leahy, B. Full-Scale and Laboratory-Scale Anaerobic
Treatment of Citric Acid Production Wastewater. Biodegradation 1998, 9, 233–245. [CrossRef] [PubMed]
131. Cho, J.; Song, K.G.; Yun, H.; Ahn, K.H.; Kim, J.Y.; Chung, T.H. Quantitative Analysis of Biological Effect
on Membrane Fouling in Submerged Membrane Bioreactor. Water Sci. Technol. 2005, 51, 9–18. [CrossRef]
[PubMed]
125
Energies 2018, 11, 1675
132. Colleran, E.; Finnegan, S.; Lens, P. Anaerobic Treatment of Sulfate-Containing Waste Streams. Antonie Van
Leeuwenhoek Int. J. Gen. Mol. Microbiol. 1995, 67, 29–46. [CrossRef]
133. Wang, Z.; Ma, J.; Tang, C.Y.; Kimura, K.; Wang, Q. Membrane Cleaning in Membrane Bioreactors: A Review.
J. Membr. Sci. 2014, 468, 276–307. [CrossRef]
134. Zhang, Q.; Singh, S.; Stuckey, D.C. Fouling Reduction Using Adsorbents/Flocculants in a Submerged
Anaerobic Membrane Bioreactor. Bioresour. Technol. 2017, 239, 226–235. [CrossRef] [PubMed]
135. Meng, F.; Chae, S.R.; Drews, A.; Kraume, M.; Shin, H.S.; Yang, F. Recent Advances in Membrane Bioreactors
(MBRs): Membrane Fouling and Membrane Material. Water Res. 2009, 43, 1489–1512. [CrossRef] [PubMed]
136. Lee, J.; Ahn, W.; Lee, C. Comparison of the Filtration Characteristics Between Attached and Suspended
Growth Microorganisms in Submerged Membrane Bioreactor. Water Res. 2001, 35, 2435–2445. [CrossRef]
137. Di Bella, G.; Durante, F.; Torregrossa, M.; Viviani, G.; Mercurio, P.; Cicala, A. The Role of Fouling Mechanisms
in a Membrane Bioreactor. Water Sci. Technol. 2007, 55, 455–464. [CrossRef] [PubMed]
126
energies
Article
Biogas Quality across Small-Scale Biogas Plants:
A Case of Central Vietnam
Hynek Roubík 1 , Jana Mazancová 1 , Phung Le Dinh 2 , Dung Dinh Van 2 and Jan Banout 1, *
1 Department of Sustainable Technologies, Faculty of Tropical AgriSciences, Czech University of Life Sciences
Prague, Kamýcká 129, 165 00 Prague, Czech Republic; [email protected] (H.R.); [email protected] (J.M.)
2 Department of Animal Sciences and Veterinary Medicine, Hue University of Agriculture and Forestry,
Hue University, 102 Phung Hung Hue City Vietnam, Thua Thien Hue 53000, Vietnam;
[email protected] (P.L.D.); [email protected] (D.D.V.)
Received: 25 May 2018; Accepted: 26 June 2018; Published: 9 July 2018
Abstract: Production of bioenergy by the fermentation reaction is gaining attraction due to its easy
operation and the wide feedstock selection. Anaerobic fermentation of organic waste materials
is generally considered a cost-effective and proven technology, allowing simultaneous waste
management and energy production. Small-scale biogas plants are widely and increasingly used
to transform waste into gas through anaerobic fermentation of organic materials in the developing
world. In this research, the quality of biogas produced in small-scale biogas plants was evaluated,
as it has a direct effect on its use (as fuel for biogas cookers), as well as being able to influence a
decision making process over purchasing such technology. Biogas composition was measured with
a multifunctional portable gas analyser at 107 small-scale biogas plants. Complementary data at
household level were collected via the questionnaire survey with the owners of biogas plants (n = 107).
The average daily biogas production equals 0.499 m3 , not covering the demand of rural households
which are using other sources of energy as well. Related to the biogas composition, the mean content
of methane (CH4 ) was 65.44% and carbon dioxide (CO2 ) was 29.31% in the case of biogas plants
younger than five years; and CH4 was 64.57% and CO2 was 29.93% for biogas plants older than five
years. Focusing on the age of small-scale biogas plants there are no, or only minor, differences among
tested values. In conclusion, the small-scale biogas plants are sustaining a stable level of biogas
quality during their life-span.
Keywords: anaerobic digestion; methane; carbon dioxide; small-scale biogas plants; developing countries
1. Introduction
In developing countries, environmental pollution and access to energy sources still represent
challenges, especially in relation to human and environmental health and economic development [1].
Energy influences the status and pace of development; hence, a current challenge for the developing
world lies in the available supply of affordable and sustainable eco-friendly renewable energy
(SDG 7) [2]. Energy poverty is exhibited by a lack of access to electricity and clean cooking facilities,
which are two elements that are essential to meeting basic human needs [3]. Therefore, bioenergy
production by fermentation reaction is gaining attraction due to its easy operation and a wide selection
of organic wastes feedstock [4]. Anaerobic fermentation of organic materials is generally considered as
a major cost-effective and matured technology [5] with its dual benefits as a waste management tool
and simultaneous energy production [6]. Small-scale biogas plants are widely and increasingly used
to transform waste into valuable gas in the developing world [7,8] and may represent an economically
feasible technology [9,10], which is producing biogas as a main product and simultaneously producing
digestate (which may be used as fertilizer) as a by-product through waste degradation [9,11].

Energies 2018, 11, 1794
There are many factors affecting household energy consumption (as well as CO2 emissions),
such as socio-economic factors, household characteristics and geographic factors [12].
The small-scale biogas plants also create a number of indirect environmental, economic, and
societal benefits, such as a reduction in deforestation, fewer hours devoted to fuelwood collection
or savings on fuelwood/fossil fuels purchasing, decreasing the need to purchase propane for
cooking, the creation of jobs, the decrease of organic matter in effluent waters, the decrease of odour,
the production of less indoor smoke than other fuels, the production of digestate as a fertilizer
and a reduction in greenhouse gas emissions into the atmosphere, if used appropriately [6,7,13–15].
According to Zhang et al. [12] biogas households indicate having over 50% lower greenhouse
gas emissions than non-biogas households. Due to above-mentioned benefits, small-scale biogas
technology has been widely promoted and financially supported by governments and development
aid donors in Asia, including Vietnam [13] Long-term, stable running and maintenance are key points
to maximize the benefits of small-scale biogas plant [13,16]. However, if these key points are not met,
the benefits of this technology may be compromised [5,13,17].
In comparison to other forms of renewable energy (solar energy, biodiesel, bioethanol,
wind energy, etc.) biogas production through small-scale biogas plants is relatively simple,
decentralised, and can operate under various conditions in tropical regions, such as Southeast
Asia, particularly Vietnam. The most common feedstock material is animal dung or human faeces,
as it is usually the most problematic waste material in terms of waste management for rural
households [18]. An important advantage of small-scale biogas technology and one of the main
reasons for the government support is that the technology is a cost-effective method of reducing
greenhouse gas emissions and odours from animal manure, if used properly [13,19]. The biogas
produced is mainly used for cooking, heating and lighting, therefore replacing energy sources such
as fuelwood, dried dung, coal, or liquid petroleum gas (LPG), commonly used for these purposes
in rural households [5,9,13]. It is always difficult to adopt new and unknown digester technology
within households. Therefore, recommendations for various models implemented within the country
are needed. The design of the biogas plants varies based on geographical locations, availability of
feedstock and climatic conditions. The most common types of feedstock for chosen Asian countries can
be seen in Table 1. In Asia, the fixed dome model is the most commonly used [5,16]. However, there are
two exceptions—Indonesia, where various models were applied according to the regions and islands,
and India with a prevailing number of the floating drum model followed by the fixed dome [20].
Table 1. Most common feedstock for small-scale biogas plants in selected Asian countries.
Country Most Common Feedstock Reference

Vietnam Pig manure [7,21]
Cambodia Combination of pig manure and cow manure [22]
Bangladesh Cow and buffaloes manure [5,23]
Laos Cow manure [24]
Nepal Cow manure [25]
India Livestock manure [26]
Indonesia Cow manure [27,28]
In Vietnam, anaerobic digestion of animal manure has already been practiced since the 1960s [7,10].
Since then its popularity has grown, mainly due to the promotion of this technology by the government
and international organizations, e.g., SNV (Netherland Development Organization). The Ministry of
Agriculture and Rural Development of Vietnam (MARD), together with SNV (using 10% government
subsidy for support of capital costs of small-scale biogas technology), installed between 2003 and
2013 over 200,000 small-scale biogas plants [5,29]. The follow-up biogas programme by SNV and
MARD aimed to build 140,000 biogas digesters between 2006 and 2011. The current number of
small-scale biogas plants in Vietnam is more than 500,000 [30]. The target was reached, and digesters
128
Energies 2018, 11, 1794
serve over 600,000 people with cooking fuel with CO2 savings of around 260,000 tons per year [31].
However, in Vietnam, the biogas technology is still far below its potential for utilizing available
livestock and agricultural wastes [30].
As the small-scale biogas technology is one of the fastest growing and highly promising renewable
energy sources, mainly for rural households, the main objective of this paper is to evaluate the quality
of biogas produced in the small-scale biogas plants installed in central Vietnam in terms of chemical
and physical parameters in relation to the age of installed biogas plants. The quality of biogas has
a direct effect on its use (as fuel for biogas cookers), which may, in return, influence an individual
decision of purchasing such a technology. Furthermore, biogas quality evaluation is needed to provide
sufficient information for authorities to have their future policy decisions well supported.

The research was carried out in two districts, Huong Tra and Phong Dien, Thua Thien Hue
Province, in Central Vietnam (Figure 1). Huong Tra is a rural district in the northern part of the
central coast of Vietnam with a population of over 115,000 inhabitants covering an area of 521 km2 .
The district is located on the northern outskirts of Hue (provincial city of Thua Thien Hue province)
and can, therefore, be considered as a peri-urban area. Phong Dien has a population over 105,000
inhabitants and covers an area of 954 km2 . The district has a varied topography with mountains,
plains, and coastline.
Figure 1. Thua Thien Hue province and the target area.
2.1. Description of Biogas Technology in Target Area

The study was done on two types of small-scale biogas plants, specifically KT1 (Figure 2a) and
KT2 (Figure 2b). Both types are predominant in the target area. Both types are varieties of the Chinese
fixed dome, where KT1 is an appropriate type for a good structure of soil to be easily excavated.
KT2 is used in places where soil excavation is difficult or where high levels of groundwater or floods
are reported. Both types are unheated and usually built underground, in order to minimize the
temperature fluctuations and for space saving reasons. The digester is filled in through the inlet tank
and the inlet pipe. The produced biogas is accumulated at the upper part of the digester and the
difference between the slurry inside the digester and the digestate in the compensation tank creates
a gas pressure. The slurry flows back into the digester from the compensation tank after the gas is
released through the gas pipe. Both types and their potential problems are described in detail in the
study by [21]. The Vietnamese small-scale biogas plants operate at the temperature of the surrounding
soil as they are built underground. The time of the year significantly influences temperatures in the
129
Energies 2018, 11, 1794
air, the slurry mixing tank, the soil, and the digesters. The average summer temperatures in Central
Vietnam are around 34 ◦ C (mesophilic conditions), creating a suitable environment for the bacterial
fermentation; however, during winter time the temperature is in the range of 15–25 ◦ C, which might
cause lower biogas production [9,32].
(a)
(b)
Figure 2. (a) Small-scale biogas plant—fixed dome model (KT1). Reprinted from [21]. (b) Small-scale
biogas plant—fixed dome model (KT2). Reprinted from [21].
2.2. Data Collection—Questionnaire Survey

The questionnaire survey was carried out with the owners of small-scale biogas plants from June
to July 2013. Biogas plants were randomly selected from recipients of government subsidies (n = 107;
corresponding to 20% share of total subsidy recipients in the area) listed by the local unit of the Ministry
of Agriculture and Rural Development. The recipients were, at the same time, beneficiaries of one
of two running projects on building small-scale biogas plants—one supported by the SNV, the other
one by the Czech Development Agency (CzDA). The questionnaire included nine questions (Table 2).
Furthermore, the data results were cross-checked with the local facilitators during field trips in July
and August 2016 in order to increase their validity and reliability.
Table 2. Overview of variables in the questionnaire.
Variable Type of the Question Value Unit

Capacity of the digester Close-ended question No. m3
Investor of the construction Close-ended question SNV/CzDA -
Digester type Close-ended question KT1/KT2 -
Digester connection to the toilet Open-ended question - -
Animal stable Open-ended question - -
Feedstock materials for the BGP Open-ended question - -
No. of animals Open-ended question No. Heads
No. of other applications powered by biogas Open-ended question No. -
Digestate practices Open-ended question - -
130
Energies 2018, 11, 1794
2.3. Data Collection—Biogas Composition

There are many analytical methods for biogas quality evaluation based on its final use [33] From
the technical point of view, the most important parameter is the content of the potentially corrosive
components (oxygen, hydrogen, water, carbon dioxide, hydrogen sulphide, and chlorine and fluorine
compounds) [34,35]. Biogas composition was measured using a GA5000 multifunctional portable gas
analyser (Geotech, Leamington Spa, UK), which is adapted to measurements of CH4 , CO2 , O2 , H2 , and
H2 S with the following measurement accuracy of: CH4 (0–70 vol % ± 0.5%), CH4 (70–100 vol % ± 1.5%),
CO2 (0–60 vol % ± 0.5%), CO2 (60–100 vol % ± 1.5%), O2 (0–25 vol % ± 1.0%), H2 S (0–5000
ppm ± 2.0%), and H2 S (0–10,000 ppm ± 5.0%). For measurements of media (CH4 /CO2 ) a
dual-wavelength infrared sensor was used, for O2 /H2 S an internal electrochemical sensor was used.
The measurements were taken upstream of the H2 S filter to eliminate measurement inaccuracies.
Obtained values are the mean of three measurements in the interval of one hour at each biogas plant.
Calorific values were set as the quantity of heat produced by complete combustion of a unit of a
combustible compound. In total, the measurements were taken in 81 KT1 small-scale biogas plants and
26 KT2 small-scale biogas plants at districts Huong Tra (n = 49) and Phong Dien (n = 58), Thua Thien
Hue Province.
2.4. Data Analysis

The collected data were categorized, coded, and analysed with descriptive and inferential statistics
using the SPSS version 18. Effects of variables such as type of biogas plant, age of the biogas plant,
and size of the biogas plant on biogas composition were analysed by the analysis of covariance
model. Dummy variables included the type of biogas plant (KT1 and KT2), the age of the biogas plant
(>5 years old and <5 years old), and a continuous variable of the size of the biogas plants (m3 ).
Yi = α + βx1 + γ1 Di1 + γ2 Di2 + ε i (1)
Yi = Biogas composition;
α = Intercept;
x1 = Biogas plant size (m3 );
Di1 = Biogas plant types (Di1 = 1 = KT2; Di1 = 0 = KT1);
Di2 = Biogas plant age (Di2 = 1 5 years; Di1 = 0 ≥ 5 years);
β = Regression coefficient of biogas plant size on biogas composition;
γ1 = The difference of biogas composition between KT2 and KT1;
γ2 = The difference of biogas composition between <5 year and ≥5 year old biogas plants; and
εi = Error term.
3.1. Feedstock Used for Biogas Production

The majority of respondents (90%) are farmers producing mainly rice. However, many of them
are also involved in additional off-farm activities, such as trade, rice noodle production, and rice wine
production. All questioned households use pig slurry as their main feedstock for biogas plants and,
in all cases, pigs were housed in concrete pigpens (with a concrete floor). Feedstock manure from
other animals (65%) is also used as an additive within the surveyed households. This included chicken
manure (29%) and human excreta (36%) (Figure 3). These additives are added if they are available in
sufficient quantities. In every household, the pigpen is connected directly to the biogas plant, and in
37% of cases toilet outflows are connected to the biogas plant as well. Only in one case was a chicken
shed was connected to the biogas plant; in the rest of cases the chicken manure is put in the digester
inlet manually. Generally, the feedstock input was unified as biogas owners were recipients of one of
two running projects on building small-scale biogas plants and there were criteria on the necessary
131
Energies 2018, 11, 1794
number of pigs. In addition, further details regarding manure management practices of small-scale
farmers in Vietnam can be seen in our previous study [6].
100
% of respondents
80
60
40
20
0
Pig slurry Human excreta Chicken manure
Figure 3. Feedstock used for biogas production (Multiple choices were possible) (n = 107).
3.2. Use of Products of Small-Scale Biogas Technology within the Rural Households in Central Vietnam
Small-scale biogas plants have been applied as an optimal livestock waste treatment in Vietnam
since the 1960s and, although the history of this technology is rather old, the number of the constructed
biogas plants is still limited. With the current number of units around 500,000 [5,23] it is still far
below the real demand on livestock waste treatment [10,36] that has increased significantly in the last
decade [7]. The primary use of biogas is for cooking [21]. However, it could also be used for lighting in
remote areas where electrification is limited [29]. Furthermore, biogas plants produce residue from
the process in the form of digestate, which can be applied as organic fertilizer to enhance agriculture
production [10,17]. On a daily basis, a minimum of 20 kg of organic waste is required to operate the
smallest biogas plants (4 m3 ) in the area. Therefore, there is a required number (five growing pigs)
of pigs (manure, respectively) to meet the feedstock needs of these plants. This quota is usually met,
as the average number of animals in the area is 13–14 piglets and 2–3 sows [6,21].
3.2.1. Biogas for Cooking and Water Boiling

In developing countries, biogas produced from the household biogas plants is used mainly for
cooking [29]. This is also applied to the present study, where 100% of households use biogas primarily
for cooking, with an average time of 2.8 h/day. Biogas is usually used primarily for household
cooking and boiling water and, afterwards, for cooking of feed for pigs. An average daily production
of biogas is 0.499 m3 (±0.086). Such an amount used for cooking purposes may represent about
8–10 m3 and 96–120 m3 of biogas per month and year, respectively. However, according to the study
by [37], the biogas volume needed for a typical farming household of six people is 0.8 to 1 m3 per day.
This difference between average daily biogas productions could be explained by [38], as he states that
the fixed dome biogas digesters can annually leak around 55% of CH4 and the production of biogas is
also dependent on the temperature of the feedstock. Therefore, the majority of respondents (60%) are
still using additional energy sources in form of LPG and/or electricity (for cooking rice in rice cookers)
and fuelwood (usually only for cooking of feed for animals).
3.2.2. Biogas for Lightning and Power Generation

The other major possible application of biogas may be for lighting and power generation.
Biogas lamps are more efficient than kerosene-powered lamps, but their efficiency is quite low
compared to electric-powered lamps [18]. In addition, electricity is now widely available in Vietnam;
therefore, use of biogas lamps is very occasional. In the case of our respondents, less than 10% were
using biogas lamps. Farmers usually prefer biogas for cooking instead of lighting; also, from the reason
that 1 m3 of biogas is equal to lighting of 60–100 watt bulbs for around 6 h, or cooking 2–3 meals a day
for 5–6 persons. As observed during interviews with the farmers, power generation is favoured when
farmers have an abundance of biogas. In that case, they purchase a combustion engine, which converts
132
Energies 2018, 11, 1794
biogas into the mechanical energy in a heat engine and, consequently, the mechanical energy activates
a generator to produce electrical power.
3.2.3. Digestate
The residue remaining after treatment (anaerobic digestion) in the biogas plant is called digestate.
The use of digestate as a fertilizer is considered beneficial since it provides nutrients (N, P, K) which are
easily accessible to plants. Digestate can be applied directly through the overflow outlet or manually.
Another option is through pre-treatment (e.g., drying) before application. However, this possibility
is used only sporadically. The most common practice is the usage of the digestate directly to the
surrounding household home gardens and use of mainly solid parts of digestate as a crop fertilizer
for rice. Another way of usage (especially of liquid manure and slurry) is partly limited by a long
distance between the biogas plant location and the rice field. In 25% of cases, farmers use digestate
as a fertilizer for vegetable and home-garden, which is a very popular way because of its simplicity
and convenience for the farmers. Study also showed that usage of digestate for fish feeding is still
not adopted in this area, as none of the respondents used digestate for such purpose even though its
benefits were proven in the study by [39] focused on pig-biogas-fish systems.
3.3. Biogas Composition in Various Types of Small-Scale Biogas Plant (KT1 and KT2)
The performance of two models of biogas plants (KT1 and KT2) was observed in order to show the
differences between these two types. It was revealed that the KT2 digester has demonstrated a slightly
higher production of CH4 (66.23%) and, at the same time, lower production of carbon dioxide (28.27%)
(Table 3). Furthermore, slight differences might be seen among other variables, such as content of O2 ,
NH3 , or H2 S (Figure 4). Nitrogen, hydrogen, and water vapour were collectively marked as NHW,
where slight differences were also recognized. However, factors, such as organic input, or maintenance
might be the cause of these discrepancies. Furthermore, Table 4 shows a comprehensive review of
biogas composition from small-scale biogas plants reported by other studies from developing countries.
As is shown, CH4 varies from 50% to 75%, and CO2 from 25% to 50%. Other elements (N2 , CO, O2 ,
and H2 S) are below 2% in general and, respectively, 1% in case of H2 .
Table 3. Biogas composition according to the type of biogas plant, KT1 (n = 83) and KT2 (n = 24).
95% Confidence Interval

Variable Type of Biogas Plant Mean
Lower Bound Upper Bound
KT1 63.79 62.94 64.63
CH4 (vol %)
KT2 66.23 64.70 67.76
KT1 30.97 30.04 31.89
CO2 (vol %)
KT2 28.27 26.59 29.94
KT1 0.05 0.04 0.05
NH3 (vol %)
KT2 0.04 0.03 0.05
KT1 0.10 0.07 0.14
H2 S (vol %)
KT2 0.16 0.09 0.22
KT1 2.14 2.02 2.25
CH:CO2 index
KT2 2.24 2.03 2.44
KT1 21.60 21.31 21.89
Calorific value (MJ/m3 )
KT2 22.36 21.84 22.88
133
Energies 2018, 11, 1794
6%
5%
Content in biogas
4%
3%
2%
1%
0%
ଶ ଶ
KT1 KT2
Figure 4. Performance of two models of biogas plants.
There is also the need to take into consideration the substances that can cause operation difficulties
(dust, oils, and siloxanes) [33]. For the purposes of this study, another biogas quality indicator
represented by methane and carbon dioxide index (CH4 :CO2 ) was set up. This parameter characterises
the relation between the content of CH4 and CO2 as two major substances influencing the final quality
of biogas. From this point of view, a higher index means a higher quality of biogas.
Table 4. Biogas composition from small-scale biogas plants recognized in previous studies.
CH4 CO2 N2 CO O2 H2 H2 S
Feedstock Type of BGP Country Reference
(vol %) (vol %) (vol %) (vol %) (vol %) (vol %) (vol %)
Animal
61–72 - - - - - 0.0043–0.0084 Tubular PVC Costa Rica [40]
wastewater
Livestock
64 34 1.05 0.3 - 0.6 0.05 Floating dome Pakistan [8]
manure
Livestock Laboratory
56.2 39.51 - 1.91 - - 1.84 Nigeria [41]
manure conditions
Livestock Developing
50–75 25–45 <2 - <2 <1 <1 Not specified [42]
manure countries
Livestock
60 35–40 - - - - - Not specified Malaysia [28]
manure
Organic Developing
50–75 25–50 - - - - - Not specified [43]
waste countries
Organic Sub-Saharan
60 - - - - - - Fixed dome [44]
waste Africa
Generalized
50–75 25–50 <2 <2 <2 <1 <2 - - -
values
All surveyed small-scale biogas plants (n = 107) showed in the average content of methane (CH4 )
in biogas of 64.57% (±2.85) and the carbon dioxide (CO2 ) of 30.20% (±3.10). The average presence
of NH3 was 0.05% (±0.02) and the presence of H2 S of 0.25% (±0.12). The average value of CH4 :CO2
index was 2.20 (±0.35). The average calorific value of biogas produced by plants was 21.83 MJ/m3
(±0.96), which corresponds with the typical value of 21–24 MJ/m3 [42].
3.4. Biogas Composition According to Various Ages of Small-Scale Biogas Plants

The results presented in Table 5 is uncovering differences between small-scale biogas plants
younger than five years and older than five years and its effect on various aspects of biogas quality.
However, as shown in Table 5, there are no, or only minor, differences among tested values. This
fact leads us to the conclusion that small-scale biogas plants using pig slurry as a main feedstock are
sustaining a stable level of biogas quality during their life-span. Especially for the main indicators,
which are volume of CH4 in biogas, CH4 :CO2 index and calorific value of biogas.
134
Energies 2018, 11, 1794
Table 5. Biogas composition from small-scale biogas plants younger than five years (n = 82) and older
than five years (n = 25).
Small-Scale Biogas Plants Younger than Five Years Small-Scale Biogas Plants Older than Five Years
Variable 95% Confidence Interval 95% Confidence Interval
Mean Mean
Lower Bound Upper Bound Lower Bound Upper Bound
CH4 (vol %) 65.44 64.58 66.30 64.57 63.05 66.09
CO2 (vol %) 29.31 28.37 32.25 29.93 28.27 31.58
NH3 (vol %) 0.04 0.04 0.05 0.04 0.03 0.05
H2 S (vol %) 0.12 0.08 0.16 0.14 0.07 0.20
CH4 :CO2 index 2.26 2.14 2.38 2.12 1.91 2.32
Calorific value (MJ/m3 ) 22.08 21.79 22.39 21.87 21.36 22.39
3.5. Biogas Composition as Affected by Type, Age, and Capacity of the Biogas Plant
There was an effort to identify the factors that fundamentally influence the biogas quality (Table 6),
therefore, the effects of variables of the type of biogas plant, the age of the biogas plants, and the size of
the biogas plants on biogas composition were analysed. Firstly, the age of the biogas plant was tested
as a relevant factor potentially influencing various aspects of the biogas quality. As demonstrated in
Table 6, all biogas composition factors (including amounts of CH4 , CO2 , NH3 , H2 S, CH4 :CO2 index, and
calorific value) were not significantly affected by the age of biogas plant (p > 0.05). Secondly, the type of
the biogas plant was tested as a relevant factor, using KT1 and KT2 models for comparison. As shown
in Table 6, CH4 , CO2 , and calorific value were recognized as significantly influenced by the type
of biogas plant (p < 0.01). The results show that KT2 model demonstrated a higher percentage of
CH4 and, consequently, a higher calorific value and a lower percentage of CO2 of produced biogas.
Another factor under the examination was the digester capacity (size). The results show different CH4
contents, CH4 :CO2 index values, and calorific values according to the digester capacity (size).
Table 6. Biogas composition as affected by type, age, and capacity of the biogas plant.
Dependent Variable Parameter Coefficient Std. Error t-Value p

Intercept 62.02 1.92 32.32 0.00
Type of digester a −2.44 0.82 −2.99 0.00
CH4 (vol %)
Age b 0.87 0.81 1.08 0.28
Digester capacity (m3 ) 0.52 0.22 2.36 0.02
Intercept 30.00 2.10 14.32 0.00
Type of digester a 2.70 0.89 3.03 0.00
CO2 (vol %)
Age b −0.62 0.88 −0.70 0.49
Digester capacity (m3 ) −0.20 0.24 −0.81 0.42
Intercept 0.04 0.02 2.64 0.01
Type of digester a 0.01 0.01 0.81 0.42
NH3 (vol %)
Age b 0.00 0.01 0.67 0.51
Digester capacity (m3 ) 0.00 0.00 −0.25 0.80
Intercept 0.23 0.08 2.73 0.01
H2 S (vol %)
Age b −0.02 0.04 −0.51 0.62
Digester capacity (m3 ) −0.01 0.01 −0.87 0.39
Intercept 1.56 0.26 5.96 0.00
CH4 :CO2 index
Age b 0.14 0.11 1.29 0.20
Intercept 20.96 0.65 32.14 0.00
Calorific value Type of digester a −0.76 0.28 −2.73 0.01
(MJ/m3 ) Age b 0.21 0.27 0.77 0.45
a The base type is KT1; b the base age is >5 years.
135
Energies 2018, 11, 1794
4. Conclusions
Small-scale biogas plants can be a very useful tool for manure management and may help reduce
global warming impacts if used appropriately. This technology offers a unique set of benefits, as it is a
sustainable source of energy, it is benefiting the environment, and it provides a way to treat and reuse
manure. However, if used inappropriately, its benefits may be compromised. In this study, the most
common feedstock for a small-scale biogas plant was pig slurry, followed by a combination of pig
slurry and human excreta. The majority of biogas plants were connected with the pig stable, or by
latrine and stable. An average daily production of biogas equals to 0.499 m3 , which does not cover
the demand of rural household with six members. Hence, 60% of surveyed households are still using
other sources of energy as well. Biogas composition was measured with a multifunctional portable gas
analyser. The mean content of methane (CH4 ) was 65.44%, and for carbon dioxide (CO2 ) was 29.31%
in the case of biogas plants younger than five years and, respectively, CH4 was 64.57% and CO2 29.93%
for biogas plants older than five years. The only dependent factor influencing the biogas quality was
between biogas plant size and biogas composition, which was proven at CH4 , CH4 :CO2 index, and the
calorific value. Furthermore, type of the biogas plant affected CH4 , CO2 , and the calorific values of
the biogas. Focusing on the influence of age of small-scale biogas plants there are no, or only minor,
differences among tested qualitative biogas parameters. Concluding, that small-scale biogas plants
sustaining a stable level of biogas quality during their life-span.
Author Contributions: Conceptualization, H.R., J.M. and J.B.; Methodology, H.R., J.M. and J.B.; Data collection,
J.B., D.D.V.; Data analysis, H.R., J.B., P.L.D. and D.D.V.; Writing-Original Draft Preparation, H.R.; Writing-Review
& Editing, H.R., J.M., P.L.D., D.D.V. and J.B.; Funding Acquisition, H.R., J.M. and J.B.
Funding: This research was conducted with the financial support of project number 20165003 of Internal Grant
Agency of the Czech University of Life Sciences, Prague. Further support was provided by the Internal Grant
Agency of the Faculty of Tropical AgriSciences, Czech University of Life Sciences, Prague, and project numbers
20185010 and 20185016.
References
1. Ahuja, D.; Tatsutani, M. Sustainable energy for developing countries. Surv. Perspect. Integr. Environ. Soc.
2009, 2, 1–16.
2. Mengistu, M.G.; Simane, B.; Eshete, G.; Workneh, T.S. Factors affecting households’ decisions in biogas
technology adoption, the case of Ofla and Mecha Districts, northern Ethiopia. Renew. Energy 2016, 93, 215–227.
[CrossRef]
3. Rahman, M.M.; Hasan, M.M.; Paatero, J.V.; Lahdelma, R. Hybrid application of biogas and solar resources
to fulfil household energy needs: A potentially viable option in rural areas of developing countries.
4. Thi, N.B.D.; Lin, C.-Y.; Kumar, G. Waste-to-wealth for valorisation of food waste to hydrogen and methane
towards creating a sustainable ideal source of bioenergy. J. Clean. Prod. 2016, 122, 29–41. [CrossRef]
5. Kinyua, M.N.; Rowse, L.E.; Ergas, S.J. Review of small-scale tubular anaerobic digesters treating livestock
waste in the developing world. Renew. Sustain. Energy Rev. 2016, 58, 896–910. [CrossRef]
6. Roubík, H.; Mazancová, J.; Phung, L.D.; Banout, J. Current approach to manure management for small-scale
Southest Asian farmers—Using Vietnamese biogas and non-biogas farms as an example. Renew. Energy 2018,
115, 362–370. [CrossRef]
7. Huong, L.Q.; Madsen, H.; Anh, L.X.; Ngoc, P.T.; Dalsgaard, A. Hygienic aspects of livestock manure
management and biogas systems operated by small-scale pig farmers in Vietnam. Sci. Total Environ. 2014,
8. Mushtaq, K.; Zaida, A.A.; Askari, S.J. Design and performance analysis of floating dome type portable biogas
plant for domestic use in Pakistan. Sustain. Energy Technol. Assess. 2016, 14, 21–25. [CrossRef]
9. Cu, T.T.T.; Cuong, P.H.; Hang, L.T.; Chao, N.V.; Anh, L.X.; Trach, N.X.; Sommer, S.G. Manure management
practices on biogas and non-biogas pig farms in developing countries—Using livestock farms in Vietnam as
an example. J. Clean. Prod. 2012, 27, 64–71. [CrossRef]
136
Energies 2018, 11, 1794
10. Rodolfo, S.; Anh, L.H.; Konrad, K. Feasibility assessment of anaerobic digestion technologies for household
wastes in Vietnam. J. Vietnam. Environ. 2016, 7, 1–8. [CrossRef]
11. Chiumenti, A.; Chiumenti, R.; Chiumenti, A. Complete nitrification-denitrification of swine manure in a
full-scale, non-conventional composting system. Waste Manag. 2015, 46. [CrossRef] [PubMed]
12. Zhang, X.; Luo, L.; Skitmore, M. Household carbon emission research: An analytical review of measurement,
influencing factors and mitigation prospects. J. Clean. Prod. 2015, 103, 873–883. [CrossRef]
13. Bruun, S.; Jensen, L.S.; Vu, V.T.K.; Sommer, S. Small-scale household biogas digesters: An option for global
warming mitigation or a potential climate bomb? Renew. Sustain. Energy Rev. 2014, 33, 736–741. [CrossRef]
14. Neupane, M.; Basnyat, B.; Fischer, R.; Froeschl, G.; Wolbers, M.; Rehfuess, E.A. Sustained use of biogas fuel
and blood pressure among women in rural Nepal. Environ. Res. 2015, 136, 343–351. [CrossRef] [PubMed]
15. Mengistu, M.G.; Simane, B.; Eshete, G.; Workneh, T.S. The environmental benefits of domestic biogas
technology in rural Ethiopia. Biomass Bioenergy 2016, 90, 131–138. [CrossRef]
16. Zhang, L.X.; Wang, C.B.; Song, B. Carbon emission reduction potential of a typical household biogas system
in rural China. J. Clean. Prod. 2013, 47, 415–421. [CrossRef]
17. Truc, N.T.T.; Nam, T.S.; Ngan, N.V.C.; Bentzen, J. Factors Influencing the Adoption of Small-scale Biogas
Digesters in Developing Countries—Empirical Evidence from Vietnam. Int. Bus. Res. 2017, 10. [CrossRef]
18. Rajendran, K.; Aslanzadeh, S.; Taherzadeh, J. Household Biogas Digesters—A Review. Energies 2012,
5, 2911–2942. [CrossRef]
19. Chiumenti, R.; Chiumenti, A.; da Borso, F.; Limina, S. Anaerobic Digestion of Swine Manure in Conventional
and Hybrid Pilot Scale Plants: Performance and Gaseous Emissions Reduction. In Proceedings of the
International Syposium ASABE 2009, Reno, NV, USA, 21–24 June 2009.
20. Bhattacharya, S.C.; Jana, C. Renewable energy in India: Historical developments and prospects. Energy 2009,
34, 981–991. [CrossRef]
21. Roubík, H.; Mazancová, J.; Banout, J.; Verner, V. Addressing problems at small-scale biogas plants: A case
study from central Vietnam. J. Clean. Prod. 2016, 112, 2784–2792. [CrossRef]
22. Thy, S.; Preston, T.R.; Ly, J. Effect of retention time on gas production and fertilizer value of digester
effluent. Livest. Res. Rural Dev. 2003, 15. Available online: http://www.lrrd.org/lrrd15/7/sant157.htm
(accessed on 2 July 2018).
23. Khan, E.U.; Martin, A.R. Optimization of hybrid renewable energy polygeneration system with membrane
distillation for rural households in Bangladesh. FUEL 2015, 93, 1116–1127. [CrossRef]
24. Phanthavongs, S.; Saikia, U. Biogas Digesters in Small Pig Farming Systems in Lao PDR: Evidence of
Impact. Livest. Res. Rural Dev. 2013, 25. Available online: http://www.lrrd.org/lrrd25/12/phan25216.
htm (accessed on 2 July 2018).
25. Katuwal, H.; Bohara, A.K. Biogas: A promising renewable technology and its impact on rural households in
Nepal. Renew. Sustain. Energy Rev. 2009, 13, 2668–2674. [CrossRef]
26. Kaur, G.; Brar, Y.S.; Kothari, D.P. Potential of Livestock Generated Biomass: Untapped Energy Source in
India. Energies 2017, 10, 847. [CrossRef]
27. Usack, J.G.; Wiratni, W.; Angenent, L.T. Improved design of anaerobic digesters for household biogas
production in Indonesia: One cow, one digester, and one hour of cooking per day. Sci. World J. 2014, 1–8.
[CrossRef] [PubMed]
28. Abdeshahian, P.; Lim, J.S.; Ho, W.S.; Hashim, H.; Lee, C.T. Potential of biogas production from farm animal
waste in Malaysia. Renew. Sustain. Energy Rev. 2016, 60, 714–723. [CrossRef]
29. Ghmire, P.C. SNV supported domestic biogas programmes in Asia and Africa. Renew. Energy 2013, 49, 90–94.
[CrossRef]
30. Khan, E.U.; Martin, A.R. Review of biogas digester technology in rural Bangladesh. Renew. Sustain.
Energy Rev. 2016, 62, 247–259. [CrossRef]
31. SNV. Biogas Programme in Vietnam, SNV Vietnam. Available online: http://www.snv.org/ (accessed on
1 February 2018).
32. Pham, C.H.; Vu, C.C.; Sommer, S.G.; Bruun, S. Factors Affecting Process Temperature and Biogas Production
in Small-scale Biogas Digesters in Winter in Northern Vietnam. Asian-Australas. J. Anim. Sci. 2014,
33. Zamorska-Wojdyla, D.; Gaj, K.; Holtra, A.; Sitarska, M. Quality evaluation of biogas and selected methods of
its analysis. Ecol. Chem. Eng. 2012, 19, 77–87. [CrossRef]
137
Energies 2018, 11, 1794
34. Rasi, S. Biogas Composition and Upgrading to Biomethane; University of Jyväskylä: Jyväskylä, Finland, 2009;
ISBN 978-951-39-3618-1.
35. Rasi, S.; Lantela, J.; Rintala, J. Trace compounds affecting biogas energy utilization—A review.
36. Nguyen, V.C.N. Small-scale anaerobic digesters in Vietnam—Development and challenges.
J. Vietnam. Environ. 2011, 1, 12–18. [CrossRef]
37. Vu, Q.D.; Tran, T.M.; Nguyen, P.D.; Vu, C.C.; Vu, V.T.K.; Jensen, L.S. Effect of biogas technology on nutrient
flows for small- and medium-scale pig farms in Vietnam. Nutr. Cycl. Agroecosyst. 2012, 94, 1–13. [CrossRef]
38. Sasse, L.; Kellner, C.; Kimaro, A. Improved Biogas Unit for Developing Countries. Deutsche Gesellschaft
für Technische Zusammenarbeit (GTZ) GmbH; Vieweg & Sohn Verlagsgesellschaft Braunschweig:
Eschborn, Germany, 1991.
39. Nhu, T.T.; Dewulf, J.; Serruys, P.; Huysveld, S.; Nguyen, C.V.; Sorgeloos, P.; Schaubroeck, T. Resource usage
of integrated Pig-Biogas-Fish system: Partitioning and substitution within attributional life cycle assessment.
Resour. Conserv. Recycl. 2015, 102, 27–38. [CrossRef]
40. Lansing, S.; Botero, R.B.; Martin, J.F. Waste treatment and biogas quality in small-scale agricultural digesters.
41. Adegun, I.K.; Yaru, S.S. Cattle Dung Biogas as a Renew. Energy Source for Rural Laboratories.
J. Sustain. Technol. 2013, 4, 1–8.
42. Bond, T.; Templeton, M.R. History and future of domestic biogas plants in the developing world.
Energy Sustain. Dev. 2011, 15, 347–354. [CrossRef]
43. Surendra, K.C.; Takara, D.; Hashimoto, A.G.; Khanal, S.K. Biogas as sustainable energy source for developing
countries: Opportunities and challenges. Renew. Sustain. Energy Rev. 2014, 31, 846–859. [CrossRef]
44. Tumwesige, V.; Fulford, D.; Davidson, G.C. Biogas appliances in Sub-Saharan Africa. Biomass Bioenergy 2014,
70, 40–50. [CrossRef]
138
energies
Article
Design, Construction, and Testing of a
Gasifier-Specific Solid Oxide Fuel Cell System
Alvaro Fernandes 1, *, Joerg Brabandt 2 , Oliver Posdziech 2 , Ali Saadabadi 1 , Mayra Recalde 1 ,
Liyuan Fan 1 , Eva O. Promes 1 , Ming Liu 1 , Theo Woudstra 1
and Purushothaman Vellayan Aravind 1
1 Energy Technology Section, Department of Process and Energy, Delft University of Technology,
Leeghwaterstraat 39, 2628 CB Delft, The Netherlands; [email protected] (A.S.);
[email protected] (M.R.); [email protected] (L.F.); [email protected] (E.O.P.);
[email protected] (M.L.); [email protected] (T.W.); [email protected] (P.V.A.)
2 Sunfire GmbH, Gasanstaltstraße 2, 01237 Dresden, Germany; joerg.brabandt@sunfire.de (J.B.);
oliver.posdziech@sunfire.de (O.P.)
* Correspondence: [email protected]; Tel.: +31-(0)15-278-36-88
Received: 26 May 2018; Accepted: 26 July 2018; Published: 31 July 2018
Abstract: This paper describes the steps involved in the design, construction, and testing of a
gasifier-specific solid oxide fuel cell (SOFC) system. The design choices are based on reported
thermodynamic simulation results for the entire gasifier- gas cleanup-SOFC system. The constructed
SOFC system is tested and the measured parameters are compared with those given by a system
simulation. Furthermore, a detailed exergy analysis is performed to determine the components
responsible for poor efficiency. It is concluded that the SOFC system demonstrates reasonable
agreement with the simulated results. Furthermore, based on the exergy results, the components
causing major irreversible performance losses are identified.
Keywords: SOFC; validation; simulation; exergy; syngas
1. Introduction
The production of electricity, biofuels, and chemicals is increasingly using biomass sources.
Indeed, by 2015, Europe had installed a net maximum capacity of 35.4 GW from energy sources
including municipal waste, biogas, wood and wood residues, and other solid residues. [1].
Biomass is a storable feedstock that is being employed for power generation in biomass-fired
plants. These plants are typically steam cycle or organic Rankine cycle (ORC) power systems capable
of achieving electrical efficiencies from 15% (small plants) to 40% (large plants) [2,3].
Alternatively, biomass can be processed into gaseous fuels such as syngas and biogas for further
use in power-producing steam engines, which have a modest efficiency of approximately 20%; or gas
engines, gas turbines, and fuel cells, especially SOFC, which can achieve efficiencies up to 50% [4–7].
Nevertheless, there are major problems in the use of biomass for power generation, namely
the logistics of collection and seasonal availability, which create inefficient biomass power chains.
To overcome these issues, the installation of small-scale decentralized biomass power plants is
an economically viable and efficient solution [5,8]. At this scale, gas engines and gas turbines
suffer from lower efficiency (i.e., a reduction in power production capacity), compared with SOFCs.
Moreover, SOFCs also have the advantage of operating at very high efficiencies in part-load windows.
Furthermore, they are less susceptible to variations in fuel composition [9,10].
To accommodate the fluctuating electricity demands of both grid and off-grid installations,
SOFC systems should be capable of operating within a wide part-load window. Consequently, it is
fundamental that SOFC systems have an adequate system configuration and components. The selection

Energies 2018, 11, 1985
of components must be based on, among other factors, the material limitations, variations in gas
composition, thermal management, and carbon suppression.
System modelling is a rapid and cost-effective method for predicting system performance and
off-design operation conditions. Many studies on the simulation of SOFC systems [11] focus on the
performance under specific design conditions or the transient performance of SOFCs under a varying
electric load [12], whereas others consider the transient performance of the entire SOFC system. These
system models are usually validated by mathematical models, although some have been validated
using experimental data for a number of the components.
Rokni [13] investigated the re-powering of a steam power plant with gas turbines and SOFCs.
In his work, three system configurations were simulated: base case (steam plant), steam plant with
gas turbines, and steam plant with SOFCs. Only the latter system was calibrated with experimental
data for a planar SOFC, while for other components a similar modeling approach to other available
studies was followed. It was reported that the plant with an SOFC system could achieve an optimized
efficiency above 66% for an operating temperature of 1013 K, current density of 200 A/cm2 , and a fuel
utilization of 80%.
Similarly, Ugartemendia et al. [14] validated his dynamic model of an SOFC-steam cycle with
SOFC experimental data from the literature. Other components such as heat exchangers were assumed
to have constant thermal effectiveness. The authors concluded that the operating temperature of 1173
K and a fuel utilization of 65% were the optimal conditions for achieving the higher power output.
Chung et al. [15] studied the influence of operating parameters on the plant efficiencies of a
methane gas-fed SOFC power generation system. The results obtained from a mathematical model
simulated under the design conditions revealed that the air-to-fuel ratio (A/F) was the most important
parameter in terms of system efficiency. The pre-reforming rate of fuel was found to be relatively
insignificant in terms of efficiency, but could be used as an auxiliary tuner for the operating temperature
of the SOFCs, in addition to A/F.
Chitsaz et al. [16] conducted a thermodynamical evaluation of an integrated tri-generation system
driven by an SOFC. Through steady-state system simulations based on mathematical correlations,
and partially validated by experimental data obtained from an SOFC setup, a maximum system exergy
efficiency of 46% was achieved. The main sources of irreversibility were observed to be the air heat
exchanger, SOFC, and afterburner.
A similar analysis was performed by Stamatis et al. [17] for an SOFC and a hybrid SOFC-gas
turbine system fuelled by ethanol. The models were partially validated by available experimental data
from the literature for the SOFC component. A system efficiency of up to 60% was achieved under
certain operating conditions. It was also disclosed that the SOFC and burner-reformer components
were the major sources of irreversibilities within the systems.
Xu et al. [18] investigated the influence of various design parameters on the SOFC thermal
behaviour and system performance of a natural-gas-fuelled 1 kW conceptual design for a residential
combined heat and power system. This system was also modelled based on mathematical correlations
and partially validated by experimental data obtained from an SOFC experimental setup. The results
indicated that the cell output voltage, system inlet fuel flow rate, and SOFC stack inlet air temperature
had a dramatic effect on the electrical efficiency and cogeneration efficiency.
Somekawa et al. [19] investigated the influence of various design parameters on a manufactured
multi-stack SOFC system coupling an anode regenerator between stacks. The regenerator consisted of
a CO2 absorber and water vapour condenser to selectively remove these compounds from the anode
off-gas mixture. The system models were validated by experimental data collected from an SOFC
setup and a hot module designed especially for the study. A remarkable total fuel utilization of 92.0%
and an electrical efficiency of 77.8% were achieved with this design.
A diesel-fed SOFC power system is being designed and developed for maritime applications
under the SchIBZ-project [20]. In a first stage, a system model was prepared and validated for the
reformer and SOFC components with experimental data available in the literature. The authors will
140
Energies 2018, 11, 1985
further validate the whole system with experimental lab work. The system model will then be used
for system performance analysis and the determination of optimal operation conditions through an
exergetic analysis.
The aforementioned studies all use hydrogen, methane, or natural gas as the fuel. In a recent study,
D’Andrea et al. [21] performed a dynamic simulation for the proof-of-concept of a biogas-fed SOFC
polygeneration system. The SOFC model was primarily validated by reproducing similar test
conditions as used by the manufacturer and comparing with data provided by the manufacturer.
A second validation was performed by including the SOFC model in the system model for comparison
with data collected from the proof-of-concept. This model was developed to investigate both the stack
and balance-of-plant (BoP) thermal behaviour under abnormal operation conditions, namely fast load
current ramps, fault cathode air, and different rates of internal reforming of the fuel in the SOFC.
They concluded that in the event of the two first abnormal conditions, the SOFC may overheat and be
damaged. To prevent this, the system control should shift the system to the open circuit condition.
It was also concluded that the percentage of fuel that is internally reformed in the SOFC can be adjusted
to control the temperature.
To the best of the authors’ knowledge, this paper is the first to describe all of the steps involved
in the design, construction, and testing of a gasifier-specific SOFC system. For the first time, system
testing was conducted to compare the recorded performance with the predicted thermodynamic
performance obtained by system simulation, aiming at validating the exergy flow model of the system.
In addition to the design point analysis, we investigated part-load and off-design conditions to develop
a deeper understanding of the variation in system performance.
2. Selection of the Design of SOFC System

The SOFC system was developed to be integrated with a plasma gasifier and a gas cleanup
unit (Figure 1) at TU Delft. Prior to determining the system configuration, TU Delft performed a
thermodynamic analysis of the power plant in which two suitable configurations for the SOFC system
were evaluated [22]. The main difference in the configurations was the mechanism for suppressing
carbon formation.
Both system configurations considered the SOFC materials, approaches for suppressing carbon
formation, SOFC temperature control strategy, and thermal management.
2.1. SOFC Materials

Nickel–Gadolinium-doped Ceria (Ni/GDC) anode cells were selected for their anticipated
advantages with hydrocarbon fuels [23,24]. It has been reported that Ni/GDC is more resistant
to carbon deposition and poisoning by typical syngas contaminants than Nickel/Yttria-stabilized
Zirconia (Ni/YSZ) anodes [25–27].
2.2. Carbon Suppression Technique

The major difference between the system configurations in the study of Liu et al. [22] is the
technique employed for suppressing carbon formation. One system uses steam produced in a heat
recovery steam generator unit that is mixed with the anode flow before it enters the SOFC module,
whereas the other uses a catalytic partial oxidation unit (CPOx) to suppress carbon formation and
pre-reform some of the hydrocarbons before it enters the SOFC module. In our system, the last
technique is employed.
2.3. SOFC Temperature Control Strategy

In our system, the SOFC temperature is controlled by varying the cathode flow rate. An additional
degree of control is offered by changing the fuel utilization.
141
Energies 2018, 11, 1985
2.4. Thermal Management

The syngas exits the gas cleanup unit at a relatively low temperature (approximately 673 K) and
the cathode air enters the system at environment temperature. Therefore, both flows must be heated
before they enter the SOFC module. This is accomplished by using heat exchangers and a CPOx unit
to suppress carbon formation. Flue gas produced in an afterburner is used as heat source to heat the
cathode air, and is also used to pre-heat/heat the syngas. For the system configuration using steam as
the carbon suppression, the flue gas is also used as heat source to generate steam in a heat recovery
steam generator. The flue gas subsequently moves to the dryer and is used to pre-dry the feedstock
that is fed to the gasifier.
,ĞĂƚ
,ƵŵĂŶǁĂƐƚĞ WůĂƐŵĂ ^ƵƌƉůƵƐŚĞĂƚ

ƌǇĞƌ 'ĂƐĐůĞĂŶƵƉ ^K&ƐǇƐƚĞŵ
;&ĂĞĐĂůŵĂƚƚĞƌͿ ŐĂƐŝĨŝĐĂƚŝŽŶ ĂŶĚĞůĞĐƚƌŝĐŝƚǇ
DŝĐƌŽǁĂǀĞ ůĞĐƚƌŝĐŝƚǇ
ŐĞŶĞƌĂƚŽƌ
Figure 1. Flow of the integrated plasma gasifier-gas cleanup-solid oxide fuel cell (SOFC) system.
Reproduced with permission from [22], Copyright Elsevier, 2018.
Previous results suggest that the system configuration with a CPOx unit could achieve higher
electrical efficiency [22]. As this unit also increases the flexibility in terms of thermal integration, it was
selected for construction.
3. Description of the SOFC System

In this section, the components in the SOFC system and the construction process are briefly
described. The equipment used to record data for the test runs and the safety control strategies are
also listed.
3.1. System Construction Design

The SOFC system (Figure 2) consists of two hot boxes: a BoP hot box and an SOFC hot box [28];
as well as a make-up gases panel. The hot boxes are connected by four insulated pipes (i.e., anode and
cathode inflows/outflows). The BoP hot box is also connected to the supply gases, syngas, and intake
air, and has an exhaust flue gas pipe. Each hot box is thermally insulated with low-thermal-conductivity
panels to reduce heat losses to the surroundings. Moreover, the BoP hot box is filled with a granulated
microporous insulation material that offers good resistance to heat transfer between BoP components.
The BoP hot box contains the air heater, fuel pre-heater, CPOx unit, and afterburner. Both the
fuel pre-heater and air heater are counter-flow plate heat exchangers made of high-temperature alloy.
Under the design conditions, the air heater increases the cathode air temperature to 923 K. The fuel
pre-heater increases the fuel temperature to approximately 523–623 K, after which the fuel is supplied
to the CPOx unit. This component increases the oxygen-to-carbon (O/C) ratio to prevent carbon
formation, pre-reforms hydrocarbons, and increases the fuel temperature to 973–1073 K. Both the
CPOx and afterburner are catalytic burners impregnated with adequate catalysts.
The SOFC hot box contains the SOFC module, consisting of two towers of SOFC stacks that
are electrically connected in parallel. Each tower is composed of three 30-cell stacks that are
electrically connected in series. This module was tested and achieved a nominal power of 4 kW
in reference conditions (40% H2 /60% N2 , fuel utilization (Uf ) = 0.75, cell voltage (V cell ) > 0.65 V,
T cathode outlet < 1133 K). A reduction in the electrical output of approximately 10–15% is expected
142
Energies 2018, 11, 1985
for syngas. The make-up gases panel contains the hydrogen and nitrogen rotameters, which provide
the forming gas during system start-up, and the hydrogen, nitrogen, and air (CPOx air) mass flow
controllers, which provide the gases for the test runs. Other gases such as carbon monoxide and carbon
dioxide are measured by mass flow controllers located in the gas cleanup unit and supplied by the
syngas inlet pipe.
The intake air is fed to the system using two blowers (i.e., a cathode blower and a CPOx blower).
The cathode blower can provide 900 nL of air per minute under reference conditions. A filter and a
flow measurement device are located on the discharge side. The CPOx blower provides 15 nL of air
per minute under reference conditions and is coupled with a filter on the admission side. The mass
flow controller is located on the discharge side of the CPOx blower.
Figure 2. The SOFC 3.5 KWe system at TU Delft with description of the various parts. The air blowers
are located behind the make-up bottle gases panel.
3.2. Safe Operation Parameters

The safe operating parameters are primarily intended to protect expensive components and users
from risks associated with the operation of the SOFC system. This is accomplished by implementing
safety functions in a programmable logic controller (PLC) that are capable of shifting the system to a
safe operation mode when required. The safety functions process information acquired from installed
measurement devices in the system and signals from external safeguards such as the gas detection and
ventilation systems.
The parameters that can limit the extent of the system part-load operation are illustrated in
Figure 3. The safe upper limit operating temperature, maximum flows of cathode air and CPOx air,
and minimum cell operating voltage are the main parameters that influence the part-load operation
window. Safe upper limit operating temperatures are primarily determined by material limitations,
whereas the air flows limit the cooling capacity and the capability for suppressing carbon formation.
143
Energies 2018, 11, 1985
,QIORZV 2XWIORZV %R3+RWER[ 1373 K 62)&KRWER[

1123 K
$IWHUEXUQHU
31

&DWKRGH
1118 K
EORZHU $LU 1103 K 24
+HDWHU 923 K
298 K
,QWDNHDLU
21 22 23 &DWKRGH

900 nl/min
700 nl/min $UHD P
8I 14 1103 K
32 723 K : .:H
873 K 1173 K 9FHOO !9
673 K 713 K 1073 K 9FHOO 9
%LRV\QJDV &32[
11 12 13 $QRGH

110 nl/min )XHO
SUHKHDWHU
1.6 < O/C < 2.0

33 713 K
42
)OXHJDV
&32[
EORZHU

298 K
,QWDNHDLU
41

15 nl/min
13.4 nl/min
Figure 3. Limiting operation and preliminary design parameters of the SOFC system. Red parameters:
maximum volume flows and temperatures; blue parameters: preliminary design parameters. BoP:
balance-of-plant; CPOx: catalytic partial oxidation unit.
4. System Modelling
The system was simulated using Cycle-Tempo R
software (version 5.0, Delft University of
Technology, Delft, The Netherlands), a Fortran-based software package designed for analysing the first
and second laws of thermodynamics of power plants [29].
Figure 4 illustrates the SOFC system model. Three inlet flows are defined: syngas (source 100),
cathode air (source 200), and CPOx air (source 400). There is one outlet flow: flue gas (sink 300).
The operating parameters of the SOFC are inserted to accurately determine the amount of chemical
energy converted into electricity. Specifically, these parameters are the fuel utilization (Uf ), equivalent
resistance (Req ), and electric power (ẆSOFC ). Other parameters such as the CPOx O/C ratio and SOFC
outflows temperature must also be specified.
The CPOx air mole flow (Φ41 ) is calculated as

Φ11 · y11.CO · (O/C − 1) + y11.CO2 · (O/C − 2)
Φ41 = , (1)
2 · y41.O2
where ypipe.species is the mole fraction of the species (CO, CO2 , or O2 ) in the mixture (pipe) and Φ11 is
the inlet flow (mol/s) of the syngas in the system (pipe 11). Assuming a gas composition constant in
the cross-section and that the process occurs at constant temperature and pressure, Φ11 is calculated in
the SOFC component by
I
Φ11 = − Φ41 · y41.N2 , (2)
z · F · Uf · (y13.H2 + y13.CO )
Φ13 · (y13.H2 + y13.CO ) − Φ14 · (y14.H2 + y14.CO )

Uf = , (3)
Φ13 · (y13.H2 + y13.CO )
ẆSOFC
I= , (4)
Vcell
Ix
Vcell = VNernst,x − · Req , (5)
Area x
144
Energies 2018, 11, 1985

go R̄ · Tcell yH2 ,x · y0.5
O2 x
VNernst,x = − + · ln · p0.5
cell , (6)
z·F z·F yH2 O,x
where I and V cell are the current produced by all cells and the cell voltage in amperes and volts,
respectively, z is the number of electrons involved in a single reaction, F is Faraday’s constant (96485
C/mol), VNernst,x is the local Nernst potential (V), and the ratio of Ix and Area x is the local current
density (A/cm2 ). More detailed information can be found in [29,30].
The cathode air mole flow (Φ21 ) is calculated by an energy balance in the SOFC as
Φ24 · h̄24 + Φ14 .h̄14 − Φ13 · h̄13 + ẆSOFC

Φ21 = , (7)
h̄23
where h̄ is the enthalpy in kJ/mol.
The modelling of the heat exchangers is based on the first law of thermodynamics for the energy
balance in the heat exchanger,
Q̇Trans = Φm,h · c p,h · ( Th,in − Th,out ) = Φm,c · c p,c · ( Tc,out − Tc,in ) , (8)
where c p is the specific heat (kJ/(kg· K)) of the medium, Φm is the flow of the fluids (kg/s), h is the hot
medium, and c is the cold medium. The transmitted heat flow in kW (Q̇Trans ) can be calculated as
Q̇Trans = U · A · LMTD, (9)
where U is the overall heat transfer coefficient (W/m2 ·K) and A is the heat transfer surface area (m2 ).
The logarithm mean temperature difference (LMTD) for a counter-flow heat exchanger is
( Th,in − Tc,out ) − ( Th,out − Tc,in )

LMTD = T − Tc,out
. (10)
ln( Th,in )
h,out − Tc,in
The off-design calculations of Cycle-Tempo R

use the approach developed by Miedema [31],
in which the variation in the overall heat transfer coefficient multiplied by the heat transfer surface
area is proportional to the variation of the mass flow:
ηcf
Φm,i
(U.A)i = (U.A)D · , (11)
Φm,D
where i refers to an off-design operating point, D refers to the design operating point of the heat
exchanger, and ηcf is an exponential correction factor. Knowing the inlet temperatures of both media
in the heat exchanger, Cycle-Tempo R
iteratively calculates the outlet temperature of both flows.
The input parameters are listed in Table 1.
145
Energies 2018, 11, 1985
Table 1. Input parameters used in the system modelling.
Item Id Parameters
Syngas (100) 15.6% H2 , 18.3% CO, 6.1% CO2 , 60% N2 (vol. basis)
T11 = 298 K
SOFC (103) Req = 1.1 Ω·cm2 Area = 2300 cm2
T14,24 = 1073 K, Uf = 0.75
Intake air (200; 400) 79% N2 , 21% O2 (vol. basis)
p21,41 =101.3 kPa and T21,41 = 298 K
Air heater (202) Φm,D = 13.41 kg/s; (U.A)D = 55.87 W/K; η = 1.45
Fuel pre-heater (101) Φm,D = 1.4 kg/s; (U.A)D = 3.44 W/K; η = 0.8
Blowers (201; 401) ηis = 0.6; ηmec,el = 0.6
All components - Δp = 5% · pstream
5. Test Runs with the SOFC System

The SOFC system was designed and developed to integrate a gasifier-gas cleaning unit-SOFC plant
fed by faeces from human waste. The faeces, after being separated from urine in the sanitation system,
were sent to a pre-drying unit to reduce the moisture content from approximately 80% to acceptable
levels for the plasma gasification process (20–40%) [22]. The syngas composition used for the design
of the SOFC system was based up on the model developed and presented in [22]. The model was
based on equilibrium assumptions for the gasifier considering that very-high-temperature gasification
processes approach equilibrium conditions.
The SOFC system was tested over the range allowed by the limited testing conditions in the
laboratory. The syngas composition was modified from the initial design composition by increasing the
nitrogen concentration by 20% to enhance the accuracy of the nitrogen flow measurements. The syngas
composition and flow rate were controlled by mixing bottled gases that were individually measured
in mass flow controllers before entering the system. As a consequence, the syngas was fed in at the
environmental temperature on a dry basis, which also diverged from the design conditions. Thus,
some parameters like the flow rate of the CPOx blower or cathode air flow rate did limit the window
of operation points in the performed experiments.
Operating points were taken at constant intervals of 10 mA/cm2 current density. Operating points
were only considered for analysis if the system achieved the steady state without crossing any of the
safe operating parameters.
During the test runs, the following operating parameters that are calculated or displayed in the
interface of the software were maintained:
• Constant SOFC fuel utilization of 0.75 ± 0.01,

• Constant O/C ratio of 1.6 ± 0.002 in the CPOx unit,
• Constant outlet temperature of 1073 ± 10 K at the SOFC cathode.
Table 2 lists the measurement devices installed in the system and their accuracy.
146
Energies 2018, 11, 1985
Table 2. List of measurement devices and their accuracy.
Measurement Type Measurement Device Accuracy
Temperature Thermocouples type N ±1.1 K or 0.4% of Rd ±0.3% terminal (PLC)

Cell voltage In-situ wiring ±0.3% signal analog input (PLC)
CPOx air flow Mass flow controller ±0.5% of Rd ±0.1 of Fs (16 nL/min)
H2 flow Mass flow controller ±0.5% of Rd ±0.1 of Fs (20 nL/min)
N2 flow Mass flow controller ±0.5% of Rd ±0.1 of Fs (20 nL/min)
CO flow Mass flow controller ±0.8% of Rd ±0.2 of Fs (5 nL/min)
CO2 flow Mass flow controller ±1% of Rd ±1% of Fs (1 nL/min)
Cathode air Flow meter ±2% of Rd ±0.3 signal analog input (PLC)
Current Electronic load ±0.2% of Fs (160 A)
Rd: reading value; Fs: full scale/nominal; PLC: programmable logic controller.

The results of this study are presented in four main parts. Firstly, the system model results
for different syngas supply conditions and SOFC temperatures are compared with those under the
design conditions. Secondly, data recorded in the test runs are compared with results obtained by the
system model. Thirdly, the system efficiency is analysed to determine the major sources of inefficiency.
Finally, based on the results, some considerations for improved design strategies are given.
6.1. Thermodynamic Comparison for Different Syngas Supply Conditions and SOFC Temperatures
A thermodynamic model was prepared considering both the design conditions and the
experimental conditions. The results are depicted in Figure 4 and Table 3.
.
62)&KRWER[ .
$IWHUEXUQHU

62)&
9FHOO 9
M $P . ,QIORZV
$ & .
9FHOO 9 2XWIORZV
M $P . . . QOPLQ
. . . QOPLQ
.
.

+
$LUKHDWHU
&DWKRGHEORZHU ,QWDNHDLU
.
.
. . QOPLQ
. . QOPLQ
2&
2&
+
)XHO
&32[ %LRV\QJDV
SUHKHDWHU
.
. .
.

)OXHJDV
&32[
. QOPLQ
EORZHU . QOPLQ

%R3KRWER[ ,QWDNHDLU
Figure 4. SOFC system model. Blue text refers to system simulation with design conditions, green text
refers to experimental conditions for 3.5 kW production by the SOFC module.
147
Energies 2018, 11, 1985
Table 3. Comparison of the results of the main streams as well as system model efficiencies.
Simulation Considering Design Conditions
Pipe number
11 13 14 23 24 31 41
mol % H2 26.95 21.70 5.57

mol % CO 22.00 18.55 4.54
mol % CO2 7.01 8.01 22.05 3.90
mol % N2 37.85 43.09 43.04 79.0 81.04 76.06 79.0
mol % O2 21.0 18.96 15.57 21.0
mol % H2 O 6.19 8.60 24.81 4.46
Mole flow (mol/s) 0.082 0.089 0.089 0.537 0.523 0.608 0.009
Vol. flow (nL/min) 110.0 119.9 119.9 721.8 703.3 817.2 12.7
Exergy (kW) 10,382 10,118 3984 4773 6958 10,287 ~0
Auxiliaries consumption = 0.092 kW System electric efficiency = 32.8%
Simulation Considering Experimental Conditions
Pipe number
11 13 14 23 24 31 41
mol % H2 15.60 8.71 2.23

mol % CO 18.36 13.21 3.25
mol % CO2 6.10 7.84 17.80 4.38
mol % N2 60.0 65.51 65.50 79.0 80.70 77.99 79.0
mol % O2 21.0 19.30 14.82 21.0
mol % H2 O 4.74 11.22 2.80
Mole flow (mol/s) 0.146 0.169 0.169 0.661 0.647 0.811 0.029
Vol. flow (nL/min) 195.8 227.0 227.0 888.0 869.2 1090 39.5
Exergy (kW) 12,545 11,946 5166 5381 8129 12,606 ~0
Auxiliaries consumption = 0.116 kW System electric efficiency = 27.0%
The supply of a syngas composition with no water vapor and higher nitrogen content
(experimental conditions) resulted in an increase of all other flows. Higher syngas flow was fed
to compensate the loss of chemical energy in the CPOx unit by oxidation with air to achieve the specific
O/C ratio. For the effect, a higher amount of CPOx air was also supplied (Figure 4). As a major
consequence, a higher amount of heat was produced that resulted in higher outlet temperature of
pipe 13 of the CPOx unit, which was also higher than the SOFC temperature. Therefore, the cooling
requirements of the SOFCs were enhanced and, subsequently, a higher cathode air flow was needed.
The higher cathode flow resulted in a lower operating temperature in the afterburner (53 K) and
lower temperature of the cathode inflow (pipe 23) of approximately 35 K.
The syngas fed at 298 K was heated in the fuel pre-heater to 100 K less than with the preliminary
design conditions, even though higher heat flow was transmitted by the flue gas. As a consequence,
the temperature of the flue gas leaving the system (pipe 33) was slightly reduced.
Finally, the system performance was substantially lower for the experimental conditions of 5%.
A substantial amount of chemical energy converted in the CPOx and, consequently, higher syngas
flow was required to produce equal electric power. Nonetheless, no significant deviations were found
in terms of SOFC performance and temperatures and, therefore, the model showed to be reliable for
comparison and analysis of the experimental work.
148
Energies 2018, 11, 1985
6.2. Comparison of System Simulation with Test Runs

In the test runs, only three steady-state operating points were achieved, at 70, 80 and 90 mA/cm2 ,
resulting in a part-load operating window of 37–47%. For lower current densities than 70 mA/cm2 ,
a steady state could not be reached because of the safety protection for the fuel pre-heater (T33 should
not exceed 573 K), whereas for higher current densities than 90 mA/cm2 , no sufficient CPOx air could
be provided to achieve the specified O/C ratio.
Table 4 compares the data measured in the test run at a current density of 90 mA/cm2 with the
values calculated in the system simulations. Good agreement between the results can be observed.
Figure 5 compares the SOFC performance (average cell voltage) and cathode air flow required to
maintain an outlet cathode temperature of 1073 K for experimental and simulation results. A good
match in the average cell voltage was seen, and thus equal power was produced. The cell voltage
was also in accordance with values reported in many studies in syngas-fed-SOFCs [32,33]. Moreover,
as equal fuel flow rate and composition were supplied, it can be concluded that the heat production in
the subsystem was similar between test runs and the simulation model. The temperatures of the main
streams in the BoP box are shown in Figures 6–8. There was reasonable agreement between the values
measured in the real system during the test runs and those provided by the system simulations. In
general, the maximum deviations were on the order of 10 K. The exception was the temperature of the
syngas after the CPOx (T13 ), where the deviation was approximately 20 K for a current density of 90
mA/cm2 .
These deviations were caused by various factors, such as simplifications and constraints imposed
in the system simulation (non-heat-losses approach, equilibrium calculations for the CPOx component,
and inaccuracies in the heat exchanger design), and the accuracy of measurement devices in the system,
among others.
Table 4. Comparison of results for the 90 mA/cm2 operating point.
Apparatus/Pipe Parameter Units Simulation Test Runs
11 Syngas flow nL/min 75.6 75.7 ± 0.54

21 Cathode air flow nL/min 631.8 633.2 ± 18
41 CPOx air flow nL/min 15.3 15.3 ± 0.09
103 Cell voltage, V cell V 0.789 0.791 ± 0.0023

103 SOFC power, ẆSOFC kW 1.632 1.631 ± 0.013
13 Temperature after CPOx, T13 K 1049.9 1065.4 ± 5.5

23 Cathode inlet temp., T23 K 996.7 996.5 ± 5.0
24 Temperature cathode outlet, T24 K 1073 1073 ± 5.6
31 Temp. after afterburner, T31 K 1126.1 1127.8 ± 6.0
33 Temp. exhaust flue gas, T33 K 516.9 518.9 ± 1.8
6.3. System Efficiency

The system performance (electric efficiency) was evaluated at the three operating points obtained
in the test runs. The thermodynamic calculations are similar to those described in [34]. Figure 9 shows
the distribution (percentage) of the total input exergy (syngas flow) across the various types of exergy
output by the system, namely electricity, exergy destruction, and exergy loss.
149
Energies 2018, 11, 1985

9ROXPHIORZ>Q/PLQ@

6LPXODWLRQ
7HVWUXQV

&XUUHQWGHQVLW\>P$FP@
(a)

$YHUDJHFHOOYROWDJH>9@
7HVWUXQV
6LPXODWLRQ

(b)
Figure 5. Comparison of parameters acquired in the test runs with simulated ones. (a) Cathode volume
flow; (b) Average cell voltage (Vcell ).
3DUWORDG>@

7HPSHUDWXUH>R&@

7HPSHUDWXUH>.@

7HVWUXQV>7@ 7HVWUXQV>7@
6LPXODWLRQ>7@ 6LPXODWLRQ>7@
Figure 6. Comparison of simulated results with those acquired in the test runs. Syngas from the CPOx
(T13 ) Flue gas from afterburner (T31 ) .
150
Energies 2018, 11, 1985
3DUWORDG>@

7HPSHUDWXUH>R&@
7HPSHUDWXUH>.@

7HVWUXQV>7@
6LPXODWLRQ>7@

Figure 7. Comparison of simulated results with those acquired in the test runs. Cathode air from air
heater (T23 ) .
3DUWORDG>@

7HVWUXQV>7@
7HPSHUDWXUH>R&@
7HPSHUDWXUH>.@
6LPXODWLRQ>7@

Figure 8. Comparison of simulated results with acquired ones in the test runs. Temperature of flue gas
leaving the system (T33 ) .

)OXHJDV
RIWRWDOLQSXWH[HUJ\V\QJDVIORZ
62)&
$IWHUEXUQHU

&32[

$LUKHDWHU

)XHOSUHKHDWHU
(OHFWULFSRZHU

Figure 9. Exergy distribution into the various types of exergy produced in the system.
* Net electric power; ** Exergy destruction per component; *** Exergy loss carried by the flue gas.
151
Energies 2018, 11, 1985
The air heater contributed to a 25–30% reduction in the efficiency of the system, making it the
major source of irreversibility. The impact of this type of component on the system performance has
been highlighted in many studies (e.g., [35,36]), and is caused by heat transfer at limited temperature
as well as great temperature difference between hot and cold media. Flue gas losses appeared as the
second contributor, carrying approximately 20% of the exergy. The high physical exergy carried by
the stream due to high exhaust temperature and mass flow rate were the aspects contribution for
such value. The afterburner and CPOx also caused a significant reduction in system performance of
approximately 14% as a consequence of the irreversible nature of the oxidation process. The latter
component also contributed to the great exergy contained in flue gas. Although it is not clearly
demonstrated in Figure 9, a significant amount of the chemical exergy of the syngas was converted
into heat exergy in the CPOx. The main consequence was that, for equal electric power production, a
larger amount of syngas needed to be fed into the system, which subsequently reduced the system
efficiency. Finally, the SOFC and fuel pre-heater were minor sources of irreversibility. Nevertheless,
they collectively corresponded to approximately 6% in exergy destruction. This resulted in a modest
system electrical efficiency of 33.7–34.5%.
Table 5. System performance of various stationary SOFC systems. Modified after [37].
System Manufacturer Output Power (kW) η (%) Fuel Processing Ref.
Bloom energy 250 AC >53 - [38]

Wärtsilä 24 DC 47 SR/AOGR [39]
FZJ 20 DC 41 External SR [37]
VTT(2010) 7 AC 43 SR/AOGR [40]
VTT(2011) 8 AC 49 SR/AOGR [41]
CFCL 1.5 AC 60 SR/WR [42]
ENE-farm 0.7 AC 41–47 SR/WR [43,44]
Hexis 1 AC 35 CPOx [45]

SOFCpower 1 AC 32 CPOx [46]
IKTS 1.26 AC 39 CPOx [47]
SR: steam reforming; AOGR: anode off-gas recycling; WR: water recycling.
The system performance was comparable with other systems employing a CPOx unit, as can be
seen in Table 5. It can also be observed that all SOFC systems employing steam reforming as the fuel
processing had higher system efficiency. This aspect is associated with the steam reforming process,
which increases the chemical energy of the fuel by using available heat exergy in the system.
Nevertheless, it is should also be highlighted that the system performance is affected by various
aspects such as the fuel composition, selected load operation as well as the system design choices. In
the next section, some considerations for improving the design are given.
7. Future Considerations for Improving the Design

As mentioned previously, high-temperature flue gas is exhausted as a consequence of the use of
a CPOx unit for suppressing carbon formation. The replacement of this unit with a steam reformer
would potentially enhance the electrical efficiency of the system. The produced steam would be mixed
with the supply syngas. To enable this, a heat recovery steam generator should be installed and the heat
exchangers should be rearranged. In this new arrangement, the flue gas from the afterburner would
first be used as a heat source in the fuel pre-heater in order to increase the anode flow temperature to
the specified value at inlet of the SOFC anode, then to the air heater, and finally to the heat recovery
steam generator.
152
Energies 2018, 11, 1985
8. Summary
In this study, we investigated the design, construction, and testing of a gasifier-specific SOFC
system. This paper makes the following contributions to the study of advanced SOFC technologies:
1. System development: this is the first study to describe the design, development, and testing
of an SOFC system to be integrated with a gasifier. The gasifier considered in this study is a
plasma reactor with the capacity to process 8.84 kW of human waste (before pre-drying) [22,48].
The SOFC system was designed based on discussions between TU Delft and Sunfire GmbH.
2. Calculated results exhibited good agreement with experimentally recorded values under different
operating conditions. This clearly demonstrates the advantage of a rigorous thermodynamic
model of new fuel cell power systems when they are being designed and built.
3. The validated model clearly indicates where the thermodynamic losses are occurring and provides
indications on how to minimize these losses in such a system, resulting in improved designs in
the future.
4. System efficiencies of 33.7–34.5% were estimated. The CPOx unit and heat exchangers, especially
the air heater, were identified as the major contributors to reductions in efficiency.
Author Contributions: The experimental work was carried out by A.S., M.R., L.F. and E.O.P. The SOFC simulation
was supported by M.L. O.P. and J.B. jointly with M.L., E.O.P. and P.V.A. decided the conceptual design SOFC
system. J.B. and O.P. led and supervised the construction of the SOFC system. T.W., P.V.A. and J.B. contributed in
the results discussion. A.F. led the experimental work, developed the SOFC models, and prepared the first draft
and final version of the paper. The paper was corrected and reviewed by J.B. and P.V.A.
Funding: This research was founded by The Bill & Melinda Gates foundation to build the SOFC system.
Acknowledgments: The Bill & Melinda Gates Fundation is thanked for the financial support. Also, the Portuguese
Fundação para a Ciência e Tecnologia (FCT), through the Grant—SFRH/BD/77042/2011, and the Ecuador
Secretaría de Educación Superior, Ciencia, Tecnología e Innovación (SENESCY) are thanked for the partial
financial support to the first authors.
References
1. Eurostat. The Key to European Statistics. 2017. Available online: http://ec.europa.eu/eurostat/web/
energy/data/database (accessed on 22 April 2017).
2. Guercio, A.; Bini, R. 15—Biomass-fired Organic Rankine Cycle combined heat and power systems. In Organic
Rankine Cycle (ORC) Power Systems; Macchi, E., Astolfi, M., Eds.; Woodhead Publishing: Sawston, UK, 2017;
pp. 527–567. [CrossRef]
3. Hurskainen, M.; Vainikka, P. 7—Technology options for large-scale solid-fuel combustion. In Fuel Flexible
Energy Generation; Oakey, J., Ed.; Woodhead Publishing: Boston, MA, USA, 2016; pp. 177–199. [CrossRef]
4. Chacartegui, R.; Sanchez, D.; de Escalona, J.M.; Monje, B.; Sanchez, T. On the effects of running existing
combined cycle power plants on syngas fuel. Fuel Process. Technol. 2012, 103, 97–109. [CrossRef]
5. Asadullah, M. Barriers of commercial power generation using biomass gasification gas: A review.
6. Santhanam, S.; Schilt, C.; Turker, B.; Woudstra, T.; Aravind, P. Thermodynamic modeling and evaluation of
high efficiency heat pipe integrated biomass Gasifier-Solid Oxide Fuel Cells-Gas Turbine systems. Energy
2016, 109, 751–764. [CrossRef]
7. Bellomare, F.; Rokni, M. Integration of a municipal solid waste gasification plant with solid oxide fuel cell
and gas turbine. Renew. Energy 2013, 55, 490–500. [CrossRef]
8. Singh, J. Management of the agricultural biomass on decentralized basis for producing sustainable power in
India. J. Clean. Prod. 2017, 142, 3985–4000. [CrossRef]
9. Bae, C.; Kim, J. Alternative fuels for internal combustion engines. Proc. Combust. Inst. 2017, 36, 3389–3413.
[CrossRef]
10. Hagos, F.Y.; Aziz, A.R.A.; Sulaiman, S.A. Study of syngas combustion parameters effect on internal
combustion engine. Asian J. Sci. Res. 2013, 6, 187–196. [CrossRef]
153
Energies 2018, 11, 1985
11. Barelli, L.; Bidini, G.; Ottaviano, A. Part load operation of SOFC/GT hybrid systems: Stationary analysis.
Int. J. Hydrogen Energy 2012, 37, 16140–16150. [CrossRef]
12. Andersson, D.; Aberg, E.; Yuan, J.; Sunden, B.; Eborn, J. Dynamic Modeling of a Solid Oxide Fuel Cell System
in Modelica. In Proceedings of the ASME 2010 8th International Conference on Fuel Cell Science, Engineering
and Technology, Brooklyn, NY, USA, 14–16 June 2010; American Society of Mechanical Engineers: New York,
NY, USA, 2010; pp. 65–72.
13. Rokni, M. Performance comparison on repowering of a steam power plant with gas turbines and solid oxide
fuel cells. Energies 2016, 9, 399. [CrossRef]
14. Ugartemendia, J.; Ostolaza, J.X.; Zubia, I. Operating point optimization of a hydrogen fueled hybrid solid
oxide fuel cell-steam turbine (SOFC-ST) plant. Energies 2013, 6, 5046–5068. [CrossRef]
15. Chung, T.D.; Hong, W.T.; Chyou, Y.P.; Yu, D.D.; Lin, K.F.; Lee, C.H. Efficiency analyses of solid oxide fuel
cell power plant systems. Appl. Therm. Eng. 2008, 28, 933–941. [CrossRef]
16. Chitsaz, A.; Mahmoudi, S.M.S.; Rosen, M.A. Greenhouse gas emission and exergy analyses of an integrated
trigeneration system driven by a solid oxide fuel cell. Appl. Therm. Eng. 2015, 86, 81–90. [CrossRef]
17. Stamatis, A.; Vinni, C.; Bakalis, D.; Tzorbatzoglou, F.; Tsiakaras, P. Exergy analysis of an intermediate
temperature solid oxide fuel cell-gas turbine hybrid system fed with ethanol. Energies 2012, 5, 4268–4287.
[CrossRef]
18. Xu, H.; Dang, Z.; Bai, B.F. Analysis of a 1 kW residential combined heating and power system based on solid
oxide fuel cell. Appl. Therm. Eng. 2013, 50, 1101–1110. [CrossRef]
19. Somekawa, T.; Nakamura, K.; Kushi, T.; Kume, T.; Fujita, K.; Yakabe, H. Examination of a high-efficiency
solid oxide fuel cell system that reuses exhaust gas. Appl. Therm. Eng. 2017, 114, 1387–1392. [CrossRef]
20. Huerta, G.V.; Álvarez Jordán, J.; Dragon, M.; Leites, K.; Kabelac, S. Exergy analysis of the diesel pre-reforming
solid oxide fuel cell system with anode off-gas recycling in the SchIBZ project. Part I: Modeling and validation.
Int. J. Hydrogen Energy 2018. [CrossRef]
21. D’Andrea, G.; Gandiglio, M.; Lanzini, A.; Santarelli, M. Dynamic model with experimental validation of a
biogas-fed SOFC plant. Energy Convers. Manag. 2017, 135, 21–34. [CrossRef]
22. Liu, M.; Woudstra, T.; Promes, E.; Restrepo, S.; Aravind, P. System development and self-sustainability
analysis for upgrading human waste to power. Energy 2014, 68, 377–384. [CrossRef]
23. Din, Z.U.; Zainal, Z. The fate of SOFC anodes under biomass producer gas contaminants. Renew. Sustain.
Energy Rev. 2016. [CrossRef]
24. Cassidy, M.; Connor, P.; Irvine, J.; Savaniu, C. 5—Anodes. In High-Temperature Solid Oxide Fuel Cells for the
21st Century (Second Edition), 2nd ed.; Kendall, K., Kendall, M., Eds.; Academic Press: Boston, MA, USA,
2016; pp. 133–160. [CrossRef]
25. Kuramoto, K.; Hosokai, S.; Matsuoka, K.; Ishiyama, T.; Kishimoto, H.; Yamaji, K. Degradation behaviors of
SOFC due to chemical interaction between Ni-YSZ anode and trace gaseous impurities in coal syngas. Fuel
Process. Technol. 2017, 160, 8–18. [CrossRef]
26. Liu, M.; van der Kleij, A.; Verkooijen, A.; Aravind, P. An experimental study of the interaction between tar
and SOFCs with Ni/GDC anodes. Appl. Energy 2013, 108, 149–157. [CrossRef]
27. Zhang, L.; Jiang, S.P.; He, H.Q.; Chen, X.; Ma, J.; Song, X.C. A comparative study of H2 S poisoning on
electrode behavior of Ni/YSZ and Ni/GDC anodes of solid oxide fuel cells. Int. J. Hydrogen Energy 2010,
35, 12359–12368. [CrossRef]
28. GmbH, S. Power Core—Efficient Electricity Generator. 2017. Available online: http://www.sunfire.de/en/
products-technology/power-core (accessed on 6 October 2017).
29. Delft, T. Cycle-Tempo 5.0. 2015. Available online: http://www.asimptote.nl/software/cycle-tempo/
(accessed on 22 September 2017).
30. De Groot, A. Advanced Energy Analysis of High Temperature Fuel Cell Systems. Ph.D. Thesis,
Delft University of Technology, Delft, The Netherlands, 2004.
31. Miedema, J.A. CYCLE: A general computer code for thermodynamic cycle computations. Studies of
cogeneration in district heating systems. In NASA STI/Recon Technical Report N; NASA: Washington, DC,
USA, 1981; Volume 82.
32. Suwanwarangkul, R.; Croiset, E.; Entchev, E.; Charojrochkul, S.; Pritzker, M.; Fowler, M.; Douglas, P.;
Chewathanakup, S.; Mahaudom, H. Experimental and modeling study of solid oxide fuel cell operating
with syngas fuel. J. Power Sources 2006, 161, 308–322. [CrossRef]
154
Energies 2018, 11, 1985
33. Baldinelli, A.; Cinti, G.; Desideri, U.; Fantozzi, F. Biomass integrated gasifier-fuel cells: Experimental
investigation on wood syngas tars impact on NiYSZ-anode Solid Oxide Fuel Cells. Energy Convers. Manag.
2016, 128, 361–370. [CrossRef]
34. Fernandes, A.; Woudstra, T.; van Wijk, A.; Verhoef, L.; Aravind, P. Fuel cell electric vehicle as a power plant
and SOFC as a natural gas reformer: An exergy analysis of different system designs. Appl. Energy 2016,
173, 13–28. [CrossRef]
35. Lee, Y.D.; Ahn, K.Y.; Morosuk, T.; Tsatsaronis, G. Exergetic and exergoeconomic evaluation of an
SOFC-Engine hybrid power generation system. Energy 2018, 145, 810–822. [CrossRef]
36. Hosseinpour, J.; Chitsaz, A.; Eisavi, B.; Yari, M. Investigation on performance of an integrated
SOFC-Goswami system using wood gasification. Energy 2018, 148, 614–628. [CrossRef]
37. Peters, R.; Blum, L.; Deja, R.; Hoven, I.; Tiedemann, W.; Küpper, S.; Stolten, D. Operation experience with a
20 kW SOFC system. Fuel Cells 2014, 14, 489–499. [CrossRef]
38. Bloom Energy Corporation. Bloom Energy Server ES5—250 kW. Available online: https://www.
bloomenergy.com/ (accessed on 8 July 2018).
39. Noponen, M.; Hottinen, T.; Sandstrom, C. WFC20 Biogas Unit Operation. In Proceedings of the 9th European
Solid Oxide Fuel Cell Forum, Lucerne, Switzerland, 29 June–2 July 2010; pp. 2–90.
40. Halinen, M.; Rautanen, M.; Saarinen, J.; Pennanen, J.; Pohjoranta, A.; Kiviaho, J.; Pastula, M.; Nuttall, B.;
Rankin, C.; Borglum, B. Performance of a 10 kW SOFC demonstration unit. ECS Trans. 2011, 35, 113–120.
41. Halinen, M.; Pennanen, J.; Himanen, O.; Kiviaho, J.; Silvennoinen, P.; Backman, J.; Salminen, P.;
Pohjoisranta, A. Sofc Power 2007–2012 Project—The Finnish SOFC System Research Flagship. J. Fuel
Cell Technol. 2013, 13, 32–39.
42. Payne, R.; Love, J.; Kah, M. Generating electricity at 60% electrical efficiency from 1–2 kWe SOFC products.
ECS Trans. 2009, 25, 231–239.
43. Suzuki, M.; Takuwa, Y.; Inoue, S.; Higaki, K. Durability verification of residential SOFC CHP system.
ECS Trans. 2013, 57, 309–314. [CrossRef]
44. Yoshida, H.; Seyama, T.; Sobue, T.; Yamashita, S. Development of residential SOFC CHP system with flatten
tubular segmented-in-series cells stack. ECS Trans. 2011, 35, 97–103.
45. Mai, A.; Iwanschitz, J.; Schuler, A.; Denzler, R.; Nerlich, V.; Schuler, A. Hexis and the SOFC System
Galileo 1000 N—Past, Present, Future. In Proceedings of the 11th European SOFC SOE Forum A, Lucerne,
Switzerland, 1–4 July 2014; pp. 3–11.
46. Bucheli, O.; Bertoldi, M.; Modena, S.; Ravagni, A. Development and manufacturing of SOFC-based products
at SOFCpower SpA. ECS Trans. 2013, 57, 81–88. [CrossRef]
47. Pfeifer, T.; Chakradeo, A.; Ahire, N.; Baade, J.; Barthel, M.; Dosch, C.; Näke, R.; Hartmann, M. Development
of a SOFC Power Generator for the Indian Market. Fuel Cells 2017, 17, 550–561. [CrossRef]
48. Sturm, G.S.J.; Muñoz, A.N.; Aravind, P.V.; Stefanidis, G.D. Microwave-Driven Plasma Gasification for
Biomass Waste Treatment at Miniature Scale. IEEE Trans. Plasma Sci. 2016, 44, 670–678. [CrossRef]
c 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
155
energies
Article
Prospecting for Oleaginous and Robust Chlorella spp.
for Coal-Fired Flue-Gas-Mediated
Biodiesel Production
Bohwa Kim 1,† , Ramasamy Praveenkumar 2,† , Eunji Choi 3 , Kyubock Lee 4 , Sang Goo Jeon 3
and You-Kwan Oh 5, *
1 Advanced Biomass R&D Center, Korea Advanced Institute of Science and Technology (KAIST),
Daejeon 34141, Korea; [email protected]
2 Aquatic Biology Laboratory, and Renewable Materials and Nanotechnology Group, KU Leuven,
Campus Kulak, 8500 Kortrijk, Belgium; [email protected]
3 Biomass and Waste Energy Laboratory, Korea Institute of Energy Research (KIER), Daejeon 34129, Korea;
[email protected] (E.C.); [email protected] (S.G.J.)
4 Graduate School of Energy Science and Technology, Chungnam National University (CNU), Daejeon 34134,
Korea; [email protected]
5 School of Chemical and Biomolecular Engineering, Pusan National University (PNU), Busan 46241, Korea
* Correspondence: [email protected]; Tel.: +82-51-510-2395; Fax: +82-51-512-8563
† These authors contributed equally to this work.
Received: 2 July 2018; Accepted: 1 August 2018; Published: 3 August 2018
Abstract: Prospecting for robust and high-productivity strains is a strategically important step
in the microalgal biodiesel process. In this study, 30 local strains of Chlorella were evaluated in
photobioreactors for biodiesel production using coal-fired flue-gas. Three strains (M082, M134,
and KR-1) were sequentially selected based on cell growth, lipid content, and fatty acid composition
under autotrophic and mixotrophic conditions. Under autotrophic conditions, M082 and M134
showed comparable lipid contents (ca. 230 mg FAME [fatty acid methyl esters derived from microalgal
lipids]/g cell) and productivities (ca. 40 mg FAME/L·d) versus a reference strain (KR-1) outdoors
with actual flue-gas (CO2 , 13%). Interestingly, under mixotrophic conditions, M082 demonstrated,
along with maximal lipid content (397 mg FAME/g cell), good tolerance to high temperature (40 ◦ C).
Furthermore, the fatty acid methyl esters met important international standards under all of the
tested culture conditions. Thus, it was concluded that M082 can be a feedstock of choice for coal-fired,
flue-gas-mediated biodiesel production.
Keywords: Chlorella; coal-fired flue-gas; screening; biodiesel property; mixotrophic cultivation
1. Introduction
Microalgae have attracted much global attention for their potential as biodiesel feedstock [1,2].
This interest reflects not only microalgae’s higher photosynthetic efficiency and lipid yield compared
with conventional oil crops but also their ability to mitigate atmospheric CO2 and to grow in non-arable
land and a variety of wastewaters [3–7]. Moreover, microalgae are easily isolated: they are ubiquitously
present in nature and represent a vast diversity of (>50,000) species existing over a wide range of
environment conditions [1,8].
Despite its several advantages, microalgae-based biodiesel production, when considered on a
commercial scale, remains challenging. First, for competitive and sustainable biodiesel production,
the cost of the culture system needs to be significantly reduced, and lipid productivity needs to
be substantially improved [1]. Industrial exhaust flue-gases are an inexpensive and rich source of
CO2 (~14% v/v), the utilization of which for microalgal biomass production would be a fiscally and

Energies 2018, 11, 2026
environmentally better option than compressed CO2 [9,10]. However, high concentrations of CO2
and the presence of inhibitory compounds such as NOx, Sox, and CO in flue-gases can significantly
suppress the metabolic activities of microalgae [6,10]. Hence, for the successful application of flue-gases
to microalgal biomass production, the selection of a flue-gas-tolerant and high-lipid-yield strain
is warranted.
On the other hand, several microalgae, if not all, can grow mixotrophically by utilizing organic
carbons along with light, thereby yielding high lipid productivity. Liang et al. reported that a
mixotrophic cultivation of Chlorella vulgaris with 1–2% glucose yielded a 10-to-20-fold increase
in the biomass (254 mg/L·d) and lipid (54 mg/L·d) productivities relative to an autotrophic
cultivation [11]. Exploiting such a mixotrophic cultivation system with a suitable microalga would
open up opportunities to utilize waste organics such as carbon sources and, at the same time, help to
bring down cultivation costs.
Amongst the several oleaginous microalgae, Chlorella is one of the most widely studied genus
for its potential to grow in outdoors at high cell densities and accumulate high levels of intracellular
triacylglycerol (TAG), a desirable biodiesel feedstock [11–15]. However, the literature indicates that the
growth rates and lipid contents of Chlorella are both species-specific and possibly significantly variant
according to culture conditions. Recently, Sun et al. reported the differences in biomass productivity
(0.41–0.58 g/L·d) and lipid content (36–49 wt. %) among 9 Chlorella spp. cultured with artificial CO2
gas under autotrophic conditions [15].
What also must be noted is the fact that many highly productive microalgae selected under
laboratory conditions fail to withstand the actual, fluctuating environmental conditions and
contamination outdoors [7,16,17]. While accounting for such differential behavior among the
Chlorella spp., it is imperative to select a native isolate that is appropriate not only in having a
high lipid content but also in its capacity to adapt to environmental fluctuations under a suitable
cultivation system.
The aim of the present work was to screen various native Chlorella spp. to identify a robust,
fast-growing, and high-lipid-accumulating strain for high-quality biodiesel production under
autotrophic/mixotrophic systems mediated by coal-fired flue-gas (CO2 , 13%). The employed screening
strategies included comparisons of growth rate, lipid content, high-temperature tolerance, and fatty
acids quality with respect to key biodiesel properties. The robustness of each of the selected
high-productivity stains was finally evaluated under outdoor culture conditions mediated by
coal-fired flue-gas.
2.1. Preliminary Screening Based on Growth Rate

In a preliminary screening, 30 local strains of Chlorella were evaluated for their growth rates in
bubble-column photobioreactors (b-PBRs) with 10% CO2 (v/v) under autotrophic and mixotrophic
conditions. The mixotrophic cultures were supplemented with 5 g glucose/L. After 50 h incubation,
and the growths (OD660 ) of the strains were compared (Figure 1). Most of the strains showed relatively
higher growths under the mixotrophic than under the autotrophic conditions. This phenomenon
is consistent with other studies that have reported organic carbon’s induction of growth promotion
in Chlorella [11,18,19]. The 30 stains, as based on their growth characteristics, were divided into the
following three groups.
157
Energies 2018, 11, 2026
Figure 1. Growth of Chlorella spp. (30 strains) under autotrophic and mixotrophic conditions.
The optical density was measured at 660 nm at the end of 50 h cultivation in b-PBRs. The mixotrophic
cultures were supplemented with 5 g glucose/L.
Group 1: Twelve strains of the tested Chlorella showed a very low growth rate under
both autotrophic and mixotrophic conditions. This delayed growth was expected, based on their
demonstrated intolerance to the high concentrations of CO2 (10% v/v in air) applied in this study.
Many microalgae reportedly tolerate CO2 concentrations up to 5% (v/v in air), whereas concentrations
above this can be harmful, not to mention inhibitive of microalgal growth [9]. Hence, these strains
might not be suitable for flue-gas-mediated cultivation, in which CO2 concentrations often exceed 10%.
Group 2: Interestingly, this group of 3 strains (N043, N052, and M138) exhibited decreased growth
under mixotrophic, compared with autotrophic, conditions. Lower growth rate under mixotrophy
could possibly be due to the inhibitory effects of glucose on the microalgal cells. Such effects are
species-dependent; high concentrations of glucose have been reported to substantially inhibit Chlorella
and Nannochloropsis growths [11,18]. The three group 2 strains, seemingly sensitive to organic carbon
supplementation, were not further studied, considering their poor future prospects for utilization in
lipid-productivity-improvement using waste organics.
Group 3: For these 15 strains of Chlorella, the mixotrophic cultures resulted in higher growth
rates than the autotrophic condition. That is to say, these 15 strains utilized glucose efficiently to
achieve impressive growth. Among them, six (M082, M134, M189, M194, M751, and KR-1) exhibited
the highest growth under mixotrophic conditions and also appeared to tolerate a high concentration
of CO2 (10% v/v in air). Hence, these six strains, thus identified as fast-growing, were held over for
further investigation.
2.2. Secondary Screening Based on Lipid Content and Composition

Growth rate and lipid content are the most important factors in assessing prospective microalga for
biodiesel production [1,20]. Additionally, the composition of the fatty acids produced by the microalga
determines the quality of the biodiesel produced [21,22]. Hence, the six selected fast-growing strains
(M082, M134, M189, M194, M751, and KR-1) were allowed to grow for a longer period of time (until
96 h), and then they were further evaluated for their lipid contents (in this study, expressed as FAME,
i.e., amount of fatty acid methyl esters derived from microalgal lipids) and their compositions under
autotrophic and mixotrophic conditions (Figure 2). After 96 h of cultivation, the microalgal cells would
have already experienced nitrogen depletion [6]. Indeed, despite their fast-growing nature, strains
M082 and M134 accumulated high levels of intracellular lipids: 283 and 298 mg FAME/g DCW (dry
cell weight), respectively, under autotrophic conditions. Moreover, under mixotrophic conditions,
these strains accumulated 397 and 310 mg FAME/g DCW, respectively, which is slightly less than the
corresponding content for control strain KR-1 (411.3 mg FAME/g DCW). Notwithstanding, owing to
158
Energies 2018, 11, 2026
their higher biomass productions, the small differences in the lipid contents of M082 and M134 could
be very well compensated by their higher biomass productivities (Figure S1). As for Chlorella sp. KR-1,
it has been well studied in the past for its fast-growth and high-lipid-accumulation potential under
mixotrophic conditions [6,23]. Thus, M082 and M134 strains, based on their indoor performances,
could be suggested to be potentially comparable with KR-1.
Figure 2. Lipid contents of six fast-growing Chlorella strains under autotrophic and mixotrophic
conditions. The lipid contents of the cultures were estimated by FAME analysis after 96 h cultivation.
The mixotrophic cultures were supplemented with 5 g glucose/L.
Further, the FAME compositions of the biodiesel produced from lipids by the six fast-growing
strains also were examined (Table 1). The data were similar under both autotrophic and mixotrophic
conditions: the C16 and C18 species, the common fatty acids in vegetable oils, were the main
constituents [15,22,24]. Also, at different distribution (%) levels, major fatty acid methyl esters such as
palmitate (C16:0), stearate (C18:0), oleate (C18:1n9C), linoleate (C18:2n6C), and linolenate (C18:3n3)
were observed (Figure S2). Among these, palmitate, oleate, and linoleate accounted for more than 70%
of the total fatty acids identified. The balance between these fatty acids determines the key qualities of
biodiesel, such as oxidative stability, cold flow point, lubricity, viscosity, cetane number, iodine number,
heat of combustion, and NOx emission [15,22,24]. The details are discussed in a later section below
(see Section 2.4.—estimation of biodiesel properties from fatty acids profiles). In consideration of the
high lipid contents and FAME compositions of the 3 Chlorella strains M082, M134, and KR-1, these
were chosen for further investigation in an outdoor culture mediated by coal-fired flue-gas.
2.3. Outdoor Flue-Gas Cultivation and High-Temperature Tolerance Experimentation

The 3 high-productivity strains (M082, M134, and KR-1) were cultivated outdoors in 1 L b-PBRs
for 160 h with a supply of actual flue-gas collected from a 2 MW demonstration-scale coal-burning
power plant [6]. The changes in the light intensity and temperature were continuously monitored;
the maximum sunlight was 1000 μmol/m2 ·s with an average daylight of 500 μmol/m2 ·s; meanwhile,
the day temperature ranged between 30 and 38 ◦ C and came down to 19 ◦ C during the night (Figure 3a).
M082 and M134 were tolerant to flue-gas and grew well in the outdoor conditions: M134, like the
reference strain KR-1, readily adapted to the conditions, starting to grow in quick time. Irrespective of
the sharp lag phase, M082, very similarly to the other two strains, grew rapidly to reach the final OD660
(Figure 3b). At the end of 160 h of cultivation, the lipid contents of M082 and M134 were estimated
to be 228 and 235 mg FAME/g cell, for FAME productivities of 39 and 42 mg/L·d, respectively
(Figure 3c). These values were comparable with the lipid productivity of the reference strain KR-1
(41 mg FAME/L·d). Under the optimized culture conditions, the FAME productivity of KR-1 was
shown to have increased as high as 155 mg/L·d [6], which highlights the likelihood of improving the
159
Energies 2018, 11, 2026
productivities of the selected strains M082 and M134. Thus, despite growing outdoors, with a supply
of actual flue-gas and under fluctuating environmental conditions, the productivities of these two
strains did not deteriorate; moreover, the FAME compositions remained less altered when compared
with the indoor cultivations (Figure S3).
(a)
(b)
(c)
Figure 3. (a) Time-course changes in light intensity and temperature during cultivation; (b) growth;
and (c) lipid content of three high-productivity Chlorella strains in outdoor b-PBR culture supplied with
coal-fired flue-gas.
160
Table 1. FAME compositions of lipid extracted from six fast-growing Chlorella strains under autotrophic and mixotrophic conditions.
Ratio (%) of Each FAME to Total FAME (Mean ± Standard Deviation)

Fatty Acid Methyl Ester M082 M134 M189 M194 M751 KR-1
Auto Mixo Auto Mixo Auto Mixo Auto Mixo Auto Mixo Auto Mixo
Caprate C10:0 0.05 ± 0.07 0.08 ± 0.00 nd 0.1 ± 0.01 nd nd nd nd nd nd nd 0.19 ± 0.01
Tridecanoate C13:0 0.26 ± 0.01 nd 0.26 ± 0.01 nd 0.34 ± 0.03 0.18 ± 0.00 0.37 ± 0.01 0.14 ± 0.01 0.49 ± 0.06 0.32 ± 0.00 0.23 ± 0.03 0.11 ± 0.01
Energies 2018, 11, 2026
Myristate C14:0 nd 0.94 ± 0.01 nd 0.57 ± 0.10 0.61 ± 0.03 0.59 ± 0.00 0.64 ± 0.03 0.48 ± 0.00 0.36 ± 0.02 0.51 ± 0.51 0.23 ± 0.00 0.22 ± 0.00
Pentadecanoate C15:0 0.12 ± 0.00 0.19 ± 0.01 0.1 ± 0.00 0.22 ± 0.00 nd 0.15 ± 0.00 0.21 ± 0.01 0.14 ± 0.00 nd 0.16 ± 0.00 nd 0.13 ± 0.00
Palmitate C16:0 23 ± 0.08 26.78 ± 0.06 25.99 ± 0.05 26.79 ± 0.06 22.36 ± 0.99 24.55 ± 0.37 22.21 ± 1.34 24.6 ± 0.08 20.96 ± 0.16 23.91 ± 0.52 24.49 ± 0.30 28.55 ± 0.02
Palmitoleate C16:1 0.49 ± 0.01 0.29 ± 0.00 0.55 ± 0.01 0.37 ± 0.00 0.35 ± 0.01 0.4 ± 0.01 0.31 ± 0.00 0.36 ± 0.00 0.14 ± 0.20 0.3 ± 0.02 0.36 ± 0.01 0.32 ± 0.00
Stearate C18:0 3.72 ± 0.01 6.26 ± 0.01 4.32 ± 0.01 5.71 ± 0.02 0.78 ± 0.00 1.01 ± 0.00 0.72 ± 0.03 1.52 ± 0.35 0.5 ± 0.04 0.97 ± 0.00 4.57 ± 0.02 4 ± 0.02
Oleate C18:1n9c 21.45 ± 0.02 21.8 ± 0.05 20.76 ± 0.02 19.25 ± 0.03 11.54 ± 0.03 16.06 ± 0.10 10.07 ± 0.22 22.02 ± 0.03 7.28 ± 0.01 15.26 ± 0.05 19.07 ± 0.10 18.7 ± 0.02
Linoleate C18:2n6c 24.64 ± 0.04 25.42 ± 0.17 23.66 ± 0.07 27.47 ± 0.14 20.61 ± 0.55 21.43 ± 0.18 20.2 ± 0.69 21.6 ± 0.02 18.3 ± 0.21 19.43 ± 0.25 25.6 ± 0.30 24.08 ± 0.03
γ-Linoleate C18:3n6 0.34 ± 0.00 0.31 ± 0.00 0.32 ± 0.00 0.35 ± 0.01 0.4 ± 0.03 0.35 ± 0.03 0.43 ± 0.01 0.33 ± 0.01 0.4 ± 0.02 0.38 ± 0.00 0.35 ± 0.01 0.29 ± 0.01
Linolenate C18:3n3 6.69 ± 0.00 4.22 ± 0.06 6.22 ± 0.01 4.43 ± 0.01 12.87 ± 0.38 12.12 ± 0.16 13.09 ± 0.56 9.89 ± 0.01 14.52 ± 0.19 12.18 ± 0.21 6.5 ± 0.09 7.49 ± 0.02
Arachidate C20:0 0.14 ± 0.19 0.35 ± 0.00 0.25 ± 0.01 0.15 ± 0.01 nd nd nd nd nd nd 0.32 ± 0.03 0.26 ± 0.01
cis-11-iocosenoate C20:1 0.14 ± 0.20 0.05 ± 0.07 nd nd nd nd nd nd nd nd nd nd
Behenate C22:0 0.12 ± 0.01 0.19 ± 0.00 nd 0.05 ± 0.07 nd nd nd nd nd nd nd 0.1 ± 0.00
Lignocerate C24:0 nd 0.15 ± 0.00 nd 0.15 ± 0.05 nd nd nd nd nd nd nd nd
Saturated 27.41 ± 0.19 34.95 ± 0.07 30.91 ± 0.06 33.75 ± 0.16 24.1 ± 1.05 26.49 ± 0.36 24.14 ± 1.40 26.62 ± 0.05 22.31 ± 0.09 25.87 ± 0.53 29.85 ± 0.32 33.55 ± 0.04
Monounsaturated 22.08 ± 0.21 22.14 ± 0.02 21.31 ± 0.03 19.63 ± 0.03 11.89 ± 0.03 16.47 ± 0.11 10.38 ± 0.22 22.38 ± 0.04 7.42 ± 0.22 15.55 ± 0.07 19.43 ± 0.09 19.02 ± 0.02
Polyunsaturated 31.66 ± 0.05 29.95 ± 0.22 30.2 ± 0.07 32.25 ± 0.14 33.89 ± 0.90 33.9 ± 0.32 33.73 ± 1.26 31.82 ± 0.01 33.22 ± 0.39 31.99 ± 0.46 32.45 ± 0.40 31.85 ± 0.03
Others 18.85 ± 0.04 12.96 ± 0.13 17.58 ± 0.02 14.37 ± 0.01 30.12 ± 0.19 23.15 ± 0.15 31.75 ± 0.36 19.18 ± 0.09 37.05 ± 0.51 26.59 ± 0.13 18.26 ± 0.01 15.57 ± 0.09
Total 100 100 100 100 100 100 100 100 100 100 100 100
161
Auto: autotrophic condition; Mixo: mixotrophic condition; nd: not detected.
Energies 2018, 11, 2026
The three strains were further tested for their tolerances to high temperatures (35–40 ◦ C) under
indoor culture condition. Selection of a high-temperature-tolerant strain carries an added advantage for
practical application of microalgal culture to higher-temperature flue-gas from emission sites, especially
in summer [23]. Although Chlorella sp. KR-1 has been extensively studied due to its high biomass/lipid
productivity and flue-gas tolerance [2,5,6,23], its lowered metabolic activities, especially at 40 ◦ C,
should be noticed. Interestingly, Chlorella sp. M082 demonstrated better tolerance to temperatures up
to 40 ◦ C than those of M134 and the reference strain KR-1 (Figure S4). The results indicated, thereby,
that Chlorella sp. M082, because it is endowed with not only a high temperature/flue-gas tolerance but
also high lipid productivity, might be a highly robust strain.
2.4. Theoretical Estimation of Biodiesel Properties from Fatty Acid Profiles

Based on fatty acid profiles (Table 1) and literature-based correlations [21,24], the biodiesel
properties (especially viscosity, specific gravity, cloud point, cetane number, iodine number,
and HHV-high heating value) of the six Chlorella strains (M082, M134, M189, M194, M751, and KR-1)
grown under indoor and three strains under outdoor cultures were estimated (Table 2). The values
of these key properties were then compared with the Europe (EN14214) and USA (ASTM D6751)
standards and a variety of vegetable oil values.
Kinematic viscosity (mm/s2 ) is one of the important properties of a biodiesel: the value must be
high enough to provide sufficient lubrication for the engine parts, but also low enough not to affect the
flow properties at the operational temperature. The viscosities of the Chlorella strains under the indoor
and outdoor culture conditions were determined to be approximately 4.6 and 4.5 mm/s2 , respectively,
which meet the specifications established by the international standards. These values, it should also be
emphasized, are very similar to the viscosities of the vegetable oils. Sun et al. (2015), similarly, reported
that the kinematic viscosities of 9 Chlorella strains all measured around 4.4 mm/s2 [15]. Indeed, in the
present study, no significant viscosity value changes were observed between the indoor and outdoor
culture conditions, which suggests that these conditions have a negligible effect on such changes,
especially in saturated fatty acids. The specific gravity of a biodiesel is the ratio of the density of the
FAME to that of water at 15 ◦ C. A denser biodiesel has a higher energy content and thus will deliver
better engine performance. Irrespective of the culture conditions, the biodiesels from all of the tested
strains tested in the present investigation measured a specific gravity of 0.88, which is similar to the
values of the vegetable oils and petroleum diesel, and which, in any cases, falls within the limits of
international standards.
The cloud point (◦ C) is defined as the temperature at which crystals appear in biodiesel as it
becomes cooler. The cloud point increases with increased saturated fatty acids. The lower the cloud
point, the better the fuel quality. Insignificant cloud point differences were observed among the strains
under the different culture conditions; however, there were no common trends. In general, the cloud
point values were within the 6.7–8.4 ◦ C range indoors and the 5.4–6.5 ◦ C range outdoors, whose
values are slightly higher than those of the Jatropa oil that is commonly used as a biodiesel in many
countries. Cloud point values ranging from 3.6 to 5.5 ◦ C have been reported for different Chlorella
strains tested under autotrophic conditions [15]. Nevertheless, the addition of anti-gel additives is
widely recommended in order to lower the cloud point of biodiesel sufficiently to prevent filters from
clogging under lower temperatures [24]. The cetane number (CN) represents the combustion quality
of a biodiesel during compressor ignition. In a particular diesel engine, higher-CN fuels will have
shorter ignition delay periods compared with lower-CN ones. The CN values of the tested strains were
within a 56.26–57.10 range indoors and a 55.58–56.16 range in the flue-gas-mediated outdoor cultures,
whose values were well above the standard specifications. Moreover, the microalgal biodiesels had
higher CN values compared with Jatropa and most of the other vegetable oils.
The iodine value (IV) indicates the total unsaturation degree regardless of the relative proportions
of mono to polyunsaturated compounds. A high IV value for a fuel represents a decreased oxidation
stability, which causes the formation of various degradation products that can negatively affect
162
Energies 2018, 11, 2026
engine operability. The IV values of the tested Chlorella strains ranged from 77.05 to 94.50 under the
different culture conditions and were under the maximum limit of 120 established in the European
biodiesel standard (no limits have been specified for IV in the U.S. standard). The higher heating
value (HHV, MJ/kg) shows the energy content, otherwise known as the calorific value, of biodiesel.
Hence, a higher HHV generally is recommended for diesel automobiles. The HHV of a given biodiesel
increases with increasing fatty acid chain length and decreases with unsaturation [25]. The HHVs
of the present tested strains were unaffected by the culture conditions and were around 40 MJ/kg
(Table 2). Similar values of HHV (40.5 MJ/kg) had been reported among 9 Chlorella strains under
indoor cultivation conditions [15]. Moreover, these values were almost the same as those of other
vegetable oils. There are no specifications in the international standards. It should be noted that overall,
irrespective of the culture conditions, the biodiesel properties of all of the tested Chlorella strains were
within the ranges specified in the international standards.
163
Table 2. Biodiesel properties of Chlorella strains under autotrophic and mixotrophic conditions indoors and outdoors compared with literature and standards.
Chlorella Viscosity [mm/s2 ] *Specific Gravity Cloud Point [◦ C] Cetane Number Iodine Value HHV [MJ/kg]
Strain Auto Mixo Auto Mixo Auto Mixo Auto Mixo Auto Mixo Auto Mixo
KR-1 4.63 ± 0.00 4.63 ± 0.00 0.88 ± 0.00 0.88 ± 0.00 7.81 ± 0.07 7.91 ± 0.01 56.79 ± 0.04 56.84 ± 0.00 80.53 ± 0.42 80.02 ± 0.05 40.14 ± 0.01 40.13 ± 0.00
M082 4.62 ± 0.00 4.66 ± 0.00 0.88 ± 0.00 0.88 ± 0.00 7.67 ± 0.01 8.44 ± 0.05 56.72 ± 0.00 57.10 ± 0.03 81.35 ± 0.05 77.05 ± 0.29 40.16 ± 0.00 40.05 ± 0.01
M134 4.65 ± 0.00 4.64 ± 0.01 0.88 ± 0.00 0.88 ± 0.00 8.21 ± 0.01 8.12 ± 0.03 56.99 ± 0.00 56.94 ± 0.02 78.34 ± 0.06 78.83 ± 0.16 40.09 ± 0.00 40.09 ± 0.00
Indoor culture
M189 4.62 ± 0.01 4.59 ± 0.00 0.88 ± 0.00 0.88 ± 0.00 7.58 ± 0.20 7.07 ± 0.06 56.68 ± 0.10 56.42 ± 0.04 81.83 ± 1.11 84.65 ± 0.35 40.17 ± 0.03 40.23 ± 0.01
Energies 2018, 11, 2026
M194 4.63 ± 0.01 4.60 ± 0.00 0.88 ± 0.00 0.88 ± 0.00 7.79 ± 0.27 7.14 ± 0.00 56.78 ± 0.14 56.46 ± 0.00 80.66 ± 1.52 84.28 ± 0.02 40.14 ± 0.04 40.23 ± 0.00
M751 4.64 ± 0.01 4.62 ± 0.00 0.88 ± 0.00 0.88 ± 0.00 8.14 ± 0.11 7.69 ± 0.11 56.95 ± 0.05 56.74 ± 0.05 78.74 ± 0.61 81.20 ± 0.56 40.09 ± 0.02 40.15 ± 0.02
KR-1 4.52 ± 0.00 na 0.88 ± 0.00 Na 5.37 ± 0.04 na 55.58 ± 0.02 na 94.12 ± 0.23 na 40.46 ± 0.01 na
¶ Outdoor culture M082 4.55 ± 0.00 na 0.88 ± 0.00 Na 6.02 ± 0.00 na 55.90 ± 0.00 na 90.50 ± 0.01 na 40.37 ± 0.00 na
M134 4.57 ± 0.00 na 0.88 ± 0.00 Na 6.54 ± 0.02 na 56.16 ± 0.01 na 87.64 ± 0.12 na 40.31 ± 0.00 na
† Jatropa 4.48 0.88 4.67 55.23 98.02 40.55
‡ Palm 4.61 0.87 14.00 61.90 54.00 40.60
Vegetable oil ‡ Rapeseed 4.50 0.88 -3.00 53.70 116.10 41.10
‡ Soy bean 4.26 0.88 0.00 51.30 125.50 39.70
‡ Petroleum diesel 2-3 0.85 Country specific 40-45 na na
International EN14214 3.5–5.0 0.86–0.9 na Minimum 51.0 Maximum 120 na
standard ASTM
1.9–6.0 0.85–0.9 na Minimum 47.0 na na
D6751
* Specific gravity = (1/1000) density at 15 ◦ C; ¶ outdoor autotrophic are mediated by coal-fired flue-gas; † values from literature [21]; ‡ values from literature [24]; na: not available.
164
Energies 2018, 11, 2026
3.1. Microalgal Strains and Culture Medium

Twenty-nine strains of Chlorella were evaluated for biodiesel production in comparison with a
reference strain Chlorella sp. KR-1 (Table 3). These strains were obtained from the Korean Collection
for Type Cultures (KCTC; Daejeon, Korea). Meanwhile, Chlorella sp. KR-1, which has been extensively
studied for its high lipid productivity and flue-gas tolerance, was maintained at the Korea Institute of
Energy Research (KIER; Daejeon, Korea) [2,5,6,23,26,27]. The modified N8 medium used in this study
was prepared and filter-sterilized through a 0.2 μm membrane, and its pH was adjusted to 6.5 [16].
Table 3. List of tested Chlorella strains and their origins.
No. Strain Species Origin

1 N002 Chlorella vulgaris Paddy field/Ssukgol, Gyeonggi
2 N006 C. vulgaris Paddy field/Suwon, Gyeonggi
3 N032 C. vulgaris Reservoir/Andongho, Gyeongbuk
4 N034 C. vulgaris Reservoir/Andongho, Gyeongbuk
5 N043 C. vulgaris Reservoir/Imhaho, Gyeongbuk
6 N052 C. vulgaris Reservoir/Youngsanho, Chungnam
7 N053 Chlorella sp. Reservoir/ Andong, Gyeongbuk
8 N113 Chlorella sp. Youngpyung, Gangwon
9 N309 Chlorella sp. Local fresh water body
10 M077 Chlorella sp. Local fresh water body
25 M696 C. vulgaris Reservoir/Yongsoho Jeju
26 M740 C. sorokiniana Reservoir/Daedongho, Hampyung
27 M742 C. vulgaris Reservoir/Daedongho, Hampyung
28 M746 C. vulgaris Reservoir/Daedongho, Hampyung
29 M751 C. vulgaris Local fresh water body
30 KR-1 Chlorella sp. Yeongwol, Gangwon
3.2. Seed Flask Culture

All of the strains were initially cultured in 250 mL Erlenmeyer flasks (working vol., 100 mL).
The flasks were incubated under continuous illumination (white fluorescent lamps, ca. 40 μmol
photons/m2 ·s) in a shaking incubator (IS-971RF, Lab Companion, Korea) at 150 rpm and 25 ◦ C.
3.3. Indoor Culture Using Photobioreactor with 10% (v/v) CO2

Pyrex glass bubble-column photobioreactor (b-PBRs; working vol., 500 mL) [23] was utilized
to culture Chlorella strains. The biomass from the flask cultures was used as the b-PBRs inoculum,
a concentration of which was fixed to an initial optical density (OD) of 0.1 at 660 nm. The strains
were grown under either autotrophic or mixotrophic conditions for 96 h. For the latter, the modified
N8 medium was supplemented with filter-sterilized glucose (5 g/L, Sigma, USA; 0.2 μm Minisart
165
Energies 2018, 11, 2026
High-Flow filter, 16532-K, Sartorius Stedium Biotech., Germany). The b-PBR was continuously
supplied with 10% CO2 in air (v/v) from the bottom of the reactor. The supplied gas was conveyed
by a 0.2 μm PTFE syringe filter (Minisart SRP15, Satorius Stedium Biotech., Germany) and controlled
by mass-flow controllers (MKP, Korea) and flow meters (RM Rate-Master, Dwyer instruments Inc.,
Michigan, IN, USA). The reactor was maintained under continuous illumination (white fluorescent
lamps, ca. 170 μmol/m2 ·s) in a temperature-controlled room (28–31 ◦ C). Preliminary screening of
30 Chlorella spp. was done using single PBR, whereas the rest of the experiments were carried out with
duplicate PBRs.
A high-temperature-tolerance experiment was carried out for the fast-growing three strains under
autotrophic conditions including continuous illumination (white fluorescent lamps, ca. 69 μmol/m2 s)
in a temperature-controlled growth chamber (GC-300, Jeio Tech, Korea). The b-PBR was incubated at
35 ◦ C for 45 h, and then the temperature was increased to 40 ◦ C, where it was maintained for a further
25 h.
3.4. Outdoor Culture Using Photobioreactor with Actual Flue-gas

The three strains selected under indoor experimentation were grown outdoors in b-PBRs under
autotrophic conditions for 160 h. The biomass from the flask cultures was used as an inoculum with
a fixed initial OD of 0.3 at 660 nm. The cultures were mediated by a supply of coal-fired flue-gas
(obtained from a 2 MW demonstration-scale coal-burning power plant located at KIER, Daejeon, Korea)
at a flow rate of 0.6 vvm [23]. The typical composition of the flue-gas supplied was CO2 , 13.3%; O2 ,
7.6%; CO, 39.1ppm; NOx, 6.9 ppm [6]. Details on the storage, pre-processing, and transfer of flue-gas
are available elsewhere [6]. Throughout the experiment, the cultures were maintained at ambient
temperature, and the average day time light intensity in this period was 500 μmol/m2 ·s.
3.5. Analytical Methods

Growth of the microalgal strains was evaluated based on their OD values measured at 660 nm
using a UV-Vis spectrophotometer (Optizen 2120UV, Mecasys Co., Daejeon, Korea). The dry-cell
weight (DCW) was measured by GF/C filtration, washing with distilled water, and 105 ◦ C drying.
The pH and light intensity were determined using a pH meter (DKK-TOA Co., Japan) and a
quantum meter (LI-250A, LI-COR Inc., Lincoln, NE, USA), respectively. For outdoor cultures, data
on parameters such as temperature and light intensity were continuously recorded. The temperature
probe (CIMON-SCADA V 2.10, KDT Systems, Seongnam, Korea) was immersed inside the b-PBR in
the culture medium. Changes in the light intensity were monitored through a light sensor (LI-250A,
LI-COR Inc., USA) that was fixed on the top of the b-PBR-setting panel, and the data was collected by
a data logger (LI-1400, LI-COR Inc., USA). The compositions (CO, NO, NO2 , SO2 , CO2 , and O2 ) of the
flue gas supplied for the cultures were analyzed using a portable flue-gas analyzer (Vario Plus, MRU,
Neckarsulm, Germany). The FAME (fatty acids methyl ester) contents of the dried cells were prepared
by following a modified in-situ transesterification method [18] and analyzed by gas chromatography
(GC; Agilent 6890, Agilent Technologies, Wilmington, DE, USA). The relevant detailed methodologies
have previously been reported [3,23]. All the analyses were done in triplicates, and the results were
represented as mean ± standard deviation.
3.6. Estimation of Biodiesel Properties

Important biodiesel properties including viscosity, specific gravity, cloud point, cetane number,
iodine number, and higher heating value (HHV) were estimated from the fatty acid compositions
based on the empirical formulae proposed by Tanimura et al. (2014) [21].
4. Conclusions
In this study, 30 local strains of Chlorella were evaluated for application to flue-gas-mediated
biodiesel production. Among the 30 strains, 3 strains (M082, M134, and KR-1) could be selected on the
166
Energies 2018, 11, 2026
basis of their growth rates, lipid contents, and fatty acid profiles under autotrophic (10% CO2 ) and
mixotrophic (10% CO2 plus 5 g glucose/L) nutrition modes. These strains were further tested under
outdoor conditions using the actual coal-fired flue-gas from demonstration-scale plant. M082 showed
good lipid production performances (228 mg FAME/g cell and 39 mg FAME/L·d), as well as a high,
up-to-40 ◦ C temperature tolerance. Thus, Chlorella sp. M082 can be considered a potential strain for
simultaneous utilization of CO2 from flue-gas and of waste organics for biodiesel production.
Supplementary Materials: The following are available online at http://www.mdpi.com/1996-1073/11/8/2026/s1,

Figure S1: Growth curve of 6 fast-growing Chlorella strains under autotrophic and mixotrophic conditions for
96 h of cultivations, Figure S2: FAME distributions of 6 fast-growing Chlorella strains under autotrophic and
mixotrophic indoor conditions for 96 h of cultivation, Figure S3: FAME distribution of 3 high-productivity Chlorella
strains in outdoor culture mediated by coal-fired flue-gas, and Figure S4: Growth patterns of 3 high-productivity
Chlorella strains under high-temperature conditions.
Author Contributions: Y.-K.O. and B.K. conceived and designed the experiments; B.K. and E.C. performed the
experiments; Y.-K.O. and R.P. analyzed the data; Y.-K.O., K.L. and S.G.J. contributed reagents/materials/analysis
tools; R.P., B.K. and Y.-K.O. wrote the paper.
Funding: This work was supported by a 2-Year Research Grant of Pusan National University.
References
1. Mata, T.M.; Martins, A.A.; Caetano, N.S. Microalgae for biodiesel production and other applications:
A review. Renew. Sustain. Energ. Rev. 2010, 14, 217–232. [CrossRef]
2. Praveenkumar, R.; Kim, B.; Lee, J.; Vijayan, D.; Lee, K.; Nam, B.; Jeon, S.G.; Kim, D.M.; Oh, Y.K. Mild pressure
induces rapid accumulation of neutral lipid (triacylglycerol) in Chlorella spp. Bioresour. Technol. 2016, 220,
3. Cho, S.; Luong, T.T.; Lee, D.; Oh, Y.K.; Lee, T. Reuse of effluent water from a municipal wastewater treatment
plant in microalgae cultivation for biofuel production. Bioresour. Technol. 2011, 102, 8639–8645. [CrossRef]
[PubMed]
4. Lee, K.; Lee, S.Y.; Na, J.G.; Jeon, S.G.; Praveenkumar, R.; Kim, D.M.; Chang, W.S.; Oh, Y.K. Magnetophoretic
harvesting of oleaginous Chlorella sp. by using biocompatible chitosan/magnetic nanoparticle composites.
5. Lee, K.; Lee, S.Y.; Praveenkumar, R.; Kim, B.; Seo, J.Y.; Jeon, S.G.; Na, J.G.; Park, J.Y.; Kim, D.M.; Oh, Y.K.
Repeated use of stable magnetic flocculant for efficient harvest of oleaginous Chlorella sp. Bioresour. Technol.
2014, 167, 284–290. [CrossRef] [PubMed]
6. Praveenkumar, R.; Kim, B.; Choi, E.; Lee, K.; Park, J.Y.; Lee, J.S.; Lee, Y.C.; Oh, Y.K. Improved biomass
and lipid production in a mixotrophic culture of Chlorella sp. KR-1 with addition of coal-fired flue-gas.
7. Praveenkumar, R.; Shameera, K.; Mahalakshmi, G.; Akbarsha, M.A.; Thajuddin, N. Influence of nutrient
deprivations on lipid accumulation in a dominant indigenous microalga Chlorella sp., BUM11008: Evaluation
for biodiesel production. Biomass Bioenergy 2012, 37, 60–66. [CrossRef]
8. Remmers, I.M.; Wijffels, R.N.; Barbosa, M.; Lamers, P.P. Can We Approach Theoretical Lipid Yields in
Microalgae? Trend Biotechnol. 2018, 36, 265–276. [CrossRef] [PubMed]
9. Chiu, S.Y.; Kao, C.Y.; Huang, T.-T.; Lin, C.J.; Ong, S.C.; Chen, C.D.; Chang, J.S.; Lin, C.S. Microalgal biomass
production and on-site bioremediation of carbon dioxide, nitrogen oxide and sulfur dioxide from flue gas
using Chlorella sp. cultures. Bioresour. Technol. 2011, 102, 9135–9142. [CrossRef] [PubMed]
10. Hende, S.V.D.; Vervaeren, H.; Boon, N. Flue gas compounds and microalgae:(Bio-) chemical interactions
leading to biotechnological opportunities. Biotechnol. Adv. 2012, 30, 1405–1424. [CrossRef] [PubMed]
11. Liang, Y.; Sarkany, N.; Cui, Y. Biomass and lipid productivities of Chlorella vulgaris under autotrophic,
heterotrophic and mixotrophic growth conditions. Biotechnol. Lett. 2009, 31, 1043–1049. [CrossRef] [PubMed]
12. Guccione, A.; Biondi, N.; Sampietro, G.; Rodolfi, L.; Bassi, N.; Tredici, M.R. Chlorella for protein and biofuels:
From strain selection to outdoor cultivation in a green wall panel photobioreactor. Biotechnol. Biofuel 2014, 7,
167
Energies 2018, 11, 2026
13. Muthuraj, M.; Kumar, V.; Palabhanvi, B.; Das, D. Evaluation of indigenous microalgal isolate Chlorella sp. FC2
IITG as a cell factory for biodiesel production and scale up in outdoor conditions. J. Ind. Microbiol. Biotechnol.
2014, 41, 499–511. [CrossRef] [PubMed]
14. Nascimento, I.A.; Marques, S.S.I.; Cabanelas, I.T.D.; Pereira, S.A.; Druzian, J.I.; de Souza, C.O.; Vich, D.V.; de
Carvalho, G.C.; Nascimento, M.A. Screening microalgae strains for biodiesel production: Lipid productivity
and estimation of fuel quality based on fatty acids profiles as selective criteria. Bioenergy Res. 2013, 6, 1–13.
[CrossRef]
15. Sun, Z.; Zhou, Z.G.; Gerken, H.; Chen, F.; Liu, J. Screening and characterization of oleaginous Chlorella strains
and exploration of photoautotrophic Chlorella protothecoides for oil production. Bioresour. Technol. 2015, 184,
16. Li, L.; Cui, J.; Liu, Q.; Ding, Y.; Liu, J. Screening and phylogenetic analysis of lipid-rich microalgae. Algal Res.
2015, 11, 381–386. [CrossRef]
17. Xia, L.; Song, S.; He, Q.; Yang, H.; Hu, C. Selection of microalgae for biodiesel production in a scalable
outdoor photobioreactor in north China. Bioresour. Technol. 2014, 174, 274–280. [CrossRef] [PubMed]
18. Cheirsilp, B.; Torpee, S. Enhanced growth and lipid production of microalgae under mixotrophic culture
condition: Effect of light intensity, glucose concentration and fed-batch cultivation. Bioresour. Technol. 2012,
19. Kim, B.; Praveenkumar, R.; Lee, J.; Nam, B.; Kim, D.M.; Lee, K.; Lee, Y.C.; Oh, Y.K. Magnesium
aminoclay enhances lipid production of mixotrophic Chlorella sp. KR-1 while reducing bacterial populations.
20. Abomohra, A.E.-F.; El-Sheekh, M.; Hanelt, D. Screening of marine microalgae isolated from the hypersaline
Bardawil lagoon for biodiesel feedstock. Renew. Energy 2017, 101, 1266–1272. [CrossRef]
21. Tanimura, A.; Takashima, M.; Sugita, T.; Endoh, R.; Kikukawa, M.; Yamaguchi, S.; Sakuradani, E.; Ogawa, J.;
Shima, J. Selection of oleaginous yeasts with high lipid productivity for practical biodiesel production.
22. Vello, V.; Phang, S.M.; Chu, W.L.; Majid, N.A.; Lim, P.E.; Loh, S.K. Lipid productivity and fatty
acid composition-guided selection of Chlorella strains isolated from Malaysia for biodiesel production.
J. Appl. Phycol. 2014, 26, 1399–1413. [CrossRef]
23. Praveenkumar, R.; Kim, B.; Choi, E.; Lee, K.; Cho, S.; Hyun, J.; Park, J.; Lee, Y.; Lee, H.; Lee, J. Mixotrophic
cultivation of oleaginous Chlorella sp. KR-1 mediated by actual coal-fired flue gas for biodiesel production.
Bioproc. Biosyst. Eng. 2014, 37, 2083–2094. [CrossRef] [PubMed]
24. Hoekman, S.K.; Broch, A.; Robbins, C.; Ceniceros, E.; Natarajan, M. Review of biodiesel composition,
properties, and specifications. Renew. Sustain. Energy Rev. 2012, 16, 143–169. [CrossRef]
25. Valdez-Ojeda, R.; González-Muñoz, M.; Us-Vázquez, R.; Narváez-Zapata, J.; Chavarria-Hernandez, J.C.;
López-Adrián, S.; Barahona-Pérez, F.; Toledano-Thompson, T.; Garduño-Solórzano, G.; Medrano, R.M.E.-G.
Characterization of five fresh water microalgae with potential for biodiesel production. Algal Res. 2015, 7,
33–44. [CrossRef]
26. Sung, K.D.; Lee, J.S.; Shin, C.S.; Park, S.C.; Choi, M.J. CO2 fixation by Chlorella sp. KR-1 and its cultural
characteristics. Bioresour. Technol. 1999, 68, 269–273. [CrossRef]
27. Yun, M.; Oh, Y.K.; Praveenkumar, R.; Seo, Y.S.; Cho, S. Contaminated bacterial effects and qPCR application
to monitor a specific bacterium in Chlorella sp. KR-1 culture. Biotechnol. Bioprocess. Eng. 2017, 22, 150–160.
[CrossRef]
168
energies
Article
Thermophilic Anaerobic Digestion: Enhanced and
Sustainable Methane Production from Co-Digestion
of Food and Lignocellulosic Wastes
Aditi David 1 , Tanvi Govil 1 , Abhilash Kumar Tripathi 1 , Julie McGeary 2 , Kylie Farrar 3
and Rajesh Kumar Sani 1,4, *
1 Department of Chemical and Biological Engineering, South Dakota School of Mines and Technology,
Rapid City, SD 57701, USA; [email protected] (A.D.); [email protected] (T.G.);
[email protected] (A.K.T.)
2 Biology Teacher, Central High School, Rapid City, SD 57701, USA; [email protected]
3 Science Teacher, Russell Middle School, Colorado Springs, CO 80918, USA; [email protected]
4 BuG ReMeDEE Consortium, South Dakota School of Mines and Technology, Rapid City, SD 57701, USA
Abstract: This article aims to study the codigestion of food waste (FW) and three different
lignocellulosic wastes (LW) (Corn stover (CS), Prairie cordgrass (PCG), and Unbleached paper
(UBP)) for thermophilic anaerobic digestion to overcome the limitations of digesting food waste alone
(volatile fatty acids accumulation and low C:N ratio). Using an enriched thermophilic methanogenic
consortium, all the food and lignocellulosic waste mixtures showed positive synergistic effects
of codigestion. After 30 days of incubation at 60 ◦ C (100 rpm), the highest methane yield of
305.45 L·kg−1 volatile solids (VS) was achieved with a combination of FW-PCG-CS followed by
279.31 L·kg−1 VS with a mixture of FW-PCG. The corresponding volatile solids reduction for these
two co-digestion mixtures was 68% and 58%, respectively. This study demonstrated a reduced
hydraulic retention time for methane production using FW and LW.
Keywords: thermophilic anaerobic digestion; corn stover; prairie cord grass; unbleached paper;
digester performance; process stability; synergistic effects; microbial community; Methanothermobacter
1. Introduction
Anaerobic digestion for the production of biogas is an environmentally friendly multi-step process
employing complex consortia of microorganisms. These consortia comprise various facultative or
obligate anaerobic microbial groups which work synergistically and convert complex organic substrates
into biogas. Biogas with 60–70% of its component being methane is a combustible renewable energy
that can be used as an alternative energy source to replace fossil fuels, either by direct combustion to
generate heat and electricity, or through upgradation to be used as vehicle fuel and injection into the
gas grid [1–4]. After the on-site demand of the produced biogas is met, the remaining biogas is usually
stored as compressed natural gas (CNG) or liquefied biomethane (LBM) for future use. While the
biogas industry in Europe is well established, with more than 10,000 biogas producing Anaerobic
digestion (AD) plants in operation, in the United States, the biogas industry is still growing. According
to the American Biogas Council, 1241 wastewater treatment plants and 236 farms have functional
Biogas plants [5]. Currently, the feedstocks used for biogas production in these anaerobic digesters are
limited, and therefore it is important to explore new substrates to meet the growing energy needs.
Food waste (FW) is an attractive substrate for AD because of its low total solids (TS) and
high content of soluble organics which make it readily biodegradable. Additionally, FW offers

Energies 2018, 11, 2058
low-cost alternatives for methane production and is abundantly available. The Food and Agriculture
Organization of the United Nations reports a wastage of 1.3 billion tons of food per year. Therefore,
using FW to produce methane appears to be an ideal way to strengthen the world’s energy security
while addressing waste management and nutrient recycling [6]. The AD of FWs, however, depends on
the delicate balance between the acidification process and methanogenesis [7]. Many studies have
attributed the low stability and efficiency of individual FW fermentation to the low C/N ratio, ammonia
accumulation, and the readily biodegradable organic fraction that causes acidification [8]. If the
acidification process is rapid, the accumulation of volatile fatty acids (VFAs) occurs, causing an abrupt
fall in pH, which stresses and inhibits the methanogenic archaea [9,10]. The reactor acidification
through reactor overload is one of the most common reasons for process deterioration in anaerobic
digesters [11]. Therefore, the codigestion of FW with other organic wastes, such as municipal
wastewater treatment plant sludge, animal manure, and agricultural biomass, has become popular, as it
provides the ability to alter the carbon and nitrogen ratios which prevent accumulation of intermediate
inhibitory products (NH4 + and VFAs) [12–14] and increase the methane yield, as well as improve the
utilization efficiency [8]. The use of co-substrates for AD has shown to produce higher biogas yields
due to positive synergisms established in the digestion medium and the supply of missing nutrients
by the co-substrates [15,16]. Mixing organic substrates often results in the formation of a codigestion
mixture with a C/N ratio included in the optimal range of 20–30% as reported in the literature [16–18].
The additional benefits of the codigestion process are (1) the dilution of the potential toxic compounds
present in any of the co-substrates involved; (2) the adjustment of the moisture content and pH;
(3) supply of the necessary buffer capacity to the mixture; (4) the increase of the biodegradable material
content; and (5) the widening the range of bacterial strains taking part in the process [17].
Like food waste, lignocellulosic wastes (LW), which include agricultural residues (corn stover,
rice straw), herbaceous crops (switchgrass, prairie cordgrass) and waste paper, are extensively
available low-cost substrates. Lignocellulosic wastes are in fact the most abundant renewable organics,
reaching, annually, over 150 billion tons in the form of plant biomass [19]. Due to their chemical
composition based on sugars and other compounds of interest, they could be utilized to produce
several value-added products, such as ethanol, biogas, food additives, organic acids, enzymes,
and others [20–23]. Therefore, besides the environmental problems caused by their accumulation,
the non-utilization of these materials constitutes a loss of potentially valuable sources [24]. Previous
studies report that the methane generation potential is expected to be much higher if lignocellulosic
biomass resources are used. It is estimated that 4.2 trillion cubic feet per year (about 4,318 trillion
British thermal units [25]) LW are available, the biogasification of which can displace about 46% of
current natural gas consumption in the electric power sector and the entire natural gas consumption in
the transportation sector [26]. Moreover, being a second-generation feedstock for biofuel, the use of
lignocellulosic residues does not compete for arable land [27,28]. Thus, utilizing LW in the AD process
is a promising option. Still, the AD of LW alone has several limitations, such as the high C:N ratio
creating nitrogen deficiency, the risk of producing inhibitors (e.g., furfural and hydroxymethylfurfural),
a less digestible biomass, a high heat demand [29,30], and long digestion time due to their low cellulose,
hemicellulose, and lignin conversion rates [8].
Performing anaerobic codigestion of FW with LW can overcome the limitations of their respective
mono-digestions. Previous studies have also demonstrated the role of codigestion of FW and LW
in the adjustment of C:N ratio [31], in the reduction of the start-up time and volatile fatty acid
(VFA) accumulation [32], and in the enhancement of methane yield. Yong et al., 2015 examined the
potential of co-digestion of food waste and straw at 35 ◦ C. The study showed increased methane
yield reaching 0.392 m3 ·kg−1 VS at an optimal mixing ratio of 5:1 (FW: straw), which marked a 39.5%
and 149.7% increase in methane yield compared with individual digestion results of food waste
and straw, respectively [8]. In another study, Jabeen et al., 2015, codigested FW with rice husk in
a pilot-scale mesophilic (37 ◦ C) anaerobic reactor. They obtained a daily biogas production of 196 L·d−1 ,
at an organic loading rate of 6 kg VS·m−3 ·d−1 , which decreased to 136 L·d−1 at a loading rate of
170
Energies 2018, 11, 2058
9 kg VS/m3 /d. They attributed the decrease in biogas production, reactor stability, and volatile solids
removal efficiency to the increase in organic loading rate [33]. These effects of high organic loading
rate are more profound in mesophilic anaerobic reactors as compared to thermophilic reactors.
In all the studies mentioned above, the anaerobic co-digestion of FW and LW improved process
performance and increased biogas production. However, a major disadvantage of using LW is
the recalcitrant nature of their cell wall. The intricate composition of lignocellulosic materials,
where cellulose fiber is tightly linked to hemicellulose and lignin, hinders their biodegradability and
thus limits their use as co-substrates for AD. The recalcitrant nature of the LW necessitates the inclusion
of the pretreatment step (physical, chemical or biological) for their efficient conversion into biofuels [34].
The use of thermophilic digesters can eliminate the use of costly and sometimes hazardous pretreatment
steps. Thermophilic anaerobic digestion (TAD) allows for the better degradation of LW by increasing
the microbial hydrolysis rate which is considered the slowest as well as rate limiting step [35].
Besides aiding methanogenesis through increased hydrolysis rate, TAD also decreases the level of
pathogens. This reduces the potential health hazards for the biogas plant operators as well as ensures
safe disposal of the digestate after process completion [36].
Most of the studies to date on the codigestion of FW and LW have been conducted under
mesophilic conditions [31,33,37–39] which require higher hydraulic retention times (HRTs) and
lower organic loading rates. Here, however, we attempt to assess the methane production from
codigestion of FW and LW under thermophilic conditions. The objectives of this study were to
(1) evaluate the methane potential and biodegradability of FW, corn stover (CS), prairie cordgrass
(PCG), and unbleached paper (UBP) at 60 ◦ C and (2) describe the trends of methane yield, process
stability, and digester performance, when these are codigested using a pre-acclimatized thermophilic
methanogenic consortium. In South Dakota, PCG is abundant and can offer local alternatives to corn
for the production of biofuels. This study is the first attempt at investigating the co-digestion of FW
with PCG under thermophilic conditions (60 ◦ C). The results of this study could provide baseline data
for the adoption of Thermophilic Anaerobic Digestor (TAD) using readily available PCG as feedstocks
for biogas generation.
2.1. Feedstock and Inoculum

The FW and three different types of LW-CS, PCG, and UBP—were used as feedstocks for TAD.
CS and PCG were kindly provided by Dr. K. Muthukumarappan from South Dakota State University,
Brookings, SD. The FW and UBP were collected from the cafeteria at the South Dakota School of
Mines and Technology and stored at 4 ◦ C. LW were reduced in size using a cutting mill and sieved
through sieve between 15 and 10 mm pores (SM 200, Rectch GmbH, Haan, Germany). The particle
size of the FW was also reduced by crushing it in an electrical kitchen blender, and the resultant FW
slurry was sieved to remove coarse particles larger than 15 mm. The effect of shredding paper had
been tested previously and did not influence the AD performance [40], so the UBP was hand cut.
Inoculum used in this study was effluent procured from an anaerobic digester at the Waste Water
Reclamation Plant, Rapid City, SD which was enriched in our lab. The enrichment details are included
in our previous report [41]. This enriched thermophilic methanogenic consortia (called TMC) was used
as the inoculum for this study and stored in an air-tight container at −20 ◦ C. The characteristics of
substrates and inocula are presented in Table 1. Each data point presented was the average of triplicate
measurements of the same feedstock.
2.2. Experimental setup for Batch Tests

Batch tests were conducted in 500 mL serum bottles with a working volume of 200 mL containing
180 mL of anaerobic culture medium and 20 mL (10%, v/v) TMC. The composition of the anaerobic
culture medium is given in Table 1. To analyze the feasibility of the individual FW and LW for
171
Energies 2018, 11, 2058
anaerobic digestion, different organic loadings (1%, 2%, 5%, and 10% w/v VS) of these samples were
first tested, which constituted 16 mono-digestion tests (Table 2). Based on this, seven FW samples
(1 g VS each) were separately mixed with CS, PCG, or UBP for the anaerobic co-digestion study in
mixing ratios as given in Table 3. The substrate-free serum bottles containing inoculum, and the media
served as controls. The experiments were repeated three times under the same conditions using the
same co-digestion mixing ratios. The average pH of digestion mixtures T1, T2, T3, T4, T5, T6, and T7
after mixing with inoculum were 7.1, 7.1, 7.2, 7.1, 7.2, 7.3, and 7.3, respectively.
The substrates (according to Tables 1 and 2), media and 10% (v/v) inoculum were added in the
serum bottles to make a total working volume of 200 mL. These digester bottles were capped with
butyl rubber stoppers, crimped with aluminum caps, and nitrogen gas was purged into them for
20 min to simulate anaerobic conditions. The bottles were then incubated at 60 ◦ C (100 rpm) for up to
30 days. Methane production was monitored at different time intervals, and the methane yield was
reported as mL CH4 per gram of Volatile solids (mL CH4 /g VS).
Table 1. The composition of the anaerobic culture medium.
Components Composition (g/L)

K2 HPO4 0.30
KH2 PO4 0.30
NaCl 0.10
CaCl2 0.05
NH4 Cl 1.00
MgCl2 ·6H2 O 0.50
KCl 0.30
Cysteine.HCl 0.50
Yeast extract 0.05
Na2 S·9H2 O 0.003
NaHCO3 20 mM
Nitsch trace element 2.5 mL
Table 2. The feedstock composition in the batch tests with individual wastes.
Individual Substrate Food Waste Corn Stover Prairie Cordgrass Unbleached Paper
Type (FW) (CS) (PCG) (UBP)
Organic Loading (%) s 1% 2.5% 5% 10% 1% 2.5% 5% 10% 1% 2.5% 5% 10% 1% 2.5% 5% 10%
Composition
2 5 10 20 2 5 10 20 2 5 10 20 2 5 10 20
(g VS)
Table 3. The feedstock composition in batch tests (60 ◦ C) with codigestion mixtures.
Test Bottles T1 T2 T3 T4 T5 T6 T7
Composition FW FW + CS FW + PCG FW + UBP FW + PCG + CS FW + PCG + UBP FW + UBP + CS
FW (g VS) 2.5 2.5 2.5 2.5 2.5 2.5 2.5
CS (g VS) - 1.5 - - 0.75 - 0.75
PCG (g VS) - - 1.5 - 0.75 0.75 -
UBP (g VS) - - - 1.5 - 0.75 0.75
Mixing Ratio - 5:3 5:3 5:3 5:1.5:1.5 5:1.5:1.5 5:1.5:1.5
C:N 19:1 23.9 25.6 28.2 24.4 26.8 25.6
(Note: FW = Food wastes; CS = Corn stover; PCG = Prairie cordgrass; UBP = Unbleached paper).

Total Solids (TS) and Volatile solids (VS) of the inoculum and feedstock were analyzed according
to APHA Standard Methods [42]. The cellulose, hemicellulose, and lignin content were determined
according to standard NREL analytical procedures [43]. The product gas was sampled using a 100 μL
gas-tight glass syringe (Hamilton Company, Reno, NV, USA), and measured using gas chromatography
(Agilent Technologies 7890A) equipped with a thermal conductivity detector at 200 ◦ C and a Supelco
172
Energies 2018, 11, 2058
Porapak Q stainless steel packed column (6 ft × 1/8 in). Nitrogen was the carrier gas at a rate of
10 mL/min. The conditions for CH4 analysis were as follows: injector temperature: 70 ◦ C; detector
temperature: 100 ◦ C; and oven temperature: 1 min at 35 ◦ C, followed by a 5 ◦ C/min ramp to 50 ◦ C
with a hold time of 2 min (total run time 4.5 min). The VFAs (acetic acid, propionic, and butyric acid)
were measured using an Aminex HPX-87 H column with 0.005 N sulfuric acid as the mobile phase.
Liquid samples were sent to Atlantic Microlab, Norcross, GA for elemental analysis.
Methane yield (mL/g VS) was calculated as the volume of methane as produced per g of VS
feedstock loaded into the digester bottles initially and corrected by subtracting the methane yield
obtained from the control bottle [44]. VS represents the organic portion of the material solids that
can be digested, while the remainder of the solids is considered as fixed. The ‘fixed’ solids are
non-biodegradable [45]. The pH value of each mixture before and after the digestion process was
measured with an Oakton pH 700 meter. Process stability was assessed in terms of final pH and
VFA accumulation, and performance was measured in terms of methane production and VS removal
efficiencies. The carbon to nitrogen ratio was calculated using the following equation [37]:
(VS ∗ TOC) FW + (VS ∗ TOC) LW

C:N =
(VS ∗ TN) FW + (VS ∗ TN) LW
where TOC = the total organic carbon (%VS) and TN = total nitrogen (%VS).
2.4. Statistical Analysis

The batch tests were conducted in triplicates, and each data point was the average of triplicate
readings for all the chemical analysis and measurements. The statistical significance of the results was
analyzed by Analysis of Variance (Two-way ANOVA) with a 95% level of confidence (p = 0.05) [46].
Statistical analysis was done using GraphPad Prism version 7.04 (GraphPad Software, San Diego, CA,
USA). The results of the analysis are given in the supplementary files.
3.1. Feedstock and Inoculum Characteristics

The results of the feedstock characterization are shown in Table 4. FW had a moisture content
of 84.8% and a VS/TS ratio of 90.8%. These results comply with previous reports in the literature where
the moisture content and VS/TS ratio of FW was 69–93% and 85–95%, respectively [47]. FW containing
a high amount of digestible organic matter is suitable for anaerobic microbial growth and attaining
high methane yield. However, elemental analysis showed that FW had a C:N ratio of 13:1, which was
too low to maintain nutrient balance in the anaerobic digester. TS and VS contents of the three LW (CS,
PCG, UBP) were higher than that of FW, and their VS/TS ratios were between 85% to 95% making
them suitable feedstocks for TAD (Table 1). The C:N ratio was found to be the highest for UBP
(124:1), followed by CS (55:1) and PCG (41:1), which was higher than the optimum range required
(20–30:1) [48]. The cellulose content was the highest in UBP (84.5%) and lowest in PCG (30.3%). CS had
the highest hemicellulose content among the tested feedstocks and PCG had the highest lignin content.
UBP had the lowest hemicellulose and lignin content.
The inoculum (TMC) used in this study was enriched through sub-culturing techniques [41].
Briefly, 10% (v/v) effluent from the wastewater reclamation plant was inoculated in 200 mL of anaerobic
medium (Table 1) in 500 mL serum bottles. One-gram VS of mixed wastes (containing equal amounts of
FW, CS, PCG, and UBP) was used as the substrate and the bottles were incubated at 60 ◦ C and 100 rpm.
When methane production reached a stable level, 10% (v/v) of the actively growing anaerobic culture
was transferred into fresh media (200 mL) containing another 1 g of mixed waste. After 10 serial
transfers, a methanogenic consortium growing at 60 ◦ C was obtained that produced methane using
FW, CS, PCG, and UBP as a carbon and energy source. The enriched TMC had a VS content, TS content,
and C:N ratio of 6.1%, 8.3%, and 2.7, respectively. During optimization, a pH of 7.5, 2–3% (w/v)
173
Energies 2018, 11, 2058
substrate loading, and 10% (v/v) inoculum density had a profound effect on consortium growth and
subsequent methane production. The microbial community analysis of the inoculum highlighted
the role of the bacterial orders—Clostridiales, Bacillales, Bacteroidales, and Thermoanaerobacteriales—in
the anaerobic degradation of the complex organic polymers present in FW and LW. As expected,
the hydrogenotrophic pathway was found to be the dominant pathway of methanogenesis where
Methanothermobacter were the predominant archaea in TAD (Supplementary Figure S2A–D).
Table 4. The characteristics of inoculum and feedstock a .
Parameters Inoculum FW CS PCG UBP

TS (% w/w) 8.3 15.2 94.3 87.6 96.2
VS (% w/w) b 6.1 13.8 92.6 76.9 84.7
Ash (% w/w) b 93.9 86.2 7.4 23.1 15.3
VS/TS (%) 37.2 90.8 98.2 87.8 88.0
C:N c 2.7 13:1 55:1 41:1 124:1
VFA/alkalinity 1.3 ND ND ND ND
Lignin (%) b ND ND 13.5 22.9 2.1
Cellulose (%) b ND ND 38.2 30.3 84.5
Hemicellulose (%) b ND ND 32.4 25.7 11.2
ND = Not determined; a Data shown are average values based on duplicate runs; b Based on the TS of the sample;
c Based on the total weight of the sample.
3.2. Thermophilic Anaerobic Digestion of Individual Waste Substrates

Cumulative methane production for organic loadings of 1, 2.5, 5, and 10% (w/v VS) of FW is
shown in Figure 1. The highest methane yield of 321.5 L·kg−1 VS was obtained with an organic
loading of 1% FW. The further increase of organic loading resulted in a significant decrease in methane
production (p < 0.05), indicating inhibition of methanogenesis. This is due to the rapid rise in the
concentration of VFAs with increasing organic load, causing an abrupt fall in pH (<5.3) and the
consequent acidification of the anaerobic system (Figure 2). Previous studies have also reported
a similar trend, in which the accumulation of superfluous VFAs was found to be inhibitory towards
methanogens, thus causing a drop in their activity [49,50]. The instability of the FW digesting system
can also be attributed to the low C:N ratio of FW. A low C:N ratio causes ammonia accumulation,
reduced substrate degradation, and even an inhibition of methanogenesis [8].
The codigestion of FW with lignocellulosic biomass has shown to solve the problem by bringing
in an opportunity to balance the nutrients content [15,51]. However, the complex structures of lignin
and other cell wall polysaccharides make lignocellulosic waste materials hard to biodegrade and to
be used by anaerobic microorganisms, leading to a lower biogas production. Therefore, to assess the
biodegradability of LW by the anaerobic consortia used in this study, CS, PCG, and UBP were digested
individually at different organic loading rates (1, 2.5, 5, and 10%; w/v VS). Figure 1 presents the results
of the mono-digestion study. The methane yield and solid content for all the substrates tested showed
an inverse relationship where 1% w/v substrate gave the highest methane yield. Cumulative methane
yields obtained with 1% LW were 236.14 L·kg−1 VS for UBP followed by CS (111.05 L·kg−1 VS) and
PCG (94.87 L·kg−1 VS). The methane yields obtained in this study were comparable with values (80 to
530 L·kg−1 VS) reported by others for the same or similar agricultural and energy crops [39,52–54].
Substrate concentrations of more than 1% w/v showed a significant decrease in methane production
(p < 0.05). Nevertheless, these results indicated that the TMC can efficiently degrade LW anaerobically.
174
Energies 2018, 11, 2058
Figure 1. The methane yield from different organic loadings of food waste (FW), corn stover (CS),
prairie cord grass (PCG), and unbleached paper (UBP). The data represented are the average values of
the triplicate bottles. The error bars represent the standard deviation.
Figure 2. The Total Volatile Fatty Acid (TVFA) concentration and pH at the end of batch tests
(the data represented are the average values of the triplicate bottles. The error bars represent the
standard deviation).
3.3. Effect of Codigestion on Methane Yield and Digester Performance

High organic loading rates are desirable to reduce the digester volume and minimize capital
losses, however, increased organic loading rates often result in VFA accumulation, nutrient imbalance,
and sometimes the direct inhibition of methanogenesis, leading to complete digester failure [55–57].
Codigestion experiments were carried out with an organic loading of 2.5% FW to analyze the efficacy
of codigestion in overcoming the limitations of high organic loadings observed in mono-digestion
experiments. In the first set of tests, FWs (2.5% VS) were codigested with LW individually in three
separate serum bottle digesters (T2, T3, and T4) where the concentration of CS, PCG, and UBP was
1.5% w/v (VS basis). As shown in Figure 3a, methane volume produced in these co-digestion studies
with two waste mixtures was considerably higher than that obtained from the mono-digestion of
FW (p < 0.05). This demonstrates positive synergistic effects among codigested wastes. The highest
methane yield of 279.31 L·kg−1 was observed in T3 (FW + PCG), followed by 251.90 L·kg−1 in T2
(FW + CS) and 177.48 L·kg−1 in T4 (FW + UBP). Because a higher reduction in VS is indicative of
better digester performance, a direct correlation between methane production and VS reduction was
observed in this study (p < 0.05). Digester bottles with higher VS reduction yielded higher methane.
175
Energies 2018, 11, 2058
The highest VS reduction of 45.7% was achieved in T3 (FW + PCG) followed by 38.7% in T2 (FW + CS),
and 38.3% in T4 (FW + UBP). The lowest biodegradability of the FW + UBP mixture is reflected in the
mixture’s lower VS removal and consequently in its lowest methane yield.
The second set of experiments where the co-digestion mixtures comprised of 2.5% w/v FW and
two LW (0.75 % w/v each on VS basis) showed a similar pattern of results (Figure 3b). The highest
methane yield of 305.45 L·kg−1 VS and a VS reduction of 68.5% were obtained with FW:CS:PCG (T5)
mixture, while the lowest yield of 219.9 L·kg−1 VS and a VS reduction of 49.8% was observed with FW
+ UBP + CS (T7) mixture. The higher methane production for some codigestion mixtures over others
can be attributed to the C/N ratio for these mixtures which were in the recommended range of 20:1
and 30:1 [18,58].
At a constant organic loading of 4 g VS, the highest methane yield among all the codigested
mixtures was obtained with FW + PCG + CS (305.4 L·kg−1 VS), followed by FW + PCG (279.3 L·kg−1
VS). This marked a 94% and 74% increase in methane production compared to the mono-digestion of
FW. Previous studies have also shown similar results but were performed at mesophilic temperatures
and had higher hydraulic retention times (>30 days) in comparison to our study (18 days). Xu and Li,
2012 reported the highest methane yield of 304.4 L/kg VS feed with a codigestion mixture containing
an equal amount of corn stover and dog food, 109% compared to methane yield from mono-digestion
dog food, respectively [32]. In another study conducted by Brown and Li, 2013, the increased methane
yields and volumetric productivities were observed with the co-digestion of food waste with yard
waste. Food waste to yard waste mixture of 1:9 gave the highest volumetric productivity of 8.6 L
methane L−1 , with a 43% VS reduction [31]. They attributed nutrient balance as one of the reasons for
enhanced synergism during co-digestion.
(a)
(b)
Figure 3. (a,b): methane production from the co-digestion of food and lignocellulosic waste (the data
represented are the average values of triplicate bottles. The error bars represent the standard deviation).
176
Energies 2018, 11, 2058
3.4. Effect of Codigestion on Process Stability

Reactor acidification due to the buildup of VFAs is one of the most common reasons for the process
deterioration and instability in anaerobic digesters [48]. VFAs are naturally produced by acidogenic and
acetogenic bacteria and accumulate during conditions of substrate overload in the reactor. Additionally,
a high VFA concentration is usually associated with a drop in pH and a breakdown of the buffering
capacity of the reactor [31,59]. Hence, the final pH values in batch tests are another measure of
system stability. Maintenance of the system pH in the proper range (6.5 and 7.6) is required for
efficient anaerobic digestion [60]. Analysis of VFA concentrations and pH were done at the end of
the batch experiments, and the results are shown in Figure 4. While acetic acid represented major
portion of the total VFAs produced, propionic acid and butyric acid were produced in lower amounts
(Supplementary Figure S1). The total VFA concentration was highest for T1 (FW) followed by T4 (FW
+ UBP) and T7 (FW + UBP + CS) which dropped the pH below 6 in all the three digesters creating
instability. This instability directly affected the methane production which is evident from the lowest
methane yield observed in T1, T4, and T7 (Table 5). The results indicated that these anaerobic reactors
were operating under higher loading rates. Statistical analysis also showed that VFA concentration
has a significant effect on methane yield with a p value of <0.05. The serum bottles having VFA
concentrations of 1600–3300 mg/L did not show a drastic decrease in the pH of the system.
Figure 4. The Total Volatile Fatty acids (TVFA) and pH at the end of the codigestion tests (the data
represented are the average values of triplicate bottles. The error bars represent the standard deviation).
Table 5. The effect of codigestion on digester performance.
Test Bottles T1 T2 T3 T4 T5 T6 T7
Methane yield (L·kg−1 VS) 159.8 ± 2.4 251.9 ± 26.6 279.3 ± 7.2 177.5 ± 16.1 305.4 ± 23.8 243.6 ± 20.9 219.9 ± 10.9
VS reduction (%) 44 ± 1.4 57.4 ± 2.7 58.0 ± 0.9 47.4 ± 1.8 68.5 ± 2.2 51.1 ± 2.6 49.8 ± 1.5
Initial pH 7.1 ± 0.0 7.1 ± 0.2 7.2 ± 0.1 7.1 ± 0.1 7.2 ± 0.0 7.3 ± 0.3 7.3 ± 0.2
Final pH 5.3 ± 0.0 6.8 ± 0.1 7.4 ± 0.2 5.8 ± 0.1 7.3 ± 0.2 6.8 ± 0.1 5.9 ± 0.0
4. Conclusions
Biogas production using the abundant wastes as the substrate is a promising technology both in
renewable energy and solid waste management sectors. FW and LW are attractive co-substrates for
TAD as their complementary characteristics can overcome the limitations faced by their mono-digestion.
The results suggested that codigesting FW with LW not only improved the system stability but also
enhanced methane production, thus improving overall digester performance. At a constant organic
loading of 4 g VS, the highest methane yield of 305.4 L·kg−1 VS was obtained with FW + PCG +
CS followed by 279.3 L·kg–1 VS with FW + PCG. This marked a 94% and 74% increase in methane
177
Energies 2018, 11, 2058
production compared to the mono-digestion of FW. Since the digester bottles with these codigestion
mixtures showed increased system stability as well, they will be employed for scale-up studies.
The pretreatment of the LW increases the amount of available sugars for AD, but also increases
the bioprocessing cost. As this study was conducted without employing any pretreatment, consortium
development became a crucial step when the codigestion of LW was performed. The use of
thermophilic temperature may increase the costs of the overall process but gives the added advantage
of digesting higher organic loading of wastes at reduced hydraulic retention times. In addition,
the surplus biogas produced can be used for heating purposes to maintain the reactor temperature in
the thermophilic range. Nevertheless, the consortium developed in this study is significant for the
co-digestion and can be used for scale-up studies. Codigestion mixtures showing the best performance
and stability in this study will be used for scale-up studies to conduct the cost analysis. The present
batch study can also direct further studies for setting up semi-continuous or continuous commercial
scale plants for the codigestion of regionally abundant PCG with food waste and help in tackling the
conditions limiting methane production from higher organic loadings.
Supplementary Materials: The following are available online at http://www.mdpi.com/1996-1073/11/8/2058/s1.

Author Contributions: A.D. conceived and designed the experiments. A.D., J.M., and K.F. performed the
experiments. T.G. and A.K.T. analyzed the data, and A.D. and T.G. wrote the paper. R.K.S. contributed in the idea,
experiment planning, writing, validity and interpretation of the results.
Funding: This research was supported by the US Air Force under the Biological Waste to Energy Project (Award
#FA4819-14-C-0004) and Sustainable Development-Research Experience for Teachers (SD-RET) (Award #1711946).
National Science Foundation also provided financial support in the form of the BuG ReMeDEE initiative (Award
#1736255) The authors are grateful to the Department of Chemical and Biological Engineering at the South Dakota
School of Mines and Technology for the research support.
Nomenclature
FW Food wastes
LW Lignocellulosic wastes
CS Corn Stover
PCG Prairie cordgrass
UBP Unbleached paper
AD Anaerobic digestion
VFA Volatile Fatty acids
C/N Carbon to nitrogen
TAD Thermophilic anaerobic digestion
VS Volatile solids
TS Total solids
TMC Thermophilic methanogenic consortia
References
1. Qian, Y.; Sun, S.; Ju, D.; Shan, X.; Lu, X. Review of the state-of-the-art of biogas combustion mechanisms and
applications in internal combustion engines. Renew. Sustain. Energy Rev. 2017, 69, 50–58. [CrossRef]
2. Börjesson, P.; Mattiasson, B. Biogas as a resource-efficient vehicle fuel. Trends Biotechnol. 2008, 26, 7–13.
[CrossRef] [PubMed]
3. Hengeveld, E.J.; van Gemert, W.J.T.; Bekkering, J.; Broekhuis, A.A. When does decentralized production of
biogas and centralized upgrading and injection into the natural gas grid make sense? Biomass Bioenergy 2014,
67, 363–371. [CrossRef]
4. Ptak, M.; Koziołek, S.; Derlukiewicz, D.; Słupiński, M.; Mysior, M. Analysis of the Use of Biogas as Fuel
for Internal Combustion Engines. In Proceedings of the 13th International Scientific Conference, Wrocław,
Poland, 22–24 June 2016; Rusiński, E., Pietrusiak, D., Eds.; Springer International Publishing: Cham,
Switzerland, 2017; pp. 441–450.
178
Energies 2018, 11, 2058
5. Dahl, R. A Second Life for Scraps: Making Biogas from Food Waste. Environ. Health Perspect. 2015, 123,
A180–A183. [CrossRef] [PubMed]
6. Islam, M.; Park, K.-J.; Yoon, H.-S. Methane Production Potential of Food Waste and Food Waste Mixture
with Swine Manure in Anaerobic Digestion. J. Biosyst. Eng. 2012, 37, 100–105. [CrossRef]
7. Braguglia, C.M.; Gallipoli, A.; Gianico, A.; Pagliaccia, P. Anaerobic bioconversion of food waste into energy:
A critical review. Bioresour. Technol. 2018, 248, 37–56. [CrossRef] [PubMed]
8. Yong, Z.; Dong, Y.; Zhang, X.; Tan, T. Anaerobic co-digestion of food waste and straw for biogas production.
9. Staley, B.F.; de los Reyes, F.L.; Barlaz, M.A. Effect of Spatial Differences in Microbial Activity, pH, and
Substrate Levels on Methanogenesis Initiation in Refuse. Appl. Environ. Microbiol. 2011, 77, 2381–2391.
[CrossRef] [PubMed]
10. Manyi-Loh, C.E.; Mamphweli, S.N.; Meyer, E.L.; Okoh, A.I.; Makaka, G.; Simon, M. Microbial Anaerobic
Digestion (Bio-Digesters) as an Approach to the Decontamination of Animal Wastes in Pollution Control
and the Generation of Renewable Energy. Int. J. Environ. Res. Public Health 2013, 10, 4390–4417. [CrossRef]
[PubMed]
11. Franke-Whittle, I.H.; Walter, A.; Ebner, C.; Insam, H. Investigation into the effect of high concentrations of
volatile fatty acids in anaerobic digestion on methanogenic communities. Waste Manag. 2014, 34, 2080–2089.
[CrossRef] [PubMed]
12. Lin, J.; Zuo, J.; Gan, L.; Li, P.; Liu, F.; Wang, K.; Chen, L.; Gan, H. Effects of mixture ratio on anaerobic
co-digestion with fruit and vegetable waste and food waste of China. J. Environ. Sci. 2011, 23, 1403–1408.
[CrossRef]
13. Zhang, C.; Xiao, G.; Peng, L.; Su, H.; Tan, T. The anaerobic co-digestion of food waste and cattle manure.
14. Fitamo, T.; Boldrin, A.; Boe, K.; Angelidaki, I.; Scheutz, C. Co-digestion of food and garden waste with
mixed sludge from wastewater treatment in continuously stirred tank reactors. Bioresour. Technol. 2016, 206,
15. Mata-Alvarez, J.; Dosta, J.; Macé, S.; Astals, S. Codigestion of solid wastes: A review of its uses and
perspectives including modeling. Crit. Rev. Biotechnol. 2011, 31, 99–111. [CrossRef] [PubMed]
16. Mao, C.; Feng, Y.; Wang, X.; Ren, G. Review on research achievements of biogas from anaerobic digestion.
17. Esposito, G.; Frunzo, L.; Panico, A.; d’Antonio, G. Mathematical modelling of disintegration-limited
co-digestion of OFMSW and sewage sludge. Water Sci. Technol. 2008, 58, 1513–1519. [CrossRef] [PubMed]
18. Wang, X.; Lu, X.; Li, F.; Yang, G. Effects of temperature and carbon-nitrogen (C/N) ratio on the performance
of anaerobic co-digestion of dairy manure, chicken manure and rice straw: Focusing on ammonia inhibition.
PLoS ONE 2014, 9, e97265. [CrossRef] [PubMed]
19. Zhu, S.; Wu, Y.; Yu, Z.; Zhang, X.; Li, H.; Gao, M. The effect of microwave irradiation on enzymatic hydrolysis
of rice straw. Bioresour. Technol. 2006, 97, 1964–1968. [CrossRef] [PubMed]
20. Zabed, H.; Sahu, J.N.; Boyce, A.N.; Faruq, G. Fuel ethanol production from lignocellulosic biomass:
An overview on feedstocks and technological approaches. Renew. Sustain. Energy Rev. 2016, 66, 751–774.
[CrossRef]
21. Zheng, Y.; Zhao, J.; Xu, F.; Li, Y. Pretreatment of lignocellulosic biomass for enhanced biogas production.
Prog. Energy Combust. Sci. 2014, 42, 35–53. [CrossRef]
22. Ravindran, R.; Jaiswal, A.K. Microbial Enzyme Production Using Lignocellulosic Food Industry Wastes as
Feedstock: A Review. Bioengineering 2016, 3, 30. [CrossRef] [PubMed]
23. Bellasio, M.; Mattanovich, D.; Sauer, M.; Marx, H. Organic acids from lignocellulose: Candida lignohabitans
as a new microbial cell factory. J. Ind. Microbiol. Biotechnol. 2015, 42, 681–691. [CrossRef] [PubMed]
24. Kumar, M.; Revathi, K.; Khanna, S. Biodegradation of cellulosic and lignocellulosic waste by
Pseudoxanthomonas sp R-28. Carbohydr. Polym. 2015, 134 (Suppl. C), 761–766. [CrossRef] [PubMed]
25. The National Petroleum Council. Renewable Natural Gas for Transportation: An Overview of the Feedstock
Capacity, Economics, and GHG Emission Reduction Benefits of RNG as a Low-Carbon Fuel; The National Petroleum
Council: Washington, DC, USA, 2013; p. 44.
26. (EIA), U.S. Energy Information Administration. Monthly Energy Review: Energy Consumption by Sector;
U.S. Energy Information Administration: Washington, DC, USA, 2013.
179
Energies 2018, 11, 2058
27. Valentine, J.; Clifton-Brown, J.; Hastings, A.; Robson, P.; Allison, G.; Smith, P. Food vs. fuel: The use of land
for lignocellulosic ‘next generation’ energy crops that minimize competition with primary food production.
GCB Bioenergy 2012, 4, 1–19. [CrossRef]
28. Awais, M.; Alvarado-Morales, M.; Tsapekos, P.; Gulfraz, M.; Angelidaki, I. Methane Production and Kinetic
Modeling for Co-digestion of Manure with Lignocellulosic Residues. Energy Fuels 2016, 30, 10516–10523.
[CrossRef]
29. Kabir, M.M.; del Pilar Castillo, M.; Taherzadeh, M.J.; Sárvári Horváth, I. Effect of the
N-Methylmorpholine-N-Oxide (NMMO) Pretreatment on Anaerobic Digestion of Forest Residues.
BioResources 2013, 8, 5409–5423. [CrossRef]
30. Achinas, S.; Achinas, V.; Euverink, G.J.W. A Technological Overview of Biogas Production from Biowaste.
Engineering 2017, 3, 299–307. [CrossRef]
31. Brown, D.; Li, Y. Solid state anaerobic co-digestion of yard waste and food waste for biogas production.
32. Xu, F.; Li, Y. Solid-state co-digestion of expired dog food and corn stover for methane production.
33. Jabeen, M.; Zeshan; Yousaf, S.; Haider, M.R.; Malik, R.N. High-solids anaerobic co-digestion of food waste
and rice husk at different organic loading rates. Int. Biodeterior. Biodegr. 2015, 102, 149–153. [CrossRef]
34. Mulat, D.G.; Huerta, S.G.; Kalyani, D.; Horn, S.J. Enhancing methane production from lignocellulosic
biomass by combined steam-explosion pretreatment and bioaugmentation with cellulolytic bacterium
Caldicellulosiruptor bescii. Biotechnol. Biofuels 2018, 11, 19. [CrossRef] [PubMed]
35. Gerardi, M.H. The Microbiology of Anaerobic Digesters; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2003;
p. 177.
36. Lloret, E.; Pastor, L.; Pradas, P.; Pascual, J.A. Semi full-scale thermophilic anaerobic digestion (TAnD) for
advanced treatment of sewage sludge: Stabilization process and pathogen reduction. Chem. Eng. J. 2013, 232,
42–50. [CrossRef]
37. Haider, M.R.; Zeshan; Yousaf, S.; Malik, R.N.; Visvanathan, C. Effect of mixing ratio of food waste and
rice husk co-digestion and substrate to inoculum ratio on biogas production. Bioresour. Technol. 2015, 190,
38. Li, Y.; Zhang, R.; Liu, X.; Chen, C.; Xiao, X.; Feng, L.; He, Y.; Liu, G. Evaluating Methane Production from
Anaerobic Mono- and Co-digestion of Kitchen Waste, Corn Stover, and Chicken Manure. Energy Fuels 2013,
27, 2085–2091. [CrossRef]
39. Li, L.; Yang, X.; Li, X.; Zheng, M.; Chen, J.; Zhang, Z. The Influence of Inoculum Sources on Anaerobic
Biogasification of NaOH-treated Corn Stover. Energy Sources Part A 2010, 33, 138–144. [CrossRef]
40. Pommier, S.; Llamas, A.M.; Lefebvre, X. Analysis of the outcome of shredding pretreatment on the anaerobic
biodegradability of paper and cardboard materials. Bioresour. Technol. 2010, 101, 463–468. [CrossRef]
[PubMed]
41. Dhiman, S.S.; Shrestha, N.; David, A.; Basotra, N.; Johnson, G.R.; Chadha, B.S.; Gadhamshetty, V.; Sani, R.K.
Producing methane, methanol and electricity from organic waste of fermentation reaction using novel
microbes. Bioresour. Technol. 2018, 258, 270–278. [CrossRef] [PubMed]
42. Rice, E.W.; Bridgewater, L. Standard Methods for the Examination of Water and Wastewater; American Public
Health Association: Washington, DC, USA, 2012.
43. Sluiter, A.; Hames, B.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Templeton, D.; Crocker, D. Determination of Structural
Carbohydrates and Lignin in Biomass; National Renewable Energy Laboratory: Golden, CO, USA, 2008.
44. Kaparaju, P.; Serrano, M.; Thomsen, A.B.; Kongjan, P.; Angelidaki, I. Bioethanol, biohydrogen and biogas
production from wheat straw in a biorefinery concept. Bioresour. Technol. 2009, 100, 2562–2568. [CrossRef]
[PubMed]
45. Dhar, H.; Kumar, P.; Kumar, S.; Mukherjee, S.; Vaidya, A.N. Effect of organic loading rate during anaerobic
digestion of municipal solid waste. Bioresour. Technol. 2016, 217, 56–61. [CrossRef] [PubMed]
46. Bibra, M.; Kumar, S.; Wang, J.; Bhalla, A.; Salem, D.R.; Sani, R.K. Single pot bioconversion of prairie cordgrass
into biohydrogen by thermophiles. Bioresour. Technol. 2018, 266, 232–241. [CrossRef] [PubMed]
47. Zhang, L.; Lee, Y.-W.; Jahng, D. Anaerobic co-digestion of food waste and piggery wastewater: Focusing on
the role of trace elements. Bioresour. Technol. 2011, 102, 5048–5059. [CrossRef] [PubMed]
180
Energies 2018, 11, 2058
48. Tanimu, I.; Mohd Ghazi, T.; Razif Harun, M.; Idris, A. Effect of Carbon to Nitrogen Ratio of Food Waste
on Biogas Methane Production in a Batch Mesophilic Anaerobic Digester. Int. J. Innovation Technol. 2014, 5,
116–119.
49. Zhai, N.; Zhang, T.; Yin, D.; Yang, G.; Wang, X.; Ren, G.; Feng, Y. Effect of initial pH on anaerobic co-digestion
of kitchen waste and cow manure. Waste Manag. 2015, 38, 126–131. [CrossRef] [PubMed]
50. Zhang, T.; Mao, C.; Zhai, N.; Wang, X.; Yang, G. Influence of initial pH on thermophilic anaerobic co-digestion
of swine manure and maize stalk. Waste Manag. 2015, 35, 119–126. [CrossRef] [PubMed]
51. Zainab Ziad Ismail, A.R.T. Assessment of anaerobic co-digestion of agro wastes for biogas recovery: A bench
scale application to date palm wastes. Int. J. Energy Environ. 2014, 5, 591–600.
52. Tong, X.; Smith, L.H.; McCarty, P.L. Methane fermentation of selected lignocellulosic materials. Biomass 1990,
21, 239–255. [CrossRef]
53. Ge, X.; Xu, F.; Li, Y. Solid-state anaerobic digestion of lignocellulosic biomass: Recent progress and
perspectives. Bioresour. Technol. 2016, 205, 239–249. [CrossRef] [PubMed]
54. Shiralipour, A.; Smith, P.H. Conversion of biomass into methane gas. Biomass 1984, 6, 85–92. [CrossRef]
55. Schmidt, T.; Pröter, J.; Scholwin, F.; Nelles, M. Anaerobic digestion of grain stillage at high organic loading
rates in three different reactor systems. Biomass Bioenergy 2013, 55, 285–290. [CrossRef]
56. Ferguson, R.M.W.; Coulon, F.; Villa, R. Organic loading rate: A promising microbial management tool in
anaerobic digestion. Water Res. 2016, 100, 348–356. [CrossRef] [PubMed]
57. Cantrell, K.B.; Ducey, T.; Ro, K.S.; Hunt, P.G. Livestock waste-to-bioenergy generation opportunities.
58. Esposito, G.; Frunzo, L.; Giordano, A.; Liotta, F.; Panico, A.; Pirozzi, F. Anaerobic co-digestion of organic
wastes. Rev. Env. Sci. Biotechnol. 2012, 11, 325–341. [CrossRef]
59. Ahring, B.K. Methanogenesis in thermophilic biogas reactors. Antonie Leeuwenhoek 1995, 67, 91–102.
[CrossRef] [PubMed]
60. Labatut, R.A.; Gooch, C.A. Monitoring of Anaerobic Digestion Process to Optimize Performance and Prevent System
Failure; Department of Biological and Environmental Engineering, Cornell University: Ithaca, NY, USA, 2014.
181
energies
Article
Enhancing Biochemical Methane Potential and
Enrichment of Specific Electroactive Communities
from Nixtamalization Wastewater using Granular
Activated Carbon as a Conductive Material
David Valero 1 , Carlos Rico 2 , Blondy Canto-Canché 3 , Jorge Arturo Domínguez-Maldonado 1 ,
Raul Tapia-Tussell 1 , Alberto Cortes-Velazquez 4 and Liliana Alzate-Gaviria 1, *
1 Renewable Energy Unit, Yucatan Center for Scientist Research, Mérida CP 97302, Mexico;
[email protected] (D.V.); [email protected] (J.A.D.-M.); [email protected] (R.T.-T.)
2 Department of Water and Environmental Science and Technologies, University of Cantabria, 39005
Santander, Spain; [email protected]
3 Biotechnology Unit, Yucatan Center for Scientist Research, Mérida CP 97205, Mexico; [email protected]
4 GeMBio Laboratory, Yucatan Center for Scientist Research, Mérida CP 97205, Mexico; [email protected]
Abstract: Nejayote (corn step liquor) production in Mexico is approximately 1.4 × 1010 m3 per
year and anaerobic digestion is an effective process to transform this waste into green energy.
The biochemical methane potential (BMP) test is one of the most important tests for evaluating
the biodegradability and methane production capacity of any organic waste. Previous research
confirms that the addition of conductive materials significantly enhances the methane production
yield. This study concludes that the addition of granular activated carbon (GAC) increases methane
yield by 34% in the first instance. Furthermore, results show that methane production is increased
by 54% when a GAC biofilm is developed 10 days before undertaking the BMP test. In addition,
the electroactive population was 30% higher when attached to the GAC than in control reactors.
Moreover, results show that electroactive communities attached to the GAC increased by 38% when
a GAC biofilm is developed 10 days before undertaking the BMP test, additionally only in these
reactors Geobacter was identified. GAC has two main effects in anaerobic digestion; it promotes direct
interspecies electron transfer (DIET) by developing an electro-active biofilm and simultaneously it
reduces redox potential from −223 mV to −470 mV. These results suggest that the addition of GAC
to biodigesters, improves the anaerobic digestion performance in industrial processed food waste.
Keywords: biochemical methane potential; redox potential reduction; direct interspecies electron
transfer; electroactive biofilm; Nejayote; granular activated carbon
1. Introduction
The Mexican corn tortilla industry produces between 2.2–3.5 m3 ·t−1 of wastewater from processed
corn [1,2]. The wastewater (Nejayote) presents an environmental problem for Mexican society due
to its malodorous nature and its composition. It is imperative to have an explicit waste management
strategy to deal with the unpleasant by-products, and it is also essential to ascertain the potential
uses of the wastewater in order to harness and capitalise on the large quantities that are produced
through nixtamalization. Nixtamalization is the ancient Aztec process of cooking corn with lime to
produce corn masa [3]. Water, corn, and lime are cooked at 80 ◦ C during three hours and is then
steeped for 15 hours [4]. Nowadays, the process was carried out on an industrial scale but based on
the process carried out by the Aztecs known as nixtamalization. This complex waste has a pH in the

Energies 2018, 11, 2101
9–14 range, a chemical oxygen demand (COD) between 3.5 and 40 g·L−1 and a phenol concentration
of 4.2 g·L−1 [3–6]. This wide range is a result of the proportions of water corn and lime that vary
in different nixtamal plants [4,5,7]. Its organic matter content makes nejayote a suitable waste for
anaerobic digestion process rather than any aerobic process due to the huge amount of energy that is
required to degrade its high organic load [8–10]. A biochemical methane potential (BMP) test is the
most efficient means of identifying the methane production capacity, in anaerobic digestion, for any
organic waste [11–13]. Additionally, it can be useful to identify future inhibition and adaptation
problems before scaling the process [14]. Given the lack of previous studies in relation to BMP testing
of nejayote, this article is the first step in determining the biochemical methane potential via direct
interspecies electro transfer (DIET) of nejayote in anaerobic digestion and it provides an opportunity
to develop a perspective on its possible performance on a larger scale.
Anaerobic digestion process is based in four main steps: hydrolysis, acidogenesis, acetogenesis,
and methanogenesis. In methanogenesis, step electrons are transported from acetoclastic bacteria
to hydrogenotrophic archaea to reduce carbon dioxide to methane. This transport can be done by
shuttle moleculas as hydrogen or formate. Traditionally, interspecies hydrogen transfer (IHT) and
interspecies formate transfer (IFT) have been intrinsically linked with interspecies electron transfer
(IET). However, these can present problems in relation to H2 partial pressure and formate concentration,
respectively [15]. More recently, there has been a greater production of studies referring to DIET, which
has become more widely accepted. DIET is another form of IET that can be produced through pili or
employing conductive materials [16–19]. Exoelectrogenic bacteria and electrotrophic methanogens
are the partners that participate in DIET [20–22]. Short chain volatile fatty acids (VFA) and alcohols
are the biodegradable compounds in the syntrophic association between exoelectrogenic bacteria
and electrotrophic methanogens [23,24]. For the first time in 2014, the DIET between exoelectrogenic
bacteria and electrotrophic methanogens was evidenced through pili in a Geobacter-Methanosaeta
co-culture [18]. Not all exoelectrogenic bacteria are able to generate pili; however, the addition of
conductive materials increases the genera of bacteria that are able to supply pili the genera of bacteria
that are able to participate in DIET. In cases where the bacteria are not able to generate pili, conductive
materials fulfil the role of pili in terms of facilitating electron exchange [23,25,26]. Granular activated
carbon has shown high performance in promoting DIET in different studies [26,27]. Moreover, its low
cost [28,29] and its high conductivity (3000 μS·cm−1 ) [21] when compared to other conductive materials
such as biochar (4 μS·cm−1 ) [30], magnetite (160 μS·cm−1 ) [31], or stainless steel (667 μS·cm−1 ) [32],
make it a preferential material to promote DIET in order to improve methane production yields [17,33].
Additionally, metallic materials, such as stainless steel, have corrosion problems [34].
Within the past three years, through the microorganism identification of biofilms
developed in conductive materials, new exoelectrogenic bacteria have been discovered, such
as Thauera, Sporanaerobacter, Enterococcus, Pseudomonas, Anaerolinaceae, Bacteroides, Streptococcus,
Syntrophomonas, Sulfurospirillum, Caloramator, Tepidoanaerobacter, Coprothermobacter, Clostridium senso
stricto, Peptococcaceae, and Bacillaceae, also new electrotrophic methanogen genera have been found
able to participate in DIET, such as Methanosarcina, Methanobacterium, Methanolinea, Methanothrix,
Methanoregula, and Methanospirillum [17,27,35–44]. However, the number of bacteria that are capable
of participating in DIET through conductive materials is still unknown [45].
Industrial processed food is one of the most important industries that employ anaerobic digestion
to degrade water pollution due to its high COD [10]. Most previous studies have used simple substrates
such as ethanol or short chain VFA, or commercial waste, such as dog food [17,23,27]. At present, there
are no studies that use a substrate of industrial food waste using conductive materials to promote DIET.
This study is the first step towards developing an understanding of the behaviour and growth of
electroactive communities in a real wastewater, and thus providing an indication of the improvements
in methane production and VFA and COD degradation, which can be made by implementing DIET
through conductive materials on a large scale.
183
Energies 2018, 11, 2101
2.1. Nejayote and Inoculum

Nejayote was collected from an industrial corn flour plant in Yucatan State (Mexico). Nejayote
was preserved in a cold box at 4 ◦ C from the corn flour plant to the laboratory, which is in accordance
with the Standard Test Method 1060 for the collection and preservation of samples [46]. Prior to
undertaking the BMP test, the nejayote was characterized according to the following parameters:
pH, total suspended solids (TSS), total solids (TS) and volatile solids (VS), COD, total nitrogen (TN),
phosphates (PO4 −3 ), ammonia nitrogen (NH3 N), sulfate (SO4 −2 ), and total alkalinity (TA).
The inoculum composition employed was made according to Poggi Varaldo et al. [47], while using
30 g·L−1 of deep soil, 300 g·L−1 of cattle manure, 150 g·L−1 of pig manure, 1.5 g·L−1 of commercial
Na2 CO3 , and 1 L of tap water. The inoculum was characterized for TA, TS, and VS.
2.2. BMP Test

The trial was carried out in triplicate in 110 mL serum bottles that were capped with rubber
septum sleeve stoppers. Three experimental conditions were assayed. In controls, the reactors (N)
were filled with 60 g of inoculum and nejayote at a ratio of 2 based on VS [12]. In three reactors (N0) 3 g
of GAC was added [33] at the beginning of the BMP test. In the other three reactors (N10), inoculum
and 3 g of GAC were added to serum bottles ten days before undertaking the BMP test. After this
time, nejayote was introduced in anoxic conditions. All of the tests were carried out for 30 days at
38 ◦ C with an automatic agitation of 100 rpm [48]. All reactors were flushed with nitrogen in an anoxic
chamber to avoid the presence of oxygen. Three blanks with 60 g of inoculum were tested to measure
the inoculum methane potential. The methane production measurements are expressed at 0 ◦ C and
standard pressure of 760 mmHg (NCTP) in dry conditions.
Gas production measurement and numerical calculation was according to Valero et al. [49].
Headspace pressure was measured employing a digital pressure transducer (ifm Germany type PN78
up to 2 bars) with a syringe that was connected to pierce the septum. Statistical analysis was performed
with the Analysis ToolPak in Microsoft Excel (Excel 2016 (v16.0)).
Once the BMP test was finished, COD, VFA, and redox potential were measured for all of
the assays.
2.3. GAC Conductivity

GAC was introduced in PVC tubes, of specific lengths and known surface areas, to calculate its
conductivity. A source meter (Keithley 2400, Cleveland, OH, USA) was used to measure the resistance
of GAC. Once the resistance was determined, conductivity was calculated employing the following
equations:
A
R = ρ (1)
L
1
K = (2)
ρ
where R is the resistance measured with Keithley 2400, A is the area of PVC tube, L is the tube length,
ρ is the GAC resistivity, and K is the GAC conductivity. The process was undertaken in triplicate and
the result was 3690 μS·cm−1 [50].

Biogas characterization was determined on a Molesieve column (30 m long, 0.53 mm internal
diameter, and 0.25 μm film thickness) in a gas chromatograph (Clarus 500-Perkin Elmer, Waltham, MA,
USA) with the thermal conductivity detector (TCD). Nitrogen was used as gas carrier and temperatures
of 75, 30 and 200 ◦ C were used for the injector, oven and detector, respectively. VFA were determined
184
Energies 2018, 11, 2101
by gas chromatography using (Clarus 500-Perkin Elmer, Waltham, MA, USA). The column employed
was Agilent J&W (30 m long, 0.53 mm internal diameter) with a flame ionization detector (FID). Before
determinating VFA by chromatography, samples were filtered, acidified with phosphoric acid, and
then centrifuged. Conductivity, temperature, pH, TSS, TS, and VS were analyzed, following standard
methods [46]. Colorimetric methods (Hach Company DR-890, Loveland, CO, USA) were used to
determine COD, TN, PO4 −3 , NH3 -N, and SO4 −2 . Redox potential was determined employing a redox
potential sensor (Extech RE300, Nashua, NH, USA).
2.5. Microbial Community Analysis

After the BMP test, sludge from N, N0, and N10 and GAC from N0 and N10 was kept at
−80 ◦ C [27]. DNA was extracted from the five samples N sludge (S), N0 sludge (S0), N10 sludge (S10),
GAC biofilm from N0 (C0), and GAC biofilm from N10 (C10). Three DNA extraction of each sample
were undertook, and then the DNA from each of samples were mixed before perform the sequencing.
DNA from each sample was isolated with a DNA extraction kit ZymoBIOMICSTM DNA miniprep Kit,
(Zymo Research, Irvine, CA, USA), in accordance with manufacturer´s protocol. The DNA pellet was
resuspended in 50 μL of TE buffer (10mM Tris-HCl (pH 8) and 1mM EDTA (pH 8)). DNA purity and
DNA concentration were checked by employing a NanoDrop ND-1000 spectrophotometer (NanoDrop
Technologies, Seattle, WA, USA). The A260/280 ratios were over 1.6 and DNA concentration in the
range of 50–100 ng·μL−1 . The primers employed for amplifying 16S rRNA genes (bacteria and archaea)
were 341F and 805R. DNA was sent to Macrogen Inc. (Seoul, Korea) who performed sequencing on an
Illumina MiSeq platform. MG-RAST software was used to analyze bacteria and archaeal communities
through GREENGENES and RDP II databases [51].
2.6. Scanning Electron Microscopy (SEM)

GAC was analyzed by SEM (SEM, model JSM-6360LV, JEOL, Tokyo, Japan). In order to establish
the existence of microbial communities attached to GAC, three samples of N10 were mounted on
a metallic stub using double-sided adhesive tape, each being coated with a 15 nm gold layer and
observed at 20 kV.
3. Results

Nejayote collected from industrial corn flour plant showed basic pH values of 10.2. The inoculum,
developed in the laboratory, had a basic pH but with values that were lower than nejayote, in this case,
the pH reached 8.1. The VS percentage obtained for nejayote was 0.85 ± 0.01%, whereas inoculum
values reached 2.69 ± 0.01%. TS results were 1.22 ± 0.01% and 4.62 ± 0.01% for nejayote and inoculum,
respectively. Nejayote and inoculum VS/TS ratios that were obtained were 0.69 and 0.55, respectively.
TA concentration results were similar for nejayote and inoculum, reaching values of 1799 ± 116 and
1652 ± 147 mg CaCO3 L−1 , respectively. COD concentration determined for nejayote was 15,433 ±
827 mg·L−1 . TSS levels for nejayote of 2676 ± 512 mg·L−1 were recorded. Other parameters that were
characterized for nejayote were NH3 -N 4.6 ± 0.1 mg·L−1 , TN 95 ±4 mg·L−1 , SO4 −2 22 ± 2 mg·L−1 ,
and PO4 −3 59 ± 1.2 mg·L−1 .
3.2. BMP Test

Methane production did not peak for the three assays N, N0, and N10 until the third day of BMP
testing. The rate at which methane was produced spiked during the third day and it resulted in 39,
46, and 45 L CH4 kg−1 VS day−1 for N, N0, and N10, respectively. As shown in Figure 1A, the first
significant difference between N10 reactors and N and N0 reactors, was on the fifth day (t-student
p < 0.05). N10 reactors resulted in higher methane production levels of 26 L CH4 kg−1 VS day−1 , whilst
N and N0 reactors produced 17 and 10 L CH4 kg−1 VS day−1 , respectively.
185
Energies 2018, 11, 2101
Figure 1. Biochemical Methane Potential (BMP) curves for control reactor (N), reactors with granular
activated carbon (GAC) (N0) and reactors with biofilm GAC developed before undertaking the BMP
test (N10). (A) Daily methane production (B) Cumulative methane production (C) methane biogas
percentage. Mean values ± SD from triplicate assays.
DIET or IHT occurs in the last step of the anaerobic digestion process, however adding GAC
promotes DIET instead of IHT in N0 and N10 reactors [45]. The only compounds that can be oxidized
by exoelectrogenic bacteria are alcohols and VFA [24,52]. When these compounds are available to
stimulate the production of methane through DIET, a significant difference can be seen between DIET
reactors through GAC and those reactors without conductive materials (t-student analyses p < 0.05).
As shown in Figure 1A, from the tenth day until the fifteenth day of experimentation, N0 and N10
186
Energies 2018, 11, 2101
consistently produced methane over 10 L CH4 kg−1 VS day−1 , while N methane production was
under 10 L CH4 kg−1 VS day−1 . Moreover, on one occasion during the same period, N0 and N10
actually reached more than 22 L CH4 kg−1 VS day−1 , while N did not produce values above 10 L CH4
kg−1 VS day−1 .
As highlighted in Figure 1B, the cumulative methane production results draws focus to the
significant differences between N, N0, and N10 (t-student analysis p < 0.05). The final methane
generation for N, N0 and N10 reached values of 222 ± 23 L CH4 kg−1 VS, 297 ± 10 L CH4 kg−1 VS and
342 ± 29 L CH4 kg−1 VS, respectively. N0 methane production was 34% higher than N, however the
difference between N10 and N soared staggeringly to register an increase of 54%. Additionally, results
show that the N10 methane volume generated was 15% higher than N0.
Methane production was predicted relative to biochemical oxygen demand (BOD) concentration.
Nejayote BOD is around 5870 ± 1900 mg L−1 [3,5,53,54]. Theoretically, 1 kg of BOD produces 350 L
methane [55]. Thus, the methane production from nejayote expected through anaerobic digestion
process without DIET promotion was 73.96 mL CH4 (C6 H12 O6 (DBO) → 3 CO2 + 3 CH4 ) [55]. In this
study, real methane production in N reactors was 69.12 ± 8.96 mL CH4 . Likewise, through the DIET
process an increase between 20–33% in the production of methane has been reported due to a greater
recovery of electrons in the reduction of CO2 to CH4 through VFA (acetate, propionate, and butyrate)
and alcohol (ethanol) degradation [18,25,38,56–58]. According to these increments, the methane
production promoting DIET expected was in the range of 88.75–98.36 mL CH4 . In this study, real
methane production for N0 and N10 was 92.71 ± 5.38 and 106.69 ± 8.15 mL CH4 , respectively. These
values are in the expected range.
The first difference in the biogas composition came after the sixth day, when N10 methane
percentage was recorded as 5% higher than N and N0, as illustrated in Figure 1C. This difference
remained constant between N10 and N0 throughout the BMP test.
Between days eleven and eighteen, N0 methane composition was 5% higher than N. Thus, N10
was 10% higher than N. The maximum methane concentration was reached on the fifteenth day with
values of 64%, 71%, and 77% for N, N0, and N10, respectively. After the eighteenth day, there was
no further biogas production in all of the reactors, and therefore the biogas composition plateaued to
show similar levels of methane concentration in all reactors.
Once the BMP test was completed, the contents of the different reactors were analyzed to
determine pH, redox potential, VFA, and COD, as shown in Table 1. There were no significant
differences in pH values between the three assays. A huge reduction of redox potential was produced
in reactors with GAC, with the measured value for N being recorded as −222 ± 7 mV, when compared
with −466 ± 1 mV and −471 ± 2 mV for N0 and N10, respectively. VFA composition was formed by
acetic acid 203 mg·L−1 and butyric acid 90 mg·L−1 in N reactors, while in N0 and N10 assays the VFA
composition was made up solely of acetic acid because butyric acid was only present in N reactors.
COD concentration was higher in N reactors than in N0 and N10, with values of 532 ± 10 mg·L−1 ,
307 ± 21 mg·L−1 , and 218 ± 14 mg·L−1 , respectively.
Table 1. Reactors analysis after BMP test.
Parameter N N0 N10
pH 7.3 ± 0.11 7.2 ± 0.03 7.3 ± 0.02
Redox potential (mV) −223 ± 5 −467 ± 1 −471 ± 2
Acetic acid (mg·L−1 ) 203 ± 12 139 ± 9 109 ± 8
Butyric acid (mg·L−1 ) 91 ± 5
COD (mg·L−1 ) 532 ± 10 307 ± 21 218 ± 14
Control reactor (N), Reactor with granular activated carbon (N0), reactor with GAC added 10 days before
undertaking biochemical methane potential test.
187
Energies 2018, 11, 2101
3.3. Microbial Community Analysis

As shown in Figure 2, the four main bacteria phylum in all of the samples were Actinobacteria,
Bacteroidetes, Firmicutes, and Proteobacteria. Firmicutes was the dominant phylum in S, comprising
almost 50% of the total bacteria. In sludges S0 and S10, there was a lower percentage of Firmicutes,
forming 26% and 25% of the total bacteria, respectively. Furthermore, the percentage of Firmicutes was
even lower in C0 and C10, measuring 15% and 16%, respectively. In contrast, Bacteroidetes follows the
opposite trend to Firmicutes. Bacteroidetes had its highest percentage in C0 and C10, attached to the
carbon, reaching 48% in C0 and 43% in C10, while in S it accounted for 18% of the microbial community.
The Proteobacteria phylum presented a difference between the samples: it was detected at 17% in the
C10 biofilm; whilst, for the other samples, it never reached 10%.
Figure 2. Bacteria phylum relative abundance. Phylum level with relative abundance lower than 1%
were included in unclassified group. Sludge from control reactor (S), sludge from reactor with granular
activated carbon (GAC) (S0), sludge from reactor with GAC biofilm developed before undertaking the
BMP test (S10), GAC biofilm from reactor with GAC (C0), and GAC biofilm from reactor with GAC
biofilm developed before undertaking the BMP test (C10).
Figure 3A illustrates the classification of bacteria by genus proportion. Clostridium anaerobic

fermentative bacteria was highly developed in S, comprising 46% of the bacterial community, whereas
it was less than 21% in S0 and S10, and interestingly, it accounted for 7% of the bacterial community
attached to GAC in C0 and C10. Parabacteroides were developed in GAC biofilm and the sludge
of GAC reactors with a relative abundance of 8%, 13%, 12%, and 14% for S0, S10, C0, and C10,
respectively; however, there was no presence detected in S. Parabacteroides that is is able to produce
VFA as an end product of fermentation. This, in turn, can be oxidized by exoelectrogenic bacteria.
Additionally, an enrichment of this genus in microbial fuel cells [31] and microbial electrolysis cells [32]
has been observed.
Geobacter, an exoelectrogenic bacteria widely accepted as an exoelectrogen DIET partner [7,8],
was only detected in the C10 sample, where the biofilm was developed 10 days before undertaking
the BMP test. The amount of Geobacter in C10, whilst not dominant, remains a substantive 14% of the
overall bacteria community and 82% of Proteobacteria phylum. The most dominant genus on the surface
of both GAC samples C0 and C10 was Prolixibacter, with 31% and 21%, respectively. Prolixibacter had a
relative abundance of 16%, 18%, and 11% in S, S0, and S10, respectively. It must be mentioned that
Prolixibacter has been reported as a possible exoelectrogenic bacteria in microbial electrolysis cells [33]
and in marine-sediment fuel cells [34].
188
Energies 2018, 11, 2101
Figure 3. Bacteria (A) and Archaea community structure (B) at genus level. Genus level with relative
abundance lower than 1% were included in unclassified groups. Sludge from control reactor (S),
sludge from reactor with granular activated carbon (GAC) (S0), sludge from reactor with GAC biofilm
developed before undertaking the BMP test (S10), GAC biofilm from reactor with GAC (C0) and GAC
biofilm from reactor with GAC biofilm developed before undertaking the BMP test (C10).
A low enrichment for Syntrophomonas and Bacteroides was detected in the microbial community
analysis. Both showed a relative abundance of 3% in C0 and C10, but contrastingly they were not
detected in S. Both genera have only been reported as exoelectrogen DIET partners in studies with
complex waste [19], as similar to nejayote.
In this study, the total amount of electroactive bacteria (Geobacter, Proxilibacter, Bacteroides, and
Syntrophomonas) was 41%, 37%, 17%, 22%, and 16% for C10, C0, S10, S0, and S, respectively. This
indicates an exoelectrogenic bacteria enrichment of 24% in C10 when compared with the control
reactor, and an 18% higher electroactive bacteria community in C0 than S.
The archaeal community was dominated by Methanosaeta in all of the reactors, attached to the
carbons and suspended in all sludges. It was most prevalent in C10, reaching 89% of the archaeal
community, while in S, its percentage was 78%. Methanosaeta has been established as one of the
main methanogens able to accept electrons in DIET [7]. Methanolinea, another methanogen associated
with electrotrophic capacity in DIET [35], only appeared on the carbon surface of C10 with a relative
abundance of less than 5% of the archaeal community.
The similiarities of the microbial community structures in the five samples were analyzed at the
genus level by hierarchically clustered heat analysis (Figure 4). The heatmap includes 17 genera. C10
is the sample with the highest concentration of different genera. The communities attached to GAC,
C10, and C0 clustered together finally. These samples share a high similitude when compared with
sludge samples (S, S0, S10).
189
Energies 2018, 11, 2101
Figure 4. Heatmap of bacteria at genus level. Genus level with relative abundance lower than 1% were
included in unclassified groups. Sludge from control reactor (S), sludge from reactor with granular
activated carbon (GAC) (S0), sludge from reactor with GAC biofilm developed before undertaking
the BMP test (S10), GAC biofilm from reactor with GAC (C0), and GAC biofilm from reactor with
GAC biofilm developed before undertaking the BMP test (C10). Heatmap showing the 17 genera with
significant difference of relative abundances among five samples. Heatmap is color-coded based scale
(from red (−1) less abundance to green (+1) more abundance).
3.4. Scanning Electron Microscopy (SEM)

Through SEM, it was possible to examine the carbon surface and its porosity free of microbial
communities before adding GAC to the reactors (Figure 5A). Additionally, microbial communities
attached to GAC are clearly visible in Figure 5B. The exoelectrogen partner Geobacter was identified in
the microbiological analysis attached to the GAC in sample C10 with a relative abundance of 14% in
the bacterial community. Geobacter has been described as rod-shaped bacteria with a length between
2–4 μm [59]. In Figure 5C, bacteria can be visualized with the same morphological characteristics
described above. This, coupled with the high percentage of Geobacter attached to GAC, makes
it plausible to identify them in the Figure 5C. Methanosaeta is a rod-shaped archaea with average
dimensions between 2 and 6 μm in length by single cell. Cells are enclosed inside an annular, striated
sheathed structure, and they are separated by partitions forming a filamentous structure. Methanosaeta
filaments can reach 100 μm of length [60]. In Figure 5D, it is possible to see Methanosaeta communities.
Besides Methanosaeta was the dominant archaea in the microbial communities analysis, making its
identification easy by SEM.
190
Energies 2018, 11, 2101
Figure 5. Scanning electron micrographs of the granular activated carbon surface (A), biofilm attached
granular activated carbon (B), Geobacter communities attached to granular activated carbon (C), and
Methanosaeta communities attached to granular activated carbon (D) after biochemical methane
potential test (30 days).
4. Discussion

Table 2 shows the results of the physicochemical characterization of nejayote, as compared with
the results that were obtained in other studies. As it is a wastewater that may come from an industrial
process or from a more traditional process, which is not always done in the same way, the nejayote
presents a high level of heterogeneity in the results of its composition, as seen in the wide range
that is given for the majority of parameters. According to Ibarra-Mendívil et al. [61], different lime
concentrations of 0%, 0.5%, and 1% were used for nixtamalization and pH values of 4.75, 7.72, and
11.01 were obtained, respectively. Small modifications in proportions of lime, cooking, and rest times,
and different varieties of corn can give rise to notable variations in the composition of nejayote.
The presence of lime in the process leads to the nejayote having a basic pH. This study is in the range
expected with a value recorded of 10.2.
191
Energies 2018, 11, 2101
Table 2. Nejayote characterization compared with other studies.
Parameter Results Other Studies

pH 10.2 6.3–11.6 [1,3–5,62,63]
COD (mg·L−1 ) 15,433 ± 826 3430–40,058 [1,4,5,63]
NH3 -N (mg·L−1 ) 4.65 ± 0.05 2 [1]
TN (mg·L−1 ) 95.33 ± 3.7 209–428 [1,5]
SO4 −2 (mg·L−1 ) 22.5 ± 2.5 13 [1]
PO4 −3 (mg·L−1 ) 58.75 ± 1.2 7.6–1321 [1,5]
TA (mgCaCO3 ·L−1 ) 1799 ± 116 5768 [4]
TSS (mg·L−1 ) 2676 ± 512 1810–8340 [3,4]
TS% 1.22 ± 0.01 0.34–2.5 [1,4,63]
VS% 0.84 ± 0.01 0.24–1.55 [1,63]
VS/TS 0.69 ± 0.01 0.70 [63]
The COD concentration of nejayote is within the range observed in the different studies (Table 2),
with a value of 15,433 ± 826 mg·L−1 , coming mainly from the pericarp tissue of corn. The content
in TN was 95.33 ± 3.7 mg·L−1 , lower than the composition analyzed in other studies, although this
concentration is not sufficient to cause inhibition in anaerobic digestion. Concentrations between 1.7
and 14 g·L−1 cause a 50% reduction in methane production [64]. Both nitrogen and phosphorus are
necessary for the good performance of an anaerobic reactor, since they are indispensable nutrients
for bacteria. Concentrations lower than 0.3 mg·L−1 of phosphorus prevent the formation of the
microbial community [65]. The nejayote that is characterized in this study has a concentration of 58.75
± 1.25 mg·L−1 , which will aid a good performance in the BMP test. The concentration of SO4 is 22.5 ±
2.5 mg·L−1 , which is a little higher than the previously reported 13 mg·L−1 (Table 2). The COD/SO4
ratio is 686, so the presence of sulfur-reducing products and possible inhibition effects in the anaerobic
digestion process due to the byproducts generated by them will be non-existent [66]. The percentages
in both VS and TS are within the range referenced in Table 2. The values are 0.85 ± 0.01% VS and
1.22 ± 0.01% TS.
The inoculum has an alkalinity of 1652.5 ± 147 mg CaCO3 ·L−1 . This contribution of alkalinity to
the medium will absorb possible acidification processes that occur during anaerobic digestion due to
the formation of VFAs [67]. The VS percentage of the inoculum is 2.67% ± 0.1. This percentage aids
the good performance in the anaerobic digestion process [68].
4.2. BMP Test

The results of this study clearly demonstrate that generating a previous GAC biofilm before
starting the wastewater anaerobic digestion enhances methane biogas production and results
in an overall improved anaerobic digestion performance. The syntrophic relationship between
exoelectrogenic bacteria and electrotrophic archaea is essential to enable DIET. Recently, it has been
widely recognised that exolectrogenic bacteria can only oxidize alcohols and VFA to send electrons to
reduce carbon dioxide to methane within DIET [24,52]. The advantages of promoting DIET instead of
IHT in methane production yield do not become obviously apparent until there are increased alcohols
and VFA concentrations and they are available to be oxidized by exoelectrogenic bacteria. While
examining the results of the study, it is important to state that the first difference between control
reactors and GAC reactors in methane production did not appear until the seventh day in N0 and
N10 (Figure 1A), due to the insufficient concentration of VFA and alcohols, which are available in the
medium to be oxidized. In previous studies undertaken in batch assays with conductive materials,
the difference in methane production emerges earlier due to the substrate employed. In these studies,
the carbon sources used were synthetic wastewater, which has alcohols and VFA readily available
from the beginning of experimentation [36,38]. Nejayote, on the other hand, is a complex waste that
requires hydrolysis and acidogenesis steps before VFA and alcohols are available, which explains the
delay in DIET.
192
Energies 2018, 11, 2101
After the BMP test was concluded, the COD concentration was higher in N reactors than N0 and
N10 reactors. This fact explains part of the difference in accumulated methane production. Butyrate
was detected in N medium, whilst there was no presence of butyrate in N0 and N10 reactors. The total
VFA concentration of N was more than double that of N0 and N10, which can, in part, go some way
towards explaining the difference between the methane potentials in the tests; however, it can not be
held accountable for all of the differences. These differences in methane yield could also be due to
some remnant organic matter that has not been yet converted in VFA in N reactors [49].
In Table 3, the results are compared with studies that were developed with conductive materials.
It is noted in Table 3 that the percentage in the increase of methane production is within plausible
values. An increase of 54% when the biofilm is previously generated in GAC and 34% without the
previous formation of GAC biofilm are reasonable results.
Table 3. Methane production increase with conductive materials.
Reactor Conductive Methane Production Time

Substrate Reference
Volume (mL) Material Increase (%) (Days)
120 Nejayote GAC 54 30 This study
120 Nejayote GAC 34 30 This study
250 Ethanol Graphene 25 12 [38]
500 Synthetic wastewater GAC 86 43 [17]
120 Synthetic wastewater Magnetite 32 20 [56]
250 Glucose Graphene 51 15 [36]
250 Sludge treatment plant GAC 17 20 [33]
4.3. GAC Conductivity

DIET was stimulated by the addition of GAC in the reactors. GAC works as a support
for microorganism communities and as electron conductor between exoelectrogenic bacteria and
electrotrophic archaea. DIET using stainless steel as conductive support had a kinetic advantage
that is 108 times greater than IHT to compete for the electron donor [69]. Stainless steel conductivity
is 667 μS·cm−1 , while the GAC conductivity employed in this study was 3600 μS·cm−1 . Similar
conductive materials employed before, such as magnetite [56] and biochar [30], possessed
conductivities of 160 μS·cm−1 and 5 μS·cm−1 , respectively. GAC has substantially higher values
in terms of conductivity; making it an ideal material to use given its surface area, its high conductivity,
and it is economically low cost [28,29].
4.4. Redox Potential

One of the most important effects, which GAC causes in the improvement of the methanogenic
performance, is the change that it produces in redox potential. Redox potential must be under −200 mV
for methanogenic activity to be possible. When redox potential is higher than −200 mV, this activity
is virtually negligible [70]. All of the reactors in this study were under this barrier. As shown in the
results, redox potential changed from −222 ± 7 mV to −471 ± 2 mV from control reactors N to N0
and N10 reactors. GAC had a direct effect on reducing redox potential by generating a reductive
atmosphere, which is supported by evidence in other studies [71] where carbon nanotubes caused a
reduction in redox potential in an abiotic environment. The potential energy difference (ΔEo ) between
the electron donor and electron acceptor is directly proportional to the energy (ATP) that a microbe
could employ from the metabolic process [72]. Consequently, a high reducing micro-environment,
produced by the GAC presence in the anaerobic reactor, expedites electron donation to provide a
thermodynamic driving force for accomplishing the electrophilic attack in terms of reduction process,
increasing carbon dioxide reduction to methane, which could be one of the key points for improving
the reactor performance [71].
193
Energies 2018, 11, 2101
4.5. Relationship Between Redox Potential, Methane Production and Archaea Communities
There was evidence that low redox potential stimulated the syntrophic relationship between
exoelectrogenic bacteria and electrotrophic methanogens, resulting in elevated levels in methane
concentration, which can be observed in Figure 1C. A previous study observed that a pure
methanogenic archaeal culture with carbon materials had higher methane yield than a control reactor
without carbon material, due to lower redox potential [71]. That means that, without an exoelectrogenic
partner to send electrons through DIET to electrotrophic methanogens, methane yield is improved
due to redox potential reduction. In a pure culture of hydrogenotrophic methanogens, the redox
potential was under −450 mV, causing an increase in its population [70]. This change may be a result
of the fact that it requires less thermodynamic energy to undertake an electrophilic attack by these
methanogens to reduce carbon dioxide to methane. The low redox potential that was obtained in
this study (−471 mV), by adding GAC, caused an increase of 10% in the acetoclastic methanogen
(Methanosaeta) population.
4.6. Microbial Analysis

Only Methanosaeta and Methanosarcina are able to generate methane from acetate [73]. Additionally,
Methanosaeta is recognised as an electron acceptor partner in DIET, reducing carbon dioxide to methane,
hence in the control reactor, Methanosaeta played a vital role as acetate consumer to produce methane.
However, in N0 and N10 reactors, due to its attachment to the GAC surface, Methanosaeta worked as an
electron acceptor to reduce carbon dioxide to methane. Methanobacterium is a hydrogenotrophic archaea
that tended to be in suspended solution instead of being attached to the carbon surface [74], although
Methanobacterium has been detected in conductive support as an electron acceptor in DIET [44].
Methanobacterium had a relative abundance over 10% in the N reactor, whereas when detected in
the sludges of N0 and N10, it was consistently under 5%. In this study, Methanobacterium was
responsible for reducing hydrogen partial pressure through IHT in control reactor N. This is because
Methanobacterium needs to be attached to conductive material to accept electrons. Methanolinea,
which is a hidrogenotrophic archaea, was only detected in C10 samples with a proportion of 3%.
Methanolinea and Methanobacterium can accept electrons when they are attached to a conductive
support; therefore, they are able to work as electrotrophic archaea in DIET [17]. In this study, the two
different hidrogenotrophic archaea played distinct roles: one (Methanobacterium) was suspended in the
bulk and participated in IHT by controlling hydrogen partial pressure, and other (Methanolinea) was
attached to the GAC surface and accepted electrons to reduce carbon dioxide to methane.
A variety of electroactive bacteria were enriched in the GAC surface; Geobacter, Bacteroides, and
Syntrophomonas. All of these bacteria genera have been acknowledged as participants in DIET as
electron donors [40,58]. Syntrophomonas and Bacteroides were detected with a relative abundance of 2%
and 3% for C0 and C10, respectively. However, there was no presence of Syntrophomonas and Bacteroides
in N reactors. Previous studies have shown that Bacteroides have the potential to grow significantly on
the anode of a bioelectrochemical system and on the surface of carbon cloth in anaerobic reactors. This
genus is able to extracellularly transfer electrons to ferric iron and donate electrons to electrotrophic
methanogens via DIET [40,75]. There is no evidence that Syntrophomonas are able to transport electrons
extracellularly, but notable growth has been reported when acting as a DIET partner in anaerobic
systems [16].
Geobacter were only detected in C10 with a relative abundance of 13%. It must be concluded that
this was because the biofilm was developed before introducing nejayote into the reactor and Geobacter
could then grow under favorable conditions. The absence of Geobacter in C0 can be explained by the
fact that Geobacter cannot efficiently degrade complex organic waste [76]. An additional supposition
is that Geobacter is not viable in complex waste under hard conditions, as relative high salinity
environment [27,40]. Geobacter enrichment was produced during the ten days before undertaking the
BMP test in N10 reactors. This enrichment was not possible in N0 reactors due to the fact that GAC
194
Energies 2018, 11, 2101
was added to the reactors at the same time as nejayote, and other exoelectrogenic bacteria coming from
nejayote or inoculum, overtaking Geobacter and adapting better to the new environmental conditions.
Other enrichment produced in the BMP test involved Prolixibacter. This genus has not been
reported before in any study that is related with DIET in anaerobic digestion as exoelectrogenic
bacteria. The capacity of Prolixibacter to transport electrons extracellularly has not yet been concluded,
but it has been reported as a fermentative bacterium in sediment fuel cells that could be able to
transfer electrons extracellularly [77]. In addition, its enrichment in microbial electrolysis cells has
been reported [78].
The GAC biofilm developed previously in N10 reactors had two positive effects in the
exoelectrogenic bacteria community. C10 exoelectrogenic community was 4% higher than the C0
exoelectrogenic community and the Geobacter genus only appeared in C10 samples.
Microbial analysis has shown a relative abundance of 12% and 14% of Parabacteroides attached
to N0 GAC and N10 GAC, respectively, while it was not detected in the N reactor. This genus
is known for its capacity to degrade polysaccharides to acetate [79]. The main function of these
fermentative bacteria was to convert organic matter to compounds that were available to be oxidized by
exoelectrogenic bacteria. Parabacteroides played an important role in DIET, producing the essential VFA
required by the exoelectrogenic bacteria to degrade and donate electrons to electrotrophic methanogens.
Parabacteroidetes growth has been reported in microbial fuel cells [80], assuming the same role as
in this study, however it has not been reported in studies that are related with the promotion of
DIET in anaerobic reactors. This is a result of the fact that most of these studies employ synthetic
wastewater that is easily converted to methane. One of the difficulties previously detected in DIET
was the hydrolysis and acidogenesis limit of VFA and alcohol concentration that was required to take
advantage of DIET instead of IHT [52]. Parabacteroidetes growth in the conductive material greatly
assists in resolving this issue, by generating the crucial carbon source for exoelectrogenic bacteria to
oxidize. The electroactive community in C10 was 8% higher than in C0 and 38% higher than S.
5. Conclusions
GAC addition had a positive impact on the BMP test. GAC had four main resulting effects
on the anaerobic digestion process; increased methane production, redox potential reduction, DIET
promotion, and electroactive biofilm development. A low redox potential made an electrotrophic attack
to reduce carbon dioxide to methane easier. Additionally, DIET was promoted instead of IHT resulting
in higher methane concentration and greater methane production. Furthermore, the electroactive
community population intensified when the biofilm was developed 10 days before the assay, and the
reactors with previous biofilm obtained a better yield than those without. This study was the first step
to a better understanding of the relationship between DIET promotion and a complex wastewater,
such as nejayote. Undertaking a study using bigger reactors is necessary to further support these
promising results, in order to fully realise the potential of testing these conclusions on a large scale.
Author Contributions: D.V. design of experiments, performed physico-chemical parameter analysis and BMP
tests, participated in the evaluation of microorganisms communities and drafted the manuscript. C.R. helped
co-ordinate the project and revised the manuscript. B.C.-C. developed the DNA extraction protocol, participated
in the evaluation of microorganisms communities and revised the manuscript. J.A.D.-M. performed VFA analysis
and participated in physico-chemical parameter analysis and BMP test. R.T.-T. developed the DNA extraction
protocol and performed MG RAST analysis. A.C.-V. participated in developing the DNA extraction protocol
and performed DNA extraction. L.A.-G. designed and coordinated this work alongside helping to draft the
manuscript. All authors read and approved the final manuscript.
Funding: Mexican Council for Science (CONACYT) for the financial support granted to carry out this study
through grant 738499 awarded for doctoral studies.
Acknowledgments: The authors thank Q.I Tanit Toledano Thompson for technical assistance in SEM micrographs.
Also we would like to thank M.S. Miguel Alonso Tzec Simá for technical assistance in DNA extraction.
Conflicts of Interest: The authors declare that they have no competing interests.
195
Energies 2018, 11, 2101
References
1. España-Gamboa, E.; Domínguez-Maldonado, J.A.; Tapia-Tussell, R.; Chale-Canul, J.S.; Alzate-Gaviria, L.
Corn industrial wastewater (nejayote): A promising substrate in Mexico for methane production in a coupled
system (APCR-UASB). Environ. Sci. Pollut. Res. 2017, 25, 712–722. [CrossRef] [PubMed]
2. Salmerón-Alcocer, A.; Rodríguez-Mendoza, N.; Pineda-Santiago, V.; Cristiani-Urbina, E.; Juárez-Ramírez, C.;
Ruiz-Ordaz, N.; Galíndez-Mayer, J. Aerobic treatment of maize-processing wastewater (nejayote) in a
single-stream multi-stage bioreactor. J. Environ. Eng. Sci. 2003, 2, 401–406. [CrossRef]
3. Rosentrater, K.A. A review of corn masa processing residues: Generation, properties, and potential utilization.
Waste Manag. 2006, 26, 284–292. [CrossRef] [PubMed]
4. Valderrama-Bravo, C.; Gutiérrez-Cortez, E.; Contreras-Padilla, M.; Rojas-Molina, I.; Mosquera, J.C.;
Rojas-Molina, A.; Beristain, F.; Rodríguez-García, M.E. Constant pressure filtration of lime water (nejayote)
used to cook kernels in maize processing. J. Food Eng. 2012, 110, 478–486. [CrossRef]
5. Meraz, K.A.S.; Vargas, S.M.P.; Maldonado, J.T.L.; Bravo, J.M.C.; Guzman, M.T.O.; Maldonado, E.A.L.
Eco-friendly innovation for nejayote coagulation–flocculation process using chitosan: Evaluation through
zeta potential measurements. Chem. Eng. J. 2016, 284, 536–542. [CrossRef]
6. García-Zamora, J.L.; Sánchez-González, M.; Lozano, J.A.; Jáuregui, J.; Zayas, T.; Santacruz, V.; Hernández, F.;
Torres, E. Enzymatic treatment of wastewater from the corn tortilla industry using chitosan as an adsorbent
reduces the chemical oxygen demand and ferulic acid content. Process Biochem. 2015, 50, 125–133. [CrossRef]
7. Argun, M.S.; Argun, M.E. Treatment and alternative usage possibilities of a special wastewater: Nejayote.
J. Food Process Eng. 2018, 41, e12609. [CrossRef]
8. Desloover, J.; De Clippeleir, H.; Boeckx, P.; Du Laing, G.; Colsen, J.; Verstraete, W.; Vlaeminck, S.E. Floc-based
sequential partial nitritation and anammox at full scale with contrasting N2 O emissions. Water Res. 2011, 45,
9. Massara, T.M.; Komesli, O.T.; Sozudogru, O.; Komesli, S.; Katsou, E. A Mini Review of the
Techno-environmental Sustainability of Biological Processes for the Treatment of High Organic Content
Industrial Wastewater Streams. Waste Biomass Valoriz. 2017, 8, 1665–1678. [CrossRef]
10. Alexandropoulou, M.; Antonopoulou, G.; Lyberatos, G. Food industry waste’s exploitation via anaerobic
digestion and fermentative hydrogen production in an up-flow column reactor. Waste Biomass Valoriz. 2016,
7, 711–723. [CrossRef]
11. Veluchamy, C.; Kalamdhad, A.S. Biochemical methane potential test for pulp and paper mill sludge with
different food/microorganisms ratios and its kinetics. Int. Biodeterior. Biodegradation 2017, 117, 197–204.
[CrossRef]
12. Raposo, F.; Fernández-Cegrí, V.; De la Rubia, M.A.; Borja, R.; Béline, F.; Cavinato, C.; Demirer, G.;
Fernández, B.; Fernández-Polanco, M.; Frigon, J.C. Biochemical methane potential (BMP) of solid organic
substrates: Evaluation of anaerobic biodegradability using data from an international interlaboratory study.
J. Chem. Technol. Biotechnol. 2011, 86, 1088–1098. [CrossRef]
13. Kafle, G.K.; Chen, L. Comparison on batch anaerobic digestion of five different livestock manures and
prediction of biochemical methane potential (BMP) using different statistical models. Waste Manag. 2016, 48,
14. Hansen, T.L.; Schmidt, J.E.; Angelidaki, I.; Marca, E.; la Cour Jansen, J.; Mosbæk, H.; Christensen, T.H.
Method for determination of methane potentials of solid organic waste. Waste Manag. 2004, 24, 393–400.
[CrossRef] [PubMed]
15. Stams, A.J.M.; Plugge, C.M. Electron transfer in syntrophic communities of anaerobic bacteria and archaea.
Nat. Rev. Microbiol. 2009, 7, 568–577. [CrossRef] [PubMed]
16. Zhao, Z.; Zhang, Y.; Quan, X.; Zhao, H. Evaluation on direct interspecies electron transfer in anaerobic sludge
digestion of microbial electrolysis cell. Bioresour. Technol. 2016, 200, 235–244. [CrossRef] [PubMed]
17. Lee, J.-Y.; Lee, S.-H.; Park, H.-D. Enrichment of specific electro-active microorganisms and enhancement of
methane production by adding granular activated carbon in anaerobic reactors. Bioresour. Technol. 2016, 205,
196
Energies 2018, 11, 2101
18. Rotaru, A.-E.; Shrestha, P.M.; Liu, F.; Shrestha, M.; Shrestha, D.; Embree, M.; Zengler, K.; Wardman, C.;
Nevin, K.P.; Lovley, D.R. A new model for electron flow during anaerobic digestion: Direct interspecies
electron transfer to Methanosaeta for the reduction of carbon dioxide to methane. Energy Environ. Sci. 2014,
7, 408–415. [CrossRef]
19. Baek, G.; Kim, J.; Kim, J.; Lee, C. Role and Potential of Direct Interspecies Electron Transfer in Anaerobic
Digestion. Energies 2018, 11, 107. [CrossRef]
20. Dubé, C.-D.; Guiot, S.R. Direct Interspecies Electron Transfer in Anaerobic Digestion: A Review. In Biogas
Science and Technology; Guebitz, M.G., Bauer, A., Bochmann, G., Gronauer, A., Weiss, S., Eds.; Springer
International Publishing: Cham, Switzerland, 2015; pp. 101–115. ISBN 978-3-319-21993-6.
21. Liu, F.; Rotaru, A.-E.; Shrestha, P.M.; Malvankar, N.S.; Nevin, K.P.; Lovley, D.R. Promoting direct interspecies
electron transfer with activated carbon. Energy Environ. Sci. 2012, 5, 8982–8989. [CrossRef]
22. Chen, S.; Rotaru, A.-E.; Shrestha, P.M.; Malvankar, N.S.; Liu, F.; Fan, W.; Nevin, K.P.; Lovley, D.R. Promoting
Interspecies Electron Transfer with Biochar. Sci. Rep. 2014, 4, 5019. [CrossRef] [PubMed]
23. Zhao, Z.; Zhang, Y.; Holmes, D.E.; Dang, Y.; Woodard, T.L.; Nevin, K.P.; Lovley, D.R. Potential enhancement
of direct interspecies electron transfer for syntrophic metabolism of propionate and butyrate with biochar in
up-flow anaerobic sludge blanket reactors. Bioresour. Technol. 2016, 209, 148–156. [CrossRef] [PubMed]
24. Lovley, D.R. Happy together: Microbial communities that hook up to swap electrons. ISME J. 2016, 11,
25. Chen, S.; Rotaru, A.-E.; Liu, F.; Philips, J.; Woodard, T.L.; Nevin, K.P.; Lovley, D.R. Carbon cloth stimulates
direct interspecies electron transfer in syntrophic co-cultures. Bioresour. Technol. 2014, 173, 82–86. [CrossRef]
[PubMed]
26. Dang, Y.; Sun, D.; Woodard, T.L.; Wang, L.-Y.; Nevin, K.P.; Holmes, D.E. Stimulation of the anaerobic
digestion of the dry organic fraction of municipal solid waste (OFMSW) with carbon-based conductive
materials. Bioresour. Technol. 2017, 238, 30–38. [CrossRef] [PubMed]
27. Dang, Y.; Holmes, D.E.; Zhao, Z.; Woodard, T.L.; Zhang, Y.; Sun, D.; Wang, L.-Y.; Nevin, K.P.; Lovley, D.R.
Enhancing anaerobic digestion of complex organic waste with carbon-based conductive materials.
28. Hesas, R.H.; Arami-Niya, A.; Daud, W.M.A.W.; Sahu, J.N. Preparation and characterization of activated
carbon from apple waste by microwave-assisted phosphoric acid activation: Application in methylene blue
adsorption. BioResources 2013, 8, 2950–2966.
29. Arami-Niya, A.; Daud, W.M.A.W.; Mjalli, F.S.; Abnisa, F.; Shafeeyan, M.S. Production of microporous palm
shell based activated carbon for methane adsorption: Modeling and optimization using response surface
methodology. Chem. Eng. Res. Des. 2012, 90, 776–784. [CrossRef]
30. Zhao, Z.; Zhang, Y.; Yu, Q.; Dang, Y.; Li, Y.; Quan, X. Communities stimulated with ethanol to perform direct
interspecies electron transfer for syntrophic metabolism of propionate and butyrate. Water Res. 2016, 102,
31. Parker, R.; Tinsley, C.J. Electrical conduction in magnetite. Phys. Status Solidi 1976, 33, 189–194. [CrossRef]
32. Guo, K.; Soeriyadi, A.H.; Feng, H.; Prévoteau, A.; Patil, S.A.; Gooding, J.J.; Rabaey, K. Heat-treated stainless
steel felt as scalable anode material for bioelectrochemical systems. Bioresour. Technol. 2015, 195, 46–50.
[CrossRef] [PubMed]
33. Yang, Y.; Zhang, Y.; Li, Z.; Zhao, Z.; Quan, X.; Zhao, Z. Adding granular activated carbon into anaerobic
sludge digestion to promote methane production and sludge decomposition. J. Clean. Prod. 2017, 149,
1101–1108. [CrossRef]
34. Mo, J.; Steen, S.M.; Zhang, F.-Y.; Toops, T.J.; Brady, M.P.; Green, J.B. Electrochemical investigation of stainless
steel corrosion in a proton exchange membrane electrolyzer cell. Int. J. Hydrogen Energy 2015, 40, 12506–12511.
[CrossRef]
35. Hu, Q.; Sun, D.; Ma, Y.; Qiu, B.; Guo, Z. Conductive polyaniline nanorods enhanced methane production
from anaerobic wastewater treatment. Polymer 2017, 120, 236–243. [CrossRef]
36. Tian, T.; Qiao, S.; Li, X.; Zhang, M.; Zhou, J. Nano-graphene induced positive effects on methanogenesis in
anaerobic digestion. Bioresour. Technol. 2017, 224, 41–47. [CrossRef] [PubMed]
37. Jing, Y.; Wan, J.; Angelidaki, I.; Zhang, S.; Luo, G. iTRAQ quantitative proteomic analysis reveals the
pathways for methanation of propionate facilitated by magnetite. Water Res. 2017, 108, 212–221. [CrossRef]
[PubMed]
197
Energies 2018, 11, 2101
38. Lin, R.; Cheng, J.; Zhang, J.; Zhou, J.; Cen, K.; Murphy, J.D. Boosting biomethane yield and production rate
with graphene: The potential of direct interspecies electron transfer in anaerobic digestion. Bioresour. Technol.
2017, 239, 345–352. [CrossRef] [PubMed]
39. Xu, H.; Wang, C.; Yan, K.; Wu, J.; Zuo, J.; Wang, K. Anaerobic granule-based biofilms formation
reduces propionate accumulation under high H2 partial pressure using conductive carbon felt particles.
40. Lei, Y.; Sun, D.; Dang, Y.; Chen, H.; Zhao, Z.; Zhang, Y.; Holmes, D.E. Stimulation of methanogenesis in
anaerobic digesters treating leachate from a municipal solid waste incineration plant with carbon cloth.
41. Zhang, J.; Lu, Y. Conductive Fe3 O4 nanoparticles accelerate syntrophic methane production from butyrate
oxidation in two different lake sediments. Front. Microbiol. 2016, 7, 1316. [CrossRef] [PubMed]
42. Yan, W.; Shen, N.; Xiao, Y.; Chen, Y.; Sun, F.; Tyagi, V.K.; Zhou, Y. The role of conductive materials in the
start-up period of thermophilic anaerobic system. Bioresour. Technol. 2017, 239, 336–344. [CrossRef] [PubMed]
43. Yamada, C.; Kato, S.; Ueno, Y.; Ishii, M.; Igarashi, Y. Conductive iron oxides accelerate thermophilic
methanogenesis from acetate and propionate. J. Biosci. Bioeng. 2015, 119, 678–682. [CrossRef] [PubMed]
44. Zhuang, L.; Tang, J.; Wang, Y.; Hu, M.; Zhou, S. Conductive iron oxide minerals accelerate syntrophic
cooperation in methanogenic benzoate degradation. J. Hazard. Mater. 2015, 293, 37–45. [CrossRef] [PubMed]
45. Barua, S.; Dhar, B.R. Advances towards understanding and engineering direct interspecies electron transfer
in anaerobic digestion. Bioresour. Technol. 2017, 244, 698–707. [CrossRef] [PubMed]
46. American Public Health Association (APHA). Standard Methods for the Examination of Water and Wastewater;
APHA: Washington, DC, USA, 2005.
47. Poggi-Varaldo, H.M.; Valdés, L.; Esparza-Garcia, F.; Fernández-Villagómez, G. Solid substrate anaerobic
co-digestion of paper mill sludge, biosolids, and municipal solid waste. Water Sci. Technol. 1997, 35, 197–204.
[CrossRef]
48. Wang, B.; Björn, A.; Strömberg, S.; Nges, I.A.; Nistor, M.; Liu, J. Evaluating the influences of mixing strategies
on the Biochemical Methane Potential test. J. Environ. Manag. 2017, 185, 54–59. [CrossRef] [PubMed]
49. Valero, D.; Montes, J.A.; Rico, J.L.; Rico, C. Influence of headspace pressure on methane production in
Biochemical Methane Potential (BMP) tests. Waste Manag. 2016, 48, 193–198. [CrossRef] [PubMed]
50. Bowler, N. Four-point potential drop measurements for materials characterization. Meas. Sci. Technol. 2011,
22, 12001. [CrossRef]
51. Wirth, R.; Kovács, E.; Maróti, G.; Bagi, Z.; Rákhely, G.; Kovács, K.L. Characterization of a biogas-producing
microbial community by short-read next generation DNA sequencing. Biotechnol. Biofuels 2012, 5, 41.
[CrossRef] [PubMed]
52. Zhao, Z.; Li, Y.; Quan, X.; Zhang, Y. Towards engineering application: Potential mechanism for enhancing
anaerobic digestion of complex organic waste with different types of conductive materials. Water Res. 2017,
53. Trejo-González, A.; Fería-Morales, A.; Wild-Altamirano, C. The role of lime in the alkaline treatment of corn
for tortilla preparation. Adv. Chem. 1982, 198, 245–263.
54. González-Martinez, S. Biological treatability of the wastewaters from the alkaline cooking of maize
(Indian corn). Environ. Technol. 1984, 5, 365–372. [CrossRef]
55. Eddy, M. Wastewater Engineering: Treatment, Disposal and Reuse; McGraw-Hill: New York, NY, USA, 1991.
56. Cruz Viggi, C.; Rossetti, S.; Fazi, S.; Paiano, P.; Majone, M.; Aulenta, F. Magnetite particles triggering a faster
and more robust syntrophic pathway of methanogenic propionate degradation. Environ. Sci. Technol. 2014,
57. Liu, Y.; Zhang, Y.; Zhao, Z.; Ngo, H.H.; Guo, W.; Zhou, J.; Peng, L.; Ni, B.-J. A modeling approach to
direct interspecies electron transfer process in anaerobic transformation of ethanol to methane. Environ. Sci.
Pollut. Res. 2016, 1–9. [CrossRef] [PubMed]
58. Rotaru, A.-E.; Shrestha, P.M.; Liu, F.; Markovaite, B.; Chen, S.; Nevin, K.P.; Lovley, D.R. Direct interspecies
electron transfer between Geobacter metallireducens and Methanosarcina barkeri. Appl. Environ. Microbiol.
2014, 80, 4599–4605. [CrossRef] [PubMed]
59. Lovley, D.R.; Phillips, E.J.P. Novel mode of microbial energy metabolism: Organic carbon oxidation coupled
to dissimilatory reduction of iron or manganese. Appl. Environ. Microbiol. 1988, 54, 1472–1480. [PubMed]
198
Energies 2018, 11, 2101
60. Kamagata, Y.; Kawasaki, H.; Oyaizu, H.; Nakamura, K.; Mikami, E.; Endo, G.; Koga, Y.; Yamasato, K.
Characterization of three thermophilic strains of Methanothrix (“Methanosaeta”) thermophila sp. nov. and
rejection of Methanothrix (“Methanosaeta”) thermoacetophila. Int. J. Syst. Evol. Microbiol. 1992, 42, 463–468.
[CrossRef] [PubMed]
61. Ibarra-Mendívil, M.H.; Gallardo-Navarro, Y.T.; Torres, P.I.; Ramírez Wong, B. Effect of processing conditions
on instrumental evaluation of nixtamal hardness of corn. J. Texture Stud. 2008, 39, 252–266. [CrossRef]
62. Rosentrater, K.A.; Flores, R.A.; Richard, T.L.; Bern, C.J. Physical and nutritional properties of corn masa
by-product streams. Appl. Eng. Agric. 1999, 15, 515–523. [CrossRef]
63. Krishnan, R.; Ríos, R.; Salinas, N.; Durán-de-Bazúa, C. Treatment of Maize Processing Industry Wastewater
by Percolating Columns. Environ. Technol. 1998, 19, 417–424. [CrossRef]
64. Chen, Y.; Kurt, S.; Creamer, J.J.C. Inhibition of anaerobic digestion process: A review. Bioresour. Technol. 2007,
65. Singh, R.P.; Kumar, S.; Ojha, C.S.P. Nutrient requirement for UASB process: A review. Biochem. Eng. J. 1999,
3, 35–54. [CrossRef]
66. Omil, F.; Lens, P.; Visser, A.; Hulshoff Pol, L.W.; Lettinga, G. Long-term competition between sulfate reducing
and methanogenic bacteria in UASB reactors treating volatile fatty acids. Biotechnol. Bioeng. 1998, 57, 676–685.
[CrossRef]
67. Gerardi, M.H. The Microbiology of Anaerobic Digesters; John Wiley & Sons: Hoboken, NJ, USA, 2003; ISBN
0471468959.
68. Lettinga, G.; Hobma, S.W.; Pol, L.W.H.; De Zeeuw, W.; De Jong, P.; Grin, P.; Roersma, R. Design operation
and economy of anaerobic treatment. Water Sci. Technol. 1983, 15, 177–195. [CrossRef]
69. Li, Y.; Zhang, Y.; Yang, Y.; Quan, X.; Zhao, Z. Potentially direct interspecies electron transfer of methanogenesis
for syntrophic metabolism under sulfate reducing conditions with stainless steel. Bioresour. Technol. 2017,
70. Hirano, S.; Matsumoto, N.; Morita, M.; Sasaki, K.; Ohmura, N. Electrochemical control of redox potential
affects methanogenesis of the hydrogenotrophic methanogen Methanothermobacter thermautotrophicus.
Lett. Appl. Microbiol. 2013, 56, 315–321. [CrossRef] [PubMed]
71. Salvador, A.F.; Martins, G.; Melle-Franco, M.; Serpa, R.; Stams, A.J.M.; Cavaleiro, A.J.; Pereira, M.A.;
Alves, M.M. Carbon nanotubes accelerate methane production in pure cultures of methanogens and in a
syntrophic coculture. Environ. Microbiol. 2017, 19, 2727–2739. [CrossRef] [PubMed]
72. Thrash, J.C.; Coates, J.D. Direct and indirect electrical stimulation of microbial metabolism. Environ. Sci. Technol.
2008, 42, 3921–3931. [CrossRef] [PubMed]
73. Zinder, S.H. Physiological ecology of methanogens. In Methanogenesis; Springer: Boston, MA, USA 1993;
pp. 128–206.
74. Xu, S.; He, C.; Luo, L.; Lü, F.; He, P.; Cui, L. Comparing activated carbon of different particle sizes on
enhancing methane generation in upflow anaerobic digester. Bioresour. Technol. 2015, 196, 606–612. [CrossRef]
[PubMed]
75. Wang, A.; Liu, L.; Sun, D.; Ren, N.; Lee, D.-J. Isolation of Fe (III)-reducing fermentative bacterium Bacteroides
sp. W7 in the anode suspension of a microbial electrolysis cell (MEC). Int. J. Hydrogen Energy 2010, 35,
3178–3182. [CrossRef]
76. Wu, D.; Wang, T.; Huang, X.; Dolfing, J.; Xie, B. Perspective of harnessing energy from landfill leachate
via microbial fuel cells: Novel biofuels and electrogenic physiologies. Appl. Microbiol. Biotechnol. 2015, 99,
77. Holmes, D.E.; Nevin, K.P.; Woodard, T.L.; Peacock, A.D.; Lovley, D.R. Prolixibacter bellariivorans gen.
nov., sp. nov., a sugar-fermenting, psychrotolerant anaerobe of the phylum Bacteroidetes, isolated from a
marine-sediment fuel cell. Int. J. Syst. Evol. Microbiol. 2007, 57, 701–707. [CrossRef] [PubMed]
78. Huang, L.; Jiang, L.; Wang, Q.; Quan, X.; Yang, J.; Chen, L. Cobalt recovery with simultaneous methane and
acetate production in biocathode microbial electrolysis cells. Chem. Eng. J. 2014, 253, 281–290. [CrossRef]
199
Energies 2018, 11, 2101
79. Hodgson, D.M.; Smith, A.; Dahale, S.; Stratford, J.P.; Li, J.V.; Grüning, A.; Bushell, M.E.; Marchesi, J.R.;
Avignone Rossa, C. Segregation of the anodic microbial communities in a microbial fuel cell cascade.
Front. Microbiol. 2016, 7, 699. [CrossRef] [PubMed]
80. Toczyłowska-Mamińska, R.; Szymona, K.; Król, P.; Gliniewicz, K.; Pielech-Przybylska, K.; Kloch, M.;
Logan, B.E. Evolving Microbial Communities in Cellulose-Fed Microbial Fuel Cell. Energies 2018, 11,
124. [CrossRef]
200
energies
Article
Nitric Acid Pretreatment of Jerusalem Artichoke
Stalks for Enzymatic Saccharification and
Bioethanol Production
Urszula Dziekońska-Kubczak *, Joanna Berłowska, Piotr Dziugan, Piotr Patelski,
Katarzyna Pielech-Przybylska and Maria Balcerek
Institute of Fermentation Technology and Microbiology, Faculty of Biotechnology and Food Sciences,
Lodz University of Technology, Wolczanska 171/173, 90-924 Lodz, Poland; [email protected] (J.B.);
[email protected] (P.D.); [email protected] (P.P.); [email protected] (K.P.-P.);
[email protected] (M.B.)
Received: 8 June 2018; Accepted: 13 August 2018; Published: 17 August 2018
Abstract: This paper evaluated the effectiveness of nitric acid pretreatment on the hydrolysis and
subsequent fermentation of Jerusalem artichoke stalks (JAS). Jerusalem artichoke is considered a
potential candidate for producing bioethanol due to its low soil and climate requirements, and
high biomass yield. However, its stalks have a complexed lignocellulosic structure, so appropriate
pretreatment is necessary prior to enzymatic hydrolysis, to enhance the amount of sugar that can
be obtained. Nitric acid is a promising catalyst for the pretreatment of lignocellulosic biomass due
to the high efficiency with which it removes hemicelluloses. Nitric acid was found to be the most
effective catalyst of JAS biomass. A higher concentration of glucose and ethanol was achieved after
hydrolysis and fermentation of 5% (w/v) HNO3 -pretreated JAS, leading to 38.5 g/L of glucose after
saccharification, which corresponds to 89% of theoretical enzymatic hydrolysis yield, and 9.5 g/L of
ethanol. However, after fermentation there was still a significant amount of glucose in the medium.
In comparison to more commonly used acids (H2 SO4 and HCl) and alkalis (NaOH and KOH), glucose
yield (% of theoretical yield) was approximately 47–74% higher with HNO3 . The fermentation of 5%
nitric-acid pretreated hydrolysates with the absence of solid residues, led to an increase in ethanol
yield by almost 30%, reaching 77–82% of theoretical yield.
Keywords: Jerusalem artichoke; lignocellulose; acid pretreatment; nitric acid; alkali pretreatment;
enzymatic hydrolysis; ethanol fermentation
1. Introduction
According to forecasts by the International Energy Agency (IEA), as the population increases
(1.3-fold) between 2009 and 2050, energy consumption will grow even more quickly (2-fold), reaching
15–18 billion tons of oil equivalent (TOE) in 2035 [1,2]. Thus far, world energy demand has been met
mainly by burning fossil fuels. However, due to the depletion of coal, oil, and natural gas reserves,
as well as increasing public awareness of the environmental impact of emissions, more attention
is being focused on developing renewable energy sources, such as biofuels [3]. First-generation
biofuels are mostly produced from starch- and sugar-based biomass, derived from food crops grown
on agricultural land using standard processes. This can affect food supply and prices. Interest has
therefore been growing in second generation biofuels, produced using different feedstocks.
Second generation bioethanol production requires energy crops with high biomass yields per unit
land area, resistance to changing climatic conditions, diseases and pests, and the ability to grow in poor
soil conditions. Jerusalem artichoke (JA) corresponds to all these requirements [4,5]. An herbaceous

Energies 2018, 11, 2153
perennial plant belonging to the Asteraceae (also called Compositae) family, JA consists mainly of small
underground tubers with 1-3 m tall stalks. Originally used in food and feed, JA tubers contain 75–79%
water, 15–16% carbohydrates (in the form of inulin), and 2–3% protein [6,7]. In the past few decades,
JA has been cultivated extensively for fructose, oligofructose, and inulin extraction (as functional food
ingredients, and more recently to produce bioethanol from the tuber inulin [8–11]. It is also possible
to obtain other valuable biochemicals from lignocellulosic biomass, including biobutanol [12] and
dimethylfuran [13]. Jerusalem artichoke tubers are considered an attractive feedstock, for producing
these chemicals [14,15].
Much less attention has been given to the opportunities for producing bioethanol from JA stalks
(JAS). Like all lignocellulosic material, JAS consist mainly of cellulose, hemicellulose, and lignin, bound
together in a complex matrix which makes the raw material resistant to enzymatic hydrolysis [16,17].
Appropriate pretreatment must therefore be applied, to break down this recalcitrant structure and
make the cellulose more accessible to hydrolyzing enzymes [16]. Many pretreatment methods exist,
including physical, chemical, physicochemical, biological, and combined methods. However, the best
method for pretreating JAS has not yet been found. Only a few papers describe bioethanol production
from JAS. Kim et al. [18] and Kim and Kim [19] suggest using whole JA plants for bioethanol production.
Using SSF and CBP processes, respectively, these authors treated the tubers as a source of nitrogen for
Kluyveromyces marxianus yeasts. Khatun et al. [20] used inulinase producing Saccharomyces cerevisiae
yeasts, for the fermentation of JAS pretreated with NaOH. Song et al. [21] investigated the effect of
hydrogen peroxide-acetic acid (HPAC) pretreatment. All these studies have had a significant impact
on research into bioethanol production from JAS. However, many issues regarding the pretreatment of
JAS remain unresolved.
Dilute acid pretreatment is the most commonly used method for preparing lignocellulosic
biomass and has great potential for industrial applications. This method is recommended by the
National Renewable Energy Laboratory (NREL), as it enables the removal and recovery of most of
the hemicelluloses, in the form of dissolved sugars, so higher yields of glucose can be achieved [22].
The removal of hemicelluloses facilitates access to the cellulose. However, this method does not result
in significant delignification [23,24]. Dilute sulfuric acid, has been widely used to pretreat various
types of lignocellulosic biomass [25–27]. Much less attention has been given to hydrochloric, and
especially nitric acid. In comparison to sulfuric acid, nitric acid causes less equipment corrosion,
and has been found to be much faster and effective for hydrolysis of lignocellulosic biomass [28–30].
Unlike acid pretreatment, alkaline pretreatment removes lignin, as well as acetyl and uronic acid
groups present in hemicelluloses. It thus leads to the solubilization of lignocellulosic complexes.
The main advantage of alkaline methods is that the reaction is performed under normal pressure and
temperature, so less energy input is required [31]. Alkaline treatment has been applied with success
to various lignocellulosic feedstocks, including switchgrass [32] and sweet sorghum bagasse [33],
enabling a high rate of cellulose digestibility.
Given the advantages of JA as a feedstock for biofuel production, and the limited research
regarding the use of JAS in second-generation processes, this paper assessed the impact of various
pretreatments on this type of biomass. Each kind of lignocellulosic biomass requires the selection of
the most appropriate pretreatment method. This study investigated the effect of different alkaline
(sodium and potassium hydroxide) and acidic (sulfuric, hydrochloric and nitric acid) pretreatments, on
the yields of subsequent enzymatic hydrolysis and fermentation. Based on our previous research [34],
as well as on literature review [35,36], the effect of the acids and alkalis was evaluated under uniform
conditions, i.e., 121 ◦ C for 60 min; as these conditions were found to be sufficient to release the structure
of the lignocellulosic complex, and thus enhance enzymatic hydrolysis. It was assumed that the use
a of a less severe method of pretreatment would also result in the production of fewer inhibitors.
To the best knowledge of the authors, this was the first study to evaluate the effect on JAS biomass,
of pretreatment with hydrochloric and nitric acids or potassium hydroxide.
202
Energies 2018, 11, 2153
2.1. Chemical Composition of Raw and Pretreated Jerusalem Artichoke Stalks

The composition of whole plant JA biomass is highly variable and depends on climatic and
cultivation conditions [18]. To evaluate the potential of JAS as a feedstock for second generation
bioethanol production, the biomass was analyzed for total solids [37], cellulose [38], hemicellulose [39],
and lignin content [40]. The raw JAS were composed of 89.05 ± 0.55% dry matter, 34.95 ± 2.59%
d.m. cellulose, 12.69 ± 3.23% d.m. hemicellulose, and 20.24 ± 2.13% d.m. lignin. The content of
cellulose, the most important component, was comparable to that of other feedstock commonly used
for bioethanol production, e.g., 36.3% for corn stover [41], 36.9% for sweet sorghum bagasse [33], 35.8%
for rice straw [42], and 38.7% for wheat straw [43].
To increase the susceptibility of lignocellulosic biomass to enzymatic hydrolysis, appropriate
pretreatment methods must be applied. In our study, the effect of different concentrations of acids
(2% and 5% w/v HCl, H2 SO4 , HNO3 ) and alkalis (2% and 5% w/v NaOH, KOH) on JAS was evaluated
by analyzing the solid and liquid fractions obtained after each pretreatment. The solid residues were
analyzed after washing for cellulose, hemicellulose, and lignin content. The dry weight loss was also
measured, to calculate the percentage of recovered solids. In the liquid fraction, the concentrations of
inhibitory compounds (formic acid, acetic acid, 5-hydroxymethylfurfural—HMF, furfural, and total
phenolics) and of sugars released during pretreatment (cellobiose, glucose, xylose, and arabinose) were
measured using the HPLC method. The rate of hemicellulose and cellulose hydrolysis was calculated,
based on the concentrations of xylose and glucose in the hydrolysates and the initial hemicellulose
and cellulose contents in the raw JAS. The results are presented in Tables 1–3.
Table 1. Solids recovery and chemical composition of pretreated Jerusalem artichoke stalks (JAS).
Pretreatment Solid Recovery Cellulose Hemicellulose Lignin

Method (%) (% d.m.)
2% HCl 74.75 ± 2.13e 1 56.03 ± 1.37bcd 5.41 ± 0.41bcd 23.66 ± 1.66bcd
5% HCl 59.51 ± 2.01ab 54.53 ± 1.99bc 2.31 ± 0.22abc 25.74 ± 1.20d
2% H2 SO4 73.52 ± 1.82d 53.09 ± 2.46b 4.93 ± 1.02bcd 24.98 ± 2.35cd
5% H2 SO4 60.27 ± 1.98ab 42.16 ± 1.80a 1.91 ± 0.23ab 27.41 ± 1.47d
2% HNO3 56.62 ± 1.18a 60.31 ± 1.67cde 5.78 ± 0.27cd 22.96 ± 1.09bcd
5% HNO3 66.27 ± 1.42c 77.27 ± 1.32f 1.31 ± 0.16a 20.70 ± 2.02bc
2% NaOH 74.83 ± 2.65d 59.30 ± 1.97cde 9.64 ± 0.66e 6.47 ± 1.22a
5% NaOH 66.29 ± 2.03c 62.75 ± 2.34e 8.17 ± 0.70de 4.77 ± 0.36a
2% KOH 72.48 ± 1.66d 56.28 ± 1.98bcd 10.43 ± 1.50e 6.90 ± 0.93a
5% KOH 64.86 ± 2.02bc 59.98 ± 2.35de 8.70 ± 0.61de 5.23 ± 0.64a
1 Different lower-case letters in the columns indicate a significant difference (p < 0.05), as analyzed by the Tukey’s
post-hoc test.
203
Table 2. Concentration of sugars in liquid fraction after acid or alkaline pretreatment of JAS, with hemicellulose and cellulose solubilization rate.
Cellobiose Glucose Xylose Arabinose

Pretreatment Method Hemicellulose Solubilization Rate (%) Cellulose Solubilization Rate (%)
(g/L)
2% HCl 0.41 ± 0.04bcd 1 2.20 ± 0.19a 9.38 ± 0.81b 0.53 ± 0.05ab 68.76 ± 5.92b 5.64 ± 0.49a
5% HCl 0.61 ± 0.12de 2.73 ± 0.53a 11.96 ± 1.32bc 0.84 ± 0.16c 87.70 ± 7.04bc 7.02 ± 1.36a
Energies 2018, 11, 2153
2% H2 SO4 0.64 ± 0.09e 2.63 ± 0.39a 10.84 ± 1.21bc 0.47 ± 0.07ab 79.52 ± 6.72bc 6.76 ± 1.00a
5% H2 SO4 0.54 ± 0.06cde 2.80 ± 0.30a 11.35 ± 0.90bc 0.68 ± 0.07bc 83.23 ± 5.81bc 7.19 ± 0.76a
2% HNO3 0.09 ± 0.01a 2.45 ± 0.22a 9.41 ± 0.85b 0.55 ± 005ab 69.01 ± 6.26b 6.30 ± 0.57a
5% HNO3 1.12 ± 0.10f 5.70 ± 0.52b 13.20 ± 1.19c 1.29 ± 0.12d 96.77 ± 8.75c 14.66 ± 1.33b
2% NaOH 0.42 ± 0.05bcd 2.95 ± 0.36a n.d. 2 0.62 ± 0.07abc 0.00 ± 0.00a 7.60 ± 0.92a
5% NaOH 0.23 ± 0.01ab 2.12 ± 0.07a 0.29 ± 0.01a 0.56 ± 0.02ab 2.15 ± 0.07a 5.45 ± 0.17a
2% KOH 0.37 ± 0.03bc 2.51 ± 0.18a 0.13 ± 0.01a 0.43 ± 0.03a 0.95 ± 0.07a 6.45 ± 0.47a
5% KOH 0.30 ± 0.02abc 2.99 ± 0.17a 0.31 ± 0.02a 0.52 ± 0.03abc 2.26 ± 0.13a 7.69 ± 0.44a
1 Different lower-case letters in the columns indicate a significant difference (p < 0.05), as analyzed by the Tukey’s post-hoc test. 2 n.d.—not detected.
Table 3. Concentration of inhibitors in liquid fraction after acid or alkaline pretreatment of JAS.
Pretreatment Formic Acid Acetic Acid TPC 1 HMF 2 Furfural
204
Method (g/L) (mg/L)
2% HCl 0.19 ± 0.02ab 3 1.16 ± 0.10a 0.69 ± 0.06b 201.61 ± 17.34c 95.81 ± 8.24a
5% HCl 0.32 ± 0.06abc 2.92 ± 0.57b 0.55 ± 0.05a 136.62 ± 26.55bc 523.79 ± 81.79c
2% H2 SO4 0.13 ± 0.02a 2.80 ± 0.41b 0.87 ± 0.12c 739.26 ± 108.93d 95.49 ± 14.07a
5% H2 SO4 0.36 ± 0.04bc 3.25 ± 0.34bc 1.08 ± 0.08de 220.65 ± 23.36c 293.09 ± 31.03b
2% HNO3 0.48 ± 0.04c 1.76 ± 0.16a 0.51 ± 0.06a 153.95 ± 13.97bc 8.33 ± 0.75a
5% HNO3 1.18 ± 0.11d 3.19 ± 0.29bc 0.46 ± 0.04a 14.00 ± 1.26a 41.63 ± 3.76a
2% NaOH 1.16 ± 0.14d 3.42 ± 0.41bc 1.39 ± 0.11f 1.42 ± 0.17a 0.98 ± 0.12a
5% NaOH 2.86 ± 0.09f 5.02 ± 0.16de 1.12 ± 0.05e 2.46 ± 0.07a 16.43 ± 0.52a
2% KOH 0.97 ± 0.07d 4.07 ± 0.30cd 1.23 ± 0.10e 1.04 ± 0.07a 1.94 ± 0.08a
5% KOH 1.97 ± 0.11e 6.05 ± 0.34e 1.25 ± 0.11ef 2.92 ± 0.17a 10.17 ± 0.57a
1 TPC—total phenolics concentration. 2 HMF—hydroxymethylfurfural. 3 Different lower-case letters in the columns indicate a significant difference (p < 0.05), as analyzed by the Tukey’s
post-hoc test.
Energies 2018, 11, 2153
Analysis of the solid fractions revealed that when acid pretreatment was applied, the hemicellulose
content in the biomass decreased significantly, from 12.69% d.m. to 1.31% d.m. (nitric acid
pretreatment), while the cellulose and lignin content increased. This could be explained by the
fact that the main reaction which occurred during pretreatment in acidic environments was the
decomposition of hemicelluloses, especially the xylan fraction. As a result, higher recovery of
hemicellulosic sugars is possible, and the cellulose present in the solid fraction is more susceptible to
enzymatic hydrolysis [44,45]. The highest cellulose content in the biomass (77.27% d.m.) was achieved
when 5% nitric acid was applied (Table 1). Under the same conditions, 96% of the hemicelluloses
were hydrolyzed (Table 2). However, biomass recovery was significantly lower (p < 0.05) (66% for
2% w/v HNO3 and 57% for 5% w/v HNO3 ) than when other pretreatment methods were used.
Rodríguez-Chong et al. [28] reported that nitric acid exhibits much higher efficiency for hemicellulose
removal, and the whole process is faster than when hydrochloric or sulfuric acids are used. Under
optimal conditions (i.e., 6% HNO3 , 122 ◦ C, 9.3 min), after pretreatment these authors obtained a
liquor containing 18.6 g xylose/L, 2.87 g glucose/L, 2.04 g arabinose/L, 0.9 g acetic acid/L, and 1.32 g
furfural/L. Zhang et al. [29] evaluated the effect of low concentrations of nitric acid on the hydrolysis
of hemicelluloses. The optimum conditions were reported as 0.6% HNO3 , 150 ◦ C for 1 min, releasing
22.01 g xylose/L, 1.91 g glucose/L, 2.90 g arabinose/L, 2.42 g acetic acid/L, and 0.21 g furfural/L.
However, these authors did not perform enzymatic hydrolysis of the solids obtained, so the effects on
cellulose saccharification are still unclear.
The recovery of solids after pretreatment with HCl or H2 SO4, NaOH or KOH at 2% concentration,
was approximately 72–74%. Increasing the acid or alkali concentration to 5% resulted in solids recovery
in the range of 59–66%. In the liquid fraction obtained after acid pretreatment, glucose was present at a
similar level (2.2–2.8 g/L), regardless of the type of acid used, except for 5% w/v nitric acid, with which
5.70 g/L of glucose was obtained (corresponding to 14.66% of cellulose hydrolysis) (Table 2). After acid
pretreatment, large amounts of furan derivatives were found in the liquid fraction, reaching 0.74 g/L
for HMF and 0.52 g/L for furfural (Table 3). Herrera et al. [46] obtained 2.5 g/L and 1.0 g/L of acetic
acid and furfural, respectively, in hydrolysate from 2% w/v HCl-pretreated sorghum straw at 100 ◦ C.
When alkaline treatment is applied, the main reactions are saponification and solvation, causing lignin
to be removed from the lignocellulosic matrix. Unlike acid treatment, the deployment of alkaline
reagents leads to lower hemicellulose hydrolysis and lower production of inhibitory compounds [45].
In our study, the application of NaOH or KOH resulted in the removal of 66–76% of the lignin,
whilst the cellulose and hemicellulose contents were approximately 56–63% d.m. and 8–10% d.m.,
respectively (Table 1).
The concentration of glucose in the liquid fraction was at a similar level to that with acid
pretreatment (p > 0.05). However, the concentrations of xylose and furan derivatives were significantly
lower, by up to 0.31 g/L and 0.98–16.43 mg/L, respectively (Tables 2 and 3). Nevertheless,
the concentration of aliphatic carboxylic acids was higher than in the acid-treated samples (p < 0.05),
due to the removal of acetic and other uronic acids, which were substituted for hemicelluloses because
of alkaline treatment [47]. The concentration of phenolic compounds was also significantly higher
(p < 0.05) in alkaline-treated samples than in acid-treated ones, and ranged from 1.12 to 1.39 g/L.
The application of acidic pretreatment led to the release of 0.46 to 1.08 g/L of TPC. Moreover, there
were significant differences (p < 0.05) in obtained level of TPC depending on acid used, with the
lowest concentration of phenols obtained for nitric acid pretreatment, and the highest for sulfuric acid
pretreatment. However, as reported by Kim et al [48], phenolic compounds are released from biomass
during enzymatic hydrolysis, regardless of whether the pretreatment step was applied or not. Phenolic
compounds are strong inhibitors, which are proven to inhibit enzymes used in the saccharification step,
as well as being responsible for their deactivation. Ximenes et al. [49,50], have shown that phenolic
compounds are important inhibitors and deactivators of cellulolytic enzymes, especially β-glucosidase,
with the strongest negative effect exhibited by a polymeric compound—tannin acid. Authors imply
205
Energies 2018, 11, 2153
that the removal of inhibitors, in particular, phenolic compounds, prior to the saccharification process,
ought to be performed to ensure high enzyme activity and to achieve better hydrolysis yield.
Taking into consideration the results obtained in our study, the pretreatment using nitric acid,
resulted in significantly lower (p < 0.05) concentrations of inhibitors in liquid fractions. Therefore,
acid treatment using nitric acid is a promising pretreatment method for JAS biomass, enabling a large
cellulose fraction to be obtained from the solid material.
2.2. Effect of Pretreatment on Enzymatic Hydrolysis

To evaluate the susceptibility of acid and alkali pretreated JAS to enzymatic hydrolysis,
the biomass obtained after each pretreatment was subjected to saccharification at 50 ◦ C for 72 h.
During that time, samples were collected every 24 h and the concentrations of glucose and other
saccharides (cellobiose, xylose, and arabinose) were assessed. The results are presented in Figure 1
and Table 4. The hydrolysis of untreated JAS resulted in the formation of a small amount of glucose
(approximately 2.5 g/L). This concentration was achieved after 24 h of hydrolysis and remained at an
almost constant level until the end of the process. The best results from among the acid pretreatment
methods, were obtained when nitric acid was used. Not only did it provide the highest glucose
yield after hydrolysis (89% of the theoretical yield) (Table 4), but more than 92% of this glucose was
formed within the first 48 h of hydrolysis (Figure 1). A similar situation was observed in the case
of 5% (w/v) HCl. However, the total amount of glucose and the glucose yield were significantly
lower (p < 0.05) (17 g/L and 55.5%, respectively). The application of 2% (w/v) HCl, or H2 SO4 (at both
concentrations—2% and 5% (w/v)), required more than 48 h of hydrolysis (Figure 1). The application
of sulfuric acid resulted in the formation of 12.5 and 13.5 g/L of glucose, respectively, when 2% and
5% (w/v) of reagent was used. Increasing the acid concentration did not lead to a significant increase
in released glucose (p > 0.05). However, taking into consideration the cellulose content in pretreated
biomass, the yield of cellulose hydrolysis was significantly (p < 0.05) higher when 5% (w/v) H2 SO4
was applied (58% of the theoretical yield), as compared with 2% (w/v) H2 SO4 (42% of the theoretical
yield) (Table 4).
An inverse relation was observed with regard to xylose formation. When 2% (w/v) sulfuric acid
was used, the concentration of released xylose was 1.90 g/L, in comparison to 1.28 g/L for 5% (w/v)
acid, which was probably caused by greater hemicellulose decomposition during pretreatment with
higher acid concentrations (Table 4). The application of hydrochloric acid showed a similar pattern
to H2 SO4 for glucose formation, i.e., increasing the concentration of HCl did not cause a significant
increase. When 2% (w/v) nitric acid was used, the concentration of glucose was significantly higher
than for other acidic reagents (p < 0.05) and reached 17.87 g/L (54% yield), after 72 h of enzymatic
hydrolysis. However, the use of a higher concentration of HNO3 (i.e., 5% w/v) resulted in a more than
two-fold increase in the amount of released glucose, up to 38.47 g/L with 89% cellulose hydrolysis
efficiency. Acid treatment converted most hemicelluloses into monomeric sugars, leading to a higher
rate of cellulose bioconversion. Martin et al. [51] achieved almost 60% cellulose conversion after
enzymatic hydrolysis of Jatropha curcas shells, pretreated with 1.5% sulfuric acid at 110 ◦ C for 60 min.
206
Energies 2018, 11, 2153
(a) (b)
(c) (d)
(e)
Figure 1. Formation of glucose during enzymatic hydrolysis of JAS pretreated with (a) hydrochloric
acid, (b) sodium hydroxide, (c) sulfuric acid, (d) potassium hydroxide and (e) nitric acid at concentration
of 2% (blue square) or 5% (purple triangle), in relation to untreated JAS (orange diamond). The dotted
lines indicate the theoretical maximum glucose concentration that can be achieved for 2% (blue) and
5% (purple) acid/alkali concentrations.
Alkali treatment leads to the degradation of side ester and glycoside chains, as well as to cellulose
swelling and decrystalization [52,53]. This enables better access by cellulolytic enzymes. In the present
study, two alkaline reagents, sodium and potassium hydroxide, were used. The results showed that the
application of NaOH resulted in higher glucose concentrations (18.85 g/L and 23.63 g/L, respectively,
for 2% and 5% (w/v) NaOH), when compared with KOH (16.33 g/L and 20.67 g/L) (Table 4). However,
after pretreatment using potassium hydroxide, the maximum level of glucose was obtained in a shorter
time. The amount of cellulose converted to glucose was also significantly higher (p < 0.05) after NaOH
treatment, than after pretreatment with KOH. Moreover, a significant amount of xylose, ranging from
4.99 g/L to 8.47 g/L, was released via enzymatic saccharification of JAS after alkali pretreatment.
NaOH was found to be an effective method, for the pretreatment of Coastal Bermuda grass, at 121 ◦ C
207
Energies 2018, 11, 2153
with 0.75% NaOH for 15 min [54]. Under these conditions, conversion efficiencies for glucan and xylan,
were 90.43% and 65.11%, respectively.
Table 4. Concentration of monomeric sugars and glucose yield after hydrolysis of acid- or
alkali-pretreated JAS.
Cellobiose Glucose Xylose Arabinose

Pretreatment Method Glucose Yield (%)
(g/L)
Untreated 0.01 ± 0.001a 1 2.62 ± 0.36a 1.42 ± 0.28ab 0.05 ± 0.01c 13.27 ± 1.21a
2% HCl 0.05 ± 0.01ab 15.70 ± 1.84b 1.53 ± 0.45ab 0.02 ± 0.00ab 50.81 ± 1.71c
5% HCl 0.05 ± 0.02ab 16.99 ± 2.50bc 1.54 ± 0.16ab 0.01 ± 0.01a 55.57 ± 1.81cde
2% H2 SO4 0.02 ± 0.01a 12.47 ± 0.88b 1.90 ± 0.29ab n.d. 2 41.64 ± 1.48b
5% H2 SO4 0.04 ± 0.01ab 13.55 ± 1.99b 1.28 ± 0.14a 0.01 ± 0.00a 57.64 ± 1.55de
2% HNO3 0.06 ± 0.10ab 17.87 ± 1.09bc 2.56 ± 0.46bc 0.01 ± 0.00a 53.90 ± 1.33cd
5% HNO3 0.21 ± 0.05b 38.47 ± 2.97d 2.68 ± 0.09c n.d. 89.32 ± 2.24g
2% NaOH 0.10 ± 0.04ab 18.85 ± 2.88bc 6.83 ± 0.44e 0.04 ± 0.01c 57.53 ± 1.67de
5% NaOH 0.10 ± 0.03ab 23.63 ± 4.07c 4.99 ± 0.60d n.d. 67.60 ± 1.51f
2% KOH 0.01 ± 0.00a 16.33 ± 1.64b 6.69 ± 0.68e 0.02 ± 0.00ab 52.14 ± 1.94c
5% KOH 0.01 ± 0.00ab 20.67 ± 2.52bc 8.47 ± 1.03e 0.02 ± 0.00ab 62.27 ± 1.68ef
1 Different lower-case letters in the columns indicate a significant difference (p < 0.05), as analyzed by the Tukey’s
post-hoc test. 2 n.d.—not detected.
2.3. Concentration of Inhibitors and Fermentation of Hydrolysates

The next step in the process of obtaining bioethanol, after enzymatic hydrolysis of cellulose, is
microbial fermentation of the released glucose to ethanol, conducted mainly by Saccharomyces cerevisiae
(S. cerevisiae) yeasts. Unfortunately, these microbes are very sensitive to the presence in the hydrolysates,
of products from the degradation of lignocellulose. These products include a wide range of substances
formed either from the decomposition of sugars, i.e., furan derivatives and aliphatic carboxylic acids,
or from lignin, i.e., phenolic compounds [45]. The furan derivatives (furfural and HMF), are formed by
the dehydration of pentose and hexose sugars, respectively, and are responsible for reducing yeast
growth, as well as ethanol yield and productivity. Acetic acid is formed because of the hydrolysis
of the acetyl groups in hemicellulose, whilst formic acid is a product of the degradation of furan
aldehydes [45,55]. A variety of phenolic compounds are formed from lignin, when pretreatment is
applied. They can affect the growth of microorganisms, as well as fermentation efficiency, and their
toxicity is most probably related with specific functional groups [55].
After pretreatment, the biomass was washed thoroughly with water to remove inhibiting
compounds. However, there was a risk that small amounts of inhibiting compounds could remain
bound to solid fractions and affect downstream processes. To determine the concentration of inhibitors
in the hydrolysates, HPLC analysis was performed. The results are summarized in Table 5. In the
raw JAS hydrolysates, only acetic acid was present, at a concentration of 1.11 mg/L. No formic acid,
furan derivatives, or phenolic compounds were detected. The concentration of aliphatic carboxylic
acids in hydrolysates obtained from pretreated biomass was in the range of 2.57–22.23 mg/L and
8.31–37.44 mg/L, respectively, for formic and acetic acids. The only exceptions were the hydrolysates
obtained from HNO3 -pretreated biomass, in which the concentration of acetic acid was significantly
higher (p < 0.05) than in hydrolysates of JAS treated with other chemicals, reaching 128.58 mg/L
(2 mM). However, this concentration is still not considered to be toxic to S. cerevisiae. In fact, according
to Larsson et al. [56], below 100 mM acetic acid can have a stimulatory effect on ethanol fermentation.
Furan compounds, are the other large group of inhibitory compounds present in lignocellulosic
hydrolysates. It has been reported that, as with carboxylic aliphatic acids, moderate amounts of furan
aldehydes in fermented media can enhance the fermentation efficiency of recombinant xylose-utilizing
S. cerevisiae [55]. In our study, the concentrations of HMF and furfural in hydrolysates, were in the
ranges of 1.68–12.29 mg/L and 2.83–7.78 mg/L, respectively.
208
Energies 2018, 11, 2153
Table 5. Fermentation inhibitors in acid and alkali pretreated JAS hydrolysates.
Pretreatment Formic Acid Acetic Acid HMF 1 Furfural TPC 2

Method (mg/L)
Untreated n.d. 3 1.11 ± 0.09a n.d. n.d. n.d.
2% HCl 6.41 ± 0.65bc 4 29.99 ± 3.05f 2.87 ± 0.22ab 7.78 ± 0.48e 162.11 ± 19.15cde
5% HCl 8.26 ± 0.66cd 35.04 ± 2.25g 3.40 ± 0.38b 6.68 ± 0.44de 178.09 ± 13.34ef
2% H2 SO4 12.29 ± 0.66e 12.36 ± 0.82bc 12.29 ± 1.41d 5.40 ± 0.42bcd 173.83 ± 12.56de
5% H2 SO4 8.11 ± 0.55cd 37.44 ± 1.34g 5.26 ± 0.66c 6.18 ± 0.72cde 155.96 ± 15.42cd
2% HNO3 14.60 ± 0.76f 128.58 ± 2.98i 2.54 ± 0.25ab 4.53 ± 0.12b 196.91 ± 8.27f
5% HNO3 9.48 ± 1.04d 68.76 ± 2.97h n.d. 4.93 ± 0.90bc 80.63 ± 8.11a
2% NaOH 19.62 ± 1.11g 14.09 ± 1.27cd n.d. 6.39 ± 0.41cde 128.04 ± 18.20b
5% NaOH 22.23 ± 0.72h 8.31 ± 0.65b n.d. 2.83 ± 0.31a 161.34 ± 14.13cde
2% KOH 2.57 ± 0.25a 18.56 ± 1.08de 2.75 ± 0.35ab 7.15 ± 0.73e 151.21 ± 19.11c
5% KOH 5.37 ± 0.99b 20.32 ± 1.68e 1.86 ± 0.24a 6.86 ± 0.75de 170.12 ± 17.15cde
1 HMF—hydroxymethylfurfural. 2 TPC—total phenolics concentration. 3 n.d.—not detected. 4 Different lower-case
letters in the columns indicate a significant difference (p < 0.05), as analyzed by the Tukey’s post-hoc test.
The concentration of phenolic compounds ranged from 80.63 to 196.91 g/L, respectively, when 5%
w/v and 2% w/v HNO3 were applied. In hydrolysates obtained with the use of other reagents, the TPC
value was between 128 and 178 mg/L. Despite the significant differences between the samples prepared
using different pretreatment methods, no direct correlation was observed between the concentrations
of furan or phenolic compounds, and the initial pretreatment methods used. This implied that washing
the solids before enzymatic hydrolysis, had successfully removed all analyzed inhibitors which had
formed during the pretreatment step.
To evaluate the susceptibility of the hydrolysates to be converted into bioethanol, fermentations
were performed using commercial distiller’s yeast at a dose of 0.5 g/L of hydrolysate, supplemented
with ammonium phosphate (0.3 g/L) as a nutrient. The results are presented in Figures 2 and 3.
The application of dilute sulfuric acid, resulted in the formation at most of 5 g/L of ethanol (Figure 2).
It was also observed that when a higher concentration of acid was used, the yield of ethanol decreased,
from 80% to 51% respectively, for 2% and 5% (w/v) H2 SO4 (Figure 3). This was despite the fact that
the initial level of sugars in the hydrolysates was at a similar level. Moreover, not all the available
glucose was converted by the yeast, following pretreatment with 5% (w/v) H2 SO4 . Pretreatment
with HCl or KOH (at concentrations of both 2% and 5%), as well as 2% NaOH, resulted in similar
ethanol concentrations (p > 0.05), ranging from 6.2 g/L to 7.2 g/L, with corresponding ethanol yields
in the rage of 68–78% of the theoretical yield. The utilization of glucose during fermentation in these
samples was almost complete. Pretreatment with nitric acid had a significantly different impact on the
fermentation process, compared with the other tested chemicals. The application of 5% (w/v) HNO3 ,
led to the formation of 38.5 g/L of glucose during saccharification. However, only 9.5 g/L of ethanol
was obtained after fermentation, which was only 46% of the theoretical yield. Moreover, the amount of
glucose that remained in the medium after fermentation was significant (11.5 g/L), which may imply
that there were still some unidentified inhibitors present in hydrolysate.
As shown in Table 5, the concentrations of acetic and formic acids, as well as of furfural and
HMF, were not at levels considered toxic to yeasts. The concentration of phenolic compounds was also
relatively low. However, the low yield of ethanol may have been caused by the presence of inhibitory
compounds, which were released from the pretreated biomass when the ethanol was produced, or due
to synergic interactions between inhibitors. In addition, the exact composition of phenolic compounds
in tested hydrolysates is unknown. Despite the low TPC level, the concentration of individual phenols
can be close to inhibitory to the yeast fermentation process. Nitric acid is a strong oxidizing agent,
and for that reason, some compounds that are products of the oxidative reactions of commonly
found inhibitors may have been present in the medium [57]. Luo et al. [58] found 2-furancarboxylic
acid, 2-furanacetic acid, and 5-hydroxymethylfurancarboxylic acid, of which furfural and HMF are
precursors, in nitric-acid hydrolysate of aspen chips. Several phenolic compounds are also formed
209
Energies 2018, 11, 2153
during the degradation of lignin, and some of these could have remained bound to the biomass.
Phenolic compounds are much stronger inhibitors of S. cerevisiae than aliphatic acids or furans, so even
at low concentrations, they can decrease the ethanol yield significantly. Modelska et al. [59] found the
minimal inhibitory concentration (MIC) of vanillin to be 0.25% (2.5 g/L). Adeboye et al. [60] reported
that the most toxic inhibitor of S. cerevisiae Ethanol Red, was 4-hydroxy-3-methoxycinnaldehydyde
(coniferyl aldehyde), for which the toxicity limit was as low as 1.8 mM (0.32 g/L). The results presented
in this paper, are in line with those of other authors working with dilute nitric acid pretreatment.
Following the fermentation of wheat straw pretreated with 1% nitric acid at 130 ◦ C, Tutt et al. [61]
obtained an ethanol yield of 95 g/kg of biomass, which corresponded to 59.2% fermentation efficiency.
Figure 2. Concentration of ethanol in hydrolysates pretreated using different reagents after 72 h of

fermentation and concentration of residual glucose after fermentation. Different lower-case letters,
indicate significant differences (p < 0.05) between the mean values of the ethanol concentration (Tukey’s
post-hoc test). Different capital letters, indicate significant differences (p < 0.05) between the mean
values of the residual glucose concentration (Tukey’s post-hoc test).
Figure 3. Yield of ethanol (% of theoretical) from hydrolysates pretreated using different reagents, after
72 h of fermentation. Different lower-case letters, indicate significant differences (p < 0.05) between the
mean values of the ethanol yield (Tukey’s post-hoc test).
210
Energies 2018, 11, 2153
Fermentation of Nitric Acid Pretreated Hydrolysates—The Effect of Yeast Strain, Inoculum Size and
Presence of Solids Residues
To increase the ethanol yield in nitric acid pretreated JAS hydrolysates, additional fermentation
trials were performed. The impact of a yeast strain (Thermosacc Dry, Ethanol Red), and applied
yeast inoculum (0.5 and 1.5 g/L) were assessed. In addition, since all previous experiments were
performed with the presence of solid residues during the fermentation process, the effect of separation
of unhydrolyzed biomass before fermentation was also evaluated. The results are presented in Figure 4.
Figure 4. Yield of ethanol (% of theoretical) from nitric acid pretreated hydrolysates fermented with the
use of different yeast strains and inoculum. Different lower-case letters, indicate significant differences
(p < 0.05) between the mean values of the ethanol yield (Tukey’s post-hoc test).
The obtained results showed that the ethanol yield was not affected by the yeast strain or
the inoculum used. The fermentation trials of hydrolysates in the presence of solids, with use of
Thermosacc Dry and Ethanol Red, both at a dose of 0.5 and 1.5 g/L, resulted in similar (p > 0.05)
ethanol yield ranging from 49.22% to 53.85% of the theoretical yield. According to Azhar et al. [62],
the concentration of inoculum has an impact on sugars consumption rate and ethanol productivity,
but not on the final ethanol yield. The separation of unhydrolyzed solids before fermentation, resulted
in a significant (p < 0.05) increase in fermentation efficiency, reaching 77.61–81.51% of the theoretical
yield. However, the obtained yield was not influenced by yeast strain or inoculum size under applied
conditions. Usually, the fermentation of lignocellulosic hydrolysates is carried out in the presence of
solids, which contain mostly lignin [63]. Nonetheless, as reported by Monot et al. [64], it is a good
practice to remove the solid parts before fermentation, unless hydrolysis and fermentation processes
occur simultaneously. As shown in this paper, it is very important when fermentation of nitric acid
pretreated biomass is carried out. The separation of solids before the fermentation allowed us to
obtain almost 30% higher ethanol yield. This may have been the result of the presence of some toxic
compounds, in the unhydrolyzed solid. These compounds may have been released from biomass
when ethanol, even at low concentration, appeared in the medium.
211
Energies 2018, 11, 2153
3.1. Materials
3.1.1. Plant Biomass

Jerusalem artichoke (Helianthus tuberosus L.) stalks were harvested from organic ‘Żeń-szeń’
breeding plants (Kalinówka, Poland), towards the end of October 2017, and delivered fresh. The stalks
were air dried and chopped into 0.2–0.5 cm lengths before chemical pretreatment.
3.1.2. Enzymes and Chemicals

Enzymatic hydrolysis was carried out with use of Cellic CTec2 (Novozymes A/S, Basgsværd,
Denmark), a blend of cellulases, β-glucosidases, and hemicellulases with reported activities of
120 FPU/mL cellulase, 2371 U/g β-glucosidase, and 161 mg/g protein [65].
Hydrochloric acid (38% w/w), nitric acid (65% w/w), sulfuric acid (95% w/w), sodium hydroxide
(pellets), and potassium hydroxide (pellets), all from Chempur (Piekary Slaskie, Poland) were used to
prepare the pretreatment solutions. Glucose, xylose, arabinose, cellobiose, acetic acid, formic acid, HMF,
and furfural from Sigma Aldrich (St. Louis, MO, USA) were used as standards for high-performance
liquid chromatography (HPLC) analysis. All the reagents were at least analytical grade, and all the
aqueous solutions were prepared using high-purity deionized water (18.2 MΩ·cm, Simplicity Millipore
Waters, Milford, MA, USA). Penicillin and streptomycin (Sigma Aldrich, St. Louis, MO, USA) were
used to prevent microbial growth during saccharification.
3.1.3. Yeasts
Fermentation was conducted using the commercial dry distiller’s yeasts Saccharomyces cerevisiae
strains: Thermosacc Dry (Lallemand Ethanol Technology, Montreal, QC, Canada) and Ethanol Red
(Fermentis Ltd., Marcq-en-Baroeul, France). Both selected yeast strains, were declared by the producers
to have the average number of live cells of 2 × 1010 per gram and approx. 95% of dry matter. Depending
on experiments (see below), yeast dose was 0.5 or 1.5 g/L of hydrolysate. Before the inoculation of
hydrolysates, dry yeast was hydrated in 10 times the volume of water, at approximate 35 ◦ C for 15 min,
to ensure higher activity and homogenization [66].
3.2. Pretreatment
The biomass pretreatments were carried out with 10% w/v dry solid loadings. For acid
pretreatment, JAS was soaked in 150 mL of a 2% or 5% (w/v) solution of sulfuric acid, hydrochloric
acid or nitric acid, whilst for alkali pretreatment, 150 mL of a 2% or 5% (w/v) solution of sodium
hydroxide or potassium hydroxide was used. The samples were autoclaved at 121 ◦ C for 60 min.
After pretreatment, the samples were filtered, and the liquid fraction was analyzed after neutralization
using the HPLC method to determine the individual sugars, as well as the products of lignocellulose
degradation. The solid biomass was washed thoroughly with water until a neutral pH (7.0 ± 0.2) value
was achieved. The content of cellulose, hemicellulose, and lignin was then measured, before the solid
biomass was subjected to enzymatic hydrolysis.
3.3. Enzymatic Hydrolysis and Fermentation

The pretreated material was subjected to enzymatic hydrolysis with 5% w/v solid loading, based
on dry matter, in 100 mL of 0.05 M citric buffer solution (pH 4.8). Cellic CTec2 preparation was
used with a dose of 10 FPU/g cellulose (15.43 mg protein/g cellulose). A mixture of penicillin and
streptomycin (50 U/mL and 50 μg/mL, respectively) was added. Hydrolysis experiments were
performed at 50 ◦ C for 72 h. Samples were taken every 24 h, and the supernatant was analyzed
for monomeric sugar concentrations and lignocellulose degradation products using HPLC. After
hydrolysis, the whole hydrolysate was used in the fermentation stage. The dry distiller’s yeast
212
Energies 2018, 11, 2153
S. cerevisiae Thermosacc Dry was added to each sample at a dose of 0.5 g/L, along with ammonium
phosphate as a source of nutrients (0.3 g/L). Fermentation trials were conducted for 72 h at 32 ◦ C,
in 250 mL flat-bottomed flasks equipped with fermentation tubes. At the same time, untreated JAS
was enzymatically hydrolyzed and fermented as a reference.
The additional fermentation trials of nitric acid pretreated hydrolysates were conducted with the
use of S. cerevisiae Thermosacc Dry and Ethanol Red strains, at a standard dose of 0.5 as well as 1.5 g/L.
Fermentations were carried out directly in whole hydrolysate (including solids), or after separation
of solid residues (fermentation of liquids only). The other parameters of the fermentation process
(nitrogen nutrient, temperature, and time) remained unchanged.
3.4. Analysis and Calculations

The raw and pretreated JAS, were analyzed to determine their content of total solids according
to National Renewable Energy Laboratory (NREL) protocol [37]; cellulose according to the
Kurschner-Hoffer method [38]; hemicellulose according to the Ermakov method [39]; and lignin
according to the NREL protocol [40]. The concentrations of sugars, organic acids, ethanol, and furan
aldehydes (HMF, furfural) were analyzed using an HPLC system (Agilent 1260 Infinity, Agilent
Technologies, Santa Clara, CA, USA). The compounds were separated using a Hi-Plex H column
(7.7 × 300 mm, 8 μm) (Agilent Technologies, Santa Clara, CA, USA), equipped with a refractive index
detector (RID). The column and detector temperatures were maintained at 60 ◦ C and 55 ◦ C, respectively.
As a mobile phase, 0.05 M H2 SO4 was used at a flow rate of 0.7 mL/min, with a sample volume of
20 μL. The samples were filtered through 0.22 μm syringe filters to remove any particulate matter.
The total concentration of phenolic compounds was determined using a modified Folin-Ciocalteu
method, as described by Antolak et al. [67]. Gallic acid (at concentrations of 0–250 mg/L) was used to
prepare a standard curve. Samples of enzymatic hydrolysates were analyzed directly, whilst samples
of a liquid fraction after pretreatment were diluted ten times before analysis.
The recovery of solid after pretreatment was calculated based on the following equation:
dry weight of biomass after pretreatment

Solid recovery = × 100% (1)
dry weight of untreated biomass
The hemicellulose solubilization rate was calculated based on the following equation:
g
Xylose concentration in liquid fraction × 0.88
Hemicellulose solubilization = L
g × 100% (2)
Hemicellulose in untreated biomass L
The cellulose solubilization rate was calculated based on the following equation:
g
Glucose concentration in liquid fraction L × 0.90
Cellulose solubilization = g × 100% (3)
Cellulose in untreated biomass L
The glucose yield (cellulose hydrolysis efficiency) after enzymatic hydrolysis, was calculated
based on the following equation:
g
Glucose concentration in hydrolysate × 0.90
Glucose yield = Lg × 100% (4)
Cellulose in pretreated biomass L
The ethanol yield, after fermentation, was calculated based on the following equation:
g
Ethanol concentration L
Ethanol yield = g × 100% (5)
Glucose in hydrolysate L × 0.511
213
Energies 2018, 11, 2153
All tests were performed in triplicate. Statistical analysis was performed using Statistica
10 software (Tibco Software, Palo Alto, CA, USA), with the results expressed as a mean ± standard
deviation. The data were analyzed using a one-way or two-way ANOVA, followed by the Tukey’s
post-hoc test, with a significance level of 0.05.
4. Conclusions
Pretreatment of JAS, using acid and alkaline reagents, successfully enhances the efficiency of
enzymatic hydrolysis. The highest concentration of glucose (38.5 g/L, 89% of theoretical yield), after
72 h of enzymatic hydrolysis was obtained from material pretreated with 5% (w/v) HNO3 . However,
due to the fact that pretreatments were performed under the same conditions (121 ◦ C for 1 h), there
is a possibility that other acid (besides HNO3 ) and alkali reagents can reveal higher efficiency, when
optimal conditions for them are used.
The highest ethanol concentration (9.5 g/L) was produced when 5% (w/v) HNO3 was applied.
However, almost 30% of the available glucose remained after fermentation, and the yield of ethanol
was low (46% of the theoretical yield). Separation of unhydrolyzed solids before fermentation, allows
obtaining of significantly higher ethanol yield (77–82% of the theoretical). Regardless of using for
fermentation S. cerevisiae yeast strain (Ethanol-Red or Thermosacc Dry), an increase in inoculum size
(from 0.5 to 1.5 g/L) did not influence the obtained ethanol yield.
Author Contributions: Conceptualization, U.D.-K., J.B. and P.D.; Formal analysis, U.D.-K. and K.P.-P.; Funding
acquisition, J.B. and P.D.; Investigation, U.D.-K., P.P. and K.P.-P.; Methodology, U.D.-K. and P.P.; Project
administration, J.B. and P.D.; Resources, J.B. and P.D.; Validation, U.D.-K. and M.B.; Writing—original draft,
U.D.-K. and P.P.; Writing—review & editing, U.D.-K. and M.B.
Funding: This research was funded by the Polish National Center for Research and Development under grant
number BIOSTRATEG2/296369/5/NCBR/2016.
References
1. Eyidogan, M.; Kilic, F.C.; Kaya, D.; Coban, V.; Cagman, S. Investigation of Organic Rankine Cycle (ORC)
technologies in Turkey from the technical and economic point of view. Renew. Sustain. Energy Rev. 2016, 58,
2. Zecchina, A. Energy sources and carbon dioxide waste. Rend. Fis. Acc. Lincei 2014, 25, 113–117. [CrossRef]
3. Rastogi, M.; Shrivastava, S. Recent advances in second generation bioethanol production: An insight to
pretreatment, saccharification and fermentation processes. Renew. Sustain. Energ. Rev. 2017, 80, 330–340.
[CrossRef]
4. Wilson, P.; Glithero, N.J.; Ramsden, S.J. Prospects for dedicated energy crop production and attitudes towards
agricultural straw use: The case of livestock farmers. Energy Policy 2014, 74, 101–110. [CrossRef] [PubMed]
5. Long, X.H.; Shao, H.B.; Liu, L.; Liu, L.P.; Liu, Z.P. Jerusalem artichoke: A sustainable biomass feedstock for
biorefinery. Renew. Sustain. Energ. Rev. 2016, 54, 1382–1388. [CrossRef]
6. Yang, L.; He, Q.S.; Corscadden, K.; Udenigwe, C.C. The prospects of Jerusalem artichoke in functional food
ingredients and bioenergy production. Biotechnol. Rep. 2015, 5, 77–88. [CrossRef] [PubMed]
7. Negro, M.J.; Ballesteros, I.; Manzanares, P.; Oliva, J.M.; Sáez, F.; Ballesteros, M. Inulin containing biomass for
ethanol production: Carbohydrate extraction and ethanol fermentation. Appl. Biochem. Biotechnol. 2006, 132,
8. Panchev, I.; Delchev, N.; Kovacheva, D.; Slavov, A. Physicochemical characteristics of inulins obtained from
Jerusalem artichoke (Helianthus tuberosus L.). Eur. Food Res. Technol. 2011, 233, 889–896. [CrossRef]
9. Roberfroid, M.B. Functional food: Concepts and application to inulin and oligofructose. Br. J. Nutr. 2002, 87,
S139–S143. [CrossRef] [PubMed]
10. Wang, Y.Z.; Zou, S.M.; He, M.L.; Wang, C.H. Bioethanol production from the dry powder of Jerusalem
artichoke tubers by recombinant Saccharomyces cerevisiae in simultaneous saccharification and fermentation.
J. Ind. Microbiol. Biotechnol. 2015, 42, 543–551. [CrossRef] [PubMed]
214
Energies 2018, 11, 2153
11. Matías, J.; Encinar, J.M.; González, J.; González, J.F. Optimisation of ethanol fermentation of Jerusalem
artichoke tuber juice using simple technology for a decentralised and sustainable ethanol production.
Energy Sustain. Dev. 2015, 25, 34–39. [CrossRef]
12. Guan, W.; Xu, G.; Duan, J.; Shi, S. Acetone–Butanol–Ethanol Production from Fermentation of
Hot-Water-Extracted Hemicellulose Hydrolysate of Pulping Woods. Ind. Eng. Chem. Res. 2018, 57, 775–783.
[CrossRef]
13. Braun, M.; Antonietti, M. A continuous flow process for the production of 2,5-dimethylfuran from fructose
using (non-noble metal based) heterogeneous catalysis. Green Chem. 2017, 19, 3813–3819. [CrossRef]
14. Hailong, C.; Qishun, L.; Shuguang, L.; Zhongbao, Z.; Yuguang, D. Helianthus tuberosus—A good kind of
biomass source for dimethylfuran production. J. Biotechnol. 2008, 136, 271–272. [CrossRef]
15. Sarchami, T.; Rehmann, L. Optimizing Acid Hydrolysis of Jerusalem Artichoke-Derived Inulin for
Fermentative Butanol Production. Bioenergy Res. 2015, 8, 1148. [CrossRef]
16. Kumar, P.; Barrett, D.M.; Delwiche, M.J.; Stroeve, P. Methods for pretreatment of lignocellulosic biomass for
efficient hydrolysis and biofuel production. Ind. Eng. Chem. Res. 2009, 48, 3713–3729. [CrossRef]
17. Xiao, W.; Yin, W.; Xia, S.; Ma, P. The study of factors affecting the enzymatic hydrolysis of cellulose after
ionic liquid pretreatment. Carbohydr. Polym. 2012, 87, 2019–2023. [CrossRef]
18. Kim, S.; Park, J.M.; Kim, C.H. Ethanol production using whole plant biomass of Jerusalem artichoke by
Kluyveromyces marxianus CBS1555. Appl. Biochem. Biotechnol. 2013, 169, 1531–1545. [CrossRef] [PubMed]
19. Kim, S.; Kim, C.H. Evaluation of whole Jerusalem artichoke (Helianthus tuberosus L.) for consolidated
bioprocessing ethanol production. Renew. Energy 2014, 65, 83–91. [CrossRef]
20. Khatun, M.M.; Li, Y.H.; Liu, C.G.; Zhao, X.Q.; Bai, F.W. Fed-batch saccharification and ethanol fermentation
of Jerusalem artichoke stalks by an inulinase producing Saccharomyces cerevisiae MK01. RSC Adv. 2015, 5,
107112–107118. [CrossRef]
21. Song, Y.; Wi, S.G.; Kim, H.M.; Bae, H.-J. Cellulosic bioethanol production from Jerusalem artichoke
(Helianthus Tuberosus L.) using hydrogen peroxide-acetic acid (HPAC) pretreatment. Bioresour. Technol.
2016, 214, 30–36. [CrossRef] [PubMed]
22. Yang, B.; Wyman, C. Pretreatment: The key to Unlocking Low-Cost Cellulosic Ethanol. Biofuels Bioprod.
Biorefin. 2008, 2, 26–40. [CrossRef]
23. Mosier, N.; Wyman, C.; Dale, B.; Elander, R.; Lee, Y.Y.; Holtzapple, M.; Ladisch, M. Features of promising
technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 2005, 96, 673–686. [CrossRef]
24. Sannigrahi, P.; Kim, D.H.; Jung, S.; Ragauskas, A. Pseudo-lignin and pretreatment chemistry.
Energy Environ. Sci. 2011, 4, 1306–1310. [CrossRef]
25. Digman, M.F.; Shinners, K.J.; Casler, M.D. Optimizing on-farm pretreatment of perennial grasses for fuel
ethanol production. Bioresour. Technol. 2010, 101, 5305–5314. [CrossRef] [PubMed]
26. Karapatsia, A.; Pappas, I.; Penloglou, G.; Kotrotsiou, O.; Kiparissides, C. Optimization of dilute acid
pretreatment and enzymatic hydrolysis of Phalaris aquatica L. lignocellulosic biomass in batch and fed-batch
processes. Bioenergy Res. 2017, 10, 225–236. [CrossRef]
27. Deshavath, N.N.; Dasu, V.V.; Goud, V.V.; Rao, P.S. Development of dilute sulfuric acid pretreatment method
for the enhancement of xylose fermentability. Biocatal. Agric. Biotechnol. 2017, 11, 224–230. [CrossRef]
28. Rodíguez-Chong, A.; Ramírez, J.A.; Garrote, G.; Vázquez, M. Hydrolysis of sugar cane bagasse using nitric
acid: A kinetic assessment. J. Food Eng. 2004, 61, 143–152. [CrossRef]
29. Zhang, R.; Lu, X.; Sun, Y.; Wang, X.; Zhang, S. Modeling and optimization of dilute nitric acid hydrolysis on
corn stover. J. Chem. Technol. Biotechnol. 2011, 86, 306–314. [CrossRef]
30. Kim, I.; Lee, B.; Park, J.; Choi, S.; Han, J. Effect of nitric acid on pretreatment and fermentation for enhancing
ethanol production of rice straw. Carbohydr. Polym. 2014, 99, 563–567. [CrossRef] [PubMed]
31. Maurya, D.P.; Singla, A.; Negi, S. An overview of key pretreatment processes for biological conversion of
lignocellulosic biomass to bioethanol. 3 Biotech 2015, 5, 597–609. [CrossRef] [PubMed]
32. Guan, W.; Shi, S.; Blersch, D. Effects of Tween 80 on fermentative butanol production from alkali-pretreated
switchgrass. Biochem. Eng. J. 2018, 135, 61–70. [CrossRef]
33. Umagiliyage, A.L.; Choudhary, R.; Liang, Y.; Haddock, J.; Watson, D.G. Laboratory scale optimization of
alkali pretreatment for improving enzymatic hydrolysis of sweet sorghum bagasse. Ind. Crops Prod. 2015, 74,
215
Energies 2018, 11, 2153
34. Berłowska, J.; Pielech-Przybylska, K.; Balcerek, M.; Dziekońska-Kubczak, U.; Patelski, P.; Dziugan, P.;
Kr˛egiel, D. Simultaneous saccharification and fermentation of sugar beet pulp for efficient bioethanol
production. BioMed Res. Int. 2016, 2016, 3154929. [CrossRef]
35. Gonzales, R.R.; Sivagurunathan, P.; Kim, S.H. Effect of severity on dilute acid pretreatment of lignocellulosic
biomass and the following hydrogen fermentation. Int. J. Hydrogen Energy 2016, 41, 21678–21684. [CrossRef]
36. McIntosh, S.; Vancov, T. Enhanced enzyme saccharification of Sorghum bicolor straw using dilute alkali
pretreatment. Bioresour. Technol. 2010, 101, 6718–6727. [CrossRef] [PubMed]
37. Sluiter, A.; Hames, B.; Hyman, D.; Payne, C.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Templeton, D. Determination
of Total Solids in Biomass and Total Dissolved Solids in Liquid Process Samples; Technical Report No.
NREL/TP-510-42621; NREL: Golden, CO, USA, 2008.
38. Kurschner, K.; Hoffer, A. Cellulose and cellulose derivatives. Fresenius J. Anal. Chem. 1993, 92, 145–154.
[CrossRef]
39. Arasimovich, V.V.; Ermakov, A.I. Measurement of the total content of hemielluloses. In Methods for Biochemical
Studies of Plants; Ermakov, A.I., Ed.; Agropromizdat: Saint Petersburg, Russia, 1987; pp. 164–165.
Carbohydrates and Lignin in Biomass; Technical Report No. NREL/TP-510-42618; NREL: Golden, CO,
USA, 2012.
41. Ma, S.; Wang, H.; Wang, Y.; Bu, H.; Bai, J. Bio-hydrogen production from cornstalk wastes by orthogonal
design method. Renew. Energy 2011, 36, 709–713. [CrossRef]
42. Imman, S.; Arnthong, J.; Burapatana, V.; Champreda, V.; Laosiripojana, N. Fractionation of rice straw
by a single-step solvothermal process: Effects of solvents, acid promoters, and microwave treatment.
43. Valdez-Vazquez, I.; Pérez-Rangel, M.; Tapia, A.; Buitrón, G.; Molina, C.; Hernández, G.; Amaya-Delgado, L.
Hydrogen and butanol production from native wheat straw by synthetic microbial consortia integrated by
species of Enterococcus and Clostridium. Fuel 2015, 159, 214–222. [CrossRef]
44. Hendriks, A.T.W.M.; Zeeman, G. Pretreatments to enhance the digestibility of lignocellulosic biomass.
45. Jönsson, L.J.; Martín, C. Pretreatment of lignocellulose: formation of inhibitory by-products and strategies
for minimizing their effects. Bioresour. Technol. 2016, 199, 103–112. [CrossRef] [PubMed]
46. Herrera, A.; Tellez-Luis, S.J.; Gonzalez-Cabriales, J.J.; Ramírez, J.A.; Vazquez, M. Effect of the hydrochloric
acid concentration on the hydrolysis of sorghum straw at atmospheric pressure. J. Food Eng. 2004, 63,
47. Chang, V.S.; Holtzapple, M.T. Fundamental factors affecting biomass enzymatic reactivity. Appl. Biochem.
Biotechnol. 2000, 84–86, 5–37. [CrossRef]
48. Kim, D.; Orrego, D.; Ximenes, E.A.; Ladisch, M.R. Cellulose conversion of corn pericarp without pretreatment.
49. Ximenes, E.; Kim, Y.; Mosier, N.; Dien, B.; Ladisch, M. Inhibition of cellulases by phenols. Enzym. Microb.
Technol. 2010, 46, 170–176. [CrossRef]
50. Ximenes, E.; Kim, Y.; Mosier, N.; Dien, B.; Ladisch, M. Deactivation of cellulases by phenols. Enzym. Microb.
Technol. 2011, 48, 54–60. [CrossRef] [PubMed]
51. Martín, C.; García, A.; Schreiber, A.; Puls, J.; Saake, B. Combination of water extraction with dilute-sulphuric
acid pretreatment for enhancing the enzymatic hydrolysis of Jatropha curcas shells. Ind. Crop. Prod. 2015, 64,
52. Cheng, Y.S.; Zheng, Y.; Yu, C.W.; Dooley, T.M.; Jenkins, B.M.; Gheynst, J.S.V. Evaluation of high solids alkaline
pretreatment of rice straw. Appl. Biochem. Biotechnol. 2010, 162, 1768–1784. [CrossRef] [PubMed]
53. Ibrahim, M.M.; El-Zawawy, W.K.; Abdel-Fattah, Y.R.; Soliman, N.A.; Agblevor, F.A. Comparison of alkaline
pulping with steam explosion for glucose production from rice straw. Carbohydr. Polym. 2011, 83, 720–726.
[CrossRef]
54. Wang, Z.; Keshwani, D.R.; Redding, A.P.; Cheng, J.J. Sodium hydroxide pretreatment and enzymatic
hydrolysis of coastal Bermuda grass. Bioresour. Technol. 2010, 101, 3583–3585. [CrossRef] [PubMed]
55. Jönsson, L.J.; Björn, A.; Nilvebrant, N.O. Bioconversion of Lignocellulose: Inhibitors and Detoxification.
Biotechnol. Biofuels 2013, 6, 16. [CrossRef] [PubMed]
216
Energies 2018, 11, 2153
56. Larsson, S.; Palmqvist, E.; Hahn-Hägerdal, B.; Tengborg, C.; Stenberg, K.; Zacchi, G.; Nilvebrant, N.O. The
generation of fermentation inhibitors during dilute acid hydrolysis of softwood. Enzym. Microb. Technol.
1999, 24, 151–159. [CrossRef]
57. Farías-Sánchez, J.C.; Velázquez-Valadez, U.; Pineda-Pimentel, M.G.; López-Miranda, J.; Castro-Montoya, A.J.;
Carrillo-Parra, A.; Vargas-Santillán, A.; Rutiaga-Quiñones, J.G. Simultaneous Saccharification and
Fermentation of Pine Sawdust (Pinus pseudostrobus L.) Pretreated with Nitric Acid and Sodium Hydroxide
for Bioethanol Production. Bioresources 2016, 12, 1052–1063. [CrossRef]
58. Luo, C.; Brink, D.L.; Blanch, H.W. Identification of potential fermentation inhibitors in conversion of hybrid
poplar hydrolysate to ethanol. Biomass Bioenergy 2002, 22, 125–138. [CrossRef]
59. Modelska, M.; Berlowska, J.; Kregiel, D.; Cieciura, W.; Antolak, H.; Tomaszewska, J.; Binczarski, M.;
Szubiakiewicz, E.; Witonska, I.A. Concept for Recycling Waste Biomass from the Sugar Industry for Chemical
and Biotechnological Purposes. Molecules 2017, 22, 1544. [CrossRef] [PubMed]
60. Adeboye, P.T.; Bettiga, M.; Olsson, L. The chemical nature of phenolic compounds determines their toxicity
and induces distinct physiological responses in Saccharomyces cerevisiae in lignocellulose hydrolysates.
AMB Express 2014, 4, 46. [CrossRef] [PubMed]
61. Tutt, M.; Kikas, T.; Olt, J. Influence of different pretreatment methods on bioethanol production from wheat
straw. Agron. Res. 2012, S1, 269–276.
62. Azhar, S.H.M.; Abdulla, R.; Jambo, S.A.; Marbawi, H.; Gansau, J.A.; Faik, A.A.M.; Rodrigues, K.F. Yeasts in
sustainable bioethanol production: A review. Biochem. Biophys. Rep. 2017, 10, 52–61. [CrossRef]
63. Häggström, C.; Rova, U.; Brandberg, T.; Hodge, D. Integration of Ethanol Fermentation with Second
Generation Biofuels Technologies. In Biorefineries: Integrated Biochemical Processes for Liquid Biofuels;
Qureshi, N., Hodge, D., Vertès, A., Eds.; Elsevier: Amsterdam, The Netherlands, 2014; Chapter 8; pp. 161–187.
ISBN 978-0-444-59498-3.
64. Monot, F.; Duplan, J.L.; Alazard-Toux, N.; His, S. Biofuels. In Renewable Energies; Sabonnadière, J.C., Ed.;
ISTE Ltd.: London, UK; John Wiley & Sons, Inc.: Hoboken, NY, USA, 2009; Chapter 10; pp. 329–395.
ISBN 9781848211353.
65. Cannella, D.; Jørgensen, H. Do new cellulolytic enzyme preparations affect the industrial strategies for high
solids lignocellulosic ethanol production? Biotechnol. Bioeng. 2014, 111, 59–68. [CrossRef] [PubMed]
66. Balcerek, M.; Pielech-Przybylska, K.; Dziekońska-Kubczak, U.; Patelski, P.; Strak,
˛ E. Fermentation results and
chemical composition of agricultural distillates obtained from rye and barley grains and the corresponding
malts as a source of amylolytic enzymes and starch. Molecules 2016, 21, 1320. [CrossRef] [PubMed]
67. Antolak, H.; Czyzowska, A.; Sakač, M.; Mišan, A.; Đuragić, O.; Kregiel, D. Phenolic compounds contained in
little-known wild fruits as antiadhesive agents against the beverage-spoiling bacteria Asaia spp. Molecules
2017, 22, 1256. [CrossRef] [PubMed]
217
energies
Article
Bamboo Fiber and Sugarcane Skin as a
Bio-Briquette Fuel
Anna Brunerová 1, *, Hynek Roubík 2 and Milan Brožek 1
1 Faculty of Engineering, Department of Material Science and Manufacturing Technology, Czech University of
Life Sciences Prague, Kamýcká 129, 165 00 Prague, Czech Republic; [email protected]
2 Faculty of Tropical AgriSciences, Department of Sustainable Technologies, Czech University of Life Sciences
Prague, Kamýcká 129, 165 00 Prague, Czech Republic; [email protected]
Abstract: The present study deals with the issue of bio-briquette fuel produced from specific
agriculture residues, namely bamboo fiber (BF) and sugarcane skin (SCS). Both materials originated
from Thừa Thiên Huế province in central Vietnam and were subjected to analysis of their
suitability for such a purpose. A densification process using a high-pressure briquetting press
proved its practicability for producing bio-briquette fuel. Analysis of fuel parameters exhibited a
satisfactory level of all measured quality indicators: ash content Ac (BF—1.16%, SCS—8.62%) and
net calorific value NCV (BF—16.92 MJ·kg−1 , SCS—17.23 MJ·kg−1 ). Equally, mechanical quality
indicators also proved satisfactory; bio-briquette samples’ mechanical durability DU occurred at an
extremely high level (BF—97.80%, SCS—97.70%), as did their bulk density ρ (BF—986.37 kg·m−3 ,
SCS—1067.08 kg·m−3 ). Overall evaluation of all observed results and factors influencing the
investigated issue proved that both waste biomass materials, bamboo fiber and sugarcane skin,
represent suitable feedstock materials for bio-briquette fuel production, and produced bio-briquette
samples can be used as high-quality fuels.
Keywords: waste biomass; Vietnam; solid biofuel; calorific value; mechanical durability
1. Introduction
With the increasing prices and various environmental impacts created by the use of fossil fuels,
the importance of biofuel production has subsequently increased. Such production has reached
unprecedented volumes over the last 20 years [1]. In addition, the world’s population is expected
to continue growing, with the total population calculated to be almost ten billion by mid-2050 [2].
This population growth will lead to a deepening requirement for energy, and therefore, global energy
demand will subsequently increase. Energy is considered a key source for the future and plays an
important role in socioeconomic development because affordable energy is an essential ingredient for
such development [3,4].
With deforestation comprising a major problem in many parts of the developing world, there is
increasing demand for fuelwood for household cooking. This phenomenon particularly affects remote
rural communities that have no access to fuels such as liquid petroleum gas (LPG) and that depend
substantially on burning locally-collected biomass [5]. Such a demand for fuel could be covered by
bio-briquettes, which may provide necessary energy from waste materials because biomass is globally
recognized as a renewable and sustainable energy source [6]. In addition, agricultural residues have
recently been posited as a major fuel source for many potential bio-energy projects in developing
countries [1].
Biomass is now considered a primary global energy resource, providing 14% of the world’s energy
needs. It comprises at least one third of energy consumption in some developing countries. In addition,

Energies 2018, 11, 2186
biomass combustion can be considered CO2 neutral because during production, it removes CO2
from the atmosphere by photosynthesis and is later released during combustion [7]; this aspect was
however contradicted in Cherubini et al. [8], who analyzed CO2 emissions from biomass combustion
for bioenergy.
1.1. Bamboo and Sugarcane as a Source of Herbaceous Biomass

Bamboo (Bambusoideae spp.) and sugarcane plants (Saccharum officinarum) are members of the grass
family Poaceae (also known as Gramineae). Both grasses are widely spread in tropical and subtropical
climate regions. As a strongly growing species, their cultivation provides several advantages.
Bamboo plants are the fastest growing plant species on Earth (approximately 91–122 cm per day);
thus, significant potential for herbaceous biomass production is indicated. Moreover, the subfamily
Bambusoideae contains 1100–1500 different bamboo species, with their classification and identification
being complicated; approximately 69 naturally-occurring species were monitored in Vietnam
alone [9,10].
Historically, bamboo is a traditional and widely-used plant in Asia; its utilization has lasted for
centuries in different industries, for building various parts of houses, furniture production, as well as
medicinal, food [10], clothing and home craft purposes, among others [11]. Concurrently, sugarcane plants
cultivated for agricultural purposes provide 70% of sugar worldwide (having the most calories per unit
area of cultivation of any plant), with a large amount of the herbaceous waste biomass being left behind
after plant processing, for example sugarcane skin [12]. Such material is one of the by-products of the
sugarcane processing industry [13]. Today, biological residues from sugarcane plants are a particular point
of interest of different scientific fields [14,15]; thus, its potential seems significant. However, sugarcane
skin reuse in Southeast Asia, namely in Vietnam, is not ensured; the open field burning or ejection into
municipal waste is a common practice in rural areas when individuals or small processing plants are
treating sugarcane crop. In an attempt to meet increasing energy demands and maintain an appropriate
waste management strategy, waste biomass originated from processing of both mentioned grasses could be
reused and utilized; for example, as a feedstock material for various biotechnological processes [13,16,17].
Bamboo plants commonly provide waste cuttings of bamboo logs or bamboo fiber (BF), while sugarcane
processing results in the production of sugarcane bagasse. Utilization of such residues is not secured,
and they are often left behind as useless waste or burned without a purpose. Occasionally, bamboo fiber is
used as a fire starter and sugarcane skin for combustion purposes.
1.2. Biomass Availability in Vietnam

In Vietnam, biomass is considered an important source of energy, comprising approximately 90% of
domestic energy consumption in rural areas [18], as well as being an important source of energy for small
industries and farms, often located in rural areas [19]. Further, it has significant potential as a renewable
energy source, even though the agricultural sector currently accounts only for approximately 20% of the
country’s GDP [19]. Agricultural residues, as described in Schirmer [19], and animal waste, as described in
Roubík et al. [20], are offering significant potential for the generation of electricity, sources that have so far
been insufficiently tapped. Currently, biomass is mainly used by households, and waste biomass is usually
not used at all, or very ineffectively and with potential harmful consequences.
Agricultural residues (including bamboo fiber and sugarcane skin) are important sources of
biomass in Vietnam and can be taken from residues left directly on fields or from the processing
of agricultural products. The wider use of such residues may be relatively difficult because this
agricultural waste usually occurs locally (for example, rice husk residues are produced in large
quantities only in the local rice mills). As mentioned by Schirmer [19], there is a need for a reliable
supply of biomass, its efficient collection, transportation and storage, meaning well-established supply
chains, which currently do not exist.
219
Energies 2018, 11, 2186
Agricultural Waste Utilization through Briquetting

One technology for the utilization of agricultural wastes is biomass briquetting [21,22], which
involves the densification of the biomass through the use of pressure [23,24]. The advantages of
briquetting include the following: increased bulk density of the material, making it easier to transport
and store, higher energy content per unit volume and the production of homogenous product fuel
from various raw materials [22,25]. Therefore, as concluded by Chen et al. [23], densified solid biofuel
(meaning bio-briquettes) could be an important route for efficient utilization of agro-residues.
2. Methodology
The present section details research activities performed for the collection of selected waste biomass
(both materials represented by herbaceous biomass) in raw form and its initial processing in the target area
of Huế city and surrounding villages in Thừa Thiên Huế province, central Vietnam. There is a detailed
description of the investigated waste biomass laboratory preparation (drying, grinding) and laboratory
testing of its fuel properties. Further, the utilization of selected waste biomass as a feedstock material
for bio-briquette fuel production using a high-pressure briquetting press is described, together with the
determination of their suitability for such a purpose. Finally, the complete evaluation of stated mechanical
and chemical quality indicators of produced bio-briquette samples is performed.
In general, the production of bio-briquette fuel for commercial purposes is conducted to
mandatory technical standards (stated by the country of bio-briquette fuel production) ensuring
the quality and safety of such products. Thus, all procedures of experimental testing were performed
in accordance to related European technical mandatory standards stated by the European Committee
for Standardization, namely: EN 14918 (2010) [26], ISO 1928 (2010) [27], EN 15234-1 (2011) [28], EN
ISO 16559 (2014) [29], EN ISO 17225-1 (2015) [30], EN ISO 17831-2 (2015) [31], EN ISO 18122 (2015) [32],
EN ISO 18134-2 (2015) [33], EN ISO 16948 (2016) [34] and EN ISO 18123 (2016) [35].
2.1. Materials and Samples

Both investigated waste materials, bamboo fiber (Bambusoideae spp.) and sugarcane skin
(Saccharum officinarum), were identified as herbaceous biomass due to their membership in the grass
family Poaceae. Collection activities were performed at small processing plants in central Vietnam
in June of 2017. Bamboo fiber samples originated from rural areas in Thừa Thiên Huế province;
the collections were performed during several field trips; thus, samples originated from different
villages, whereas sugarcane skin samples originated from one specific processing plant in Huế city.
Nevertheless, unprocessed raw sugarcane stalks were harvested in surrounding rural areas of Huế city.
The target area of the materials’ collection is illustrated in Figure 1.
Vietnam
Huế city
Figure 1. Area of sample collection: Huế district, Thừa Thiên Huế province.
220
Energies 2018, 11, 2186
Both investigated materials were chosen because they represented commodities frequently
processed in large quantities in the target area, and their processing results in the production of
a significant amount of waste biomass.
Bamboo stalks were already sundried prior to the fiber being manually separated in processing
plants; thus, final waste biomass occurred in ideal conditions for combustion purposes (low level of
moisture content); as shown in Figure 2.
(a) (b)
Figure 2. Bamboo processing in Thừa Thiên Huế province, central Vietnam: (a) manual removing of
fiber; (b) produced waste biomass: bamboo fibers.
In contrast, sugarcane stalks occurred in the raw state before their skin was manually removed
in the processing plant; thus, produced waste biomass occurred in its initial high moisture content,
as shown in Figure 3.
(a) (b)
Figure 3. Sugarcane processing in Huế city, central Vietnam: (a) manual removing of skin; (b) produced
waste biomass: sugarcane skin.
The drying process of sugarcane skin was thus necessary prior to its subsequent utilization,
and therefore, the material was sundried in open fields in local conditions. Thereby, the drying process
was ensured without the need for any other energy source investment. Moreover, local insects were
not interested in such materials, which represents a significant advantage if considering that such
materials contained residual sugars. Unfortunately, precise initial moisture content of both materials
was not determined due to the fact that primary research activities were performed in conditions of
221
Energies 2018, 11, 2186
rural areas of Vietnam without proper measuring equipment. Further, both materials were properly
stabilized and stored in special hermetically-sealed bags for transportation.
After the materials’ transportation to the laboratories at the target destination of Prague, Czech
Republic, both materials were appropriately prepared for subsequent experimental measurements;
their particle size and moisture content were modified to the required form. Primarily, both materials
were dried in a laboratory dryer LAC, Type S100/03 (Rajhrad, Czech Republic), for 24 hours at a
temperature of 105 ◦ C until their moisture content was constant. Disintegration of feedstock materials
was also performed using a grinding hammer mill Taurus, Type VM 7,5 (Chrudim, Czech Republic),
with a vertical shaft with eight hammers as a working unit (see Figure 4).
Engine
ne
Feed
hopper
Free
swinging
hammer
hammers
Output
hole
(a) (b)
Figure 4. The disintegration device used, the hammer grinding mill: (a) scheme; (b) in practice.
A sieve with holes of 8 mm in diameter was used to unify the faction of both disintegrated
feedstock materials. The final form of properly prepared samples is expressed in Figure 5.
(a) (b)
Figure 5. The form of samples prepared for chemical and mechanical analysis: (a) bamboo fiber; (b)
sugarcane skin.
Nevertheless, the bamboo material occurred in the fibrous form, and hence, the fibers that
fell through sieve holes were prevalently slightly longer than 8 mm. In contrast, sugarcane skin
(SCS) material also contained tiny particles in the form of dust; thus, the prevalent particle size was
222
Energies 2018, 11, 2186
smaller than 8 mm. Due to such differences between the investigated materials, brief microscopic
measurements were performed to describe the particle size and shapes of investigated materials
(measurement scale 5 mm). A stereoscopic microscope Arsenal, Type 347 SZP 11-T Zoom (Prague,
Czech Republic) was used for image analysis; the resulting images are shown in Figure 6.
(a) (b)
Figure 6. Microscopic analysis of investigated samples: (a) bamboo fiber; (b) sugarcane skin.
Despite the fact that both investigated materials were represented by the top layer of plant stem
(the epidermis), the variability in materials’ behavior during disintegration and particle size were
detected during microscopic analyses (Figure 6); such variability was related to the taxonomy of
processed plants.
2.2. Fuel Analysis of Feedstock Materials

Samples were subjected to analysis of their fuel properties, which defined their basic chemical
parameters, energy potential and elementary composition. Such measurements were performed
to determine the materials’ suitability for the process of direct combustion. At least two or three
repetitions of all of the following tests were performed for all investigated samples to ensure the
correctness of measurement processes and observed data; data were recorded and evaluated using
supporting software. Final data exhibited in the Results and Discussion Section are expressed as
average values of all observed data with standard deviations.
2.2.1. Basic Parameters and Energy Potential (Mc , Ac , GCV, NCV)

Samples were initially milled into the required particle size (<0.1 mm) using a cutting mill.
Moisture Mc (%) and ash content Ac (%) analysis was performed with a thermogravimetric
analyzer LECO, Type TGA 701 (Saint Joseph, MO, USA). Experimental tests complied with
related standards, namely EN 18134-2 (2015): Solid biofuels—Determination of moisture
content—Oven dry method—Part 2: Total moisture-Simplified method and ISO 18122 (2015): Solid
biofuels—Determination of ash content.
The principal parameters of fuel properties, gross calorific value (GCV) (MJ·kg−1 ) and net calorific
value NCV (MJ·kg−1 ) express the energy potential of the investigated materials. Gross calorific values
(GCV) (MJ·kg−1 ) were primarily measured using an isoperibol calorimeter LECO, Type AC 600 (Saint
Joseph, USA) and secondarily analyzed by related software. Net calorific values (NCV) (MJ·kg−1 ) were
calculated using relations between those two parameters. The complete process of analyses used standards
EN 14918 (2010): Solid biofuels—Determination of calorific value and ISO 1928 (2010): Solid mineral
fuels—Determination of gross calorific value by the bomb calorimetric method and calculation of NCV.
223
Energies 2018, 11, 2186
2.2.2. Elementary Composition (C, H, N, S, O)

Materials content of carbon C (%), hydrogen H (%), nitrogen N (%), sulfur S (%) and oxygen O (%)
represented the last measured fuel parameters. Experimental testing was performed using laboratory
equipment LECO, Type CHN628+S (Saint Joseph, MO, USA), while the testing process corresponded
to requirements stated by related standard EN ISO 16948 (2016): Solid biofuels—Determination of
total content of carbon, hydrogen and nitrogen.
2.3. Mechanical Analysis of Bio-Briquette Samples

When the suitability of investigated materials for the process of direct combustion was proven,
they were utilized as a feedstock for bio-briquette fuel production. Subsequently, the produced
bio-briquette samples were subjected to the determination of their mechanical quality represented
by the following indicators: bulk density ρ, mechanical durability DU and rupture force RF.
Specific methods of the mentioned experimental measurements are described in the following sections.
The present section is divided into four sections: The first section is related to the production factors of
the densification process and to the entire process of bio-briquette sample production. The remaining
three sections describe bio-briquette sample quality testing and determination of specific mechanical
quality indicator results.
2.3.1. Bio-Briquette Sample Production

As mentioned previously, high-pressure briquetting technology was used for densification of two
different types of herbaceous waste biomass: bamboo fiber and sugarcane skin. Namely, a laboratory
hydraulic piston high-pressure briquetting press Briklis, Type BrikStar 30-12 (Malšice, Czech Republic),
was used as a pressing device (shown in Figure 7).
(a) (b)
Figure 7. Briklis high-pressure briquetting press: (a) scheme; (b) in practice.
This press was operated with an automatic setting during feedstock compression with a specific
focus on ensuring a similar level of bulk density ρ of all produced bio-briquette samples. However, such
settings resulted in different lengths of final products. The device is equipped with a pressing chamber
and matrix with a diameter of 50 mm, thus producing bio-briquette samples (expressed in Figure 8) of
a diameter equal to approximately 50 mm; the dimensions of bio-briquette samples are provided in
Table 1.
224
Energies 2018, 11, 2186
(a) (b)
Figure 8. Produced bio-briquette fuel samples from: (a) bamboo fiber; (b) sugarcane skin.
Table 1. Basic mechanical properties of investigated briquette samples (±standard deviation).
Sample Weight (g) Height (mm) Diameter (mm)

Bamboo fiber 79.68 ± 20.71 37.73 ± 10.61 51.55 ± 0.37
Sugarcane skin 120.61 ± 12.85 53.35 ± 5.55 51.96 ± 5.40
In general, the dimensions of bio-briquette fuel influence its final mechanical quality.
Thus, its monitoring represents an important production parameter that needs to be considered.
2.3.2. Bulk Density ρ

The first investigated indicator, bulk density ρ, describes both the final mechanical quality of
produced bio-briquette samples, as well as the ability of tested feedstock materials to be densified.
Thus, it evaluates the efficiency of the densification process in the case of such specific feedstock
materials. Calculation of bio-briquette sample bulk density ρ (kg·m−3 ) used its dimensions, namely its
volume V (m3 ) and mass m (kg), and the following Equation (1).
m
ρ= (1)
V
where ρ is the bulk density (kg·m−3 ), V the bio-briquette samples’ volume (m3 ) and m the bio-briquette
sample mass (kg).
2.3.3. Mechanical Durability DU

The produced bio-briquette samples were subjected particularly to the determination of
mechanical durability DU (%), which is the main indicator of bio-briquette fuel mechanical quality
and which is defined by standard EN ISO 17831-2 (2015): Solid biofuels—Determination of mechanical
durability of pellets and briquettes—Part 2: Briquettes. This standard states several quality levels of
bio-briquette fuel and establishes the lowest acceptable level (DU > 90%) for commercial bio-briquette
fuel production. Experimental observation and measurements were performed repeatedly using an
electrically-powered special dust–proof rotating drum (Figure 9) with a rectangular steel partition.
The principle of such a test consists of projected impacts of tested bio-briquette samples in the rotating
drum; thereby, the samples prove their durability.
225
Energies 2018, 11, 2186
Engine
Depth: 598 mm
(a) (b)
Figure 9. Equipment used for the determination of mechanical durability: (a) in practice; (b) scheme.
In practice, several groups of bio-briquette samples of a specific weight (2 ± 0.1 kg) were separately
placed inside the drum and subjected to controlled impacts for a specific time (4 minutes 17 seconds),
which was equal to the specific number of rotations (105 ± 5 rotations). Subsequently, bio-briquette
samples were weighed prior to and after such testing, and their final resistance was determined by
Equation (2).
ma
DU = ·100 (2)
me
where DU is mechanical durability (%), me the sample weight before testing (g) and ma the sample
weight after testing (g).
2.3.4. Rupture Force RF

The methodology of the last tested mechanical quality indicator, the rupture force RF, does
not correspond to any mandatory technical standard. It was developed on the basis of previous
scientific research focused on the pressing technology and measurement of the physical and mechanical
properties of pressed items [36–38]. The stated methodology determined the hardness of tested
bio-briquette samples, thus simulating their possible damage in practice.
The principle of measurement lies in the plate-loading test (Figure 10), which was performed
using a universal hydraulic machine WPM, Type ZDM 5 (Leipzig, Germany) with a loading speed v
equal to 20 mm·min−1 . The measurement ended when bio-briquette samples disintegrated due to the
influence of force loading; thus, the maximal force loaded to the samples before their disintegration
was noted. The number of tested bio-briquette samples was equal to 64 in the case of bamboo fiber
and to 57 in the case of sugarcane skin.
226
Energies 2018, 11, 2186
F
Upper
pressure
plate
Bio–briquette
Bottom
pressure
F plate
(a) (b)
Figure 10. The equipment used for the determination of rupture force RF: (a) in practice; (b) plate,
loading test principle (F compressive force (N)).

The complex evaluation of the investigated bio-briquette production efficiency was based on the
set of quality indicators of both feedstock materials and produced bio-briquette samples. All chosen
indicators described the suitability of the investigated bio-briquette fuel for commercial production.
Therefore, required indicator levels occurred at a high level in an attempt to ensure the highest
quality of produced bio-briquette fuel. Nevertheless, mandatory requirements for solid biofuel
commercial production differ within each country; the obtained results were compared with the
European standards.
The present research aimed to answer the question of whether the chosen waste materials were
suitable for bio-briquette production. In general, it was a complex process containing several necessary
steps and could be influenced by many factors. Thus, before answering the main question posed,
the following specific issues must be answered: (a) Is there a sufficient amount of such waste materials?
Is there any functional existing way of its subsequent utilization? (b) Are those materials suitable for
direct combustion processes? Is their fuel analysis positive for such purposes? (c) Are the materials
suitable for drying and disintegration processes or do such steps involve difficulties? Is it possible to
produce bio-briquette fuel from those materials? What is the final mechanical quality of such biofuel?
Answers to all of these questions are noted in the following sections or were explained in the
Methodology Section.
3.1. Feasibility and Practicability of Production

The conditions of investigated materials production and potential reuse were primarily monitored
in the target area. Both materials originated from plants that abundantly grow and are processed in
the target area, with the only difference being that sugarcane plants are purposely cultivated as an
agriculture crop, while bamboo plants grow wildly throughout the target area. Specifically, processed
bamboo culms originated from the forests in the surrounding areas.
Both plants were nevertheless processed in the target area at a large-scale. Sugarcane plants
were used for juice production by individual sellers; plants stems were skinned before squeezing
of juice, which resulted in the production of sugarcane skin. Dried bamboo stems were manually
skinned by small storekeepers and subsequently used for the production of various commercial items.
As observed, both produced waste materials, bamboo fiber and sugarcane skin, were not reused
properly in the target area (Figure 11).
227
Energies 2018, 11, 2186
(a) (b)
Figure 11. Untapped waste biomass potential in the target area: (a) sugarcane skin in municipal waste;
(b) bamboo fiber in a fireplace.
Bamboo processing plants were located in rural areas. Thus, bamboo fiber was burned as waste
without any purpose in the fields or individuals used it as a fire lighter at a small-scale. Sugarcane plants
were processed directly in the streets of Huế city, thus leading to part of the produced sugarcane skin
ending up in garbage bins for municipal waste (Figure 11). It is worth mentioning that the generation
and amount of monitored waste materials were considerable and daily. In contrast, subsequent
utilization of materials was rarely observed.
Bamboo plant populations or their cultivation were difficult to monitor because prevalent parts
of the population are growing wildly. Focused on sugarcane plants, specific data of its production in
Vietnam were available and are shown in Table 2. Meanwhile, according to a statistical database of
the United Nations Food and Agriculture Organization (FAO), world sugarcane crop cultivation in
2016 provided the following data: harvested area 26,774,304 ha, production quantity 1,890,661,751 t
and yield 706,148 hg·ha−1 .
Table 2. Sugarcane production in Vietnam in 2010–2014.
Year Harvested Area (ha) Production Quantity (t) Yield (hg·ha−1 )

2016 256,322 16,313,145 636,432
2015 284,262 18,337,227 645,083
2014 304,969 19,822,851 649,996
2013 310,264 20,131,088 648,757
t, tonne; ha, hectare; hg, hectogram. Source: FAO, 2018.
As observed in this table, sugarcane production in Vietnam has occurred at a high level in recent
years. Thus, the production of sugarcane residues has also increased. Sugarcane plants are prevalently
used for production of juice or sugar in Vietnam, and thus, sugarcane skin represents a considerable
percentage ratio of produced waste biomass.
3.2. Fuel Analysis of Feedstock Materials

The present section describes the chemical parameters of the investigated materials, the required
levels of which must be respected in practice, otherwise their burning (in the form of bio-briquette
fuel) could cause environmental pollution. With respect to the data expressed in Table 3, it is clear
that moisture content Mc (%) during experimental measurements occurred at a suitable level for
bio-briquette production (i.e., Mc < 15%) [30]. Mentioned values of moisture content Mc do not
represent a material’s moisture content in their initial form (at the moment of creation or collection)
228
Energies 2018, 11, 2186
due to the necessary treatment within their transportation. However, as mentioned in the Methodology
Section, bamboo fiber was removed from already sundried culms; thus, their (required) lower level
of moisture content Mc was expected. Such a factor evaluated the amount of energy input positively
within the entire process of bio-briquette production, which in contrast to others, also contains feedstock
drying. Moreover, if moisture content Mc exceeds the suitable level, it complicates the densification
process or makes it completely impossible to realize it.
Table 3. Basic chemical parameters of waste biomass samples (in w.b.) (± standard deviation).
Sample Mc (%) Ac (%) GCV (MJ·kg−1 ) NCV (MJ·kg−1 )

Bamboo fiber 7.62 ± 0.47 1.16 ± 0.13 18.15 ± 0.03 16.92
Sugarcane skin 8.05 ± 0.52 8.62 ± 0.18 18.47 ± 0.50 17.23
Mc , moisture content; Ac , ash content; GCV, gross calorific value; NCV, net calorific value; w.b., wet basis.
When comparing the values of ash content Ac (%), it is clear that bamboo fiber exhibited a good
result, while sugarcane skin exhibited a result that was at the margin of tolerance (i.e., Ac < 10%) [30].
Such a result represents a disadvantage of sugarcane material; however, it is still accepted by the
mandatory requirements. Commonly, when a high level of ash content is detected, it can be caused by
contamination of material due to external impurities (dust, soil). Nevertheless, sugarcane skin samples
tested in the present research did not contain any impurities. Table 4 provides a comparison of fuel
property data obtained within the present research with data from other studies.
Table 4. Fuel properties of bamboo and sugarcane residues.
Ac (%) GCV (MJ·kg−1 ) Research

3.74 19.62 [39]
Bamboo culm
3.70 18.32 [40]
6.16 18.20 [41]
Sugarcane bagasse
11.27 17.33 [42]
Ac , ash content; GCV, gross calorific value.
A higher level of ash content Ac is commonly related to complications during biomass burning,
which can result in low burning efficiency or damage to burning devices. In general, both materials
represented herbaceous biomass, which prevalently exhibited a higher level of ash content Ac in
comparison to wood biomass [43,44]. However, the results of bamboo fiber samples were very good,
thus proving the significant potential of such material, which were qualitatively comparable with
wood biomass [45,46].
Energy potential expressed by calorific value is the most important indicator of fuel chemical
quality, which indicates the amount of energy released from fuel during burning [47]. The obtained
NVC data for both tested materials exhibited a high level of such an indicator; thus, these results were
satisfactory. For commercial sale, bio-briquette fuel must exhibit NVCs of at least 14.50 MJ·kg−1
for non-woody bio-briquette fuel and at least 15.00 MJ·kg−1 for woody bio-briquette fuel [30].
Every biomass kind exhibits different levels of calorific value, as well as the requirements on each
specific biomass kind of bio-briquette fuel differing (Table 5).
229
Energies 2018, 11, 2186
Table 5. Comparison of various biomass kind energy potential.
Biomass Kind Feedstock Material NCV (MJ·kg−1 ) Research

Vine pruning 19.19 [48]
Pine sawdust 18.14 [46]
Woody
Sycamore wood 15.62
[49,50]
Wild cherry wood 15.55
Banana leaf 17.76 [41]
Wheat straw 17.30 [45]
Herbaceous
Bamboo leaf 16.71 [51]
Rice straw 15.07 [41]
Grape seeds 20.39 [52]
Oil palm empty fruit bunch 18.16 [53]
Fruit
Banana peel 18.89 [54]
Durian peel 17.61 [55]
Aquatic Water hyacinth 16.65 [41]
Rice husk, oil palm sludge 21.68 [56]
Mixed Tropical hardwood sawdust 18.94
[57]
Rice straw, sugarcane leaves 17.83
NCV, net calorific value.
The results of elementary composition (Table 6) primarily proved the higher level of oxygen
content O (%) in the case of bamboo fiber, which is undesirable, while results for sugarcane skin were
at a satisfactory level. In general, biomass elementary composition can influence final calorific value of
produced biofuel, as well as influencing biofuel behavior during combustion. Higher levels of oxygen
O (%) influence consumption of air during biofuel burning and production of flue gas [58].
Table 6. Elementary composition of waste biomass samples (in w.b.) (±standard deviation).
Sample C (%) H (%) N (%) S (%) O (%)

Bamboo fiber 46.31 ± 0.14 6.19 ± 0.03 0.44 ± 0.02 0.04 ± 0.03 42.15
Sugarcane skin 45.85 ± 0.11 6.36 ± 0.02 0.96 ± 0.03 0.18 ± 0.01 34.02
C, carbon; H, hydrogen; N, nitrogen; S, sulfur; O, oxygen; w.b., wet basis.
Observed values were also converted into dry ash-free state to express results without the
influence of ash presence (Table 7).
Table 7. Composition of samples in dry ash-free state (d.a.f.).
Sample C (%) H (%) N (%) S (%) O (%) GCV (MJ·kg−1 ) NCV (MJ·kg−1 )
Bamboo Fiber 51.01 5.82 0.49 0.04 42.65 20.00 18.73
Sugarcane Skin 54.95 6.46 1.15 0.21 37.23 22.14 20.73
C, carbon; H, hydrogen; N, nitrogen; S, sulfur; O, oxygen; GCV, gross calorific value; NCV, net calorific value; d.a.f.,
dry ash-free state.
3.3. Mechanical Analysis of Briquette Samples

Both investigated materials proved their suitability for bio-briquette production. For the densification
process, it was possible to produce bio-briquette samples from them. Such a statement may not be a matter
of course because the success of bio-briquette sample production is not guaranteed in advance.
Specific data evaluating the efficiency of the densification process were observed immediately
following bio-briquette sample production. Thus, their bulk density ρ (kg·m−3 ) was stated. The results
of the first investigated mechanical quality indicator are noted in Table 8, together with the results of
other investigated indicators.
230
Energies 2018, 11, 2186
Table 8. Mechanical quality indicators of the investigated bio-briquette samples (±standard deviation).
Sample ρ (kg·m−3 ) DU (%) RF (N·mm−1 )

Bamboo fiber 986.37 ± 181.18 97.80 ± 0.04 143.30 ± 1.70
Sugarcane skin 1067.08 ± 39.08 97.70 ± 0.08 46.50 ± 0.80
ρ, bulk density; DU, mechanical durability; RF, rupture force.
Observed bulk density ρ data proved satisfactory results in both cases, which indicates high
quality bio-briquette fuel. As noted previously, bio-briquette fuel mechanical quality increases with
increasing bulk density ρ [49,59,60] because it indicates bio-briquette fuels’ longer burning time and
larger amount of produced heat [61]. According to other published research, the level of bulk density
ρ of high quality bio-briquette fuels should occur at approximately 1000 kg·m−3 [56,62,63].
The results of the next investigated indicator, bio-briquette sample DU (%), were obtained during
experimental laboratory testing. A statement regarding bio-briquette fuel DU (%), also termed abrasion
resistance, is necessary and required within commercial production and represents a major deciding
factor [49,62,64]; thus, its performance was a pivotal point of mechanical analysis. Data noted in
Table 8 expressed extremely good results; measured DU exceeded 97% for both materials, while the
mandatory required level is a DU ≥ 90% [31]. Figure 12 illustrates the conditions of bio-briquette
samples after abrasion testing.
(a) (b)
Figure 12. Bio-briquette fuel samples after mechanical durability testing: (a) bamboo fiber; (b)
sugarcane skin.
As observed in Figure 12, tested bio-briquette samples were in very good condition; only the
edges of their bodies were abraded. Such positive results were caused by the structural characteristics
of the materials, which were able to create strong bonds between particles.
Such excellent results correspond to the highest level of DU stated by related standards (i.e.,
DU ≥ 95%). Table 9 shows a comparison of the result values of other investigated bio-briquette fuels
sorted according to specific mechanical durability levels.
231
Energies 2018, 11, 2186
Table 9. Mechanical durability DU of different bio-briquette fuel kinds.
DU (%) Feedstock Material Research

Big bluestem sawdust
[59]
Canola straw
<90
Switchgrass
[65]
Corn stover
Pine sawdust [66]
Canary grass [67]
>90
Rice husk
[64]
Corncob
Eucalyptus sawdust,
[68]
paper
≥95 Soybean stalk [69]
Cotton stalk [70]
Digestate [71]
DU, mechanical durability.
The level of DU is prevalently influenced by the specifications of pressed feedstock materials,

but also, a considerable influence of the forming pressure used was proven. DU clearly increased with
increasing forming pressure applied by pressing the briquetting press into the feedstock material [65].
Applying such a process in practice could improve the unsuitable lower levels of other specific
bio-briquette’s DU.
The last monitored indicator was rupture force RF, which is not defined by any mandatory
standard. Thus, the evaluation of the observed results was performed only by comparison between
the investigated materials. Data in Table 9 indicate the marked differences observed. BF bio-briquette
samples had an RF = 143.30 N·mm−1 , while sugarcane skin bio-briquette samples exhibited an RF
equal to 46.50 N·mm−1 . The conditions of bio-briquette samples after RF testing are visible in Figure 13.
(a) (b)
Figure 13. Bio-briquette fuel samples after rupture force testing: (a) bamboo fiber; (b) sugarcane skin.
The considerably better strength of bamboo fiber bio-briquette samples was caused by the positive
behavior of fibers during pressing; above that, fibers made strong bonds already during feedstock
preparation. Thus, it can be concluded that fibrous materials have significant potential for bio-briquette
production. In addition, the overall evaluation of both tested materials proved satisfactory when
compared with other published data [72–74].
232
Energies 2018, 11, 2186
4. Conclusions
The main question investigated in the present research was to determine whether specific
agriculture waste residues, namely bamboo fibers and sugarcane skin, are suitable for bio-briquette
fuel production. The answer to this question, based on all observed results, is ‘Yes’. Both investigated
materials proved suitable as complete feedstocks for bio-briquette fuel production. Monitoring of their
production (quantity) and poor practice in terms of their reuse (in local conditions of the target area)
indicated high potential of such waste biomass for any further meaningful utilization.
Fuel parameter analysis found satisfactory levels of all tested quality indicators; the only identified
limitation was observed in the case of sugarcane skin ash content (Ac = 8.05%), but even that result
is acceptable. Comparatively good results were obtained during determination of bio-briquette fuel
mechanical quality indicators. Extremely high levels of mechanical durability (DU > 97%) strongly
supported the statement that utilization of investigated materials for bio-briquette fuel production will
result in high quality biofuel production.
In general, all measurements and tests performed within the present research provided satisfactory
results, and hence, the utilization of bamboo fiber and sugarcane skin for bio-briquette fuel production
can be highly recommended.
It is hoped that this research will also extend knowledge regarding appropriate waste management
principles and reuse of all potential waste biomass for the production of environmentally-friendly
solid biofuels.
Author Contributions: A.B. conceived of and designed the experiments, performed the sample collection and
experiments, analyzed the results and wrote the majority of the paper. H.R. performed the sample collection
and contributed to analyzing the results and to the writing of the paper. M.B. performed the experiments and
provided comments on the paper. All authors have read and approved the final manuscript.
Acknowledgments: This research was supported by the Internal Grant Agency of the Czech University Life
Sciences Prague 20173005 (31140/1313/3108) and by the Internal Grant Agency of the Faculty of Engineering,
Czech University of Life Sciences Prague, Number 2018:31140/1312/3111.
Nomenclature
Ac Ash content (%)
ρ Bulk density of bio-briquette samples (kg·m−3 )
C Carbon (%)
DU Mechanical durability (%)
F Compressive force (N)
GCV Gross calorific value (MJ·kg− 1 )
H Hydrogen (%)
m Mass of bio-briquette samples (kg)
ma Mass of bio-briquette samples after DU testing (g)
Mc Moisture content (%)
me Mass of bio-briquette samples before DU testing (g)
N Nitrogen (%)
NCV Net calorific value (MJ·kg− 1 )
O Oxygen (%)
RF Rupture force (N·mm− 1 )
S Sulfur (%)
V Volume of bio-briquette samples (m3 )
v Loading speed (mm·min− 1 )
233
Energies 2018, 11, 2186
References
1. Popp, J.; Lakner, Z.; Harangi-Rákos, M.; Fári, M. The effect of bioenergy expansion: Food, energy,
and environment. Renew. Sustain. Energy Rev. 2014, 32, 559–578. [CrossRef]
2. World Population Prospect. 2017. Available online: http://www.un.org/en/development/desa/
population/publications/index.shtml (accessed on 28 May 2018).
3. Owusu, P.A.; Asumadu-Sarkodie, S. A review of renewable energy sources, sustainability issues and climate
change mitigation. Cogent Eng. 2016, 3, 1167990. [CrossRef]
4. Mboumboue, E.; Njomo, D. Potential contribution of renewables to the improvement of living conditions of
poor rural households in developing countries: Cameroon’s case study. Renew. Sustain. Energy Rev. 2016, 61,
5. Liao, H.; Tang, X.; Wei, Y.M. Solid fuel use in rural China and its health effects. Renew. Sustain. Energy Rev.
2016, 60, 900–908. [CrossRef]
6. Liu, Z.; Mi, B.; Jiang, Z.; Fei, B.; Cai, Z.; Liu, X. Improved bulk density of bamboo pellets as biomass for
energy production. Renew. Energy 2016, 86, 1–7. [CrossRef]
7. Werther, J.; Saenger, M.; Hartge, E.-U.; Ogada, T.; Siagi, Z. Combustion of agricultural residues. Prog. Energy
Combust. Sci. 2000, 26, 1–27. [CrossRef]
8. Cherubini, F.; Peters, G.P.; Berntsen, T.; Stromman, A.H.; Hertwich, E. CO2 emissions from biomass
combustion for bioenergy: Atmospheric decay and contribution to global warming. Glob. Chang.
Biol. Bioenergy 2011, 3, 413–426. [CrossRef]
9. Ohrnberger, D. The Bamboos of the World: Annotated Nomenclature and Literature of the Species and the Higher and
Lower Taxa; Elsevier Press: New York, NY, USA, 1999.
10. Akinlabi, E.T.; Anane-Fenin, K.; Akwada, D.R. (Eds.) Bamboo: The Multipurpose Plant; Springer International
Publishing: Cham, Switzerland, 2017; pp. 1–37. Available online: https://www.springer.com/la/book/
9783319568072 (accessed on 6 July 2018).
11. Owen, A. Bamboo!! Improving island economy and resilience with Guam College students. J. Mar. Isl. Cult.
2015, 4, 65–75. [CrossRef]
12. Hofsetz, K.; Silva, M.A. Brazilian sugarcane bagasse: Energy and non-energy consumption. Biomass Bioenergy
2012, 46, 564–573. [CrossRef]
13. Pandey, A.; Soccol, C.R.; Nigam, P.; Soccol, V.T. Biotechnological potential of agro-industrial residues. I:
Sugarcane bagasse. Bioresour. Technol. 2000, 74, 69–80. [CrossRef]
14. Ramirez, J.A.; Brown, R.; Rainey, T.J. Techno-economic analysis of the thermal liquefaction of sugarcane
bagasse in ethanol to produce liquid fuels. Appl. Energy 2018, 224, 184–193. [CrossRef]
15. Xiong, W. Bagasse composites: A review of material preparation, attributes, and affecting factors.
J. Thermoplast. Compos. Mater. 2018, 31, 1112–1146. [CrossRef]
16. Brienzo, M.; Siqueira, A.F.; Milagres, A.M.F. Search for optimum conditions of sugarcane bagasse
hemicellulose extraction. Biochem. Eng. J. 2009, 46, 199–204. [CrossRef]
17. Jayapal, N.; Samanta, A.K.; Kolte, A.P.; Senani, S.; Sridhar, M.; Suresh, K.P.; Sampath, K.T. Value addition
to sugarcane bagasse: Xylan extraction and its process optimization for xylooligosaccharides production.
Ind. Crops Prod. 2013, 42, 14–24. [CrossRef]
18. Zwebe, D. Biomass Business Opportunities Viet Nam, SNV Netherlands Development Organization Vietnam.
2012. Available online: https://english.rvo.nl/sites/default/files/2013/12/Biomass_Opportunities_Viet_
Nam.pdf (accessed on 10 July 2018).
19. Schirmer, M. Biomass and waste as a renewable and sustainable energy source in Vietnam.
J. Vietnam. Environ. 2014, 6, 4–12. [CrossRef]
20. Roubík, H.; Mazancová, J.; Phung, L.D.; Dung, D.V. Quantification of biogas potential from livestock waste
in vietnam. Agron. Res. 2017. Available online: http://agronomy.emu.ee/wp-content/uploads/2017/05/
Vol15nr2_Roubik.pdf (accessed on 29 May 2018).
21. Maninder; Kathuria, R.S.; Grover, S. Using agricultural residues as a biomass briquetting: An alternative
source of energy. IOSR J. Electr. Electron. Eng. 2012, 1, 11–15. Available online: http://www.iosrjournals.org/
iosr-jeee/Papers/vol1-issue5/C0151115.pdf (accessed on 29 May 2018).
22. Brand, M.A.; Jacinto, R.C.; Antunes, R.; da Cunha, A.B. Production of briquettes as a tool to optimize the use
of waste from rice cultivation and industrial processing. Renew. Energy 2017, 111, 116–123. [CrossRef]
234
Energies 2018, 11, 2186
23. Chen, L.; Xing, L.; Han, L. Renewable energy from agro-residues in China: Solid biofuels and biomass
briquetting technology. Renew. Sustain. Energy Rev. 2009, 13, 2689–2695. [CrossRef]
24. Zhang, G.; Sun, Y.; Xu, Y. Review of briquette binders and briquetting mechanism. Renew. Sustain. Energy Rev.
2018, 82, 477–487. [CrossRef]
25. Arelano, G.M.T.; Kato, Y.S.; Bacani, F.T. Evaluation of Fuel Properties of Charcoal Briquettes Derived from
Combinations of Coconut Shell, Corn Cob and Sugarcane Bagasse. DLSU Research Congress, Manila,
Philippines. 2015. Available online: http://www.dlsu.edu.ph/conferences/dlsu_research_congress/2015/
proceedings/SEE/033-SEE_Bacani_FT.pdf (accessed on 12 July 2018).
26. EN 14918. Solid Biofuels—Determination of Calorific Value; ISO: Geneva, Switzerland, 2010.
27. ISO 1928. Solid Mineral Fuels—Determination of Gross Calorific Value by the Bomb Calorimetric Method,
and Calculation of Net Calorific Value; ISO: Geneva, Switzerland, 2010.
28. EN 15234-1. Solid Biofuels—Fuel Quality Assurance—Part 1: General Requirements; ISO: Geneva, Switzerland, 2011.
29. EN ISO 16559. Solid Biofuels—Terminology, Definitions and Descriptions; ISO: Geneva, Switzerland, 2014.
30. EN ISO 17225-1. Solid Biofuels—Fuel Specifications and Classes—Part 1: General Requirements; ISO: Geneva,
Switzerland, 2015.
31. EN ISO 17831-2. Solid Biofuels—Determination of Mechanical Durability of Pellets and Briquettes—Part 2:
Briquettes; ISO: Geneva, Switzerland, 2015.
32. ISO 18122. Solid Biofuels—Determination of Ash Content; ISO: Geneva, Switzerland, 2015.
33. EN 18134-2. Solid Biofuels—Determination of Moisture Content—Oven Dry Method—Part 2: Total
Moisture—Simplified Method; ISO: Geneva, Switzerland, 2015.
34. EN ISO 16948. Solid Biofuels—Determination of Total Content of Carbon, Hydrogen and Nitrogen; ISO: Geneva,
Switzerland, 2016.
35. EN ISO 18123. Solid Biofuels—Determination of the Content of Volatile Matter; ISO: Geneva, Switzerland, 2016.
36. Seifi, M.R. The moisture content effect on some Physical and Mechanical Properties of Corn. J. Agric. Sci.
2010, 2, 125–134. [CrossRef]
37. Altuntas, E.; Yıldız, M. Effect of moisture content on some physical and mechanical properties of faba bean
(Vicia faba L.) grains. J. Food Eng. 2007, 78, 174–183. [CrossRef]
38. Yahya, A.; Hamdan, K.; Ishola, T.A.; Suryanto, H. Physical and Mechanical Properties of Jatropha curcas L.
Fruits from Different Planting Densities. J. Appl. Sci. 2013, 13, 1004–1012. [CrossRef]
39. Dietenberger, R. Eigenschaftsuntersuchungen von Bambus Zur Energetischen Verwertung [Analysis of Bamboo
Characteristics for Energy Recovery]. Ph.D. Thesis, University of Freiburg, Freiburg, Germany, 2009.
40. Engler, B.; Schoenherr, S.; Zhong, Z.; Becker, G. Suitability of Bamboo as an Energy Resource: Analysis of
Bamboo Combustion Values Dependent on the Culm’s Age. Int. J. For. Eng. 2012, 23, 114–121. [CrossRef]
41. Phichai, K.; Pragrobpondee, P.; Khumpart, T.; Hirunpraditkoon, S. International Prediction Heating
Values of Lignocellulosics from Biomass Characteristics. J. Chem. Mol. Eng. 2013, 7, 532–535.
Available online: https://waset.org/Publications/prediction_heating_values_of_lignocellulosics_from_
biomass_characteristics/16408 (accessed on 12 July 2018).
42. Guar, S.; Reed, T.B. Thermal data for Natural and Synthetic Fuels; Marcel Dekker, Inc.: New York, NY, USA, 1998.
43. Gürdil, G.A.K.; Selvi, K.C.; Malat’ák, J.; Pinar, Y. Biomass Utilization for Thermal Energy. Am. Agric. Mech.
Asia Afr. Lat. Am. 2009, 40, 80–85. Available online: https://www.researchgate.net/publication/290695799_
Biomass_Utilization_for_Thermal_Energy (accessed on 10 May 2018).
44. Johansson, L.S.; Leckner, B.; Gustavsson, L.; Cooper, D.; Tullin, C.; Potter, A. Emission characteristics of
modern and old-type residential boilers fired with wood logs and wood pellets. Atmos. Environ. 2004, 38,
4183–4195. [CrossRef]
45. McKendry, P. Energy production from biomass (part 1): Overview of biomass. Bioresour. Technol. 2002, 83,
37–46. [CrossRef]
46. Stolarski, M.J.; Szczukowski, S.; Tworkowski, J.; Krzyzaniak, M.; Gulczyński, P.; Mleczek, M. Comparison of
Quality and Production Cost of Briquettes Made from Agricultural and Forest Origin Biomass. Renew. Energy
2013, 57, 20–26. [CrossRef]
235
Energies 2018, 11, 2186
47. Tumuluru, J.S.; Sokhansanj, S.; Wright, C.T.; Boardman, R.D.; Yancey, N. A Review on Biomass
Classification and Composition, Co-Firing Issues and Pretreatment Methods. 2011. Available online:
https://s3.amazonaws.com/academia.edu.documents/43368530/A_review_on_biomass_classification_
and_c20160304-21801-1dvq0xm.pdf?AWSAccessKeyId=AKIAIWOWYYGZ2Y53UL3A&Expires=
1531690192&Signature=qz4Y%2FadxiRqfCDsquKYyE6dKRDE%3D&response-content-disposition=
inline%3B%20filename%3DA_review_on_biomass_classification_and_c.pdf (accessed on 15 May 2018).
48. Nasser, R.A.; Salem, M.Z.M.; Al-Mefarrej, H.A.; Abdel-Aal, M.A.; Soliman, S.S. Fuel Characteristics
of Vine Prunings (Vitis vinifera L.) as a Potential Source for Energy Production. 2014.
Available online: https://bioresources.cnr.ncsu.edu/wp-content/uploads/2016/06/BioRes_09_1_482_
Nasser_SMAS_Fuel_Charac_Vine_Prunings_Alt_Wood_Source_4840.pdf (accessed on 11 July 2018).
49. Karunanithy, C.; Wang, Y.; Muthukumarappan, K.; Pugalendhi, S. Physiochemical characterization of
briquettes made from different feedstocks. Biotechnol. Res. Int. 2012, 165202. [CrossRef] [PubMed]
50. Telmo, C. and Lousada, J. The Explained Variation by Lignin and Extractive Contents on Higher Heating
Value of Wood. Biomass Bioenergy 2011, 35, 1663–1667. [CrossRef]
51. Brunerová, A.; Roubík, H.; Brožek, M.; Velebil, J. Agricultural residues in Indonesia and Vietnam and their
potential for direct combustion: With a focus on fruit processing and plantation crops. Agron. Res. 2018, 16,
52. Gravalos, I.; Xyradakis, P.; Kateris, D.; Gialamas, A.T. An Experimental Determination of Gross Calorific
Value of Different Agroforestry Species and Bio-Based Industry Residues. Nat. Resour. 2016, 7, 57–68.
[CrossRef]
53. Nasrin, A.B.; Choo, Y.M.; Lim, W.S.; Joseph, L.; Michael, S.; Rohaya, M.H.; Astimar, A.A.; Loh, S.K.
Briquetting of Empty Fruit Bunch Fibre and Palm Shell as a Renewable Energy Fuel. J. Eng. Appl. Sci. 2011,
6, 446–451. [CrossRef]
54. Wilaipon, P.; Trirattansirichai, K.; Tangchaichit, K. Moderate Die–Pressure Banana–Peel Briquettes. Int. J.
Renew. Energy 2007, 2, 53–58. Available online: http://www.sert.nu.ac.th/IIRE/V2N1(6).pdf (accessed on
10 July 2018).
55. Brunerová, A.; Roubík, H.; Brožek, M.; Herák, D.; Šleger, V.; Mazancová, J. Potential of Tropical Fruit Waste
Biomass for Production of Bio-Briquette Fuel: Using Indonesia as an Example. Energies 2017, 10, 2119.
[CrossRef]
56. Obi, O.F.; Okongwu, K.C.H. Characterization of fuel briquettes made from a blend of rice husk and palm oil
mill sludge. Biomass Convers. Biorefinery 2016, 6, 449–456. [CrossRef]
57. Jittabut, P. Physical and Thermal Properties of Briquette Fuels from Rice Straw and Sugarcane Leaves by
Mixing Molasses. Energy Procedia 2015, 79, 2–9. [CrossRef]
58. Demirbas, A. Combustion characteristics of different biomass fuels. Prog. Energy Combust. Sci. 2004, 30,
59. Tumuluru, J.S.; Tabil, L.G.; Song, Y.; Iroba, K.L.; Meda, V. Impact of process conditions on the density and
durability of wheat, oat, canola, and barley straw briquettes. Bioenergy Res. 2015, 8, 388–401. [CrossRef]
60. Lindley, J.A.; Vossoughi, M. Physical properties of biomass briquettes. Trans. ASABE 1989, 32, 361–366.
[CrossRef]
61. Obernberger, I.; Thek, G. Physical characterisation and chemical composition of densified biomass fuels with
regard to their combustion behaviour. Biomass Bioenergy 2004, 27, 653–669. [CrossRef]
62. Kaliyan, N.; Morey, R.V. Factors affecting strength and durability of densified biomass products.
Biomass Bioenergy 2009, 33, 337–359. [CrossRef]
63. Wakchaure, G.C.; Mani, I. Effect of binders and pressures on physical quality of some biomass briquettes.
J. Agric. Eng. 2009, 46, 24–30. Available online: https://www.researchgate.net/profile/G_C_Wakchaure/
publication/235944947_Effect_of_Binders_and_Pressures_on_Physical_Quality_of_Some_Biomass_
Briquettes/links/02e7e51499eca04887000000.pdf (accessed on 10 July 2018).
64. Muazu, R.I.; Stegemann, J.A. Effects of operating variables on durability of fuel briquettes from rice husks
and corn cobs. Fuel Process. Technol. 2015, 133, 137–145. [CrossRef]
65. Kaliyan, N.; Morey, R.V. Densification characteristics of corn stover and switchgrass. Trans. ASABE 2009, 52,
66. Li, Y.D.; Liu, H. High-pressure densification of wood residues to form an upgraded fuel. Biomass Bioenergy
2000, 19, 177–186. [CrossRef]
236
Energies 2018, 11, 2186
67. Repsa, E.; Kronbergs, E.; Pudans, E. Durability of compacted energy crop biomass. Eng. Rural. Dev. 2014,
13, 436–439. Available online: http://www.tf.llu.lv/conference/proceedings2014/Papers/74_Repsa_E.pdf
68. Wachira, G.G.; Gitau, A.N.; Kimani, M.W.; Njoroge, B.N.K. Mechanical Properties of Saw Dust Briquettes of
EucalyptusTree Species of Different Binders and Press Machines. Int. J. Emerg. Technol. Adv. Eng. 2015, 5,
532–538. Available online: https://pdfs.semanticscholar.org/100b/555644b59e9473181650d73509c067a7454a.pdf
69. Rajkumar, D.; Venkatachalam, P. Physical properties of agro residual briquettes produced from Cotton,
Soybean and Pigeon pea stalks. Int. J. Power Eng. Energy 2013, 4, 414–417. Available online: http://infomesr.
org/attachments/JO-P-0039.pdf (accessed on 10 July 2018).
70. Eissa, A.H.A.; Alghannam, A.R.O. Quality Characteristics for Agriculture Residues to Produce Briquette.
World Acad. Sci. Eng. Technol. 2013, 78, 166–171. Available online: https://www.researchgate.net/profile/
Ayman_Amer_Eissa/publication/257357240_WASET_AyA2013v78-29/links/0046352500e5a91346000000.pdf
71. Brunerová, A.; Pecen, J.; Brožek, M.; Ivanova, T. Mechanical durability of briquettes from digestate in
different storage conditions. Agron. Res. 2016, 14, 327–336. Available online: http://agronomy.emu.ee/wp-
content/uploads/2016/05/Vol14-_nr2_Brunerova.pdf#abstract-4144 (accessed on 12 July 2018).
72. Brunerová, A.; Brožek, M.; Müller, M. Utilization of waste biomass from post–harvest lines in the form of
briquettes for energy production. Agron. Res. 2017, 15, 344–358. Available online: http://agronomy.emu.ee/
wp-content/uploads/2017/05/Vol15nr2_Brunerova.pdf (accessed on 10 July 2018).
73. Brožek, M. Evaluation of selected properties of briquettes from recovered paper and board. Res. Agric. Eng.
2015, 61, 66–71. [CrossRef]
74. Brožek, M. The effect of moisture of the raw material on the properties briquettes for energy use. Acta Univ.
Agric. Silvic. Mendel. Brun. 2016, 64, 1453–1458. [CrossRef]
237
energies
Article
Biodiesel by Transesterification of Rapeseed Oil
Using Ultrasound: A Kinetic Study of
Base-Catalysed Reactions
José María Encinar 1, *, Ana Pardal 2 , Nuria Sánchez 1 and Sergio Nogales 1
1 Department of Chemical Engineering and Physical Chemistry, Extremadura University,
Avenida de Elvas s/n., 06071 Badajoz, Spain; [email protected] (N.S.); [email protected] (S.N.)
2 Department of Applied Sciences and Technology, IP Beja, Rua Pedro Soares s/n., 7800 Beja, Portugal;
[email protected]
Abstract: The objective of this work was to study the acceleration that ultrasound causes in the rate of
biodiesel transesterification reactions. The effect of different operating variables, such as ultrasound
power, catalyst (KOH) concentration and methanol:oil molar ratio, was studied. The evolution of
the process was followed by gas chromatography, determining the concentration of methyl esters at
different reaction times. The biodiesel was characterized by its density, viscosity, saponification and
iodine values, acidity index, water content, flash and combustion points, cetane index and cold filter
plugging point (CFPP), according to EN 14214 standard. High methyl ester yield and fast reaction
rates were obtained in short reaction times. Ultrasound power and catalyst concentration had a
positive effect on the yield and the reaction rate. The methanol:oil molar ratio also increased the yield
of the reaction, but negatively influenced the process rate. The reaction followed a pseudo-first order
kinetic model and the rate constants at several temperatures were determined. The activation energy
was also determined using the Arrhenius equation. The main conclusion of this work is that the use
of ultrasound irradiation did not require any additional heating, which could represent an energy
savings for biodiesel manufacture.
Keywords: fatty acid methyl ester; catalyst; viscosity; iodine value; acidity index
1. Introduction
The future development of world economy makes finding renewable sources of energy that can
replace fossil fuels necessary. For years, biodiesel has been a real alternative to fossil fuels used in
internal combustion engines [1]. As it is known, biodiesel is a fuel consisting of monoalkyl esters of
long-chain fatty acids (FAME) derived from renewable lipid feedstocks, and it is generally produced
via transesterification [2,3]. A lot of different raw materials have been used to obtain biodiesel. Edible
vegetable oils such as canola and soybean oil in the USA, palm oil in Malaysia or rapeseed oil in Europe
have been used for biodiesel production and found to be good substitutes for diesel. Non-edible
vegetable oils, such as Pongamia pinnata (karanja or honge), Jatropha curcas (jatropha or ratanjyote),
Citrus reticulata (mandarin) and Madhuca iondica (mahua) have also been found to be suitable for
biodiesel production [4–6]. Concerning the different types of vegetable oils and their composition,
fatty acids with high unsaturation levels usually imply lower fluidity at low temperatures, leading
to solidification. Thus, oils with a high ratio of monounsaturated fatty acid (and a low ratio of
polyunsaturated fatty acids) usually show good performance at low temperatures. As rapeseed oil has
such a composition, that is the reason why it is frequently used for biodiesel production in Europe.
Rapeseed oil usually contains around 98% of triglycerides. The main components of this oil are oleic

Energies 2018, 11, 2229
acid (monounsaturated), that is present in more than 60%, and linoleic acid (doubly unsaturated) that
exceeds 20% [4]. For these reasons, rapeseed oil was chosen to carry out this work.
Although there are other possibilities, transesterification has been reported as the most common
way to produce biodiesel from vegetable oil. Alcohols such as methanol or ethanol are the more
frequently used alcohols. A catalyst is necessary to increase the reaction rate and the conversion
yield. The catalysts used can be homogeneous and heterogeneous and they can be acid or basic in nature.
Generally, basic homogenous catalysis is the preferred option to provide high reaction rates. Additionally,
basic catalysts are readily available and very cheap. In the literature there is comprehensive information
about the catalytic processes involved [3,7–9].
The mechanism of the transesterification of vegetable oils by means of basic catalysis is well
known and published [10,11]. The mechanism comprises four stages. In the first step the base (catalyst)
reacts with the alcohol, giving an alkoxide and the protonated catalyst. The second step consists in
the nucleophilic attack of the alkoxide at the carbonyl group of the triglyceride, generating the alkyl
ester and the corresponding anion of the diglyceride (third step). For the final step, this diglyceride
deprotonates the catalyst, making it active and able to react with another alcohol, starting a new
transtererification cycle. Diglycerides and monoglycerides are equally converted (to a mixture of alkyl
esters and glycerol) by this mechanism.
The transesterification reaction is initially heterogeneous because methanol is only partially
miscible with triglycerides at room temperature. For this reason, the reaction initially is slow and
only takes place at the alcohol-oil interphase, and the process is dependent of mass transfer. When
the stirring rate is high, emulsions are usually generated. These emulsions are caused, mainly,
by the intermediate monoglycerides and diglycerides, which have both polar (hydroxyl groups) and
non-polar (hydrocarbon chains) parts. Therefore, when a critical concentration of these intermediates
is exceeded, emulsion takes place. Due to the low miscibility between methanol, vegetable oils
and methyl esters, this emulsion is not stable and breaks. The use of ultrasonic irradiation could
avoid this problem. Ultrasound produces special chemical and physical effects due to the collapse of
cavitation bubbles. Low frequency sound waves can produce emulsions between immiscible liquids,
being useful for the transesterification of triglycerides with alcohol [12–14]. Ultrasound has several
effects on transesterification: acoustic streaming mixing or changes in sound pressure, leading to fast
movement of fluids and cavitation bubbles that apply negative pressure gradients on liquids. The most
important effect is the formation and collapse of cavitation bubbles, providing high temperature
and pressure, with a significant influence on reaction rates, mass transfer and catalytic surface areas.
As a consequence, the use of sonochemical reactors can favor the chemical reaction and propagation
by way of enhanced mass transfer and interphase mixing between the phases and also can lower the
requirements of the operating conditions (in terms of temperature and pressure) [15].
The main advantages of ultrasonic irradiation process are: shorter reaction times, lower
alcohol/oil molar ratios, less energy consumption (50%), lower concentration of catalyst, higher
reaction rate and conversion, improved yield, simpler separation and purification processes, and higher
quality glycerol production [7,9,11,14].
In contrast, this process has some disadvantages such as: the reaction temperature was slightly
higher for long reactions and the ultrasonic power must be under control due to the possibility of
soap formation in fast reactions [16]. Nevertheless, as it is indicated, energetically the process is
highly favorable, because of the formation of micro jets and neither localized temperature increases
no agitation or heating are required to produce biodiesel with ultrasound application [17], therefore,
according to the literature, the beneficial effect of ultrasound is due to the generation of a fine emulsion
between methanol and fatty acids, increasing the surface area for the necessary chemical reactions.
No adverse effects, such as the generation of free radicals and the subsequent chain reactions caused
by them, were observed [18]. Thus, the use of ultrasound could imply an improvement in biodiesel
production, increasing the yield of the product.
239
Energies 2018, 11, 2229
Therefore, it could be said that ultrasound radiation provides energy to form a mixture and
overcome the energy activation barrier that is necessary for the process. On the other hand, high
temperatures and pressures can be achieved locally, which could imply an intense mixture between
the reagents. All these circumstances favor the reaction progress.
Considering the abovementioned facts, the objective of this work was to study the KOH-catalysed
transesterification reaction of rapeseed oil using an ultrasonic mixing technique, determining the
suitable reaction conditions to carry out the process in a short time and with low energy consumption.
Also, a kinetic study of the reaction in the presence of ultrasonic radiation was carried out, determining
the parameters necessary for the reactor design. In addition, the obtained results were compared to the
previously obtained ones in the transesterification of rapeseed oil in the absence of radiation [19,20].

The raw material (rapeseed oil) was provided by the Research Center “La Orden-Valdesequera”
(Badajoz, Spain), Section of Non-Edible Crops. The rapeseed oil was characterized by density, viscosity,
water content, acid, iodine and saponification values and fatty acid profile (Table 1).
Table 1. Rapeseed oil fatty acid profile and properties.
Fatty Acid Percentage (%)

C16:0 Palmitic 3.5
C18:0 Stearic 0.9
C18:1 Oleic 64.4
C18:2 Linoleic 22.3
C18:3 Linolenic 8.2
Other minority acids 0.7
Properties
Density at 15 ◦ C (kgm−3 ) 906.2
Viscosity at 40 ◦ C (cSt) 36.3
Water content (wt, %) 0.8
Acidity index (mgKOH ·goil −1 ) 2.7
Iodine value (gI2 ·100 g−1 ) 113.5
Saponification value (mgKOH ·goil −1 ) 194.7
Potassium hydroxide (KOH, pellets GR for analysis), used as a catalyst, was supplied by Merck
(Darmstadt, Germany) and methanol, 99%, was supplied by Panreac (Castellar del Vallés, Spain). The other
chemicals were obtained commercially (Merck) and were of analytical grade. The experimental design
is shown in Figure 1. The transesterification was carried out in a 500 mL spherical reactor, with a
temperature sensor, sampling outlet and condensation systems, using a sonicator (Digital Sonifier,
model 450, Branson, MO, USA), as shown in Figure 2. This sonicator has a fixed working frequency of
20 kHz and a power of 400 W, with adjustable levels for the latter between 0 and 100%.
Firstly, the reactor was charged with oil. Different amounts of catalyst were dissolved in different
amounts of methanol and the resulting solution was added to the reactor. At this point, the sonicator
was placed at the different power values and then the reaction started, taking place for 15–20 min.
The different reaction conditions are specified in Table 2.
240
Energies 2018, 11, 2229
Figure 1. Schematic flow diagram of the biodiesel production from rapeseed oil.
Figure 2. Schematic representation of the experimental setup.
Table 2. Experimental conditions. Properties of biodiesel produced from rapeseed oil and comparison
with the EN 14214 standard.
MeOH:Oil a KOH b Ultrasound Density 15 ◦ C, Viscosity

Run Tmax , (◦ C) c Yield, %
Molar Ratio (%) Power, %, (W) kg/m3 40 ◦ C, cSt
1 9:1 0.7 20, (80 W) 49 (±2) 55.6 (±1.1) 879.4 (±2.1) 8.8 (±0.3)
2 9:1 0.7 40, (160 W) 60 (±2) 86.6 (±0.9) 872.0 (±2.0) 4.9 (±0.2)
3 9:1 0.7 60, (240 W) 64 (±1) 92.0 (±0.7) 867.9 (±1.7) 4.8 (±0.1)
4 9:1 0.7 80, (320 W) 68 (±1) 94.9 (±0.9) 866.3 (±1.2) 4.4 (±0.2)
5 9:1 0.7 100, (400 W) 70 (±2) 96.6 (±0.8) 861.9 (±1.5) 4.7 (±0.2)
6 9:1 0.3 100, (400 W) 69 (±1) 41.1 (±1.6) 888.5 (±2.1) 13.9 (±0.5)
7 9:1 0.5 100, (400 W) 70 (±2) 94.0 (±1.0) 866.2 (±1.3) 4.8 (±0.2)
8 9:1 1.0 100, (400 W) 70 (±2) 93.6 (±1.0) 866.5 (±0.9) 4.4 (±0.1)
9 3:1 0.7 100, (400 W) 80 (±1) 70.8 (±1.3) 875.5 (±1.4) 6.9 (±0.2)
10 6:1 0.7 100, (400 W) 75 (±1) 97.4 (±1.2) 870.9 (±0.8) 4.9 (±0.2)
11 12:1 0.7 100, (400 W) 70 (±1) 98.0 (±1.1) 866.5 (±1.3) 4.4 (±0.1)
12 9:1 0.7 0, (0 W) 55 (±2) (ISO) 61.2 (±1.4) 872.0 (±1.2) 7.6 (±0.2)
EN-14214 - - - - 96.5 860–900 3.5–5.0
a The catalyst percentage is referred to the initial mass of oil; b The values in parentheses specify the applied power
in watts; c The yield is referred to the mass of methyl esters in the total biodiesel mass.
241
Energies 2018, 11, 2229
After an appropriate reaction time, the mixture was placed in a separatory funnel for 24 h to
ensure that the separation of biodiesel and glycerol phases was complete. The glycerol (bottom) phase
was removed and left in a container. Methyl esters (biodiesel) were heated, at 85 ◦ C to remove excess
methanol. The remaining catalyst was extracted by successive rinses with distilled water. Finally,
traces of water were eliminated by heating at 110 ◦ C.
The evolution of the process (methyl ester content) was followed by gas chromatography on a
VARIAN 3900 chromatograph (Varian, Palo Alto, California, USA), provided with an FID, employing
a silica capillary column of 30 m length, 0.32 mm ID, and 0.25 mm film thickness. Heptane was used
as a solvent, and the carrier gas was helium at a flow rate of 0.7 mL/min. The injector and the detector
temperatures were kept at 270 ◦ C and 300 ◦ C, respectively. The temperature ramp started at 200 ◦ C,
and then increased 20 ◦ C/min up to 220 ◦ C. The calibration curve of the peak areas versus the quantity
of biodiesel was linear. The samples were taken out from the reaction mixture, neutralized and heated
to remove methanol, centrifuged for 5 min at 6000 rpm, and then analyzed by gas chromatography.
Thus, the yield was referred to the mass of FAME in the total biodiesel mass, as follows (Equation (1)):
∑ mi
Yield (%) = ∗ 100 (1)
mT
where mi is the mass of each FAME, and mT is the mass of biodiesel, both in mg. The analytical
methods used to determine the characteristics of the biodiesel were basically those recommended
by the European Organization for Normalization (CEN), for its use in motor vehicles. Also the
recommended standards for the EN 14214 norm were employed. Details of the procedures used can be
found in previous papers [10,19]. Most experiments were done in triplicate, and standard deviations
are shown numerically in tables or as error bars in figures.

Before starting the study of the influence of variables, prior experiments in order to determine
the optimal position of the probe in the reaction medium were carried out. In these experiments the
methanol:oil molar ratio was 9:1, the concentration of catalyst (KOH) was 0.7% (w/w) and the power
or amplitude of the probe was fixed at 40%. Three different positions were investigated: the probe in
the oil phase, the probe in the methanol-oil interphase and the probe in the methanol phase. It was
observed that the degree of conversion is small when the probe was in the oil phase. The conversion
increased when the probe was placed in the methanol-oil interphase, reaching the top value when the
probe is in the methanol phase. These results can be due to the extent cavitational intensity generated
as a consequence of the presence of the probe in the oil phase, methanol phase or in the methanol-oil
interphase. These circumstances affected the physicochemical properties, mainly viscosity, surface
tension and density. As it is established in literature, methanol favors the generation of cavitation
conditions and, as a consequence, the maximum conversion was obtained in this case [15]. In view of
the above, the probe was placed in the methanol phase in all the subsequent experiments.
The generation of microturbulence for the cavitation bubbles located in the proximity of the
interphase methanol-oil originated an emulsion of the two liquids. The dispersion of methanol in the
oil depended on the intensity of the microturbulences that were generated by the cavitation bubbles in
methanol and vice versa. The intensity of the microturbulences depended on the physical properties of
the liquid medium such as density, viscosity and surface tension and also the amplitude of the acoustic
waves driving the bubble motion. The methanol and oil phases have different physical properties
and, therefore, the intensity of the microturbulences in the two phases was different. Consequently,
the extension of the dispersion of methanol in oil cannot be the same as the dispersion of oil in
methanol. The uniform dispersion of methanol in the oil phase originated the necessary interfacial
area for the reaction to take place. This dispersion was produced, as it has been indicated, by the high
level of microturbulences generated by the cavitation of the bubbles of methanol in the proximity with
242
Energies 2018, 11, 2229
the interphase and, as it was manifested by Kalva [18], the dispersion increased with the methanol:oil
molar ratio. As a consequence, the influence of this variable on the process should be notable.
Another aspect to consider is the reaction temperature. In the experiments, heating was not
applied, but it was verified that the reaction temperature increased, due to irradiation, achieving
maximum temperatures of up to 80 ◦ C (run 9). Naturally, the experimental results showed that,
for low temperatures, the extension of the conversion was also low and when temperature increased,
the conversion also increased, showing, to a certain extent, a positive impact of temperature. But it is
necessary to consider other factors. An increase in temperature implied an increase of the solubility
of methanol in the oil phase and, also, an increase of the reaction rate. These two factors contributed
to increase the conversion of the reaction. But at high temperatures, the extension of the cavitational
effects was dampened and, on the other hand, methanol leaked of the reaction medium to surpass the
boiling temperature. Hence, there was an optimal temperature that is dependent on all the previous
aspects [15].
3.1. Effect of Ultrasound Power in Methyl Ester Conversion

In order to reduce the production costs of biodiesel it is necessary to optimize the amount of
energy supplied to the reaction mixture. The objective was to obtain a maximum yield and a high
formation rate with the lowest energy consumption. In this sense, it is known that when the intensity
of the radiation (i.e., ultrasound power/irradiation area) increased, a more violent collapse of the
bubbles of cavitation took place, producing a bigger mixing intensity in the methanol-oil interphase.
These circumstances originated the formation of a very fine emulsion that favored the mass-transfer
coefficient, causing a high formation of biodiesel [7].
Power had an influence on the size and number of bubbles, maximum live time, pressure that
break these bubbles, elevation of the temperature inside the liquid, generation of cavitational effects
and over the final collapse intensity. Also the intensity blending depended on the density of energy.
Thus, generally, high levels of power dissipation promoted cavitational effects, but at times, these
depended on the geometry of the reactor. For this reason, an optimal value of power dissipation
is often observed due to phenomena of acoustic disengaging [21]. As a result, for high powers,
a damping effect, which originates a decrease in the energy transfer and a low cavitational activity,
was observed [15].
The results obtained in the study of the ultrasonic power are shown in Figure 3. The power was
varied from 20 to 100% (experiments 1 to 5). Also, in order to draw a comparison, an experiment
(run 12), carried out at 55 ◦ C in isothermal conditions, with 0% of ultrasonic power, is enclosed. Also,
in order to draw a comparison, an experiment (run 12), carried out at 55 ◦ C in isothermal conditions,
with 0% ultrasonic power, is enclosed. In this experiment, a heating power of 500 W was supplied
in order to get the indicated conditions. The conditions of this experiment were chosen considering
that are standard and common conditions of transesterification processes carried out by conventional
heating methods. There was a positive effect of power, so that, for a given reaction time, the conversion
was greater as the power was higher. The curves showed an induction period that diminished when
power increased. Differences in conversion and induction times were smaller as the power level
increased so, for powers above 60%, respective curves tended to an overlap situation. Thus, in the
current situation, it took 15 min to achieve very high conversions (around 90%), for the experiments
carried out with an ultrasound power between 160 and 400 W. An aspect that should be considered
is the little variation of the maximum conversion with the applied power. In effect, runs 3, 4 and 5
lead to similar conversions even though the applied power differed considerably (60 to 100%). In this
sense, the observations of Shing et al. [12] should be considered, showing that at higher amplitudes,
ester yields were drastically reduced. This was attributable to cracking followed by FAME oxidation to
aldehydes, ketones, and lower-chained organic fractions.
243
Energies 2018, 11, 2229
Figure 3. Effect of ultrasound power on the extent of conversion (MeOH:Oil 9:1, 0.7% KOH).
In addition, Figure 4 shows the evolution of temperature with reaction time for ultrasound
powers between 40 and 100%. In all cases the same trend was observed: in the first minutes of reaction,
the temperature increased very quickly, and later it stabilized at a constant value. The time required to
achieve this constant value was between 3 and 4 min, and the maximum values of temperature were
proportional to the applied power (60, 64, 68 and 70 ◦ C for 40, 60, 80 and 100% powers respectively).
Therefore, a balance between the heat generated as a consequence of the ultrasound radiation and
the heat lost through the walls of the reactor took place. The result is a process that occurs near the
isothermal regime. Another aspect to consider is the energy consumption. Ultrasonic powers of
160–240 W gave 90% yields. These results are in accordance with other research data, with similar
power ranges and over 90% yields [22]. However, thermal powers of 500 W only produced a 55% yield.
This implies a considerable energy savings.
Figure 4. Temperature vs time. Effect of ultrasound power (methanol:oil 9:1, 0.7% KOH).
3.2. Effect of Catalyst Amount on Methyl Ester Conversion

Figure 5 shows the influence of the catalyst amount (run 5 to 8). The positive effect of this variable
on the reaction rate of the process can be easily seen.
244
Energies 2018, 11, 2229
Figure 5. Effect of catalyst concentration on conversion (MeOH:Oil 9:1, 100% ultrasound power).
For very low concentrations, there was a long induction period. At high concentrations this
induction period did not exist. For concentrations of catalyst under 0.7%, the conversion increased
with concentration. Higher values of concentration lead to very similar conversions (run 5 and 8). This
resulted in catalyst savings, which is positive from an economic point of view, and also facilitated
the final product washing and/or the recovery of the catalyst, which is positive from a technical and
environmental point of view.
Like in other works, it is possible to see two areas. In the first one, which can extend up to
7–12 min, a positive influence on the amount of catalyst on the rate of formation of methyl esters
was observed. Indeed, as the amount of catalyst increased, it reached peak production sooner, that is,
the asymptotic zone of the curve. The second area of the curve (the asymptote) was very similar in all
cases (except for 0.3%).
Except for the experiment carried out with 0.3% catalyst (run 6) the short time necessary to
achieve a maximum conversion was noticeable. In this sense, the ultrasound helps reduce the amount
of catalyst needed due to the chemical activity due to cavitation. Thus, the necessary quantity of
catalyst to achieve a given conversion was smaller when the reaction was carried out in the presence
of ultrasound. A benefit was the resulting increase of the glycerol purity [23].
3.3. Effect of Methanol:Oil Ratio on Methyl Ester Conversion

One of the most important variables affecting the yield of esters is the alcohol:triglyceride molar
ratio. As it is known, the stoichiometric ratio for transesterification requires three moles of alcohol
and one mole of triglyceride to yield three moles of fatty acid alkyl esters and one mole of glycerol.
However, transesterification is an equilibrium reaction in which a large excess of alcohol is required to
drive the reaction to the right (products). However, the high molar ratio of alcohol to vegetable oil
interferes with the separation of glycerin because there was an increase in solubility. When glycerin
remained in solution, it helped to drive the equilibrium to the left (reactive), lowering the yield of
esters [24]. Consequently, the methanol:oil molar ratio is a variable that should be optimized.
Figure 6 shows the results obtained (experiments 5, 9 to 11). Competing considerations are
observed. On one hand, the higher initial reaction rate happened in the experiment with the lowest
ratio of 3:1 and the lowest in the experiment with 12:1. On the other hand, the maximum conversion
achieved was superior with increasing the methanol: oil molar ratio.
245
Energies 2018, 11, 2229
Figure 6. Effect of methanol:oil molar ratio on the extent of conversion (0.7% KOH, 100% ultrasound power).
This fact suggests that oil and methanol easily became an emulsion when there was a small
amount of methanol, resulting in a higher rate of methanolysis. This implies that, under ultrasonic
irradiation, small droplets of methanol were generated rapidly and easily achieved emulsion formation
in the oil phase, leading to quick emulsion formation. In contrast, the droplet size of the emulsion was
greater when there was a higher amount of methanol, and this resulted in a slower initial reaction rate.
Furthermore, when there was a larger excess of methanol, the probability of the small size droplets
of methanol encountering each other to agglomerate into larger droplets was larger than when there
was a smaller excess of methanol. This trend towards large droplet formation may result in a slow
transesterification reaction rate [25].
With an increase in the molar ratio, the quantity of methanol in the reaction mixture increased,
which mainly affected the cavitation intensity. Excess of methanol provided additional cavitation
events in the reactor, leading to the formation of enhanced emulsion quality (smaller drop sizes),
providing additional area for the reaction and hence increased conversion [26].
Also, in connection with the previous figure, it can be noted that in the experiment with a molar
ratio of 3:1, the reaction was stopped at 9 min due to the fast and excessive increase in temperature
(80 ◦ C at the end), which made the control of this variable difficult. In addition to the above, the low
effect of temperature in ultrasonic system is related to the collapse of bubbles caused by cavitations
that produced intense local heating and high pressures, with very short lifetimes, which have a much
higher effect than elevating the temperature of the liquid media. Furthermore, temperature affected
the vapor pressure, surface tension, and viscosity of the liquid medium. While higher temperature
increased the number of bubble cavitations, the collapse of bubbles was cushioned by the higher vapor
pressure, reducing the effect of ultrasound application in the reaction [27].
3.4. Biodiesel Properties

Table 2 shows the yield, density and viscosity obtained for all the experiments. Table 3 shows
other parameters of the biodiesel corresponding with the two experiments with the highest yield in
methyl esters. For comparison, in Tables 2 and 3, the values of the EN-14214 standard have been
added. As it can be seen in Table 2, only in some experiments, it was possible to achieve the minimum
conversion to methyl esters, required by the standard EN 14214. However, most of the experiments
reached values close to the requirements. Only in experiments in which the applied power, or catalyst
concentration, or the molar ratio of methanol: oil were minimal, was the conversion low.
246
Energies 2018, 11, 2229
Table 3. Properties of biodiesel produced from rapeseed oil in the best conditions of this study.
Parameter Run 5 Run 10 EN-14214

FAME content, % 96.6 (±0.8) 97.4 (±1.2) 96.5
Water content, % 0.06 (±0.01) 0.06 (±0.01) <0.05
Saponification value, gKOH ·goil −1 190.6 (±0.8) 188.9 (±0.6)
Iodine value, wt % 106.9 (±0.3) 106.1 (±0.1) ≤120
Acidity Index, mgKOH ·goil −1 0.49 (±0.02) 0.48 (±0.01) ≤0.5
CFPP, ◦ C −7 (±1) −7 (±1)
Flash point, ◦ C 178 (±1) 175 (±1) ≥120
Combustion point, ◦ C 185 (±1) 189 (±1)
Cetane index 44.1 (±0.2) 43.8 (±0.3)
Density (kgm−3 ) 861.9 (±1.1) 870.9 (±0.8) 860–900
Viscosity at 40 ◦ C (cSt) 4.7 (±0.2) 4.9 (±0.2) 3.5–5.0
In relation to density, the same ranges between 861.9 and 888.5 kg/m3 were found, within
the limits specified by the standard. Viscosity varied between 4.4 and 13.9 cSt. Except for the
three experiments with low conversion, the presented viscosity values are in accordance with the
standard. One overall effect was the relationship between conversion and viscosity: the more viscosity
the less conversion. Consequently, most of the biodiesels obtained complied with the EN-14214
specifications, especially in those cases where the conversions obtained were high. For the other
parameters (see Table 3), no significant differences were observed when high conversions were reached.
Furthermore, it can be seen that, in general, the values of the standard were met.
3.5. Kinetic Study

As it has been indicated previously, the physical mechanism responsible for the beneficial
action of ultrasound is the formation of a fine emulsion between oil and methanol that enhances
the interface area for the reaction. No chemical effect of ultrasound, i.e., production of radical
species and induction/acceleration of the reaction by these species, seemed to play any role [18].
In addition, the ultrasound radiation allowed one to get high temperature process in a short time.
These temperatures (Figure 4) reached a maximum value (3 or 4 min) and at a later time they were
constant. Therefore, the reaction, in the presence of ultrasound radiation, was similar to an isothermal
regime process.
Besides, the transesterification reaction was complex, because secondary reactions of saponification
or neutralization can happen. Additionally, the initial heterogeneous character of the reaction made the
process difficult [28,29]. In our case, rapeseed oil was refined, and the free acidity was small, that is
why saponification or neutralization reactions did not happen. The concentration of the catalyst did
not change to get the asymptotic zone of the curves, and we assumed that the reaction-formations
of monoglycerides, diglycerides and triglycerides were very fast and that there was no intermediate
of reaction. That is, the transesterification reaction was considered as an only and global reaction,
therefore, considering the above and in line with the literature [30], the reaction would be represented
by Equation (2), where TG is the triglyceride, MeOH is methanol, ME is methyl-esters and G is glycerin.
TG + 3MeOH ↔ 3ME + G (2)
The reaction (Equation (2)) is reversible, but in practice an excess of methanol was used and it can
be considered irreversible. In addition, the kinetic study only was carried out until the curve showed
the asymptotic trend. Under these conditions the inverse reaction lacks importance. The reaction rate
is expressed by Equation (3), where α and β are the orders of reaction in relation to triglycerides and
methanol respectively, k’ is a constant that includes the effect of the catalyst, and t is the reaction time:
−d[TG]/dt = k’[TG]α [MeOH]β (3)
247
Energies 2018, 11, 2229
Equation (3) can be rewritten like Equation (4), where k regroups k’ and the concentration
of methanol. This concentration can be considered constant because a very large concentration of
methanol was used:
−d[TG]/dt = k[TG]α (4)
Expressing the concentration of triglycerides in terms of conversion (X), Equation (4) takes the
appearance of Equation (5), where TG0 represents the initial concentration of triglycerides:
dX/dt = k[TG0 ](1 − X)α (5)
The integration of (5), with the most widely-used hypothesis in literature (pseudo first order
kinetic model), that is, with α = 1, leads to Equation (6):
ln(1 − X) = −kt (6)
Equation (1) has been applied to Experiments 2, 3, 4 and 5, in which the ultrasound power varied.
In these experiments, the methanol:oil molar ratio was always 9:1 and the catalyst concentration 0.7%.
In Table 4, the pseudo kinetic constants and R2 coefficient, obtained by means of regression analysis of
Equation (1) are shown.
Table 4. Kinetic analysis of the transesterification process.
Temperature, ◦ C Ultrasound Power (%) Pseudo First-Order Kinetic, min−1 R2

60 40 0.1669 0.98
64 60 0.1872 0.98
68 80 0.2103 0.98
70 100 0.2170 0.99
The temperature shown for each experiment was the temperature that the system reached after
3 or 4 min (final temperature). The high values of R2 confirmed that this system followed a pseudo-
first kinetic model. The relationship among the specific reaction rate constant (k), absolute temperature
(T) and activation energy (Ea) is given by Arrhenius Equation (7), where A is the frequency factor and
R is the universal gas constant:
k = Aexp (−Ea/RT) (7)
The linear regression of Equation (7) provides the value of the activation energy. The best fit is
shown in Equation (8). These data lead to a value of Ea = 25.51 kJ/mol. This value was relatively small,
which reveals a great catalytic activity of the basic catalysts (KOH):
lnk = −3068.5/T + 7.428; R2 = 0.993 (8)
The activation energy was similar to that obtained in rapeseed oil transesterification in the
presence of co-solvents (21.88 kJ/mol) [31] and slightly lower than in the case of Jatropha curcas
transesterification by ultrasound (31.29 and 57.33 kJ/mol) [32]. On the other hand, the activation
energy in other catalysis-base transesterification through thermal heat were superior. For instance,
for palm oil transesterification an activation energy of 105 kJ/mol was found [33].
4. Conclusions
The suitability of the use of ultrasound for rapeseed oil transesterification was studied. The main
findings of this research were:
Ultrasonic irradiation facilitated the rapeseed transesterification, since high biodiesel yields
were achieved after short reaction times (20 min). It was observed, during experimentation, that the
temperature of the reaction mixture increased, as a result of ultrasound radiation.
248
Energies 2018, 11, 2229
The use of ultrasound radiation did not need additional heating, what can suppose an energy
savings. Nevertheless, it is necessary to consider the energy consumed in the generation of ultrasound.
The influence of the ultrasound power and the catalyst concentration was positive, so the yield of
the process and the reaction rate increased as these variables increased. The methanol:oil molar ratio
also lead to a bigger yield of the process, but the lower values of the methanol:oil molar ratio lead to
the biggest reaction rate.
The obtained results showed that ultrasonic powers of 320 W, a catalyst concentration of 0.7%
and a methanol:oil molar ratio of 9:1 are enough to achieve biodiesel yields of 95%. These conditions
can be considered adequate to carry out the process. The characteristics of the biodiesel, determined
by the EN 14214 standard, revealed that, in general, they met the established limits. The final product
had similar properties to a diesel-oil.
The transesterification reactions followed a pseudo-first order kinetic model and the rate constants
at several temperatures were determined. Also, the activation energy was determined by the
Arrhenius equation.
Author Contributions: Conceptualization, J.M.E., A.P. and N.S.; Methodology, J.M.E., A.P., N.S. and S.N.;
Validation, J.M.E., A.P., N.S. and S.N.; Formal Analysis, A.P. and N.S.; Investigation, A.P. and N.S.; Resources,
J.M.E., A.P., N.S. and S.N.; Data Curation, J.M.E., A.P., N.S. and S.N.; Writing-Original Draft Preparation, J.M.E.,
A.P., N.S. and S.N.; Writing-Review & Editing, J.M.E. and S.N.; Supervision, J.M.E. and S.N.
Acknowledgments: The authors would like to thank the “Junta de Extremadura” (“Ayudas para la realización de
actividades de investigación y desarrollo tecnológico, de divulgación y de transferencia de conocimiento por los
Grupos de Investigación de Extremadura”) and the FEDER “Fondos Europeos de Desarrollo Regional” for the
financial support received.
References
1. Caldeira-Pires, A.; da Luz, S.M.; Palma-Rojas, S.; Oliveira Rodrigues, T.; Chaves Silverio, V.; Vilela, F.;
Barbosa, P.C.; Alves, A.M. Sustainability of the biorefinery industry for fuel production. Energies 2013, 6,
2. Guo, M.; Song, M.; Buhain, J. Bioenergy and biofuels: History, status, and perspective. Renew. Sustain. Energy Rev.
2015, 42, 712–725. [CrossRef]
3. Ma, F.; Hanna, M.A. Biodiesel production: A review. Bioresour. Technol. 1999, 70, 1–15. [CrossRef]
4. Pinzi, S.; Garcia, I.L.; Lopez-Gimenez, F.J.; Luque de Castro, M.D.; Dorado, G.; Dorado, M.P. The Ideal
Vegetable Oil-based Biodiesel Composition: A Review of Social, Economical and Technical Implications.
Energy Fuel 2009, 23, 2325–2341. [CrossRef]
5. Jurac, Z.; Zlatar, V. Optimization of raw material mixtures in the production of biodiesel from vegetable and
used frying oils regarding quality requirements in terms of cold flow properties. Fuel Proc. Technol. 2013, 106,
6. Azad, A.K. Biodiesel from Mandarin Seed Oil: A surprising source of alternative fuel. Energies 2017, 10, 1689.
[CrossRef]
7. Mahamuni, N.N.; Adewuyi, Y.G. Optimization of the Synthesis of Biodiesel via Ultrasound-Enhanced
Base-Catalyzed Transesterification of Soybean Oil Using a Multifrequency Ultrasonic Reactor. Energy Fuels
2009, 23, 2757–2766. [CrossRef]
8. Demirbas, A. Progress and recent trends in biodiesel fuels. Energy Convers. Manag. 2009, 50, 14–34. [CrossRef]
9. Ramachandran, K.; Suganya, T.; Nagendra Gandhi, N.; Renganathan, S. Recent developments for biodiesel
production by ultrasonic assist transesterification using different heterogeneous catalyst: A review.
10. Encinar, J.M.; González, J.F.; Pardal, A. Transesterification of castor oil under ultrasonic irradiation conditions.
Preliminary results. Fuel Proc. Technol. 2012, 103, 9–15. [CrossRef]
11. Singh, A.K.; Fernando, S.D.; Hernandez, R. Base-Catalyzed Fast Transesterification of Soybean Oil Using
Ultrasonication. Energy Fuels 2007, 21, 1161–1164. [CrossRef]
12. Hanh, H.D.; Dong, N.T.; Okitsu, K.; Nishimura, R.; Maeda, Y. Biodiesel production through transesterification
of triolein with various alcohols in an ultrasonic field. Renew. Energy 2009, 34, 766–768. [CrossRef]
249
Energies 2018, 11, 2229
13. Ji, J.; Wang, J.; Li, Y.; Yu, Y.; Xu, Z. Preparation of biodiesel with the help of ultrasonic and hydrodynamic
cavitation. Ultrasonics 2006, 44, e411–e414. [CrossRef] [PubMed]
14. Veljković, V.B.; Avramović, J.M.; Stamenković, O.S. Biodiesel production by ultrasound-assisted transesterification:
State of the art and the perspectives. Renew. Sustain. Energy Rev. 2012, 16, 1193–1209. [CrossRef]
15. Hingu, S.M.; Gogate, P.R.; Rathod, V.K. Synthesis of biodiesel from waste cooking oil using sonochemical
reactors. Ultrason. Sonochem. 2010, 17, 827–832. [CrossRef] [PubMed]
16. Talebian-Kiakalaieh, A.; Amin, N.A.S.; Mazaheri, H. A review on novel processes of biodiesel production
from waste cooking oil. Appl. Energy 2013, 104, 683–710. [CrossRef]
17. Santos, F.F.P.; Rodrigues, S.; Fernandes, F.A.N. Optimization of the production of biodiesel from soybean oil
by ultrasound assisted methanolysis. Fuel Proc. Technol. 2009, 90, 312–316. [CrossRef]
18. Kalva, A.; Sivasankar, T.; Moholkar, V.S. Physical mechanism of ultrasound-assisted synthesis of biodiesel.
Ind. Eng. Chem. Res. 2009, 48, 534–544. [CrossRef]
19. Encinar, J.M.; Pardal, A.; Martínez, G. Transesterification of rapeseed oil in subcritical methanol conditions.
20. Martínez, G.; Sánchez, N.; Encinar, J.M.; González, J.F. Fuel properties of biodiesel from vegetable oils and
mixtures. Influence of methyl esters distribution. Biomass Bioenergy 2014, 63, 22–32. [CrossRef]
21. Gole, V.L.; Gogate, P.R. A review on intensification of synthesis of biodiesel from sustainable feed stock
using sonochemical reactors. Chem. Eng. Process. 2012, 53, 1–9. [CrossRef]
22. Sheng Ho, W.W.; Kiat Ng, H.; Gan, S. Advances in ultrasound-assisted transesterification for biodiesel
production. Appl. Therm. Eng. 2016, 100, 553–563.
23. Kumar, D.; Kumar, G.; Poonam; Singh, C.P. Fast, easy ethanolysis of coconut oil for biodiesel production
assisted by ultrasonication. Ultrason. Sonochem. 2010, 17, 555–559. [CrossRef] [PubMed]
24. Meher, L.C.; Vidya Sagar, D.; Naik, S.N. Technical aspects of biodiesel production by
transesterification—A review. Renew. Sustain. Energy Rev. 2006, 10, 248–268. [CrossRef]
25. Thanh, L.T.; Okitsu, K.; Sadanaga, Y.; Takenaka, N.; Maeda, Y.; Bandow, H. Ultrasound-assisted production
of biodiesel fuel from vegetable oils in a small scale circulation process. Bioresour. Technol. 2010, 101, 639–645.
[CrossRef] [PubMed]
26. Deshmane, V.G.; Gogate, P.R.; Pandit, A.B. Ultrasound-Assisted Synthesis of Biodiesel from Palm Fatty Acid
Distillate. Ind. Eng. Chem. Res. 2009, 48, 7923–7927. [CrossRef]
27. Santos, F.F.P.; Malveira, J.Q.; Cruz, M.G.A.; Fernandes, F.A.N. Production of biodiesel by ultrasound assisted
esterification of Oreochromis niloticus oil. Fuel 2010, 89, 275–279. [CrossRef]
28. Veljkovic, V.B.; Stamenkovic, O.S.; Todorovic, Z.B.; Lazic, M.L.; Skala, D.U. Kinetics of sunflower oil
methanolysis catalyzed by calcium oxide. Fuel 2009, 88, 1554–1562. [CrossRef]
29. Stamenkovic, O.S.; Todorovic, Z.B.; Lazic, M.L.; Veljkovic, V.B.; Skala, D.U. Kinetics of sunflower oil
methanolysis at low temperatures. Bioresour. Technol. 2008, 99, 1131–1140. [CrossRef] [PubMed]
30. Ramezami, K.; Rowshanzamir, S.; Eikani, M.H. Castor oil transesterification reaction: A kinetic study and
optimization of parameters. Energy 2010, 35, 4142–4148. [CrossRef]
31. Encinar, J.M.; Pardal, A.; Sánchez, N. An improvement to the transesterification process by the use of
co-solvents toproduce biodiesel. Fuel 2016, 166, 51–58. [CrossRef]
32. Choudhury, H.A.; Srivastava, P.; Moholkar, V.S. Single-step ultrasonic synthesis of biodiesel from crude
Jatropha curcas oil. AIChE J. 2014, 60, 1572–1581. [CrossRef]
33. Permsuwan, A.; Tippayawong, N.; Tanongkiat, T.; Thararux, C.; Wangkarn, S. Reaction Kinetics of
Transesterification between Palm Oil and Methanol under Subcritical Conditions. Energy Sci. Technol.
2011, 2, 35–42.
250
energies
Article
A Study of the Production and Combustion
Characteristics of Pyrolytic Oil from Sewage Sludge
Using the Taguchi Method
Guan-Bang Chen 1, *, Jia-Wen Li 2 , Hsien-Tsung Lin 2 , Fang-Hsien Wu 2 and Yei-Chin Chao 2, *
1 Research Center for Energy Technology and Strategy, National Cheng Kung University, Tainan 701, Taiwan
2 Department of Aeronautics and Astronautics, National Cheng Kung University, Tainan 701, Taiwan;
[email protected] (J.-W.L.); [email protected] (H.-T.L.);
[email protected] (F.-H.W.)
* Correspondence: [email protected] (G.-B.C.); [email protected] (Y.-C.C.);
Tel.: +886-6-2757575 (ext. 51030) (G.-B.C.); +886-6-2757575 (ext. 63690) (Y.-C.C.)
Abstract: Sewage sludge is a common form of municipal solid waste, and can be utilized as a
renewable energy source. This study examines the effects of different key operational parameters
on sewage sludge pyrolysis process for pyrolytic oil production using the Taguchi method.
The digested sewage sludge was provided by the urban wastewater treatment plant of Tainan,
Taiwan. The experimental results indicate that the maximum pyrolytic oil yield, 10.19% (18.4% on
dry ash free (daf) basis) by weight achieved, is obtained under the operation conditions of 450 ◦ C
pyrolytic temperature, residence time of 60 min, 10 ◦ C/min heating rate, and 700 mL/min nitrogen
flow rate. According to the experimental results, the order of sensitivity of the parameters that affect
the yield of sludge pyrolytic oil is the nitrogen flow rate, pyrolytic temperature, heating rate and
residence time. The pyrolysis and oxidation reactions of sludge pyrolytic oil are also investigated
using thermogravimetric analysis. The combustion performance parameters, such as the ignition
temperature, burnout temperature, flammability index and combustion characteristics index are
calculated and compared with those of heavy fuel oil. For the blend of sludge pyrolytic oil with
heavy fuel oil, a synergistic effect occurs and the results show that sludge pyrolytic oil significantly
enhances the ignition and combustion of heavy fuel oil.
Keywords: sewage sludge; pyrolytic oil; Taguchi method; thermogravimetric analysis; synergistic effect
1. Introduction
Even now, a large part of the global energy supply still depends on fossil fuels, resulting in rapid
depletion of these resources and enormous GHG (greenhouse gas) emissions. In order to mitigate
the problems associated with this fossil fuel depletion and climate warming, it is necessary to either
improve the efficiency of fossil fuel utilization [1,2], or partially replace the use of such fossil fuels with
zero or neutral carbon footprint alternative energy supplies [3,4]. Among the alternative fuels that are
now being considered, biomass is widely recognized as a promising, eco-friendly source of renewable
energy, which has the advantage of being readily available around the world.
Biomass derives from botanical or biological sources, or from a combination of these.
Ordinary sources of biomass include agricultural solid waste, forestry residues, municipal waste,
energy crops, and biological waste. Sewage sludge, the major municipal waste from the wastewater
treatment plants, is a complicated mixture of undigested organics, inorganic materials, and moisture.
The traditional treatment of sewage sludge includes agricultural use, incineration, and landfill.
However, these methods have several shortcomings. With regard to agricultural use, sewage sludge

Energies 2018, 11, 2260
contains pathogenic bacteria, viruses, and parasitic helminths, which can harm the health of humans,
animals, and plants. For incineration, the heating value of dewatered sewage sludge is low. For landfill,
sludge contains a certain amount of organic substance, which generates a biogas rich in methane.
Moreover, the cost of this approach is also increasing, because it is more and more difficult to
find appropriate areas for landfill. In addition, with recent increases in the sewage treatment rate,
the volume of sludge has grown rapidly, and thus alternative methods of sludge management are
urgently needed. Thermochemical processes, such as pyrolysis or gasification, have been studied and
recommended as suitable alternatives [5]. Thermochemical processes can produce energy from the
sludge and have little influence on the environment.
For conversion into bio-fuel from biomass, biochemical conversion (BCC) and thermochemical
conversion (TCC) are the two general broad approaches. BCC often uses fermentation and biological
conversion processes. Direct combustion, gasification, pyrolysis, and liquefaction are generally used
for TCC [6,7]. The resulting products can be used directly as fuels, or further processed for specialty
chemicals [8,9].
Among the various thermochemical conversion processes, thermal pyrolysis has attracted
more interest with regard to yielding liquid fuel products from various biomass species, such as
woody biomass [10], bagasse [11], straws [12], miscanthus [13], castor meal [14], oil palm fiber [15],
fruit waste [16], and municipal solid waste [6,17]. Even for waste oil, pyrolysis approaches offer some
advantages over existing methods (transesterification, hydrotreating, gasification, solvent extraction,
and membrane technology) in producing useful pyrolysis products for future reuse [18]. Products from
pyrolysis processes depend strongly on the heating rate, heating temperature, gas residence time,
particle size of biomass, and water content of the biomass. In fact, great work related to the conversion
of biomass to liquid fuels was performed during the oil crisis of the 1970s [19], and most studies on
sludge liquid fuel production are based on pyrolysis for lignocellulosic materials. In general, in order
to increase the yield of bio-oil, the biomass in the fast pyrolysis process is rapidly heated to a high
temperature (425–600 ◦ C) and has a short gas residence time (<3 s) in the absence of oxygen [20].
The review work of sewage sludge pyrolysis by Fonts et al. [21] placed specific emphasis on
liquid sewage sludge pyrolytic fuel production, and detailed comparisons of pyrolytic oils between
sewage sludge and lignocellulose were also made. Several studies have performed sewage sludge
pyrolysis to produce liquid fuel, which is composed of aqueous and organic phase compounds [22–27].
Some papers emphasize the influence of operational conditions and composition on the sludge
pyrolysis product, especially the liquid yield. There are few studies related to the maximum yield
of sludge pyrolytic oil and its combustion characteristics. Therefore, in this study we examined
the influence of different key parameters on sewage sludge pyrolysis process for pyrolytic oil using
the Taguchi method. As an example, the method was applied in determining the pyrolysis process
parameters of the digested sewage sludge provided by the urban wastewater treatment plant of
Tainan, Taiwan. The combustion characteristics of the resulting pyrolytic oil are also investigated by
thermogravimetric analysis. The experimental results are also compared with those for heavy fuel oil,
since the co-firing of sludge pyrolytic oil with heavy fuel oil is a promising potential alternative for
practical applications.
2. Methodology
2.1. Experimental Setup

The feedstock used for the studies is typical digested sewage sludge from a wastewater treatment
plant in Tainan, Taiwan (Figure 1). Even though the sludge is dewatered in the wastewater treatment
plant, its water content is still higher than 30%. Therefore, before the pyrolysis experiment, the sewage
sludge was preheated to 110 ◦ C in an oven until the water content of the sewage sludge was lower
than 10% (see Table 1). The drying step is important and can even affect the conversion of a waste
into other products with high added value [28]. Figure 2a depicts the experimental thermal pyrolysis
252
Energies 2018, 11, 2260
apparatus. In the experiment, the dried sewage sludge was mixed and ground to pass through a
100-mesh sieve, and 90 g of sewage sludge were packed in a cylindrical holder made of quartz and
put into the furnace. The covers of the cylindrical holder were made of stainless steel with many
holes. For pyrolysis, the tubular furnace was first vacuumed, and then flushed with N2 to ambient
pressure. The furnace was then heated to the preset temperature and maintained at the temperature
for a designated residence time. The residence time here is different from the gas residence time,
which represents the residence time for pyrolytic vapor inside the pyrolysis furnace. In addition,
the gas residence time is inversely proportional to the nitrogen flow rate. The heated carrier gas
can flow into the holder and interact with the sludge. The generated volatile gas was collected and
delivered to a condensing system kept at 25 ◦ C to form a liquid product. Figure 2b shows the thermal
pyrolysis operating process in a fixed-bed tubular furnace and it must undergo two steps: being
heated up from room temperature to the target temperature at a specific heating rate and holding at
the targeted temperature for a designated residence time. The production of pyrolytic oil occurs in a
certain conditions and is influenced by the combination of four operational parameters: heating rate,
target pyrolytic temperature, residence time, and nitrogen flow rate. The Taguchi method will be used
to analyze the influence of this parameter combination on the pyrolytic oil production from sewage
sludge using a fixed-bed tubular furnace. After the experiment, the condensing bottle was held for 24 h,
and the pyrolytic oil was separated from the aqueous matter using a separatory funnel. This pyrolytic
oil has a high heating value, and was the target product in the study. As to the aqueous product,
it could be used as a source for fertilizers or chemicals due to the nitrogen-containing compounds [29],
or as a source of triacetonamine [30].
Figure 1. Photograph of sewage sludge, (a) sludge from the wastewater treatment plant and (b) sludge
after preheating.
Table 1. Proximate analysis of sewage sludge.
Proximate Analysis Sewage Sludge Sewage Sludge [21]

Moisture (wt %) 6.94 1.5–7.1
Volatiles (wt %) 45.45 38.3–66.8
Ash (wt %) 37.67 22.6–52.0
Fixed carbon (wt %) 9.94 0.8–19.7
253
Energies 2018, 11, 2260
Figure 2. Schematic diagram of (a) experimental setup and (b) operating process for thermal pyrolysis.
Simultaneous thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC)

measurement were performed using the thermal analyzer (PerkinElmer, Waltham, MA, USA: STA
8000). The temperature range was 30–1200 ◦ C, with a heating rate of 10 ◦ C/min. A 145-μL alumina
crucible was used to hold the testing sample, and the flow rate of air or nitrogen environmental gas
was 50 mL/min.
A portable pH meter (ROCKER, New Taipei City, Taiwan: EC-210) was used to measure the pH
value of sludge pyrolytic oil, with an accuracy of about 0.01 ± 1 digit. The pH value of sludge pyrolytic
oil can be measured directly because of the low water content. In addition, a bomb calorimeter (Parr
Instrument, Moline, IL, USA: Model 6200) was used to measure the heating values of different samples,
while an elemental analyzer (Elementar, Langenselbold, Hesse, Germany: vario EL III) was used to
analyze the C, H, O, N and S composition of the sewage sludge, pyrolytic sludge oil and heavy fuel oil.
The chemical compounds contained in the sludge pyrolytic oil were analyzed by Agilent 7890A-GC
and Agilent 5975C-MSD (Agilent Technologies, Santa Clara, CA, USA).
2.2. Taguchi Method

The Taguchi method was used to find the maximum yield of sludge pyrolytic oil from the pyrolysis
experiments. The feature of the Taguchi method uses an orthogonal array experimental design with
a simple analysis of variance. In Taguchi method, the variables are assigned to each column, called
orthogonal arrays for the design of experiments [31], of a matrix related to experimental parameters
and experiment levels. In general, the Taguchi method is not aimed at finding the optimal conditions
as a full factorial method, but it can analyze the best trend with fewer experimental data, and thus is
commonly applied to optimize industrial processes.
Taguchi method was previously applied to analyze caster meal pyrolysis by Chen et al. and this
group [14]. The Taguchi method uses the S/N ratio, the signal to noise ratio , to measure the quality
characteristics deviating from the desired value [32]. Different strategic categories of nominal—the
best (NTB), smaller the better (STB), and larger the better (LB)—are used to describe the S/N ratio
characteristics [33]. In this study, for example, to obtain the maximum yield of pyrolytic oil, the “larger
the better” characteristic must be adopted. The expression of S/N ratio is shown as follows:

1 N 1 1 3s2
N i∑
S/N LB = −10 log 2
= −10 log (1 + ) , (1)
=1 y i y2 y2
where N is the test number, yi is the value of the pyrolytic oil yield in the ith test, and y is the average
yield of pyrolytic oil. The procedure for the experimental design with the Taguchi method can be
found in previous work [14].
254
Energies 2018, 11, 2260
2.3. Characteristic Combustion Parameters

Several characteristic combustion parameters can be deduced from a TG-DTG curve to assess
the combustion properties of fuels. This study will make use of the ignition temperature (Ti ), burnout
temperature (Te ), combustion characteristics index (S), and flammability index (C) to evaluate the
combustion characteristics of sludge pyrolysis oil mixed with different percentages of heavy fuel
oil. The weight loss curve is used to define the ignition temperature. There are many studies
concerning the definition of the ignition temperature using TG-DTG curves [34–36], and the method
proposed by Tognotti et al. [34] is widely used to predict the ignition temperature for different fuels.
In this, the weight loss curve of fuel in the air atmosphere is overlapped with that in inert gas
atmosphere (N2 ), and the ignition temperature (Ti ) corresponds to the first bifurcation point in the TG
curves [37]. In addition, the burnout temperature (Te ) is defined as the temperature corresponding to
99% conversion of the fuel in the TG curve of air atmosphere.
Since the ignition and burnout temperatures show only a single combustion property of the
fuel, an integrated index is required for the global combustion performance of fuels. Therefore,
the combustion characteristic index (S) and flammability index (C) are also used for benchmarking
the combustibility of fuels [38]. The combustion characteristic index (S) was first proposed by
Cheng et al. [39]. The larger the combustion characteristics index, the better the combustion
characteristics of the fuel. The combustion rate is expressed by the Arrhenius Law, as follows:
dW/dτ = A exp(−E/RT), (2)
where dW/dτ is the combustion rate, A is a pre-exponential factor, E is the activation energy, and T is
the absolute temperature.
Take the differential of Equation (2) with temperature and rearrange it as follows:

R d dW dW 1
× = × 2. (3)
E dT dτ dτ T
(dW/dτ )max (dW/dτ )mean
At ignition temperature Ti , Equation (3) is multiplied by (dW/dτ ) T =T Te
:
i

R d dW (dW/dτ )max (dW/dτ )mean (dW/dτ )max × (dW/dτ )mean
× × = , (4)
E dT dτ T = Ti ( dW/dτ ) T = Ti Te Ti 2 × Te
where (dW/dτ )max is the maximum combustion rate, which can be obtained from the peak of DTG
curve. (dW/dτ )mean is the mean combustion rate, which can be obtained from the mean of DTG
curve. The first term of Equation (4) on the left-hand side represents the reaction strength of fuel
combustion, which is related to the activation energy. The second term stands for the change rate of
fuel combustion rate at the ignition temperature and the third term refers to the ratio of maximum
combustion rate to the combustion rate at the ignition temperature. These two terms are related to the
ignition temperature. The last term is the ratio of mean combustion rate to the burnout temperature;
the larger the value, the shorter the burnout time of the fuel.
The right-hand side of Equation (4) is the combustion characteristic index (S):
(dW/dτ )max × (dW/dτ )mean

S= . (5)
Ti 2 × Te
The flammability index combines the influence of maximum combustion rate and the ignition
temperature. It can reflect the difficulty of fuel combustion and the burning out speed. The fuel with a
255
Energies 2018, 11, 2260
large flammability index will have better combustion stability and the flammability index (C) of the
sample can be expressed as follows:
C = (dW/dτ )max /Ti 2 . (6)
3.1. Thermogravimetric Analysis of Sewage Sludge

Table 1 shows the approximate analysis of the sewage sludge used in this study, and it is
composed of 6.94 wt % moisture, 45.45 wt % volatiles, 9.94 wt % fixed carbon, and 37.67 wt %
ash. This composition is well within the range of the different kinds of sewage sludge analyzed by
Fonts et al. [21].
Figure 3 shows the TGA thermograph of sewage sludge. The heating rate is 10 ◦ C/min and the
nitrogen flow rate is 50 mL/min. The purple curve shows thermogravimetry (TG), the blue curve
shows differential of thermogravimetry (DTG), the red curve shows differential scanning calorimetry
(DSC), while the green curve represents TG in air atmosphere. The mass loss of sewage sludge increases
along with the temperature, and the major reactions occur between 120 ◦ C and 500 ◦ C.
For the purple TG curve in Figure 3, there are three stages of weight loss. The first stage has
a weight loss of 9.33% in the temperature range of 30–200 ◦ C, due to the removal of the water and
organics with low boiling points. In the second stage, there are two distinct peaks at the temperatures
of 273 ◦ C and 323 ◦ C in the DTG curve. About 33.57% of the original sewage sludge is lost, and this is
mainly from the decomposition of aliphatic compound protein and carbohydrates contained in the
sewage sludge [40–43]. In the third stage, the weight loss tends to be mild when the temperature
exceeds 500 ◦ C, which might be from the decomposition of residual organic matter and the inorganic
materials, such as calcium carbonate [44]. Finally, the residue is about 46.79% of the original weight of
the sewage sludge when the temperature reaches 1000 ◦ C.
Figure 3. Thermogravimetric analysis results for sewage sludge at a heating rate of 10 ◦ C/min and
nitrogen flow rate of 50 mL/min.
3.2. Pyrolytic Parameters for Sewage Sludge
3.2.1. The Influence of Nitrogen Flow Rate on the Pyrolytic Oil Yield
Inert carrying gases such as nitrogen and carbon dioxide are often used in the pyrolysis process.
Inert gas can keep the reactor in an oxygen-free state, and causes the volatiles to move quickly away
256
Energies 2018, 11, 2260
from the system and thus avoid condensing again. Pütün et al. investigated the effects of the sweeping
gas flow rate (N2 ) on pyrolysis products, and found that increasing the sweeping gas will reduce the
gas residence time and prevent the pyrolytic products from the secondary thermal decomposition,
thus maximizing the liquid yield [45].
The influence of the N2 flow rate on the sludge pyrolytic oil was also investigated in this study,
as shown in Figure 4. The conditions were fixed as a pyrolytic temperature of 500 ◦ C, heating rate
of 20 ◦ C/min and residence time of 60 min. The pyrolytic oil yield increases with the N2 flow
rate in the range of 300–700 mL/min. However, it then decreases again after the N2 flow rate
surpasses 700 mL/min. Higher liquid yield is usually obtained in a certain range of gas residence
time. According to the volume of the quartz holder in this study, the filled volume of 90 g sludge
and the nitrogen flow rate of 700 mL/min, the gas residence time is between 2.5 s and 3 s. It is under
the range preferred for fast pyrolysis to obtain more oil product [20]. When the nitrogen flow rate
is less than 700 mL/min, the gas residence time is increased and this gas residence time may cause
condensable gases secondary decomposition, which is not conducive to oil product. Nevertheless,
the variation in the nitrogen flow rate would also vary the degree of dilution of the pyrolysis vapors.
When the dilution of the vapors increases, the vapor pressure decreases, making it more difficult to
reach saturation pressure. This affects the condensation and thus the liquid yield. It is speculated this
reason make sludge pyrolytic oil decrease after the nitrogen flow rate exceeds 700 mL/min. Therefore,
the nitrogen flow rate was chosen as 300, 500, 700 and 900 mL/min in the following work.
Figure 4. The effect of nitrogen flow rate on the yield of pyrolytic oil.
3.2.2. The Optimal Conditions for Sludge Pyrolysis

The optimal pyrolysis conditions for the maximum yield of sludge pyrolytic oil are investigated
in the fixed-bed reactor using the Taguchi method. Table 2 gives the pyrolysis parameters and
their levels. Four pyrolysis parameters are the heating rate, pyrolysis temperature, residence time,
and nitrogen flow rate. These parameters are assumed to be independent and each parameter has four
levels. According to the number of parameters and their levels, the L16 (45 ) orthogonal array will be
selected. Table 3 shows the experimental layout using an L16 (45 ) orthogonal array and the original
256 experimental conditions for the full factorial method are thus diminished significantly to only 16.
Table 4 reveals the mass balances of all 16 experimental conditions, including aqueous, oil, and solid
yields. Gas yield was obtained by calculation since it was not collected. It shows that sewage sludge
char accounts for 50–60% of total product, and the liquid yield is in the range of 18–27%. To assure the
reliability of the results, each experiment was performed three times to take the average.
257
Energies 2018, 11, 2260
Table 2. Pyrolysis control factors and their levels.
Level 1 Level 2 Level 3 Level 4

Heating rate (◦ C/min) 10 20 30 40
Temperature (◦ C) 450 500 550 600
Residence time (min) 30 60 90 120
N2 flow rate (mL/min) 300 500 700 900
Table 3. Experimental layout using an L16 (45 ) orthogonal array.
Experiment Heating Rate Temperature Residence N2 Flow Rate Cooling

No. (◦ C/min) (◦ C) Time (min) (mL/min) Temperature (◦ C)
1 10 450 30 300 25
2 10 500 60 500 25
3 10 550 90 700 25
4 10 600 120 900 25
5 20 450 60 700 25
6 20 500 30 900 25
7 20 550 120 300 25
8 20 600 90 500 25
9 30 450 90 900 25
10 30 500 120 700 25
11 30 550 30 500 25
12 30 600 60 300 25
13 40 450 120 500 25
14 40 500 90 300 25
15 40 550 60 900 25
16 40 600 30 700 25
Table 4. The mass balances of all 16 experimental conditions.
Experiment No. Aqueous (wt %) Oil (wt %) Solid (wt %) Gas (by diff.) (wt %)
1 11.42 8.71 57.02 22.85
2 12.67 8.88 55.09 23.36
3 15.06 9.87 54.48 20.60
4 13.42 7.29 52.18 27.11
5 14.93 8.76 54.92 21.40
6 14.09 7.91 53.37 24.63
7 14.31 8.62 52.45 24.62
8 9.44 9.22 54.10 27.24
9 13.41 8.31 55.80 22.49
10 12.12 9.32 54.56 24.00
11 12.93 6.59 53.10 27.38
12 13.76 7.58 52.91 25.75
13 14.70 9.20 54.43 21.66
14 13.11 5.43 54.30 27.17
15 18.41 9.22 53.33 19.03
16 15.67 7.88 51.82 24.63
After obtaining the experimental yield of sludge pyrolytic oil, the corresponding S/N ratio can be
calculated using Equation (1). Since the experimental design is orthogonal, the effect of each pyrolysis
parameter at different levels can be separated out. The mean S/N ratio for the 16 experimental
conditions is computed and the mean S/N ratio for each level of the other pyrolysis parameters can
then be calculated. The S/N response for the mean S/N ratio of the pyrolysis parameters at each level
are shown in Table 5 and it is used to determine the maximum yield of sludge pyrolytic oil.
258
Energies 2018, 11, 2260
Table 5. S/N response table for sludge pyrolytic oil.
L1 L2 L3 L4 Max-Min
(A) Heating rate −21.27 −21.29 −22.06 −22.20 0.929
(B) Temperature −21.17 −22.25 −21.43 −21.98 1.074
(C) Residence time −22.23 −21.32 −21.93 −21.34 0.910
(D) N2 flow rate −22.55 −21.52 −20.99 −21.77 1.564
Based on the experimental results and the calculated S/N ratio, the optimal operation conditions
are determined as shown in Figure 5. In Figure 5, these four pyrolysis parameters are labelled as “A”
for heating rate, “B” for pyrolytic temperature, “C” for residence time, and “D” for nitrogen flow
rate. Although the results for pyrolytic temperature and residence time are zigzag, several studies
concerning the Taguchi method also show the same phenomenon for S/N ratios [46–48]. It can be
deduced from Figure 5 that the optimal pyrolysis conditions are A1, B1, C2 and D3, which represent a
heating rate of 10 ◦ C/min, pyrolytic temperature of 450 ◦ C, residence time of 60 min, and nitrogen flow
rate of 700 mL/min. It should be mentioned that fast pyrolysis, which represents a pyrolysis reaction
carried out at high heating rates, moderate temperatures and short gas residence times, is usually
used to obtain high yield of pyrolysis oil. There are three main kinds of technologies for fast pyrolysis:
ablative pyrolysis, fluidized bed pyrolysis, and vacuum pyrolysis. However, in the study, a tubular
furnace, a type of fixed-bed reactor was used. The results show that oil yield was affected by the
combination of four operational parameters. The heating rate is not the dominant parameter due that
it was restricted by the pyrolytic temperature. Among these parameters, the nitrogen flow rate has
the most obvious effect, which is followed by the pyrolytic temperature and the residence time has
the least effect. The nitrogen flow rate (inversely proportional to the gas residence time) has a more
significant effect on the oil yield than heating rate.
Figure 5. S/N response for sludge pyrolytic oil.
3.2.3. Confirmation Experiment

In the Taguchi method, the confirmation experiment is the last step and the aim of this is to
validate the yield of sludge pyrolytic oil. After obtaining the optimum conditions and predicting the
response under these conditions, a new experiment was performed with the optimum levels of these
pyrolysis parameters. The theoretical optimal S/N ratio can be estimated as:
S/Nopt = S/NAV + (A1 − S/NAV ) + (B1 − S/NAV ) + (C2 − S/NAV ) + (D3 − S/NAV ) = −19.63, (7)
where S/NAV is the total mean of the S/N ratio and S/Nopt is the S/N ratio at the optimal level.
259
Energies 2018, 11, 2260
From the experimental results of the standard L16 (45 ) orthogonal array, the theoretical
optimization S/N value is −19.63, and it can be deduced that the theoretical maximum pyrolytic oil
yield is 0.1044 g/g sewage sludge (18.85% on dry ash free (daf) basis). The real yield of the confirmation
experiment is 0.1019 g/g-sewage sludge (18.4% on dry ash free (daf) basis). The difference is only
0.0025 g/g sewage sludge, and thus real experimental result is very close to the theoretical value.
In addition, under the optimal conditions, the solid product is about 54.83%, aqueous product is about
15% and by mass balance, the gas product is 19.98%. Although the solid product is not suitable to
be used as fuel due to high content of inorganic matter (ash), it can be used to produce adsorbents
for disposing of pollutants such as H2 S and NOx [21]. The aqueous product could be used for the
production of fertilizer since it has high content of nitrogen-containing compounds. As to the gas
product, it could be recycled and used for the drying of feedstock.
3.3. Elemental Analysis and Heating Values of the Different Fuels

Table 6 shows the elemental analysis and heating value for the sewage sludge, sludge pyrolytic oil
and heavy fuel oil on the wet basis. The sludge pyrolytic oil is obtained using the optimal conditions
from the Taguchi method. The results show that the dried sewage sludge contained 1.28% sulfur and
the sludge pyrolytic oil contained 1.25% sulfur. Compared to the sulfur content of heavy oil, it can
be seen that when using sludge pyrolytic oil as fuel more care must be taken with regard to the SOX
emissions. However, since the combustion of pure sludge pyrolytic oil appears to be impractical
because of the yield, it is more feasible to co-combust this and other fuels, such as heavy fuel oil. In this
case, the sulfur content would drop significantly and thus decrease the sulfur oxide emissions. As to the
high nitrogen content in the sludge pyrolytic oil, it is usually related to fuel NOx during the combustion
process. Therefore, the blending ratio of sludge pyrolytic oil and heavy fuel oil should be also well
selected to have less SOx and NOx emissions. Sludge pyrolytic oil also has a higher oxygen content
than heavy fuel oil. A high oxygen content directly affects the energy density, and the resulting heating
value is usually lower than that of heavy fuel oil. Although the heating value of sludge pyrolytic oil is
only about 2/3 that of heavy fuel oil (28.66 MJ/kg vs. 44.87 MJ/kg), it is still significantly higher than
the common bio-oil from lignocellulose. In addition, it does not have the corrosive character that is
often seen in other pyrolysis liquids obtained from lignocellulosic biomass, as the pH value of sludge
pyrolytic oil is between 8.86 and 9.67 in this study. Accordingly, these advantageous properties make
sludge pyrolytic oil favorable as fuel for the co-combustion with other fossil fuels, like heavy fuel
oil. In addition, sludge pyrolytic oil contains a high nitrogen and sulfur content, primarily due to the
decomposition of protein contained within the dried sewage sludge [21].
Table 6. Elemental analysis and heating values of different fuels.
Samples N% C% H% O% S% HHV(MJ/kg)
Sewage sludge 4.14 22.62 4.93 35.29 1.28 10.18
Sludge pyrolytic oil 5.92 58.44 9.08 18.98 1.25 28.66
Heavy fuel oil 0.46 86.7 11.81 0.26 0.28 44.87
Table 7 compares the properties of diesel, heavy fuel oil, and sludge pyrolytic oil. The optimal
combination conditions obtained from the Taguchi method is used for the yield of sludge pyrolytic oil.
The test standards used in this study for different properties are also shown in Table 7. The density
of sludge pyrolytic oil is larger than that of diesel and heavy fuel oils and it is a little higher than the
density of water. The viscosity of sludge pyrolytic oil is between that of diesel and heavy fuel oils.
The flash point of sludge pyrolytic oil is 80 ◦ C, which makes it safe to use as a fuel. The pour point of
sludge pyrolytic oil is lower than that of heavy fuel oil and it still retains mobility at low temperatures.
260
Energies 2018, 11, 2260
Table 7. Properties of different kinds of liquid fuels.
Properties Diesel Heavy Fuel Oil Sludge Oil (Pyrolysis) Test Method
Density (g/cm3 , @15 ◦ C) 0.8335 0.9533 1.04 ASTM D4052
Viscosity (mm2 /s @40 ◦ C) 3.024 130.3 64.86 ASTM D445
Flash Point (◦ C) 52–80 110 80 ASTM D93
Pour Point (◦ C) −9 12 6 ASTM D97
Heating Value (MJ/kg) 42.50 44.87 28.66 ASTM D240
The bio-oil characterization was carried out using GC/MS spectrometry at the optimal conditions
for the maximum yield of the pyrolytic oil. Figure 6 shows the GC/MS mass spectra of sludge
pyrolytic oil, and there are more than 100 peaks corresponding to different organic compounds
detected. However, many of them cannot be identified. The major compounds are shown in Table 8
and the area percentage in the tables represents the area fraction under the peak for that identified
component. The highest peak areas of the identified compounds were for n-hexadecanoic acid,
cholest-4-ene, octadecanoic acid, hexadecanamide, oleic acid, 4-methyl phenol, toluene, indole, and so
on. Oleic acid is a monounsaturated fatty acid, and can be found in a variety of animal and vegetable
materials. The saturated form of this acid is stearic acid, and is used in Lorenzo’s oil. Indole is generally
used in the production of synthetic jasmine oil. Huang et al. also investigated bio-oil from pyrolysis of
granular sewage sludge [49]. In their study, the GC-MS results of bio-oil also indicated n-hexadecanoic
acid (C16 H32 O2 ) is the major component at the pyrolysis temperature of 500 ◦ C, which is close to our
operation condition. The n-hexadecanoic acid accounts for a large proportion of the known ingredients
in the sludge pyrolytic oil. The other identical components in both studies include octadecanoic acid
(C18 H36 O2 ), tetradecanoic acid (C14 H28 O2 ) and hexadecanamide (C16 H33 NO). In general, carboxylic
acid is the most compound type of the known ingredients.
Figure 6. GC-MS mass spectra for sludge pyrolytic oil.
261
Energies 2018, 11, 2260
Table 8. GC-MS analysis of sludge pyrolytic oil.
Peak Retention Time Chemical Name Mol Formula Area%

1 2.178 acetic acid C2 H4 O2 0.40%
2 2.277 2-butanone C4 H8 O 0.18%
3 4.551 Pyrrole C4 H5 N 0.75%
4 4.994 Toluene C7 H8 3.63%
5 7.228 ethyl benzene C8 H10 0.93%
6 7.764 styrene C8 H8 1.47%
7 8.995 phenol C6 H6 O 3.07%
8 10.017 4-methyl phenol C7 H8 O 4.41%
9 10.839 4-ethyl phenol C8 H10 O 1.23%
10 11.474 benzene propane nitrile - 0.87%
11 11.929 indole C 8 H7 N 3.18%
12 12.603 3-methyl indole C 9 H9 N 1.31%
13 14.729 tetradecanoic acid C14 H28 O2 1.83%
14 15.504 pentadecanenitrile C15 H29 N 3.17%
15 15.793 n-hexadecanoic acid C16 H32 O2 31.90%
16 16.664 oleic acid C18 H34 O2 4.86%
17 16.739 octadecanoic acid C18 H36 O2 5.89%
18 16.855 hexadecanamide C16 H33 NO 4.97%
19 17.748 octadecanamide C18 H37 NO 1.06%
20 19.833 Unknown - 3.18%
21 19.991 cholest-2-ene C27 H46 2.08%
22 20.077 cholest-3-ene C27 H46 1.69%
23 20.16 cholest-4-ene C27 H46 9.42%
24 20.464 cholesta-3.5-dienl - 2.10%
25 21.518 unknown - 6.41%
3.4. Combustion Characteristic Parameters

Thermogravimetric analysis was performed for sludge pyrolytic oil (SPO), heavy fuel oil (HFO),
20% SPO/80% HFO, and 50% SPO/50% HFO, and the results were used to calculate the ignition
temperature, burnout temperature, combustion characteristic index (S) and flammability index (C).
In these experiments, the samples were inserted into an alumina crucible and heated up to 1200 ◦ C at
the heating rate of 10 ◦ C/min. A nitrogen flow rate of 50 mL/min is used for the pyrolysis process and
an air flow rate of 50 mL/min is used for the oxidation reaction.
Figure 7 shows the TGA burning profiles of different blending ratio for sludge pyrolytic oil/heavy
fuel oil, and the TG pyrolysis curve for sludge pyrolytic oil with N2 carrying gas is also shown to
identify the ignition temperature. In Figure 7a, the main reaction for pure sludge pyrolytic oil occurs
between 300 and 600 ◦ C. The oxidation processes can be divided into three stages, as follows. (1) The
stage of moisture release: the first peak of DTG curves occurs between 30 and 136 ◦ C, corresponding to
the endothermic reaction in the DSC curve. This mainly represents the heat adsorption by water and
the organic matter with low boiling point. The weight loss in this stage is about 21.63 wt %. (2) The
volatilization and oxidation stage of the light volatile substance, which occurs between 136 ◦ C and
471 ◦ C. The DSC curve indicates an exothermic reaction in this stage, and the weight loss is about
65.21 wt %. (3) The stage of nonvolatile combustion, which represents the combustion of heavy organic
matter after 471 ◦ C, and there is a peak in the DTG curve at the temperature of 565 ◦ C, corresponding
to a higher exothermic peak in the DSC curve. The weight loss in this stage is about 11.67 wt %. Finally,
the remaining substance after oxidation is about 1.16 wt %.
262
Energies 2018, 11, 2260
Figure 7. TGA burning profiles of sludge pyrolytic oil at a heating rate of 10 ◦ C/min and air flow rate
of 50 mL/min. (a) Pure sludge pyrolytic oil; (b) pure heavy fuel oil; (c) 20% sludge pyrolytic oil and
80% heavy fuel oil; (d) 50% sludge pyrolytic oil and 50% heavy fuel oil. (Green curve: TG; blue curve:
DTG; red curve: DSC; purple curve: TG in the pyrolysis process).
In Figure 7b, the main reaction for pure heavy fuel oil occurs between 300 and 650 ◦ C.
The oxidation processes have two stages, as follows. (1) The volatilization and oxidation stage
of light volatile substances and there are several peaks in the DTG curve between 300 and 510 ◦ C.
The DSC curve indicates an exothermic reaction in this stage, and the weight loss is about 83.41 wt
%. (2) The stage of nonvolatile combustion represents the combustion of heavy organic matter after
510 ◦ C, and there is a peak in the DTG curve at the temperature of about 600 ◦ C, corresponding to a
higher exothermic peak in the DSC curve. The weight loss in this stage is about 16.06 wt %. Finally,
the remaining residue is about 0.42 wt %.
For the blends, Figure 7c shows the TGA burning profiles of 20% sludge pyrolytic oil/80% heavy
fuel oil. The TG curve decreases faster than that of pure heavy oil due to the mixture of sludge
pyrolytic oil, which has more volatile organic compounds. Similar to the case of pure sludge pyrolytic
oil, the reaction processes can be divided into three stages, as follows. (1) The stage of moisture release,
when the first peak of the DTG curves occurs between 30 and 150 ◦ C and the weight loss in this stage is
about 5.17 wt %. (2) The volatilization and oxidation stage of the light volatile substances, which occurs
between 150 and 510 ◦ C; the overall trend is the same as for the heavy oil. The DSC curve indicates an
exothermic reaction in this stage, and the weight loss is about 82.97 wt %. (3) The stage of nonvolatile
combustion, which represents the combustion of heavy organic matter above 510 ◦ C. There is a peak
in the DTG curve at the temperature of about 600 ◦ C, corresponding to a higher exothermic peak in
the DSC curve, and the weight loss in this stage is about 11.87 wt %. There is almost no residue left
after oxidation.
When the blending ratio increases to 50% as shown in Figure 7d, The TG curve only decreases
slower than that of pure sludge pyrolytic oil in these cases, and this is primarily due to the mixture of
263
Energies 2018, 11, 2260
sludge pyrolytic oil, which has more volatile organic compounds. The reaction processes can also be
divided into three stages, as follows. (1) The stage of moisture release, when the first peak of the DTG
curves occurs between 30 and 145 ◦ C, and the weight loss in this stage is about 14.23 wt %. (2) The
volatilization and oxidation stage of the light volatile substances, which occurs between 145 and 500 ◦ C
and can be further divided into two parts. The first is close to the oxidation of sludge pyrolytic oil and
is between 145 and 415 ◦ C, while the second is close to the oxidation of heavy fuel oil and is between
415 and 500 ◦ C. The DSC curve indicates an exothermic reaction in this stage, and the weight loss
is about 72.58 wt %. (3) The stage of nonvolatile combustion, which represents the combustion of
heavy organic matter above 500 ◦ C. There is a peak in the DTG curve at the temperature of 566 ◦ C,
corresponding to a higher exothermic peak in the DSC curve, and the weight loss in this stage is about
12.3 wt %. After oxidation, the residue is about 0.75 wt %.
From Figure 7, these results show that for the blends the second DTG peak, which represents
the combustion of volatile substances and determines the ignition temperature, shifts to a lower
temperature when compared to the case of pure heavy fuel oil. As can be seen in the TGs and DTGs
for the blends of sludge pyrolytic oil and heavy fuel oil, there is an obvious interaction between the
blend components. Table 9 shows the combustion characteristic parameters of sludge pyrolytic oil,
heavy fuel oil, and the blends. Sludge pyrolytic oil has a lower ignition temperature (274 ◦ C) than
heavy fuel oil (434 ◦ C), since it has more volatile matter. In addition, a greater percentage of sludge
pyrolytic oil in the blend will reduce the ignition temperature to a greater extent. We also found
similar trends for combustion characteristics of a single droplet in the suspended droplet experimental
system [50]. A higher blending ratio of sludge pyrolytic oil will make ignition easier. However,
the burnout temperature is around 600 ◦ C in all these cases because the combustion characteristics
of heavy organic matter (the last DTG peak) are similar for these oils. The last DTG peak derived
from the oxidation of heavy organic matter determines the burnout temperature. Table 9 also shows
that the maximum combustion rate of sludge pyrolytic oil is higher than that of heavy fuel oil.
An interesting finding is that the maximum combustion rate in the blend is even higher than for
pure sludge pyrolytic oil. This indicates that there is an interaction between these blend components,
and synergistic effects can be perceived. In addition, as the blending ratio of sludge pyrolytic oil
increases, the maximum combustion rate also moves to a lower temperature and represents the
domination shifts from heavy organic matter to light volatile substances. Finally, the flammability
index and combustion characteristics index in the blend are higher than those of heavy fuel oil. Since
the sludge pyrolytic oil has a higher volatile content than heavy fuel oil, adding a certain amount of
sludge pyrolytic oil will promote the ignition performance of heavy fuel oil, which is conductive to the
stability of heavy fuel oil combustion. The co-combustion of sludge pyrolytic oil and heavy fuel oil
has better combustion characteristics than seen with pure heavy fuel oil. Moreover, it is even better
than pure sludge pyrolytic oil, especially in the case of 50% sludge pyrolytic oil mixed with 50% heavy
fuel oil.
Table 9. The combustion characteristic parameters of sludge pyrolytic oil, heavy fuel oil, and the blends.
Fuel (dW/dτ )max (dW/dτ )mean Ti ( ◦ C ) Te ( ◦ C ) S × 107 C × 105

Heavy fuel oil 3.517 1.544 434 612 0.470 1.863
20%SPO + 80%HFO 5.003 1.783 311 601 1.531 5.157
50%SPO + 50%HFO 5.464 1.562 295 600 1.631 6.257
Sludge pyrolytic oil 4.360 1.162 274 605 1.114 5.832
4. Conclusions
The Taguchi method was used in this study to optimize the pyrolysis of sewage sludge for
obtaining the maximum pyrolytic oil yield. The combustion characteristic parameters are evaluated
based on a thermogravimetric analysis to explore the combustion characteristics of sludge pyrolytic oil
and heavy fuel oil. The findings of this study are summarized as follows.
264
Energies 2018, 11, 2260
(1) The effective sequence of pyrolytic parameters for the yield of sludge pyrolytic oil is nitrogen
flow rate, pyrolytic temperature, heating rate, and residence time. The best operating conditions
for sewage sludge pyrolysis are a heating rate of 10 ◦ C/min, pyrolysis temperature of 450 ◦ C,
residence time of 60 min, and nitrogen flow rate of 700 mL/min. Under these conditions,
the obtained pyrolytic oil yield is 10.19% (18.4% on dry ash free (daf) basis), which is very close
to 10.44% (18.85% on dry ash free (daf) basis.), the ideal value from the Taguchi method.
(2) From the thermogravimetric analysis of sludge pyrolytic oil, the combustion performance
parameters, such as the ignition temperature, burnout temperature, flammability index,
and combustion characteristics index, are calculated and compared with those of heavy fuel oil.
Since the sludge pyrolytic oil has volatile substances with lower boiling temperatures, it has a
lower ignition temperature and better combustion characteristics than heavy fuel oil.
(3) For the blends of sludge pyrolytic oil and heavy fuel oil, the maximum combustion rate,
the flammability index, and combustion characteristics index all increase markedly along with
the amount of sludge pyrolytic oil. Sludge pyrolytic oil significantly enhances the combustion
of heavy fuel oil, with the mix producing even better results than those seen with pure sludge
pyrolytic oil (synergistic effects).
Author Contributions: G.-B.C. and Y.-C.C. contributed to the concept, results explanation and writing of the
paper. J.-W.L. performed the experiments and analyzed the data. H.-T.L. and F.-H.W. contributed to experimental
method and paper review.
Funding: This research was funded by the Ministry of Science and Technology of Republic of China grant number
MOST 104-ET-E-006-001-ET.
References
1. Chen, G.B.; Li, Y.H.; Cheng, T.S.; Chao, Y.C. Chemical effect of hydrogen peroxide addition on characteristics
of methane-air combustion. Energy 2013, 55, 564–570. [CrossRef]
2. Starikovskiy, A.; Aleksandrov, N. Plasma-assisted ignition and combustion. Prog. Energy Combust. Sci. 2013,
39, 61–110. [CrossRef]
3. Kaminski, W.; Marszalek, J.; Ciolkowska, A. Renewable energy source-dehydrated ethanol. Chem. Eng. J.
2008, 13, 95–102. [CrossRef]
4. Nigam, P.S.; Singh, A. Production of liquid biofuels from renewable resources. Porg. Energy Combust. Sci.
2011, 37, 52–68. [CrossRef]
5. Werther, J.; Ogada, T. Sewage sludge combustion. Prog. Energy Combust. Sci. 1999, 25, 55–116. [CrossRef]
6. Cantrell, K.B.; Ducey, T.; Ro, K.S.; Hunt, P.G. Livestock waste-to-bioenergy generation opportunities.
7. Hossain, F.M.; Kosinkova, J.; Brown, R.J.; Ristovski, Z.; Hankamer, B.; Stephens, E.; Rainey, T.J. Experimental
investigations of physical and chemical properties for microalgae HTL bio-crude using a large batch reactor.
8. Kim, K.H.; Lee, E.Y. Environmentally-Benign Dimethyl Carbonate-Mediated Production of Chemicals and
Biofuels from Renewable Bio-Oil. Energies 2017, 10, 1790. [CrossRef]
9. Rowhani, A.; Rainey, T.J. Scrap tyre management pathways and their use as a fuel—A review. Energies 2016,
9, 888. [CrossRef]
10. Karagoz, S.; Bhaskar, T.; Muto, A.; Sakata, Y.; Oshiki, T.; Kishimoto, T. Low-temperature catalytic
hydrothermal treatment of wood biomass: Analysis of liquid products. Chem. Eng. J. 2005, 108, 127–137.
[CrossRef]
11. Asadullah, M.; Rahman, M.A.; Ali, M.M.; Motin, M.A.; Sultan, M.B.; Alam, M.R. Production of bio-oil from
fixed bed pyrolysis of bagasse. Fuel 2007, 86, 2514–2520. [CrossRef]
12. Ferreira, L.C.; Nilsen, P.J.; Fdz-Polanco, F.; Perez-Elvira, S.I. Biomethane potential of wheat straw: Influence
of particle size, water impregnation and thermal hydrolysis. Chem. Eng. J. 2014, 242, 254–259. [CrossRef]
265
Energies 2018, 11, 2260
13. Lewandowski, I.; Clifton-Brown, J.C.; Scurlock, J.M.O.; Huisman, W. Miscanthus: European experience with
a novel energy crop. Biomass Bioenergy 2000, 19, 209–227. [CrossRef]
14. Chen, G.L.; Chen, G.B.; Li, Y.H.; Wu, W.T. A study of thermal pyrolysis for castor meal using the Taguchi
method. Energy 2014, 71, 62–70. [CrossRef]
15. Chen, W.H.; Lin, B.J. Characteristics of products from the pyrolysis of oil palm fiber and its pellets in nitrogen
and carbon dioxide atmospheres. Energy 2016, 94, 569–578. [CrossRef]
16. Lam, S.S.; Liew, R.K.; Lim, X.Y.; Ani, F.N.; Jusoh, A. Fruit waste as feedstock for recovery by pyrolysis
technique. Int. Biodeterior. Biodegrad. 2016, 113, 325–333. [CrossRef]
17. Islam, M.N.; Beg, M.R.A.; Islam, M.R. Pyrolytic oil from fixed bed pyrolysis of municipal solid waste and its
characterization. Renew. Energy 2005, 30, 413–420. [CrossRef]
18. Lam, S.S.; Liew, R.K.; Jusoh, A.; Chong, C.T.; Ani, F.N.; Chase, H.A. Progress in waste oil to sustainable
energy, with emphasis on pyrolysis techniques. Renew. Sustain. Energy Rev. 2016, 53, 741–753. [CrossRef]
19. Mohan, D.; Pittman, C.U.; Steele, P.H. Pyrolysis of wood/biomass for bio-oil: A critical review. Energy Fuels
2006, 208, 48–89. [CrossRef]
20. Basu, P. Biomass Gasification, Pyrolysis and Torrefaction Practial Design and Theory, 2nd ed.; Academic Press:
Cambridge, MA, USA, 2013.
21. Fonts, I.; Gea, G.; Azuar, M.; Ábrego, J.; Arauzo, J. Sewage sludge pyrolysis for liquid production: A review.
22. Inguanzo, M.; Domínguez, A.; Menéndez, J.A.; Blanco, C.G.; Pis, J.J. On the pyrolysis of sewage sludge: The
influence of pyrolysis conditions on solid, liquid and gas fractions. J. Anal. Appl. Pyrolysis 2002, 63, 209–222.
[CrossRef]
23. Kim, Y.; Parker, W. A technical and economic evaluation of the pyrolysis of sewage sludge for the production
of bio-oil. Bioresour. Technol. 2008, 99, 1409–1416. [CrossRef] [PubMed]
24. Domínguez, A.; Menéndez, J.A.; Inguanzo, M.; Bernad, P.L.; Pis, J.J. Gas chromatographic-mass spectrometric
study of the oil fractions produced by microwave-assisted pyrolysis of different sewage sludges.
J. Chromatogr. A 2003, 1012, 193–206. [CrossRef]
25. Domínguez, A.; Menéndez, J.A.; Inguanzo, M.; Pis, J.J. Investigations into the characteristics of oils produced
from microwave pyrolysis of sewage sludge. Fuel Process Technol. 2005, 86, 1007–1020. [CrossRef]
26. Domínguez, A.; Menéndez, J.A.; Inguanzo, M.; Pis, J.J. Production of bio-fuels by high temperature pyrolysis
of sewage sludge using conventional and microwave heating. Bioresour. Technol. 2006, 97, 1185–1193.
[CrossRef] [PubMed]
27. Sánchez, M.E.; Menéndez, J.A.; Domínguez, A.; Pis, J.J.; Martínez, O.; Calvo, L.F.; Bernad, P.L. Effect of
pyrolysis temperature on the composition of the oils obtained from sewage sludge. Biomass Bioenergy 2009,
33, 933–940. [CrossRef]
28. Jeguirim, M.; Dutournié, P.; Zorpas, A.A.; Limousy, L. Olive mill wastewater: From a pollutant to green fuels,
agricultural water source and bio-fertilizer—Part 1. The drying kinetics. Energies 2017, 10, 1423. [CrossRef]
29. Azuara, M.; Ábrego, J.; Fonts, I.; Gea, G.; Murillo, M.B. Ammonia Content of Bottom Phase Liquid from
Pyrolyisis of Sewage Sludge in a Bubbling Fluidized Bed. In Proceedings of the 18th European Biomass
Conference and Exhibition, Lyon, France, 3–7 May 2010.
30. Cao, J.P.; Zhao, X.Y.; Morishita, K.; Li, L.Y.; Xiao, X.B.; Obara, R.; Wei, X.Y.; Takarada, T. Triacetonamine
formation in a bio-oil from fast pyrolysis of sewage sludge using acetone as the absorption solvent.
31. Tang, B. Orthogonal array-based hypercube. J. Am. Statist. Assoc. 1993, 88, 1392–1397. [CrossRef]
32. Yang, W.H.; Tarng, Y.S. Design optimization of cutting parameters for turning operations based on the
Taguchi method. J. Mater. Process. Technol. 1998, 84, 122–129. [CrossRef]
33. Ghani, J.A.; Choudhury, I.A.; Hassan, H.H. Application of Taguchi method in the optimization of end milling
parameters. J. Mater. Process. Technol. 2004, 145, 84–92. [CrossRef]
34. Tognotti, L.; Malotti, A.; Petarca, L.; Zanelli, S. Measurement of ignition temperature of coal particles using a
thermogravimetric technique. Combust. Sci. Technol. 1985, 15, 15–28. [CrossRef]
35. Crelling, J.C.; Hippo, E.J.; Woerner, B.A.; West, D.P., Jr. Combustion characteristics of selected whole coals
and macerals. Fuel 1992, 71, 151–158. [CrossRef]
36. Huang, X.; Jiang, X.; Han, X.; Wang, H. Combustion characteristics of fine- and micro-pulverizes coal in the
misture of O2 /CO2 . Energy Fuels 2008, 22, 3756–3762. [CrossRef]
266
Energies 2018, 11, 2260
37. Wall, T.F.; Gupta, R.P.; Gururajan, V.S.; Zhang, D. The ignition of coal particles. Fuel 1991, 70, 1011–1016.
[CrossRef]
38. Jin, Y.; Li, Y.; Liu, F. Combustion effects and emission characteristics of SO2 , CO, NOx and heavy metals
during co-combustion of coal and dewatered sludge. Front. Environ. Sci. Eng. 2015, 10, 201–210. [CrossRef]
39. Cheng, J.Y.; Sun, X.X. Determination of the devolatilization index and combustion characteristic index of
pulverized coals. Power Eng. 1987, 7, 13–18.
40. Wang, Y.; Chen, G.; Li, Y.; Yan, B.; Pan, D. Experimental study of the bio-oil production from sewage sludge
by supercritical conversion process. Waste Manag. 2013, 33, 2408–2415. [CrossRef] [PubMed]
41. Thipkhunthod, P.; Meeyoo, V.; Rangsunvigit, P.; Kitiyanan, B.; Siemanond, K.; Rirksomboon, T. Pyrolytic
characteristics of sewage sludge. Chemosphere 2006, 64, 955–962. [CrossRef] [PubMed]
42. Gómez-Rico, M.F.; Font, R.; Fullana, A.; Martín-Gullón, I. Thermogravimetric study of different sewage
sludges and their relationship with the nitrogen content. J. Anal. Appl. Pyrolysis 2005, 74, 421–428. [CrossRef]
43. Kristensen, E. Characterization of biogenic organic natter by stepwise thermogravimetry (STG).
Biogeochemistry 1990, 9, 135–139. [CrossRef]
44. Li, J.J.; Qi, B.Y.; Li, A.M.; Duan, Y.; Wang, Z. Thermal analysis and products distribution of dried sewage
sludge pyrolysis. J. Anal. Appl. Pyrolysis 2014, 105, 43–48.
45. Pütün, E. Catalytic pyrolysis of biomass: Effects of pyrolysis temperature, sweeping gas flow rate and MgO
catalyst. Energy 2010, 35, 2761–2766.
46. Chen, W.H.; Chen, C.J.; Hung, C.I. Taguchi approach for co-gasification optimization of torrefied biomass
and coal. Bioresour. Technol. 2013, 144, 615–622. [CrossRef] [PubMed]
47. Das, S.P.; Gupta, A.; Das, D.; Goyal, A. Enhanced bioethanol production from water hyacinth (Eichhornia
crassipes) by statistical optimization of fermentation process parameters using Taguchi orthogonal array
design. Int. Biodeterior. Biodegrad. 2016, 109, 174–184. [CrossRef]
48. Pattanaik, A.; Satpathy, M.P.; Mishra, S.C. Dry sliding wear behavior of epoxy fly ash composite with Taguchi
optimization. Eng. Sci. Technol. Int. J. 2016, 19, 710–716. [CrossRef]
49. Huang, F.; Yu, Y.; Huang, H. Temperature influence and distribution of bio-oil from pyrolysis of granular
sewage sludge. J. Anal. Appl. Pyrolysis 2018, 130, 36–42. [CrossRef]
50. Li, J.W. A Study of Thermal Process for Sewage Sludge Pyrolytic Oil and its Combustion Characteristics.
Master’s Thesis, National Cheng Kung University, Tainan, Taiwan, 2016.
267
energies
Article
Combined Ball Milling and Ethanol Organosolv
Pretreatment to Improve the Enzymatic Digestibility
of Three Types of Herbaceous Biomass
Seong Ju Kim 1 , Byung Hwan Um 2 , Dong Joong Im 3 , Jin Hyung Lee 4 and Kyeong Keun Oh 3,5, *
1 Department of Biomolecular and Chemical Engineering, Hankyong National University, Anseong,
Gyeonggi 17579, Korea; [email protected]
2 Department of Chemical Engineering and Research Center of Chemical Technology, Hankyong National
University, Anseong, Gyeonggi 17579, Korea; [email protected]
3 R&D Center, SugarEn Co., Ltd., Yongin, Gyeonggi 16890, Korea; [email protected]
4 Korea Institute of Ceramic Engineering and Technology, Jinju, Gyeongnam 52851, Korea; [email protected]
5 Department of Chemical Engineering, Dankook University, Yongin, Gyeonggi 16890, Korea
Received: 31 August 2018; Accepted: 14 September 2018; Published: 16 September 2018
Abstract: A combined ball milling and ethanol organosolv process is proposed for the pretreatment
of three types of herbaceous biomass, giant miscanthus, corn stover, and wheat straw. The combined
pretreatment was effective at both removing lignin and increasing the glucan content. After 120 min
pretreatment, the glucan content increased to 63.09%, and 55.89% of the acid-insoluble lignin was
removed from the giant miscanthus sample. The removal of cellulose, hemicellulose, and acetyl
groups were correlated with the removal of lignin. The pretreatment of corn stover showed the highest
removal of cellulose, but this was dependent on the removal of acid-insoluble lignin. The slope
of the regression lines, which shows the correlation between the removal of lignin and cellulose,
was lower than other correlations. The changes in biomass size were analyzed using size distribution
graphs. With increasing pretreatment time, the particle size reduction improved in the three types
of herbaceous biomass. Because of the combined physicochemical pretreatment, the enzymatic
digestibility improved, and a maximum of 91% glucan digestibility was obtained from the pretreated
corn stover when 30 FPU/g-glucan enzyme was added. Finally, compositional analysis of the
recovered lignin from the remaining black liquor was investigated.
Keywords: combined pretreatment; ball mill; ethanol organosolv; herbaceous biomass;

lignin recovery
1. Introduction
Sugars derived from herbaceous biomass are important platform chemicals and can be further
converted to high-value-added chemicals and biofuels via thermochemical or biological routes [1,2].
Pretreatment is an essential step in producing sugars from herbaceous biomass. Pretreatment reduces
the amount of the structural polymer lignin, as well as the crystalline cellulose structure, thus increasing
the accessibility of enzymes to the holocellulose [3,4]. Recently, lignin has emerged as a renewable
source of aromatics for the chemical industry [5]. Therefore, both carbohydrate and lignin fractionation
should be considered when pretreatment steps are applied to herbaceous biomass.
Pretreatment consists of biological, chemical, or physical processes [6,7]. Recently, combined
physical and chemical pretreatment processes have been proposed to enhance enzymatic
hydrolysis [8–12]. These combined methods increase enzymatic accessibility and reduce energy
consumption [13].

Energies 2018, 11, 2457
Ball milling is one physical pretreatment method. Previous studies have reported that ball milling
increases the accessible surface area and pore size and reduces the crystallinity and the polymerization
degree of cellulose [14–17]. Because chemical catalysts are not used, there is no significant change
in the chemical composition [18]. As a chemical pretreatment approach, organosolv is a promising
method for the alcoholysis of lignin from herbaceous biomass [19]. This pretreatment employs organic
solvents, mainly alcohols, or their aqueous solutions, which can be later reused and recovered [20,21].
The advantage of using organosolv is that highly pure lignin is obtained as a solid from the liquid phase;
this can be used for high-value products such as adhesives, fibers, and biodegradable polymers [20,22]
or thermal fuel [23]. During the organosolv process, cellulose is recovered as a solid phase, whereas
most of the lignin and partial hemicellulose is dissolved in the organic solvent as a liquid phase.
The recovered solid, rich in cellulose, is readily hydrolyzed by enzymatic digestion because the lignin
or hemicellulose has been removed, and this material can be used for further processing [24–27].
Finally, the fractionation of lignin and partial hemicellulose reduces the recalcitrance of the material
to digestion and increases the accessibility of the material to enzymes, thereby enhancing enzymatic
digestibility [28]. However, the two-stage pretreatment (physical and chemical) requires considerable
energy, as well as additional reactors. Previous studies have reported that pretreatment combined
with mechanical size reduction is effective in reducing the energy consumption in biorefinery
plants [14,29–33].
In this study, we propose a novel pretreatment process where ball mill grinding and the
ethanol-based organosolv process are conducted simultaneously. In this process, ball milling is used as a
physical treatment to reduce the biomass size and improve enzymatic accessibility. Ethanol organosolv
was used as a chemical treatment to fractionate the lignin, the major obstacle to cellulose digestion.
In this study, the effects of the combined physicochemical pretreatment on fermentable sugar
production were investigated using three types of herbaceous biomass.
2.1. Raw Materials

Three herbaceous biomasses namely, giant miscanthus (GM), corn stover (CS), and wheat
straw (WS) were supplied by the Rural Development Administration (Wanju, Republic of Korea).
The samples were air dried at room temperature and used directly in the pretreatment processes.
The moisture contents of GM, CS, and WS were 2.92%, 3.08%, and 3.20%, based on the total wet
biomass weight, respectively. The chemical compositions of the raw materials are shown in Table 1.
Table 1. Compositions of the three types of herbaceous biomass used in this study. GM: Giant
Miscanthus; CS: corn stover; WS: wheat straw.
Compositions (%) GM CS WS
Glucan 43.77 ± 0.07 30.30 ± 0.26 31.42 ± 0.44
Xylan 21.22 ± 0.08 16.58 ± 0.19 21.38 ± 0.09
Mannan - 0.60 ± 0.21 -
Carbohydrates
Galactan - 1.53 ± 0.47 1.56 ± 0.14
Arabinan 0.50 ± 0.01 0.82 ± 0.02 1.00 ± 0.02
Sub Total 65.49 49.83 55.36
Acid soluble 0.96 ± 0.01 1.47 ± 0.03 1.11 ± 0.02
Lignin Acid insoluble 19.55 ± 0.01 12.38 ± 0.09 13.47 ± 0.24
Sub Total 20.51 13.85 14.58
Water 4.96 ± 0.71 21.14 ± 0.12 17.11 ± 0.33
Extractive Ethanol 1.47 ± 0.11 3.71 ± 0.22 2.77 ± 0.27
Sub Total 6.43 24.85 19.88
Acetyl group 3.88 ± 0.02 2.76 ± 0.06 3.02 ± 0.06
Ash 2.18 ± 0.03 7.92 ± 0.05 7.36 ± 0.10
Total 98.49 99.21 100.2
269
Energies 2018, 11, 2457
2.2. Combined Physicochemical Pretreatment

The combined ball milling and ethanol organosolv pretreatment was performed in a
rotary-pressured type reactor (Hanul Engineering Co., Gunpo, Korea). For the pretreatment, 370 g
of dried herbaceous biomass and 11 kg of alumina balls (Ø = 10 mm) were added to the reactor
with 3.7 L of 60% (v/v) ethanol solution (1:10:30 (w/v/w) for biomass, solvent, and ball, respectively).
The rotary-pressure-type reactor was kept at 170 ◦ C for 30–120 min and rotated at 50 rpm. After reaction,
it was cooled to 80 ◦ C, and the mixture was separated into solid biomass, lignin-containing liquid,
and alumina balls. The treated solid biomass was washed with distilled water and stored until use.
After pretreatment, the particle size distribution of the pretreated biomass was determined by the
shaking sieve method after drying at 45 ◦ C overnight. The sieves used were US sieve numbers 6,
14, 20, 40, 80, 170, and 270 (sieve opening sizes: 3.35, 1.4, 0.85, 0.425, 0.180, 0.09, and 0.053 mm,
respectively). Lignin was recovered from the remaining black liquor by 3-fold dilution with distilled
water. After 24 h, the precipitated lignin was collected by filtration.
2.3. Enzymatic Digestibility

Enzymatic digestibility was analyzed according to the National Renewable Energy Laboratory
(NREL) standard procedures [34]. A commercial cellulase (Cellic CTec 2, Novozymes Korea Ltd., Seoul,
Korea) was used. The enzyme hydrolysis was performed at 50 ◦ C and pH 4.8 (50 mM sodium citrate
buffer) while the solution was shaken at 150 rpm. The quantity of cellulase loaded was 15 or 30 FPU
(filter paper unit)/g-glucan. Experiments were conducted in 250-mL Erlenmeyer flasks with a total
working volume of 100 mL, and the glucan concentration was maintained at 1.0% (w/v). Samples were
taken periodically (0, 3, 6, 9, 12, 24, 48, 72, and 96 h) and boiled for 5 min to deactivate the enzymes,
followed by filtration through a 0.45 μm nylon membrane filter for glucose content analysis using a
high-performance liquid chromatography (HPLC) column (Bio-Rad Laboratories, Hercules, CA, USA).
All experiments were performed in triplicate for error analysis. Finally, the glucan digestibility was
calculated using Equation (1).
Lignin removal percentage (%) = 100 − mRe /mL × 100 (1)
where mRe refers to the dry weight mass of remaining lignin after composition analysis of solid residues
and mL represents the mass weight of lignin in the original loaded biomass (mL = the loaded biomass
mass weight × the lignin content) [35].
2.4. Compositional Analysis of the Solid and Liquid Phases

The compositions of the solid and liquid samples were determined according to
analytical procedures of the NREL (NREL/TP-510-42623 for structural carbohydrates and lignins;
NREL/TP-510-42618 for sugars in the liquids or in the hydrolyzates) [36,37]. The sugars were
determined using a HPLC (Agilent 1260 Infinity, Agilent Technologies, Santa Clara, CA, USA) equipped
with a refractive index detector (Agilent 1260 Infinity, Agilent Technologies, Santa Clara, CA, USA).
A Bio-Rad Aminex HPX-87H column (300 mm length × 7.8 mm internal diameter) and a Cation H
micro-guard cartridge (30 mm length × 4.6 mm internal diameter, Bio-Rad Laboratories Inc., Hercules,
CA, USA) were used for sugar, organic acid, and decomposition product analysis. The samples were
filtered using syringe filters having a 0.45-μm pore size before analysis, and 5 mM sulfuric acid was
used as the mobile phase. The temperature of the column was set at 65 ◦ C and the flow rate of the
mobile phase was 0.5 mL/min. The remaining solid was filtered, dried, and burned in a muffle furnace
at 575 ◦ C for lignin analysis. The ash content of the samples was determined by complete combustion
in a muffle furnace (DMF12, Romax, Co., Seoul, Korea) equipped with a temperature controller and by
running a temperature ramp program according to the NREL/TP-510-42622 method. The remaining
residue in the crucible was taken as the ash content [38]. The compositional analysis was performed
in triplicate.
270
Energies 2018, 11, 2457
3. Results
3.1. Compositions of Herbaceous Biomasses Used in This Study

The compositions of the herbaceous biomass samples used in this study are presented in Table 1.
GM had the highest carbohydrate and lignin content among the three types of herbaceous biomass:
43.8% glucan, 21.22% xylan, mannan, and galactan (XMG), and 20.51% lignin (acid soluble + acid
insoluble). In contrast, CS and Wheat straw (WS) contained approximately 50% carbohydrates and
14% lignin. However, the extractive content of GM was much lower than those of CS and WS.
The differences in biomass composition could be due to the characteristics of each herbaceous biomass
source. These differences in composition result in unique physical/chemical properties. Therefore,
it is important to find the optimum pretreatment conditions that are tailored to the properties of each
biomass type.
3.2. Composition Changes after Combined Physicochemical Pretreatment

The goals of pretreatment are to reduce the lignin content to improve the accessibility of
the enzyme and to increase the glucan and XMG content. Wanting to optimize the combined
physicochemical pretreatment conditions, we investigated the conditions where lignin removal was
maximal while the cellulose content was maintained.
Table 2 presents the compositional changes with increasing reaction time of the GM sample after
pretreatment. The combined pretreatment by ball mill and ethanol organosolv was effective in both
removing acid-insoluble lignin (AIL) and increasing glucan content in GM (Table 2). With increasing
reaction time, the glucan content increased to 63.09% after 120 min reaction. Removal of acid-insoluble
lignin increased with reaction time. The XMG contents decreased when the reaction time increased
from 90 to 120 min (Table 2), and the acetyl group contents also decreased. However, the AIL content
did not decrease but the glucan content increased by about 6% when the remaining solid is considered.
Table 2. Changes in composition of giant miscanthus after the combined ball mill and organosolv
pretreatment performed at 170 ◦ C at 50 rpm rotation. XMG: xylan, mannan, and galactan; AIL:
acid-insoluble lignin.
Time Solid Remaining Glucan XMG 1 Acetyl Group AIL

AIL Removal
(min) (%) (%) (%) (%) (%)
Percentage (%)
Initial 100 43.77 21.22 3.88 19.53
30 79.25 50.01 23.03 3.73 16.28 33.95
60 72.29 52.22 21.79 3.36 15.89 41.18
90 65.41 56.90 20.72 3.04 14.30 52.11
120 57.44 63.09 15.14 2.10 15.00 55.89
Table 3 presents the changes in the composition of corn stover before and after the combined
physicochemical pretreatment with increasing reaction time. The AIL removal yield of corn stover
linearly increased to 52.04% during 90 min reaction. The acetyl group content continuously decreased
from 2.76% to 1.65% and glucan content steadily increased over 120 min reaction. However, the XMG
content changed from 23.89% to 22.08% from 30 to 90 min of reaction time, respectively. The reason
for this change could be the removal of extractives by ethanol. As shown for giant miscanthus,
the AIL removal slightly increased to 53.91% after 120 min reaction when compared to that after
90 min reaction.
271
Energies 2018, 11, 2457
Table 3. Change in composition of corn stover after the combined ball mill and organosolv pretreatment
performed at 170 ◦ C with 50 rpm rotation.
Time Solid Remaining Glucan XMG Acetyl Group AIL

AIL Removal
(min) (%) (%) (%) (%) (%)
Percentage (%)
Initial 100 30.30 18.71 2.76 12.23
30 56.98 46.51 23.89 2.62 14.61 31.92
60 48.32 48.58 22.91 2.18 14.97 40.84
90 42.24 51.73 22.08 1.93 13.89 52.04
120 39.28 53.16 20.27 1.65 14.35 53.91
Table 4 shows the compositional changes of wheat straw. The removal percentage of AIL reached
48.39% after 120 min, which is less than those of GM and CS. Like the other biomass types, the XMG
and acetyl group contents steadily decreased but the glucan content increased with reaction time, up to
120 min. The AIL content did not seem to change significantly. However, considering the relative
content in the remaining solid, the AIL content was greatly decreased.
Table 4. Change in composition of wheat straw after the combined ball mill and organosolv
pretreatment performed at 170 ◦ C with 50 rpm rotation.
Time Solid Remaining Glucan XMG Acetyl Group AIL

AIL Removal
(min) (%) (%) (%) (%) (%)
Percentage (%)
Initial 100 31.42 22.94 3.02 14.61
30 64.47 41.36 27.83 2.61 14.53 35.90
60 58.07 44.93 26.76 2.45 15.19 39.62
90 52.06 48.78 24.13 2.29 15.55 44.58
120 50.70 50.15 22.71 1.97 14.87 48.39
In previous studies, the pretreatment of lignocellulosic biomasses was carried out using acidic
catalysts. El Hage et al. obtained 62% AIL removal of Miscanthus samples when using 44% EtOH with
0.5% H2 SO4 catalyst [39]. Huijgen et al. removed lignin from wheat straw using an acid catalyst and
obtained a 56% removal yield [40]. However, using an acid catalyst caused the dissolution of glucan
into the liquid, and the acidic solution corroded the pretreatment equipment. In contrast, the combined
physicochemical pretreatment used in this study did not cause a loss of glucan and showed good
lignin removal yields.
After 30 min reaction time, of the three herbaceous biomass types, the highest percentage
of remaining solid was found for GM. The percentages of remaining solid were 79.25%, 56.98%,
and 64.47% for GM., CS, and WS, respectively. The remaining solid could be related to the extractive
content. The extractive contents (water + ethanol) were 6.43%, 24.85%, and 19.88% for GM, CS, and WS,
respectively, after 30 min pretreatment. The extractives were drawn into the liquid by ethanol during
the combined physicochemical pretreatment. The remaining solid further decreased with reaction
time, and the lowest value was 39.28% in the case of CS because it had the highest extractive content
and, thus, underwent the highest removal of the main components.
3.3. Correlation of Lignin Removal with the Removal of Other Compounds

Using the combined physicochemical pretreatment, lignin was mostly removed,
and cellulose/hemicellulose were partially removed. The robustness (or recalcitrance) of the
lignocellulosic biomass is attributable to the crosslinking between the polysaccharides (cellulose and
hemicellulose) and lignin via ester and ether linkages [41]. During the combined physicochemical
pretreatment, the polysaccharides were first disconnected from lignin. The newly freed celluloses
became dissolved in the ethanol and were broken down into monosaccharides. Simultaneously,
272
Energies 2018, 11, 2457
lignin and hemicellulose were concurrently fractionated. The correlation between lignin removal
with the removal of the various components is presented in Figure 1. It was found that the removal
of cellulose, hemicellulose, and acetyl groups is correlated with lignin removal. As shown by the
regression lines in Figure 1a, the dissolution of cellulose into the liquor occurred simultaneously
with the removal of AIL. With increasing AIL removal, the cellulose dissolution also increased.
The regression line for CS has a different slope compared to those of the other herbaceous biomass
samples (Figure 1). CS showed higher removal of cellulose depending on the removal of AIL.
For example, the percentages of removed cellulose were 66%, 35%, and 29% for CS, WS, and GM,
respectively, at the 100% AIL removal point on the regression lines (Figure 1a). The correlation
between lignin removal and hemicellulose removal is different from the correlation between the
removal of lignin and cellulose. The slopes of the regression lines, which show the correlation
between the removal of lignin and hemicellulose, is higher than that of lignin and cellulose because
the hemicellulose is more easily removed because of the weak and amorphous structure. From our
results, we determined that hemicellulose was rapidly removed when the AIL was dissolved during
the combined ball milling and organosolv pretreatment. The correlation between the removal of
lignin and acetyl groups also showed a similar pattern to that of hemicellulose. The cleavage of acetyl
groups occurred rapidly, and the acetyl groups were also removed when AIL was removed (Figure 1c).
Using the combined physicochemical pretreatment, the removal of crosslinking between lignin and
the other compounds (cellulose, hemicellulose, and acetyl group) occurred initially. Subsequently,
the components were dissolved and removed. However, complete AIL removal was difficult because
of the strong crosslinks between the polysaccharides and lignin.
D E F

*LDQW0LVFDQWKXV *LDQW0LVFDQWKXV *LDQW0LVFDQWKXV
&RUQ6WRYHU &RUQ6WRYHU &RUQ6WRYHU
:KHDW6WUDZ :KHDW6WUDZ :KHDW6WUDZ

+HPLFHOOXORVHUHPRYDO
$FHW\OJURXSUHPRYDO
&HOOXORVHUHPRYDO

$FLGLQVROXEOHOLJQLQUHPRYDO $FLGLQVROXEOHOLJQLQUHPRYDO $FLGLQVROXEOHOLJQLQUHPRYDO
Figure 1. Correlation between the removal of lignin and other compounds: (a) cellulose;
(b) hemicellulose; and (c) acetyl group by the combined physicochemical pretreatment.
3.4. Changes in Biomass Particle Size

The change in the size of the biomass particles with pretreatment time was investigated and
expressed by calculating the number of particles that passed through sieves of different mesh sizes.
All untreated herbaceous biomass samples showed positive skewness (Figure 2). With increasing
pretreatment time, the size distribution curves became negatively skewed. The particle sizes after
30 and 60 min pretreatment did not significantly change when compared to those of the untreated
samples (data not shown). However, significant changes to the size distribution curves were found
after 90 min pretreatment time. Many untreated GM particles were collected after sieving through a
20-mesh sieve. With increasing pretreatment time, the relative amount of GM increased between the
40- and 170-mesh sieves. After 120 min pretreatment, a significant increase was found for the 40-mesh
and 170-mesh sieves, indicating that the GM particles had become finer. Figure 2b shows that the
relative amount of CS particles passing through the 80-mesh sieve drastically increased from 0.02% to
0.4% after 120 min pretreatment. Meanwhile, the relative number of CS particles larger than 20 mesh
273
Energies 2018, 11, 2457
was reduced by the pretreatment. Similar results were also found for WS (Figure 2c). In this study,
the ball milling process ground the biomass because of the friction between the ball and the biomass.
Based on our results, ball milling was effective in pretreating the three types of herbaceous biomass
used in this study.
D E F

PLQ8QWUHDWHG PLQ8QWUHDWHG PLQ8QWUHDWHG
PLQ PLQ PLQ
PLQ PLQ PLQ
5HODWLYHDPRXQW
5HODWLYHDPRXQW
5HODWLYHDPRXQW

! !! !! !! !! !! ! ! !! !! !! !! !! !! ! !! !! !! !! !! !
3DUWLFOHVL]HPHVK 3DUWLFOHVL]HPHVK 3DUWLFOHVL]HPHVK
Figure 2. Biomass particle size distributions before and after the combined physicochemical
pretreatment at different reaction times: (a) GM; (b) CS; and (c) WS.
3.5. Enzymatic Digestibility of Pretreated Herbaceous Biomasses

To evaluate the effect of pretreatment on cellulose accessibility, the enzyme digestibility was
compared before and after the combined physicochemical pretreatment. Untreated GM showed very
low glucan digestibility, less than 10%. Even when the enzyme dosage was doubled, the glucose
production did not increase significantly. However, the combined physicochemical pretreatment
significantly improved the glucan digestibility. For the pretreated samples, a five-fold increase in
glucan digestibility was obtained compared to that of the untreated sample (Figure 3a). In addition,
the glucan digestibility increased with increasing enzyme dosage. Untreated CS showed relatively
higher glucan digestibility than the other two types of herbaceous biomass. However, the glucan
digestibility drastically increased after the combined physicochemical pretreatment. Specifically,
treated CS showed 91% glucan digestibility when the enzyme dose was 30 FPU/g-glucan (Figure 3b).
Untreated WS showed less than 20% glucan digestibility, even after the addition of 30 FPU/g-glucan
(Figure 3c). The treated WS showed 79% and 88% glucan digestibility for enzyme doses of 15 and
30 FPU/g-glucan, respectively, as shown in Figure 3c. These results indicate that the combined
physicochemical pretreatment used in this study increased the cellulose surface area available for
reaction with cellulase, improving enzyme accessibility.
D E F

8QWUHDWHG)38JJOXFDQ
8QWUHDWHG)38JJOXFDQ
7UHDWHG)38JJOXFDQ

(Q]\PDWLFGLJHVWLELOLW\
7UHDWHG)38JJOXFDQ

8QWUHDWHG)38JJOXFDQ
8QWUHDWHG)38JJOXFDQ
7UHDWHG)38JJOXFDQ 8QWUHDWHG)38JJOXFDQ
7UHDWHG)38JJOXFDQ 8QWUHDWHG)38JJOXFDQ
7UHDWHG)38JJOXFDQ
7UHDWHG)38JJOXFDQ

7LPHK 7LPHK 7LPHK
Figure 3. Glucan digestibility before and after pretreatment depending on enzyme dosage: (a) GM;
(b) CS; and (c) WS.
274
Energies 2018, 11, 2457
3.6. Composition of Recovered Lignin from the Black Liquor

Lignin is byproduct in this process, but it is a very useful compound having biopolymer
applications [42]. With increasing pretreatment time from 30 to 120 min in GM, the recovered lignin
(AIL and acid-soluble lignin (ASL)) increased from 57.61% to 80.26% (Table 5). In contrast, the contents
of glucan and XMG decreased. The contents of recovered lignin were 71.29% and 64.86% for CS
and WS, respectively, after 120 min pretreatment. With increasing pretreatment time, the quantity of
recovered lignin increased significantly for GM but increased only slightly for CS and WS.
Table 5. Composition of lignin obtained from the black liquor depending on pretreatment time. ASL:
acid-soluble lignin; AIA: acid-insoluble ash.
Biomass Time (min) AIL (%) ASL (%) AIA (%) Glucan (%) XMG (%) Acetyl Group (%)
30 55.62 1.99 5.39 14.57 12.60 2.59
60 59.02 1.85 5.41 10.74 13.00 2.65
GM
90 62.27 2.06 6.57 11.34 13.96 2.71
120 77.94 2.32 1.58 2.66 11.87 2.73
30 59.37 3.07 2.90 5.83 5.56 1.14
60 64.28 3.29 2.34 5.76 6.23 1.37
CS
90 56.34 3.78 2.32 6.14 7.28 1.42
120 67.81 3.48 2.39 4.68 6.28 1.24
30 50.91 3.65 2.97 10.47 12.24 1.93
60 54.42 2.78 4.30 10.10 11.44 1.69
WS
90 55.38 3.03 3.22 8.16 12.97 1.97
120 61.66 3.20 2.32 5.94 12.24 1.90
4. Conclusions
A combined ball milling and ethanol organosolv pretreatment was effective in both increasing the
glucan content and removing the acid-insoluble lignin. With increasing pretreatment time, the glucan
content and lignin recovery were improved in all three types of herbaceous biomass. Lignin removal
was correlated with the removal of cellulose, hemicellulose, and acetyl groups but the slope of the plot
of cellulose removal to lignin removal was lower than those of the others. In summary, the combined
pretreatment significantly improved the enzyme digestibility of all herbaceous biomass samples tested
in this study.
Author Contributions: S.J.K., B.H.U., D.J.I., J.H.L., and K.K.O. contributed equally to this work. S.J.K. and D.J.I.
contributed to the experimental process for combined fractionation process. B.H.U. and K.K.O. contributed to the
project administration and experimental design. In addition, J.H.L. contributed to providing methodology and
data validation. All the authors contributed to the writing and review of this document.
Funding: This work was funded by a grant the Korea Institute of Energy Technology Evaluation and Planning
(KETEP) (Grant No. 2015010091990).
References
1. Wettstein, S.G.; Alonso, D.M.; Gürbüz, E.I.; Dumesic, J.A. A roadmap for conversion of lignocellulosic
biomass to chemicals and fuels. Curr. Opin. Chem. Eng. 2012, 1, 218–224. [CrossRef]
2. Sheldon, R.A. Green and sustainable manufacture of chemicals from biomass: State of the art. Green Chem.
2014, 16, 950–963. [CrossRef]
3. Sun, Y.; Cheng, J. Hydrolysis of lignocellulosic materials for ethanol production: A review. Bioresour. Technol.
2002, 83, 1–11. [CrossRef]
4. Kumar, P.; Barrett, D.M.; Delwiche, M.J.; Stroeve, P. Methods for pretreatment of lignocellulosic biomass for
efficient hydrolysis and biofuel production. Ind. Eng. Chem. Res. 2009, 48, 3713–3729. [CrossRef]
5. Cazacu, G.; Capraru, M.; Popa, V.I. Advances concerning lignin utilization in new materials. In Advances in
Natural Polymers; Springer: Berlin, Germany, 2013; pp. 255–312.
275
Energies 2018, 11, 2457
6. Menon, V.; Rao, M. Trends in bioconversion of lignocellulose: Biofuels, platform chemicals & biorefinery
concept. Prog. Energy Combust. Sci. 2012, 38, 522–550. [CrossRef]
7. Taherzadeh, M.J.; Karimi, K. Pretreatment of lignocellulosic wastes to improve ethanol and biogas production:
A review. Int. J. Mol. Sci. 2008, 9, 1621–1651. [CrossRef] [PubMed]
8. Zakaria, M.R.; Hirata, S.; Hassan, M.A. Combined pretreatment using alkaline hydrothermal and ball milling
to enhance enzymatic hydrolysis of oil palm mesocarp fiber. Bioresour. Technol. 2014, 169, 236–243. [CrossRef]
[PubMed]
9. Kim, S.M.; Dien, B.S.; Tumbleson, M.E.; Rausch, K.D.; Singh, V. Improvement of sugar yields from corn
stover using sequential hot water pretreatment and disk milling. Bioresour. Technol. 2016, 216, 706–713.
[CrossRef] [PubMed]
10. Barakat, A.; Chuetor, S.; Monlau, F.; Solhy, A.; Rouau, X. Eco-friendly dry chemo-mechanical pretreatments
of lignocellulosic biomass: Impact on energy and yield of the enzymatic hydrolysis. Appl. Energy 2014, 113,
11. Yuan, Z.; Long, J.; Wang, T.; Shu, R.; Zhang, Q.; Ma, L. Process intensification effect of ball milling on
the hydrothermal pretreatment for corn straw enzymolysis. Energy Convers. Manag. 2015, 101, 481–488.
[CrossRef]
12. Deng, A.; Ren, J.; Wang, W.; Li, H.; Lin, Q.; Yan, Y.; Sun, R.; Liu, G. Production of xylo-sugars from corncob
by oxalic acid-assisted ball milling and microwave-induced hydrothermal treatments. Ind. Crops Prod. 2016,
79, 137–145. [CrossRef]
13. Huo, D.; Fang, G.; Yang, Q.; Han, S.; Deng, Y.; Shen, K.; Lin, Y. Enhancement of eucalypt chips’ enzymolysis
efficiency by a combination method of alkali impregnation and refining pretreatment. Bioresour. Technol.
2013, 150, 73–78. [CrossRef] [PubMed]
14. Da Silva, A.S.; Inoue, H.; Endo, T.; Yano, S.; Bon, E.P.S. Milling pretreatment of sugarcane bagasse and straw
for enzymatic hydrolysis and ethanol fermentation. Bioresour. Technol. 2010, 101, 7402–7409. [CrossRef]
[PubMed]
15. Inoue, H.; Yano, S.; Endo, T.; Sakaki, T.; Sawayama, S. Combining hot-compressed water and ball milling
pretreatments to improve the efficiency of the enzymatic hydrolysis of eucalyptus. Biotechnol. Biofuels 2008,
1, 2. [CrossRef] [PubMed]
16. Lin, Z.; Huang, H.; Zhang, H.; Zhang, L.; Yan, L.; Chen, J. Ball milling pretreatment of corn stover for
enhancing the efficiency of enzymatic hydrolysis. Appl. Biochem. Biotechnol. 2010, 162, 1872–1880. [CrossRef]
[PubMed]
17. Silva, G.G.D.; Couturier, M.; Berrin, J.-G.; Buléon, A.; Rouau, X. Effects of grinding processes on enzymatic
degradation of wheat straw. Bioresour. Technol. 2012, 103, 192–200. [CrossRef] [PubMed]
18. Lee, J.H.; Kwon, J.H.; Kim, T.H.; Choi, W. Impact of planetary ball mills on corn stover characteristics and
enzymatic digestibility depending on grinding ball properties. Bioresour. Technol. 2017, 241, 1094–1100.
[CrossRef] [PubMed]
19. Zhang, K.; Pei, Z.; Wang, D. Organic solvent pretreatment of lignocellulosic biomass for biofuels and
biochemicals: A review. Bioresour. Technol. 2016, 199, 21–33. [CrossRef] [PubMed]
20. Pan, X.; Arato, C.; Gilkes, N.; Gregg, D.; Mabee, W.; Pye, K.; Xiao, Z.; Zhang, X.; Saddler, J. Biorefining of
softwoods using ethanol organosolv pulping: Preliminary evaluation of process streams for manufacture of
fuel-grade ethanol and co-products. Biotechnol. Bioeng. 2005, 90, 473–481. [CrossRef] [PubMed]
21. Zhao, X.; Cheng, K.; Liu, D. Organosolv pretreatment of lignocellulosic biomass for enzymatic hydrolysis.
Appl. Microbiol. Biotechnol. 2009, 82, 815. [CrossRef] [PubMed]
22. Zakzeski, J.; Bruijnincx, P.C.A.; Jongerius, A.L.; Weckhuysen, B.M. The catalytic valorization of lignin for the
production of renewable chemicals. Chem. Rev. 2010, 110, 3552–3599. [CrossRef] [PubMed]
23. De la Torre, M.; Moral, A.; Hernández, M.; Cabeza, E.; Tijero, A. Organosolv lignin for biofuel. Ind. Crops Prod.
2013, 45, 58–63. [CrossRef]
24. Chang, V.S.; Holtzapple, M.T. Fundamental factors affecting biomass enzymatic reactivity. In Twenty-first
Symposium on Biotechnology for Fuels and Chemicals; Humana Press: Totowa, NJ, USA, 2000; pp. 5–37.
25. Lee, D.; Yu, A.H.C.; Saddler, J.N. Evaluation of cellulase recycling strategies for the hydrolysis of
lignocellulosic substrates. Biotechnol. Bioeng. 1995, 45, 328–336. [CrossRef] [PubMed]
26. Öhgren, K.; Bura, R.; Saddler, J.; Zacchi, G. Effect of hemicellulose and lignin removal on enzymatic
hydrolysis of steam pretreated corn stover. Bioresour. Technol. 2007, 98, 2503–2510. [CrossRef] [PubMed]
276
Energies 2018, 11, 2457
27. Zheng, Y.; Zhang, S.; Miao, S.; Su, Z.; Wang, P. Temperature sensitivity of cellulase adsorption on lignin and
its impact on enzymatic hydrolysis of lignocellulosic biomass. J. Biotechnol. 2013, 166, 135–143. [CrossRef]
[PubMed]
28. Koo, B.-W.; Kim, H.-Y.; Park, N.; Lee, S.-M.; Yeo, H.; Choi, I.-G. Organosolv pretreatment of Liriodendron
tulipifera and simultaneous saccharification and fermentation for bioethanol production. Biomass Bioenergy
2011, 35, 1833–1840. [CrossRef]
29. Barakat, A.; De Vries, H.; Rouau, X. Dry fractionation process as an important step in current and future
lignocellulose biorefineries: A review. Bioresour. Technol. 2013, 134, 362–373. [CrossRef] [PubMed]
30. Mathew, A.K.; Chaney, K.; Crook, M.; Humphries, A.C. Alkaline pre-treatment of oilseed rape straw for
bioethanol production: Evaluation of glucose yield and pre-treatment energy consumption. Bioresour. Technol.
2011, 102, 6547–6553. [CrossRef] [PubMed]
31. Lee, S.-H.; Teramoto, Y.; Endo, T. Enzymatic saccharification of woody biomass micro/nanofibrillated by
continuous extrusion process I—Effect of additives with cellulose affinity. Bioresour. Technol. 2009, 100,
32. Hideno, A.; Inoue, H.; Tsukahara, K.; Fujimoto, S.; Minowa, T.; Inoue, S.; Endo, T.; Sawayama, S. Wet disk
milling pretreatment without sulfuric acid for enzymatic hydrolysis of rice straw. Bioresour. Technol. 2009,
33. Zhu, J.Y.; Pan, X.J. Woody biomass pretreatment for cellulosic ethanol production: Technology and energy
consumption evaluation. Bioresour. Technol. 2010, 101, 4992–5002. [CrossRef] [PubMed]
34. Selig, M.; Weiss, N.; Ji, Y. Enzymatic Saccharification of Lignocellulosic Biomass: Laboratory Analytical Procedure;
National Renewable Energy Laboratory: Golden, CO, USA, 2008.
35. Zhang, L.; Yan, L.; Wang, Z.; Laskar, D.D.; Swita, M.S.; Cort, J.R.; Yang, B. Characterization of lignin
derived from water only and dilute acid flowthrough pretreatment of poplar wood at elevated tempeatures.
Biotechnol. Biofuels 2015, 8, 203–217. [CrossRef] [PubMed]
Carbohydrates and Lignin in Biomass: Laboratory Analytical Procedure; National Renewable Energy Laboratory:
Golden, CO, USA, 2008.
37. Sluiter, A.; Hames, B.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Templeton, D. Determination of Sugars, Byproducts,
and Degradation Products in Liquid Fraction Process Samples: Laboratory Analytical Procedure; National Renewable
Energy Laboratory: Golden, CO, USA, 2012.
38. Sluiter, A.; Hames, B.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Templeton, D. Determination of Ash in Biomass:
Laboratory Analytical Procedure; National Renewable Energy Laboratory: Golden, CO, USA, 2008.
39. El Hage, R.; Chrusciel, L.; Desharnais, L.; Brosse, N. Effect of autohydrolysis of Miscanthus x giganteus
on lignin structure and organosolv delignification. Bioresour. Technol. 2010, 101, 9321–9329. [CrossRef]
[PubMed]
40. Huijgen, W.J.J.; Smit, A.T.; De Wild, P.J.; Den Uil, H. Fractionation of wheat straw by prehydrolysis,
organosolv delignification and enzymatic hydrolysis for production of sugars and lignin. Bioresour. Technol.
2012, 114, 389–398. [CrossRef] [PubMed]
41. Jeffries, T.W. Biodegradation of lignin-carbohydrate complexes. In Physiology of Biodegradative Microorganisms;
Springer: Dordrecht, The Netherlands, 1991; pp. 163–176.
42. Zadeh, E.M.; O’Keefe, S.F.; Kim, Y.-T. Utilization of lignin in biopolymeric packaging films. ACS Omega 2018,
3, 7388–7398. [CrossRef]
277
energies
Article
Biodiesel Production from a Novel Nonedible
Feedstock, Soursop (Annona muricata L.) Seed Oil
Chia-Hung Su *, Hoang Chinh Nguyen, Uyen Khanh Pham, My Linh Nguyen and Horng-Yi Juan
Graduate School of Biochemical Engineering, Ming Chi University of Technology, New Taipei City 24301,
Taiwan; [email protected] (H.C.N.); [email protected] (U.K.P.);
[email protected] (M.L.N.), [email protected] (H.-Y.J.)
Abstract: This study investigated the optimal reaction conditions for biodiesel production from
soursop (Annona muricata) seeds. A high oil yield of 29.6% (w/w) could be obtained from soursop
seeds. Oil extracted from soursop seeds was then converted into biodiesel through two-step
transesterification process. A highest biodiesel yield of 97.02% was achieved under optimal
acid-catalyzed esterification conditions (temperature: 65 ◦ C, 1% H2 SO4 , reaction time: 90 min,
and a methanol:oil molar ratio: 10:1) and optimal alkali-catalyzed transesterification conditions
(temperature: 65 ◦ C, reaction time: 30 min, 0.6% NaOH, and a methanol:oil molar ratio: 8:1).
The properties of soursop biodiesel were determined and most were found to meet the European
standard EN 14214 and American Society for Testing and Materials standard D6751. This study
suggests that soursop seed oil is a promising biodiesel feedstock and that soursop biodiesel is a viable
alternative to petrodiesel.
Keywords: Annona muricata; biodiesel production; seed oil; soursop; two-step process
1. Introduction
Fossil fuel depletion and environmental concerns have stimulated the search for alternative fuels
from renewable sources. Biodiesel, a biomass-derived fuel, is renewable, exhibits superior combustion
properties, and is completely suitable for diesel engines [1,2]. Furthermore, the use of biodiesel results
in relatively low environmental pollution because biodiesel is sulfur free and emits minimal carbon
monoxide and hydrocarbons [3–5]. Because of these merits, biodiesel has been developed worldwide
to replace petrodiesel.
Biodiesel has been mainly produced from edible oil using an acid, alkali, or enzyme catalyst [6,7].
Nevertheless, the use of edible oil as a feedstock increases the production cost of biodiesel [8], thus
limiting the commercialization of biodiesel. Furthermore, the use of edible feedstock for fuel purpose
may cause adverse effects on food supply [9–11]; therefore, alternative feedstocks must be identified
for biodiesel synthesis. Numerous cheap and nonedible feedstocks, including microalgae oil [12–14],
Jatropha oil [15,16], waste cooking oil [17–19], insect fat [20–22], Chinese tallow tree seed oil [23],
tobacco seed oil [24], sweet basil seed oil [25], Brucea javanica seed oil [26], spent coffee grounds [27],
and food waste [28] have been investigated as vital feedstocks for biodiesel synthesis. Two-step
transesterification (acid-catalyzed esterification followed by alkali-catalyzed transesterification) is a
promising method to produce biodiesel from high free fatty acid oils [29,30]. The acid oils (fatty acid
content >1%, w/w) should be esterified using an acid catalyst to lower the oil acidity before applying an
alkali catalyst to transesterify the oil into biodiesel [31–33]. This two-step process not only minimizes
soap formation but also enhances biodiesel yield [31,34,35].
Soursop (Annona muricata L.), which belongs to the Annonaceae family, is an economically critical
crop worldwide [36,37]. A. muricata is native to North and South America and is popularly distributed

Energies 2018, 11, 2562
in the tropical and subtropical areas of Western Africa, Central America, the Caribbean, and the
Asian continent [37–39]. The A. muricata tree is approximately 5–8-m tall with low branches [39],
and the trees yield up to 10 tons of fruit per hectare [38]. The oval or heart-shaped fruit is 15–30-cm
in length, 10–20-cm in width, and, on average, weighs up to 4.0 kg [36,39]. The edible white pulp
of soursop fruit comprises about 80% water, 18% carbohydrate, 1% protein, and 1% fiber content
and contains beneficial vitamins [40]. The mesocarp contains numerous black seeds, which are each
approximately 2 cm long and 1 cm wide [36]. The soursop is mainly cultivated for its fruit, which
is used in fresh and processed forms in the production of juice, ice cream, sherbet, beverages, and
candy [37,41,42]. The use of soursop fruit in food production results in various waste materials,
including seeds which account for 5–8.5% of the fruit [37,38,43]. The seeds are usually discarded
and cannot be used as animal feed because they contain toxic substances such as annonacin and
acetogenins [37,43,44]. Studies have shown that oil comprises up to 40% of the soursop seed [37].
The seed oil mainly comprises palmitic acid, oleic acid, and linoleic acid [37,45]. This composition is
similar to that of other biodiesel feedstocks [24,29,31]. Therefore, the nonedible soursop seed oil is
a promising and cheap biodiesel feedstock. In addition, the use of this seed for biodiesel synthesis
can resolve the problematic surplus of the seed in the food industry. Only few studies have reported
the potential use of this seed oil for biodiesel production [46,47]. However, no optimization study on
reaction conditions has investigated for biodiesel production from A. muricata seed oil.
This study optimized the reaction factors for producing biodiesel from A. muricata seed
oil. Because of the presence of a high level of free fatty acid (FFA) in the seed oil, a two-step
process was used to convert FFAs and triglycerides into biodiesel. The effects of reaction factors
(molar ratio of methanol to oil, temperature, catalyst amount, and reaction time) on esterification
and transesterification were investigated to optimize the reaction conditions. The biodiesel’s
properties were finally characterized according to the American Society for Testing and Materials
(ASTM) methods.
2.1. Materials
Ripe soursop fruits were purchased from Thu Duc Agromarket (Ho Chi Minh City, Vietnam).
The seeds were removed from the fruit and air-dried at room temperature for 3 days. The soursop seed
kernels were then separated from the hulls, ground with a blender (EUPA TSK-935BAP, Tsann Kuen
Enterprise Co., Ltd., Taipei, Taiwan), and stored at room temperature. Methanol, sulfuric acid, sodium
hydroxide, n-hexane, and other reagents used in this study were of analytical grade (≥99.0% purity)
and were obtained from Tedia Company, Inc. (Fairfield, CT, USA).
2.2. Extraction of Crude Oil

Soursop seed powder was immersed in n-hexane (1:4, w/v) at room temperature for 2 days, and
stirred to extract the oil from soursop seed. After extraction, the hexane layer was separated from
solid residue by filtration (Advantec No. 5C filter paper). The n-hexane was then removed using a
R300 Buchi Rotary Evaporator (Büchi Labortechnik, Flawil, Switzerland), and the soursop seed crude
oil was obtained. The crude oil’s properties, such as acidity, saponification, and iodine values, were
measured using the standard method [48,49].
2.3. Production of Biodiesel through Two-Step Process
2.3.1. Esterification Step

An H2 SO4 -catalyzed pretreatment was employed to reduce the oil’s acidity and convert its FFAs
into biodiesel. To study the influence of reaction factors on esterification, several experimental trials
were conducted in a sealed reactor with stirring under different conditions: molar ratios of methanol to
279
Energies 2018, 11, 2562
oil (4:1–12:1), temperatures (45–85 ◦ C), catalyst amounts (0.25–2.0%), and reaction times (30–150 min).
After each reaction, the samples were withdrawn to evaluate the FFA conversion.
2.3.2. Transesterification Step

The oil pretreated through H2 SO4 -catalyzed esterification was used for the transesterification
step. The esterified reaction mixture was kept in a funnel for phase separation. After two phases were
completely separated, the crude oil and biodiesel (upper layer) was poured into a sealed reactor and
subsequently transesterified into biodiesel using NaOH as catalyst. A set of experiments with various
methanol to oil molar ratios (4:1–12:1), temperatures (45–85 ◦ C), catalyst amounts (0.4–1.2%), and
reaction times (15–75 min) were studied for their effects on the conversion yield. After each reaction,
the reactor was placed at room temperature for phase separation. The mixture’s upper layer containing
biodiesel was collected to determine the biodiesel yield.
2.4. Analysis
The acid value (AV) of the oil sample was measured using a titration method reported
previously [50,51]. The FFA conversion was then calculated as follows:
AV1 − AV2
FFA conversion (%) = × 100 (1)
AV1
where AV1 is the initial acid value, and AV2 is the acid value after the esterification reaction.
The composition of the biodiesel was quantified using a Shimadzu GC-2014 gas chromatograph
system (Shimadzu Corp., Kyoto, Japan) equipped with a Stabilwax capillary column (Restek Corp.,
Bellefonte, PA, USA) and a flame ionization detector (Shimadzu Corp., Kyoto, Japan) according to the
procedure reported in our previous study [10]. The fatty acid profiles of the soursop biodiesel were
characterized based on Supelco 37 Component FAME Mix reference standards (Sigma-Aldrich Corp.,
St. Louis, MO, USA). The biodiesel content was quantified by comparing the peak areas of fatty acid
methyl esters with those of the internal standard, methyl pentadecanoate. The soursop biodiesel yield
was then calculated as follows [10]:
Asample W W
Biodiesel yield (%) = × standard × biodiesel × 100 (2)
Astandard Wsample Woil
where Asample is peak area of biodiesel sample, Astandard is peak area of international standard, W sample
is weigh of biodiesel sample, W standard is weight of internal standard, W biodiesel is weight of total
biodiesel product, and W oil is weight of oil used.
The acid value, viscosity, sulfur content, water content, ester content, cetane index, density, and
flash point of the produced biodiesel were determined using the ASTM D664, D445, D5453, D95,
D7371, D613, D1480, and D93 methods, respectively [52].
3.1. Properties of Soursop Seed Oil

Table 1 shows the characteristics of soursop seed oil. The oil extracted from soursop seeds reached
the yield of 29.6%, demonstrating the soursop seed’s high oil content and subsequent potential as
an oil source. The saponification value of soursop seed oil was 244.7 mg KOH/g, indicating that the
average molecular weight of soursop seed oil was 884.4 g/mol. The acid value of the soursop seed oil
was 54.4 mg KOH/g, indicating a high FFA content. Biodiesel production processes must be refined
to maximize the value of materials and minimize costs [25,53,54]. To maximize the biodiesel yield
from oils with high FFA levels, esterification must be performed to reduce the level of FFAs prior to
280
Energies 2018, 11, 2562
transesterification [31,34,35]. Therefore, the two-step process of acid-catalyzed esterification followed

by alkali-transesterification was selected for biodiesel synthesis from soursop seed oil in this study.
Table 1. Properties of crude soursop seed oil.
Fat Yield (%) Acid Value (mg KOH/g) Saponification Value (mg KOH/g)
29.6 ± 0.2 54.4 ± 0.4 244.7 ± 1.6
3.2. Conversion of FFAs into Biodiesel through Acid-Catalyzed Esterification
3.2.1. Effect of Methanol to Oil Molar Ratio

Esterification pretreatment enhances the biodiesel yield by reducing the oil’s acidity and
converting its FFAs into biodiesel [29,51]. To optimize reaction conditions, this study investigates the
influences of methanol to oil molar ratio, temperature, H2 SO4 amount, and reaction time on the FFAs
conversion. First, esterification was performed at 75 ◦ C with 1% H2 SO4 (w/w) and various methanol
to oil molar ratios (4:1–12:1) for 60 min. As shown in Figure 1a, FFA conversion was greater at higher
methanol to oil molar ratios. This result corresponds with those of other studies [35,55]. The molar
ratio of methanol to oil is a critical factor affecting the efficiency of esterification reactions. A high
methanol to oil molar ratio is required to drive esterification reactions toward completion [55]. In this
work, the highest FFA conversion occurred at the methanol:oil molar ratio of 12:1. However, the
FFA conversion had insignificant differences between the methanol:oil molar ratios of 12:1 and 10:1.
Therefore, the methanol:oil molar ratio of 10:1 was chosen for the next experiments.
(a) (b)
(c) (d)
Figure 1. Effects of (a) molar ratio of methanol to oil (with a fixed temperature of 75 ◦ C, 1% H2 SO4 , and
a reaction time of 60 min); (b) temperature (with a fixed methanol:oil molar ratio of 10:1, 1% H2 SO4 ,
and a reaction time of 60 min); (c) catalyst amount (with a fixed methanol:oil molar ratio of 10:1, a
temperature of 65 ◦ C, and a reaction time of 60 min); and (d) reaction time (with a fixed methanol:oil
molar ratio of 10:1, a temperature of 65 ◦ C, and 1% H2 SO4 ) on FFA conversion in soursop seed oil.
3.2.2. Effect of Temperature

To investigate the impact of temperature on the efficiency of esterification, the reaction was
performed at various temperatures (45–85 ◦ C), whilst keeping the other factors constant. As shown in
281
Energies 2018, 11, 2562
Figure 1b, FFA conversion was enhanced from 84.86 to 96.16% when the temperature was increased
from 45 to 85 ◦ C. This result is reasonable because a high temperature enhances the reaction rate [51,56].
However, no significant differences were found in the proportion of FFA conversion at temperatures
of 65, 75, and 85 ◦ C. Therefore, to reduce energy consumption, 65 ◦ C was selected as the optimal
temperature for the esterification reaction.
3.2.3. Effect of Catalyst Amount

The results in Figure 1c reflect the effect of sulfuric acid amount on FFA conversion.
FFA conversion significantly improved when the catalyst levels increased from 0.25 to 1.0%.
Nevertheless, increasing the catalyst load to 1.5% resulted in only a slight increase in FFA conversion,
and a 2% catalyst load caused a slight decrease in conversion efficiency. This slight decrease at a 2%
catalyst load is similar to of the results of other studies [35,55]. Excess H2 SO4 catalyst can activate the
polymerization of unsaturated FFA, causing the product’s darkened color due to the oxidation and
decarboxylation of FFA [55]. Therefore, 1.0% H2 SO4 was chosen as the optimal catalyst amount for
further experiments.
3.2.4. Effect of Reaction Time

Various reaction times (30–150 min) were tested for esterification performed at 65 ◦ C with a
methanol:oil molar ratio of 10:1 and 1% H2 SO4 (w/w). As shown in Figure 1d, FFA conversion increased
from 85.71 to 96.61% when increasing reaction time from 30 to 90 min. Increases in reaction times
beyond 90 min resulted in insignificant increases in FFA conversion, indicating that reactions reached
equilibrium at 90 min. In conclusion, the optimal conditions for the H2 SO4 -catalyzed esterification
were determined to be a methanol:oil molar ratio: 10:1, a temperature: 65 ◦ C, 1% H2 SO4 (w/w), and
90 min. These conditions were thus used in this study for the esterification step in biodiesel production.
3.3. Conversion of Triglyceride into Biodiesel through Alkali-Catalyzed Transesterification
3.3.1. Effect of Methanol to Oil Molar Ratio

The esterified oil was used as the material for producing biodiesel through alkali-catalyzed
transesterification. To optimize transesterification conditions, the influences of methanol to oil
molar ratio, temperature, NaOH amount, and reaction time on the biodiesel yield were examined.
First, transesterification was performed at 65 ◦ C with 0.8% NaOH (w/w) and various methanol to oil
molar ratios (4:1–12:1) for 30 min. As can be seen from Figure 2a, the biodiesel yield increased from 81%
to 96.37% when the molar ratio of methanol to oil was increased from 4:1 to 8:1. Nevertheless, a higher
methanol to oil molar ratio caused a decrease in the biodiesel yield. This result is consistent with that
reported in the study of biodiesel synthesis from Jatropha curcus seed oil [31] and Croton megalocarpus
oil [57]. A high level of methanol may have increased the glycerol solubility in the solution, driving
the equilibrium to a reverse reaction and thus lowering the biodiesel yield [57]. Therefore, this study
selected the methanol:oil molar ratio of 8:1 as the optimal reactant ratio for transesterification.
282
Energies 2018, 11, 2562
(a) (b)
(c) (d)
Figure 2. Effects of (a) molar ratio of methanol to oil (with a fixed temperature of 65 ◦ C, 0.8% NaOH,
and a reaction time of 30 min); (b) temperature (with a fixed methanol:oil molar ratio of 8:1, 0.8%
NaOH, and a reaction time of 30 min); (c) catalyst amount (with a fixed methanol:oil molar ratio of 8:1,
a temperature of 65 ◦ C, and a reaction time of 30 min); and (d) reaction time (with a fixed methanol:oil
molar ratio of 8:1, a temperature of 65 ◦ C, and 0.6% NaOH) on transesterification of soursop seed oil.
3.3.2. Effect of Temperature

To investigate the influence of temperature on the biodiesel yield, transesterification was carried
out at various temperatures (45–85 ◦ C) with a methanol to oil molar ratio, NaOH amount, and reaction
time maintained at 8:1, 0.8%, and 30 min, respectively. As can be seen from Figure 2b, the biodiesel yield
increased when temperature was increased from 45 to 65 ◦ C. Nevertheless, the biodiesel yield reduced
at temperatures greater than 65 ◦ C. This result is in agreement with those of other studies [57,58].
A high temperature may have enhanced side reactions, including saponification, thus resulting in
a lower biodiesel yield [57]. Based on this result, 65 ◦ C was chosen as the optimal temperature for
the transesterification.
3.3.3. Effect of Catalyst Amount

Catalyst amount is a critical factor affecting the efficiency of transesterification. In this study,
various NaOH amounts were tested with other factors maintained as constant to evaluate the influence
of NaOH amount on the biodiesel yield. Results revealed that the biodiesel yield increased when the
amount of catalyst increased from 0.4 to 0.6% (Figure 2c). However, when the catalyst amount is higher
than 0.6%, the biodiesel yield reduced. Excess catalyst favored the saponification reaction, enhancing
the formation of an emulsion and gel, thus lowering the biodiesel yield [31,55,57]. Therefore, 0.6%
NaOH was selected for use in further experiments.
283
Energies 2018, 11, 2562
3.3.4. Effect of Reaction Time

Finally, the influence of reaction time on the biodiesel yield was examined using the optimal
reaction conditions obtained in previous experiments. Transesterification was performed at 65 ◦ C
with a methanol:oil molar ratio of 8:1, 0.6% NaOH (w/w), and various reaction times (15–75 min).
As indicated in Figure 2d, the biodiesel yield increased from 77.85 to 97.02% when the reaction time
increased from 15 to 30 min. A reaction time longer than 30 min caused a slight decrease in the biodiesel
yield. Therefore, a reaction time of 30 min was determined as sufficient for the transesterification
reaction. In conclusion, the highest biodiesel yield of 97.02% was achieved under the following optimal
transesterification conditions: 65◦ C, 0.6% NaOH (w/w), 30 min, and methanol:oil molar ratio of 8:1.
3.4. Fatty Acid Profiles of Soursop Biodiesel

Table 2 illustrates the fatty acid profiles of soursop biodiesel in comparison with rapeseed biodiesel.
Nine fatty acid methyl esters were identified in the soursop biodiesel, among which oleic acid methyl
ester (43.68%), linoleic acid methyl ester (32.45%), and palmitic acid methyl ester (18.14%) were present
in the highest amounts. The synthesized biodiesel was found to comprise 78.07% unsaturated fatty
acids and 21.93% saturated ones. The saturated fatty acid level in soursop biodiesel was higher than
that of rapeseed biodiesel (4.3%) [51], indicating a higher cetane index for the soursop biodiesel. This is
because the high level of saturated fatty acid increases the cetane index of a fuel [59]. In addition,
because saturated fatty acid methyl esters exhibit higher oxidative stability than unsaturated ones [4,60],
the soursop biodiesel can have more oxidative stability than rapeseed biodiesel. These results suggest
that soursop seed oil is a suitable feedstock for biodiesel synthesis.
Table 2. Fatty acid methyl ester compositions of soursop biodiesel compared with rapeseed biodiesel.
Composition Rapeseed Biodiesel a (%) Soursop Biodiesel b (%)

Palmitic acid methyl ester (C16:0) 3.5 18.14
Palmitoleic acid methyl ester (C16:1) na c 0.81
Stearic acid methyl ester (18:0) 0.8 3.79
Oleic acid methyl ester (C18:1) 64.4 43.68
Linoleic acid methyl ester (C18:2) 22.3 32.45
Linolenic acid methyl ester (C18:3) 8.2 1.13
a Data obtained from Reference [51]; b This study; c na = none reported.
3.5. Properties of Soursop Biodiesel

The soursop biodiesel’s properties were characterized using ASTM standard methods and were
compared with the corresponding properties of rapeseed biodiesel [51]. As indicated in Table 3, most
soursop biodiesel’s properties were similar to those of rapeseed biodiesel. Remarkably, most properties
of synthesized biodiesel, sulfur content (0.04%), ester content (98.6%), viscosity (5.5 mm2 /s), water
content (300 mg/kg), cetane number (53), density (868 kg/m3 ), and flash point (123◦ C), met the
standards ASTM D6751 [52] and EN 14,214 [61]. These results indicate that the synthesized biodiesel
may serve as an alternative to petrodiesel. Moreover, the high ester content in the biodiesel indicates
that the conditions identified in this study are optimal for the esterification and transesterification
reactions. However, the acid value of synthesized biodiesel was 0.8, which was higher than the
standards EN 14,214 and ASTM D6751: this could be due to the presence of free fatty acid in the
biodiesel product. A further purification step is therefore required in order to reduce the acid value [62].
284
Energies 2018, 11, 2562
Table 3. Soursop biodiesel’s properties compared with those of rapeseed biodiesel, the standards
ASTM D6751, and EN 14214.
ASTM ASTM Rapeseed

Properties EN 14214 This Study
Method D6751 a Biodiesel b
Acid value (mg KOH/g) D664 <0.5 <0.5 0.31 <0.8
Sulfur content (wt. %) D5453 <0.05 <0.05 <0.01 0.04
Ester content (%) D7371 na c >96.5 na c 98.6
Viscosity at 40 ◦ C (mm2 /s) D445 1.9–6.0 3.5–5.0 6.35 5.5
Water content (mg/kg) D95 na c <500 300 300
Cetane number D613 >47 >51 45 53
Density (kg/m3 ) D1480 na c 860–900 880 868
Flash point (closed cup) (◦ C) D93 100–170 >120 na c 123
a Data obtained from Reference [52]; b Data were obtained from Reference [51]; c na = none reported.
3.6. The Feasibility of Soursop Seed Oil as Biodiesel Feedstock

With an increasing demand for renewable energy, biodiesel has been widely produced to replace
petrodiesel. To reduce the production cost, various non-edible feedstocks including microbial
oil [12–14], waste cooking oil [17–19], insect fat [20–22], and plant seed oil [23–25], have been
studied for biodiesel production because of their low-price. However, the availability of those
oils is still a major concern for large-scale production [10]. Therefore, efforts have been made to
search for new low-cost biodiesel feedstocks. In recent years, the use of plant seed obtained from
fruit production industry for biodiesel production has attracted much attention due to its low cost
and availability [25,26]. In the state of Bahia (Brazil), approximately 20 thousand tons of soursop
fruit are produced annually, thus producing about 1.7 thousand tons of soursop seed each year [37].
Since soursop is an economically important crop worldwide [37,41], the global food production from
soursop fruit can cause a problematic surplus of this seed. To add value to the soursop seeds, the
effective method of recycling these seeds is being investigated. In the present study, soursop seeds
were used as a non-edible feedstock for biodiesel production—a solution to the soursop seed disposal
problem. The results suggested that soursop seed can be a potential biodiesel feedstock along with
other non-edible plant seed oils.
4. Conclusions
This paper investigates the use of soursop seed oil for biodiesel production. In the current study,
the reaction conditions of acid-catalyzed esterification and alkali-catalyzed transesterification were
optimized to maximize biodiesel yield. Under optimized conditions, 97.02% biodiesel yield was
obtained. The properties of the soursop biodiesel were determined and were found to meet the ASTM
D6751 and EN 14214. The results of this study suggest that soursop seed oil is a potential biodiesel
feedstock and the soursop biodiesel can serve as an alternative for petrodiesel.
Author Contributions: Conceptualization, C.-H.S. and H.C.N.; methodology, C.-H.S. and H.C.N.; validation,
H.-J.Y. and U.K.P.; formal analysis, U.K.P and M.L.N.; investigation, U.K.P. and M.L.N; resources, C.-H.S.;
writing—original draft preparation, C.-H.S. and H.C.N.; writing—review and editing, C.-H.S.; supervision,
C.-H.S.; project administration, C.-H.S.; funding acquisition, C.-H.S.
Funding: This research was funded by Ministry of Science and Technology (MOST) of Taiwan, grant
number 106-2221-E-131-028.
References
1. Mosarof, M.; Kalam, M.; Masjuki, H.; Alabdulkarem, A.; Ashraful, A.; Arslan, A.; Rashedul, H.; Monirul, I.
Optimization of performance, emission, friction and wear characteristics of palm and Calophyllum inophyllum
biodiesel blends. Energy Convers. Manag. 2016, 118, 119–134. [CrossRef]
285
Energies 2018, 11, 2562
2. Damanik, N.; Ong, H.C.; Tong, C.W.; Mahlia, T.M.I.; Silitonga, A.S. A review on the engine performance and
exhaust emission characteristics of diesel engines fueled with biodiesel blends. Environ. Sci. Pollut. Res. 2018,
25, 15307–15325. [CrossRef] [PubMed]
[CrossRef]
4. Knothe, G. “Designer” biodiesel: Optimizing fatty ester composition to improve fuel properties. Energy Fuels
2008, 22, 1358–1364. [CrossRef]
5. Abdul Malik, M.S.; Shaiful, A.I.M.; Mohd Ismail, M.S.; Mohd Jaafar, M.N.; Mohamad Sahar, A. Combustion
and emission characteristics of coconut-based biodiesel in a liquid fuel burner. Energies 2017, 10, 458.
[CrossRef]
6. Leung, D.Y.; Wu, X.; Leung, M. A review on biodiesel production using catalyzed transesterification.
7. Bhuyan, M.S.U.S.; Alam, A.H.M.A.; Chu, Y.; Seo, Y.C. Biodiesel production potential from littered edible
oil fraction using directly synthesized S-TiO2 /MCM-41 catalyst in esterification process via non-catalytic
subcritical hydrolysis. Energies 2017, 10, 1290. [CrossRef]
8. Mardhiah, H.H.; Ong, H.C.; Masjuki, H.; Lim, S.; Lee, H. A review on latest developments and future
prospects of heterogeneous catalyst in biodiesel production from non-edible oils. Renew. Sustain. Energy Rev.
2017, 67, 1225–1236. [CrossRef]
9. Martindale, W.; Trewavas, A. Fuelling the 9 billion. Nat. Biotechnol. 2008, 26, 1068–1070. [CrossRef] [PubMed]
10. Nguyen, H.C.; Liang, S.H.; Doan, T.T.; Su, C.H.; Yang, P.C. Lipase-catalyzed synthesis of biodiesel from black
soldier fly (Hermetica illucens): Optimization by using response surface methodology. Energy Convers. Manag.
2017, 145, 335–342. [CrossRef]
11. Tuntiwiwattanapun, N.; Monono, E.; Wiesenborn, D.; Tongcumpou, C. In-situ transesterification process
for biodiesel production using spent coffee grounds from the instant coffee industry. Ind. Crops Prod. 2017,
102, 23–31. [CrossRef]
12. El Shimi, H.I.; Moustafa, S.S. Biodiesel production from microalgae grown on domestic wastewater:
Feasibility and Egyptian case study. Renew. Sustain. Energy Rev. 2018, 82, 4238–4244. [CrossRef]
13. Chia, S.R.; Ong, H.C.; Chew, K.W.; Show, P.L.; Phang, S.M.; Ling, T.C.; Nagarajan, D.; Lee, D.J.; Chang, J.S.
Sustainable approaches for algae utilisation in bioenergy production. Renew. Energy 2018, 129, 838–852.
[CrossRef]
14. Ghorbani, A.; Rahimpour, M.R.; Ghasemi, Y.; Raeissi, S. The biodiesel of microalgae as a solution for diesel
demand in Iran. Energies 2018, 11, 950. [CrossRef]
15. Kamel, D.A.; Farag, H.A.; Amin, N.K.; Zatout, A.A.; Ali, R.M. Smart utilization of jatropha (Jatropha curcas
Linnaeus) seeds for biodiesel production: Optimization and mechanism. Ind. Crops Prod. 2018, 111, 407–413.
[CrossRef]
16. Lin, J.J.; Chen, Y.W. Production of biodiesel by transesterification of Jatropha oil with microwave heating.
J. Taiwan Inst. Chem. Eng. 2017, 75, 43–50. [CrossRef]
17. Milano, J.; Ong, H.C.; Masjuki, H.H.; Silitonga, A.S.; Chen, W.H.; Kusumo, F.; Dharma, S.; Sebayang, A.H.
Optimization of biodiesel production by microwave irradiation-assisted transesterification for waste cooking
oil-Calophyllum inophyllum oil via response surface methodology. Energy Convers. Manag. 2018, 158, 400–415.
[CrossRef]
18. Hossain, M.N.; Siddik Bhuyan, M.S.U.; Alam, A.H.M.A.; Seo, Y.C. Biodiesel from hydrolyzed waste cooking
oil using a S-ZrO2 /SBA-15 super acid catalyst under sub-critical conditions. Energies 2018, 11, 299. [CrossRef]
19. Poudel, J.; Karki, S.; Sanjel, N.; Shah, M.; Oh, S.C. Comparison of biodiesel obtained from virgin cooking oil
and waste cooking oil using supercritical and catalytic transesterification. Energies 2017, 10, 546. [CrossRef]
20. Nguyen, H.C.; Liang, S.H.; Chen, S.S.; Su, C.H.; Lin, J.H.; Chien, C.C. Enzymatic production of biodiesel from
insect fat using methyl acetate as an acyl acceptor: Optimization by using response surface methodology.
21. Zheng, L.; Hou, Y.; Li, W.; Yang, S.; Li, Q.; Yu, Z. Exploring the potential of grease from yellow mealworm
beetle (Tenebrio molitor) as a novel biodiesel feedstock. Appl. Energy 2013, 101, 618–621. [CrossRef]
22. Surendra, K.; Olivier, R.; Tomberlin, J.K.; Jha, R.; Khanal, S.K. Bioconversion of organic wastes into biodiesel
and animal feed via insect farming. Renew. Energy 2016, 98, 197–202. [CrossRef]
286
Energies 2018, 11, 2562
23. Barekati-Goudarzi, M.; Muley, P.D.; Clarens, A.; Nde, D.B.; Boldor, D. Continuous microwave-assisted
in-situ transesterification of lipids in seeds of invasive Chinese tallow trees (Triadica sebifera L.): Kinetic and
thermodynamic studies. Biomass Bioenergy 2017, 107, 353–360. [CrossRef]
24. García-Martínez, N.; Andreo-Martínez, P.; Quesada-Medina, J.; de los Ríos, A.P.; Chica, A.; Beneito-Ruiz, R.;
Carratalá-Abril, J. Optimization of non-catalytic transesterification of tobacco (Nicotiana tabacum) seed oil
using supercritical methanol to biodiesel production. Energy Convers. Manag. 2017, 131, 99–108. [CrossRef]
25. Amini, Z.; Ong, H.C.; Harrison, M.D.; Kusumo, F.; Mazaheri, H.; Ilham, Z. Biodiesel production by
lipase-catalyzed transesterification of Ocimum basilicum L. (sweet basil) seed oil. Energy Convers. Manag.
2017, 132, 82–90. [CrossRef]
26. Hasni, K.; Ilham, Z.; Dharma, S.; Varman, M. Optimization of biodiesel production from Brucea javanica
seeds oil as novel non-edible feedstock using response surface methodology. Energy Convers. Manag. 2017,
149, 392–400. [CrossRef]
27. Tuntiwiwattanapun, N.; Tongcumpou, C. Sequential extraction and reactive extraction processing of spent
coffee grounds: An alternative approach for pretreatment of biodiesel feedstocks and biodiesel production.
Ind. Crops Prod. 2018, 117, 359–365. [CrossRef]
28. Sakuragi, K.; Li, P.; Otaka, M.; Makino, H. Recovery of bio-oil from industrial food waste by liquefied
dimethyl ether for biodiesel production. Energies 2016, 9, 106. [CrossRef]
29. Chen, L.; Liu, T.; Zhang, W.; Chen, X.; Wang, J. Biodiesel production from algae oil high in free fatty acids by
two-step catalytic conversion. Bioresour. Technol. 2012, 111, 208–214. [CrossRef] [PubMed]
30. Silva, L.N.; Cardoso, C.C.; Pasa, V.M. Production of cold-flow quality biodiesel from high-acidity on-edible
oils—Esterification and transesterification of Macauba (Acrocomia aculeata) oil using various alcohols.
BioEnergy Res. 2016, 9, 864–873. [CrossRef]
31. Berchmans, H.J.; Hirata, S. Biodiesel production from crude Jatropha curcas L. seed oil with a high content of
free fatty acids. Bioresour. Technol. 2008, 99, 1716–1721. [CrossRef] [PubMed]
32. Suresh, R.; Antony, J.V.; Vengalil, R.; Kochimoolayil, G.E.; Joseph, R. Esterification of free fatty acids in
non-edible oils using partially sulfonated polystyrene for biodiesel feedstock. Ind. Crops Prod. 2017, 95, 66–74.
[CrossRef]
33. Nguyen, H.C.; Huong, D.T.M.; Juan, H.Y.; Su, C.H.; Chien, C.C. Liquid lipase-catalyzed esterification of oleic
acid with methanol for biodiesel production in the presence of superabsorbent polymer: Optimization by
using response surface methodology. Energies 2018, 11, 1085. [CrossRef]
34. Çaylı, G.; Küsefoğlu, S. Increased yields in biodiesel production from used cooking oils by a two step process:
Comparison with one step process by using TGA. Fuel Process. Technol. 2008, 89, 118–122. [CrossRef]
35. Hayyan, A.; Alam, M.Z.; Mirghani, M.E.; Kabbashi, N.A.; Hakimi, N.I.N.M.; Siran, Y.M.; Tahiruddin, S.
Reduction of high content of free fatty acid in sludge palm oil via acid catalyst for biodiesel production.
36. Coria-Téllez, A.V.; Montalvo-Gónzalez, E.; Yahia, E.M.; Obledo-Vázquez, E.N. Annona muricata: A
comprehensive review on its traditional medicinal uses, phytochemicals, pharmacological activities,
mechanisms of action and toxicity. Arab. J. Chem. 2016, 11, 662–691. [CrossRef]
37. Schroeder, P.; do Nascimento, B.P.; Romeiro, G.A.; Figueiredo, M.K.K.; da Cunha Veloso, M.C. Chemical and
physical analysis of the liquid fractions from soursop seed cake obtained using slow pyrolysis conditions.
J. Anal. Appl. Pyrolysis 2017, 124, 161–174. [CrossRef]
38. Fasakin, A.; Fehintola, E.; Obijole, O.; Oseni, O. Compositional analyses of the seed of sour sop,
Annona muricata L., as a potential animal feed supplement. Sci. Res. Essays 2008, 3, 521–523.
39. Moghadamtousi, S.Z.; Fadaeinasab, M.; Nikzad, S.; Mohan, G.; Ali, H.M.; Kadir, H.A. Annona muricata
(Annonaceae): A review of its traditional uses, isolated acetogenins and biological activities. Int. J. Mol. Sci.
2015, 16, 15625–15658. [CrossRef] [PubMed]
40. de Lima, M.C.; Alves, R. Soursop (Annona muricata L.). In Postharvest Biology and Technology of Tropical
and Subtropical Fruits: Mangosteen to White Sapote; Yahia, E., Ed.; Woodhead Publishing: Cambridge, UK,
2011; pp. 363–392.
41. Gajalakshmi, S.; Vijayalakshmi, S.; Devi Rajeswari, V. Phytochemical and pharmacological properties of
Annona muricata: A review. Int. J. Pharm. Pharm. Sci. 2012, 4, 3–6.
42. Awan, J.; Kar, A.; Udoudoh, P. Preliminary studies on the seeds of Annona muricata Linn. Plant Foods
Hum. Nutr. 1980, 30, 163–168. [CrossRef]
287
Energies 2018, 11, 2562
43. Badrie, N.; Schauss, A.G. Soursop (Annona muricata L.): Composition, nutritional value, medicinal uses,
and toxicology. In Bioactive Foods in Promoting Health; Watson, R.R., Preedy, V.R., Eds.; Academic Press:
Oxford, UK, 2010; pp. 621–643.
44. Zafra-Polo, M.C.; González, M.C.; Estornell, E.; Sahpaz, S.; Cortes, D. Acetogenins from Annonaceae,
inhibitors of mitochondrial complex I. Phytochemistry 1996, 42, 253–271. [CrossRef]
45. Elagbar, Z.A.; Naik, R.R.; Shakya, A.K.; Bardaweel, S.K. Fatty acids analysis, antioxidant and biological
activity of fixed oil of Annona muricata L. seeds. J. Chem. 2016, 6. [CrossRef]
46. Phoo, Z.W.M.M.; Ilham, Z.; Goembira, F.; Razon, L.; Saka, S. Physico-chemical properties of biodiesel
from various feedstocks. In Zero-Carbon Energy Kyoto 2012; Yao, T., Ed.; Springer: Berlin, Germany,
2013; pp. 113–121.
47. Schroeder, P.; dos Santos Barreto, M.; Romeiro, G.A.; Figueiredo, M.K.K. Development of energetic
alternatives to use of waste of Annona muricata L. Waste Biomass Valorization 2018, 9, 1459–1467. [CrossRef]
48. Su, C.H.; Fu, C.C.; Gomes, J.; Chu, I.; Wu, W.T. A heterogeneous acid-catalyzed process for biodiesel
production from enzyme hydrolyzed fatty acids. AIChE J. 2008, 54, 327–336. [CrossRef]
49. Vicente, G.; Martınez, M.; Aracil, J. Integrated biodiesel production: A comparison of different homogeneous
catalysts systems. Bioresour. Technol. 2004, 92, 297–305. [CrossRef] [PubMed]
50. Su, C.H. Recoverable and reusable hydrochloric acid used as a homogeneous catalyst for biodiesel production.
51. Li, Q.; Zheng, L.; Cai, H.; Garza, E.; Yu, Z.; Zhou, S. From organic waste to biodiesel: Black soldier fly,
Hermetia illucens, makes it feasible. Fuel 2011, 90, 1545–1548. [CrossRef]
52. Annual Book of ASTM Standards. Available online: https://www.astm.org/BOOKSTORE/BOS/index.html
(accessed on 26 September 2018).
53. Ramadhas, A.; Jayaraj, S.; Muraleedharan, C. Characterization and effect of using rubber seed oil as fuel in
the compression ignition engines. Renew. Energy 2005, 30, 795–803. [CrossRef]
54. Emil, A.; Yaakob, Z.; Kumar, M.S.; Jahim, J.M.; Salimon, J. Comparative evaluation of physicochemical
properties of Jatropha seed oil from Malaysia, Indonesia and Thailand. J. Am. Oil Chem. Soc. 2010, 87, 689–695.
[CrossRef]
55. Ramadhas, A.S.; Jayaraj, S.; Muraleedharan, C. Biodiesel production from high FFA rubber seed oil. Fuel
2005, 84, 335–340. [CrossRef]
56. Nguyen, H.C.; Liang, S.H.; Li, S.Y.; Su, C.H.; Chien, C.C.; Chen, Y.J.; Huong, D.T.M. Direct transesterification
of black soldier fly larvae (Hermetia illucens) for biodiesel production. J. Taiwan Inst. Chem. Eng. 2018,
85, 165–169. [CrossRef]
57. Kafuku, G.; Mbarawa, M. Biodiesel production from Croton megalocarpus oil and its process optimization.
Fuel 2010, 89, 2556–2560. [CrossRef]
58. Meng, X.; Chen, G.; Wang, Y. Biodiesel production from waste cooking oil via alkali catalyst and its engine
test. Fuel Process. Technol. 2008, 89, 851–857. [CrossRef]
59. Yang, S.; Li, Q.; Gao, Y.; Zheng, L.; Liu, Z. Biodiesel production from swine manure via housefly larvae
(Musca domestica L.). Renew. Energy 2014, 66, 222–227. [CrossRef]
60. Ramos, M.J.; Fernández, C.M.; Casas, A.; Rodríguez, L.; Pérez, Á. Influence of fatty acid composition of raw
materials on biodiesel properties. Bioresour. Technol. 2009, 100, 261–268. [CrossRef] [PubMed]
61. Automotive Fuels-Fatty Acid Methyl Esters (FAME) for Diesel Engines-Requirements and Test
Methods. Available online: http://agrifuelsqcs-i.com/attachments/1598/en14214.pdf (accessed on
25 September 2018).
62. Bala, V.S.S.; Thiruvengadaravi, K.V.; Kumar, P.S.; Premkumar, M.P.; Kumar, M.H.; Sivanesan, S. Removal of
free fatty acids in Pongamia pinnata (Karanja) oil using divinylbenzene-styrene copolymer resins for biodiesel
production. Biomass Bioenergy 2012, 37, 335–341. [CrossRef]
288
energies
Article
Optimisation of Second-Generation Biodiesel
Production from Australian Native Stone Fruit Oil
Using Response Surface Method
Mohammad Anwar 1 , Mohammad G. Rasul 1 , Nanjappa Ashwath 2 and Md Mofijur Rahman 1, *
1 School of Engineering and Technology, Central Queensland University, North Rockhampton,
Queensland 4702, Australia; [email protected] (M.A.); [email protected] (M.G.R.)
2 School of Health, Medical and Applied Sciences, Central Queensland University, North Rockhampton,
Queensland 4702, Australia; [email protected]
* Correspondence: [email protected] or [email protected]; Tel.: +617-4930-6371
Abstract: In this study, the production process of second-generation biodiesel from Australian
native stone fruit have been optimised using response surface methodology via an alkali catalysed
transesterification process. This process optimisation was performed varying three factors, each
at three different levels. Methanol: oil molar ratio, catalyst concentration (wt %) and reaction
temperature were the input factors in the optimisation process, while biodiesel yield was the key
model output. Both 3D surface plots and 2D contour plots were developed using MINITAB 18
to predict optimum biodiesel yield. Gas chromatography (GC) and Fourier transform infrared
(FTIR) analysis of the resulting biodiesel was also done for biodiesel characterisation. To predict
biodiesel yield a quadratic model was created and it showed an R2 of 0.98 indicating the satisfactory
performance of the model. Maximum biodiesel yield of 95.8% was obtained at a methanol: oil molar
ratio of 6:1, KOH catalyst concentration of 0.5 wt % and a reaction temperature of 55 ◦ C. At these
reaction conditions, the predicted biodiesel yield was 95.9%. These results demonstrate reliable
prediction of the transesterification process by Response surface methodology (RSM). The results
also show that the properties of the synthesised Australian native stone fruit biodiesel satisfactorily
meet the ASTM D6751 and EN14214 standards. In addition, the fuel properties of Australian native
stone fruit biodiesel were found to be similar to those of conventional diesel fuel. Thus, it can be said
that Australian native stone fruit seed oil could be used as a potential second-generation biodiesel
source as well as an alternative fuel in diesel engines.
Keywords: response surface methodology; RSM; second-generation biodiesel; stone fruit; optimisation;
biodiesel testing; transesterification
1. Introduction
Global climate change and the resulting desire for renewable energy sources has generated the
interests for using biofuel in the transport sector [1]. Due to the higher production of biofuel in recent
years, it currently contributes 1.5% global transportation fuel. It has been reported that nearly 40%
of the total worldwide biofuel supply comes from emerging and developing countries. However,
the expansion of biofuel production around the world has raised major concerns, for example the
existence of several first-generation biofuels. Biofuels that are produced from edible sources are
termed first-generation biofuels [2], and these have been increasingly questioned over some concerns
such as food-fuel controversy, environmental pollution, and climate change. The increasing concern
regarding the sustainability of several first-generation biofuels has led to investigations into the
potential of producing biodiesel from non-food crops which are termed as second-generation biodiesel.

Energies 2018, 11, 2566
The potential benefits offered by the second-generation biodiesels are that they consume waste oils,
make use of abandoned land and do not compete with food crops [3].
In addition, second-generation biofuel from locally available sources can play a great role
in economic development of rural and emerging region of a country [4]. Despite significant
socio-economic advantages and continuous support from government and non-government
organisations, the market for biofuel production around the world has not expanded very much
over the last few years. Many countries have announced second-generation biofuel support policies,
e.g., the United States has adopted the policies to produce 60 billion litres by 2022 and the European
Union set their target to use 20% renewable energy in the transport sector by 2020 [5]. Both the US
and EU policies could play an important role for the worldwide biofuel development because of
their market size and considerable amount of biofuel imports. In addition, the Australian Federal
Government and its State Governments have developed relevant policies to promote a sustainable
biofuel industry to ensure Australian’s long-term energy security. Leading oil companies such as
Caltex, Shell, BP, and Exxon Mobile are also coming forward in second-generation biofuel research
with more investment. A few plants with research activities are going to be established soon in
emerging countries.
Biodiesel is one of the biofuels and has proved its potential as an alternative fuel worldwide.
Biodiesel is biodegradable, renewable and environmentally friendly [6,7]. The feedstock selection of
biodiesel is very important as 75% of the total cost of biodiesel production is associated with obtaining
feedstocks alone. A high oil yield of any feedstock ensures a commercial scale biodiesel production
at reasonable prices [8]. Feedstock security of supply, feedstock cost of supply and feedstock storage
are the important factors to consider when choosing the biodiesel feedstock [9]. In addition, biodiesel
should be produced from the feedstock that is consistently available, economically viable and locally
available. Currently, the major feedstocks for biodiesel production in Australia are waste cooking oil,
animal tallow, macadamia, beauty leaf, canola and mustard oils [10]. However, stone fruit such as
Prunus armeniaca L. is widely cultivated in Australia, and it yields 22–38% of kernels which contain
up to 54.2% oil. Australia produces about 100,000 tons of summer stone-fruit from October to April
each year and, in 2008, about 16,917 tons of Prunus armeniaca L. fruit were produced from all six
mainland states in Australia. This could therefore be a potential second-generation biodiesel feedstock
in Australia. The main aim of this research was to investigate and optimise the production process
of second-generation biofuel from this Australian native feedstock as the research on it is still far
behind that into other feedstocks. This biodiesel could overcome the limitations associated with
first-generation biodiesels and be used as an alternative to conventional fossil fuels.
2. Literature Review
Stone fruit is similar to a small peach, generally 1.5–2.5 cm in diameter, with its colour varying
from yellow to orange or red. Its single seed is enclosed in a hard stony shell. During fruit processing,
the seeds are discarded due to the presence of hydro-cyanic acid [11]. To utilise this waste product,
it is important to optimise the procedures involved in oil extraction and its conversion into biodiesel.
Many researchers [12–17] have optimised the production of biodiesel from different first- and
second-generation feedstocks using various methods. For example, Saydu et al. [13] optimised the
process of biodiesel production from hazelnut and sunflower oil using single step transesterification
with methanol, and employing potassium hydroxide as a catalyst. Razack and Duraiarasan [12]
optimised the waste cooking oil biodiesel production process using response surface methodology
using encapsulated mixed enzyme as a catalyst. Dharma et al. [17] optimised the biodiesel production
process of Jatropha curcas and Ceiba pentandra oil using response surface methodology as also did
Ong et al. [16] for the Calophyllum inophyllum biodiesel production process.
A few studies have been done on the optimisation of the stone fruit oil (SFO) biodiesel production
process but none of them used any statistical modelling. For instance, Gumus et al. [18] used
alkali transesterification with methanol and potassium hydroxide catalyst for producing SFO methyl
290
Energies 2018, 11, 2566
ester. Abdelrahman [19] produced SFO biodiesel via alkali transesterification with 0.75% potassium
hydroxide catalyst and at a methanol: oil molar ratio of 6:1. Faizan et al. [20] showed that the wild
Prunus Armeniaca L. oil can be transesterified by a single step process via the use of sodium hydroxide
catalyst at a methanol: oil molar ratio of 6:1, and reported a biodiesel yield of 93%. Ashok et al. [21]
performed single step alkali transesterification using 1% potassium hydroxide as a catalyst at 55 ◦ C and
60 min reaction time with a constant stirring at 400 rpm and obtained a biodiesel yield of 96.5%. Thus,
many process parameters, including reaction temperature, catalyst type and catalyst concentration,
type of alcohol used, the oil to methanol molar ratio, reaction time and agitation speed have been
found to influence the optimum transesterification process [22–26]. From the above literature, it is
obvious that no/limited investigation has been done on the optimisation of second-generation biodiesel
production process from Prunus Armeniaca L. oil using any statistical modelling. Thus, this study
has explored optimisation of the biodiesel production processes from Australian native stone fruit oil
using response surface methodology.
3.1. Materials
Stone fruit (Prunus armeniaca L. species) seed oil was purchased from a local producer named
Chromium Group Pty Ltd. of Eumundi, Queensland, Australia. The chemicals used in this study
were methanol (99.9% purity), potassium hydroxide (KOH pellets, 99% purity) and sodium hydroxide
(NaOH pellets, 99% purity). All were of analytical reagent grade (AR) and were procured from the
School of Engineering and Technology, Central Queensland University, Rockhampton, Australia.
A three-neck laboratory reactor (1 L) along with a reflux condenser and a thermocouple placed on a
magnetic heater/stirrer were used in the SFO biodiesel conversion experiments. In this experiment,
methanol, KOH, NaOH and Whatman 541 grade filter paper (pore size 22 μm) were used.
3.2. Oil Extraction

The stone fruits were collected and the fleshy parts were separated manually for drying purposes.
The seeds were separated for kernel collection and oil extraction. For easy breaking of the hard shell,
seeds were softened by immersing in water for 10–20 min. The broken shell can be used as fertilizer or
firewood after the oil extraction process [27]. Kernels were separated from the broken shells and were
crushed using a pestle and mortar and sieved through a 40 mesh or 0.8 mm sieve [19,28]. The ground
kernel was placed in a Soxhlet apparatus and the oil was extracted using petroleum ether (40–60 ◦ C)
over 6–8 h until the extraction was completed [19]. After oil extraction, the petroleum ether was
evaporated using a rotary evaporator at 25 ◦ C [29]. The oil was placed in an oven at 60 ◦ C for 60 min to
remove the remaining solvent. The oil was filtered using Whatman 541 filter paper. After filtration, the
SFO was kept in a sealed container for characterisation. The oil yield was calculated using Equation (1).
weight of oil extracted (g)

Oil Yield (%wt/wt dry kernels) = × 100 (1)
weight of dry kernels used (g)
3.3. Biodiesel Production

The acid value of raw SFO was determined as 1.65 mg KOH/g. After transesterification using
KOH catalyst transesterification, the acid value of SFO biodiesel was found to be 0.25 mg KOH/g.
This trial experiment suggested that only the single stage alkyl catalyst transesterification process was
satisfactory for SFO biodiesel production. Thus, in each experiment, the experiment was performed by
reacting a known quantity of SFO with methanol and the catalyst.
Initially, the SFO was poured into a three-neck laboratory reactor and heated to the desired
temperature. The measured quantities (molar basis) of methanol and catalyst (KOH) were poured
into a separate beaker and stirred vigorously using a magnetic stirrer at 50 ◦ C at 600 rpm for 10 min
291
Energies 2018, 11, 2566
to produce methoxide. This solution was slowly poured into the three-neck reactor containing SFO.
The blend was agitated continuously at 600 rpm and the temperature and reaction time were varied as
per the experimental design. At the end of the transesterification reaction, the blend was transferred
into a separating funnel. Although the separation of glycerol and biodiesel occurred instantaneously,
the funnel was left undisturbed for 24 h. Two separate liquid phases were formed, with the top layer
being methyl ester (biodiesel) and the bottom layer of red viscous glycerol and impurities. The bottom
layer was drained off while the top layer was collected and washed with warm (50 ◦ C) distilled water.
This moist biodiesel was then heated to 110 ◦ C for 15 min to remove residual water that would have
been retained by the biodiesel during the washing process. The Whatman® qualitative Grade 1 filter
paper was used to filter the biodiesel and finally stored in an airtight container at room temperature
until its characterisation. Biodiesel yield was calculated using Equation (2) and its composition was
determined using a Gas Chromatogram [22]. The graphical illustration of the SFO production process
is shown in Figure 1.
weight of SFO biodiesel

SFO Biodiesel Yield = FAME percent from GC analysis × (2)
weight of Stone fruit oil
Figure 1. Graphical representation of producing biodiesel from SFO biodiesel.
292
Energies 2018, 11, 2566
3.4. Physicochemical Characterisation of SFO

This section discusses the international standards used in characterising the SFO biodiesel and its
composition via Gas Chromatography (GC) and Fourier Transform Infrared (FTIR) spectroscopy.
3.4.1. Fuel Properties

The physicochemical properties and fatty acid compositions of crude SFO and SFO biodiesel were
tested according to ASTM and EN standards. The properties studied were density at 15 ◦ C (ASTM
D1298), kinematic viscosity at 40 ◦ C (ASTM D445), acid value (ASTM D664), calorific value (ASTM
D240), flash point (ASTM D93) and oxidation stability (ASTM D2274). The fatty acid compositions
were determined using a gas chromatograph according to EN 14103.
Fuel properties calculated based on the fatty acid composition of the SFO biodiesel were cetane
number (CN), saponification value (SV), iodine value (IV), long-chain saturated factor (LCSF) and
degree of unsaturation (DU). The numerical calculations were determined using the following
equations [30]:
5458
CN = 46.3 + − (0.225 × IV) (3)
SV
(560 × Ai )
SV = ∑ MWi
(4)
(254 × D × Ai )
IV = ∑ MWi
(5)
LCSF = 0.1 × (C16 : 0 wt %) + 0.5 × (C18 : 0 wt %) + 1 × (C20 : 0 wt %)

(6)
+ 1.5 × (C22 : 0 wt %) + 2.0 × (C24 : 0 wt %)
DU = ∑ MUFA + (2 × PUFA) (7)
where D indicates the number of double bonds, Ai is the percentage of each fatty acid in the FAME,
and MWi is the molecular mass of each component. MUFA denotes monounsaturated fatty acid and
PUFA refers to polyunsaturated fatty acid. The degree of unsaturation was calculated using both
MUFA and PUFA concentrations. Fatty acids of C16:0, C18:0, C20:0, C22:0 and C24:0 stand for palmitic
acid, stearic acid, arachidic acid, behenic acid and lignoceric acid, respectively, and were used for
measuring the long chain saturated factor.
3.4.2. Gas Chromatography

According to EN14103, a gas chromatograph (GC) (Thermo Scientific Trace 1310 GC) was used to
determine the fatty acid composition of the SFO. 25 mg of the SFO biodiesel was dissolved in high
purity hexane (10 mL). Then, this solution was poured to 2 mL auto-sampler vials. The equipment
for the GC test included Thermo Scientific Trace 1310 GC with a split/split less (SSL) injector, flame
ionisation detector, and TriPlus auto-sampler. At 240 ◦ C, a 1 μL sample was injected in split mode
(40:1) by maintaining a constant helium flow of 1.2 mL/min. The conditions for separating FAMEs
were: using a BPX-70 column (60 m × 250 μm × 0.25 μm film) with a temperature program: 110 ◦ C
(4 min); 10 ◦ C/min, 150 ◦ C; 3.9 ◦ C/min, 230 ◦ C (5 min). In this study, SFO biodiesel individual
components were identified by retention time compared to a standard FAME mixture that had certified
concentrations, namely Supelco CRM18920 (FAME C8-C22). Chromeleon 7.2 software was used for
data acquisition and processing.
3.4.3. Fourier Transform Infrared (FTIR) Spectroscopy

The various functional groups present in the crude oil and the biodiesel sample were determined
with Fourier transform infrared (FTIR) spectroscopy. The Spectrum 100 FTIR spectrometer with a
universal Attenuated total reflectance (ATR) sampling accessory (Perkin Elmer, Melbourne, Australia)
293
Energies 2018, 11, 2566
was used to record ATR infrared spectra. SFO biodiesel samples were placed directly on the ATR
window at approximately 40% transmission to record the spectra with four scans, 4000–650 cm−1 .
After ATR correction, Spectrum 6.2.0 software was used to acquire data and processing.
3.5. Design of Experiments

Box-Behnken is one of the most commonly used responses surface methodology designs.
This design was used for designing and statistical analysis of this experiment. The Box-Behnken
design matrix was utilised to find the optimum conditions for maximum biodiesel yield production.
The experimental optimisation was achieved via analysis of variance (ANOVA) using Minitab 18
software. The effects of process factors such as methanol: oil molar ratio, KOH catalyst concentration,
and reaction temperature were tested. Using these three factors at three levels required a total of
15 runs for identifying the optimum conditions for transesterification. The coded symbols, ranges,
and levels of the investigated factors are listed in Table 1. The design matrix for the three factors was
varied at three levels, namely −1, 0 and +1. The range levels of the factors investigated were chosen by
considering the initial tests carried out on the effect of individual factors on biodiesel yield as well as
the operating limits of the biodiesel production process conditions as evidenced from the literature.
Table 1. Experimental range and levels coded for independent factors.
Range and Levels

Factors/Variables Unit Symbol Coded
−1 0 +1
Methanol: Oil ratio mol/mol M 4:1 5:1 6:1
KOH catalyst
wt % C 0.5 1.00 1.5
concentration
Temperature ◦C T 45 55 65
Methanol: oil molar ratio ranged from 4:1 to 6:1, catalyst concentrations were 0.5–1.5% by weight
of oil and the reaction temperature was varied from 45 ◦ C to 65 ◦ C (boiling point of methanol). Once
the experiments were completed, the response factor (biodiesel yield) was applied in a full quadratic
model to correlate the response factor to the independent factors. The general form of the full quadratic
model is shown in Equation (8).
Y = P0 + P1 Q1 + P2 Q2 + P3 Q3 + P1,2 Q1 Q2 + P1,3 Q1 Q3 + P2,3 Q2 Q3 + P1,1 Q1 2 + P2,2 Q2 2 + P3,3 Q3 2 (8)
where Y is the response factor (biodiesel yield, %); P0 is a constant; P1 , P2 , and P3 are regression
coefficients; P1,1 , P1,2 , P1,3 , P2,2 , P2,3 , and P3,3 are quadratic coefficient; and Q1 , Q2 , and Q3 are
independent variables.

This section includes the results of the characterisation of both crude SFO and SFO biodiesel,
fatty acid compositions of SFO biodiesel, optimisation of reaction conditions by response surface
methodology and response surface plots for SFO biodiesel production.
4.1. Characterisation of Crude SFO

The properties of crude stone fruit seed oil used in this study were evaluated prior to the
optimisation process. Physicochemical properties are the most important features to check the quality
of any crude oil. The SFO was characterised by viscosity, density, specific gravity, acid value, calorific
value, saponification number and iodine value. The properties of SFO from this study along with those
from other studies and those of petro diesel were compared and are presented in Table 2. The density
of the oil was found to be 910 kg/m3 which matches with that reported in the literature. Again, the acid
value of SFO was determined to be 1.65 mg KOH/g, indicating the presence of low levels of free fatty
294
Energies 2018, 11, 2566
acids in the oil. The kinematic viscosity of the oil was found to be 34.54 m2 /s and the calorific value
was 38.45 MJ/kg, which is within the values found in the literature [18]. Based on above results, it is
clear that Australian native SFO oil have similar fuel properties including fatty acid, calorific value
and viscosity with the data of other researchers, thus it is expected that Australian native SFO may
serve as a good feedstock for biodiesel production.
Table 2. Physical and chemical properties of SFO.
Properties Units SFO This Study SFO [18] SFO [21] SFO [11] Petro Diesel
Kinematic Viscosity @ 40 ◦ C m2 /s 34.54 34.82 20.53 26.22 3.23
Density kg/m3 910 920 913 916.6 827.2
Specific Gravity @ 15 ◦ C g/cm3 0.91 0.91 0.91 0.91 0.83
Acid value mg KOH/g 1.65 2.60 0.68 0.05
Calorific value MJ/kg 38.4 39.6 31.5 45.3
Saponification number mg KOH/g 173 188 187
Iodine value mgI2 /100 g 103 90 101
4.2. Properties and Qualities of SFO Biodiesel

The physical and chemical properties of SFO biodiesel from this study, along with results
from other researchers’ work on SFO biodiesel, are compared with other non-edible biodiesels and
petrodiesel in Table 3. It was found that all the properties and qualities of the SFO biodiesel fulfilled the
international standards (USA ASTM D6751 and European Union EN14214). Many researchers [31–34]
showed that densities of biodiesels do not vary considerably, as the density of methanol and oil are close
to the density of produced biodiesel, which usually varies between 850 and 900 kg/m3 . The density
of SFO and papaya seed oil (PSO) was found to be 855 and 840 kg/m3 respectively, whereas that of
petrodiesel was found to be 827.2 kg/m3 . Densities of other SFO biodiesels also matched with the
international standards. Karanja biodiesel has a density of 931 kg/m3 , which is outside the ASTM
and EN standards specification, thus limiting its efficiency of fuel atomisation in airless combustion
systems [35]. However, other biodiesels (Table 3) have slightly higher densities than petrodiesel fuel,
but they are within the range of the international standards. The viscosity of SFO biodiesel was
determined to be 4.26 mm2 /s and other biodiesels (except Karanja) ranged from 1.9–6.0 mm2 /s and
also fulfil the requirements of the standards. The viscosity of SFO biodiesels of other studies was
found to be within the range as well. The acid values of biodiesels (except Neem) were also in line
with requirements of ASTM and EN biodiesel standards which are less than 0.5 mg KOH/g. Higher
acid values can cause corrosion of IC engines and internal metal parts. Cetane number is an important
fuel property for diesel engines. A higher speed diesel engine works more efficiently with a fuel with a
higher cetane number. A lower cetane number fuel has longer ignition delays providing more time
to complete the combustion process. The cetane number of biodiesel increases with an increase in
fatty acid proportion. Longer fatty acid chains and higher saturated fatty acid content will lead to
a higher cetane number [36,37]. Moringa biodiesel has the highest cetane number of 67.1 compared
with all other biodiesels in Table 3 and is within the international standard limit. All calorific values
are lower than those of petrodiesel fuel (45.3 MJ/kg). The SFO biodiesel calorific value was found to
be 39.64 MJ/kg in this study, thus meeting the minimum EN standard requirement of 35 MJ/kg. All
other SFO biodiesels have similar calorific values as well. The flashpoint of this SFO biodiesel was
found to be 105 ◦ C, whereas the ASTM standard specifies 100–170 ◦ C and petrodiesel fuel is 68.5 ◦ C.
This suggests that SFO biodiesel fuel is safer to handle and store than petroleum diesel. The iodine
value for this SFO was recorded as 104.7 mgI2 which met the range of the EN standard. The higher
the iodine value, the more unsaturated double bonds are present in the methyl ester, leading to better
biodiesel fuel quality. Biodiesel with higher oxidation stability is preferable as low oxidation stability
can affect the quality of biodiesel [38]. This SFO biodiesel has an oxidation stability of 7.15 h, which
falls above the minimum values of both the ASTM (minimum 3 h) and EN standards (minimum 6 h).
295
Energies 2018, 11, 2566
Some biodiesels with poor oxidation stability such as Tobacco, Cottonseed, Jatropha and Moringa
biodiesel can be easily remedied by adding antioxidants.
Table 3. Comparison of SFO biodiesel with other non-edible biodiesels.
Viscosity Cetane Calorific Oxidation

Non-Edible Density Acid Value, Flash Iodine Value
at 40 ◦ C, Number Value, Stability
Biodiesels (kg/m3 ) mg KOH/g Point, ◦ C (IV) mgl2 /100 g
mm2 /s (CN) MJ/Kg (OS), h
SFO this study 855.0 4.26 0.25 50.45 39.64 105 104.70 7.15
SFO [18] 884.3 4.92 39.95 111
SFO [21] 857.0 5.20 0.32 58.70 38.93 180 100.70 6.30
SFO [11] 879.4 4.21 0.08 39.12 170 100.66
Petro diesel [39] 827.2 3.23 0.05 48.00 45.30 68.5 38.3 39.0
Tobacco [23,38] 888.5 4.23 51.60 165.4 136 0.80
PSO [39] 840.0 3.53 0.42 48.29 38.49 112 115.89 5.61
Jatropha [38] 879.5 4.80 0.40 51.60 39.23 135 104 2.30
Rapeseed [23] 882.0 4.43 54.40 37.00 170 7.60
Cottonseed [23] 875.0 4.07 0.16 54.13 40.43 150 1.83
Neem [23,38] 868.0 5.21 0.65 39.81 76 7.10
Karanja [38] 931.0 6.13 0.42 55.00 43.42 95
Moringa [23] 883.0 5.00 0.18 67.1 160 74 2.3
ASTM D6751 880.0 1.9~6.0 maximum 0.5 minimum 47 93~170 minimum 3
EN14214 860~900 3.5~5.0 maximum 0.5 minimum 51 35 >120 maximum 120 minimum 6
4.3. The Fatty Acid Composition of SFO Biodiesel

The fatty acid composition of any biodiesel feedstock is an important fuel property. The fatty
acid composition is highly dependent on the quality of the feedstock, its growth condition and the
geographic location in which the plant has grown. The chromatogram of the SFO biodiesel produced
in this study shows the existence of derivatives of C16:0 (palmitic acid), C18:0 (Stearic acid), C18:1
(oleic acid), C18:2 (linoleic acid), C18:3 (linolenic acid), and C22:1 (behenic acid) in Figure 2. GC
chromatogram of SFO biodiesel ensured the formation of methyl ester.

Figure 2. GC Chromatogram of SFO biodiesel.
The fatty acid compositional analysis of SFO biodiesel produced in this study is shown in Table 4,
which indicates a high level (89.7%) of unsaturated fatty acids made up of both polyunsaturated
and monounsaturated fatty acids. Saturated fatty acids such as palmitic acid, stearic acid, and
behenic acid were found to be present at 5.85%, 2.51%, and 0.66%, respectively. Monounsaturated
fatty acids (MUFA) such as oleic acid were found to be the dominant fatty acids (63.84%), whereas
polyunsaturated fatty acids (PUFA) such as linoleic acid and linolenic acid were 25.34% and 0.51%,
respectively. Pedro et al. [40] indicated that the degree of unsaturation of biodiesel did not significantly
affect the engine performance and the start of injection, but it had a significant influence on combustion
characteristics and emissions. The degree of unsaturation was recorded as 115.54% in this study.
296
Energies 2018, 11, 2566
Altun [41] indicated that degree of unsaturation and cetane number of biodiesel could highly influence
NOx formation in biodiesel-fuelled diesel engines. Generally, higher degrees of saturation relate to
higher cetane numbers of biodiesel. Unsaturated biodiesel produces higher NOx and lowers HC
emissions than saturated biodiesels [40]. Furthermore, higher degrees of unsaturation in crude oil
results in the production of less viscous biodiesel.
Table 4. The fatty acid composition of SFO.
Fatty Acids. Formula Molecular Weight Structure wt %

Palmitic C16 H32 O2 256 16:0 5.85
Stearic C18 H36 O2 284 18:0 2.51
Oleic C18 H34 O2 282 18:1 63.8
Linoleic C18 H32 O2 280 18:2 25.3
Linolenic C18 H30 O2 278 18:3 0.51
Behenic C22 H44 O2 340 22:0 0.66
Others 1.29
Total Saturated Fatty Acids (SFA) 9.02
Total Monounsaturated Fatty Acids (MUFA) 63.84
Total Polyunsaturated Fatty Acids (PUFA) 25.85
The degree of Unsaturation (DU) 115.5
Long Chain Saturated Factor (LCSF) 2.83
The distribution of the main fatty acids of some non-edible biodiesel feedstocks (including the SFO
in this study and those from other research works on SFO) are shown in Table 5. Gas chromatography
(GC) analysis of those biodiesels shows that the most abundant fatty acids are oleic acid, linoleic acid,
and palmitic acid. The SFO biodiesel produced in this study shows similar results compared with
those of previous research [18,21,28,42,43] on SFO biodiesel fuel. The presence of higher MUFA and
PUFA can contribute to lower oxidative stability. As mentioned earlier, low oxidative stability could
affect the quality of produced biodiesel. This low oxidation has a negative impact on both kinematic
viscosity and acid value [44]. However, the presence of high contents of MUFA and PUFA ensures
biodiesel with good fuel flow properties (especially in cold prone countries) compared with saturated
fatty acids (SFA).
Table 5. Variations in the main fatty acid compositions of selected biodiesel feedstocks.
Fatty Acids (% w/w)

Non-Edible Biodiesels Ref.
C16:0 C18:0 C18:1 C18:2 C18:3 Others
SFO this study 5.85 2.51 63.84 25.34 0.51 1.95
SFO 5.62 1.27 67.31 24.68 0.08 1.04 [18]
SFO 4.20 2.32 71.00 20.15 1.20 1.13 [21]
SFO 3.87 0.92 67.21 27.12 0.11 0.77 [28]
SFO 3.79 1.01 65.23 28.92 0.14 0.91 [42]
SFO 66.20 28.20 5.60 [43]
PSO 6.07 3.13 47.73 37.25 1.78 4.04 [39]
Tobacco 8.90 3.50 14.10 70.10 1.00 2.40 [45]
Jatropha 16.20 8.20 38.40 36.80 0.40 0 [46]
Rapeseed 3.49 0.85 64.40 22.30 8.23 0.73 [47]
Cottonseed 28.70 0.90 13.00 57.40 0 0 [48]
Neem 12.00 10.00 61.00 16.00 0 1.00 [49]
Karanja 9.80 6.20 72.20 11.80 0 0 [50]
Moringa 6.50 6.00 72.20 1.0 0.65 13.65 [51]
The comparison of different non-edible biodiesels and SFO biodiesel fatty acid compositions are
shown in Figure 3. It can be seen that MUFA-oleic acid (C18:1) dominates most of the feedstocks (except
tobacco and cottonseed), followed by PUFA-linoleic acid (C18:2) and SFA-palmitic acid (C16:0). Both
297
Energies 2018, 11, 2566
tobacco and cottonseed had high proportions of PUFA and it can be seen in Table 3 that their oxidation
stabilities are very poor (0.8 h and 1.83 h, respectively). Oxidation occurs due to the presence of high
proportions of unsaturated fatty acid chains and double bonds, i.e., PUFA in the parent molecule reacts
with oxygen as soon as it is exposed to air [23,52]. Therefore, biodiesels with high linoleic acid (C18:2)
and linolenic acids (C18:3) such as tobacco and cottonseed tend to have lower oxidation stabilities.

6)2 362 7REDFFR -DWURSKD &RWWRQVHHG 1HHP .DUDQMD 0RULQJD 5DSHVHHG
3DOPLWLFDFLG& 6WHDULFDFLG& 2OHLFDFLG&

/LQROHLFDFLG& /LQROHQLFDFLG& 2WKHUV
Figure 3. Comparison of the fatty acid composition of SFO and other non-edible vegetable oils.
Figure 4 shows the FTIR spectrum of the SFO methyl ester produced in this study. The advantages
of using the FTIR method compared with GC is the ability to analyse whole samples, including
precipitated fractions, without any further preparation [53].
Figure 4. Fourier transforms infrared (FTIR) spectrum of SFO biodiesel.
The most important functional groups, wave number, band assignment and absorption intensity
of absorption peaks detected in the FTIR spectrum of the SFO methyl ester are presented in Table 6.
The peak at 1435.8 cm−1 corresponds to the asymmetric stretching of –CH3 present in the SFO biodiesel
sample, which is shown in Figure 4. The peak in the region of 2800–3000 cm−1 represents the CH3
asymmetric stretching vibration. The peak of stretching of the carbonyl group (–C=O) is 1742 cm−1
located in the region of 1800–1700 cm−1 which is common for esters. The fingerprint region of
1500–900 cm−1 is the major spectrum from the SFO methyl ester which has a peak at 1244.8 cm−1 ,
corresponding to the bending vibration of –CH3 [5]. These results reflect the conversion of triglycerides
to methyl ester.
298
Energies 2018, 11, 2566
Table 6. Functional groups of SFO biodiesel detected in the FTIR spectrum.
Wavenumber Group Functional Absorption

Vibration Type
(cm−1 ) Attribution Groups Intensity
2924 =C–H Asymmetric stretching vibration Alkyl Strong
2854 –CH2 Symmetric stretching vibration Aromatic Strong
1742 –C=O Stretching Carbonyl Strong
1460.5 –CH2 Shear-type vibration Alkanes Weak
1244.8 –CH3 Bending vibration Alkanes Weak
1195.7 C–O–C Anti-symmetric stretching vibration Ethers Middling
1169.7 C–O–C Anti-symmetric stretching vibration Ethers Middling
1120.4 C–O–C Anti-symmetric stretching vibration Ethers Weak
1017.2 C–O–C Anti-symmetric stretching vibration Ethers Weak
722.7 –CH2 Plane rocking vibration Aromatic Weak
4.4. Optimisation of Reaction Conditions by RSM

The results of the Box–Behnken design model to optimise biodiesel production process parameters
are shown in Table 7. In the transesterification experiments, the SFO biodiesel yield ranged from 75.2%
to 95.8%. This design matrix also shows the experimental run order, experimental yields and predicted
yields. These results show that the biodiesel yield varies with the production process. To avoid
systematic errors, all run orders were randomised.
Table 7. Experimental matrix and results for Box-Behnken design model. The combination in italics
shows the best combination for SFO biodiesel production.
Exp. Run Methanol: Oil KOH Temp SFO Biodiesel Yield (%)
M C T
Number Order (Molar Ratio) (wt %) (◦ C) Experimental Predicted
1 7 0 −1 −1 5 0.5 45 86.65 85.27
2 13 1 −1 0 6 0.5 55 95.75 95.89
3 14 0 0 0 5 1 55 91.65 92.98
4 12 −1 0 −1 4 1 45 80.34 80.85
5 15 1 0 −1 6 1 45 87.71 88.95
6 6 −1 −1 0 4 0.5 55 75.24 76.12
7 9 −1 1 0 4 1.5 55 82.27 82.13
8 11 0 0 0 5 1 55 93.65 92.98
9 5 0 0 0 5 1 55 93.65 92.98
10 8 0 1 −1 5 1.5 45 84.58 84.21
11 4 1 1 0 6 1.5 55 82.11 81.23
12 3 1 0 1 6 1 65 89.33 88.82
13 2 −1 1 1 4 1 65 79.30 78.06
14 1 0 −1 1 5 0.5 65 86.71 87.08
15 10 0 1 1 5 1.5 65 78.10 79.48
The predicted biodiesel yield values were generated from a quadratic regression model as obtained
from Minitab software version 18.0 through response surface methodology (RSM) statistical analysis of
the experimental data. The Minitab 18 program was used to calculate the effects of each parameter and
its interactions with other parameters. The response parameter (biodiesel yield %) was correlated with
other parameters using a full quadratic regression model shown in Equation (9). The model represents
SFO biodiesel predicted yield (Y) as a function of methanol: oil molar ratio (M), catalyst concentration
(wt %) (C) and reaction temperature (T).
Y = 92.983 + 4.719M − 2.161C − 0.730T − 4.490M2 − 4.650C2 − 4.323T2 − 5.168MC

(9)
+ 0.665MT − 1.635CT
299
Energies 2018, 11, 2566
Figure 5a shows the line of a perfect fit with points corresponding to zero error as indicated
by the comparison of experimental and predicted biodiesel yields. Points closer to the straight line
indicate a good agreement between the experimental and predicted values. Figure 5b shows that there
is an adequate correlation between RSM predicted values and experimental values, which verifies the
acceptability of the model. The model represents a relatively good description of the experimental
data regarding the SFO yield.

(a) (b)
Figure 5. Biodiesel yield values (%): (a) Experimental versus (b) RSM predicted.
The linear, quadratic and interaction effects of the parameters were considered to investigate the
impacts on the biodiesel yield. Table 8 displays the significance of those parameters in terms of the
probability value (p value). It also summarises the resulting regression coefficients and computed
T-values. In the model, positive coefficients M and MT showed a linear effect on biodiesel yield,
whereas quadratic terms of M2 , C2 , T2 , MC, and CT had adverse effects that reduce the biodiesel yield.
At the 95% confidence level, the p values were less than 0.05, indicating significant effects of those
parameters. The analysis of variance (ANOVA) was used to determine the significance and fitness of
the quadratic model.
Table 8. Regression coefficient of the predicted quadratic polynomial model.
Term. Coefficients Standard Errors Computed T-Value p Value

Constant 92.983 0.893 104.17 0.000
M 4.719 0.547 8.63 0.000
C −2.161 0.547 −3.95 0.011
T −0.730 0.547 −1.34 0.239
M×M −4.490 0.805 −5.58 0.003
C×C −4.650 0.805 −5.78 0.002
T×T −4.323 0.805 −5.37 0.003
M×C −5.168 0.773 −6.69 0.001
M×T 0.665 0.773 0.86 0.429
C×T −1.635 0.773 −2.12 0.088
Table 9 shows the level of significance of individual terms and their interactions on the selected
response. The quadratic regression model has higher F value (24.76) and lower p value (0.001) than
significance level (p < 0.05), which indicates that the model is significant at the 99.9% confidence level.
The p value represents the probability of error and is used to check the significance of each regression
coefficient. The interaction effect of each cross product can be revealed through the p value. It is found,
M (methanol: oil molar ratio), C (catalyst concentration), M2 (quadratic effect of methanol amount),
C2 (quadratic effect of catalyst concentration), T2 (quadratic effect of reaction temperature) and MC
300
Energies 2018, 11, 2566
(methanol amount with catalyst) have significant effects on SFO biodiesel production. Among all other
parameters, methanol: oil molar ratio (M) has the lowest p value (0.000) and highest F value (74.53).
These results show that M is the most important parameter in SFO biodiesel production. According
to the regression model in Equation (9), M has a positive effect and both C and reaction temperature
(T) have negative effects on SFO biodiesel yield. This implies that increasing M will increase the
speed of the transesterification process. However, increase in C and T will slow the speed of the
transesterification reaction. The square term of T2 was also significant although it has a smaller F value
compared to its corresponding linear term which indicated its weaker influence in the model. Again,
the ANOVA results showed that the linear term of T with p value was not significant (more than 0.05)
and its quadratic term T2 with p value was significant (less than 0.05). ANOVA also showed that both
C and C2 terms were significant with their F values (15.63 and 27) which indicated their medium effect
in the model. The Lack of Fit was also determined for this regression model. F value and p value of
Lack of Fit parameters were found to be 2.32 and 0.315 respectively. The p value (0.315) of the Lack of
Fit parameter is greater than 0.050, which indicated the quadratic model has an insignificant Lack of
Fit, i.e., the model sufficiently described a relationship between independent parameters such as M, C
and T, with the dependent parameter (SFO biodiesel yield). The coefficient of determination (R2 ) was
employed to identify the quality of the model fitness. R2 also indicates the good correlation between
the independent parameters. In this study, R2 was found to be 97.8% and the adjusted coefficient of
determination (Adj. R2 ) was 93.9%. This means that the model explains 97.8% of the variation in the
experimental data. In conclusion, the regression model developed for SFO biodiesel yield was valid
and showed a satisfactory experimental relationship between the response and parameters.
Table 9. ANOVA results for SFO biodiesel.
Sum of Degree of Mean

Source F-Value p-Value Remarks
Squares Freedom Square
Model 532.603 9 59.178 24.76 0.001 Significant
M-Methanol 178.123 1 178.123 74.53 0.000 Highly significant
C-Catalyst 37.364 1 37.364 15.63 0.011 Significant
T-Temperature 4.263 1 4.263 1.78 0.239 Not significant
M2 74.447 1 74.447 31.15 0.003 Significant
C2 79.847 1 79.847 27.00 0.002 Significant
T2 69.004 1 69.004 28.87 0.003 Significant
MC 106.823 1 106.823 44.69 0.001 Significant
MT 1.769 1 1.769 0.74 0.429 Not significant
CT 10.693 1 10.693 4.47 0.088 Not significant
Lack of Fit 9.284 3 3.095 2.32 0.315 Not Significant
Pure Error 2.667 2 1.333
Total 544.553 14
R2 = 0.9781 Adj R2 = 0.9386
4.5. Response Surface Plots for SFO Biodiesel Production

The interactive effect of the two factors on the transesterification process for biodiesel production
is necessary for interpreting the impact of independent variables used in the optimisation process.
Figures 6–8 show both surface plots and contour plots of SFO biodiesel yield obtained by the regression
model in Equation (9). Surface plots were produced by plotting three-dimensional (3D) surface curves
against any two independent variables while keeping the other variables fixed at their medium values.
The interaction effect of the two parameters plotted while the third parameter was fixed at a medium
value in the contour plot. Contour plots can identify the variation in biodiesel yield with any change
in experimental conditions.
301
Energies 2018, 11, 2566
4.5.1. Interaction Effect of Methanol: Oil Molar Ratio and Catalyst Concentration
Figure 6a shows the 3D response surface for SFO biodiesel yield production as a function of
methanol: oil molar ratio, and KOH catalyst concentration under the current conditions of Box-Behnken
design matrix. With an increase of methanol: oil molar ratio up to 6:1 (highest) and 0.5 wt % of catalyst
concentration (lowest), biodiesel yield percentage increases. The maximum SFO biodiesel yield of
95.8% was found for KOH 0.5 wt % (Run 13). Table 7 design matrix indicated that highest KOH
concentration at 1.5 wt % and unchanged methanol: oil ratio at 6:1 resulted in lower SFO biodiesel
yield to 82.1% (Run 4). When the methanol:oil molar ratio remains unchanged at 6:1, and the catalyst
concentration is at the highest value of 1.5 wt %, the SFO biodiesel decreases to 82.1% (Run 4). When
the methanol: oil molar ratio was reduced to 4:1 (lowest level), and with the highest value of catalyst
concentration of 1.5 wt %, the biodiesel yield was found to be 82.3% (Run 9). Again, at methanol: oil
molar ratio of 6:1, and with the mid-value of catalyst concentration of 1wt %, the yield was found to be
89.3% (Run 3). On the other hand, when the methanol: oil molar ratio was reduced to 4:1, and with
the mid-value of catalyst concentration of 1 wt %, the yield dropped to 80.3% (Run 12). Methanol: oil
molar ratio affected total biodiesel yield production. ANOVA from Table 9 confirmed that both M, C
and MC interaction were significant. The 2D contour plot with MC interaction along with biodiesel
yield is shown in Figure 6b. It is easy to identify the optimum operating conditions and the related
response values (yield) through the 2D contour plot. Therefore, both M and C are significant factors
for higher biodiesel yield. Although literature showed that higher amount of C could result in less
biodiesel yield production, and eventually produce emulsion and phase separation [54].

(a) 3D response surface plot (b) 2D contour plot
Figure 6. Interaction effect of methanol: oil molar ratio (M) and catalyst concentration (C) on the SFO
biodiesel yield.
4.5.2. Interaction Effect of Catalyst Concentration and the Reaction Temperature

The interaction effect of catalyst concentration, C and reaction temperature, T on the SFO biodiesel
yield in 3D surface plots is shown in Figure 7a. An increase in T to mid-level (55 ◦ C), and the low-level
of C (0.5 wt %) can enhance the SFO biodiesel yield up to 95.8% (Run 13), considering M of 6:1 and
reaction time of 60 min constant, which is presented in Table 7 of the Box–Behnken design matrix.
It is found that increasing the T to 65 ◦ C and C to 1 wt % resulted in the decline in biodiesel yield
to 89.3% (Run 3). Similarly, increasing C to the highest level of 1.5 wt % and increasing T to the
mid-level of 55 ◦ C resulted in a stepwise decline in biodiesel yield to 82.1% (Run 11). Figure 7a displays
the interaction between C and T on biodiesel yield production up to 92%, keeping the experimental
conditions of T at 55 ◦ C and C at 1 wt %. ANOVA results in Table 9 confirms that the interaction
between C and T is not significant. Moreover, the higher amount of C and higher T might induce
saponification of triglycerides as well as form soap at the end [55]. Figure 7b shows the 2D contour
plots of CT interaction, which is not significant for SFO biodiesel yield production.
302
Energies 2018, 11, 2566

Figure 7. Interaction effect of reaction temperature (T) and catalyst concentration (C) on SFO
biodiesel yield.
4.5.3. Interaction Effect of Reaction Temperature and Methanol: Oil Molar Ratio
Figure 8a shows the 3D surface plot of reaction temperature, T and methanol: oil molar ratio,
M with SFO biodiesel yield. The T at mid-level (55 ◦ C) and mid-level M of 5:1 shows the maximum
yield. Table 7 Box–Behnken design matrix shows that, at the mid-level T and mid-level M, a biodiesel
yield of 93% can be achieved. However, at mid-level T and the highest level M, biodiesel yield was
optimised and found to be 95.8%. The overall biodiesel yield decreased significantly to 89.3% when T
reached 65 ◦ C (Run 3). Any change in T either by an increase or decrease from its mid-level (55 ◦ C)
resulted in reduced biodiesel yield. The optimum M was found to be 6:1 and any decreasing the molar
ratio (<6) lowered the biodiesel yield. Figure 8b shows the contour plot of interaction between M
and T. It shows that, at the mid-level of both T and M, biodiesel yield is maximum. Therefore, ANOVA
results in Table 9 confirms that both T and TM are not significant in SFO biodiesel production.
Figure 8. Interaction effect of methanol: oil molar ratio (M) and reaction temperature (T) on SFO
biodiesel yield.
From ANOVA results in Table 9, it is found that M is the only highly significant process factor
that affects the production of SFO biodiesel. Interaction of MC is also a significant process factor
for biodiesel production. Both the M and MC have higher F values of 74.5 and 44.7, respectively.
Therefore, the optimum reaction conditions are M of 6:1, KOH C of 0.5 wt % and a T of 55 ◦ C and the
optimum SFO biodiesel yield is predicted to be 95.9%. To check the validity of the regression model
(Equation (9)), experiments were carried out under predicted optimum conditions. The results of the
experimental values under the optimum conditions indicated the highest (95.8%) conversion of SFO to
303
Energies 2018, 11, 2566
SFO biodiesel. This matches very closely with the predicted value (95.9%). Finally, this small degree of
error (<0.5%) indicates the high accuracy of the model.
5. Conclusions
A response surface methodology-based Box–Behnken design matrix was employed to achieve the
optimum operating parameters for second-generation biodiesel production from SFO. Three major
parameters were varied individually within different ranges to anticipate biodiesel yield in that matrix.
Based on the results, optimum operating parameters for transesterification of stone fruit seed oil were
found to be methanol: oil molar ratio of 6:1, catalyst concentration 0.5 wt %, and a reaction temperature
of 55 ◦ C, considering both reaction time and reaction agitation speed were fixed at 60 min and 600 rpm.
The maximum biodiesel yield under such conditions was 95.8%, which also confirmed the RSM model
prediction of 95.9%. ANOVA statistics of this study confirmed that methanol: oil molar ratio has the
most significant effect on the stone fruit biodiesel yield, whereas catalyst concentration and reaction
temperature does not seem to have any significant impact. The results show a significant improvement
in fuel properties of stone fruit biodiesel with kinetic viscosity 4.26 (mm2 /s), density 0.855 (kg/m3 ),
acid value 0.25 (mg/KOH/g), flash point 105 (◦ C), cloud point −4 (◦ C), pour point −8 (◦ C), higher
heating value 39.04 (MJ/kg), cetane number 50.45 and oxidation stability 7.15 (h), all of which meet
both the ASTM D6751 and EN14214 standards. In conclusion, stone fruit oil is a potential for biodiesel
production, and this environment-friendly biodiesel can be used as an alternative to diesel fuel.
Author Contributions: M.A. collected oil, produced and characterised biodiesel, optimised the biodiesel, and
drafted the manuscript; M.G.R. contributed to the experimental design and thoroughly revised the paper; N.A.
helped revise and improve the paper; and M.M.R. checked and revised the manuscript.
Funding: This research received no external funding.
Acknowledgments: The authors would like to acknowledge Tim Mcsweeney, Adjunct Research Fellow, Tertiary
Education Division at Central Queensland University, Australia for his contribution in proofreading of this article.
References
1. Rahman, M.M.; Rasul, M.G.; Hassan, N.M.S.; Azad, A.K.; Uddin, M.N. Effect of small proportion of butanol
additive on the performance, emission, and combustion of Australian native first-and second-generation
biodiesel in a diesel engine. Environ. Sci. Pollut. Res. 2017, 24, 22402–22413. [CrossRef] [PubMed]
2. Rahman, M.; Rasul, M.; Hassan, N. Study on the tribological characteristics of Australian native first
generation and second generation biodiesel fuel. Energies 2017, 10, 55. [CrossRef]
3. Bhuiya, M.M.K.; Rasul, M.G.; Khan, M.M.K.; Ashwath, N.; Azad, A.K.; Hazrat, M.A. Second generation
biodiesel: Potential alternative to-edible oil-derived biodiesel. Energy Procedia 2014, 61, 1969–1972. [CrossRef]
4. Bhuiya, M.M.K.; Rasul, M.G.; Khan, M.M.K.; Ashwath, N.; Azad, A.K.; Hazrat, M.A. Prospects of 2nd
generation biodiesel as a sustainable fuel—Part 2: Properties, performance and emission characteristics.
Renew. Sustain. Energy Rev. 2016, 55, 1129–1146.
5. Popp, J.; Lakner, Z.; Harangi-Rákos, M.; Fári, M. The effect of bioenergy expansion: Food, energy, and
environment. Renew. Sustain. Energy Rev. 2014, 32, 559–578. [CrossRef]
6. NguyenThi, T.; Bazile, J.-P.; Bessières, D. Density measurements of waste cooking oil biodiesel and diesel
blends over extended pressure and temperature ranges. Energies 2018, 11, 1212. [CrossRef]
7. Yusop, A.; Mamat, R.; Yusaf, T.; Najafi, G.; Yasin, M.; Khathri, A. Analysis of particulate matter (pm) emissions
in diesel engines using palm oil biodiesel blended with diesel fuel. Energies 2018, 11, 1039. [CrossRef]
8. Ali, O.; Mamat, R.; Najafi, G.; Yusaf, T.; Safieddin Ardebili, S. Optimization of biodiesel-diesel blended
fuel properties and engine performance with ether additive using statistical analysis and response surface
methods. Energies 2015, 8, 14136–14150. [CrossRef]
9. Silitonga, A.S.; Atabani, A.E.; Mahlia, T.M.I.; Masjuki, H.H.; Badruddin, I.A.; Mekhilef, S. A review on
prospect of jatropha curcas for biodiesel in Indonesia. Renew. Sustain. Energy Rev. 2011, 15, 3733–3756.
[CrossRef]
304
Energies 2018, 11, 2566
10. Rahman, M.; Rasul, M.; Hassan, N.; Hyde, J. Prospects of biodiesel production from macadamia oil as an
alternative fuel for diesel engines. Energies 2016, 9, 403.
11. Sharma, R.; Gupta, A.; Abrol, G.S.; Joshi, V.K. Value addition of wild apricot fruits grown in north-west
himalayan regions: A review. J. Food Sci. Technol. 2012, 51, 2917–2924. [CrossRef] [PubMed]
12. Razack, S.A.; Duraiarasan, S. Response surface methodology assisted biodiesel production from waste
cooking oil using encapsulated mixed enzyme. Waste Manag. 2016, 47, 98–104. [CrossRef] [PubMed]
13. Saydut, A.; Erdogan, S.; Kafadar, A.B.; Kaya, C.; Aydin, F.; Hamamci, C. Process optimization for production
of biodiesel from hazelnut oil, sunflower oil and their hybrid feedstock. Fuel 2016, 183, 512–517. [CrossRef]
14. Martínez, S.L.; Romero, R.; Natividad, R.; González, J. Optimization of biodiesel production from sunflower
oil by transesterification using Na2 O/NaX and methanol. Catal. Today 2014, 220, 12–20. [CrossRef]
15. Kostić, M.D.; Djalović, I.G.; Stamenković, O.S.; Mitrović, P.M.; Adamović, D.S.; Kulina, M.K.; Veljković, V.B.
Kinetic modeling and optimization of biodiesel production from white mustard (Sinapis alba L.) seed oil by
quicklime-catalyzed transesterification. Fuel 2018, 223, 125–139. [CrossRef]
16. Ong, H.C.; Masjuki, H.H.; Mahlia, T.M.I.; Silitonga, A.S.; Chong, W.T.; Leong, K.Y. Optimization of biodiesel
production and engine performance from high free fatty acid Calophyllum inophyllum oil in Ci diesel engine.
17. Dharma, S.; Masjuki, H.H.; Ong, H.C.; Sebayang, A.H.; Silitonga, A.S.; Kusumo, F.; Mahlia, T.M.I.
Optimization of biodiesel production process for mixed Jatropha curcas–Ceiba pentandra biodiesel using
response surface methodology. Energy Convers. Manag. 2016, 115, 178–190. [CrossRef]
18. Gumus, M.; Kasifoglu, S. Performance and emission evaluation of a compression ignition engine using a
biodiesel (apricot seed kernel oil methyl ester) and its blends with diesel fuel. Biomass Bioenergy 2010, 34,
19. Fadhil, A.B. Evaluation of apricot (Prunus armeniaca L.) seed kernel as a potential feedstock for the production
of liquid bio-fuels and activated carbons. Energy Convers. Manag. 2017, 133, 307–317. [CrossRef]
20. Ullah, F.; Nosheen, A.; Hussain, I.; Bano, A. Base catalyzed transesterification of wild apricot kernel oil for
biodiesel production. Afr. J. Biotechnol. 2009, 8, 3309–3313.
21. Yadav, A.K.; Pal, A.; Dubey, A.M. Experimental studies on utilization of Prunus armeniaca L. (wild apricot)
biodiesel as an alternative fuel for CI engine. Waste Biomass Valoriz. 2017, 1–9. [CrossRef]
22. Hamze, H.; Akia, M.; Yazdani, F. Optimization of biodiesel production from the waste cooking oil using
response surface methodology. Process Saf. Environ. Prot. 2015, 94, 1–10. [CrossRef]
23. Atabani, A.E.; Silitonga, A.S.; Badruddin, I.A.; Mahlia, T.M.I.; Masjuki, H.H.; Mekhilef, S. A comprehensive
review on biodiesel as an alternative energy resource and its characteristics. Renew. Sustain. Energy Rev. 2012,
16, 2070–2093. [CrossRef]
24. Atadashi, I.M.; Aroua, M.K.; Abdul Aziz, A.R.; Sulaiman, N.M.N. The effects of water on biodiesel production
and refining technologies: A review. Renew. Sustain. Energy Rev. 2012, 16, 3456–3470. [CrossRef]
25. Banerjee, A.; Chakraborty, R. Parametric sensitivity in transesterification of waste cooking oil for biodiesel
production—A review. Resour. Conserv. Recycl. 2009, 53, 490–497. [CrossRef]
26. Yaakob, Z.; Mohammad, M.; Alherbawi, M.; Alam, Z.; Sopian, K. Overview of the production of biodiesel
from waste cooking oil. Renew. Sustain. Energy Rev. 2013, 18, 184–193. [CrossRef]
27. Targais, K.; Stobdan, T.; Yadav, A.; Singh, S.B. Extraction of apricot kernel oil in cold desert Ladakh, India.
Indian J. Tradit. Knowl. 2011, 10, 304–306.
28. Wang, L. Properties of manchurian apricot (Prunus mandshurica Skv.) and siberian apricot (Prunus sibirica L.)
seed kernel oils and evaluation as biodiesel feedstocks. Ind. Crops Prod. 2013, 50, 838–843. [CrossRef]
29. Fan, S.; Liang, T.; Yu, H.; Bi, Q.; Li, G.; Wang, L. Kernel characteristics, oil contents, fatty acid compositions
and biodiesel properties in developing siberian apricot (Prunus sibirica L.) seeds. Ind. Crops Prod. 2016, 89,
30. Rahman, M.M.; Hassan, M.H.; Kalam, M.A.; Atabani, A.E.; Memon, L.A.; Rahman, S.M.A. Performance and
emission analysis of Jatropha curcas and Moringa oleifera methyl ester fuel blends in a multi-cylinder diesel
engine. J. Clean. Prod. 2014, 65, 304–310. [CrossRef]
31. Kafuku, G.; Mbarawa, M. Alkaline catalyzed biodiesel production from moringa oleifera oil with optimized
production parameters. Appl. Energy 2010, 87, 2561–2565. [CrossRef]
32. Alptekin, E.; Canakci, M. Determination of the density and the viscosities of biodiesel–diesel fuel blends.
305
Energies 2018, 11, 2566
33. Encinar, J.M.; González, J.F.; Rodríguez-Reinares, A. Biodiesel from used frying oil. Variables affecting the
yields and characteristics of the biodiesel. Ind. Eng. Chem. Res. 2005, 44, 5491–5499. [CrossRef]
34. Graboski, M.S.; McCormick, R.L. Combustion of fat and vegetable oil derived fuels in diesel engines.
Prog. Energy Combust. Sci. 1998, 24, 125–164. [CrossRef]
35. Silitonga, A.S.; Ong, H.C.; Mahlia, T.M.I.; Masjuki, H.H.; Chong, W.T. Biodiesel conversion from high FFA crude
Jatropha curcas, Calophyllum inophyllum and Ceiba pentandra oil. Energy Procedia 2014, 61, 480–483. [CrossRef]
36. Silitonga, A.S.; Ong, H.C.; Masjuki, H.H.; Mahlia, T.M.I.; Chong, W.T.; Yusaf, T.F. Production of biodiesel
from Sterculia foetida and its process optimization. Fuel 2013, 111, 478–484. [CrossRef]
37. Keera, S.T.; El Sabagh, S.M.; Taman, A.R. Transesterification of vegetable oil to biodiesel fuel using alkaline
catalyst. Fuel 2011, 90, 42–47. [CrossRef]
149, 392–400. [CrossRef]
39. Anwar, M.; Rasul, M.G.; Ashwath, N. Production optimization and quality assessment of papaya (Carica papaya)
biodiesel with response surface methodology. Energy Convers. Manag. 2018, 156, 103–112. [CrossRef]
40. Benjumea, P.; Agudelo, J.R.; Agudelo, A.F. Effect of the degree of unsaturation of biodiesel fuels on engine
performance, combustion characteristics, and emissions. Energy Fuels 2011, 25, 77–85. [CrossRef]
41. Altun, Ş. Effect of the degree of unsaturation of biodiesel fuels on the exhaust emissions of a diesel power
generator. Fuel 2014, 117, 450–457. [CrossRef]
42. Wang, L.; Yu, H. Biodiesel from siberian apricot (Prunus sibirica L.) seed kernel oil. Bioresour. Technol. 2012,
43. Kate, A.E.; Lohani, U.C.; Pandey, J.P.; Shahi, N.C.; Sarkar, A. Traditional and mechanical method of the oil
extraction from wild apricot kernel: A comparative study. Res. J. Chem. Environ. Sci. 2014, 2, 54–60.
44. Atabani, A.E.; César, A.D.S. Calophyllum inophyllum L.—A prospective non-edible biodiesel feedstock. Study
of biodiesel production, properties, fatty acid composition, blending and engine performance. Renew. Sustain.
45. García-Martínez, N.; Andreo-Martínez, P.; Quesada-Medina, J.; de los Ríos, A.P.; Chica, A.; Beneito-Ruiz, R.;
Carratalá-Abril, J. Optimization of non-catalytic transesterification of tobacco (Nicotiana tabacum) seed oil
using supercritical methanol to biodiesel production. Energy Convers. Manag. 2017, 131, 99–108. [CrossRef]
46. Goyal, P.; Sharma, M.P.; Jain, S. Optimization of conversion of high free fatty acid Jatropha curcas oil to
biodiesel using response surface methodology. ISRN Chem. Eng. 2012, 2012, 8. [CrossRef]
47. Lin, L.; Zhou, C.; Vittayapadung, S.; Shen, X.; Dong, M. Opportunities and challenges for biodiesel fuel.
48. Demirbaş, A. Biodiesel fuels from vegetable oils via catalytic and non-catalytic supercritical alcohol
transesterifications and other methods: A survey. Energy Convers. Manag. 2003, 44, 2093–2109. [CrossRef]
49. Shankar, V.; Jambulingam, R. Waste crab shell derived CaO impregnated Na-ZSM-5 as a solid base catalyst
for the transesterification of neem oil into biodiesel. Sustain. Environ. Res. 2017, 27, 273–278. [CrossRef]
50. Patel, R.L.; Sankhavara, C.D. Biodiesel production from karanja oil and its use in diesel engine: A review.
51. Rashid, U.; Anwar, F.; Moser, B.R.; Knothe, G. Moringa oleifera oil: A possible source of biodiesel.
52. Silitonga, A.S.; Masjuki, H.H.; Mahlia, T.M.I.; Ong, H.C.; Chong, W.T.; Boosroh, M.H. Overview properties of
biodiesel diesel blends from edible and non-edible feedstock. Renew. Sustain. Energy Rev. 2013, 22, 346–360.
[CrossRef]
53. Gupta, J.; Agarwal, M.; Dalai, A.K. Optimization of biodiesel production from mixture of edible and
nonedible vegetable oils. Biocatal. Agric. Biotechnol. 2016, 8, 112–120. [CrossRef]
306
Energies 2018, 11, 2566
54. Silva, G.F.; Camargo, F.L.; Ferreira, A.L.O. Application of response surface methodology for optimization
of biodiesel production by transesterification of soybean oil with ethanol. Fuel Process. Technol. 2011, 92,
55. Vicente, G.; Martínez, M.; Aracil, J. Optimisation of integrated biodiesel production. Part I. A study of the
biodiesel purity and yield. Bioresour. Technol. 2007, 98, 1724–1733. [CrossRef] [PubMed]
307
energies
Article
Effects of Organosolv Pretreatment Using
Temperature-Controlled Bench-Scale Ball Milling on
Enzymatic Saccharification of Miscanthus × giganteus
Tae Hoon Kim 1,2 , Dongjoong Im 2 , Kyeong Keun Oh 2,3, * and Tae Hyun Kim 1, *
1 Department of Materials Science and Chemical Engineering, Hanyang University, Ansan,
Gyeonggi-do 15588, Korea; [email protected]
2 R&D Center, SugarEn Co., Ltd., Yongin, Gyeonggi-do 16890, Korea; [email protected]
3 Department of Chemical Engineering, Dankook University, Yongin, Gyeonggi-do 16890, Korea
* Correspondence: [email protected] (T.H.K.); [email protected] (K.K.O);
Tel.: +82-31-400-5222 (T.H.K.); +82-31-330-5222 (K.K.O)
Received: 16 August 2018; Accepted: 1 October 2018; Published: 5 October 2018
Abstract: The effect of organosolv pretreatment was investigated using a 30 L bench-scale ball
mill reactor that was capable of simultaneously performing physical and chemical pretreatment.
Various reaction conditions were tried in order to discover the optimal conditions for the minimal
cellulose loss and enhanced enzymatic digestibility of Miscanthus × giganteus (MG), with conditions
varying from room temperature to 170 ◦ C for reaction temperature, from 30 to 120 min of reaction time,
from 30% to 60% ethanol concentration, and a liquid/solid ratio (L/S) of 10–20 under non-catalyst
conditions. The pretreatment effects were evaluated by chemical compositional analysis, enzymatic
digestibility test and X-ray diffraction of the treated samples. The pretreatment conditions for
the highest glucan digestibility yield were determined as 170 ◦ C, reaction time of 90 min, ethanol
concentration of 40% and L/S = 10. With these pretreatment conditions, the XMG (xylan + mannan
+ galactan) fractionation yield and delignification were 84.4% and 53.2%, respectively. The glucan
digestibility of treated MG after the aforementioned pretreatment conditions was 86.0% with 15 filter
paper units (FPU) of cellulase (Cellic® CTec2) per g-glucan enzyme loading.
Keywords: lignocellulosic biomass; Miscanthus; mechanical pretreatment; organosolv pretreatment
1. Introduction
Lignocellulosic biomass is considered to be an abundant and attractive resource in the biorefinery
system. However, lignocellulosic biomass is a complex organic polymer, which is composed mainly of
carbohydrate polymers (cellulose and hemicellulose) and lignin. In addition, chemical and physical
barriers such as lignin and the crystalline structure of cellulose inhibit the hydrolytic reaction of
enzymes on the cellulose substrate. Thus, pretreatment is an indispensable process to enhance the
enzymatic saccharification of cellulose, and various pretreatment methods have been attempted
extensively. Pretreatment methods can be divided into two different categories: physical and chemical
pretreatment methods and biological pretreatment methods. The purpose of pretreatment is to break
down the structure of the lignocellulosic biomass to improve enzymatic digestibility and thus to enable
the efficient bioconversion of monomeric sugars to biochemicals [1–3].
Chemical pretreatment generally uses various catalysts such as acids, alkalis, ionic liquids,
oxidizing agents, or organic solvents. Among them, acid and alkali catalysts effectively affect the
biomass by either reducing the degree of polymerization or removing lignin [3,4]. Despite its many
advantages, chemical pretreatment has encountered major problems such as equipment corrosion,
long reaction times, high input of solvents, and sugar degradation due to the high reactivity and sever

Energies 2018, 11, 2657
reaction conditions [5,6]. In addition, the neutralization and washing of the treated solid is essential to
remove the residual chemicals; therefore, a large amount of water is required. As one type of chemical
pretreatment, the organosolv pretreatment method uses a variety of organic or aqueous solvents
under catalytic or non-catalytic conditions. The solvents being used in the organosolv pretreatment
include ethanol, methanol, acetone, formic acid, acetic acid, and glycerol [7,8]. In addition, acid
catalysts such as sulfuric acid and hydrochloric acid are generally applied to lower the pH, because
the carbohydrate–lignin complex efficiently cleaved at low pH. The organosolv pretreatment can
effectively cleave the lignin-to-lignin and lignin-to-hemicellulose bonds, which can recover high purity
lignin [8,9]. As a result, the organosolv-pretreated biomass with low hemicellulose and lignin contents
can be expected to have increased digestibility. However, organosolv pretreatment usually requires
a high reaction temperature range (170–210 ◦ C), long reaction time (30–120 min), and high chemical
loading (50–75% in the case of ethanol and 0.5–2.0 wt.% of acid catalyst (such as H2 SO4 or HCl)) [10–13].
However, in case of non-acid catalytic conditions (organosolv-only without additional acid), it typically
requires more severe reaction condition. For example, a study of J Wildschut [14] (using a wheat straw)
showed very low enzymatic digestibility of 31.7% at 170 ◦ C, 60 min, and 60% ethanol concentration.
When the reaction temperature was increased by 30 ◦ C (200 ◦ C, 60 min, 60% ethanol concentration),
the digestibility was only 44.4%. On the other hand, when the most severe reaction conditions (210 ◦ C,
90 min, 50% ethanol concentration) were applied, the enzymatic digestibility increased to 85.9%.
Meanwhile, mechanical pretreatment methods commonly include grinding, milling,
and comminution. They usually reduce the particle size and increase the surface area, which improves
the contact surface of the biomass exposed to the enzyme [14]. Among the mechanical pretreatment
methods, ball milling pretreatment severely reduces the crystallinity of cellulose by breaking hydrogen
bonds in the substrate [15]. It increases the amorphous portion of cellulose, which improves the
accessibility of the substrate to the enzyme [16]. In some studies on the tendency of biomass size
reduction, the negative effect of milling on enzymatic digestibility was reported that the grinded
MG (using hammer mill to 80 μm) resulted in only 22.5% digestibility [17]. In addition, ball mill
pretreatment has previously received less attention because of its high energy consumption and long
reaction time. However, integrating the mechanical and chemical pretreatment has the potential to
overcome the shortcomings of chemical pretreatment method [18–20].
The main purpose of this study was to improve the pretreatment effects for Miscanthus × giganteus
(MG) by combining mechanical and chemical pretreatment methods. The various reaction conditions
(reaction temperature, time, and ethanol concentration) were evaluated to reduce the reaction time
and decrease the reaction temperature of the organosolv pretreatment. Ethanol was used because
ethanol is not only cheaper than other organic solvents, but it also has a low boiling point, which
makes it easy to recover [21,22]. In general, organosolv pretreatment is supplemented with organic
or inorganic acid to control the pH level for improved hydrolytic reaction, but this study was carried
out under non-catalytic conditions to avoid the problems with an acid catalyst, such as equipment
corrosion and the formation of toxic substances such as levulinic acid, 5-(hydroxymethyl)furfural
(5-HMF) and furfural [1]. To the best of our knowledge, there have been few studies on the combined
pretreatment with organosolv and ball milling in a large-scale reactor system. According to a study
by Y. Teramoto [23] (ball milling and organosolv treatment were performed separately), enzymatic
digestibility was improved to 67.4% under the pretreatment conditions of 200 ◦ C, 60 min, and 75%
ethanol concentration.
In this study, one of the representative mechanical pretreatments, the ball mill pretreatment
with alumina balls, was applied. A 30-L reactor was custom designed to demonstrate the combined
effects of chemical and physical pretreatment. In addition, this large-scale ball mill can be operated
at a maximum temperature of 200 ◦ C and pressure of 20 kgf/cm2 . Prior to testing the 30-L ball mill,
small lab scale tests were performed.
309
Energies 2018, 11, 2657
2.1. Effects of Ethanol Concentration on Compositions of Pretreated Solids and Liquids Using Small Scale
Batch Reactor
In a previous study of organosolv pretreatment, high ethanol concentration and loading were
applied to achieve effective pretreatment effects; therefore, this method was expected to increase the
overall processing cost. In general, the use of a lower ethanol concentration to reduce the processing
cost of the organosolv pretreatment process. A preliminary experiment was conducted to determine
the adequate ethanol concentration, and the chemical compositions of the organosolv-treated MG
are shown in Figure 1. The reaction temperature and reaction time were fixed at 170 ◦ C for 60 min,
which were pre-selected based on the previous studies [10–13].
Figure 1. Effect of ethanol concentration on (a) solid content of untreated and pretreated
Miscanthus × giganteus (MG); (b) monomeric sugar and byproduct concentrations of liquid hydrolysate.
Reaction conditions: 170 ◦ C, 60 min, liquid/solid (L/S) ratio = 10. Data in the figure are based on
untreated biomass (wt.%).
As shown in Figure 1a, solubilizations of glucan (10.2–13.2%) were generally lowers than those
(10.5–29.8%) of XMG (sum of xylan, mannan, and galactan). In addition, the glucan and XMG contents
of the treated solids were more preserved as the ethanol concentration increased from 30% to 60%.
The reason for this phenomenon was because of the different hydrogen ion concentrations in water
produced at high temperatures.
A low ethanol concentration can produce more hydronium ion (H3 O+ ) than a high ethanol
concentration does due to the greater water content [24]. Therefore, the pH value of the low ethanol
concentration becomes lower than that of the high ethanol concentration, and this promotes the
depolymerization of hemicellulose [25]. In contrast, lignin content showed no important change
as ethanol concentrations rose from 30% to 60%. The hemicellulose is bound with lignin in the
cell wall of the biomass, which is more susceptible to hydrolysis under severe conditions [26,27].
Concentrated ethanol is known to be effective for lignin solubilization. However, reaction under harsh
conditions is generally necessary for the effective cleavage of the lignin–hemicellulose complex,
which was assumed to be related to pH. The extent of delignification with increased ethanol
concentration was shown to be consistent at the tested ethanol concentration (Figure 1a). The tendency
of sugars in the residual solid upon various ethanol concentrations was similar to that of the solubilized
sugar in liquid hydrolyzate shown in Figure 1b. As shown in Figure 1b, the formations of XMG and
byproducts (formic acid, acetic acid, levulinic acid, 5-HMF, and furfural) in liquid hydrolyzate were
decreased with increased ethanol concentration. However, the amount of glucan decomposition was
310
Energies 2018, 11, 2657
not affected by ethanol concentration. At 30% ethanol concentration, 21.1% of XMG was solubilized.
Under the same conditions, byproduct was formed including 0.4 g/L of formic acid, 0.8 g/L of acetic
acid, and 0.4 g/L of furfural (levulinic acid and 5-HMF were < 0.1 g/L). At 60% ethanol concentration,
11.9% of XMG was solubilized. Under these conditions, byproduct was formed and contained 0.3 g/L
of formic acid, 0.4 g/L of acetic acid and 0.2 g/L of furfural (c.f. levulinic acid and 5-HMF were
< 0.1 g/L). In general, organosolv pretreatment has more advantage for delignification with high
ethanol concentration. Based on these results in Figure 1b, the ethanol concentration for maximum
XMG solubilization was determined to be 40%, which can achieve maximum XMG solubilization and
possibly more delignification.
2.2. Organosolv Pretreatment Using Bench-Scale Ball-Milling Reactor
2.2.1. Effects of Reaction Temperature and Time

The various effects of organosolv pretreatment were evaluated using a bench scale ball-milling
reactor. Table 1 presents the effects of organosolv pretreatment on the chemical compositions of the
treated solids and liquid hydrolyzate. As the reaction temperature increased from room temperature
(RT) to 170 ◦ C, the XMG removal and lignin removal (delignification) were highly increased for all
pretreatment times. However, the changes in XMG solubilization from the treated solids at RT and
130 ◦ C were inconsiderable at all reaction times.
Table 1. Effects of reaction temperature and time on chemical compositions of treated solids and
liquid hydrolyzate.
Conditions Treated Solids Liquid Hydrolyzate Total Cons. 1
Temp. Time Glucan XMG 2 Lignin 3 Glucan XMG Glucan XMG

[◦ C] [min] [wt.%] [wt.%] [wt.%] [wt.%] [wt.%] [wt.%] [wt.%]
Untreated 43.8 21.2 20.5 - - - -
30 43.5 20.9 20.2 0.4 0.5 99.7 99.1
RT 4 60 43.3 21.0 20.0 0.4 0.7 99.5 99.8
120 43.0 20.7 20.3 0.6 0.8 98.9 98.5
30 41.7 20.1 18.3 0.7 2.1 96.0 97.0
130 60 42.3 20.2 18.5 0.9 2.6 97.5 97.7
120 42.7 20.0 16.7 0.8 3.4 98.5 97.7
30 42.1 19.0 15.4 1.1 8.4 97.3 98.0
150 60 41.5 17.6 14.1 1.4 15.4 96.3 98.4
120 40.7 14.4 12.3 2.0 30.0 95.1 97.8
30 41.2 10.1 10.6 1.9 40.9 96.0 88.3
170 60 40.8 4.2 10.0 1.9 65.3 95.0 85.3
120 35.4 2.0 7.8 4.1 74.1 85.0 83.6
1Total conservation: [(remaining (wt.%) in treated solid + fractionated (wt.%) in liquid)/untreated (wt.%) in
MG]×100; 2 XMG: xylan + mannan + galactan; 3 Lignin: acid-insoluble lignin; 4 room temperature (~25 ◦ C).
Reaction conditions: 40% ethanol concentration and L/S ratio = 10.
At 150 ◦ C for 120 min, XMG removal was 32.1%, whereas the delignification was 40%. On the other
hand, with the 170 ◦ C treatment, the XMG removal and delignification differed considerably between
the 30 min and 120 min treatments. At 170 ◦ C for 120 min, the XMG removal and delignification yields
were 90.6% and 62.0%, respectively, which were considerably increased from the 52.4% XMG removal
and 48.3% delignification of the 30 min treatment. The glucan contents were preserved well at all
treatment conditions (>80.8%).
The composition of liquid hydrolyzates had similar trends. As the reaction conditions became
more severe, the XMG solubilization yields in the liquid hydrolyzate were proportionally increased.
The solubilization yield of XMG was less than 30.0% at RT, 130 ◦ C, and 150 ◦ C, while substantial
311
Energies 2018, 11, 2657
changes were observed at 170 ◦ C treatment. The maximum recovery yield of XMG was 74.1% at 170 ◦ C
for 120 min.
The glucan and XMG contents were decreased as the reaction temperature and time became more
severe. Under the most severe conditions of 170 ◦ C and 120 min, total conservation of glucan and
XMG were 85.0% and 83.6%, respectively. This suggested that appropriate reaction conditions must be
selected to accomplish the effective fractionation of sugars and lignin.
Figure 2 summarizes the byproduct formations (mainly formic acid, acetic acid, and furfural)
at two different reaction temperatures (150 ◦ C and 170 ◦ C). The byproduct formations (formic acid,
acetic acid, levulinic acid, 5-HMF, and furfural) at RT and 130 ◦ C were not substantial, which were
less than 0.3 g/L (data not shown). Byproducts were produced by the decomposition of cellulose
and hemicellulose, which typically occurred at the more severe reaction conditions [28]. Therefore,
this is a good indicator to determine the severity of the reaction. At the most severe conditions
(i.e., 170 ◦ C for 120 min), the productions of formic acid, acetic acid, and furfural were 1.2 g/L, 3.5 g/L
and 2.0 g/L, respectively.
Figure 2. Effects of reaction temperature and time on concentrations of byproducts in liquid

hydrolyzate. Reaction conditions: 40% ethanol concentration and 10 L/S.
2.2.2. Effects of Liquid/Solid (L/S) Ratio

The effects of ethanol concentrations (40% and 60%), L/S ratio (10 and 20), and reaction times
(from 30 to 120 min) on the chemical compositions of treated solids at 170 ◦ C are summarized in
Table 2. The results indicate that the run No. 1 (ethanol concentration of 40% and L/S ratio of 10)
conditions seem to be more severe than those of run No. 2 (ethanol concentration of 40% and L/S
ratio of 20) and run No. 3 (ethanol concentration of 60% and L/S ratio of 10), because the retention
of XMG in treated solid in run No. 1 was lower than those of the other runs. For the decomposition
of sugar, in particular, XMG was depending on the pretreatment reaction severity, and the observed
results in Table 2 were assumed to be related to the generation of acid products such as formic acid
and acetic acid. Because the acetyl group of hemicellulose is cleaved as the reaction progresses, acetic
acid is generated, which would drop the pH of the solution during the reaction and consequently
increase the reaction severity, therefore promoting the hydrolysis of hemicellulose. Regardless of
the amount generated, the acid concentration at an L/S of 10 is expected to be higher than that of
20. After treatment, the pH values of the liquid hydrolyzate of runs No. 1 and 2 at a reaction time
of 120 min were 3.32 and 4.17, respectively. This means that L/S = 10 is a more severe condition
than L/S = 20, and this can further promote the hydrolysis of MG during the reaction. On the other
hand, the high ethanol concentration (run No. 3) resulted in the formation of a low concentration of
hydronium ion; therefore, the hydrolysis effect was not improved. In addition, the delignifications
312
Energies 2018, 11, 2657
of runs No. 1, 2, and 3 were 58.0%, 61.0%, and 62.0%, respectively. This indicated that a 40% ethanol
concentration could effectively remove the lignin (~60%).
Table 2. Chemical compositions of solids and liquid hydrolyzates and total conservation yield of ball
mill organosolv pretreatment under various reaction conditions.
Conditions Treated Solid Liquid Hydrolyzate Total Conservation

No. EtOH L/S Time Glucan XMG Lignin Glucan XMG Glucan XMG
[%] [-] [min] [%] [%] [%] [%] [%] [%] [%]
Untreated 43.8 21.2 20.5 - - - -
30 41.2 10.1 10.6 1.9 40.9 96 88.3
60 40.8 4.2 10 1.9 65.3 95 85.3
1 40 10
90 37.5 3.3 9.6 1.8 68.6 87.6 84
120 35.4 2 7.8 4.1 74.1 85 83.6
30 39.6 12.8 11.8 1.7 35.5 92.3 95.8
60 38.8 8.8 9.1 2.3 56.5 91 98
2 40 20
90 37.8 6.3 8.5 2.2 57.4 88.5 87.1
120 36.6 5.3 8 2.4 59.2 86.1 84.1
30 42.3 18.3 12.9 1.8 9.8 98.5 95.8
60 39.5 15.7 11.5 1.9 16.8 92.2 91
3 60 10
90 38.8 13.6 9.4 2 24.5 90.6 88.4
120 37.5 8.7 8.6 2.1 45.3 87.8 86.3
Total conservation: [(remaining (wt.%) in treated solid + fractionated (wt.%) in liquid)/untreated (wt.%) in MG]×100;
EtOH, ethanol; XMG: xylan + mannan + galactan; Lignin: acid-insoluble lignin. Reaction temperature: 170 ◦ C.
2.2.3. Enzymatic Digestibility Testing of Pretreated Solids

The enzymatic digestibility tests were performed on untreated and treated MG, and the glucan
digestibilities are summarized in Figure 3. Figure 3a compares the digestibility of untreated MG
and treated MG under various reaction conditions with the fixed L/S ratio of 10, because L/S = 10
required less solvent than L/S = 20 and effective hydrolysis could be expected. As shown in Figure 3a,
the digestibility at three treatment temperatures, RT, 130 ◦ C and 150 ◦ C, did not exceed 40% for 72 h
of hydrolysis. The digestibility of the solids treated at 150 ◦ C for 120 min was only 39.4% at 72 h of
hydrolysis. However, the enzyme digestibility of the sample treated at 170 ◦ C was increased to the
extent to which it was similar to that using α-cellulose. The 72-h digestibility values of the samples
treated for 60 and 120 min were 75.7% and 96.9%, respectively. In particular, the 72-h digestibility of
the sample treated at 170 ◦ C for 120 min was 8.9% higher than that 88.0% of α-cellulose. The 72-h
digestibility values of untreated MG and α-cellulose as the reference and control were 6.5% and
88.0%, respectively.
The effects of ethanol concentration and L/S ratio on digestibility at the 90 min reaction time
are presented in Figure 3b. The 72-h digestibility of treated MG with 60% ethanol was considerably
lower than that of the treated solids with 40% ethanol. Although the digestibility of the two treatment
conditions using 40% ethanol were close to 80% in 24 h, the hydrolysis rate was faster than that
of α-cellulose (70.2%) with 15 filter paper units (FPU)/g-glucan enzyme loading. Considering the
aforementioned results (Table 2 and Figure 3), the following conditions were selected for the effective
pretreatment and fractionation of MG: 170 ◦ C reaction temperature, 90 min reaction time, 40% ethanol
concentration and 10 L/S ratio. The 72-h digestibility of the organosolv-treated sample (L/S of 10 with
40% ethanol concentration) was 86.0%. In addition, during pretreatment reaction at 170 ◦ C (Table 2 and
Figure 3b), it was observed that lignin removals were not substantial (<58.5% of delignification) at all
tested conditions. It indicated that lignin could melt or dissolve and re-precipitate back on the solids,
which was to suggest that removal of lignin to the liquid phase was only a partial measure of what is
happening with the lignin fraction.
313
Energies 2018, 11, 2657
Figure 3. Enzymatic digestibility profiles of treated and untreated Miscanthus × giganteous at: (a) 40%
ethanol concentration and L/S = 10 and (b) 170 ◦ C for 90 min. RT (room temperature); EtOH (ethanol);
L/S (liquid/solid); enzymatic hydrolysis conditions: 15 FPU of Cellic® CTec2/g-glucan, pH 4.8, 50 ◦ C,
and 150 rpm.
2.3. Crystallinity Index (CrI) of Untreated and Treated MG

The X-ray diffraction (XRD) patterns and CrI of untreated and treated MG at various reaction
temperatures (RT, 130 ◦ C, 150 ◦ C, and 170 ◦ C) are presented in Figure 4. Several studies have shown
that as the reaction severity increases, the CrI of treated solid increases due to the removal of the
amorphous parts (xylan and lignin) [29]. However, as shown in Figure 4, the crystallinities of the
samples treated at RT (59.1) and 130 ◦ C (62.9) were lower than that of the untreated sample (64.1).
We speculated that the decrease in CrI for the samples treated at RT and 130 ◦ C was because of the
severe size reduction effect due to the ball milling pretreatment. In contrast, for the samples treated
at 150 and 170 ◦ C, the CrIs were 67.3 and 67.0, respectively. As the reaction temperature increased
to >150 ◦ C, XMG and lignin were removed and the CrI increased to more than that of the untreated
sample, which may be due to the more removal of the amorphous portions.
Figure 4. The diffraction patterns and CrIs (crystallinity indices) of untreated and treated
Miscanthus × giganteous at different reaction temperatures. Reaction conditions: 60 min reaction time,
40% ethanol concentration and L/S = 10.
314
Energies 2018, 11, 2657
3.1. Materials
The air-dried MG, harvested in Wanju, Jeollabuk-do, Korea in 2016, was supplied by the
National Institute of Crop Science (Wanju, Jeollabuk-do, Korea). The MG was ground (Blender 7012s,
Warning Commercial Inc., Stamford, CT, USA) and screened to a nominal size of 14–45 mesh
(from 0.36 to 1.4 mm). The average moisture content of the MG was determined to be 5.6%
(by oven-drying at 105 ◦ C for 24 h). The chemical composition of the untreated MG was determined
using the laboratory analytical procedure (LAP) developed by the National Renewable Energy
Laboratory (NREL, 2008).
The initial composition of the MG was: 43.8 wt.% glucan, 21.2 wt.% XMG, 0.5 wt.% arabinan,
3.9 wt.% acetyl group, 1.0 wt.% acid-soluble lignin, 19.6 wt.% acid-insoluble lignin, 1.5 wt.% organosolv
extractives, 5.0 wt.% water-soluble extractives, and 2.2 wt.% ash (n = 3, standard deviations < 0.2).
The mass closure of the untreated MG reached 98.5 wt.% (oven-dried biomass weight). The α-cellulose
(catalog number C8002) and ethanol (95.0%, denatured, catalog number E0219) were purchased
from Sigma-Aldrich Korea (Yongin, Gyeonggi-do, Korea) and Samchun Pure Chemical Co., Ltd.
(Pyeongtaek, Gyeonggi-do, Korea), respectively.
3.2. Experimental Setup and Operation
3.2.1. Small Batch Organosolv Pretreatment Reactor

The small batch reactor (Figures 5 and A1) consisted of a control box, a reaction bath (molten salt
bath and silicon oil bath), and a cooling bath (water bath). The bomb tubular reactor was made
of SS-316L (internal diameter of 1.0 cm, length of 20.0 cm, and internal volume of 15.7 cm3 ).
The temperature of the reaction bath was measured using a high temperature junction-type
thermocouple (catalog number HY-72D, HANYOUNG NUX, Inchoen, Gyeonggi-do, Korea).
The control box had a timer, a movement controller, and a digital temperature controller.
Figure 5. Schematic diagram of the small batch reactor setup: (1) timer and counter; (2) control box;
(3) voltmeter; (4) ammeter; (5) thermo-controller; (6) temperature indicator; (7) salt bath; (8) silicon oil
bath; (9) cooling bath; (10) batch reactor; (11) electric motor; (12) main switch.
First, 0.6 g (oven dry weight) MG and 6.0 mL ethanol (30%, 40%, 50% or 60%, w/v) were packed
into the tubular-shaped batch reactor. The reactor was placed in a molten salt bath to preheat it to
the target reaction temperature (170 ◦ C). To shorten the preheating time, the molten salt bath was
set to 50 ◦ C higher than the target temperature. The preheating time was less than 1 min. When the
315
Energies 2018, 11, 2657
target temperature was reached, the reactor was transferred to the silicon oil bath set at the target
temperature (170 ◦ C) targeted time (60 min). At the end of the reaction time, it was transferred to
a water bath and rapidly cooled. After cooling, liquid samples were removed from the reactor then
then diluted 3-fold with DI (deionized) water—it facilitated the lignin precipitation. This liquid sample
was subjected to the evaporation (at 55 ◦ C for 4 h), then acid hydrolysis was carried out to analyze the
liquid sample by adding 72% sulfuric acid to a final sulfuric acid concentration of 4.0%. The solids
removed from the reactor were washed with ethanol (30%, 40%, 50% or 60%, w/v) and water, and were
then dried in a 45 ◦ C convection oven. The dried solids were measured for weight loss and subjected
to composition analysis. Each experiment was performed in triplicate.
3.2.2. Bench Scale Ball-Milling Pretreatment Reactor

A schematic diagram of the bench scale ball-milling reactor is shown in Figures 6 and A2. The main
reactor was constructed using SS-316L (internal diameter of 30.0 cm, depth of 42.0 cm and internal
volume of 30 L) with a wall thickness of 10.0 mm. The pressure, temperature, and rpm could be driven
up to 20 kgf/cm2 , 200 ◦ C, and 60 rpm, respectively. A metal spiral gasket was installed between the
head and reactor at each of the two sides (front head and back head) to prevent pressure leakage.
The front head was equipped with a hand hole 7.5 cm in diameter to feed in the reactant (biomass balls
and ethanol) and to discharge the product. The reactor heating system was controlled of a mantle type
electrical heater and capacity of heater was 10 kW. The motor (380 V, 3 Phase and 0.5 Hp) was installed
to rotating for reactor and slip ring was installed with a rotating connector that could supply power
or signals to the rotating equipment without twisting the line. The reactor was mounted on a frame
with a turning roller to assist rotation and was equipped with a tilting system. The reactor was based
on ball milling, but it was also designed to enable its operation under high temperature and high
pressure. The alumina ball (HD, sphere type, 10 mm diameter and 3.6 g/cm3 density) was purchased
from NIKKATO Co., (Osaka, Japan). The alumina balls, MG, and ethanol (40% or 60%, w/v) were
placed in the bench scale ball milling pretreatment reactor. The ratio of ball/MG/solvent was 30:1:10
(w/w/v). The reactor was preheated for 1 h until it reached the target temperature (130 ◦ C, 150 ◦ C or
170 ◦ C), and the reactor was rotated during preheating. The pretreatment at RT was not preheated.
When the target temperature was reached, the reaction was continued for a designated time (30, 60,
90 or 120 min). At the completion of the reaction, liquid samples were removed from the reactor then
subjected to the evaporation (at 55 ◦ C for 3 h), which was then diluted 3-fold with DI-water, which
facilitated the lignin precipitation. This liquid sample was subjected to the evaporation (at 55 ◦ C for
4 h), then acid hydrolysis was carried out to analyze the sugars and lignin in the liquid sample by
adding 72% sulfuric acid to a final sulfuric acid concentration of 4.0%. The solids removed from the
reactor were washed with ethanol (40% or 60%, w/v) and water and separated into two portions.
One portion was dried in a 45 ◦ C convection oven to determine weight loss (remaining solids) and
subjected to composition analysis. The other portion was washed with DI-water and stored in a tightly
sealed container until it underwent an enzymatic digestibility test. Each experiment was performed
in duplicate.
316
Energies 2018, 11, 2657
Figure 6. Schematic diagram of bench scale ball milling pretreatment reactor setup. PG (pressure gauge);
TG (temperature gauge); DI (deionized) water.
3.3. Enzymatic Digestibility Test

The enzymatic digestibility of untreated and treated MG was determined according to the NREL
Technical Report 510-42629 [30]. The tests were performed under the following conditions: 50 ◦ C,
150 rpm, and pH 4.8 in a shaking incubator (BioFree, model BF-175SI, Co., Ltd., Dongdaemun-gu,
Seoul, Korea). Enzyme loadings were 15 FPU/g-glucan of commercial cellulase enzyme Cellic CTec2
(Novozymes, A/S Bagsvaerd, Denmark). The initial glucan concentration was 1.0% (w/v) based on
100 mL (50 mM sodium citrate buffer) of total liquid in 250-mL Erlenmeyer flasks. To prevent microbial
contamination, 1.0 mL of 20 mg/mL sodium azide was added. Samples were taken at the appropriate
sampling times (6, 12, 24, 48 and 72 h) and analyzed for glucose content using high performance
liquid chromatography (HPLC) with an HPX-87H column (cat. No. 125-0098, Bio-Rad Laboratories
Inc., Hercules, CA, USA). The total glucose released after 72 h of hydrolysis was used to calculate the
enzymatic digestibility for glucan as follows:

Total released glucose (g) × 0.9
Glucan digestibility (%) = × 100 (1)
Initial glucan loading (g)
where 0.9 is the conversion factor of glucose to the equivalent in glucan. Untreated MG and α-cellulose
were put through the same procedure as the reference and the control, respectively.

The extractives, chemical compositions, and liquid fractions of the solid and liquid samples
were determined by following the procedures of NREL-LAP [31]. Glucose, XMG and byproducts
(formic acid, acetic acid, levulinic acid, 5-HMF, and furfural) were analyzed by HPLC. The HPLC
system (1260 Infinity, Agilent Technologies Inc., Santa Clara, CA, USA) was used for the measurement
of monomeric sugars and byproducts. The analytical column and detector used were an Aminex
HPX-87H organic acid column (Bio-Rad) and a refractive index detector (1260 RID, Agilent
Technologies Inc.), respectively. The mobile phase was 5 mM sulfuric acid in deionized water filtered
through 0.2-μm filters and degassed by sonication. The operating conditions for the HPLC column
were 65 ◦ C with a mobile phase with a volumetric flow rate of 0.5 mL/min. External standards
consisting of glucose, XMG, arabinose, formic acid, acetic acid, levulinic acid, 5-HMF, and furfural
were used for quantification.
317
Energies 2018, 11, 2657
3.5. Crystallinity Measurement

Crystallinities of the untreated and treated MG were determined by XRD (Rigaku Co.,
Tokyo, Japan) operated at 40 kV and 40 mA. The samples were scanned at 4◦ /min (2θ = 10–35◦ , 0.02
increment). The crystallinity indices of the samples were calculated using the following equation [32].

I002 − I18
CrI = × 100 (2)
I002
where I002 is the peak intensity corresponding to the 002 lattice plane of the cellulose molecule observed
at 2θ equal to 22.5◦ , and I18 (at 2θ = 18◦ ) is the peak intensity corresponding to amorphous cellulose.
4. Conclusions
In this study, we demonstrated that the combined mechanical and chemical pretreatment
effectively reduced the recalcitrance of the MG. We also confirmed that the integration of mechanical
and chemical pretreatment can be implemented in a large-scale reactor system. Reduced XMG, lignin,
and cellulose crystallinity with minimal sugar loss could improve the enzymatic digestibility of
the treated solids. For the improved enzymatic digestibility of glucan using commercial cellulase,
pretreatment conditions were determined (170 ◦ C reaction temperature, 90 min reaction time,
40% ethanol concentration, and 10 L/S ratio).
Author Contributions: T.H.K. as the first author conducted all experiments, summarized the data, and drafted
the manuscript. D.J.I. as the co-author conducted a part of experiments and analyzed the data. T.H.K. and K.K.O.
designed the reactor system as well as the overall study and experiments, interpreted the results, wrote the
manuscript, and finalized the manuscript. T.H.K. and K.K.O., as the co-corresponding authors, contributed
equally. All authors read and approved the final manuscript.
Funding: This work was supported by the New & Renewable Energy Core Technology Program of the Korea
Institute of Energy Technology Evaluation and Planning, and was granted financial resources from the Ministry of
Trade, Industry & Energy of the Republic of Korea (No. 20153030091050).
Appendix
Figure A1. Small batch reactor setup for organosolv only pretreatment.
318
Energies 2018, 11, 2657
Figure A2. Bench scale ball milling pretreatment reactor setup for ball-milling organosolv pretreatment.
References
1. Chen, H.; Liu, J.; Chang, X.; Chen, D.; Xue, Y.; Liu, P.; Lin, H.; Han, S. A review on the pretreatment of
lignocellulose for high-value chemicals. Fuel Process. Technol. 2017, 160, 196–206. [CrossRef]
2. Kim, T.H.; Kim, T.H. Overview of technical barriers and implementation of cellulosic ethanol in the US.
Energy 2014, 66, 13–19. [CrossRef]
3. Zabed, H.; Sahu, J.N.; Boyce, A.N.; Faruq, G. Fuel ethanol production from lignocellulosic biomass:
An overview on feedstocks and technological approaches. Renew. Sustain. Energy Rev. 2016, 66, 751–774.
[CrossRef]
4. Szczodrak, J.; Fiedurek, J. Technology for conversion of lignocellulosic biomass to ethanol. Biomass Bioenergy
1996, 10, 367–375. [CrossRef]
5. Rao, L.V.; Goli, J.K.; Gentela, J.; Koti, S. Bioconversion of lignocellulosic biomass to xylitol: An overview.
Bioresour. Technol. 2016, 213, 299–310.
6. Hassan, S.S.; Williams, G.A.; Jaiswal, A.K. Emerging Technologies for the Pretreatment of Lignocellulosic
Biomass. Bioresour. Technol. 2018, 262, 310–318. [CrossRef] [PubMed]
7. Salapa, I.; Katsimpouras, C.; Topakas, E.; Sidiras, D. Organosolv pretreatment of wheat straw for efficient
ethanol production using various solvents. Biomass Bioenergy 2017, 100, 10–16. [CrossRef]
8. Zhang, K.; Pei, Z.; Wang, D. Organic solvent pretreatment of lignocellulosic biomass for biofuels and
biochemicals: A review. Bioresour. Technol. 2016, 199, 21–33. [CrossRef] [PubMed]
9. Pan, X.; Kadla, J.F.; Ehara, K.; Gilkes, N.; Saddler, J.N. Organosolv ethanol lignin from hybrid poplar as
a radical scavenger: Relationship between lignin structure, extraction conditions, and antioxidant activity.
J. Agric. Food Chem. 2006, 54, 5806–5813. [CrossRef] [PubMed]
10. Wildschut, J.; Smit, A.T.; Reith, J.H.; Huijgen, W.J. Ethanol-based organosolv fractionation of wheat straw
for the production of lignin and enzymatically digestible cellulose. Bioresour. Technol. 2013, 135, 58–66.
[CrossRef] [PubMed]
11. Pan, X.; Gilkes, N.; Kadla, J.; Pye, K.; Saka, S.; Gregg, D.; Ehara, K.; Xie, D.; Lam, D.; Saddler, J.
Bioconversion of hybrid poplar to ethanol and co-products using an organosolv fractionation process:
Optimization of process yields. Biotechnol. Bioeng. 2006, 94, 851–861. [CrossRef] [PubMed]
12. Geng, A.; Xin, F.; Ip, J.Y. Ethanol production from horticultural waste treated by a modified organosolv
method. Bioresour. Technol. 2012, 104, 715–721. [CrossRef] [PubMed]
13. Sun, Y.C.; Wen, J.L.; Xu, F.; Sun, R.C. Structural and thermal characterization of hemicelluloses isolated by
organic solvents and alkaline solutions from Tamarix austromongolica. Bioresour. Technol. 2011, 102, 5947–5951.
[CrossRef] [PubMed]
14. Lee, J.H.; Kwon, J.H.; Kim, T.H.; Choi, W.I. Impact of planetary ball mills on corn stover characteristics
and enzymatic digestibility depending on grinding ball properties. Bioresour. Technol. 2017, 241, 1094–1100.
[CrossRef] [PubMed]
319
Energies 2018, 11, 2657
15. Avolio, R.; Bonadies, I.; Capitani, D.; Errico, M.E.; Gentile, G.; Avella, M. A multitechnique approach to
assess the effect of ball milling on cellulose. Carbohydr. Polym. 2012, 87, 265–273. [CrossRef]
16. Zhao, H.; Kwak, J.H.; Wang, Y.; Franz, J.A.; White, J.M.; Holladay, J.E. Effects of crystallinity on dilute acid
hydrolysis of cellulose by cellulose ball-milling study. Energy Fuels 2006, 20, 807–811. [CrossRef]
17. Khullar, E.; Dien, B.S.; Rausch, K.D.; Tumbleson, M.E.; Singh, V. Effect of particle size on enzymatic hydrolysis
of pretreated Miscanthus. Ind. Crops Prod. 2013, 135, 11–17. [CrossRef]
18. Liu, H.; Pang, B.; Zhao, Y.; Lu, J.; Han, Y.; Wang, H. Comparative study of two different alkali-mechanical
pretreatments of corn stover for bioethanol production. Fuel 2018, 221, 21–27. [CrossRef]
19. Shi, Z.; Liu, Y.; Xu, H.; Yang, Q.; Xiong, C.; Kuga, S.; Matsumoto, Y. Facile dissolution of wood pulp in
aqueous NaOH/urea solution by ball milling pretreatment. Ind. Crops Prod. 2018, 118, 48–52. [CrossRef]
20. Dai, L.; Li, C.; Zhang, J.; Cheng, F. Preparation and characterization of starch nanocrystals combining ball
milling with acid hydrolysis. Carbohydr. Polym. 2018, 180, 122–127. [CrossRef] [PubMed]
21. Zhao, X.; Cheng, K.; Liu, D. Organosolv pretreatment of lignocellulosic biomass for enzymatic hydrolysis.
Appl. Microbiol. Biotechnol. 2009, 82, 815. [CrossRef] [PubMed]
22. Nitsos, C.; Rova, U.; Christakopoulos, P. Organosolv fractionation of softwood biomass for biofuel and
biorefinery applications. Energies 2017, 11, 50. [CrossRef]
23. Teramoto, Y.; Tanaka, N.; Lee, S.H.; Endo, T. Pretreatment of eucalyptus wood chips for enzymatic
saccharification using combined sulfuric acid-free ethanol cooking and ball milling. Biotechnol. Bioeng.
2008, 99, 75–85. [CrossRef] [PubMed]
24. Yan, J.; Joshee, N.; Liu, S. Utilization of hardwood in biorefinery: A kinetic interpretation of pilot-scale
hot-water pretreatment of Paulownia elongata woodchips. J. Biobased Mater. Bioenergy 2016, 10, 339–348.
[CrossRef]
25. Liu, S. Woody biomass: Niche position as a source of sustainable renewable chemicals and energy and
kinetics of hot-water extraction/hydrolysis. Biotechnol. Adv. 2010, 28, 563–582. [CrossRef] [PubMed]
26. Yuan, T.Q.; You, T.T.; Wang, W.; Xu, F.; Sun, R.C. Synergistic benefits of ionic liquid and alkaline pretreatments
of poplar wood. Part 2: Characterization of lignin and hemicelluloses. Bioresour. Technol. 2013, 136, 345–350.
[CrossRef] [PubMed]
27. Weinwurm, F.; Turk, T.; Denner, J.; Whitmore, K.; Friedl, A. Combined liquid hot water and ethanol
organosolv treatment of wheat straw for extraction and reaction modeling. J. Clean. Prod. 2017, 165,
1473–1484. [CrossRef]
28. Kim, T.H.; Jeon, Y.J.; Oh, K.K.; Kim, T.H. Production of furfural and cellulose from barley straw using
acidified zinc chloride. Korean J. Chem. Eng. 2013, 30, 1339–1346. [CrossRef]
29. Kim, J.S.; Lee, Y.Y.; Kim, T.H. A review on alkaline pretreatment technology for bioconversion of
lignocellulosic biomass. Bioresour. Technol. 2016, 199, 42–48. [CrossRef] [PubMed]
30. Selig, M.; Weiss, N.; Ji, Y. Enzymatic Saccharification of Lignocellulosic Biomass; NREL/TP-510–42629;
National Renewable Energy Laboratory: Golden, CO, USA, 2018.
31. Sluiter, A.; Hames, B.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Tmpleton, D.; Crocker, D. Determination of Structural
Carbohydrates and Lignin in Biomass; NREL/TP-510–42618; National Renewable Energy Laboratory: Golden,
CO, USA, 2012.
32. Cao, Y.; Tan, H. Study on crystal structures of enzyme-hydrolyzed cellulosic materials by X-ray diffraction.
Enzyme Microb. Technol. 2005, 36, 314–317. [CrossRef]
320
energies
Review
A Critical Analysis of Bio-Hydrocarbon Production in
Bacteria: Current Challenges and Future Directions
Ziaur Rahman 1, *, Javed Nawab 2 , Bong Hyun Sung 3 and Sun Chang Kim 4, *
1 Department of Microbiology, Abdul Wali Khan University Mardan, Mardan 23200, Pakistan
2 Department of Environmental Sciences, Abdul Wali Khan University Mardan, Mardan 23200, Pakistan;
[email protected]
3 Cell Factory Research Center, Korea Research Institute of Bioscience and Biotechnology, Daejeon 34141,
Korea; [email protected]
4 Department of Biological Sciences, Korea Advanced Institute of Science and Technology,
Daejeon 34141, Korea
* Correspondence: [email protected] (Z.R.); [email protected] (S.C.K.);
Tel.: +92-340-901-2316 (Z.R.); +82-42-350-2619 (S.C.K.)
Received: 10 September 2018; Accepted: 3 October 2018; Published: 6 October 2018
Abstract: As global fossil reserves are abruptly diminishing, there is a great need for bioenergy.
Renewable and sustainable bioenergy products such as biofuels could fulfill the global energy
demand, while minimizing global warming. Next-generation biofuels produced by engineered
microorganisms are economical and do not rely on edible resources. The ideal biofuels are alcohols
and n-alkanes, as they mimic the molecules in fossil fuels and possess high energy densities. Alcohols
and n-alkane hydrocarbons (C2 to C18 ) have been produced using engineered microorganisms.
However, it is difficult to optimize the complex metabolic networks in engineered microorganisms to
obtain these valuable bio-hydrocarbons in high yields. Metabolic engineering results in drastic and
adverse cellular changes that minimize production yield in microbes. Here, we provide an overview
of the progress in next-generation biofuel (alcohols and n-alkanes) production in various engineered
microorganisms and discuss the latest tools for strain development that improve biofuel production.
Keywords: microbial biofuel; metabolic engineering; alkanes; alcohols
1. Introduction
The global scarcity and predicted depletion of fossil fuel reserves is a threat to future energy
demand [1–3]. Excess consumption of transportation fuel leads to ever-increasing global warming [1,4].
There is an urgent need to replace the current fossil fuel technology with renewable, sustainable,
and environmentally friendly alternatives [5,6]. Biofuels are considered sustainable if they are produced
from biomass through biological or non-biological methods. The ideal biofuels are bio-hydrocarbons
and these have been suggested as replacements for fossil fuels. Bio-hydrocarbons produced in
bacteria, such as alcohols (C2 –C18 ) and n-alkanes (C10 –C20 ), are of great importance [7–9]. These
bio-hydrocarbons have advantages over fossil fuels, as they possess higher energy densities and
emit fewer toxic chemicals into the environment [10,11]. These ready-to-use bio-hydrocarbons do
not require the replacement of existing engines and are compatible with current technologies [12].
The current methods used for production of these biofuels are non-economical and are based on
the conversion of plant and vegetable oil using the Fischer-Tropsch method [13,14]. The genetically
engineered microorganisms could utilize cheaper raw materials such as cellulosic sugars and carbon
dioxide to produce bio-hydrocarbons more economically.
Bio-hydrocarbons (alcohols and n-alkanes) are produced in engineered bacteria by modifying their
central metabolism, including glycolytic, tricarboxylic, and fatty acid pathways (Figure 1). The current

Energies 2018, 11, 2663
production yield of bio-hydrocarbons from modified bacteria is far below commercial scale due to the
challenges of metabolic engineering. Current challenges include cellular toxicity, growth retardation,
cofactor exhaustion and metabolic pathway imbalance. In this paper, the metabolic engineering of
bacteria for bio-hydrocarbon production is reviewed. In addition, we discuss current problems and
challenges in strain optimization for bio-hydrocarbon production and provide future directions for
this technology.
Figure 1. Metabolic pathways of short- and medium-chain alcohols and hydrocarbon production in an
engineered bacterial cell. Glucose is metabolized to acetyl-CoA, a precursor of the tricarboxylic acid
(TCA) cycle and fatty acid biosynthesis pathways. Short-chain alcohols (ethanol, propanol, butanol) are
produced from acetyl-CoA and TCA. Fatty alcohols, fatty acid ethyl esters and n-alkanes are produced
by utilizing fatty acyl-ACPs of the fatty acid machinery. The native pathways are shown in black, while
heterologous pathways are shown in red. CoA: Coenzyme A; ACP: Acyl Carrier Protein; OAA: Oxalo
Acetic Acid; CIT: Citrate; ICT: Iso Citrate; OGA: Oxo Glutaric Acid; SUC: Succinic Acid; FUM: Fumaric
Acid; MAL: Malic Acid; FAME: Fatty Acid Methyl Ester; FAEE: Fatty Acid Ethyl Ester.
2. Engineering Metabolic Pathways for Hydrocarbon Production in Bacteria

Initial work on engineering microbes for biofuel production involved introducing the
ethanol-producing pathway from Zymomonas mobilis into Escherichia coli [15]. The progress of most
biofuels produced this way is discussed in detail in previous reviews [16–18]. In this section, we
present an overview of alcohol and n-alkane biofuels produced by engineered microorganisms.
2.1. Alcohols Production in Bacteria

Next-generation alcohols are a group of compounds used in chemical industries and as
transportation biofuels [19,20]. They can be grouped into short, medium, and long-chain alcohols,
ranging from C2 to C6 , C6 to C12 , and C12 to C18 chain lengths, respectively [21–23]. They can be
used as 100% solutions or blended with petroleum. Custom alcohols are currently produced by
distillation during petroleum processing or chemically from fatty acids, which makes their production
unviable in the long run. The sustainable approach to their production is to use fermenters, utilizing
microorganisms that can easily convert cellulosic material into ready-to-use biofuels. Ethanol is a
renewable liquid fuel and its annual production is greater than 100 billion liters, with an increase of 10%
per year. More than 60% of ethanol is produced with fermentation technology using microorganisms [24].
322
Energies 2018, 11, 2663
From an engineering perspective, microbial strain selection is an important consideration. The best
strain is one that suits and resists the characteristics of the fermenter environment, such as pH,
temperature, acidity, agitation, and biofuel toxicity.
Wild-type Escherichia coli produces ethanol under fermentative conditions by converting pyruvate
to acetyl-CoA and formate through the enzyme pyruvate formate lyase, encoded by pfl gene [15].
This process involves a two-step reaction: acetyl-CoA is converted to acetaldehyde, followed by
ethanol synthesis. The overall reaction is catalyzed by a bifunctional aldehyde dehydrogenase (ADH),
encoded by the adhE gene (Figure 1). Similarly, butanol is produced by the fermentative anaerobe,
Clostridium acetobutylicums, through the acetone-butanol-ethanol (ABE) pathway [19]. The problem
with wild-type strains is the low production titer. This is the main focus of most research into enhancing
yields using metabolic engineering tools.
Pichia pastoris used for industrial production of chemicals was also engineered for butanol
production [25]. The amino acid synthesis pathways of Pichia pastoris, specifically the endogenous
L-valine biosynthesis pathway, was overexpressed to produce 2.22 g/L of butanol from glucose as
a carbon source [25]. Furthermore, using cheaper raw materials for the production of renewable
biofuel could minimize the final cost. Glycerol is one of the cost effective byproducts of the biodiesel
production industry. Engineered microbes could utilize glycerol as a carbon source. To obtain the
phenotype, Saini et al. (2017) rewired the central metabolism of Escherichia coli to utilize glycerol as
a carbon source, introduced a butanol synthesis pathway, and obtained a titer of 6.9 g/L butanol
from 20 g/L glycerol (Table 1) [26]. This highest titer was obtained because of forcing glycolytic flux,
enabling gluconeogenic pathway, improving the breakdown of glycerol and moderately suppressing
tricarboxylic acid pathway [26]. To enhance production yield, Liew et al. (2017) introduced a novel
pathway enzyme, aldehyde/alcohol ferredoxin reductase (AOR), to replace the native bifunctional
aldehyde dehydrogenase in Clostridium autoethanogenum [9]. The strain was engineered using the
allelic mutagenesis approach and as a result, the final strain produced ethanol at a titer of up to 180%
(0.28 g/L) from CO2 as a carbon source in batch fermentation experiments [9].
Furthermore, the yeast strains were found to be promising for butanol production. Pichia pastoris used
industrial chemical production and has also been engineered for butanol production [25]. The endogenous
L-valine biosynthesis pathway of P. pastoris was overexpressed to produce 2.22 g/L of butanol from
glucose as a carbon source [25]. Furthermore, using cheaper raw materials for renewable biofuel
production could minimize the final cost. Glycerol, a byproduct of biodiesel production and other
industrial processes, is a cost-effective raw material and engineered microbes could utilize it as a carbon
source. Saini et al. (2017) rewired the central metabolism of Escherichia coli to utilize glycerol as the sole
carbon source and introduced a butanol synthesis pathway that gives a production titer of 6.9 g/L of
butanol from 20 g/L of glycerol (Table 1) [26]. This high titer was obtained by forcing glycolytic flux,
enabling the gluconeogenic pathway, improving breakdown of glycerol, and moderately suppressing
the tricarboxylic acid pathway [26]. The breakthrough report for butanol production was obtained
using a Clostridium tyrobutyricum strain [27]. The complex regulatory pathway of butyrate production
was manipulated by using the CRISPR-Cas system to delete cat1 (regulatory component) and to
integrate adhE1 (alcohol producing) genes. The resultant strain achieved a record butanol titer of
26.2 g/L in batch fermentation using CO2 as the carbon source (Table 1) [27].
2.2. Alkanes Production in Bacteria

High-energy renewable-fuel-like n-alkanes are direct replacements for petroleum fuel. These
molecules range in chain length from C8 to C18 . Chemically, n-alkanes are produced by the
Fischer–Tropsch process, making them uneconomical. Biologically, n-alkanes are produced in
Escherichia coli by modifying the lipid biosynthesis pathway and introducing aldehyde reductase
(AAR) and aldehyde-deformylating oxygenase (ADO) from cyanobacteria [28]. Fatty acid synthesis
plays a vital role in the production of n-alkanes. The prokaryotic fatty acid pathway is composed
of individual components/enzymes. Acetyl-CoA enters the fatty acid pathway and a subsequent
323
Energies 2018, 11, 2663
condensation reaction leads to the elongation of fatty acids. The alkane’s production is directly related
to fatty acid biosynthesis in bacteria. The fatty acid synthesis machinery of E. coli produces fatty
acids from malonyl-CoA condensation. The first step in the fatty acid biosynthesis is carried out by
acetyl-CoA carboxylase (ACC), a tetramer enzyme which produces malonyl-CoA from acetyl-CoA.
The malonyl-CoA then transfers to acyl carrier protein (ACP) an elongation unit to get malonyl-ACP
with the help of FabD, malonyl-CoA: ACP transcylase enzyme. The condensation cycle composes of
four reactions catalyzed by FabH (3-oxooacyl-ACP synthase III), FabG (3-oxooacyl-ACP reductase),
FabA/FabZ (enoyl-ACP synthase), and FabI (enoyl-ACP reductase), responsible for the production
of 3-ketoacyl-ACP, 3-hydroxyacyl-ACP, enoyl-ACP, and acyl-ACP, respectively [16]. The fatty acids
are then released from acyl carrier proteins with the help of thioesterases. Free fatty acids are then
converted into fatty aldehydes and n-alkanes by AAR and ADO, respectively. The details of microbial
biofuel production are shown in Table 1 and Figure 2.
Figure 2. Recent studies on the production of short-chain alcohols, medium-chain alcohols and n-alkanes
in microorganisms. The data represent batch fermentation experiments. For more detail, see Table 1.
3. Comparative Analysis of Bio-Hydrocarbons Production in Bacteria

Metabolic engineering plays a vital role in enhancing the production yield of bio-hydrocarbons.
The ethanol produced in wild type E. coli with aldehyde dehydrogenase enzyme is far below the
commercial threshold. To improve production yield, the genes of sleepy beauty mutase operon
containing methyl malonyl CoA mutase, methyl malonyl CoA decarboxylase propinoyl CoA: Succinate
transferase enzymes were expressed in E. coli that successfully produced 31 and 7 g/L ethanol and
propanol, respectively in a fed-batch fermentation (Tables 1 and 2). The pathway for the production of
propanol using sleeping beauty mutase (Sbm) was different when compared to propanol production
using the keto acid pathway [21]. The highest production of propanol by Sbm pathway may be due to
anaerobic conditions as compared to the aerobic keto acid pathway [21,29].
324
Energies 2018, 11, 2663
The C4 alcohol (butanol) is considered better than ethanol in terms of hygroscopic properties and
high energy densities [21]. The Clostridium acetobutylicum is a natural producer of butanol. The main
problem in the wild type strain is butyrate production that lowers butanol synthesis. Zhang et al.
(2018) speculated that by interrupting the butyrate synthesis genes, the butanol production may
be enhanced. The spoA and cat1 genes of butyrate synthesis were selectively deleted by using the
CRISPR-Cas technology and as a result the highest titer of 26.2 g/L butanol in batch fermentation was
achieved (Tables 1 and 2). The breakthrough report for butanol production was obtained in Clostridium
tyrobutyricum strain [27]. The complex regulatory pathway of butyrate production was manipulated
with the CRISPR-Cas system by deletion of cat1 (regulatory component) and integration of adhE1
(alcohol producing) genes [27]. The resultant strain achieved record production of butanol up to
26.2 g/L in a batch fermentation experiment using CO2 as a carbon source (Figure 2) [27].
The fatty alcohol’s production is linked with fatty acid synthesis pathways [16]. The fermentative
products such as lactate, acetate, pyruvate, and formate are also synthesized in fatty alcohols producing
strains that lower the final production titer [8]. To overcome these problems, Fatima et al. (2018)
developed a constraint-based modeling approach and identified the competing pathways, such as
formate, lactate, and acetate synthesis. By blocking these pathways, 1.5 and 12.5 g/L fatty alcohols
were obtained in batch and fed-batch fermentation, respectively (Tables 1 and 2).
The alkane hydrocarbons are considered as a direct replacement for fossil petroleum [20]. The bio
production of alkanes remains a major challenge because of slow acting aldehyde deformylating
oxygenase, an alkane-producing enzymes (Figure 1) [28]. Currently, a number of efforts have been
made to boost the alkane production but it is still far below the commercial threshold. The second
problem is the toxicity of intermediate fatty aldehyde to the host cells. The toxicity of fatty
aldehydes was overcome by applying DNA scaffolds to enzymes AAR and ADO and enhanced
8-fold alkanes production 0.04 g/L [28]. Cao et al. (2016) blocked the competing pathways such
as endogenous reductases, and engineered the electron transfer system towards alkane-producing
enzymes. The resultant strains produced 0.1 and 1.3 g/L alkanes in a batch and fed-batch fermentation,
respectively [30]. The details about current production yields and experimental conditions can be
found in Tables 1 and 2.
Table 1. Biofuel production in engineered bacteria.
Biofuel Substrate Microorganism Metabolic Engineering Yield (g/L) References

Heterologous expression of sleeping beauty
Glycerol Escherichia coli 0.20 a , 31 b [29]
mutase operon
Ethanol Allelic mutagenesis of aldehyde dehydrogenase
Clostridium
CO2 (adhE) and expression of aldehyde ferredoxin 0.28 a [9]
autoacetobutylicum
reductase (AOR)
Heterologous expression of sleeping beauty
Propanol Glycerol Escherichia coli 0.15 a , 7 b [29]
mutase operon
Targeted deletion of sopA and cat1 (butyrate
Clostridium
CO2 production) and integration of adhE gene 26.2 a [27]
acetobutylicum
by CRISPR-Cas9
Enhanced glycolytic flux, prevention of
Glycerol Escherichia coli 6.9 a [26]
glycerol breakdown
Butanol
2-keto
Pichia pastoris Expression of keto acid degradation pathway 0.28 a [25]
isovalerate
Overexpression of endogenous L-Valine
Pichia pastoris 0.89 a [25]
biosynthesis pathway
Glucose Fine tuning of expression levels using
Pichia pastoris 2.22 a [25]
episomal plasmids
325
Energies 2018, 11, 2663
Table 1. Cont.
Biofuel Substrate Microorganism Metabolic Engineering Yield (g/L) References

In silico modelling of metabolic pathways for
Fatty
Glucose Escherichia coli alkane production to remove bottlenecks 1.5 a , 12.5 b [8]
alcohols
identified in pentose phosphate pathway
Expression of alkane-producing pathway,
Glucose Yarrowia lipolytica consisting of aldehyde reductase (AAR) and 2.15 a [20]
aldehyde deformylating oxygenase (FAD)
Expression of alkane-producing pathway
Saccharomyces Bottlenecks were identified by flux balance
Lignocellulose 1.2 a , 6 b [7]
cerevisiae analysis and bacterial fatty aldehyde reductases
(FARs) were replaced with mouse FARs
Heterologous expression of alkane-producing
pathway (AAR & aldehyde-deformylating
Escherichia coli 0.06 a [30]
oxygenase (ADO)) from cyanobacteria, with no
further modification
Deletion of adhE (endogenous reductase)
Alkanes Glucose Expression of alkane-producing pathway and

Escherichia coli 0.1 a , 1.3 b [30]
modification of electron transfer system
Expression of alkane-producing pathway and
DNA scaffolding of cyanobacterial AAR & ADO
Yarrowia lipolytica 0.02 a [20]
(FAR & FAD), with no further modification
In silico modelling of metabolic pathways for
Escherichia coli alkane production to help remove bottlenecks in 0.43 a , 2.5 b [8]
pentose phosphate pathway
Notes: a : batch fermentation; b : fed-batch fermentation.
Table 2. Comparison of experimental parameters of biofuel production in bacteria.
Biofuel Microorganism Experimental Conditions Yield (g/L) References

Substrate Cell Density Reaction
Substrate pH Temperature [29]
Concentration Initial-Final Time
Escherichia coli Glycerol 30 g/L 7.0 30 ◦ C 0.4–1.8 g/DCW/L 28 h 0.20 a , 31 b [9]
Ethanol Clostridium
CO2 50 kPa 5.8 37 ◦ C 0.2–1.8 O.D600 30 h 0.28 a [29]
autoacetobutylicum
Propanol Escherichia coli Glycerol 30 g/L 7.0 30 ◦ C 1.8 g/DCW/L 28 h 0.15 a , 7 b [27]
Clostridium ◦C a
CO2 - 6.0 37 - 30 h 26.2 [26]
acetobutylicum
Escherichia coli Glycerol 30 g/L 7.0 37 ◦ C 0.2–7.0 O.D550 40 h 6.9 a [25]
Butanol 2-keto
Pichia pastoris 4 g/L - 30 ◦ C 0.05–12 O.D550 72 h 0.28 a [25]
isovalerate
Pichia pastoris Glucose 20 g/L - 30 ◦ C 0.05–1.2 O.D550 72 h 0.89 a [25]
Pichia pastoris Glucose 20 g/L - 30 ◦ C 0.05–2.0 O.D550 72 h 2.22 a [8]
Fatty ◦C a, b
Escherichia coli Glucose 20 g/L 7.0 30 0.02-O.D550 48 h 1.5 12.5 [20]
alcohols
Yarrowia lipolytica Glucose 100 g/L 5.5 28 ◦ C 1–80 O.D600 120 h 2.15 a [7]
Saccharomyces
Lignocellulose 20 g biomass 5.0 30 ◦ C 1–25 O.D600 240 h 1.2 a , 6 b [30]
cerevisiae
Escherichia coli Glucose 20 g/L 7.0 30 ◦ C 0.5–4.0 O.D600 50 h 0.06 a [30]
Alkanes Escherichia coli Glycerol 140 g/L 7.0 30 ◦ C 0.5–140 O.D600 50 h 0.1 a , 1.3 b [28]
Yarrowia lipolytica Glucose 100 g/L 5.5 28 ◦ C 1–80 O.D600 120 h 0.02 a [8]
Escherichia coli Glucose 20 g/L 7.0 37 ◦ C 0.02–102 O.D550 72 h 0.43 a , 2.5 b [29]
Notes: a : batch fermentation; b : fed-batch fermentation.
326
Energies 2018, 11, 2663
4. Current Challenges in Improving Hydrocarbon Production
4.1. Promoters
Promoters are important genetic elements responsible for the controlled expression of genes on
a chromosome or plasmid [31,32]. In wild-type cells, promoters are well organized and controlled
by repressors and inducible factors. From a bioengineering perspective, promoters are critical for
balancing the metabolic flux of heterologous and/or engineered pathways by controlling the expression
levels of individual genes. Therefore, choosing the correct promoter is an important step in engineering
prokaryotic and eukaryotic cells. Uncontrolled gene expression in a heterologous host exerts a
metabolic burden and results in the exhaustion of cellular biomolecules that are essential for cellular
growth, leading to undesired physiological changes (Figure 3) [33]. Therefore, the promoter needs to be
properly employed for smooth growth and productivity of the desired products. For example, when a
target protein is toxic to the cell, a low-strength promoter is needed. If the expression of multiple genes
is not optimized or if the catalytic activity of an enzyme is greater than the corresponding enzymes
of the same metabolic pathway, it is recommended to balance the pathway with native/synthetic
promoters to optimize flux towards the product. A wide range of promoters have been characterized
and are available for research and academic purposes [34]. Historically, natural constitutive and
inducible promoters have been used in this research. However, more recently, novel artificial promoters
have been designed by modifying the −35 (TTGACA) and −10 (TATA) sequences. The expression
patterns of these novel promoters are diverse [34]. Furthermore, two modes of transcription initiation
have been reported. In the positive control mode, the interaction between regulatory proteins and
their corresponding recognition sequences switches on transcription. In contrast, the negative control
mode inhibits transcription initiation by the interaction of regulatory proteins with their regulatory
regions. These synthetic promoters have been studied with the help of reporter genes coding for
fluorescent proteins (GFP, RFP), luciferase, and galactosidase [35]. The promoters can be switched
on/off using both chemical and physical stimuli. For instance, the Lac promoter is controlled by the
LacI repressor and the IPTG inducer. The Tet promoter is controlled by the TetR repressor and the Tc
inducer [36]. Therefore, selecting the appropriate promoter is critical for controlling gene expression
levels and optimizing pathway flux. This can be achieved using either single-promoter-single-gene
or single-promoter-multiple-gene expression systems. In addition, certain enzymes are required at
different times of fermentation. Heat-inducible phage promoters and constitutive expression systems
of Bacillus subtilis have proven promising for industrial and batch fermentation scales [37]. Currently,
synthetic or artificial promoters are used in bacterial, yeast and mammalian cells. Basic questions about
balancing pathway flux for biofuel production have been answered using synthetic promoters [38–40].
Another bottleneck in strain development for biofuel production is the combinatorial assembly of
building blocks (promoters, genes, and terminators). These issues have mostly been solved using
computational approaches, such as the DNA fragment compiler software, BioPartsBuilder [41].
4.2. Gene Copy Number

Gene expression level can also be regulated by controlling gene copy number (GCN). Changes in
microbial GCN have been successful at balancing the flow of metabolites towards biofuels in engineered
pathways. Enzymes in metabolic pathways have different catalytic turnover rates. Therefore, if their
expression is not normalized according to their catalytic efficiencies, physiological disturbances such
as toxicity and cell death can occur [33]. The strategies used to control the desired gene expression
levels through GCN by selecting low-, medium-, and high-copy-number plasmids are discussed
in the literature [42]. Gene expression levels can also be controlled by knocking down genes or
integrating them into a chromosome, leading to a decrease or increase in GCN, respectively [43].
Furthermore, the strategy to engineer high GCN plasmids has been successful in the fission yeast,
Saccharomyces cerevisiae. In S. cerevisiae, CEN/ARS a centromere based plasmid has been engineered to
obtain plasmid diversity with low, medium and high copy number versions [44]. For the development
327
Energies 2018, 11, 2663
of Escherichia coli strains, plasmids with low, medium and high copy number per cell have been
developed. Based on the ColE1 replicon, 15 to 20 plasmid copies per cell can be maintained.
To obtain higher gene expression levels, 200+ copies of the pMB1 replicon in the pUC plasmid
have been used [45]. More recently, traditional gene cloning approaches are being replaced with
novel self-cloning techniques, based on transcription activator-like effector nuclease (TALEN) in
Pseudochoricystis ellipsoidea [46].
Figure 3. Graphical representation of improving bio-hydrocarbon production in bacteria. The figure

shows challenges and possible solutions in transcription (A) and metabolite toxicity (B) in newly
engineered strains.
4.3. Artificial Transcription Factor (ATF) Engineering

Artificial transcription factors (ATF) control genome-wide gene expression levels in engineered
host cells to balance metabolic pathways (Figure 3). Transcription factors are composed of a DNA
binding domain and an effector domain, both of which are required for regulating the expression of
downstream genes. Specifically, the DNA binding domain is responsible for binding to the target
regulatory site of a gene and the effector domain facilitates the initiation of downstream gene expression
by RNA polymerase. Most ATFs are composed of zinc fingers (ZF) and sigma factors. ZF proteins
bind specifically to DNA and function as regulatory factors. ATFs recognize specific DNA sequences,
typically near the transcription start site of the target gene. Numerous methods have been developed
for the construction of ATFs [47]. At present, ATFs are mainly used in industrial microorganisms to
improve their tolerance to environmental stresses and to enhance the yield of chemicals and fuels.
The ATF-based strain engineering approach has proven successful in improving tolerance to various
environmental and physical stresses [48]. The T2 ATF has been shown to generate thermo tolerance
and butanol tolerance phenotypes [48]. Furthermore, DNA microarray analysis of a T2 ATF strain
showed up-regulation of the outer membrane protein gene, ompW and down-regulation of the marRAB
328
Energies 2018, 11, 2663
operon. The exact function of OmpW is not known, but a previous publication has suggested that it
has an important role during environmental stress and under physically varied conditions in industrial
bioreactors [48]. Lee et al. (2011) engineered an ATF library by fusing ZFs with a cyclic adenosine
mono phosphate receptor protein (CRP) and screened for butanol tolerance in E. coli [49]. Wild-type
E. coli could not tolerate 1% (v/v) butanol, but a butanol-tolerant strain with the ATF was able to
tolerate 1.5% (v/v) butanol (Lee et al. 2011). Similarly, lignocellulosic properties of fungal species were
studied for efficient degradation of plant cell walls. Ling et al. (2011) engineered a transcription factor
from Saccharomyces cerevisiae to generate an n-alkanes-resistant phenotype by site-directed mutagenesis
of the drug resistant transcription factors (PDR), PDR1P and PDR3P. Point mutations were generated
(PDR1, F815S and PDR3, Y276H), which resulted in high tolerance to C10 -alkanes (Ling et al. 2015).
More recently, an acetic-acid-resistant phenotype was also generated using an ATF that successfully
recognized novel genes responsible for the tolerance phenotype [50].
4.4. Ribosome Engineering

Ribosome engineering is also considered a useful tool for modulating gene expression levels.
Ochi (2007) showed that the gene for ribosomal proteins, rpsL, could be engineered to dramatically turn
on previously switched off genes [51]. Ribosomes, therefore, may be a useful tool for the production of
microbes for biofuel production. The principle of ribosomal gene modification is to mimic the natural
antibiotic resistance system developed by bacteria [52].
Ribosome engineering has proven to be a successful approach for the production of primary and
secondary metabolites. Tanaka modified the rpsL gene in Propionibacterium shermanii, resulting in more
than fivefold enhanced production of vitamin B12 [53]. The challenge with ribosome engineering is
growth retardation of engineered strains, which may be due to the unavailability of free ribosomes for
cellular growth. In addition, ribosome modification in Bacillus subtilis resulted in up to a 30% increase
in amylase and protease production in the stationary phase [54]. Besides ribosome modification,
the rapid availability of ribosomes for desired metabolic pathways has also proven to be beneficial for
the production of antibiotics. This approach, which involves the introduction of a ribosome release
factor, has been applied for improved antibiotic production in Streptomyces [55,56].
4.5. Transfer RNA Engineering

Translation is the final step in the gene expression process. Translation requires energy and
involves several steps: initiation, elongation and termination. The process is tightly regulated by a
number of factors, such as the synthesis and degradation of the messenger RNA (mRNA), activation
of transfer RNA (tRNA), and the coupling of tRNA with ribosomes. To obtain strains with high
biofuel production, individual components involved in the translation process such as mRNA, tRNA,
ribosomes, and various proteins required for translation can be modulated.
Biofuel-producing phenotypes have problems with fuel toxicity, as most proteins are denatured
by the hydrophobic biofuel. This type of issue may be solved using tRNA engineering approaches.
Recently, modified tRNAs were designed with different codon specificities to incorporate unnatural
amino acids into proteins, resulting in the formation of novel proteins with altered or enhanced
properties, such as resistance to toxic substances or enhanced catalytic efficiency [57].
Protein synthesis is based on the universal genetic code recognized by specific tRNAs. The term
“tRNA engineering” refers to the chemical modification of tRNAs with unnatural amino acids that
can then be incorporated into proteins [58]. Recently, in vivo modification of tRNA was demonstrated
using a special aminoacyl tRNA synthetase to esterify unnatural amino acids to an engineered tRNA,
i.e., tRNApyl [59,60]. More than 30 novel amino acids have been introduced into proteins using
this technology [60]. Similarly, pyrrolysine, an unnatural amino acid analogue of lysine with a
methyl-pyrroline moiety, was first identified in the active site of the enzyme monomethylamine
methyltransferase from Methanosarcina barkeri [61]. Furthermore, it was found that pyrrolysine
329
Energies 2018, 11, 2663
incorporation was due to tRNApyl containing the CUA anticodon, complementary to UAG bases.
The uncharged tRNApyl is activated by pyrrolysyl-tRNA synthetase (PylRS).
This technique can also be used to fine-tune certain enzymes to improve their function or
pharmacological properties, e.g., in vivo incorporation of p-azido-L-phenylalanine for Phe170 or
Phe281 in E. coli urate-oxidase [57]. Engineering proteins for functional metabolic pathways using this
strategy may improve the productivity of biofuel in microbes. The described technologies also provide
a path forward for biotech industries to modify their targets for improved activity [57].
Finally, advances in tRNA engineering broaden the field of phenotype engineering by allowing
the incorporation of unnatural amino acids into structural and functional proteins of the cell. However,
engineering aminoacyl tRNA synthetases and screening natural systems still present challenges for
tRNA engineering.
4.6. Cofactor Regeneration

Cofactors such as NADH, NADPH, GTP, ATP, FAD, and FADH are important partners of
catalytic reactions and enzymatic pathways. Therefore, the availability of these molecules is vital
for a chemical reaction to proceed and improving their availability can enhance the efficiency of
the pathway. The continuous supply of ATP in an engineered microbial host can boost the desired
products [62]. The glutathione metabolic pathway requires ATP and blocking the ATP-consuming nlpD
gene improves glutathione production [62]. Wang et al. (2013) showed up to a 90% production yield of
2,3-butanediol by improving NADH availability through overexpression of NADH-generating glucose
and formate dehydrogenase enzymes [63]. Similar approaches have been applied in the yeast Pichia
pastoris where blocking NADH-consuming pathways led to a boost in butanediol production [64].
4.7. Scaffold System

Nature has provided a sophisticated mechanism for enhancing pathway catalysis by using
scaffolding enzymes. Many organisms have evolved multifunctional enzyme complexes such as fatty
acid, polyketide, and alkaloid synthases for the biosynthesis of fatty acids polyketides, and alkaloids,
respectively [65–67]. Applying this approach has led to the efficient production of artemisinin, an
anti-malarial drug; n-alkanes; and biodiesel [16,28,68–70]. In our previous study, we demonstrated
the potential of a DNA scaffold system for n-alkane production in E. coli. The critical enzymes of the
n-alkane pathway, acyl-ACP reductase (AAR) and ADO, were linked to four zinc finger DNA-binding
domains, capable of binding to a 12 bp sequence on the DNA scaffold. This resulted in an eightfold
increase in n-alkane production [16]. Design in synthetic biology is usually inspired by the spatial
organization of enzymes observed in nature. This close approximation of enzymes to their desired
distances and angles is made possible using modular scaffolds. The concept of scaffolding involves the
use of small building blocks to allow the channeling of intermediates during a biochemical reaction
(Figure 3). The advantage of such a system is that it prevents the diffusion of intermediates of
sequential enzymatic reactions to nearby competing pathways, minimizing transit time and toxicity of
certain intermediates [16,28,70,71]. Types of scaffolding include the immobilization, co-localization,
compartmentalization, and spatial organization of enzymes on nucleic acids (DNA/RNA) and
protein-protein scaffolding. During immobilization and co-localization, the enzymes of interest
are targeted to the membrane via signal peptides as leading sequences. Enzyme scaffolds on nucleic
acids are constructed by tagging nucleic acid binding domains onto the desired enzymes so that they
land on the sequence of interest. Protein scaffolds can be constructed either as a fusion protein of two
or more enzymes to form a large single multi-enzyme protein or by fusing an individual enzyme with
protein binding domains that eventually lead to a complex of proteins or scaffolds. Coiled-coil proteins
have been used to create this type of protein scaffold [72].
330
Energies 2018, 11, 2663
5. Conclusions and Future Directions

The global energy crisis and ever-increasing global warming may be addressed by decreasing
reliance on fossil fuels. Alcohols and n-alkanes are currently the best replacements for petroleum fuel.
The production yields of these biofuels are far below commercial level due to technical problems in
strain optimization. In this review, gene expression optimization strategies such as the use of promoters,
artificial transcription factors and gene copy number alterations were discussed. In addition, strategies
for cofactor regeneration and scaffolding of pathway enzymes for metabolic pathway optimization
were also discussed. Future tools must focus on rapid screening technologies for the selection of single
microbial cells in real time to boost biofuel production.
Author Contributions: Z.R. conceived of this work. J.N. contributed to the manuscript text and figures. B.H.S.
and S.C.K. supervised this work.
Funding: This work was supported by grants from the Higher Education Commission of Pakistan (Grant No.
5197/KPK/NRPU/R&D/HEC/2016), the Intelligent Synthetic Biology Center of Global Frontier Project through
the National Research Foundation funded by the Ministry of Science & ICT of Korea, and the Research Initiative
Program of KRIBB.
Acknowledgments: We are very thankful to Raham Sher, Department of Biotechnology, AWKUM for providing
insights about global warming.
Conflicts of Interest: The authors have no conflicts of interest to declare.
References
1. Goldthau, A. The G20 must govern the shift to low-carbon energy. Nature 2017, 546, 203–205. [CrossRef]
[PubMed]
2. Balch, J.K.; Nagy, R.C.; Archibald, S.; Bowman, D.M.; Moritz, M.A.; Roos, C.I.; Scott, A.C.; Williamson, G.J.
Global combustion: The connection between fossil fuel and biomass burning emissions (1997–2010).
Philos. Trans. R. Soc. Lond. B Biol. Sci. 2016, 371. [CrossRef] [PubMed]
3. Den, W.; Sharma, V.K.; Lee, M.; Nadadur, G.; Varma, R.S. Lignocellulosic biomass transformations via greener
oxidative pretreatment processes: Access to energy and value-added chemicals. Front. Chem. 2018, 6, 141.
[CrossRef] [PubMed]
4. Nejat, P.; Jomehzadeh, F.; Taheri, M.M.; Gohari, M.; Majid, M.Z.A. A global review of energy consumption,
CO2 emissions and policy in the residential sector (with an overview of the top ten CO2 emitting countries).
5. Shields-Menard, S.A.; Amirsadeghi, M.; French, W.T.; Boopathy, R. A review on microbial lipids as a potential
biofuel. Bioresour. Technol. 2018, 259, 451–460. [CrossRef] [PubMed]
6. Dresselhaus, M.; Thomas, I. Alternative energy technologies. Nature 2001, 414, 332. [CrossRef] [PubMed]
7. d’Espaux, L.; Ghosh, A.; Runguphan, W.; Wehrs, M.; Xu, F.; Konzock, O.; Dev, I.; Nhan, M.; Gin, J.;
Reider Apel, A.; et al. Engineering high-level production of fatty alcohols by Saccharomyces cerevisiae from
lignocellulosic feedstocks. Metab. Eng. 2017, 42, 115–125. [CrossRef] [PubMed]
8. Fatma, Z.; Hartman, H.; Poolman, M.G.; Fell, D.A.; Srivastava, S.; Shakeel, T.; Yazdani, S.S. Model-assisted
metabolic engineering of Escherichia coli for long chain alkane and alcohol production. Metab. Eng. 2018,
9. Liew, F.; Henstra, A.M.; Kpke, M.; Winzer, K.; Simpson, S.D.; Minton, N.P. Metabolic engineering of
Clostridium autoethanogenum for selective alcohol production. Metab. Eng. 2017, 40, 104–114. [CrossRef]
[PubMed]
10. Demirbas, A. Biofuels sources, biofuel policy, biofuel economy and global biofuel projections. Energy Convers.
Manag. 2008, 49, 2106–2116. [CrossRef]
11. Pandey, R.K.; Tewari, L. Mycotechnology for lignocellulosic bioethanol production: An emerging approach
to sustainable environment. In Microbial Biotechnology in Environmental Monitoring and Cleanup; IGI Glob:
Hershey, PA, USA, 2018; pp. 28–43. [CrossRef]
12. Babu, V.; Murthy, M. Butanol and pentanol: The promising biofuels for CI engines–A review. Renew. Sustain.
Energy Rev. 2017, 78, 1068–1088.
331
Energies 2018, 11, 2663
13. No, S.-Y. Application of hydrotreated vegetable oil from triglyceride based biomass to CI engines—A review.
Fuel 2014, 115, 88–96. [CrossRef]
14. Van Vliet, O.P.; Faaij, A.P.; Turkenburg, W.C. Fischer—Tropsch diesel production in a well-to-wheel
perspective: A carbon, energy flow and cost analysis. Energy Convers. Manag. 2009, 50, 855–876. [CrossRef]
15. Ohta, K.; Beall, D.S.; Mejia, J.P.; Shanmugam, K.T.; Ingram, L.O. Genetic improvement of Escherichia coli for
ethanol production: Chromosomal integration of Zymomonas mobilis genes encoding pyruvate decarboxylase
and alcohol dehydrogenase ii. Appl. Environ. Microbiol. 1991, 57, 893–900. [PubMed]
16. Rahman, Z.; Rashid, N.; Nawab, J.; Ilyas, M.; Sung, B.H.; Kim, S.C. Escherichia coli as a fatty acid and biodiesel
factory: Current challenges and future directions. Environ. Sci. Pollut. Res. Int. 2016, 23, 12007–12018.
[CrossRef] [PubMed]
17. Lu, X. A perspective: Photosynthetic production of fatty acid-based biofuels in genetically engineered
Cyanobacteria. Biotechnol. Adv. 2010, 28, 742–746. [CrossRef] [PubMed]
18. Zhou, Y.J.; Buijs, N.A.; Siewers, V.; Nielsen, J. Fatty acid-derived biofuels and chemicals production in
Saccharomyces cerevisiae. Front Bioeng. Biotechnol. 2014, 2, 32. [CrossRef] [PubMed]
19. Lin, Y.L.; Blaschek, H.P. Butanol production by a butanol-tolerant strain of Clostridium acetobutylicum in
extruded corn broth. Appl. Environ. Microbiol. 1983, 45, 966–973. [PubMed]
20. Xu, P.; Qiao, K.; Ahn, W.S.; Stephanopoulos, G. Engineering Yarrowia lipolytica as a platform for synthesis of
drop-in transportation fuels and oleochemicals. Proc. Natl. Acad. Sci. USA 2016, 113, 10848–10853. [CrossRef]
[PubMed]
21. Atsumi, S.; Cann, A.F.; Connor, M.R.; Shen, C.R.; Smith, K.M.; Brynildsen, M.P.; Chou, K.J.; Hanai, T.;
Liao, J.C. Metabolic engineering of Escherichia coli for 1-butanol production. Metab. Eng. 2008, 10, 305–311.
[CrossRef] [PubMed]
22. Rutter, C.D.; Rao, C.V. Production of 1-decanol by metabolically engineered Yarrowia lipolytica. Metab. Eng.
2016, 38, 139–147. [CrossRef] [PubMed]
23. Elgaali, H.; Hamilton-Kemp, T.; Newman, M.; Collins, R.; Yu, K.; Archbold, D. Comparison of long-chain
alcohols and other volatile compounds emitted from food-borne and related gram positive and gram
negative bacteria. J. Basic Microbiol. 2002, 42, 373–380. [CrossRef]
24. Kohse-Hoinghaus, K.; Osswald, P.; Cool, T.A.; Kasper, T.; Hansen, N.; Qi, F.; Westbrook, C.K.; Westmoreland, P.R.
Biofuel combustion chemistry: From ethanol to biodiesel. Angew. Chem. Int. Ed. Engl. 2010, 49, 3572–3597.
[CrossRef] [PubMed]
25. Siripong, W.; Wolf, P.; Kusumoputri, T.P.; Downes, J.J.; Kocharin, K.; Tanapongpipat, S.; Runguphan, W.
Metabolic engineering of Pichia pastoris for production of isobutanol and isobutyl acetate. Biotechnol. Biofuels.
2018, 11, 1. [CrossRef] [PubMed]
26. Saini, M.; Wang, Z.W.; Chiang, C.J.; Chao, Y.P. Metabolic engineering of Escherichia coli for production of
n-butanol from crude glycerol. Biotechnol. Biofuels. 2017, 10, 173. [CrossRef] [PubMed]
27. Zhang, J.; Zong, W.; Hong, W.; Zhang, Z.T.; Wang, Y. Exploiting endogenous crispr-cas system for multiplex
genome editing in Clostridium tyrobutyricum and engineer the strain for high-level butanol production.
Metab. Eng. 2018, 47, 49–59. [CrossRef] [PubMed]
28. Rahman, Z.; Sung, B.H.; Yi, J.Y.; Bui le, M.; Lee, J.H.; Kim, S.C. Enhanced production of n-alkanes in
Escherichia coli by spatial organization of biosynthetic pathway enzymes. J. Biotechnol. 2014, 192 Pt A, 187–191.
[CrossRef]
29. Srirangan, K.; Liu, X.; Westbrook, A.; Akawi, L.; Pyne, M.E.; Moo-Young, M.; Chou, C.P. Biochemical, genetic,
and metabolic engineering strategies to enhance coproduction of 1-propanol and ethanol in engineered
Escherichia coli. Appl. Microbiol. Biotechnol. 2014, 98, 9499–9515. [CrossRef] [PubMed]
30. Cao, Y.X.; Xiao, W.H.; Zhang, J.L.; Xie, Z.X.; Ding, M.Z.; Yuan, Y.J. Heterologous biosynthesis and
manipulation of alkanes in Escherichia coli. Metab. Eng. 2016, 38, 19–28. [CrossRef] [PubMed]
31. Garcia, H.G.; Brewster, R.C.; Phillips, R. Using synthetic biology to make cells tomorrow’s test tubes.
Integr. Biol. 2016, 8, 431–450. [CrossRef] [PubMed]
32. Smanski, M.J.; Zhou, H.; Claesen, J.; Shen, B.; Fischbach, M.A.; Voigt, C.A. Synthetic biology to access and
expand nature’s chemical diversity. Nat. Rev. Microbiol. 2016, 14, 135–149. [CrossRef] [PubMed]
33. Wu, G.; Yan, Q.; Jones, J.A.; Tang, Y.J.; Fong, S.S.; Koffas, M.A.G. Metabolic burden: Cornerstones in synthetic
biology and metabolic engineering applications. Trends Biotechnol. 2016, 34, 652–664. [CrossRef] [PubMed]
332
Energies 2018, 11, 2663
34. Jensen, P.R.; Hammer, K. Artificial promoters for metabolic optimization. Biotechnol. Bioeng. 1998, 58, 191–195.
[CrossRef]
35. Polli, F.; Meijrink, B.; Bovenberg, R.A.L.; Driessen, A.J.M. New promoters for strain engineering of
Penicillium chrysogenum. Fungal. Genet. Biol. 2016, 89, 62–71. [CrossRef] [PubMed]
36. Lutz, R.; Bujard, H. Independent and tight regulation of transcriptional units in Escherichia coli via the lacr/o,
the tetr/o and arac/i1-i2 regulatory elements. Nucleic. Acids. Res. 1997, 25, 1203–1210. [CrossRef] [PubMed]
37. Ho, K.M.; Lim, B.L. Co-expression of a prophage system and a plasmid system in Bacillus subtilis. Protein Expr.
Purif. 2003, 32, 293–301. [CrossRef] [PubMed]
38. Dehli, T.; Solem, C.; Jensen, P.R. Tunable promoters in synthetic and systems biology. Subcell. Biochem. 2012,
64, 181–201. [PubMed]
39. Song, Y.; Nikoloff, J.M.; Fu, G.; Chen, J.; Li, Q.; Xie, N.; Zheng, P.; Sun, J.; Zhang, D. Promoter screening from
Bacillus subtilis in various conditions hunting for synthetic biology and industrial applications. PLoS ONE
2016, 11. [CrossRef] [PubMed]
40. Rudge, T.J.; Brown, J.R.; Federici, F.; Dalchau, N.; Phillips, A.; Ajioka, J.W.; Haseloff, J. Characterization of
intrinsic properties of promoters. ACS Synth. Biol. 2016, 5, 89–98. [CrossRef] [PubMed]
41. Yang, K.; Stracquadanio, G.; Luo, J.; Boeke, J.D.; Bader, J.S. Biopartsbuilder: A synthetic biology tool for
combinatorial assembly of biological parts. Bioinformatics 2016, 32, 937–939. [CrossRef] [PubMed]
42. Stewart, G.S.; Lubinsky-Mink, S.; Jackson, C.G.; Cassel, A.; Kuhn, J. Phg165: A PBR322 copy number
derivative of puc8 for cloning and expression. Plasmid 1986, 15, 172–181. [CrossRef]
43. Atkinson, M.R.; Savageau, M.A.; Myers, J.T.; Ninfa, A.J. Development of genetic circuitry exhibiting toggle
switch or oscillatory behavior in Escherichia coli. Cell 2003, 113, 597–607. [CrossRef]
44. Lian, J.; Jin, R.; Zhao, H. Construction of plasmids with tunable copy numbers in Saccharomyces cerevisiae and
their applications in pathway optimization and multiplex genome integration. Biotechnol. Bioeng. 2016, 113,
45. Shizuya, H.; Birren, B.; Kim, U.J.; Mancino, V.; Slepak, T.; Tachiiri, Y.; Simon, M. Cloning and stable
maintenance of 300-kilobase-pair fragments of human DNA in Escherichia coli using an f-factor-based vector.
Proc. Natl. Acad. Sci. USA 1992, 89, 8794–8797. [CrossRef] [PubMed]
46. Kasai, Y.; Oshima, K.; Ikeda, F.; Abe, J.; Yoshimitsu, Y.; Harayama, S. Construction of a self-cloning system in
the unicellular green alga Pseudochoricystis ellipsoidea. Biotechnol. Biofuels. 2015, 8, 94. [CrossRef] [PubMed]
47. Sera, T. Zinc-finger-based artificial transcription factors and their applications. Adv. Drug. Deliv. Rev. 2009,
48. Lee, J.Y.; Sung, B.H.; Yu, B.J.; Lee, J.H.; Lee, S.H.; Kim, M.S.; Koob, M.D.; Kim, S.C. Phenotypic engineering
by reprogramming gene transcription using novel artificial transcription factors in Escherichia coli. Nucleic.
Acids. Res. 2008, 36. [CrossRef] [PubMed]
49. Lee, J.Y.; Yang, K.S.; Jang, S.A.; Sung, B.H.; Kim, S.C. Engineering butanol-tolerance in Escherichia coli with
artificial transcription factor libraries. Biotechnol. Bioeng. 2011, 108, 742–749. [CrossRef] [PubMed]
50. Ma, C.; Wei, X.; Sun, C.; Zhang, F.; Xu, J.; Zhao, X.; Bai, F. Improvement of acetic acid tolerance of
Saccharomyces cerevisiae using a zinc-finger-based artificial transcription factor and identification of novel
genes involved in acetic acid tolerance. Appl. Microbiol. Biotechnol. 2015, 99, 2441–2449. [CrossRef] [PubMed]
51. Ochi, K. From microbial differentiation to ribosome engineering. Biosci. Biotechnol. Biochem. 2007, 71,
52. Santos, C.N.; Stephanopoulos, G. Combinatorial engineering of microbes for optimizing cellular phenotype.
Curr. Opin. Chem. Biol. 2008, 12, 168–176. [CrossRef] [PubMed]
53. Tanaka, Y.; Kasahara, K.; Izawa, M.; Ochi, K. Applicability of ribosome engineering to vitamin B12 production
by Propionibacterium shermanii. Biosci. Biotechnol. Biochem. 2017, 81, 1636–1641. [CrossRef] [PubMed]
54. Kurosawa, K.; Hosaka, T.; Tamehiro, N.; Inaoka, T.; Ochi, K. Improvement of alpha-amylase production by
modulation of ribosomal component protein S12 in Bacillus subtilis 168. Appl. Environ. Microbiol. 2006, 72,
55. Hosaka, T.; Xu, J.; Ochi, K. Increased expression of ribosome recycling factor is responsible for the enhanced
protein synthesis during the late growth phase in an antibiotic-overproducing Streptomyces coelicolor
ribosomal rpsl mutant. Mol. Microbiol. 2006, 61, 883–897. [CrossRef] [PubMed]
56. Janosi, L.; Shimizu, I.; Kaji, A. Ribosome recycling factor (ribosome releasing factor) is essential for bacterial
growth. Proc. Natl. Acad. Sci. USA 1994, 91, 4249–4253. [CrossRef] [PubMed]
333
Energies 2018, 11, 2663
57. Chen, M.; Cai, L.; Fang, Z.; Tian, H.; Gao, X.; Yao, W. Site-specific incorporation of unnatural amino acids
into urate oxidase in Escherichia coli. Protein Sci. 2008, 17, 1827–1833. [CrossRef] [PubMed]
58. Rothschild, K.J.; Gite, S. Trna-mediated protein engineering. Curr. Opin. Biotechnol. 1999, 10, 64–70. [CrossRef]
59. Gladyshev, V.N.; Jeang, K.T.; Stadtman, T.C. Selenocysteine, identified as the penultimate c-terminal residue in
human t-cell thioredoxin reductase, corresponds to tga in the human placental gene. Proc. Natl. Acad. Sci. USA
1996, 93, 6146–6151. [CrossRef] [PubMed]
60. Yanagisawa, T.; Ishii, R.; Fukunaga, R.; Kobayashi, T.; Sakamoto, K.; Yokoyama, S. Multistep engineering
of pyrrolysyl-trna synthetase to genetically encode n(epsilon)-(o-azidobenzyloxycarbonyl) lysine for
site-specific protein modification. Chem. Biol. 2008, 15, 1187–1197. [CrossRef] [PubMed]
61. Burke, S.A.; Lo, S.L.; Krzycki, J.A. Clustered genes encoding the methyltransferases of methanogenesis from
monomethylamine. J. Bacteriol. 1998, 180, 3432–3440. [PubMed]
62. Hara, K.Y.; Shimodate, N.; Hirokawa, Y.; Ito, M.; Baba, T.; Mori, H.; Mori, H. Glutathione production by efficient
atp-regenerating Escherichia coli mutants. FEMS Microbiol. Lett. 2009, 297, 217–224. [CrossRef] [PubMed]
63. Wang, Y.; Li, L.; Ma, C.; Gao, C.; Tao, F.; Xu, P. Engineering of cofactor regeneration enhances
(2S,3S)-2,3-butanediol production from diacetyl. Sci. Rep. 2013, 3, 2643. [CrossRef] [PubMed]
64. Geier, M.; Brandner, C.; Strohmeier, G.A.; Hall, M.; Hartner, F.S.; Glieder, A. Engineering Pichia pastoris for
improved NADH regeneration: A novel chassis strain for whole-cell catalysis. Beilstein. J. Org. Chem. 2015,
65. Ishikawa, M.; Tsuchiya, D.; Oyama, T.; Tsunaka, Y.; Morikawa, K. Structural basis for channelling mechanism
of a fatty acid beta-oxidation multienzyme complex. EMBO J. 2004, 23, 2745–2754. [CrossRef] [PubMed]
66. Jorgensen, K.; Rasmussen, A.V.; Morant, M.; Nielsen, A.H.; Bjarnholt, N.; Zagrobelny, M.; Bak, S.; Moller, B.L.
Metabolon formation and metabolic channeling in the biosynthesis of plant natural products. Curr. Opin.
Plant. Biol. 2005, 8, 280–291. [CrossRef] [PubMed]
67. Pfeifer, B.A.; Khosla, C. Biosynthesis of polyketides in heterologous hosts. Microbiol. Mol. Biol. Rev. 2001, 65,
68. Akhtar, M.K.; Dandapani, H.; Thiel, K.; Jones, P.R. Microbial production of 1-octanol: A naturally excreted
biofuel with diesel-like properties. Metab. Eng. Commun. 2015, 2, 1–5. [CrossRef] [PubMed]
69. Martin, V.J.; Pitera, D.J.; Withers, S.T.; Newman, J.D.; Keasling, J.D. Engineering a mevalonate pathway in
Escherichia coli for production of terpenoids. Nat. Biotechnol. 2003, 21, 796–802. [CrossRef] [PubMed]
70. Yu, P.; Chen, X.; Li, P. Enhancing microbial production of biofuels by expanding microbial metabolic
pathways. Biotechnol. Appl. Biochem. 2017, 64, 606–619. [CrossRef] [PubMed]
71. Jones, J.A.; Toparlak, O.D.; Koffas, M.A. Metabolic pathway balancing and its role in the production of
biofuels and chemicals. Curr. Opin. Biotechnol. 2015, 33, 52–59. [CrossRef] [PubMed]
72. Peacock, A.F. Recent advances in designed coiled coils and helical bundles with inorganic prosthetic
groups-from structural to functional applications. Curr. Opin. Chem. Biol. 2016, 31, 160–165. [CrossRef]
[PubMed]
334
energies
Article
Electrochemical Hydrogenation of Acetone to Produce
Isopropanol Using a Polymer Electrolyte
Membrane Reactor
Chen Li 1 , Ashanti M. Sallee 2 , Xiaoyu Zhang 3, * and Sandeep Kumar 1, *
1 Department of Civil & Environmental Engineering, Old Dominion University, Norfolk, VA 23529, USA;
[email protected]
2 Department of Chemistry, Hampton University, Hampton, VA 23668, USA; [email protected]
3 Department of Mechanical & Aerospace Engineering, Old Dominion University, Norfolk, VA 23529, USA
* Correspondence: [email protected] (X.Z.); [email protected] (S.K.); Tel.: +1-757-683-4913 (X.Z.);
+1-757-683-3898 (S.K.)
Received: 17 August 2018; Accepted: 2 October 2018; Published: 10 October 2018
Abstract: Electrochemical hydrogenation (ECH) of acetone is a relatively new method to produce

isopropanol. It provides an alternative way of upgrading bio-fuels with less energy consumption
and chemical waste as compared to conventional methods. In this paper, Polymer Electrolyte
Membrane Fuel Cell (PEMFC) hardware was used as an electrochemical reactor to hydrogenate
acetone to produce isopropanol and diisopropyl ether as a byproduct. High current efficiency (59.7%)
and selectivity (>90%) were achieved, while ECH was carried out in mild conditions (65 ◦ C and
atmospheric pressure). Various operating parameters were evaluated to determine their effects on
the yield of acetone and the overall efficiency of ECH. The results show that an increase in humidity
increased the yield of propanol and the efficiency of ECH. The operating temperature and power
supply, however, have less effect. The degradation of membranes due to contamination of PEMFC
and the mitigation methods were also investigated.
Keywords: acetone; electrochemical hydrogenation; isopropanol; membrane contamination;

polymer electrolyte membrane; relative humidity
1. Introduction
Propanol is an important organic raw material in chemical production, two isomers 1-propanol
and isopropanol are widely used in the paint, medicine and pesticide industries [1]. Compared to
1-propanol, isopropanol has more extensive and important applications. Along with ethanol, n-butanol,
and methanol, isopropanol belongs to the group of alcohol solvents, about 6.4 million tons of which
were utilized worldwide in 2011 [2]. Isopropanol is primarily produced by combining water and
propene in a hydration reaction, through either an indirect or direct process. In an indirect process,
propene reacts with sulfuric acid and forms a mixture of sulfate esters. Subsequent hydrolysis of those
esters by steam produces isopropanol. In a direct hydration process, propene reacts with water or
steam at high pressure (200–300 atm) and high temperatures (230–270 ◦ C), in the presence of solid
or supported acidic catalysts [3,4]. Isopropanol is produced by a direct combination of propene and
water. Both processes require intensive energy input and use of corrosive chemicals.
Thermal hydrogenation of acetone is a relatively new and more advanced method to produce
isopropanol, where acetone is hydrogenated either in the liquid or gas phase over a Raney nickel or
copper and chromium oxide mixture [5]. Compared to the aforementioned conventional methods,
thermal hydrogenation can be carried out at a lower temperature (75 ◦ C) with up to 35% yield rate.
However, an elevated temperature (350–400 ◦ C) is still required to enable fully activated catalysts.
In addition, handling corrosive chemicals remains a problem [6].

Energies 2018, 11, 2691
Electrochemical hydrogenation (ECH) provides a more energy efficient and environment-friendly

method of upgrading organics, by integrating both the electrochemical and catalytic methods [7].
The overall reaction mechanism of ECH of an unsaturated organic molecule is suggested as the
following Equations (1)–(5) [8–13].
Anode side: H2 → 2(H+ ) + 2e− ; H2 O → 1/2O2 + 2(H+ ) + 2e−
Cathode side:
(1) Protons react with electrons and generate M(H)ads (M is an adsorption site):
(H+ ) + e− + M → M(H)ads + H2 O (1)
(2) An organic molecule Y = Z is adsorbed by an adsorption site M:
Y = Z + M → M(Y = Z)ads (2)
(3) M(H)ads reacts with the adsorbed organic molecule:
M(Y = Z) ads + 2M(H)abs → M(YH-ZH)abs + 2M (3)
(4) A hydrogenated product is generated:
M(YH-ZH)abs → YH-ZH + M (4)
(5) H2 gas is also produced:
2M(H)ads → H2 (gas) +2M M(H)ads + (H+ ) + e− → H2 (gas) +M (5)
In the process, chemisorbed hydrogen M(H)ads is generated in situ on the electrocatalyst surface
through either hydrogen pumping or water electrolysis and reacts with the adsorbed organics
(Equation (1)). Note that hydrogenation also competes with hydrogen gas evolution (Equation (5)),
which results in a decrease of the current efficiency. ECH allows the reactions (Equations (2)–(4))
to happen at lower temperatures and ambient pressure. Compared to conventional hydrogenation
methods, ECH mainly uses electrical energy and all the reactions take place in a mild operating
condition (i.e., low temperature and atmospheric pressure). Therefore, intense thermal energy input is
not required in ECH. ECH also uses either water or H2 to supply protons, which eliminate the need for
any reducing agent. ECH can be conducted onsite using fuel cell stacks and renewable power sources
to produce hydrogen enriched compounds. This will avoid or minimize the storage and transportation
of corrosive and hazardous chemicals.
ECH has been widely used to upgrade unsaturated compounds to corresponding saturated
chemicals, such as furfural [14–20], aromatic compounds [7,13,21–29], soybean oil [30], edible oil [31,32],
levulinic acid [33–35], lactic acid [36], acetaldehyde [37], ethanol [37], acetylene [38], bio-oil [39,40],
cyclohexane [9], glucose [41], and lignin [42]. In all those cases, reactions take place under mild
conditions with temperatures below 100 ◦ C and atmospheric pressure. A maximum current efficiency
up to 45%, which is defined as the efficiency of electrogenerated H2 addition to unsaturated bonds,
was reported [30–32,38]. The yield of the electrocatalytic hydrogenation of organic molecules is directly
related to the processes described in Equations (1)–(5), which are determined by the capabilities of
catalysts. Compared to nickel (Ni), copper (Cu) and lead (Pb), activated carbon fibers supported
platinum (Pt) demonstrated the best catalytic activities for upgrading various organics, such as
furfural [20] and acetaldehyde [37]. The electrochemical conversion rate is also affected by the
nature of electrodes [43], current density [44], temperature [26,38,40,45], solvent compositions [9,12,45],
solution pH [9,14,21,40,46], and chemical potential [30,39,40]. The most commonly used ECH reactor
is H-type cell [7,10,14,15,19,20,22,23,27,39,42,44,46]. A typical H-type cell consists of two electrode
chambers, between which a cation exchange membrane is sandwiched. Compared to the traditional
336
Energies 2018, 11, 2691
homogenous electrolyte method of the single chamber cell [14,24,33,34,36,37], the basic anolyte and
the acidic/neutral catholyte method were applied in many H-type cells [7,14,40] to promote the proton
transmission efficiency.
Hydrogenation of unsaturated compounds using polymer electrolyte membrane fuel cell (PEMFC)
reactors was also reported [18,29–32]. PEMFC is a type of low-temperature fuel cell that takes its
name from ion conductive polymer membrane used as the electrolyte [47]. A typical PEMFC assembly
consists of an ion exchange membrane, two electrodes made of carbon layer loaded with Pt, and two
gas diffusion media. The reactor provides gas distribution, current collection, temperature control,
and mechanical support of the PEMFC assembly [47]. Alfonso et al. electrocatalytically hydrogenated
acetophenone by H2 using a PEMFC reactor. They reported the selectivity of produced 1-phenylethanol
around 90% with only methylbenzene and hydrogen as by-products [29]. Green et al. reported that the
main products from ECH of furfural were furfural alcohol (54–100% selectivity) and tetrahydrofurfuryl
alcohol (0–26% selectivity). A higher production rate was achieved by feeding pure hydrogen gas
than that from electrolysis of water [18]. Pintauro et al. and An et al. both studied ECH of soybean
oil in a PEMFC reactor at 60–90 ◦ C and atmospheric pressure [30–32]. Pintauro et al. proved that a
bimetallic cathode (Pd/Co or Pd/Fe) could increase the yield rate of the ECH process [30]. An et al.
proved Pd-black cathode worked significantly better than Pt and the best current efficiency could reach
41% [31,32]. There are many advantages of using a PEMFC reactor for ECH [48]: First, compared to
H-Cell and one chamber cell reactor, the PEMFC reactor has a smaller internal resistance, due to
its highly conductive and thin membrane electrolyte assembly (MEA), resulting in significantly less
electric energy loss. In addition, the energy consumption can be further reduced if protons are supplied
from hydrogen oxidation, rather than electrolysis. Second, since PEMFCs can be easily scaled up by
simply stacking them, the space-time yield of ECH using PEMFC reactors is superior to the other
methods. PEMFC can be easily applied in space limited area such as transportation, stationary,
and portable/micro power generation sectors [49].
In this paper, ECH of acetone to produce isopropanol was demonstrated using a PEMFC reactor at
ambient pressure. Various factors that impact the yield of propanol were investigated, including current
density, temperature, relative humidity (RH), and membrane degradation. The main objective of this
work was to evaluate the appropriate pathways of ECH of bio-oil components using a PEMFC reactor.
2. Materials and Method
2.1. Material and Experimental Setup

The experiments were performed using a standard PEMFC hardware (Scribner Associates Inc.,
USA) with an active area of 25 cm2 . Such standard PEMFC hardware has been widely used for
PEMFC evaluation tests [29,50]. Commercially available MEAs were purchased from Ion Power
Inc., New Castle, DE, USA. The MEAs consist of Nafion® 117 membranes sandwiched with porous
carbon-based electrodes, each of which has a Pt loading of 0.3 mg/cm2 . Micro-porous carbon papers
(SIGRACET® 10BC) were trimmed and used as gas diffusion layers (GDLs) for both electrodes.
Teflon gaskets were used to seal around the assembly. A pair of graphite bipolar plates with flow
patterns were used to distribute flows and enclose the assembly. All temperatures were acquired
via K-type thermocouples (OMEGA, USA). A fuel cell test station (850e, Scribner Associates Inc.,
Southern Pines, NC, USA) was used to control temperatures, flow rates, and humidity. It was also
used for data acquisition.
Ultra-high purity (99.999%) N2 (Airgas, Radnor, PA, USA), and filtered shop air were connected
to the fuel cell test station and supplied to the PEMFC through the purge line. They were used for
purging system and making current and voltage curve, respectively, and were cut off while ECH
experiment was running. Ultra-high purity (99.999%) H2 (Airgas, USA) and Deionized (DI) water
tank were purged through the anode side of the reactor, using the fuel cell test station. H2 was the
electrons donator and deionized (DI) water tank was used for adding humidity to H2 . Acetone (Fisher
337
Energies 2018, 11, 2691
Scientific, Hampton, NH, USA, Certified ACS Reagent Grade) was injected into the cathode side of the
reactor by a syringe pump (MTI Corporation, Richmond, CA, USA, EQ-300sp-LD). A direct current
(DC) power supply (Tektronix, Beaverton, OR, USA, PWS 4205) was used to supply DC power for
ECH. The positive probe was connected to the anode, while the negative probe was connected to the
cathode. While ECH was running, electrons were deprived of H2 and transferred to the cathode trough
the DC power supply. Acetone was the electron acceptor and reacted with the produced protons.
The products from the cathode outlet were collected by airbags (Tedlar® ), which were held by water
bath. Room temperature water bath was used for condensing the unreacted acetone and products.
The cell temperature was controlled by the fuel cell test station. The exhaust gases and byproducts
from the anode were condensed in a knock-out bottle prior to venting out. The schematic diagram of
the whole experimental setup is shown in Figure 1.
Figure 1. Schematic diagram of the experimental setup.
2.2. Electrochemical Hydrogenation and Characterization

All the PEMFCs were assembled in accordance with the standard assembly procedure of fuel
cell hardware [51]. Prior to each ECH experiment, a new membrane was conditioned based on the
standard protocol [51] for at least 24 h, until it reached a fully functional state. Current–voltage (I-V)
sweeps were performed to establish the baseline data for the following ECH experiments.
In each ECH experiment, the acetone and H2 flow rates were controlled at 6 mL/h and 0.25 slpm,
respectively. The voltage was consistent for each reaction; the current was recorded every 5 min.
Four different factors, namely cell temperature, RH, supplied voltage, and membrane degradation,
were investigated to identify the optimized operating conditions. The operating conditions were in the
range of 55–80 ◦ C, 35–90% and 0.01–0.02 V, respectively. I-V scans were conducted before and after
ECH to characterize the membrane degradation.
The flow chart of the ECH experimental operation is shown in Figure 2. Initially, three collection
methods, including the dead end, partially confined and an open end were evaluated to identify
the best means for accurate collection of products. In the partially confined method, products were
collected by airbags with pressure relief valves, which prevents pressure buildup while trapping
the products. In the other two collection methods, although the products were condensed by room
temperature water bath, dead end still resulted in too much back pressure accumulation, whereas open
end failed to collect enough products for analysis. Therefore, partially confined airbags were used for
all the experiments. The pressure differential of two electrodes can be controlled by airbags confined
338
Energies 2018, 11, 2691
extent. Green et al. proved that a suitable pressure differential between the anode and cathode could
decrease cross-over and increase conversion [52]. Dadda et al. believed that water transport in the
membrane of a PEMFC is influenced by a convective force, resulting from a pressure gradient [53].
Many researchers also point out the importance of the flow pressure [54].
Figure 2. Flowchart of the experiment.
The collected samples were analyzed by gas chromatography–mass spectroscopy (Shimadzu,

GC-MS-QP2010 SE), using a Shimadzu (SH-Rxi-5Sii) MS column (length: 30 m, inner diameter:
0.25 mm). Volume quantitative analysis was conducted by another gas chromatograph (SRI Instrument,
8610C), using a Restek (MXT-WAX) column (length: 30 m, inner diameter: 0.53 mm). Helium was used
as a carrier gas for both gas chromatographs.
3.1. Product Characterization

Both the liquid and gaseous products were analyzed by GC-MS. Unreacted acetone (C3 H6 O),
isopropanol (C3 H8 O), and diisopropyl ether (C6 H14 O) were detected in the liquid products.
Unreacted acetone, diisopropyl ether, and isopropanol were found in the gaseous and liquid products.
The reactions on both electrodes are catalyzed by Pt. While applying a DC voltage, protons formed
at the anode are electrochemically pumped to the cathode. The protons then react with acetone to
produce isopropanol and diisopropyl ether. The most feasible reaction pathways at the cathode are
shown below:
2H + + 2e− → 2Hads
Pt
(6)
(7)
(8)
339
Energies 2018, 11, 2691
Electrosorbed hydrogen is formed on Pt surface by reduction of H+ (Equation (6), where Hads is the
electrosorbed hydrogen). Hydrogenation of the C=O bond then proceeds as in catalytic hydrogenation
through the reaction of the acetone with the electrosorbed hydrogen (Equation (7)). As a result,
isopropanol, which is the main product, is generated. Two isopropanol molecules may also combine
and free one water, generating diisoproply ether as a byproduct is formed (Equation (8)). Note that
hydrogen gas can also be regenerated, which is an unfavorable electrochemical reaction during ECH.
Hydrogen regeneration reduces the efficiency by electrochemically pumping useless hydrogen through
the MEA, resulting in a reduction of the yield rate of products.
Each MEA had undergone at least three ECH experiments before replacement. To minimize the
impacts of MEA degradation on ECH experiments, the results of the I-V scans, which were carried
out prior to the experiments, were compared to the baseline performance of each MEA. If the I-V
curve demonstrated an obvious deflection from the baseline performance, then the MEA needed to
be replaced. The components detected in the products were isopropanol, diisopropyl ether, acetone,
and water, with their volumetric percentages ranging 12–16%, 1–2%, 69–75%, and 11–14%, respectively.
Note that abundant acetone was supplied to the cathode to prevent fuel starvation. As a result,
the maximum conversion rate of acetone to isopropanol was 23%. The selectivity of isopropanol was
calculated more than 90%. Acetone was also detected in the anode due to crossover, which is discussed
below. The produced isopropanol can be easily separated from the mixture using extraction and
distillation, which are two widely adopted methods in the industry [55] and therefore not discussed in
detail here.
Three different control parameters, including RH, operating temperature, and input voltage,
were assessed to identify the optimized operating conditions for ECH of acetone. All operating
parameters used in the experiments are shown in the Table S1 in the Supplementary Materials.
RH is a very important parameter that affects the performance of PEMFCs [54,56–60]. Figure 3a
shows a typical impact of RH on the product yield during ECH of acetone. It was obviously evidenced
that higher humidity promoted higher yield of isopropanol. The composition of isopropanol in the
products increased from 4.9% to 16.1%, while RH climbed from 35% to 90%. The reason humidity
had such a significant impact is that the MEA usually uses a perfluorosulfonic acid membrane
(e.g., Nafion® ) as the electrolyte. A high or nearly saturated humidity (RH > 80%) is usually required to
obtain practical performance because the conductivity of perfluorosulfonic acid membranes depends
on the water content. Higher humidity means higher conductivity, consequently resulting in better
performance [61–64]. Water management is critical for PEMFC operation. Sufficient water must be
absorbed into the membrane to ionize the acid groups, whereas excess water can cause flooding issues
and thus diminishing the performance [52]. The inlet RH of the electrodes must be controlled to
prevent both membranes from drying out and electrode flooding. Although better performance is
usually obtained by increasing RH, excess moisture may result in water flooding that hinders gas
transport [65]. In the present experiments, RH was maintained between 35% and 90% to prevent either
water starvation or over saturation. In fact, many researchers have investigated the mechanism of
humidity influence. It is generally believed that RH can impact electro-osmotic drag, water diffusion,
membrane ionic conductivity, and water back diffusion flux in the MEAs, which consequently influence
the performance [66,67]. Elevated RH can greatly improve the PEMFC performance, through increasing
the membrane conductivity [68,69], the catalyst activities [68,70], the electrode kinetics [71,72], and the
mass transfer rates [73,74].
340
Energies 2018, 11, 2691
Figure 3. Isopropanol yield (percentage) change as a result of: (a) RH; and (b) temperature. H2 and
acetone flow rates were 0.25 L/min and 6 mL/h, respectively. (a) The ECH experiments were conducted
five times at 65 ◦ C. Each ECH experiment was repeated twice. (b) The ECH experiments were conducted
four times at 80% RH.
Operating temperature is the second factor that was assessed in this study. Generally, the temperature
was found to have a slightly positive impact on the product yield. As shown in Figure 3b, the isopropanol
yield percentage of products varied between 9.3% and 14.3%, when the temperature increased from
55 ◦ C to 80 ◦ C. However, the temperature seemed to have minimal effect on the total efficiency.
The results agreed with the findings from the literature. Singh et al. [25] investigated ECH of phenol
by Pt accordance with increasing temperature. They believed that dehydrogenated phenol adsorbents
easily block the active sites of Pt at higher temperatures. The ECH efficiency was claimed to be directly
correlated with the adsorption properties of acetone, hydrogen, and propanol onto the Pt/C catalyst.
Murillo and Chen [75] used temperature programmed desorption (TPD) to monitor the desorption and
decomposition property of propanol in a wide temperature range on the Pt surface. According to their
research, propanol decomposition peaked at 65 ◦ C and 117 ◦ C. In the present research, the operating
temperature ranged from 55 ◦ C to 80 ◦ C, between which propanol decomposition could happen at a
higher temperature (>80 ◦ C). Decomposition of propanol was believed to cause the decrease of its yield.
Therefore, increasing the operating temperature in the range does not necessarily result in an increase in
the product yield.
Finally, the influence of applied voltage on the product yield was also investigated. Generally,
the input voltage has no obvious impact on the product yield. In the experiments, the voltage ranged
from 10 mV to 20 mV, with 5 mV increments. At 10 mV and 15 mV, the yield of isopropanol was 15.9%
and 17.0%, respectively. Diisopropyl ether was not detected in either case. However, when the input
voltage was increased to 0.02 V, the volumetric percentage of isopropanol produced was up to 16%,
and about 1% diisopropyl ether was detected.
3.2. System Analysis

Selectivity, H2 utilization, and current efficiency were selected to evaluate the hydrogenation efficiency.
Selectivity represents the yield of desirable products. As the major product, higher isopropanol selectivity
was pursued. The selectivity is calculated based on the following equation [14,20], where acetone unreacted
is excluded:
Moles o f Desired Product
Selectivity = × 100% (9)
Total Moles o f Products
In the present research, H2 was supplied to the anode to produce protons for ECH reactions
on the cathode. Due to gas diffusion resistance, gas crossover, and hydrogen regeneration on the
cathode, some H2 was wasted. The H2 utilization is directly related to the overall ECH efficiency.
Higher H2 utilization percentage is desired since more hydrogen will be involved in the ECH process.
341
Energies 2018, 11, 2691
The actual amount of H2 used to produce isopropanol can be derived from the amount of product.
The H2 utilization is calculated by the following equation:
Atomic Hydrogen Used f or Faraday Current

H2 Utilization = × 100% (10)
Total Atomic Hydrogen Supplied
During the ECH process, acetone reacts with M(H)ads to produce isopropanol and byproducts on
the cathode. Concurrently, H2 regeneration happens and is an unfavorable process simply because it
wastes energy. The H2 regeneration reaction is affected by supplied voltage, temperature, humidity,
and catalyst. Hereby, current efficiency (shown below) is used as an important parameter to determine
how efficient H2 is used for the ECH process [14,20].
Current used f or the ECH process

Current E f f iciency = × 100% (11)
Total Faraday current
The total efficiency is defined by the H2 utilization multiplying the current efficiency, as shown in
the following equation:
Total E f f iciency = H2 Utilization × Current E f f iciency (12)
Figure 4a shows the impact of temperature on H+ utilization, current efficiency, and total
efficiency in a typical set of experiments. As the operating temperature increased from 50 ◦ C to
80 ◦ C, H+ utilization increased from 1.5% to 6.0%, whereas current efficiency decreased from 28% to
18.5%. Temperature affects H+ utilization and current efficiency differently. Higher ionic mobility
and catalytic activities are achieved with higher operating temperatures, resulting in higher H2
utilization. Consequently, the electrochemical conversion and reaction rates increase with elevated
temperatures [52]. However, Figure 4a indicates that, although elevated operating temperature enabled
more hydrogen being involved in reactions, the yield of products did not increase or even decreased.
That resulted in a loss of current efficiency, which means most extra protons produced were somehow
wasted. The conclusion can also be evidenced by the curve of the total efficiency, which remained
almost flat. Note that the total efficiency was low because abundant H2 was supplied to the system
to minimize the impact of fuel starvation and gas diffusion resistance. Practically, stoichiometric
flow can be fed to the system based upon the actual current. In their experiments of ECH of
acetone, Sara et al. observed that the current efficiency increased while the cell temperature increased
from 25 ◦ C to 50 ◦ C, which seems to contradict our results [52]. However, in the present research,
PEMFCs were operated in a recommended range between 50 ◦ C and 80 ◦ C to achieve the best
performance. The reduction of current efficiency is believed mainly due to propanol decomposition,
as mentioned in Section 3.1. Another minor reason was acetone vaporization, since the boiling point of
acetone is 56 ◦ C. Acetone gasification might have negative impact on the hydrogenation reactions on
the cathode, due to increased pressure and thus higher diffusion resistance.
342
Energies 2018, 11, 2691
Figure 4. H2 Utilization, Current Efficiency, and Total efficiency calculated at different: (a) temperatures;
and (b) RH. H2 and acetone flow rates were 0.25 L/min and 6 mL/h, respectively. (a) The ECH
experiments were conducted five times at 80% RH; (b) The ECH experiments were conducted five
times at 65 ◦ C. Each ECH experiment was repeated twice.
Figure 4b shows the impact of RH on H2 utilization, current efficiency, and total efficiency.
The operating temperature was set at 65 ◦ C, and the RH was controlled by setting the humidifier’s
temperature. It is seen that higher RH resulted in better efficiencies. As the RH ranged from 35%
to 90%, the H2 utilization increased from 0.9% to 2.8%, and the total efficiency increased from 0.4%
to 1.5%. A sudden spike of the current efficiency was observed when ramping the RH from 80% to
91%. It can be concluded that higher RH is favorable for ECH of acetone and will result in a higher
yield of products. It has been proved that higher ionic conductivity can be achieved when MEAs
become more hydrated [68–70,76–79]. Practically, high RH is required to maintain the best fuel cell
performance. The higher water content in the Nafion membrane will ease proton transport, i.e., reduce
ionic resistance. As a result, more protons can be created and transported to the cathode for ECH.
Typically, >80% RH is recommended [61–64], which explains why a spike in the current efficiency was
observed when the RH surpassed 80%.
The present research shows that the performance of ECH of acetone is correlated with RH,
input voltage, and temperature, in which RH has the most obvious effect. It is suggested that
the optimized operating conditions are RH of 80% or more, the input voltage of 0.02 V or less,
and temperature of between 50 ◦ C and 55 ◦ C. The obtained maximum H2 utilization and maximum
current efficiency achieved in the present experiments were 5.9% and 59.7%, respectively. To further
increase those efficiencies, stoichiometric flow control is strongly recommended.
3.3. MEA Degradation

Long-term durability is one important factor that affects the practical applications of ECH using
PEMFC reactors. Nowadays, commercial MEAs are fairly durable for their common roles as the
power sources. The ECH process, however, involves organics that may contaminate MEAs and thus
shorten their lifetime. To our best knowledge, very limited research has been conducted to evaluate
the impacts of contaminants on the durability of ECH. In the past decade, extensive research has
been carried out on mitigating contamination of PEMFCs from impurities, including CO, CO2 , H2 S,
NOx , SOx , and hydrocarbons [80,81]. Impurities may contaminate one or more components of the
MEA, resulting in performance degradation. Three major contamination effects were identified as the
poisoning of the electrode catalysts, a decrease of the ionic conductivity, and an increase of the mass
transfer resistance.
Additionally, the crossover is another factor that negatively impacts the PEMFC performance.
Crossover of organic compounds during hydrogenation using PEMFCs has been reported [82,83].
One immediate drawback is the loss of fuel and/or products, which decreases the efficiency.
343
Energies 2018, 11, 2691
Furthermore, contaminants not only poison just one electrode but also may crossover and further
poison the catalyst on the other electrode [12].
To investigate the impacts of MEA degradation on the present hydrogenation tests, polarization
scans (V-I sweeps) were performed after each test [51]. The black curves in Figure 5 are the baseline
data recorded for the fresh MEAs prior to ECH tests. After each test, pure N2 was purged for at
least 10 h to remove all the temporary contaminants. The effects of RH and temperature on the MEA
degradation were also evaluated.
Figure 5. V-I scans performed during three sets of ECH experiments: (a) The ECH experiments were
conducted five times at 65 ◦ C and 80% RH; (b) the ECH experiments were conducted four times at
65 ◦ C but different RH (65%, 65%, 50% and 50%, respectively); (c) the ECH experiments were conducted
five times at 80% RH but different temperatures (80 ◦ C, 73 ◦ C, 57 ◦ C, and 50 ◦ C, respectively). The black
curves are the baseline data recorded for the fresh MEAs prior to ECH tests. After each test, pure N2
was purged for at least 10 h to remove all the temporary contaminants.
Figure 5a shows the results of five sets of V-I measurements performed on a PEMFC,
which underwent five 10-h ECH experiments. Both the ECH experiments and V-I measurements
shown in Figure 5a were conducted at 65 ◦ C with 80% RH. It clearly shows that the MEA performance
degraded as more ECH tests were conducted, especially after the third ECH experiment. The open
circuit voltage (OCV) dropped a lot starting from the fourth V-I scan, which indicates that crossover
became significant. It implied that pinholes might form due to degradation.
Figure 5b shows the effect of RH on the MEA degradation during ECH. For the ECH experiments
conducted in Figure 5b, RH was reduced to 65% for the first and second tests and was further reduced
to 50% for the third and fourth tests. To compare with the same baseline data (the black curve in
Figure 5b), all V-I scans were performed using the same operating conditions that have been used
for the baseline scan. The results illustrate that reducing RH was able to mitigate the degradation to
some extent. It is believed that less RH resulted in less mass transport via the MEA, which eventually
344
Energies 2018, 11, 2691
extended the lifetime of the catalysts. However, permanent damage to the MEA still existed, as seen
from the general trend of V-I scans. Similar to the observations in Figure 5a, purging with pure N2
could not remove permanent contaminants.
Finally, the effect of operating temperature on the MEA degradation was investigated, as shown
in Figure 5c. For the ECH experiments conducted in Figure 5c, the operating temperatures were 80 ◦ C,
73 ◦ C, 57 ◦ C, and 50 ◦ C, while maintaining the same RH. Again, all V-I scans were performed using the
same operating conditions as those used for the baseline scan (black curve in Figure 5c). The results
show that temperature variation has no observable impact on MEA degradation. In other words,
changing the operating temperature did not mitigate degradation.
Figure 6 shows that the trends of current efficiency, H2 utilization, and total efficiency using the
same MEA for several ECH experiments. Figure 6 shows that, as the MEA degraded, current efficiency,
H2 utilization, and total efficiency all decreased. Until a method of contamination mitigation is found,
it is acceptable to use one MEA three times.
Figure 6. H2 Utilization, Current Efficiency and Total efficiency as a function of ECH experiment times.
The ECH experiments were conducted four times at 65 ◦ C and 80% RH. H2 and acetone flow rates
were 0.25 L/min and 6 mL/h, respectively. After each test, pure N2 was purged for at least 10 h to
remove all the temporary contaminants. Each ECH experiment was repeated twice.
Usually, the lifespan of a PEMFC under steady-state operation can be very long, up to thousands
of hours [84–89]. However, catalyst contamination is the major factor that diminishes the PEMFC
performance and very likely results in significant degradation [90–92]. In fact, many organic compounds
can contaminate the MEA. Those compounds include acetaldehyde, toluene, propane, vinyl acetate,
methyl methacrylate, acetonitrile, dichloromethane, acetylene, chlorobenzene, formic acid, methanol,
ethanol, phenol, butane, acetone, and naphthalene [93–95], and the list is expanding. The main reason
that so many contaminants were found is that the catalysts used in common MEAs are Pt-based. Pt is a
premium catalyst, but also sensitive to so many contaminants. The MEAs used in the present research
contain pure Pt as the catalyst. Although developing non-Pt catalyst is beyond the scope of the present
research, to further conduct durable ECH experiments, MEAs with contamination tolerant catalysts need
to be used.
Reactant and product crossing over is another possible reason that caused the MEA degradation.
Liquids that contained mainly acetone were detected at the anode side during the ECH experiments.
Those liquids not only decrease the fuel utilization but also further contaminate the anode catalyst.
Feasible solutions to this issue include adopting thicker MEAs, feeding gaseous feedstock instead of
liquid, and using non-Pt catalyst [96].
345
Energies 2018, 11, 2691
In summary, to minimize the MEA contamination using the current setup, keeping low RH is
suggested. To solve the problem essentially, novel non-Pt catalysts need to be developed, such as Pd-
and Ni-based catalysts [97,98]. Even though a wide range of metals can be used as electrocatalysts at
the cathode, those with the strong hydrogen absorption capability are desired.
4. Conclusions
Electrochemical hydrogenation of acetone using a PEMFC reactor was successfully demonstrated
in the present research. The results proved that ECH can be a feasible way of hydrogenating acetone
to produce isopropanol in mild conditions. In the experiments, the main product obtained was
isopropanol with a selectivity of approximately 90%. A small amount (about 1%) diisopropyl ether
was also obtained as a byproduct. The mild operation conditions, including low temperature and
ambient pressure, are the greatest advantages of the proposed ECH method. The present research
suggests that the optimized conditions for ECH of acetone using a PEMFC reactor include an operating
temperature around 65 ◦ C and relatively high RH.
Contamination impact using the PEMFC reactor during ECH was also investigated. It was
concluded that organic compounds can contaminate the MEAs, resulting in serious degradation.
However, methods to mitigate contamination are limited. The present research only demonstrated that
lower RH could help reduce contamination. Eventually, novel non-Pt catalysts need to be developed
for durable ECH process.
Supplementary Materials: The following are available online at http://www.mdpi.com/1996-1073/11/10/2691/s1,

Table S1: Operating Parameters of the Experiments.
Author Contributions: Conceptualization, C.L. and S.K.; Methodology, X.Z.; Software, C.L. and A.M.S.;
Validation, C.L., X.Z. and S.K.; Formal Analysis, C.L.; Investigation, C.L. and A.M.S.; Resources, X.Z. and
S.K.; Data Curation, C.L.; Writing—Original Draft Preparation, C.L.; Writing—Review and Editing, X.Z. and S.K.;
Visualization, C.L.; Supervision, S.K.; Project Administration, X.Z.; and Funding Acquisition, X.Z.
Funding: This research was supported by the United States National Science Foundation (#1560194) and the Old
Dominion University.
Acknowledgments: The authors would like to acknowledge the technical assistance from Can Zhou, Qiang Tang,
and Yazdanshenas Elias who helped with the experiment set up. The authors are also grateful to Ali Teymouri,
Caleb Talbot, Adams Kameron, Alexander Asiedu, and Samaratunga Ashani for technical discussion and
sample analysis.
References
1. World Health Organization. Guide to Local Production: WHO-Recommended Handrub Formulations. 2011,
pp. 1123–1156. Available online: http://www.who.int/gpsc/1125may/Guide_to_Local_Production.pdf
(accessed on 2 October 2018).
2. Ceresana Inc. Market Study Solvents, 4th ed. 2012, p. 10. Available online: https://www.ceresana.com/en/
market-studies/chemicals/solvents/ (accessed on 2 October 2018).
3. Alt, C. Ullmann’s Encyclopedia of Industrial Chemistry, 1th ed.; Wiley-VCH: New York, NY, USA, 2006;
pp. 2345–2351.
4. Onoue, Y.; Mizutani, Y.; Akiyama, S.; Izumi, Y.; Ihara, H. Isopropyl alcohol by direct hydration of propylene.
Bull. Jpn. Pet. Inst. 1973, 15, 50–55. [CrossRef]
5. John, E.L.; Loke, R.A. Isopropyl Alcohol. In Kirk-Othmer Encyclopedia of Chemical Technology, 5th ed.; John, E.,
Logsdon, L.R.A., Eds.; Wiley-VCH: New York, NY, USA, 2000; Volume 6, pp. 98–100. ISBN 9780471238966.
6. Yurieva, T.M.; Plyasova, L.M.; Makarova, O.V.; Krieger, T.A. Mechanisms for hydrogenation of acetone to
isopropanol and of carbon oxides to methanol over copper-containing oxide catalysts. J. Mol. Catal. A Chem.
1996, 113, 455–468. [CrossRef]
7. Cyr, A.; Chiltz, F.; Jeanson, P.; Martel, A.; Brossard, L.; Lessard, J.; Ménard, H. Electrocatalytic hydrogenation
of lignin models at Raney nickel and palladium-based electrodes. Can. J. Chem. 2000, 78, 307–315. [CrossRef]
346
Energies 2018, 11, 2691
8. Bockris, J. Electrochemical Processing. In Comprehensive Treatise of Electrochemistry, 1th ed.; Bockris, J.,
Brian, E.C., Yeager, E., White, R.E., Eds.; Springer Science & Business Media: New York, NY, USA, 2013;
Volume 2, pp. 537–616. ISBN 1468437852.
9. Cirtiu, C.M.; Brisach-Wittmeyer, A.; Ménard, H. Electrocatalysis over Pd catalysts: A very efficient alternative
to catalytic hydrogenation of cyclohexanone. J. Catal. 2007, 245, 191–197. [CrossRef]
10. Amouzegar, K.; Savadogo, O. Electrocatalytic hydrogenation of phenol on dispersed Pt: Reaction mechanism
and support effect. Electrochim. Acta 1998, 43, 503–508. [CrossRef]
11. St-Pierre, G.; Chagnes, A.; Bouchard, N.-A.; Harvey, P.D.; Brossard, L.; Ménard, H. Rational design of
original materials for the electrocatalytic hydrogenation reactions: Concept, preparation, characterization,
and theoretical analysis. Langmuir 2004, 20, 6365–6373. [CrossRef] [PubMed]
12. Cirtiu, C.M.; Hassani, H.O.; Bouchard, N.-A.; Rowntree, P.A.; Ménard, H. Modification of the surface
adsorption properties of alumina-supported Pd catalysts for the electrocatalytic hydrogenation of phenol.
Langmuir 2006, 22, 6414–6421. [CrossRef] [PubMed]
13. Song, Y.; Chia, S.H.; Sanyal, U.; Gutiérrez, O.Y.; Lercher, J.A. Integrated catalytic and electrocatalytic
conversion of substituted phenols and diaryl ethers. J. Catal. 2016, 344, 263–272. [CrossRef]
14. Li, Z.; Kelkar, S.; Lam, C.H.; Luczek, K.; Jackson, J.E.; Miller, D.J.; Saffron, C.M. Aqueous electrocatalytic
hydrogenation of furfural using a sacrificial anode. Electrochim. Acta 2012, 64, 87–93. [CrossRef]
15. Chu, D.; Hou, Y.; He, J.; Xu, M.; Wang, Y.; Wang, S.; Wang, J.; Zha, L. Nano TiO2 film electrode for
electrocatalytic reduction of furfural in ionic liquids. J. Nanopart. Res. 2009, 11, 1805–1809. [CrossRef]
16. Parpot, P.; Bettencourt, A.; Chamoulaud, G.; Kokoh, K.; Belgsir, E. Electrochemical investigations of the
oxidation–reduction of furfural in aqueous medium: Application to electrosynthesis. Electrochim. Acta 2004,
49, 397–403. [CrossRef]
17. Chamoulaud, G.; Floner, D.; Moinet, C.; Lamy, C.; Belgsir, E. Biomass conversion II: Simultaneous
electrosyntheses of furoic acid and furfuryl alcohol on modified graphite felt electrodes. Electrochim. Acta
2001, 46, 2757–2760. [CrossRef]
18. Green, S.K.; Lee, J.; Kim, H.J.; Tompsett, G.A.; Kim, W.B.; Huber, G.W. The electrocatalytic hydrogenation
of furanic compounds in a continuous electrocatalytic membrane reactor. Green Chem. 2013, 15, 1869–1879.
[CrossRef]
19. Nilges, P.; Schroder, U. Electrochemistry for biofuel generation: Production of furans by electrocatalytic
hydrogenation of furfurals. Energy Environ. Sci. 2013, 6, 2925–2931. [CrossRef]
20. Zhao, B.; Chen, M.; Guo, Q.; Fu, Y. Electrocatalytic hydrogenation of furfural to furfuryl alcohol using
platinum supported on activated carbon fibers. Electrochim. Acta 2014, 135, 139–146. [CrossRef]
21. Ilikti, H.; Rekik, N.; Thomalla, M. Electrocatalytic hydrogenation of alkyl-substituted phenols in aqueous
solutions at a Raney nickel electrode in the presence of a non-micelle-forming cationic surfactant.
J. Appl. Electrochem. 2004, 34, 127–136. [CrossRef]
22. Laplante, F.; Brossard, L.; Ménard, H. Considerations about phenol electrohydrogenation on electrodes made
with reticulated vitreous carbon cathode. Can. J. Chem. 2003, 81, 258–264. [CrossRef]
23. Song, Y.; Gutiérrez, O.Y.; Herranz, J.; Lercher, J.A. Aqueous phase electrocatalysis and thermal catalysis for
the hydrogenation of phenol at mild conditions. Appl. Catal. B Environ. 2016, 182, 236–246. [CrossRef]
24. Santana, D.S.; Lima, M.V.F.; Daniel, J.R.R.; Navarro, M. Electrocatalytic hydrogenation of organic compounds
using current density gradient and sacrificial anode of nickel. Tetrahedron Lett. 2003, 44, 4725–4727. [CrossRef]
25. Singh, N.; Song, Y.; Gutiérrez, O.Y.; Camaioni, D.M.; Campbell, C.T.; Lercher, J.A. Electrocatalytic
Hydrogenation of Phenol over Platinum and Rhodium: Unexpected Temperature Effects Resolved. ACS Catal.
2016, 6, 7466–7470. [CrossRef]
26. Tountian, D.; Brisach-Wittmeyer, A.; Nkeng, P.; Poillerat, G.; Ménard, H. Effect of support conductivity
of catalytic powder on electrocatalytic hydrogenation of phenol. J. Appl. Electrochem. 2009, 39, 411–419.
[CrossRef]
27. Ilikti, H.; Rekik, N.; Thomalla, M. Electrocatalytic hydrogenation of phenol in aqueous solutions at a Raney
nickel electrode in the presence of cationic surfactants. J. Appl. Electrochem. 2002, 32, 603–609. [CrossRef]
28. Tsyganok, A.; Holt, C.M.; Murphy, S.; Mitlin, D.; Gray, M.R. Electrocatalytic hydrogenation of aromatic
compounds in ionic liquid solutions over WS2-on-glassy carbon and Raney nickel cathodes. Fuel 2012, 93,
347
Energies 2018, 11, 2691
29. Sáez, A.; García-García, V.; Solla-Gullón, J.; Aldaz, A.; Montiel, V. Electrocatalytic hydrogenation of
acetophenone using a Polymer Electrolyte Membrane Electrochemical Reactor. Electrochim. Acta 2013,
91, 69–74. [CrossRef]
30. Pintauro, P.N.; Gil, M.P.; Warner, K.; List, G.; Neff, W. Electrochemical Hydrogenation of Soybean Oil with
Hydrogen Gas. Ind. Eng. Chem. Res. 2005, 44, 6188–6195. [CrossRef]
31. An, W.; Hong, J.K.; Pintauro, P.N.; Warner, K.; Neff, W. The electrochemical hydrogenation of edible oils
in a solid polymer electrolyte reactor. II. Hydrogenation selectivity studies. J. Am. Oil Chem. Soc. 1999, 76,
32. An, W.; Hong, J.K.; Pintauro, P.N.; Warner, K.; Neff, W. The electrochemical hydrogenation of edible oils in a
solid polymer electrolyte reactor. I. Reactor design and operation. J. Am. Oil Chem. Soc. 1998, 75, 917–925.
[CrossRef]
33. Xin, L.; Zhang, Z.; Qi, J.; Chadderdon, D.J.; Qiu, Y.; Warsko, K.M.; Li, W. Electricity Storage in Biofuels:
Selective Electrocatalytic Reduction of Levulinic Acid to Valeric Acid or γ-Valerolactone. ChemSusChem 2013,
34. Qiu, Y.; Xin, L.; Chadderdon, D.J.; Qi, J.; Liang, C.; Li, W. Integrated electrocatalytic processing of levulinic
acid and formic acid to produce biofuel intermediate valeric acid. Green Chem. 2014, 16, 1305–1315. [CrossRef]
35. Nilges, P.; dos Santos, T.R.; Harnisch, F.; Schroder, U. Electrochemistry for biofuel generation: Electrochemical
conversion of levulinic acid to octane. Energy Environ. Sci. 2012, 5, 5231–5235. [CrossRef]
36. Dalavoy, T.S.; Jackson, J.E.; Swain, G.M.; Miller, D.J.; Li, J.; Lipkowski, J. Mild electrocatalytic hydrogenation
of lactic acid to lactaldehyde and propylene glycol. J. Catal. 2007, 246, 15–28. [CrossRef]
37. Lai, S.C.; Koper, M.T. Electro-oxidation of ethanol and acetaldehyde on platinum single-crystal electrodes.
Faraday Discuss. 2009, 140, 399–416. [CrossRef]
38. Huang, B.; Durante, C.; Isse, A.A.; Gennaro, A. Highly selective electrochemical hydrogenation of acetylene
to ethylene at Ag and Cu cathodes. Electrochem. Commun. 2013, 34, 90–93. [CrossRef]
39. Li, Z.; Kelkar, S.; Raycraft, L.; Garedew, M.; Jackson, J.E.; Miller, D.J.; Saffron, C.M. A mild approach for
bio-oil stabilization and upgrading: Electrocatalytic hydrogenation using ruthenium supported on activated
carbon cloth. Green Chem. 2014, 16, 844–852. [CrossRef]
40. Li, Z.; Garedew, M.; Lam, C.H.; Jackson, J.E.; Miller, D.J.; Saffron, C.M. Mild electrocatalytic hydrogenation
and hydrodeoxygenation of bio-oil derived phenolic compounds using ruthenium supported on activated
carbon cloth. Green Chem. 2012, 14, 2540–2549. [CrossRef]
41. Kwon, Y.; Koper, M. Electrocatalytic hydrogenation and deoxygenation of glucose on solid metal electrodes.
ChemSusChem 2013, 6, 455–462. [CrossRef] [PubMed]
42. Lam, C.H.; Lowe, C.B.; Li, Z.; Longe, K.N.; Rayburn, J.T.; Caldwell, M.A.; Houdek, C.E.; Maguire, J.B.;
Saffron, C.M.; Miller, D.J. Electrocatalytic upgrading of model lignin monomers with earth abundant metal
electrodes. Green Chem. 2015, 17, 601–609. [CrossRef]
43. Moutet, J.-C. Electrocatalytic hydrogenation on hydrogen-active electrodes. A review. Org. Prep. Proced. Int.
1992, 24, 309–325. [CrossRef]
44. Mahdavi, B.; Chapuzet, J.M.; Lessard, J. The electrocatalytic hydrogenation of phenanthrene at Raney nickel
electrodes: The effect of periodic current control. Electrochim. Acta 1993, 38, 1377–1380. [CrossRef]
45. Lipkowski, J.; Ross, P.N. Electrocatalysis, 1th ed.; John Wiley&Sons: New York, NY, USA, 1998; pp. 75–290.
ISBN 0471246735.
46. Robin, D.; Comtois, M.; Martel, A.; Lemieux, R.; Cheong, A.K.; Belot, G.; Lessard, J. The electrocatalytic
hydrogenation of fused poly cyclic aromatic compounds at Raney nickel electrodes: The influence of catalyst
activation and electrolysis conditions. Can. J. Chem. 1990, 68, 1218–1227. [CrossRef]
47. Garraín, D.; Lechón, Y.; de la Rúa, C. Polymer electrolyte membrane fuel cells (PEMFC) in automotive applications:
Environmental relevance of the manufacturing stage. Smart Grid Renew. Energy 2011, 2, 68. [CrossRef]
48. Montiel, V.; Sáez, A.; Expósito, E.; García-García, V.; Aldaz, A. Use of MEA technology in the synthesis of
pharmaceutical compounds: The electrosynthesis of N-acetyl-l-cysteine. Electrochem. Commun. 2010, 12,
49. Wang, Y.; Chen, K.S.; Mishler, J.; Cho, S.C.; Adroher, X.C. A review of polymer electrolyte membrane
fuel cells: Technology, applications, and needs on fundamental research. Appl. Energy 2011, 88, 981–1007.
[CrossRef]
348
Energies 2018, 11, 2691
50. Zhang, X.; Pasaogullari, U.; Molter, T. Influence of ammonia on membrane-electrode assemblies in polymer
electrolyte fuel cells. Int. J. Hydrogen Energy 2009, 34, 9188–9194. [CrossRef]
51. Florida Solar Energy Center, Test Protocol for Cell Performance Tests Performed Under DOE. 2009; pp. 19–20.
Available online: https://www.energy.gov/sites/prod/files/2014/03/f10/htmwg_may09_pem_single_
cell_testing.pdf (accessed on 8 October 2018).
52. Green, S.K.; Tompsett, G.A.; Kim, H.J.; Kim, W.B.; Huber, G.W. Electrocatalytic Reduction of Acetone in
a Proton-Exchange-Membrane Reactor: A Model Reaction for the Electrocatalytic Reduction of Biomass.
ChemSusChem 2012, 5, 2410–2420. [CrossRef] [PubMed]
53. Dadda, B.; Abboudi, S.; Zarrit, R.; Ghezal, A. Heat and mass transfer influence on potential variation in a
PEMFC membrane. Int. J. Hydrogen Energy 2014, 39, 15238–15245. [CrossRef]
54. Zong, Y.; Zhou, B.; Sobiesiak, A. Water and thermal management in a single PEM fuel cell with non-uniform
stack temperature. J. Power Sources 2006, 161, 143–159. [CrossRef]
55. Berg, L. Separation of Acetone from Isopropanol-Water Mixtures by Extractive Distillation. US 5897750, 1999.
Available online: https://patents.justia.com/patent/5897750 (accessed on 8 October 2018).
56. Jian, Q.; Ma, G.; Qiu, X. Influences of gas relative humidity on the temperature of membrane in PEMFC with
interdigitated flow field. Renew. Energy 2014, 62, 129–136. [CrossRef]
57. Anantaraman, A.; Gardner, C. Studies on ion-exchange membranes. Part 1. Effect of humidity on the
conductivity of Nafion® . J. Electroanal. Chem. 1996, 414, 115–120. [CrossRef]
58. Sumner, J.; Creager, S.; Ma, J.; DesMarteau, D. Proton conductivity in Nafion® 117 and in a novel bis
[(perfluoroalkyl) sulfonyl] imide ionomer membrane. J. Electrochem. Soc. 1998, 145, 107–110. [CrossRef]
59. Sone, Y.; Ekdunge, P.; Simonsson, D. Proton conductivity of Nafion 117 as measured by a four-electrode AC
impedance method. J. Electrochem. Soc. 1996, 143, 1254–1259. [CrossRef]
60. Yuan, W.; Tang, Y.; Pan, M.; Li, Z.; Tang, B. Model prediction of effects of operating parameters on proton
exchange membrane fuel cell performance. Renew. Energy 2010, 35, 656–666. [CrossRef]
61. Wong, K.; Loo, K.; Lai, Y.; Tan, S.-C.; Chi, K.T. A theoretical study of inlet relative humidity control in PEM
fuel cell. Int. J. Hydrogen Energy 2011, 36, 11871–11885. [CrossRef]
62. Lee, C.-I.; Chu, H.-S. Effects of cathode humidification on the gas–liquid interface location in a PEM fuel cell.
63. Guvelioglu, G.H.; Stenger, H.G. Flow rate and humidification effects on a PEM fuel cell performance and
operation. J. Power Sources 2007, 163, 882–891. [CrossRef]
64. Zhang, J.; Tang, Y.; Song, C.; Xia, Z.; Li, H.; Wang, H.; Zhang, J. PEM fuel cell relative humidity (RH) and its
effect on performance at high temperatures. Electrochim. Acta 2008, 53, 5315–5321. [CrossRef]
65. Jeon, D.H.; Kim, K.N.; Baek, S.M.; Nam, J.H. The effect of relative humidity of the cathode on the performance
and the uniformity of PEM fuel cells. Int. J. Hydrogen Energy 2011, 36, 12499–12511. [CrossRef]
66. Ikeda, T.; Koido, T.; Tsushima, S.; Hirai, S. MRI investigation of water transport mechanism in a membrane
under elevated temperature condition with relative humidity and current density variation. ECS Trans. 2008,
16, 1035–1040.
67. Misran, E.; Hassan, N.S.M.; Daud, W.R.W.; Majlan, E.H.; Rosli, M.I. Water transport characteristics of a
PEM fuel cell at various operating pressures and temperatures. Int. J. Hydrogen Energy 2013, 38, 9401–9408.
[CrossRef]
68. Knights, S.D.; Colbow, K.M.; St-Pierre, J.; Wilkinson, D.P. Aging mechanisms and lifetime of PEFC and
DMFC. J. Power Sources 2004, 127, 127–134. [CrossRef]
69. Abe, T.; Shima, H.; Watanabe, K.; Ito, Y. Study of PEFCs by AC Impedance, Current Interrupt, and Dew Point
Measurements I. Effect of Humidity in Oxygen Gas. J. Electrochem. Soc. 2004, 151, A101–A105. [CrossRef]
70. Saleh, M.M.; Okajima, T.; Hayase, M.; Kitamura, F.; Ohsaka, T. Exploring the effects of symmetrical and
asymmetrical relative humidity on the performance of H2 /air PEM fuel cell at different temperatures.
71. Xu, H.; Song, Y.; Kunz, H.R.; Fenton, J.M. Effect of elevated temperature and reduced relative humidity on
ORR kinetics for PEM fuel cells. J. Electrochem. Soc. 2005, 152, A1828–A1836. [CrossRef]
72. Uribe, F.A.; Springer, T.E.; Gottesfeld, S. A microelectrode study of oxygen reduction at the
platinum/recast—Nafion film interface. J. Electrochem. Soc. 1992, 139, 765–773. [CrossRef]
73. Jang, J.-H.; Yan, W.-M.; Li, H.-Y.; Chou, Y.-C. Humidity of reactant fuel on the cell performance of PEM fuel
cell with baffle-blocked flow field designs. J. Power Sources 2006, 159, 468–477. [CrossRef]
349
Energies 2018, 11, 2691
74. Broka, K.; Ekdunge, P. Oxygen and hydrogen permeation properties and water uptake of Nafion® 117
membrane and recast film for PEM fuel cell. J. Appl. Electrochem. 1997, 27, 117–123. [CrossRef]
75. Murillo, L.E.; Chen, J.G. Adsorption and reaction of propanal, 2-propenol and 1-propanol on Ni/Pt (111)
bimetallic surfaces. Surf. Sci. 2008, 602, 2412–2420. [CrossRef]
76. Takalloo, P.K.; Nia, E.S.; Ghazikhani, M. Numerical and experimental investigation on effects of inlet
humidity and fuel flow rate and oxidant on the performance on polymer fuel cell. Energy Convers. Manag.
2016, 114, 290–302. [CrossRef]
77. Jiang, R.; Kunz, H.R.; Fenton, J.M. Investigation of membrane property and fuel cell behavior with sulfonated
poly(ether ether ketone) electrolyte: Temperature and relative humidity effects. J. Power Sources 2005, 150,
78. Xu, H.; Kunz, H.R.; Fenton, J.M. Analysis of proton exchange membrane fuel cell polarization losses at
elevated temperature 120 ◦ C and reduced relative humidity. Electrochim. Acta 2007, 52, 3525–3533. [CrossRef]
79. Jiang, R.; Kunz, H.R.; Fenton, J.M. Influence of temperature and relative humidity on performance and CO
tolerance of PEM fuel cells with Nafion® –Teflon® –Zr (HPO4 ) 2 higher temperature composite membranes.
80. Wang, X.; Baker, P.; Zhang, X.; Garces, H.F.; Bonville, L.J.; Pasaogullari, U.; Molter, T.M. An experimental
overview of the effects of hydrogen impurities on polymer electrolyte membrane fuel cell performance. Int. J.
Hydrogen Energy 2014, 39, 19701–19713. [CrossRef]
81. Cheng, X.; Shi, Z.; Glass, N.; Zhang, L.; Zhang, J.; Song, D.; Liu, Z.-S.; Wang, H.; Shen, J. A review of PEM
hydrogen fuel cell contamination: Impacts, mechanisms, and mitigation. J. Power Sources 2007, 165, 739–756.
[CrossRef]
82. Benziger, J.; Nehlsen, J. A polymer electrolyte hydrogen pump hydrogenation reactor. Ind. Eng. Chem. Res.
2010, 49, 11052–11060. [CrossRef]
83. Lundin, M.D.; McCready, M.J. High pressure anode operation of direct methanol fuel cells for carbon dioxide
management. J. Power Sources 2011, 196, 5583–5590. [CrossRef]
84. Scholta, J.; Berg, N.; Wilde, P.; Jörissen, L.; Garche, J. Development and performance of a 10 kW PEMFC
stack. J. Power Sources 2004, 127, 206–212. [CrossRef]
85. Cleghorn, S.; Mayfield, D.; Moore, D.; Moore, J.; Rusch, G.; Sherman, T.; Sisofo, N.; Beuscher, U. A polymer
electrolyte fuel cell life test: 3 years of continuous operation. J. Power Sources 2006, 158, 446–454. [CrossRef]
86. St-Pierre, J.; Wilkinsor, D.; Knights, S.; Bos, M. Relationships between water management, contamination
and lifetime degradation in PEFC. J. New Mater. Electrochem. Syst. 2000, 3, 99–106.
87. Satyapal, S. Hydrogen & Fuel Cells Program Overview. In Proceedings of the Annual Merit Review and
Peer Evaluation Meeting, Arlington, VA, USA, 7–11 June 2011.
88. Wipke, K.; Sprik, S.; Kurtz, J.; Ramsden, T.; Ainscough, C.; Saur, G. VII. 1 Controlled Hydrogen Fleet and
Infrastructure Analysis. In Proceedings of the DOE Annual Merit Review and Peer Evaluation Meeting,
Washington, DC, USA, 13–17 May 2012.
89. Wu, J.; Yuan, X.Z.; Martin, J.J.; Wang, H.; Zhang, J.; Shen, J.; Wu, S.; Merida, W. A review of PEM fuel
cell durability: Degradation mechanisms and mitigation strategies. J. Power Sources 2008, 184, 104–119.
[CrossRef]
90. Ferreira, P.; Shao-Horn, Y.; Morgan, D.; Makharia, R.; Kocha, S.; Gasteiger, H. Instability of Pt/C
electrocatalysts in proton exchange membrane fuel cells a mechanistic investigation. J. Electrochem. Soc. 2005,
152, A2256–A2271. [CrossRef]
91. Xie, J.; Wood, D.L.; More, K.L.; Atanassov, P.; Borup, R.L. Microstructural changes of membrane electrode
assemblies during PEFC durability testing at high humidity conditions. J. Electrochem. Soc. 2005, 152,
A1011–A1020. [CrossRef]
92. Antolini, E. Formation, microstructural characteristics and stability of carbon supported platinum catalysts
for low temperature fuel cells. J. Mater. Sci. 2003, 38, 2995–3005. [CrossRef]
93. Murugan, A.; Brown, A.S. Review of purity analysis methods for performing quality assurance of fuel cell
hydrogen. Int. J. Hydrogen Energy 2015, 40, 4219–4233. [CrossRef]
94. St-Pierre, J.; Angelo, M.; Zhai, Y. Focusing research by developing performance related selection criteria for
PEMFC contaminants. ECS Trans. 2011, 41, 279–286.
95. St-Pierre, J.; Zhai, Y.; Angelo, M.S. Effect of selected airborne contaminants on PEMFC performance.
J. Electrochem. Soc. 2014, 161, F280–F290. [CrossRef]
350
Energies 2018, 11, 2691
96. Fukuzumi, S.; Lee, Y.M.; Nam, W. Mechanisms of two-electron versus four-electron reduction of dioxygen
catalyzed by earth—Abundant metal complexes. ChemCatChem 2018, 10, 9–28. [CrossRef]
97. Chen, Y.; Hwang, C.; Liaw, C. One-step synthesis of methyl isobutyl ketone from acetone with calcined
Mg/Al hydrotalcite-supported palladium or nickel catalysts. Appl. Catal. A Gen. 1998, 169, 207–214.
[CrossRef]
98. Nikolopoulos, A.; Jang, B.-L.; Spivey, J. Acetone condensation and selective hydrogenation to MIBK on Pd
and Pt hydrotalcite-derived MgAl mixed oxide catalysts. Appl. Catal. A Gen. 2005, 296, 128–136. [CrossRef]
351
energies
Article
A Systematic Multivariate Analysis of Carica papaya
Biodiesel Blends and Their Interactive Effect
on Performance
Mohammad Anwar 1, *, Mohammad G. Rasul 1 and Nanjappa Ashwath 2
1 School of Engineering and Technology, Central Queensland University, North Rockhampton, QLD 4702,
Australia; [email protected]
2 School of Health, Medical and Applied Sciences, Central Queensland University, North Rockhampton,
QLD 4702, Australia; [email protected]
Received: 11 October 2018; Accepted: 25 October 2018; Published: 26 October 2018
Abstract: This paper investigates the interactive relationship between three operating parameters
(papaya seed oil (PSO) biodiesel blends, engine load, and engine speed) and four responses
(brake power, BP; torque; brake specific fuel consumption, BSFC; and, brake thermal efficiency, BTE)
for engine testing. A fully instrumented four cylinder four-stroke, naturally aspirated agricultural
diesel engine was used for all experiments. Three different blends: B5 (5% PSO biodiesel +
95% diesel), B10 (10% PSO biodiesel + 90% diesel), and B20 (20% PSO biodiesel + 80% diesel)
were tested. Physicochemical properties of these blends and pure PSO biodiesel were characterised,
and the engine’s performance characteristics were analysed. The results of the engine performance
experiments showed that, in comparison with diesel, the three PSO biodiesel blends caused a slight
reduction in BP, torque, and BTE, and an increase in BSFC. The analysis of variance and quadratic
regression modelling showed that both load and speed were the most important parameters that
affect engine performance, while PSO biodiesel blends had a significant effect on BSFC.
Keywords: diesel; Carica papaya; engine performance; biodiesel; characterisation
1. Introduction
Diesel fuel has successfully contributed to all sectors of human life, especially to transportation,
industry, and agricultural sectors due to its availability, reliability, adaptability, higher combustion
efficiency, and excellent handling/storage properties. However, fossil reserves (oil, gas, and coal)
are limited, therefore, researchers have been exploring an alternative source for diesel for many
years, and, over the last decade, biofuel (i.e., mainly biodiesel) has drawn massive attention for
its excellent environmental and sustainability attributes [1,2]. Biodiesel is biodegradable, non-toxic,
non-explosive, non-flammable, renewable, and an environmentally friendly (produces fewer emissions)
fuel. Although the biodiesel energy content is about 10–12% less than diesel, it can be mixed with
diesel at specific proportions to make the blended fuel properties close to diesel [3,4]. Furthermore,
researchers are working hard to find suitable sustainable biodiesels and their blends, which can be
used as fuel in unmodified diesel engines.
Numerous researchers have explored the production of biodiesel from different feedstocks
focusing on non-edible oil, but very few investigations [5–10] have been performed on papaya seed oil
(PSO) biodiesel. Among these, most of the articles deal with only the biodiesel production process when
using different catalysts and methanol: oil molar ratios. Asokan et al. [2] examined engine performance
in a single cylinder diesel engine fuelled with a mixture of papaya seed oil and watermelon seed oil
and found that B20 performed close to diesel. Prabhakaran et al. [11] analysed engine performance

Energies 2018, 11, 2931
of a single cylinder Kirloskar-TV1 diesel engine with PSO biodiesel blends and found B25 blend has
lower BSFC than all other blends. Sundar Raj and Karthikayan [12] did diesel engine (single cylinder)
performance analysis with PSO-diesel blends with/without additives and found PSO-diesel blends
with additives have better combustion and emission characteristics. However, no other literature has
been found on engine performance analysis of a fully instrumented four cylinder four-stroke diesel
engine fuelled with PSO biodiesel.
PSO is non-edible and converting the waste product (seeds) of the fruit into biodiesel is a
sensible option. Carica papaya was originally native to the tropics of the Americas and it is now
primarily grown in tropical-subtropical climates of Asia (in particular, India with 42% of world
production), South America (mainly Brazil), Africa (Nigeria and Congo), and Polynesia [13,14].
Therefore, the evaluation of papaya seed oil as biodiesel feedstock will contribute to the development of
regional communities and their overall economy. In this study, papaya seed oil was used as feedstock
to produce biodiesel and investigate its suitability as an alternative fuel source. Determining the
fuel properties of PSO and its various blends with diesel were also undertaken. Finally, interactive
relationships between several operating parameters (PSO biodiesel blends, load, and speed) and
engine performance are evaluated and discussed.
2. Materials and Method
2.1. Materials
The raw PSO was obtained from a supplier in Eumundi, Queensland, Australia. The chemicals
that were utilised for this study were methanol (99.9% purity), potassium hydroxide (KOH pellets,
99% purity), and sodium hydroxide (NaOH pellets, 99% purity). All chemicals were of analytical
reagent grade (AR) and they were procured from the School of Engineering and Technology, Central
Queensland University. A reflux condenser and a thermocouple fitted on 0.5 L and 1 L three-neck
laboratory reactors were used for the PSO biodiesel conversion experiments.
2.2. Equipment List

The PSO properties were characterised for density, viscosity, refractive index, angular rotation,
acid value, oxidation stability, and iodine value. These properties were measured before the biodiesel
conversion and optimisation process. Table 1 summarises the equipment used in this study to measure
the properties of PSO, PSO biodiesel, and the biodiesel-diesel blends, and the relevant standards
applied. The properties studied were density at 15 ◦ C, kinematic viscosity at 40 ◦ C, acid value, calorific
value, flash point, and oxidation stability. A gas chromatograph was used to determine the fatty acid
compositions by EN 14103.
Table 1. Equipment used for measuring properties of papaya seed oil (PSO) and related products in
this study.
Property Equipment Standard Applied Accuracy

Kinematic viscosity NVB classic (Normalab, France) ASTM D445 ±0.01 mm2 /s
DM40 LiquiPhysicsTM density meter
Density ASTM D1298 ±0.1 kg/m3
(Mettler Toledo, Switzerland)
NPM 440 Pensky-Martens flash point tester
Flash point ASTM D93 ±0.1 ◦ C
(Normalab, France)
Automation titration Rondo 20 ±0.001 mg
Acid value ASTM D664
(Mettler Toledo, Switzerland) KOH/g
6100EF semi-auto bomb calorimeter
Calorific value ASTM D240 ±0.001 MJ/kg
(Perr, USA)
Oxidation stability at 110 ◦ C 873 Rancimat (Metrohm, Switzerland) ASTM D2274 ±0.01 h
Refractive index RM 40 Refractometer - ±0.0001
353
Energies 2018, 11, 2931
2.3. Biodiesel Production

The acid value of raw PSO was determined to be 0.98 mg KOH/g. After transesterification
using the KOH catalyst, the acid value of PSO biodiesel was found to be 0.42 mg KOH/g. This trial
experiment suggested that only the single-stage alkyl catalyst transesterification process was necessary
for satisfactory PSO biodiesel production. Further, only a few researchers [6–10] have found that the
single-stage transesterification process was sufficient for PSO biodiesel conversion. Based on these
researches and the trial experiment, it was decided to use single-stage alkaline transesterification for
biodiesel production. Thus, in each experiment, the conversion was performed by reacting a known
quantity of PSO with methanol and the catalyst.
In each of these experiments, PSO was poured into a three-neck laboratory reactor and was heated
to the preferred temperature. The catalyst (KOH) was dissolved in methanol in a separate beaker, and
a magnetic stirrer was used to provide vigorous agitation at 600 rpm for 10 min at 50 ◦ C to produce
methoxide. This solution was then poured into the reactor that contained heated PSO to allow the
transesterification reaction at a constant speed of agitation of 600 rpm. Other reaction conditions,
such as temperature and time, were adjusted as per the individual experimental designs determined
from the optimisation process of PSO biodiesel [14]. At the end of the transesterification reaction, the
blend was poured into a separating funnel and left to settle 24 h for layer separation. Under gravity,
two distinct liquid phases were formed; the upper layer was methyl ester (biodiesel), and the bottom
dark brown layer consisted of glycerol and impurities. The glycerol and impurities were drawn off,
whence the methyl ester layer was collected and washed with warm (50 ◦ C) distilled water. The methyl
ester was then heated to 110 ◦ C for 15 min to remove any water that would have been retained from
the washing process. A Whatman® qualitative Grade 1 filter paper was used to filter the methyl ester
(i.e., biodiesel) and was stored at laboratory temperature for characterisation. Figure 1 shows the PSO
production process.
Figure 1. PSO biodiesel production.
354
Energies 2018, 11, 2931
PSO biodiesel yield was calculated using Equation (1).
weight of PSO biodiesel

PSO Biodiesel Yield = FAME percent(from GC analysis) × (1)
weight of PSO
2.4. Fatty Acid Composition

Fatty acids with a double bond are known as unsaturated fatty acids, while those without a
double bond are called saturated fatty acids. High unsaturated fatty acid levels make biodiesel prone
to autoxidation [14]. However, high unsaturated fatty acid levels do ensure good flow properties
as compared with saturated fatty acids. The major drawback of high saturated fatty acid levels in
biodiesel is its poor fuel filterability, particularly in cold weather conditions. The PSO biodiesel fatty
acid composition was determined using a gas chromatograph (GC), (Thermo Trace 1310 GC) according
to EN14103. Table 2 shows the operating condition details of the GC.
Table 2. Gas chromatograph (GC) operating conditions.
Property Details
Brand, model Thermo Scientific Trace 1310 GC
Carrier gas Helium
Flow rates Air: 350 mL/min, H2 : 35 mL/min, N2 : 30 mL/min
Detector temperature 240 ◦ C
Column dimensions 60 m × 250 μm × 0.25 μm
Column head pressure 23.8 psi
Split injector, 40:1 ratio, split flow 48 mL/min,
Injector
constant flow 1.2 mL/min, 1 μL injection volume
Temperature ramp 1 110 ◦ C hold for 4 min
Temperature ramp 2 10 ◦ C/min to 230 ◦ C, hold for 3 min
The compositional analysis that is detailed in Table 3 shows that PSO biodiesel contains a high
level (87.5%) of unsaturated fatty acid methyl esters (FAME) made up of polyunsaturated fatty acid
(PUFA) at 39.03% and monounsaturated fatty acid (MUFA) at 48.49%. Amongst these, the dominant
fatty acids were found to be oleic acid (C18:1) at 47.7% and linoleic acid (C18:2) at 37.3%. The saturated
fatty acids included palmitic acid (C16:0) at 6%.
Table 3. PSO biodiesel fatty acid composition.
Fatty Acids Formula Molecular Weight Structure wt%

Palmitic C16 H32 O2 256 16:0 6.07
Stearic C18 H36 O2 284 18:0 3.13
Oleic C18 H34 O2 282 18:1 47.73
Linoleic C18 H32 O2 280 18:2 37.25
Linolenic C18 H30 O2 278 18:3 1.78
Eicosenoic C20 H38 O2 310 20:1 0.76
Behenic C22 H44 O2 340 22:0 0.68
Erucic C22 H42 O2 338 22:1 1.51
Others 1.09
Total Saturated Fatty Acids (SFA) 9.88
Total Monounsaturated Fatty Acids (MUFA) 48.49
Total Polyunsaturated Fatty Acids (PUFA) 39.03
Degree of Unsaturation (DU) 126.55
Long Chain Saturated Factor (LCSF) 3.19
355
Energies 2018, 11, 2931
2.5. Properties Analysis

The properties of the produced PSO methyl ester (i.e., pure biodiesel or B100) are compared with
those of diesel (B0) and PSO biodiesel-diesel blends (B5, B10, and B20) in Table 4. The properties
and qualities of PSO biodiesel and the diesel blends comply with the requirements of international
standards ASTM D6751 and EN14214. The USA’s ASTM D6751 ensures that the parameters of any
pure biodiesel (B100) satisfy the standard before being used as a blend with diesel or pure fuel, whereas
the European Union’s EN14214 defines the minimum standards for FAME [14,15]. The PSO biodiesel
was found to comply with both standards.
Table 4. Comparison of PSO biodiesel and blends with diesel.
ASTM
Properties Units PSO B100 B20 B10 B5 B0 EN14214-03
D6751-2
Density Kg/m3 885 840 829.76 828.48 827.84 827.2 870–890 860–900
Viscosity at 40 ◦ C mm2 /s 27.3 3.53 3.29 3.26 3.25 3.23 1.9–6.0 3.5–5.0
Acid value mg KOH/g 0.98 0.42 0.12 0.09 0.07 0.05 max. 1 0.5 max. 0.5
Cetane number (CN) - - 48.29 48.06 48.03 48.01 48.00 min. 2 47 min. 51
Calorific value MJ/kg - 38.49 43.94 44.62 44.96 45.30 - 35.00
Flash point (◦ C) - 112 77.20 72.85 70.68 68.50 min. 93 >120
Iodine value (IV) - 79.95 115.89 53.82 46.06 42.18 38.30 - max. 120
Oxidation stability (OS) Hour 77.97 5.61 32.32 35.66 37.33 39 min. 3 min. 6
1 max. = maximum; 2 min. = minimum.
2.6. Fourier Transform Infrared (FTIR) Analysis

The various functional groups that are present in the pure PSO biodiesel sample were determined
with FTIR spectroscopy. In this study, a Spectrum 100 series FTIR spectrometer with a universal
Attenuated total reflectance (ATR) sampling accessory was used. Samples of the PSO biodiesel
were inserted directly on the ATR window to record the spectra over the frequency range of
4000–650 cm−1 with four scans at approximately 40% transmission. Spectrum analysis program
(Spectrum version 6.2.0, Perkin-Elmer Life and Analytical Sciences, Bridgeport, CT, USA) was used to
acquire data and for processing. Figure 2 shows the FTIR spectrum of the pure PSO biodiesel (B100).
Figure 2. PSO methyl ester investigation Fourier Transform Infrared (FTIR) spectrum.
356
Energies 2018, 11, 2931
2.7. Blending of Biodiesel

The pure PSO biodiesel (B100) was mixed with pure diesel (B0) to prepare various blends.
Those blends comprised homogeneous mixtures of 5% vol. of biodiesel with 95% vol. of diesel denoted
by B5, 10% vol. of biodiesel with 90% vol. of diesel indicated by B10 and 20% vol. of biodiesel with
80% vol. of diesel meant by B20. The blends were prepared in a 2 L flask and were agitated at 2000 rpm
with a magnetic stirrer for 60 min.
2.8. Experimental Setup for Engine Testing

This experimental investigation was carried out using pure diesel (B0) and PSO biodiesel-diesel
blends (B5, B10, and B20) in an agricultural tractor four-stroke diesel engine with four cylinders
(model: Kubota V3300) coupled with an eddy current dynamometer. Table 5 details the engine
specification. This tractor engine was used for testing PSO biodiesel-diesel blends with regard to
engine performance and characteristics of exhaust emissions. The engine performance and emissions
data were collected under full load (100%) condition, keeping the throttle 100% wide open and varying
speeds in the range from 1200 to 2400 rpm with an increment of 200 rpm. A photograph of the engine
and test bed and a schematic of the setup of the engine along with the data acquisition system are
shown in Figure 3. The engine is connected with the test bed along with instrumentation consoles
that measure engine speed, torque, air, and fuel consumption and temperature. A Dynolog data
acquisition system was used to convert the console-measured data to a display in the computer
monitor. The exhaust gas emissions of NOx and HC in ppm, and CO and CO2 in vol. %, were
measured with a CODA 5 gas analyser (CODA Products Pty Ltd., Hamilton NSW 2303, Australia).
For measuring Particulate Matter (PM) emissions, a PM meter (MPM-4M) (MAHA Maschinenbau
Haldenwang GmbH & Co. KG, Haldenwang, Germany) was used. Table 6 shows the specification of
the gas analyser and PM meter. Before taking any data from the diesel engine, it was run with pure
diesel (B0) for 20 min at full load to ensure that the engine was warmed up. The biodiesel-diesel blend
(B5, B10, and B20) was then fuelled into the engine for analysis and data acquisition. At the end of
any test or experiment with blended fuel, the engine was again flushed out with pure diesel to clear
the fuel line of blended fuel and the injection system. All of the tests were repeated three times to
minimise any possible error.
There are a large number of studies available in literatures on different biodiesels, and their
engine performance and emission characteristics, but a very few researchers have analysed the engine
performance or emission characteristics in terms of multivariate analysis. This study focuses on the
multivariate analysis of PSO biodiesel blends and their interactive effect on engine performance.
Table 5. Details of the test engine.
Property Apparatus, Model

Engine Model Kubota V3300 Indirect injection
Type Vertical, 4 cycle liquid cooled diesel
Number of cylinders 4
Total displacement (L) 3.318
Bore × Stroke (mm) 98 × 110
Combustion system Spherical type (E-TVCS) (three vortex combustion system)
Intake system Natural aspired
Rated power output (KW/rpm) 53.9/2600
Rated torque (Nm/rpm) 230/1400
Compression ratio 22.6
Emissions certification Tier 2
357
Energies 2018, 11, 2931
(a) Photograph of test bed and diesel engine
(b) Schematic of test procedure setup
Figure 3. Test engine set up.
358
Energies 2018, 11, 2931
Table 6. Exhaust gas analyser, PM meter specification, and error analysis.
Measurement
Measured Gas
Range Resolution Accuracy
HC 0–30,000 ppm (n-hexane) 1 ppm ±1 ppm
CO 0–15% 0.001% ±0.02%
CO2 0–20% 0.001% ±0.3%
NOx 0–5000 ppm 1 ppm ±1 ppm
Meter Particle Size Particle Concentration Range Resolution
Particulate Matter <100 nm to >10 μm 0.1 to >700 mg/m3 ±0.1 mg/m3
Measurements Accuracy Relative Uncertainty (%) Average Reading for B0
BP ±0.41 kW 0.0105 39.20
BSFC ±5 g/kWh 0.0195 256
3. Results and discussion
3.1. Characterisation of PSO Biodiesel-Diesel Blends

While the engine performance experiments in this study investigated B5, B10, and B20 biodiesel
blends, the characteristics of a much broader series of PSO biodiesel-diesel blends has been investigated.
A total of 10 PSO biodiesel-diesel blends (from B5-B90) were prepared, and their individual fuel
properties of density, viscosity, flash point, calorific value, and oxidation stability have been analysed
and are presented in Figure 4a–e.
Density is one of the vital properties of biodiesel that affects fuel atomisation efficiency in an airless
combustion system [16]. According to relevant ASTM and EN standards, the density of biodiesel at
15 ◦ C should be in the range of 860–900 kg/m3 . The densities of B5 (827.84 kg/m3 ), B10 (828.48 kg/m3 ),
and B20 (829.76 kg/m3 ) were very close to that of B0 diesel (827.20 kg/m3 ). Generally, biodiesel has a
slightly higher density than diesel. In addition, density increased as the percentage of PSO biodiesel in
the blends increased (Figure 4a).
Some vegetable oils have a higher viscosity, which causes poor fuel flow. Raw PSO has a viscosity
of 27.3 mm2 /s, which is 8–9 times higher than diesel (B0). However, after the transesterification
process, the viscosity of PSO drops to an acceptable limit. Generally, the viscosity of biodiesel ranges
from 1.9–6.0 mm2 /s, and all PSO biodiesel-diesel blends in Figure 4b fulfill this requirement. However,
the viscosity of B5 (3.25 mm2 /s), B10 (3.26 mm2 /s), B20 (3.29 mm2 /s), and B30 (3.32 mm2 /s) were
very close to B0 (3.23 mm2 /s). Therefore, those blends (B10–B30) can be used as a diesel engine fuel
without any modifications to the engine.
According to ASTM D6751, the flashpoint should be 100–170 ◦ C, and all of the biodiesel-diesel
blends met that requirement as well. The flash point of B0 was recorded as 68.50 ◦ C, whereas B5, B10,
and B20 were found to be 70.68 ◦ C, 72.85 ◦ C, and 77.20 ◦ C, respectively. Biodiesel tends to a higher
flash point than diesel, and Figure 4c shows the pattern of the increased flash point with increased
biodiesel blends. The more the flashpoint is above 66 ◦ C is indicative of a safer fuel with better storage
ability [1,17]. From that perspective, all PSO biodiesel-diesel blends can be stored safely and they can
be used to fuel a diesel engine without any modifications.
Calorific value is another important property in the selection of any fuel. Figure 4d shows that
pure PSO biodiesel and all of its blends have slightly lower calorific values than diesel although
they are within the requirements of international standards. Biodiesel has nearly 10% more oxygen
content than diesel, and less hydrogen-carbon content hence will produce less thermal energy [17].
Calorific values of B5 (44.96 MJ/kg), B10 (44.62 MJ/kg), and B20 (43.94 MJ/kg) blends are very close
to B0 (45.30 MJ/kg).
Another vital fuel property is oxidation stability (OS), as it indicates the degree of oxidation that
occurs during prolonged storage. Hasni et al. [18] mentioned that lower oxidation stability could
adversely affect fuel quality. The higher the oxidation stability, the better the fuel quality. As the PSO
359
Energies 2018, 11, 2931
biodiesel (unsaturated fatty acid) percentage increases, the OS decreases, which is shown in Figure 4e.
As per European standard EN 590, the minimum OS of any fuel should be 20 h (indicated by the red
line on the Figure 4e). PSO biodiesel-diesel blends B5 to B50 meet that standard. B5 has an OS value of
37.33 h and B50 is 22.31 h, while B0 has an OS value of 39 h. According to ASTM D6751 and EN14214,
the minimum OS values are 3 and 6 h respectively. This study found the 100% PSO biodiesel has an
OS value of 5.61 h, which falls within the ASTM and close to EN standards.
(a) (b)
(c) (d)
(e)
Figure 4. PSO biodiesel-diesel blending effects on: (a) density, (b) kinematic viscosity, (c) flash point,
(d) calorific value, and (e) oxidation stability.
360
Energies 2018, 11, 2931
3.2. Analysis of Engine Performance
3.2.1. Brake Power (BP)

The brake power (BP) outputs from the diesel engine fuelled with diesel (B0) and PSO B5, B10,
and B20 blends are shown in Figure 5a. It can be seen that the BP gradually increases with the increase
of engine speed. Maximum BP was observed at 2400 rpm and was 45.9 kW, 45.4 kW, 45.08 kW, and
48.34 kW for B5, B10, B20, and B0, respectively. Several factors, such as calorific value and viscosity, can
have the effect of uneven combustion that results in lowering the BP value. In this study, the viscosity
of diesel (B0) and PSO biodiesel-diesel blends were very close. Only the difference in calorific values
of blends and diesel caused the variations in BP values. Average BP value reductions for B5, B10, and
B20 in comparison with B0 were 2.88%, 3.87%, and 5.13%, respectively.
(a)
(b)
Figure 5. Cont.
361
Energies 2018, 11, 2931
(c)
(d)
Figure 5. Variation of: (a) brake power, (b) torque, (c) brake specific fuel consumption (BSFC), and
(d) brake thermal efficiency (BTE) for all PSO biodiesel-diesel blends and diesel with respect to engine
speed at full load condition.
3.2.2. Torque
Figure 5b shows that torque increases initially with speed increase up to 1400 rpm when the
maximum level is achieved and then decreases continuously until the maximum speed of 2400 rpm
for all biodiesel blends and B0. B5 had the highest torque among the biodiesel blends due to its
higher calorific value and lower density and viscosity. As expected, B0 (diesel) had the highest torque
followed by B5, B10, and B20. The maximum torques were recorded at 1400 rpm, and these were
225 Nm, 221 Nm, 220 Nm, and 218 Nm for B0, B5, B10, and B20, respectively. The average torque
reduction for B5, B10, and B20 as compared with B0 was 1.37%, 2.47%, and 3.85%, respectively.
3.2.3. Brake Specific Fuel Consumption (BSFC)

Brake specific fuel consumption (BSFC) is the ratio of fuel flow rate and brake power. According
to Mofijur et al. [17], BSFC values depend on the relationship between the fuel injection system, fuel
362
Energies 2018, 11, 2931
density, calorific value, and viscosity. Figure 5c shows the variation of BSFC for all PSO biodiesel-diesel
blends and diesel with respect to engine speed. It is found that diesel has the lowest BSFC value in
comparison with PSO biodiesel-diesel blends (B5, B10, and B20). The average BSFC value of B0 was
measured as 256.5 g/kWh, whereas those for blends of B5, B10, and B20 were 265 g/kWh, 282.6 g/kWh,
and 300.4 g/kWh, respectively. As mentioned earlier, the combined effects of fuel properties, such as
density, viscosity, and calorific values of biodiesel may result in higher BSFC. Besides, biodiesel needs
more fuel for producing the same power due to its lower calorific value in comparison with diesel.
The average increase in BSFC values for B5, B10, and B20 as compared with B0 (diesel) were 3.35%,
10.16%, and 17.13%, respectively.
3.2.4. Brake Thermal Efficiency (BTE)

Brake thermal efficiency (BTE) is the ratio of the brake power and heat energy that is produced by
fuel. Figure 5d shows that BTE decreases with the increase of biodiesel in the blends from B5 to B20.
A higher BTE value (%) depends on some specific fuel properties such as higher calorific value, and
lower density and viscosity. When compared with other blends, the properties of B5 were matched
closely with diesel. The average BTE value of B5 was measured as 30.35%, whereas diesel was recorded
as 31.33%. The lower calorific values and higher fuel consumptions of both B10 and B20 resulted in
BTE values of 28.59% and 27.23%, respectively. The average reduction in BTE values for B5, B10, and
B20 as compared with B0 (diesel) were 3.1%, 8.76%, and 13.1%, respectively.
3.3. Interaction Effects of Operating Parameters on PSO Engine Performance

The complex interaction effects of operating parameters, such as biodiesel blends, load, and speed,
on each engine output response (BP, torque, BSFC, and BTE) could not be analysed independently.
The significance of each of the various parameters in the model was obtained via analysis of variance
(ANOVA). The experiments were carried out by use of the Box-Behnken response surface design.
Minitab 18 was used to carry out the statistical analysis. Table 7 shows the factors and the range and
levels of the investigated variables.
Table 7. Experimental range and levels coded for analysis of variance (ANOVA).
Range and Levels

Factors Unit Symbol Coded
−1 0 1
Biodiesel
% BL 0 10 20
blends
Load % LD 0 50 100
Speed rpm SP 1200 1800 2400
Once the experiments were completed, a full quadratic model was applied for the correlation of
the response variable to the independent variables. The form of the full quadratic model is shown in
Equation (2).
R = P0 + P1 Q1 + P2 Q2 + P3 Q3 + P1,2 Q1 Q2 + P1,3 Q1 Q3 + P2,3 Q2 Q3 + P1,1 Q1 2 + P2,2 Q2 2 + P3,3 Q3 2 (2)
where R is the response factor; P0 is a constant; P1 , P2 , P3 are regression coefficients, P1,1 , P1,2 , P1,3 , P2,2 ,
P2,3 , and P3,3 are quadratic coefficient; and, Q1 , Q2 , and Q3 are independent variables.
Consideration was given to the linear, quadratic and combined effects of operating parameters
to identify their impacts on the response. Each parameter’s the significance was evaluated by the
probability value (p-value) from ANOVA. At the 95% confidence level, the p-values less than 0.05
indicate a ‘significant’ effect of those parameters on the response. In other words, p-values more than
5% or 0.05 indicate ‘not-significant’ effects of those parameters on the response.
363
Energies 2018, 11, 2931
3.3.1. Effects of Biodiesel Blends, Load and Speed on Brake Power (BP)
The relationships between brake power and three operating parameters of biodiesel blends,
load, and speed were analysed. A quadratic regression model based on the coded parameters with
determined coefficients for statistical prediction as defined by Equation (3) was developed using
Minitab 18 to predict BP (kW) as a function of biodiesel blends (BL), load (LD), and speed (SP).
BP = 20 − 0.018 BL + 17.945 LD + 4.41 SP + 0.501 BL2 − 0.324 LD2 − 0.799 SP2 −

(3)
0.267 BL × LD − 2.177 BL × SP + 4.093 LD × SP
p-values from Table 8 show that the model is highly significant with an insignificant lack of
fit. When considering the linear, quadratic, and combined effects, BL is not significant as a linear
parameter; also, none of the quadratic terms is significant. Only the combined effects of LD and
SP are highly significant, whereas BL and SP have a significant effect on BP. The ANOVA results in
Table 8 also show that both LD and SP have the lowest p-values (<0.0001 each) and highest F-values
(3307.29 and 199.74). According to the quadratic Equation (3), both LD and SP have positive effects on
BP. This means that increasing the LD and SP will increase BP as well. The coefficient of determination
(R2 ) and the adjusted coefficient of determination (adj. R2 ) were 99.86% and 99.61%, respectively,
which indicates a high accuracy for the model.
Table 8. ANOVA results of the interactive effect of biodiesel blends, load, and speed on brake power.
Source DF Adj. SS Adj. MS F-Value p-Value Significant

Model 9 2821.89 313.54 402.53 <0.0001 Highly
Biodiesel
1 0.00 0.00 0.00 0.957 No
blends (BL)
Load (LD) 1 2576.18 2576.18 3307.29 <0.0001 Highly
Speed (SP) 1 155.58 155.58 199.74 <0.0001 Highly
BL × BL 1 0.93 0.93 1.19 0.325 No
LD × LD 1 0.39 0.39 0.50 0.512 No
SP × SP 1 2.36 2.36 3.02 0.143 No
BL × LD 1 0.29 0.29 0.37 0.571 No
BL × SP 1 18.97 18.97 24.35 0.004 Yes
LD × SP 1 66.99 66.99 86.01 <0.0001 Highly
Lack-of-Fit 3 3.58 1.19 7.71 0.117 No
Pure Error 2 0.31 0.15 - - -
Total 14 2825.79 - - - -
R2 = 0.9986 Adj. R2 = 0.9961 -
The ANOVA results in Table 8 for both LD × SP, and BL × SP interaction effects on BP are shown
graphically in Figure 6. The three-dimensional (3D) surface plot and two-dimensional (2D) contour
plot of LD and SP effects on BP are presented in Figure 6a,b respectively. BP (kW) increases with the
increase of LD up to 100% and exceeds 40 kW with SP of 2000 rpm onwards. The maximum BP was
found to be 45 kW at 2400 rpm. At 50% LD, the average BP value was recorded at about 20 kW with
minimum effects from SP. However, the combined effects of LD and SP on BP are more significant with
an increase of LD above 50%. Figure 6c,d present the 3D surface plot and 2D contour plot, respectively,
which show only minor influences of changes of BL and SP on BP. It is therefore concluded that BL has
a slight effect on BP, irrespective of SP changes.
364
Energies 2018, 11, 2931
(a) Three-dimensional (3D) surface plot of load (b) Two-dimensional (2D) contour plot of load and
and speed on brake power (BP) speed on BP
(c) 3D surface plot of biodiesel blends and speed on BP (d) 2D contour plot of biodiesel blends and speed on BP
Figure 6. Combined effects of biodiesel blends, load, and speed on BP (kW).
3.3.2. Effects of Biodiesel Blends, Load, and Speed on Torque

The relationships between torque and three operating parameters of biodiesel blend, load, and
speed were analysed. Based on the coded parameters, a quadratic regression model with determined
coefficients for statistical prediction, as defined by Equation (4), was developed using Minitab 18 to
predict torque (N.m) as a function of biodiesel blend (BL), load (LD), and speed (SP).
T = 137.47 + 4.596 BL + 95.884 LD − 8.105 SP − 16.6 BL2 − 15.06 LD2 − 18.13 SP2 −
(4)
0.53 BL × LD − 11.42 BL × SP − 12 LD × SP
p-values from Table 9 show that the model is highly significant with an insignificant lack of fit.
When considering the linear, quadratic, and combined effects, only the combined effects of BL and
LD are not significant. The ANOVA results in Table 9 show that LD is not significant, as it has the
lowest p-value and highest F-value, while BL is found to be significant and all other parameters are
highly significant. According to the quadratic Equation (4), both LD and BL have positive effects
on torque. This means that increasing the LD and BL will increase torque as well. The coefficient of
determination (R2 ) and the adjusted coefficient of determination (adj. R2 ) were 99.96% and 99.89%,
respectively, which indicates the high accuracy of the model.
365
Energies 2018, 11, 2931
Table 9. ANOVA results of the interactive effect of biodiesel blends, load, and speed on torque.

Model 9 78,005.7 8667.3 1352.54 <0.0001 Highly
BL 1 169.0 169.0 26.37 0.004 Yes
LD 1 73,549.5 73,549.5 11,477.45 <0.0001 Highly
SP 1 525.5 525.5 82.01 <0.0001 Highly
BL × BL 1 1017.2 1017.2 158.73 <0.0001 Highly
LD × LD 1 837.8 837.8 130.73 <0.0001 Highly
SP × SP 1 1213.7 1213.7 189.40 <0.0001 Highly
BL × LD 1 1.1 1.1 0.18 0.691 No
BL × SP 1 521.7 521.7 81.41 <0.0001 Highly
LD × SP 1 575.5 575.5 89.81 <0.0001 Highly
Lack-of-Fit 3 30.4 10.1 12.13 0.077 No
Pure Error 2 1.7 0.8 - - -
Total 14 78,037.7 - - - -
R2 = 0.9996 Adj. R2 = 0.9989
The ANOVA results in Table 9 for LD × SP, and BL × SP interaction effects on torque are shown
graphically in Figure 7. The 3D surface plot and 2D contour plot of LD and SP effects on torque are
presented in Figure 7a,b, respectively. Torque (Nm) decreases with the increase of SP at 100% LD.
The maximum torque was found to be 220 Nm at 1400 rpm. At 50% LD, the average torque was
recorded as about 120 Nm with minimum effects from SP. However, the combined effects of LD and SP
on torque are more significant with an increase of LD above 50%. Figure 7c,d present the 3D surface
plot and 2D contour plot, respectively, which show only minor influences of changes of BL and SP on
torque. It is therefore concluded that BL has a slight effect on torque, irrespective of SP changes.
(a) 3D surface plot of load and speed on torque (b) 2D contour plot of load and speed on torque
(c) 3D surface plot of biodiesel blends and (d) 2D contour plot of biodiesel blends and
speed on torque. speed on torque.
Figure 7. Combined effects of biodiesel blends, load, and speed on torque (Nm).
366
Energies 2018, 11, 2931
3.3.3. Effects of Biodiesel Blends, Load, and Speed on Brake Specific Fuel Consumption (BSFC)
The relationships between three operating parameters of biodiesel blend, load, and speed with
brake specific fuel consumption (BSFC) were analysed. Based on the coded parameters, a quadratic
regression model with determined coefficients for statistical prediction, as defined by Equation (5),
was developed using Minitab 18 to predict BSFC (gm/kWh) as a function of biodiesel blends (BL),
load (LD), and speed (SP).
BSFC = 252.17 + 138.13 BL − 526.25 LD − 30.43 SP + 62.5 BL2 + 488.88 LD2 + 71.66 SP2 −
(5)
125.71 BL × LD − 38.55 BL × SP + 78.67 LD × SP
p-values in Table 10 show that the model is highly significant with an insignificant lack of fit. When
considering the linear, quadratic, and combined effects, all of the parameters are highly significant.
Among all parameters, LD is the parameter with the lowest p-value and highest F-value (20,429.23).
According to the quadratic Equation (5), both LD and SP have a negative effect on BSFC. This means
that increasing the LD will decrease BSFC, whereas decreasing the LD will increase BSFC. The SD
parameter has less influence in changing BSFC. The coefficient of determination (R2 ) and the adjusted
coefficient of determination (adj. R2 ) were 99.98% and 99.95%, respectively, which indicates the high
accuracy of the model.
Table 10. ANOVA results of the interactive effect of biodiesel blends, load and speed on brake specific
fuel consumption (BSFC).

Model 9 3,358,105 373,123 3440.61 <0.0001 Highly
BL 1 152,631 152,631 1407.43 <0.0001 Highly
LD 1 2,215,481 2,215,481 20,429.23 <0.0001 Highly
SP 1 7408 7408 68.31 <0.0001 Highly
BL × BL 1 14,424 14,424 133.00 <0.0001 Highly
LD × LD 1 882,480 882,480 8137.45 <0.0001 Highly
SP × SP 1 18,960 18,960 174.83 <0.0001 Highly
BL × LD 1 63,215 63,215 582.91 <0.0001 Highly
BL × SP 1 5944 5944 54.81 0.001 Highly
LD × SP 1 24,756 24,756 228.28 <0.0001 Highly
Lack-of-Fit 3 521 174 16.38 0.058 No
Pure Error 2 21 11 - - -
Total 14 3,358,647 - - - -
R2 = 0.9998 Adj. R2 = 0.9995
The ANOVA results in Table 10 for LD × SP, BL × LD, and BL × SP interaction effects on BSFC
are shown graphically in Figure 8. The 3D surface plot and 2D contour plot of LD and SP effects on
BSFC are presented in Figure 8a,b respectively. BSFC decreases with the increase of LD, irrespective
of SP. The maximum BSFC was found to be 1450 gm/kWh at 0% LD and 1200 rpm SP. At 50% LD,
the average BSFC was recorded at about 400 gm/kWh with almost no effect from changes in SP (rpm).
Figure 8c,d present the 3D surface plot and 2D contour plot, respectively, of BL and LD effects on BSFC.
At 0% LD, BL has a significant effect on BSFC. The 20% biodiesel blend (BL20) shows the maximum
BSFC of 1450 gm/kWh at 0% LD. BSFC values decrease with the increase of LD, irrespective of BL.
From Figure 8d, it can be seen that, at 36% LD, the 0–10% biodiesel blends (BL0, BL5, and BL10) have
low BSFC values whereas, at 60% LD, BL20 also has achieved a lower BSFC. It is therefore concluded
that BL has a minor effect on BSFC, irrespective of SP changes. Figure 8e,f show the 3D surface plot
and 2D contour plot respectively of BL and SP effects on BSFC. BSFC values for BL0 have recorded
for SP values in the range of 1200 to 2400 rpm and found to be less than 400 gm/kWh. Changes in
both BL and SP influence BSFC values. From Figure 8f, it can be seen that both BL at 5% and BL at 10%
367
Energies 2018, 11, 2931
have low BSFC for SP values in the range of 1600 to 2000 rpm. Higher BL percentages with higher SP
values (up to 2000 rpm) result in higher values of BSFC.
(a) 3D surface plot of load and speed on BSFC (b) 2D contour plot of load and speed on BSFC
(c) 3D surface plot of biodiesel blends and load on BSFC (d) 2D contour plot of biodiesel blends and load on BSFC
(e) 3D surface plot of biodiesel blends and speed on BSFC (f) 2D contour plot of biodiesel blends and speed on BSFC
Figure 8. Combined effects of biodiesel blends, load, and speed on BSFC (gm/kWh).
3.3.4. Effects of Biodiesel Blends, Load, and Speed on Brake Thermal Efficiency (BTE)
The relationships between brake thermal efficiency (BTE) and three operating parameters of
biodiesel blends, load, and speed were analysed. Based on the coded parameters, a quadratic regression
model with determined coefficients for statistical prediction, as defined by Equation (6), was developed
using Minitab 18 to predict BTE (%) as a function of biodiesel blends (BL), load (LD), and speed (SP).
BTE = 32.283 − 1.36 BL + 11.752 LD − 1.71 SP − 2.795 BL2 − 12.05 LD2 − 1.9 SP2 +
(6)
1.038 BL × LD + 0.267 BL × SP − 2.097 LD × SP
p-Values in Table 11 show that the model is highly significant with an insignificant lack of
fit. When considering the linear, quadratic, and combined effects, only the BL × LD and BL × SP
combined parameters are not significant; all other parameters are significant. Only the combined
368
Energies 2018, 11, 2931
effects of LD × SP are significant, whereas the quadratic terms LD × LD, BL × BL, and SP × SP were
all found to have significant effects on BTE. The ANOVA results in Table 11 also show that LD has both
the lowest p-values (<0.0001) and highest F-value (940.76). According to the quadratic Equation (6),
only LD has a positive effect on BTE. This means that increasing the LD will increase BTE as well.
The coefficient of determination (R2 ) and the adjusted coefficient of determination (adj. R2 ) were
99.66% and 99.05%, respectively, which indicates the high accuracy of the model.
Table 11. ANOVA results of the interactive effect of biodiesel blends, load, and speed on brake thermal
efficiency (BTE).

Model 9 1717.12 190.79 162.44 <0.0001 Highly
BL 1 14.80 14.80 12.60 0.016 Yes
LD 1 1104.97 1104.97 940.76 <0.0001 Highly
SP 1 23.39 23.39 19.92 0.007 Yes
BL × BL 1 28.85 28.85 24.57 0.004 Yes
LD × LD 1 536.17 536.17 456.49 <0.0001 Highly
SP × SP 1 13.34 13.34 11.35 0.020 Yes
BL × LD 1 4.31 4.31 3.67 0.114 Not
BL × SP 1 0.29 0.29 0.24 0.642 No
LD × SP 1 17.60 17.60 14.98 0.012 Yes
Lack-of-Fit 3 5.48 1.83 9.26 0.099 No
Pure Error 2 0.39 0.20 - - -
Total 14 1722.99 - - - -
R2 = 0.9966 Adj. R2 = 0.9905
The ANOVA results in Table 11 for LD × SP interaction effects on BTE are shown graphically
in Figure 9. The 3D surface plot and 2D contour plot of LD and SP effects on BTE are presented
in Figure 9a,b, respectively. BTE increases with the increase of LD with minimum impacts from SP.
The maximum BTE was found to be 33.43% at 1200 rpm. However, the combined effects of LD and
SP on BTE are affected by an increase of LD above 50%. It is therefore concluded that SP has a minor
impact on BTE, irrespective of LD changes.
(a) 3D surface plot of load and speed on BTE (b) 2D contour plot of load and speed on BTE
Figure 9. Combined effects of biodiesel blends, load, and speed on BTE (%).
4. Conclusions
This study investigated the effect of papaya seed oil (PSO) biodiesel in an agricultural diesel engine
with various biodiesel-diesel blends and the resulting engine performance outcomes as compared with
those from the use of a reference diesel. A response surface methodology was introduced to analyse
and describe the performance of this engine. The results of this investigation can be summarised,
as follows:
369
Energies 2018, 11, 2931
• PSO biodiesel-diesel blends (B5–B50) meet the European standard EN 590 for having minimum
oxidation stability of 20 h.
• Over the entire speed range of 1200 to 2400 rpm, the average BP value reductions for B5, B10, and
B20 in comparison with B0 were 2.88%, 3.87%, and 5.13%, respectively.
• The average torque reductions for B5, B10, and B20 in comparison with B0 were 1.37%, 2.47%,
and 3.85%, respectively.
• The average increase in BSFC values for B5, B10, and B20 in comparison with B0 were 3.35%,
10.16%, and 17.13%, respectively.
• The average BTE value of B5 was measured as 30.35%, whereas B0 (diesel) was recorded as 31.33%.
• The interactive relationships between three operating parameters (biodiesel blends, load, and
speed) and four responses (BP, torque, BSFC, and BTE) were analysed. ANOVA and a statistical
regression model show that load and speed were the two most important parameters that affect
all four responses. The biodiesel blends parameter had a significant effect on BSFC.
These results show that B5 and B10 PSO biodiesel-diesel blends can be used to fuel diesel
engines without further engine modification. Therefore, it can be concluded that papaya seed oil
can be considered as a promising source of biodiesel production. However, before recommending
as a future alternative energy source in commercial scale, further research needs to be conducted in
terms of engine emissions, in-cylinder pressure, burn rate data, combustion analysis, and tribological
performance analysis.
Author Contributions: The contributions of each author are as follows: M.A. produced and characterised
biodiesel-diesel blends, conducted and analysed engine performance, and drafted the manuscript; M.G.R.
contributed to the experimental design and thoroughly revised the paper, and N.A. helped to revised and
improved the paper.
Acknowledgments: The authors would like to acknowledge Tim McSweeney, Adjunct Research Fellow, Tertiary
Education Division at Central Queensland University, Australia for his contribution in proof reading of this article.
References
1. Bhuiya, M.M.K.; Rasul, M.G.; Khan, M.M.K.; Ashwath, N.; Azad, A.K.; Hazrat, M.A. Second Generation
Biodiesel: Potential Alternative to-edible Oil-derived Biodiesel. Energy Procedia 2014, 61, 1969–1972.
[CrossRef]
2. Asokan, M.A.; Prabu, S.S.; Kamesh, S.; Khan, W. Performance, combustion and emission characteristics
of diesel engine fuelled with papaya and watermelon seed oil bio-diesel/diesel blends. Energy 2018, 145,
3. Tüccar, G.; Tosun, E.; Özgür, T.; Aydın, K. Diesel engine emissions and performance from blends of citrus
sinensis biodiesel and diesel fuel. Fuel 2014, 132, 7–11. [CrossRef]
4. Ozsezen, A.N.; Canakci, M.; Turkcan, A.; Sayin, C. Performance and combustion characteristics of a DI diesel
engine fueled with waste palm oil and canola oil methyl esters. Fuel 2009, 88, 629–636. [CrossRef]
5. Agunbiade, F.O.; Adewole, T.A. Methanolysis of Carica papaya Seed Oil for Production of Biodiesel. J. Fuels
2014, 2014, 904076. [CrossRef]
6. Wong, C.S.; Othman, R. Biodiesel Production by Enzymatic Transesterification of Papaya Seed Oil and
Rambutan Seed Oil. Int. J. Eng. Technol. 2014, 6, 2773–2777.
7. Mohan, A.; Sen, S.S.S. Emission Analysis Of The Biodiesel from Papaya and Chicken Blends. Int. J. Inform.
Futurist. Res. 2015, 2, 2258–2264.
8. De Melo, M.L.S.; Santos, N.A.; Rosenhaim, R.; Souza, A.G.; Filho, P.F.A. Use of thermal analysis techniques for
evaluation of the stability and chemical properties of papaya biodiesel (Carica papaya L.) at low temperatures.
J. Therm. Anal. Calorim. 2011, 106, 831–836. [CrossRef]
9. Daryono, E.D.; Sinaga, E.J. Rapid In Situ Transesterification of Papaya Seeds to Biodiesel with the Aid of
Co-solvent. Int. J. Renew. Energy Res. 2017, 7, 379–385.
370
Energies 2018, 11, 2931
10. Anwar, M.; Rasul, M.G.; Ashwath, N. Optimization of biodiesel production process from papaya
(Carica papaya) seed oil. In Proceedings of the 2017 IEEE 7th International Conference on Power and
Energy Systems (ICPES), Toronto, ON, Canada, 1–3 November 2017.
11. Prabhakaran, P.; Saravanan, C.G.; Aalam, C.S. Effects of Papaya Methyl Ester on DI Diesel Engine
Combustion, Emission and Performance Characteristics. Int. Res. J. Eng. Technol. 2016, 3, 319–325.
12. Sundar Raj, C.; Karthikayan, M. Effect of additive on the performance, emission and combustion
characteristics of a diesel engine run by diesel-papaya methyl ester blends. Int. J. Chem. Sci. 2016, 14,
2823–2834.
13. Sharma, S.K.; Mitra, S.K.; Saran, S. Papaya Production in India—History, Present Status and Future Prospects;
International Society for Horticultural Science (ISHS): Leuven, Belgium, 2016; Volume 1111, pp. 87–94.
[CrossRef]
15. Silitonga, A.S.; Masjuki, H.H.; Mahlia, T.M.I.; Ong, H.C.; Chong, W.T.; Boosroh, M.H. Overview properties of
biodiesel diesel blends from edible and non-edible feedstock. Renew. Sustain. Energy Rev. 2013, 22 (Suppl. C),
16. Silitonga, A.S.; Ong, H.C.; Mahlia, T.M.I.; Masjuki, H.H.; Chong, W.T. Biodiesel Conversion from High
FFA Crude Jatropha Curcas, Calophyllum Inophyllum and Ceiba Pentandra Oil. Energy Procedia 2014,
61 (Suppl. C), 480–483. [CrossRef]
17. Mofijur, M.; Masjuki, H.H.; Kalam, M.A.; Atabani, A.E. Evaluation of biodiesel blending, engine performance
and emissions characteristics of Jatropha curcas methyl ester: Malaysian perspective. Energy 2013, 55,
149 (Suppl. C), 392–400. [CrossRef]
371
energies
Article
Mathematical Modeling of Non-Premixed Laminar
Flow Flames Fed with Biofuel in Counter-Flow
Arrangement Considering Porosity and
Thermophoresis Effects: An Asymptotic Approach
Mehdi Bidabadi 1 , Peyman Ghashghaei Nejad 1 , Hamed Rasam 1 , Sadegh Sadeghi 1 and
Bahman Shabani 2, *
1 School of Engineering, Iran University of Science and Technology, Narmak, Tehran 16846-13114, Iran;
[email protected] (M.B.); [email protected] (P.G.N.); [email protected] (H.R.);
[email protected] (S.S.)
2 School of Engineering, RMIT University, Melbourne, VIC 3083, Australia
* Correspondence: [email protected]; Tel.: +61-(0)-3-9925-4353
Received: 2 October 2018; Accepted: 24 October 2018; Published: 29 October 2018
Abstract: Due to the safe operation and stability of non-premixed combustion, it can widely
be utilized in different engineering power and medical systems. The current paper suggests a
mathematical asymptotic technique to describe non-premixed laminar flow flames formed in organic
particles in a counter-flow configuration. In this investigation, fuel and oxidizer enter the combustor
from opposite sides separately and multiple zones including preheating, vaporization, flame and
post-flame zones were considered. Micro-sized lycopodium particles and air were respectively
applied as a biofuel and an oxidizer. Dimensionalized and non-dimensionalized mass and energy
conservation equations were determined for the zones and solved by Mathematica and Matlab
software by applying proper boundary and jump conditions. Since lycopodium particles have
numerous spores, the porosity of the particles was involved in the equations. Further, significant
parameters such as lycopodium vaporization rate and thermophoretic force corresponding to the
lycopodium particles in the solid phase were examined. The temperature distribution, flame sheet
position, fuel and oxidizer mass fractions, equivalence ratio and flow strain rate were evaluated
for the counter-flow non-premixed flames. Ultimately, the thermophoretic force caused by the
temperature gradient at different positions was computed for several values of porosity, fuel and
oxidizer Lewis numbers.
Keywords: porosity; thermophoretic force; biomass fuel; non-premixed combustion; counter-flow

structure; mathematical modeling
1. Introduction
Considering the progressive depletion of fossil fuels, the soaring costs of these fuels and the
greenhouse gas (GHG) emissions associated with them, significant attention has been turned toward
using alternatives to these fuels [1,2]. In recent years, bioenergy has been introduced as a reasonable
form of renewable energy with a high degree of sustainability for use in various industrial and medical
applications [3]. Bioenergy can readily be achieved through the combustion of biomass fuels [4]. On the
basis of structure, combustion is categorized into premixed, partially premixed and non-premixed
flames [5,6], among which non-premixed combustion is widely utilized in engineering combustion
power systems due to its safe operation and stability [7]. So far, different experimental and analytical
investigations have been performed by various researchers to specify the propagation of non-premixed
flames through biofuels.
Energies 2018, 11, 2945; doi:10.3390/en11112945. 372 www.mdpi.com/journal/energies

Energies 2018, 11, 2945
Joshi and Berlad [8] experimentally studied the behavior of premixed flames through a
lycopodium-air mixture and reported the flame temperature distribution in the vertical position.
In another investigation, Berlad and Joshi [9] used both analytical and experimental approaches to
model the extinction and temperature distribution of premixed flames burning lycopodium dust clouds.
Berlad and Tangirala [10] studied the effect of thermal radiation on the behavior and structure of
lycopodium-fueled flames considering a small quantity of gravity force. The behavior of a triple-zone
premixed flame containing uniformly-distributed lycopodium particles was developed by Seshadri,
applying an asymptotic method [11]. Han et al. [12] experimentally perused flame propagation
through a lycopodium–air mixture within a vertical duct and evaluated the weight loss of the
lycopodium particles and flame velocity. In further experimental work, Han et al. [13] investigated
the propagation of lycopodium dust particles in a premixed combustible system using particle image
velocimetry (PIV). Xi et al. [14] used an experimental method to study the pulsation characteristics
of a swirl non-premixed flame. Proust [15] experimentally investigated flame propagation through
lycopodium-air and sulphur flour-air mixtures. Skjold et al. [16] examined turbulent flame propagation
through lycopodium dust clouds at a constant pressure and made a comparison between the structures
of the produced premixed and non-premixed flames. Shamooni et al. [17] employed a finite-rate
scale similarity method to predict the combustion and heat release rates in non-premixed jet flames.
A two-dimensional analytical model was presented by Rahbari et al. [18] for premixed flames
propagating through a lycopodium-air mixture taking into account the particle size and equivalence
ratio parameters. Bidabadi and Esmaeilnejad et al. [19] suggested an analytical triple-zone model for
counter-flow premixed flames propagating through a lycopodium-air mixture. Bidabadi et al. [20]
proposed a non-premixed triple-zone combustion model for lycopodium dust flames in a counter
flow configuration. Xi et al. [21] experimentally and numerically simulated the flame shape and
size for a high-pressure turbulent non-premixed swirl combustion. The combustion behavior of
non-premixed micro-jet flames through a methane-air mixture in a co-flow configuration was reported
by Li et al. [22]. Bidabadi et al. [23] theoretically investigated the effect of thermal resistance on
the structure of the premixed combustion of micro-organic dust particles with air. Spijker and
Raupenstrauch [24] performed a numerical investigation to reveal the effect of the inner structure
of lycopodium particles on the propagation of premixed flames. Chen et al. [25] experimentally
described the temperature distribution of the propagation of premixed flames through a methane-air
mixture considering the porosity parameter. Bidabadi et al. [26] developed a method for the theoretical
investigation of multi-zone counter-flow premixed flames burning lycopodium dust clouds considering
the vaporization rate. Ji et al. [27] experimentally computed the maximum pressure and the maximum
rate of pressure rise of several hybrid mixtures taking into account different venting diameters and
static activation pressures. Di Benedetto et al. [28] employed a theoretical/numerical method to
determine the minimum ignition temperature of polyethylene dust at several dust concentrations in air.
Sanchirico et al. [29] evaluated the flammability and combustion behavior of several complex hybrid
mixtures, e.g., mixtures of nicotinic acid, lycopodium and methane, at different dust concentrations.
In another study, Sanchirico et al. [30] experimentally demonstrated the combustion and flammability
behavior of combustible dust mixtures including lycopodium, Nicotinic acid and Ascorbic acid.
As reviewed of the literature, numerous theoretical and experimental examinations have been
undertaken on the behavior of premixed flames propagating through organic fuels. Until now, no
comprehensive mathematical models have been available in the literature on the behavior of multi-zone
counter-flow non-premixed flames. In most of the previous studies, in which limited models were
introduced, the shape of lycopodium particles was considered to be spherical while these particles
consist of numerous minute spores that can affect the structure of premixed and non-premixed
flames [12]. In addition, the thermophoretic force, caused by a temperature gradient for transferring
micro-sized fuel particles, has not yet been investigated mathematically for non-premixed flames.
The current paper aimed to theoretically describe the structure of counter-flow non-premixed
flames through organic particles taking into account the effects of fuel particle porosity and
373
Energies 2018, 11, 2945
thermophoresis. The biofuel particles and oxidizer were separately injected into a combustor from
opposite sides. Multiple zones composed of preheating, vaporization, flame and post-flame zones,
are presented. Lycopodium particles and air were injected into the system as biomass and oxidizer,
respectively. The governing equations, including mass and energy conservation equations, were
derived for the above-mentioned zones. The impacts of fuel particle porosity, vaporization and
thermophoresis on the structure of the flames are discussed. Changes in the flame temperature and the
location of the flame sheet with fuel and oxidizer Lewis numbers were examined for several porosity
factors. Additionally, the influence of the equivalence ratio on the temperature distribution was
evaluated. Afterwards, the mass fractions and temperature distributions of the biofuel and oxidizer at
different positions were obtained. Then, variations of the flow critical strain rate with oxidizer and fuel
Lewis numbers are described. Finally, the thermophoretic force caused the temperature gradient of the
solid particles was computed for several different porosity factors, fuel and oxidizer Lewis numbers.
2. Theoretical Modeling
2.1. Lycopodium Characteristics

In the last couple of decades, the accidental ignition of flammable organic and inorganic dusts,
as well as knowledge of combustion and flammability, has gained much attention [31]. Lycopodium
particles have been introduced as a reasonable reference for volatile biomass fuels due to its excellent
dispersibility and flammability [12]. Lycopodium can efficiently be used as a reference fuel for testing
different biomass-fueled systems and recognizing the combustion processes in the systems prior to
industrial scale-up [12,29,30]. A lycopodium particle typically contains almost 50% fat oil, 2% sucrose
and 24% sporopollenin [12]. One of the main advantages of this biofuel is that almost no solid residue
is produced during the combustion process [12]. In this study, as can be found in the literature, it
was assumed that a gaseous fuel with a certain chemical composition, gaseous methane, is produced
through the vaporization of these volatile particles [11]. In this regard, the pyrolysis process was
disregarded in the current study. It should be noted that lycopodium particles are mono-disperse in
size [14] and contain numerous minute spores [12], as illustrated in Figure 1. Moreover, when the
lycopodium particles were injected into a combustible system via a nozzle, they did not undergo
breakage due to their sponge-like behavior (i.e., the numerous minute spores are “elastic”) [31].
Figure 1. Scanning electron microscopy of lycopodium biofuel particles [22].
374
Energies 2018, 11, 2945
2.2. Porosity of Lycopodium Particles

As described in Figure 1, lycopodium particles are not completely spherical-shaped particles.
Therefore, the porosity effect should be involved in the modeling of the flame structures fueled by
these particles. Figure 2 shows the porosity of the lycopodium particles considered in this study.
Figure 2. Schematic representation of the porous biofuel particle considered in the current analysis.
The volume porosity and porosity factor [32] are defined by the following equations, respectively:
Vvoid
ε= (1)
Vtotal
fe = 1 − ε (2)
where ε, Vvoid and Vtotal are the fraction of the empty volume, void volume and total
volume, respectively.
2.3. Thermophoretic Force

The thermophoresis phenomenon, first introduced by Tyndall, is due to the temperature gradient
that transfers micro-sized particles from a warmer zone to a colder zone [33–35]. This phenomenon
is notably influenced by the particle size and mean free path of the molecules [36–38]. It should be
pointed out that the ratio of the molecules mean free path to diameter is represented by the Knudsen
number that can be obtained by the following equation [36]:
L
Kn = (3)
rp
where r p is the radius of the particles, and L is the mean free path of the molecules that can be defined
as follows [36,38]:
μ
L= (4)
∅ρ g C
where ∅ is equal to 0.941 [38], and μ and ρ g are the dynamic viscosity of the fuel particles and the
density of the gas, respectively. c is the mean thermal velocity of the molecules that can be expressed
as follows [36]:
8RT
c= (5)
πM
where T and M are the temperature and molecular mass, respectively. In addition, R is the universal
gas constant, which is equal to 8.314 molJ ·K .
To assess the thermophoretic force, the following correlation can be applied [36]:
∇T
FTh = −3πμ2 r p k T (6)
ρ g T∞
375
Energies 2018, 11, 2945
where k T is calculated by the following correlation [36]:
k
2Ck ( k gp + Ct Kn)
kT = k
(7)
(1 + 3Cm Kn)(1 + 2 k gp + 2Ct Kn)
where μ, r p , k g , k p , ∇ T and T∞ are the dynamic viscosity of the fuel particles, the particle radius, the
gas thermal conductivity, the solid particle thermal conductivity, the temperature gradient and the
ambient temperature, respectively. Ck , Ct and Cm are the temperature jump, the temperature creep
and the velocity jump coefficients [37]. Table 1 lists the constant parameters used in Equation (7).
Table 1. Constant parameters used in Equation (7) [36,37].
Parameter Value
1.446538 × 10−4
kj
kp m·s·K
10−4 m·s·K
kj
kg 0.3468 ×
Ck 1.147
Ct 2.2
Cm 1.146
The Knudsen number can be obtained by the following correlation [36]:

Assuming small Knudsen numbers, the Talbot correlation (Equation (6)) can be rewritten as
follows [36,38,39]:
kg ∇T
FTh,continum = −6πμ2 r p C p (8)
k p + 2k g ρ g T∞
where C p is the specific heat of the lycopodium particles.
2.4. The Flame Structure

Figure 3 represents the structure of a non-premixed flame, in which the biofuel and oxidizer were
entrained in a counter-flow configuration. Regarding this figure, the porous biofuel particles and
oxidizer separately enter the combustor from −∞ and +∞ (opposed sides), respectively. In this analysis,
a gaseous fuel with a specific chemical composition, gaseous methane, was produced from solid biofuel
particles once they asymptotically crossed the vaporization front [11]. Subsequently, the gaseous fuel
blended with the oxidizer within the reaction zone and a huge quantity of thermal bioenergy was
released. In this investigation, an asymptotic approach was used to model the vaporization and the
reaction processes. Besides, the momentum of the fuel and oxidizer streams was considered to be
equal so a stagnation plane was placed in the middle of the fuel and oxidizer nozzles. The location of
the stagnation plane was employed as the reference coordinate for measuring the location of the flame
sheet and vaporization front, as shown in Figure 3. The mass particle concentration value at −∞ was
assumed to be less than 100 gr/cm3 . Therefore, the initial position of the flame sheet was presumed to
be on the left-hand side of the stagnation plane. It should be noted that the flame sheet position was
the axial distance from the stagnation plane at which the flame front was detected.
376
Energies 2018, 11, 2945
Figure 3. Schematic of the non-premixed flame in a counter-flow arrangement.
2.5. Mathematical Modeling of the Flame

In this paper, the vaporization rate was modeled as the amount of gaseous fuel mass per unit
volume per unit time that can be described as follows [40]:
Ys
ωv = H ( T − Tv ) (9)
τvap
where Ys , τvap , H, T and Tv are the mass fraction of solid particles, the constant vaporization
characteristic time, the Heaviside function, the fuel temperature and the vaporization temperature of
the particles. The Heaviside function is defined as follows:

0 T < Tv
H ( T − Tv ) = . (10)
1 T ≥ Tv
The thermal diffusivity (α) to mass diffusivity (D) ratio (i.e., Lewis number) is defined as
follows [41]:
α
Le = . (11)
D
Chemical kinetics were assumed to be a one-stage reaction and expressed as follows [42]:
v F [ F ] + vO [O] → v product [ P] (12)
where [ F ], [O] and [ P] denote the fuel, oxidizer and products, respectively. υ F , υO and υP are the
stoichiometric coefficients of the fuel, oxidizer and products, respectively. The one-dimensional
velocity field of the particles was considered as follows:
u = − aX (13)
where u and a represent the velocity in the X direction and the flow strain rate, respectively.
The following assumptions were considered in this analysis:
• For simplicity, it was assumed that values of density and specific heat were constant and the
momentum of the fuel and oxidizer streams were the same.
• The vaporization process occurred in a very thin zone (asymptotic limit).
377
Energies 2018, 11, 2945
• Thermal radiation and heat losses were disregarded.

• Lycopodium particles were uniformly dispersed. In other words, the size and shape of the
particles were assumed to be the same.
• A large Zeldowich number was presumed. Thus, the thickness of the flame zone would be
too small.
• The ambient temperature ( T±∞ ) was assumed to be 300 K.
• In order to analytically solve the coupled complex conservation equations of mass and energy in
the considered zones, it was assumed that a gaseous fuel with a certain chemical composition
evolved from the asymptotic vaporization of the lycopodium particles. Therefore, pyrolysis was
disregarded as clearly considered in References [11,20,43].
• No chemical interaction occurred between the solid particles before the vaporization front.
2.5.1. Dimensionalized Governing Equations

The conservation of mass equation for a fuel in the gaseous phase was derived, considering the
effect of porosity:
dY d2 YF ω ωv
− aX F = DF − fe F + fe (14)
dX dX 2 ρ ρ
where DF , YF and ω F are the fuel diffusivity, the gaseous fuel mass fraction and the chemical
reaction rate, respectively. The chemical reaction rate was obtained using a one-step Arrhenius
correlation [44,45]:
E
ω F = βρ2 v F vO YF YO exp(− ) (15)
RT
where β and E are the frequency constant and overall activation energy, respectively. YF and YO are
expressed as follows [41]:
m
YF = YF (16)
vF mF
m
YO = YO (17)
vO mO
where YO and YF represent the oxidizer and fuel volume fractions. m F and mO are the fuel and oxidizer
molecular weights, and m denotes the molecular weight of the mixture.
By neglecting the diffusion of the solid particles (no interaction exists between the solid particles),
the mass conservation equation for the solid fuel particles can be presented as follows:
dYs
− aX = − ωv (18)
dX
where Ys and ωv are the mass fraction and vaporization rate of the solid fuel particles, respectively.
The conservation equation of energy was obtained as follows:
dT d2 T Q Qv
− aX = DT 2 + f e ω F − f e ωv (19)
dX dX ρC C
where Q is the released heat per unit mass of consumed fuel, Qv is the latent heat of vaporization and
λ
DT = ρC is the thermal diffusivity coefficient. In Equation (19), C represents the specific heat capacity
of the mixture that can be calculated using the following correlation [40,46,47]:
4ρ p 3
C = Ca + f e πr C p n p (20)
3ρ
where ρ p is the density of the solid particles and n p is the particle number density per unit volume.
The density of the mixture is stated as follows, taking into account the porosity of the particles [40]:
378
Energies 2018, 11, 2945
4
ρ = ρ a + f e πr p 3 n p ρ p . (21)
3
The conservation of mass equation for the oxidizer is as follows:
dYO d2 YO ω
− aX = DO − fe ϑ F (22)
dX dX 2 ρ
where DO and YO are the oxidizer mass diffusion and mass fraction, respectively.
2.5.2. Normalization of the Governing Equations

For normalizing the governing equations, the following parameters were employed [41]:
C ( T − T∞ ) YF YO Ys X
θ= , yF = , y = , ys = , x= (23)
QYF−∞ YF−∞ O ϑYF−∞ YF−∞ λ
ρCa
where θ, y F , yO , yS and x are the normalized forms of the temperature, the fuel mass fraction, the
oxidizer mass fraction, the solid fuel mass fraction and position, respectively. YF−∞ is the initial
mass fraction of the fuel in the gaseous phase at x → −∞ (where fuel particles exit in the fuel
nozzle). By applying the non-dimensionalized parameters (Equation (23)) into Equations (14), (18),
(19) and (22) and considering the definition of a chemical reaction, the vaporization rate and Lewis
number, the following normalized equations were obtained for the gaseous fuel mass conservation
(Equation (24)), the solid fuel mass conservation (Equation (25)), the oxidizer mass conservation
equation (Equation (26)) and the energy conservation equation (Equation (27)), respectively:
1 d2 y F dy ys Ta
+ x F + fe H ( T − Tv ) = f e Dc y F yO exp (− ) (24)
Le F dx2 dx aτvap T
dys ys
x = H ( T − Tv ) (25)
dx aτvap
dyO 1 d 2 yO Ta
x + = f e Dc y F yO exp (− ) (26)
dx LeO dx2 T

d2 θ dθ q Ta
+ x − fe ys H ( T − Tv ) = − f e DC y F yO exp − (27)
dx2 dx aτvap T
where Ta is the activation temperature, which is defined as Ta = RE (E is the overall activation energy),
q is defined as q = QQv and DC is the Damkohler number, which was determined using the following
equation [41]:
DC = ρBϑO YF−∞ /WF a (28)
where ϑO is the number of moles of oxygen participating in the reaction with one mole of fuel and WF
is the molecular weight of the fuel.
2.5.3. Boundary and Jump Conditions

In order to solve the non-dimensionalized governing equations in each of the zones, proper
boundary and jump conditions were used. In this study, the following division was considered:
Preheat zone: R1 : { x | − ∞ < x ≤ x v }
Post vaporization zone: R2 : { x | x v ≤ x ≤ x f }
Oxidizer zone: R3 : { x | x f ≤ x < ∞ }
where xv and x f are locations of the vaporization front and flame sheet, respectively. As previously
379
Energies 2018, 11, 2945
mentioned, solid particles turned into the gaseous phase asymptotically once they passed the
vaporization front. Therefore, the following conditions were considered:
−∞ < x ≤ xv : ys = 1
(29)
xv ≤ x < ∞ : ys = 0
The following boundary conditions were applied in the analysis of the counter-flow
non-premixed flame:
y F = 0, yO = 0, ys = 1, θ = 0 @ x → −∞ (30)
y F = 0, yO = α, ys = 0, θ = 0 @ x → +∞ (31)
where α was obtained using the following correlation [38]:
YO,∞
α= (32)
ϑYF−∞
The above boundary conditions were defined for the temperature and mass fraction of the fuel
and oxidizer at x = −∞ and x = +∞. Nevertheless, to detect the positions of the vaporization front
(xv ) and flame sheet (x f ), and the corresponding temperature and mass fractions, and enforce the
continuity throughout the system, additional conditions were required at x = xv and x = x f . It should
be pointed out that at the flame sheet, θ = θ f and y = y F , and at the vaporization front, θ = θv and
y = y Fv , which must be solved. Since the current analysis was conducted based on an asymptotic
concept, the vaporization and reaction processes occurred within a very thin zone. At the vaporization
front, the convection and reaction terms were neglected compared to the diffusion and vaporization
terms. Moreover, at the flame sheet position, the convection and vaporization terms can be disregarded
against the reaction and diffusion terms. By using the aforementioned explanations and integrating
Equations (24), (26) and (27) from xv− to xv+ and x − +
f to x f , the following jump conditions were obtained
at x = xv and x = x f , respectively, as below:

1 dy F dθ dyO
− = = qxv , = 0, [y F ] = [yO ] = [θ ] = 0 @ x = xv (33)
Le F dx dx dx

1 dyO 1 dy F dθ
= =− , [ y F ] = [ yO ] = [ θ ] = 0 @ x = xf (34)
LeO dx Le F dx dx
where [] (square brackets) is defined as [] =( )+ − ( )− . No reaction occurred at the vaporization front.
Therefore, the following governing equations were used for the modeling of the vaporization process:
d2 θ dθ q
+ x − fe ys H ( T − Tv ) = 0 (35)
dx2 dx aτvap
1 d2 y F dy ys
+ x F + fe H ( T − Tv ) = 0 (36)
Le F dx2 dx aτvap
1 d 2 yO dy
+ x O = 0. (37)
LeO dx2 dx
By integrating the above equations from xv− to xv+ , the following equations were achieved:

1 dy F dθ
− = = qxv (38)
Le F dx dx

dyO
=0 (39)
dx
380
Energies 2018, 11, 2945
Equations (38) and (39) describe the jump conditions at the vaporization front (xv ). In order to obtain
the jump condition at the flame sheet (x f ), the vaporization terms were neglected relative to the
diffusion and reaction terms. Further, since the value of the Zeldovich number was presumed to be
very large, the thickness of the reaction zone would be very small and the convection term can be
neglected. Hence, the following equations were used to model the reaction process:

d2 θ Ta
= − f e D C y F y O exp − (40)
dx2 T
1 d2 y F Ta
= f e Dc y F yO exp (− ). (41)
Le F dx2 T
By aggregating Equations (40) and (41), the following equation was achieved:
d2 θ 1 d2 y F
2
+ = 0. (42)
dx Le F dx2
By performing a similar procedure for the oxidizer, the following equation was obtained:
1 d 2 yO Ta
= f e Dc y F yO exp (− ). (43)
LeO dx2 T
The following equation was determined by aggregating Equations (40) and (41):
d2 θ 1 d 2 yO
2
+ = 0. (44)
dx LeO dx2
Finally, by integrating Equations (42) and (44) from x − +

f to x f , the jump condition at the flame
sheet was achieved as follows:

1 dy F 1 dyO dθ
= =− . (45)
Le F dx LeO dx dx
2.5.4. Solution of the Governing Equations

In this subsection, the solution of the non-dimensionalized governing equations is presented
considering the boundary and jump conditions. It must be noted that for solving the complex equations,
Mathematica and Matlab software were employed.
Zone R1 : −∞ ≤ x ≤ xυ
Temperature distribution
In this zone, the temperature distribution was obtained by solving Equation (35):

π x
θ = C1 er f ( √ ) + C2 (46)
2 2
Considering the presented boundary conditions, C1 and C2 were calculated and the following
function was achieved for the temperature distribution:

er f√x +1
θ= θv 2 . (47)
er f √ + 1
x v
2
Oxidizer mass fraction
381
Energies 2018, 11, 2945
By solving Equation (37), the following equation was obtained for the oxidizer mass fraction:
⎛ ⎞

1 π x
yO = C1 er f ⎝ √ ⎠ + C2 . (48)
LeO 2 2 1
LeO
By applying the boundary conditions, C1 and C2 were found to be zero and therefore the mass
fraction of the oxidizer will be zero:
yO = 0. (49)
Fuel mass fraction

By solving Equation (36), the following equation was obtained for the fuel mass fraction in
this zone: ⎛ ⎞

1 π ⎝ x ⎠ + C2 .
y f = C1 er f √ (50)
Le f 2 2 1
Le f
By using the boundary conditions presented in Equation (51), the fuel mass fraction function was
calculated as presented in Equation (52).
x = −∞ → y F = 0, x = xv → y F = y Fv (51)
⎛ ⎞
y Fv [er f ⎝ √ x
⎠ + 1]
2 1
Le f
yf = ⎛ ⎞ . (52)
er f ⎝ √ xv ⎠+1
2 1
Le f
Zone R2 : xυ ≤ x ≤ x f
As previously mentioned, the solid fuel particles were immediately turned into a gaseous fuel
once they crossed the vaporization front. In this zone, the following boundary conditions were applied:
x = xv → θ = θv , x = x f → θ = θ f . (53)
By solving Equation (35) and using these boundary conditions, the temperature distribution of a
gaseous fuel was obtained as follows:

x xv θ f − θv
θ = θv + (er f √ − er f √ ) x (54)
2 2 er f √f − er f √ xv
2 2

In this zone, there was no oxidizer; thus, the mass fraction of the oxidizer was found to be zero.
yO = 0. (55)
Fuel mass fraction

The fuel mass fraction was obtained by solving Equation (36) and applying the following
boundary conditions:
x = xv → y F = y Fv , x = x f → y F = 0 (56)
382
Energies 2018, 11, 2945
⎛ ⎞ ⎛ ⎞
xf
[er f ⎝ √ x

1
⎠ − er f ⎝ √
1
⎠]
2 Le f 2 Le f
y F = y Fv ⎛ ⎞ ⎛ ⎞ . (57)
xf
[er f ⎝ √ x
v
1
⎠ − er f ⎝ √
1
⎠]
2 Le f 2 Le f
Zone R3 : x f ≤ x ≤ +∞
In this zone, no gaseous fuel existed. The temperature distribution can be achieved by calculating
Equation (35) and using the following boundary conditions:
x = x f → θ = θ f , x = +∞ → θ = 0 (58)

er f √x −1
θ = θf x2 . (59)
er f √ f
−1
2

The oxidizer mass fraction was obtained by solving Equation (37), considering the below
boundary conditions:
x = x f → yO = 0, x = +∞ → yα = 0 (60)

xf
[er f x
√ 1 − er f √ 1 ]
2 Le 2 Le
yO = α O
O
. (61)
xf
1 − er f √ 1
2 Le O
Fuel mass fraction

In this zone, there was no fuel. Hence, the fuel mass faction was found to be zero:
y F = 0. (62)
It is notable that the values of the five parameters including the flame temperature, the fuel
mass fraction, the oxidizer mass fraction and the location of the vaporization front and flame sheet
were unknown.
2.6. Flame Zone Analysis

In order to study the flame zone, as previously expressed, it was assumed that the quantity of
the Zeldovich number was very large so the thickness of the flame zone will be very small. In this
regard, employing an asymptotic approach could be a promising technique for analyzing this zone.
The Zeldovich number was calculated using the following correlation:
EQYF−∞
Ze = (63)
RCT 2f
where E, R and T are the reaction activation energy, the universal gas constant and the flame
temperature, respectively. To evaluate the critical extinction of the flame, a reduced Damkohler
number was applied. Below is the critical extinction of the Damkohler number [41]:
δ0E ≈ 2e( Z f − 2Z2f + 1.04Z3f + 0.44Z4f ) (64)
383
Energies 2018, 11, 2945
x
where Z f = 12 er f c( √f ) and δ0 are an order of unity that can be approximated using the
2
following equation:
1
δ = δ0 + O( ) (65)
Ze
where δ is the reduced Damkohler number that can be calculated using the following equation:

8π exp ( x2f ) DLeO Le f Z2f Ta
δ= exp − (66)
α2 FO2 f Ze3 Tf
where FO f is defined as follows [38]:

⎛ ⎞
1 − 0.276
t f + t2
2.15
tO f ⎜ f ⎟
FO f ≡ FO ( x f , LeO ) = ⎝ 2.15 ⎠
(67)
t LeO f 1 − tO f + tO f 2
0.276
!
tO f = 1 + 0.33333x f LeO (68)
t f = 1 + 0.33333x f . (69)
By substituting Equation (64) into Equation (66), the following correlation was obtained:

α2 FO2 f Ze3 exp Ta
Tf − x2f + 1
D0E ≈ (1 − 2Z f + 1.04Z2f + 0.44Z3f ). (70)
4πLeO Le f Z f
According to Equation (70), it can be readily implied that DOE mainly depends on the flame
sheet position (x f ) and the flame temperature (T f ). The ratio of a to a0 (for which LeO = Le F = 1) is
presented in the following:

a LeO Le f T f 6 η 2 Ta T 0f
0 2
( ) = ( ) exp 1 − + ( x 2
f − ( x f ) ) . (71)
a0 crit FO2 f T 0f d E T 0f Tf
Equation (71) describes the critical strain rate as a function of the fuel and oxidizer Lewis numbers.
In this equation, η and d E are expressed as follows [20]:
Zf
η= (72)
Z0f
1 − 2Z f + 1.04Z2f + 0.44Z3f
dE = η 2 3
. (73)
1 − 2Z f 0 + 1.04( Z f 0 ) + 0.44( Z f 0 )
2.7. Calculation of the Thermophoretic Force
Zone R1 : −∞ ≤ x ≤ xυ

er f √x +1
By applying θ = θv 2
xv
into Equation (47), the thermophoretic forcewas obtained as follows:
er f √ +1
2

x
T = ( er f √ + 1 × 1107.65) + 300 (74)
2
√ 2
dT 2 exp (− x2 )
∇T (x) = = 1107.65 1
(75)
dx π 2
384
Energies 2018, 11, 2945
x2
FTh,continuum = −3.3132 × 10−7 × exp (− ) (76)
2
According to the definition of the thermophoretic force, this force mainly depends on the
temperature gradient and the radius of the fuel particles. As previously mentioned, after the
vaporization front, the solid fuel particles were completely transformed into a gaseous fuel. In this
regard, the radius of the particles and subsequently the value of the thermophoretic force will be
almost zero.

The properties of the lycopodium biofuel used in the present study are listed in Table 2.
Table 2. Properties of the biofuel particles and oxidizer [15,41,48].
Property Value Property Value

kg kJ
ρp 1000 Cp 5.67kg·K
m3
ρa
kg
1.2 m3 Ca 1.001 kgkJ·K
kJ
Q 64,895.4 kg q 0.4
r 12 μm n 12 Giga
Tin 300 K Tvap 650 K
v 2 yO,+∞ 0.13
In this section, changes in flame temperature, the flame sheet position, the mass fractions of
the biofuel and oxidizer, the flow strain rate and the thermophoretic force with Lewis number, the
position and the equivalence ratio were studied. In order to follow the temperature and mass fraction
distributions of the fuel and oxidizer, solutions of the mass and energy conservation equations
were used considering the afore-presented boundary and jump conditions. As mentioned earlier,
the locations of the flame sheet and vaporization front were measured relative to the location of
the stagnation plane whose position was employed as the reference coordinate. As assumed in
References [11,48], the fuel supplied by the vaporization of the lycopodium particles was gaseous
methane. The complete reaction of gaseous methane with air is considered as follows:
CH4 + 2(O2 + 3.76N2 ) → CO2 + 2H2 O + (7.52)N2 . (77)
By taking into account the volatility of the biofuel, the effective gas phase equivalence ratio is
defined as follows [11,48]:
17.18YF−∞
φu = (78)
1 − YF−∞
where YF−∞ was calculated using the following equation [11,48]:
4
3 πr p n p ρ p
3
YF−∞ = . (79)
ρ
In order to include the effect of particle porosity, the following correlation was applied:
4
πr p 3 n p ρ p
YF−∞ ∗ = f e 3 (80)
ρ
Figure 4 shows the variations of flame temperature (K) with the fuel Lewis number for several
porosity factors when the oxidizer Lewis number was unified. With regard to the definition of the
Lewis number in Equation (11) (the ratio of thermal diffusivity to mass diffusivity), an increase of the
Lewis number leads to a reduction in the fuel mass fraction. Therefore, an accessible fuel reaching
the reaction zone declined, which leads to a decrease in flame temperature. As can be observed,
385
Energies 2018, 11, 2945
an increment in the fuel Lewis number from 0.2 to 1.4 resulted in a gradual increase of the flame
kg
temperature between ~1355 and 1615 K for a lycopodium mass concentration of 1000 m3 and a unity
value of the porosity factor. It is worth mentioning that decreasing the porosity factor caused a decline
in the flame temperature.
1650 fe=1
fe=0.98
1600 fe=0.92
Flame temperature (K)
1550
1500
1450
1400
1350
1300
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Fuel Lewis number
Figure 4. Flame temperature against the fuel Lewis number for several porosity factors.
Figure 5 shows the change in the flame sheet position with the fuel Lewis number for several
porosity factors assuming that a unity value of the oxidizer Lewis number. Again, considering the
definition of the Lewis number, increasing the Lewis number leads to a reduction in the fuel mass
fraction and flame temperature. Thus, the amount of available fuel for the combustion process
decreased leading to the movement of the flame sheet toward the fuel nozzle (left-hand side of the
stagnation plane). According to Figure 5, increasing the fuel Lewis number from 0.2 to 1.4 changed the
flame sheet position from 0.0967 to −0.1545 when the value of the porosity factor was unity.
fe=1
fe=0.98
0.15
fe=0.92
Flame sheet position (mm)
0.1
0.05
0
-0.05
-0.1
-0.15
-0.2
-0.25
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Fuel Lewis number
Figure 5. Flame sheet position against the fuel Lewis number for different porosity factors.
Figure 6 indicates a change in flame temperature with the oxidizer Lewis number for different
porosity factors when the fuel Lewis number was unity. As it is seen in Figure 6, an increase of the
oxidizer Lewis number caused a decline in the mass diffusivity of the oxidizer (according to the
386
Energies 2018, 11, 2945
Lewis number formula) leading to a decline in the flame temperature. For the considered range of the
oxidizer Lewis number from 0.3 to 1, the flame temperature varied from 1955 K to 1391 K.
fe=1
1955 fe=0.98
1895 fe=0.92
1835
1775
1715
1655
1595
1535
1475
1415
1355
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
Oxidizer Lewis number

Figure 6. Flame temperature against the oxidizer Lewis number for several porosity factors.
Figure 7 represents the change in the flame sheet position with the oxidizer Lewis number for
different porosity factors and a unity magnitude of the fuel Lewis number. According to the figure,
an increment in the oxidizer Lewis number caused a rise in the value of the flame sheet position.
In other words, the flame sheet shifted toward the oxidizer nozzle. Based on the Lewis number
formula, increasing the oxidizer Lewis number reduces the mass diffusivity of oxidizer. Accordingly,
the flame sheet moved toward the oxidizer zone. Regarding Figure 7, for the considered values of
the oxidizer Lewis number from 0.3 to 1, the flame sheet position changed gradually from −0.533 to
−0.119. Moreover, increasing the particle porosity factor increased the flame temperature and shifted
the flame sheet toward the oxidizer zone. With an increasing porosity factor, the available fuel mass for
the reaction decreased so the flame sheet would be closer to the fuel nozzle. For validation purposes,
the current results were compared to the numerical results reported by Wang [42] for a methane-air
mixture under the same conditions (the unity values of the porosity factor and the fuel Lewis number).
As represented in Figure 7, there was consistency between the present analytical results and the data
provided by Reference [42] for the case in which the porosity factor was unity.
387
Energies 2018, 11, 2945
2
1.8
Flame sheet position (mm)
1.6
1.4
1.2
1
fe=1
0.8
fe=0.98
0.6
fe=0.92
0.4 previous work[42]
0.2
0
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
Figure 7. Flame sheet position against the oxidizer Lewis number for different porosity factors.
In Figure 8, the effect of the fuel Lewis number on the temperature distribution of the flame
is shown for different mass particle concentrations. With regard to Figure 8, an increase of the fuel
Lewis number resulted in a decrease of the flame temperature, which can readily be explained by the
definition of the Lewis number. According to this figure, by increasing the fuel Lewis number from 0.5
to 1, the flame temperature varied from about 1380 to 1485 K. Furthermore, by increasing the value of
the mass particle concentration, the amount of available fuel for the formation of the flame increased,
resulting in an increase of the flame temperature.
p=1000
p=950
1500 p=900
1480
1460
1440
1420
1400
1380
1360
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
Fuel Lewis number

Figure 8. Flame temperature against the fuel Lewis number for several values of mass
particle concentration.
Figure 9 depicts the effect of the fuel Lewis number on the location of the flame sheet considering
several different mass particle concentrations. By increasing the fuel Lewis number from 0.5 to 1,
the flame sheet position ranged from −0.096 to −0.01765 when the mass particle concentration was
900 kg/m3 . Increasing the Lewis number of the fuel decreased the amount of fuel mass approaching
the reaction zone, which shifted the flame sheet toward the fuel nozzle. It is notable that the mass
388
Energies 2018, 11, 2945
fraction of the fuel for the reaction process increased with an increase in the mass particle concentration.
Therefore, further fuel mass diffusivity pushed the flame sheet toward the oxidizer zone.
p=1000
0 p=950
p=900
Flame sheet postion (mm)
-0.05
-0.1
-0.15
-0.2
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
Fuel Lewis number

Figure 9. Flame sheet position against the fuel Lewis number for different mass particle concentrations.
The variation of the flame temperature with an effective equivalence ratio (φu ) is drawn in
Figure 10 for different lycopodium radii. As it is observed in Figure 10, an increase of the equivalence
ratio resulted in an increment in the flame temperature. In addition, increasing the size of the fuel
particles (from 8 to 18 μm) caused a reduction in the flame temperature. An enhancement of the
particle size decreased the ratio of the particle surface to the particle volume. On this basis, a portion of
released heat during the reaction was consumed for the preheating and vaporization of the lycopodium
particles with larger diameters that led to a decrease in the flame temperature.
2000
1800
1600
1400
1200
1000
800 r=8
600 r=10
400 r=12
200 r=18
0
1 1.1 1.2 1.3 1.4 1.5 1.6
Equivalence ratio
Figure 10. Flame temperature against the effective equivalence ratio (φu ) considering several
lycopodium radii.
389
Energies 2018, 11, 2945
Figure 11 represents the change in the temperature distribution with the flame position for
various oxidizer Lewis numbers when the fuel Lewis number was unity. According to Figure 11, the
temperature of the fuel particles grew until it approached the maximum temperature corresponding
to the flame sheet position. The temperature of the oxidizer also grew until it reached the flame
temperature (right-hand side of the stagnation plane). Reducing the oxidizer Lewis number from 1 to
0.4 caused a decrease in the flame temperature and a shift of the flame sheet toward the oxidizer nozzle.
1900
Leo=1,Lef=1
1700 Leo=0.8,Lef=1
Leo=0.6,Lef=1
Temperature (K)
1500
1300 Leo=0.4,Lef=1
1100
900
700
500
300
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Position
Figure 11. Temperature distribution against position for different oxidizer Lewis numbers.
In Figure 12, changes in the gaseous fuel and oxidizer mass fractions with the flame position are
described for several different mass particle concentrations. Regarding Figure 12, the gaseous fuel
mass fraction increased until it reached the vaporization front at which the mass fraction of gaseous
fuel was the lowest. Afterwards, the gaseous fuel was oxidized and consumed in the reaction zone so
its mass fraction reduced to zero. From the opposite side, the oxidizer flowed toward the reaction zone
and was consumed for the generation of the flame. Therefore, its mass fraction decreased as it moved
closer to the flame front. Similarly to the fuel mass fraction, the oxidizer was entirely consumed at
the flame front position and its mass fraction became zero. It can also be implied that increasing the
value of the porosity factor increased the gaseous fuel mass fraction. Therefore, the vaporization front
moved toward the oxidizer nozzle.
1.2
Oxidizer and gaseous fuel mass
0.8
yo(fe=1)
fractions
yo(fe=0.98)
0.6
yo(fe=0.92)
0.4 yf(fe=1)
yf(fe=0.98)
0.2 yf(fe=0.92)
0
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Position (mm)
Figure 12. Oxidizer and gaseous fuel mass fractions against the position for different mass
particle concentrations.
390
Energies 2018, 11, 2945
Figures 13 and 14 illustrate the effects of the fuel and oxidizer Lewis numbers on the critical
strain rate ( aa0 ) for different porosity factors. According to Figure 13, for unity values of the fuel Lewis
number and the porosity factor, increasing the oxidizer Lewis number led to a decline in the ratio of aa0 .
Moreover, decreasing the value of the porosity factor decreased the value of x f leading to a decrease
of aa0 . It is notable that by increasing the value of aa0 , the flame sheet shifted toward the oxidizer
zone. Figure 14 demonstrates the variation of aa0 with the fuel Lewis number considering different
porosity factors when the oxidizer Lewis number was unity. Regarding this figure, the value of aa0
decreased as the fuel Lewis number increased. Decreasing the fuel Lewis number incremented the
value of the flame sheet position (x f ) and so a remarkable enhancement in the value of the exponential
term in Equation (71) that resulted in an increase in the value of aa0 . In order to validate the results, a
comparison was made between the results obtained for aa0 in this investigation and the results provided
by Seshadri and Trevino [41] for the critical strain rate under the same conditions (Ta = 30,000 K and the
oxidizer Lewis number was unity). With regard to Figure 14, a reasonable agreement exists between
the compared results under the considered conditions.
Figures 15 and 16 delineate the impact of the oxidizer and fuel Lewis numbers on the
thermophoretic force for the unity value of the porosity factor, respectively. According to these
figures, until reaching the vaporization front, the thermophoretic force continuously reduced with
the flame position. As the oxidizer and fuel Lewis numbers decreased, the maximum temperature
occurring at the reaction zone increased. In this regard, a temperature gradient and subsequently the
value of the thermophoretic force increased. Further, it can be seen that by decreasing the oxidizer and
fuel Lewis numbers, the location of the minimum thermophoretic force moved toward the fuel nozzle.
As there was no solid fuel after the vaporization front, the value of r p and the thermophoretic force
will be zero.
7 T_a=10000 K
fe=1
6
fe=0.98
5
Critical strain rate
fe=0.92
4
0
0 0.2 0.4 0.6 0.8 1 1.2
Figure 13. Critical flow strain rate against the oxidizer Lewis number for several porosity factors.
391
Energies 2018, 11, 2945
T_a=30000 K
14
12
fe=1
Critical Strain Rate
10 fe=0.98
fe=0.92
8
previous work[41]
6
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Fuel Lewis number
Figure 14. Critical flow strain rate against the fuel Lewis number for different porosity factors.
3E-09-09
3.00×10
-2E-09-09
-2.00×10
Thermophoretic force (N)
Leo=1,Lef=1
- -7E-09
7.00×10-09 Leo=0.8,Lef=1
Leo=0.6,Lef=1
-1.20×10-08 Leo=0.4,Lef=1
-1.2E-08
-1.70×10
-1.7E-08
-08
-2.20×10 -08
-2.2E-08
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Position
Figure 15. Thermophoretic force against the position for different oxidizer Lewis numbers.
7E-09-09
7.00×10
2E-09-09
2.00×10
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
-3E-09-09
-3.00×10
Leo=1,Lef=1
-8E-09-09
-8.00×10
Leo=1,Lef=0.8
--1.3E-08
1.30×10-08
Leo=1,Lef=0.6
-1.8E-08-08
-1.80×10 Leo=1,Lef=0.4
Position
Figure 16. Thermophoretic force against position for different fuel Lewis numbers.
392
Energies 2018, 11, 2945
The impact of porosity on the thermophoretic force is illustrated in Figure 17 taking into account
the unity values of the oxidizer and fuel Lewis numbers. According to this figure, an increase of the
porosity factor enhanced the thermophoretic force. By increasing the porosity factor, the accessible
fuel that approached the flame zone decreased leading to a decline in the value maximum flame
temperature. In this regard, the temperature gradient term in Equation (8) decreased causing a decline
in the value of the thermophoretic force. Similarly to Figures 15 and 16, since no solid fuel existed after
the vaporization front, the value of r p and then the thermophoretic force will be zero.
1.00E-08-08
1.00×10
fe=1
5.00E-09
5.00×10-09 fe=0.98
fe=0.92
0.00E+00
0.00×10+00
-5.00E-09
-5.00×10-09
-1.00×10-08
-1.00E-08
-1.50E-08 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
-1.50×10-08
Position
Figure 17. Thermophoretic force against position for different porosity factors.
4. Conclusions
Due to its appropriate dispersibility and flammability, lycopodium has been used as a reasonable
reference organic fuel for testing different combustion systems and recognizing the processes occurring
in the system prior to industrial scale-up. To provide helpful theoretical results about the effective
parameters (i.e., the porosity and thermophoretic force), this paper used a mathematical approach
to study the structure of multi-zone counter non-premixed flames fueled by lycopodium particles.
Preheat, vaporization, flame and oxidizer zones were considered for the flame. Mass and energy
equations were obtained for the zones. In this analysis, the porosity of the biofuel particles was
also taken into account. Further, the vaporization rate and thermophoresis effects were modelled.
Changes in the flame temperature and flame sheet position with fuel and oxidizer Lewis numbers were
presented by considering different porosity factors. The impacts of the effective equivalence ratio, the
flow critical strain rate, and the gaseous fuel and oxidizer mass fractions were evaluated. Finally, the
thermophoretic force caused by the temperature gradient was obtained for different porosity factors,
fuel and oxidizer Lewis numbers. For the case in which the oxidizer and fuel Lewis numbers were
equal to 0.4 and 1, a maximum temperature of the flame was found to be ~1860 K. The maximum mass
fraction of gaseous fuel that occurred at the vaporization front was found to be 0.44 when the value
of the porosity factor was unity. In this analysis, for the unity value of the fuel Lewis number, the
minimum value of the thermophoretic force was found to be −1.48 × 10−8 N, when the oxidizer Lewis
number was 0.4. On the other hand, for the unity value of oxidizer Lewis number, the minimum value
of the thermophoretic force was equal to −1.53 × 10−8 N at a fuel Lewis number of 0.4. By considering
the porosity factor, the minimum value of the thermophoretic force occurred at a unity value of the
porosity factor and its value was found to be −1.28 × 10−8 N. Studying the effects of oscillation and
393
Energies 2018, 11, 2945
instability and calculating the quenching distance for counter-flow non-premixed combustion can be
considered in future works. The main conclusions of the present study are summarized as follows:
• The flame temperature increased by decreasing the volume porosity, fuel and oxidizer
Lewis numbers.
• The flame sheet position moved toward the fuel nozzle with an increasing volume porosity and
fuel Lewis number.
• The flame sheet position shifted toward the fuel nozzle by decreasing the oxidizer Lewis number.
• The thermophoretic force increased by decreasing the volume porosity, fuel and oxidizer
Lewis numbers.
• The critical strain rate increased by decreasing the volume porosity, fuel and oxidizer
Lewis numbers.
Author Contributions: Project Administration, M.B.; Methodology and Analysis, P.G.N.; Methodology and
Validation, H.R. and S.S.; Writing-Original, S.S.; Writing-Review & Editing, B.S.; Analysis-Review M.B. & B.S.
Nomenclature
a Strain rate
C Mean thermal velocity of gas
molecular

kJ
Ca gaseous phase specific heat kg·K
Ck temperature jump coefficient

kJ
Cp Solid particle specific heat kg·K
Ct Temperature creep coefficient
Cm Velocity jump coefficient
Cs Gas velocity discontinuities coefficient
DC Damkohler number
DF Mass diffusivity coefficient of gaseous fuel (m2 /s)
DO Mass diffusivity coefficient of oxidizer (m2 /s)
D0E Critical Damkohler number
DT Thermal diffusivity coefficient (m2 /s)
E activation energy (kj)
er f i ( x ) Error function
fe Porosity factor
H Heavi side function
kJ
kg Gas thermal conductivity m·s·K
kJ
kp Lycopodium thermal conductivity m·s·K
kT Constant Defined in Equation (4)
L Mean free path
Le Lewis number
kg
m Mixture molecular mass ( mol )
kg
mf Fuel molecular mass ( mol )
kg
mO Oxygen molecular mass ( mol )
np Number of particle per volume unit
394
Energies 2018, 11, 2945

kJ
Q Reaction heat per unit of fuel mass kg
q Dimensionless heat
m3 Pa
R Universal constant of gases mol·K
rp Particle radius
T Fuel temperature (K)
Ta activation temperature (K)
Tf Flame temperature (K)
Tv Particle Start temperature of vaporization (K)
WF molecular weight of fuel
x Dimension length
xf Flame position
xv Vaporization front position
YF Gaseous fuel mass fraction
YO Oxidizer mass fraction
Ys Particle mass fraction
yF Dimensionless fuel mass fraction
yO Dimensionless oxidizer mass fraction
ys Dimensionless mass fraction of solid particles
Z Secondary coordinate axis
Ze Zeldovich number
Greek symbols
α Initial mass fraction of oxidizer
ε Volume porosity
Θ Dimensionless Temperature
kJ
λ Thermal conductivity of fuel or oxidizer m·s·K
μ Dynamic viscosity
υF Fuel stoichiometric coefficient
υO oxidizer stoichiometric coefficient
υ product Product stoichiometric
coefficient
kg
ρ Density m3
kg
ρa Gaseous phase density
m3
kg
ρp Density of Solid particle m3
τvap constant time characteristic
of vaporization

kg
ωv Particle vaporization rate
m · s2
kg
ωF Rate of chemical reaction m · s2
∅ Constant and equal to 0.941
References
1. Caputo, A.C.; Palumbo, M.; Pelagagge, P.M.; Scacchia, F. Economics of biomass energy utilization in
combustion and gasification plants: Effects of logistic variables. Biomass Bioenergy 2005, 28, 35–51. [CrossRef]
2. Wahlund, B.; Yan, J.; Westermark, M. Increasing biomass utilisation in energy systems: A comparative study
of CO2 reduction and cost for different bioenergy processing options. Biomass Bioenergy 2004, 26, 531–544.
[CrossRef]
3. Demirbas, A.H.; Demirbas, I. Importance of rural bioenergy for developing countries. Energy Convers. Manag.
2007, 48, 2386–2398. [CrossRef]
4. Hall, D.O.; Scrase, J.I. Will biomass be the environmentally friendly fuel of the future? Biomass Bioenergy
1998, 15, 357–367. [CrossRef]
5. Hu, Y.; Kurose, R. Nonpremixed and premixed flamelets LES of partially premixed spray flames using a
two-phase transport equation of progress variable. Combust. Flame 2018, 188, 227–242. [CrossRef]
6. Bhoi, P.R.; Channiwala, S.A. Emission characteristics and axial flame temperature distribution of producer
gas fired premixed burner. Biomass Bioenergy 2009, 33, 469–477. [CrossRef]
395
Energies 2018, 11, 2945
7. Akbarzadeh, M.; Birouk, M. Effect of Fuel Nozzle Geometry on the Stability of Non-Premixed Turbulent
Methane Flame. In Proceedings of the 23rd ICDERS, Irvine, CA, USA, 24–29 July 2011.
8. Josh, N.D.; Berlad, A.L. Gravitational effects on stabilized, premixed, lycopodium-air flames.
Combust. Sci. Technol. 1986, 47, 55–68. [CrossRef]
9. Berlad, A.L.; Joshi, N.D. Gravitational effects on the extinction conditions for premixed flames. Acta Astronaut.
1985, 12, 539–545. [CrossRef]
10. Berlad, A.L.; Tangirala, V.; Facca, L.R. Radiative structures of lycopodium-air flames in low gravity.
J. Propuls. Power 1991, 7, 5–8. [CrossRef]
11. Seshadri, K.; Berlad, A.L.; Tangirala, V. The structure of premixed particle-cloud flames. Combust. Flame 1992,
89, 333–342. [CrossRef]
12. Han, O.S.; Yashima, M.; Matsuda, T.; Matsui, H.; Miyake, A.; Ogawa, T. Behavior of flames propagating
through lycopodium dust clouds in a vertical duct. J. Loss Prev. Process. Ind. 2000, 13, 449–457. [CrossRef]
13. Han, O.S.; Yashima, M.; Matsuda, T.; Matsui, H.; Miyake, A.; Ogawa, T. A study of flame propagation
mechanisms in lycopodium dust clouds based on dust particles’ behavior. J. Loss Prev. Process. Ind. 2001, 14,
14. Xi, Z.; Fu, Z.; Hu, X.; Sabir, S.; Jiang, Y. An experimental investigation on flame pulsation for a swirl
non-premixed combustion. Energies 2018, 11, 1757. [CrossRef]
15. Proust, C. Flame propagation and combustion in some dust-air mixtures. J. Loss Prev. Process. Ind. 2006, 19,
16. Skjold, T.; Olsen, K.L.; Castellanos, D. A constant pressure dust explosion experiment. J. Loss Prev. Process. Ind.
2013, 26, 562–570. [CrossRef]
17. Shamooni, A.; Cuoci, A.; Faravelli, T.; Sadiki, A. Prediction of Combustion and Heat Release Rates in
Non-Premixed Syngas Jet Flames Using Finite-Rate Scale Similarity Based Combustion Models. Energies
2018, 11, 2464. [CrossRef]
18. Rahbari, A.; Shakibi, A.; Bidabadi, M. A two-dimensional analytical model of laminar flame in lycopodium
dust particles. Korean J. Chem. Eng. 2015, 32, 1798–1803. [CrossRef]
19. Bidabadi, M.; Esmaeilnejad, A. An analytical model for predicting counterflow flame propagation through
premixed dust micro particles with radiative heat loss. J. Loss Prev. Process. Ind. 2015, 35, 182–199. [CrossRef]
20. Bidabadi, M.; Ramezanpour, M.; Poorfar, A.K.; Monteiro, E.; Rouboa, A. Mathematical Modeling of a
Non-premixed Organic Dust Flame in a Counterflow Configuration. Energy Fuels 2016, 30, 9772–9782.
[CrossRef]
21. Xi, Z.; Fu, Z.; Hu, X.; Sabir, S.W.; Jiang, Y. An Investigation on Flame Shape and Size for a High-Pressure
Turbulent Non-Premixed Swirl Combustion. Energies 2018, 11, 930. [CrossRef]
22. Li, X.; Zhang, J.; Yang, H.; Jiang, L.; Wang, X.; Zhao, D. Combustion characteristics of non-premixed methane
micro-jet flame in coflow air and thermal interaction between flame and micro tube. Appl. Therm. Eng. 2017,
112, 296–303. [CrossRef]
23. Bidabadi, M.; Harati, M.; Afzalabadi, A.; Rahbari, A. Effect of Thermal Resistance on the Random Combustion
of Micro-Organic Dust Particles. J. Energy Eng. 2017, 144, 04017073. [CrossRef]
24. Spijker, C.; Raupenstrauch, H. Numerical investigation on inner particle effects in Lycopodium/Air dust
deflagrations. J. Loss Prev. Process. Ind. 2017, 49, 870–879. [CrossRef]
25. Chen, L.; Xia, Y.F.; Li, B.W.; Shi, J.R. Flame front inclination instability in the porous media combustion with
inhomogeneous preheating temperature distribution. Appl. Therm. Eng. 2018, 128, 1520–1530. [CrossRef]
26. Bidabadi, M.; Ebrahimi, F.; Bordbar, V. Modeling multi regional counter flow combustion of lycopodium
dust cloud with considering radiative heat loss. J. Cent. South Univ. 2017, 24, 2638–2648. [CrossRef]
27. Ji, W.; Yu, J.; Yu, X.; Yan, X. Experimental investigation into the vented hybrid mixture explosions of
lycopodium dust and methane. J. Loss Prev. Process. Ind. 2018, 51, 102–111. [CrossRef]
28. Di Benedetto, A.; Di Sarli, V.; Russo, P. On the determination of the minimum ignition temperature for
dust/air mixtures. Chem. Eng. Trans. 2010, 19, 189–194.
29. Sanchirico, R.; Russo, P.; Saliva, A.; Doussot, A.; Di Sarli, V.; Di Benedetto, A. Explosion of lycopodium-
nicotinic acid–methane complex hybrid mixtures. J. Loss Prev. Process. Ind. 2015, 36, 505–508. [CrossRef]
30. Sanchirico, R.; Russo, P.; Di Sarli, V.; Di Benedetto, A. On the explosion and flammability behavior of mixtures
of combustible dusts. Process. Saf. Environ. Prot. 2015, 94, 410–419. [CrossRef]
396
Energies 2018, 11, 2945
31. Sanchirico, R.; Di Sarli, V.; Russo, P.; Di Benedetto, A. Effect of the nozzle type on the integrity of dust
particles in standard explosion tests. Powder Technol. 2015, 279, 203–208. [CrossRef]
32. Kassebaum, J.L.; Chelliah, H.K. Oxidation of isolated porous carbon particles: Comprehensive numerical
model. Combust. Theory Model. 2009, 13, 143–166. [CrossRef]
33. Comtois, P. John Tyndall and the floating matter of the air. Aerobiologia 2001, 17, 193–202. [CrossRef]
34. Davies, C.N. Aerosol Science; London Academic Press: London, UK, 1966.
35. Braun, D.; Libchaber, A. Trapping of DNA by thermophoretic depletion and convection. Phys. Rev. Lett.
2002, 89, 188103. [CrossRef] [PubMed]
36. Talbot, L.R.; Cheng, R.K.; Schefer, R.W.; Willis, D.R. Thermophoresis of particles in a heated boundary layer.
J. Fluid Mech. 1980, 101, 737–758. [CrossRef]
37. Bakanov, S.P. The thermophoresis of solids in gases. J. Appl. Math. Mech. 2005, 5, 767–772. [CrossRef]
38. Rahbari, A.; Wong, K.F.; Vakilabadi, M.A.; Poorfar, A.K.; Afzalabadi, A. Theoretical investigation of particle
behavior on flame propagation in lycopodium dust cloud. J. Energy Resour. Technol. 2017, 139, 012202.
[CrossRef]
39. Bidabadi, M.; Natanzi, A.H.; Mostafavi, S.A. Thermophoresis effect on volatile particle concentration in
micro-organic dust flame. Powder Technol. 2012, 217, 69–76. [CrossRef]
40. Haghiri, A.; Bidabadi, M. Modeling of laminar flame propagation through organic dust cloud with thermal
radiation effect. Int. J. Therm. Sci. 2010, 49, 1446–1456. [CrossRef]
41. Seshadri, K.; Trevino, C. The influence of the Lewis numbers of the reactants on the asymptotic structure of
counterflow and stagnant diffusion flames. Combust. Sci. Technol. 1989, 64, 243–261. [CrossRef]
42. Wang, H.Y.; Chen, W.H.; Law, C.K. Extinction of counterflow diffusion flames with radiative heat loss and
nonunity Lewis numbers. Combust. Flame 2007, 148, 100–116. [CrossRef]
43. Bidabadi, M.; Panahifar, P.; Sadeghi, S. Analytical development of a model for counter-flow non-premixed
flames with volatile biofuel particles considering drying and vaporization zones with finite thicknesses. Fuel
2018, 231, 172–186. [CrossRef]
44. Linan, A. The asymptotic structure of counterflow diffusion flames for large activation energies.
Acta Astronaut. 1974, 1, 1007–1039. [CrossRef]
45. Fendell, F.E. Ignition and extinction in combustion of initially unmixed reactants. J. Fluid Mech. 1965, 21,
46. Huang, Y.; Risha, G.A.; Yang, V.; Yetter, R.A. Effect of particle size on combustion of aluminum particle dust
in air. Combust. Flame 2009, 156, 5–13. [CrossRef]
47. Wichman, I.S.; Yang, M. Double-spray counterflow diffusion flame model. Strain 1998, 2, 373–398. [CrossRef]
48. Rockwell, S.R.; Rangwala, A.S. Modeling of dust air flames. Fire Saf. J. 2013, 59, 22–29. [CrossRef]
397
energies
Article
Simultaneous Extraction and Emulsification of Food
Waste Liquefaction Bio-Oil
David Längauer 1,2 , Yu-Ying Lin 2 , Wei-Hsin Chen 2,3, *, Chao-Wen Wang 2 , Michal Šafář 1,2 and
Vladimír Čablík 1
1 Institute of Environmental Engineering, Faculty of Mining and Geology, VŠB—Technical University of
Ostrava, 17. listopadu 15, 708 33 Ostrava-Poruba, Czech Republic; [email protected] (D.L.);
[email protected] (M.Š.); [email protected] (V.Č.)
2 Department of Aeronautics and Astronautics, National Cheng Kung University, Tainan 701, Taiwan;
[email protected] (Y.-Y.L.); [email protected] (C.-W.W.)
3 Research Center for Energy Technology and Strategy, National Cheng Kung University, Tainan 701, Taiwan
* Correspondence: [email protected] or [email protected]; Tel.: +886-6-2004456
Received: 6 October 2018; Accepted: 31 October 2018; Published: 5 November 2018
Abstract: Biomass-derived bio-oil is a sustainable and renewable energy resource, and liquefaction is
a potential conversion way to produce bio-oil. Emulsification is a physical upgrading technology,
which blends immiscible liquids into a homogeneous emulsion through the addition of an emulsifier.
Liquefaction bio-oil from food waste is characterized by its high pour point when compared to diesel
fuel. In order to partially replace diesel fuel by liquefaction bio-oil, this study aimed to develop a
method to simultaneously extract and emulsify the bio-oil using a commercial surfactant (Atlox 4914,
CRODA, Snaith, UK). The solubility and stability of the emulsions at various operating conditions
such as the bio-oil-to-emulsifier ratio (B/E ratio), storage temperature and duration, and co-surfactant
(methanol) addition were analyzed. The results demonstrate that higher amounts of bio-oil (7 g)
and emulsifier (7 g) at a B/E ratio = 1 in an emulsion have a higher solubility (66.48 wt %). When
the B/E ratio was decreased from 1 to 0.556, the bio-oil solubility was enhanced by 45.79%, even
though the storage duration was up to 7 days. Compared to the emulsion stored at room temperature
(25 ◦ C), its storage at 100 ◦ C presented a higher solubility, especially at higher B/E ratios. Moreover,
when methanol was added as a co-surfactant during emulsification at higher B/E ratios (0.714 to 1),
it rendered better solubility (58.83–70.96 wt %). Overall, the emulsified oil showed greater stability
after the extraction-emulsification process.
Keywords: emulsification; liquefaction; bio-oils; co-surfactant; surfactant; diesel
1. Introduction
In the 21st century, food wastes in excess of one billion metric tons are generated worldwide every
year from various sources and such great amounts lead to a severe environmental problem. Failure to
recycle or treat food wastes properly will lead to the emission of methane (CH4 ) and carbon dioxide
(CO2 ) during bio-degradation. These greenhouse gases may cause human health and environmental
problems [1].
In recent years, there has been an improvement in thermochemical conversion equipment, which
can transform wastes into biofuels or bioenergy products [2,3]. Food waste is considered a valuable
source of energy due to its high biodegradability and high organic matter content [4,5]. Nevertheless,
there is a great challenge in the conversion of food wastes into biofuels by traditional thermochemical
conversions such as pyrolysis and gasification. The high moisture content of food waste requires more
time and energy for drying before the conversion proceeds.

Energies 2018, 11, 3031
Hydrothermal technology (HTL) provides opportunities for more energy efficient valorization of
wet biomass in that water is used directly as a reaction medium, thus, bypassing the drying process [6,7].
After HTL is performed, a separation procedure is needed to obtain the oil phase, also known as
the liquefaction bio-oil. The separation method includes using a gravimetric separation funnel or
extraction with an organic solvent such as tetrahydrofuran (THF), ethyl acetate, and dichloromethane
(DCM) [8,9]. The liquefaction bio-oil is a mixture of several hundred organic compounds such as acids,
alcohols, aldehydes, amides, esters, ketones, phenols, and lignin oligomers [10,11]. The liquefaction
bio-oil has undesirable features for industrial applications such as a high viscosity and high ash
content. This implies that the bio-oil composition is complex, making it relatively difficult to control
the performance during the combustion process.
Emulsification is an effective way for upgrading bio-oil and blending it with other fuels [12,13].
Emulsification technology is popular for upgrading pyrolysis oil. Since diesel and bio-oil have polar
molecules, the emulsifier or surfactants can disperse them both to form a stable and homogeneous
emulsion [12]. The emulsification of bio-oils or water in diesel using various emulsifiers is summarized in
Table 1. Michio [14] studied the emulsion of diesel oil and pyrolysis oil from hardwood and investigated
the effects of five operation parameters (bio-oil concentration, surfactant concentration, residence time,
motor speed and emulsification temperature). The results were optimized by a statistical model and the
dominant variables were identified for minimal stratification formation. These dominant parameters
were bio-oil concentration, and surfactant and power input. Chiaramonti [15] studied diesel/bio-oil
emulsions with four kinds of pyrolysis oils from Canadian oak, beech wood, California pine, and
pine wood. The results suggested that it is necessary to use fresh bio-oil and micronize these products
to obtain a stable emulsion. Guo [16] used a KDL-5 molecular distillation apparatus to separate the
middle fraction and the heavy fraction of the crude bio-oil. The blending of the diesel and bio-oil was
conducted by ultrasonic and ultrasonic-mechanical emulsification. The droplet size distribution was
measured using a Malvern nanometer particle size analyzer for the three emulsions (heavy fraction
with 40–300 nm, middle fraction with 15–60 nm and crude bio-oil with 8–25 nm) and had the same
tendency with the stability time 14 min, 216 min and 31 days, respectively. As for the emulsion on
the liquefaction bio-oil, Lijian. [17] employed micro-emulsion technologies to dissolve three kinds of
sewage sludge liquefaction bio-oil extracted by different solvents into diesel. The experiments showed
that the bio-oils obtained at a higher liquefaction temperature (360 ◦ C) benefited the micro-emulsion.
Table 1. Literature of bio-oil and liquefaction oil in diesel emulsions.
Pyrolysis
Mixing Method Emulsifier and Co-Surfactant Stability Ref.
Bio-Oil Content
Micro-emulsifier
Hypermer B246SF/2234
10–30 wt % (800–1750 rpm and 3 to 42 days [14]
(Croda International)
5–20 min)
Variable speed electrical Polymeric surfactants/short chain
25–75 wt % 3 days [15]
motor (1100 W) additives (n-octanol)
Ultrasonic (20 kHz; 2 min) Span80 (sorbitan monooleate)/Tween
and Ultrasonic-mechanical 80 (Polyoxyethylenesorbitan
10 wt % 27 to 31 days [16]
(20 kHz; 2 min; and monolaurate)/Span 85
5000 rpm; 5 min) (sorbitan trioleate)
Span 80 (sorbitan monooleate)/Tween
Homogenizer
2–40 wt % 80 (Polyoxyethylenesorbitan 5 days [18]
(8800–16,700 rpm)
monolaurate)
Liquefaction
Mixing Method Emulsifier or Co-Surfactant Stability Ref.
Bio-Oil Content
Manual shaking (1 min) and
5 wt % centrifugation (6000 rpm; Span 80 (sorbitan monooleate) - [17]
20 min)
Vortex mixer (3000 rpm; Atlox4914 (Croda This
5 to 30 wt % 7 days
2 min) International)/Methanol study
399
Energies 2018, 11, 3031
Despite some useful information provided from these studies, it should be underlined that there
is still a lack of emulsification information for the liquefaction of bio-oil from food waste. Moreover, the
mechanisms of extraction and emulsification are related to the surface tension, viscosity, and polarity
between the molecules [15–17]. Thus, this study aimed to investigate the feasibility of the simultaneous
emulsification-extraction of the food waste liquefaction bio-oil and diesel with the addition of an
emulsifier. This process can combine liquefaction bio-oil extraction and emulsification in one step,
which can reduce the experiment cost and time. The details of the proportion of the liquefaction bio-oil
in the emulsion after mixing with diesel are also addressed.
2.1. Materials
The raw liquefaction bio-oil was produced from the hydrothermal liquefaction process of food
waste, which was collected from the local food waste recycling field. It included rhizomes and peels of
several kinds of vegetables and fruits, which were obtained from livelihood waste. Before the start of
the liquefaction process, the size of the food waste (800 g) was reduced using a blender to intensify
the reaction surface area. The food waste had a high moisture content (74%) for the liquefaction
reaction, thus, dismissing additional water. Then, the food waste slurry was pretreated with a 5 wt %
potassium carbonate (K2 CO3 , alkali catalyst) by heating the mixture to 100 ◦ C for 1 h. Subsequently,
the temperature was raised to 320 ◦ C and held for 30 min as the conditions of liquefaction. The process
yielded products at three different phases: gas, liquid, and solid. The water vapor in the reactor was
removed by cooling to 105 ◦ C under atmospheric pressure. The leftover liquid/solid mixture was
heated to 100 ◦ C for 4 h in the oven to remove the excess water content in the mixture. Finally, the
remaining residue was considered as the raw liquefaction bio-oil having an average conversion rate
of 40 wt %.
The diesel fuel used for experiments was commercially available and sourced from petrol station
of China Petroleum Corporation (CPC). In the emulsification-extraction process, the liquefaction oil
was dispersed into the diesel, leading to the bulk of the emulsion volume. The emulsifier used in the
experiments was Atlox 4914 obtained from Croda International Plc. Methanol (>99 vol %) was used as
the co-surfactant. The properties of the bio-oil, diesel, Atlox 4914, and co-surfactant such as density,
viscosity, and higher heating value (HHV) are tabulated in Table 2.
Table 2. Properties of the bio-oil, diesel, surfactant, and co-surfactant.
Compound Density (kg m−3 ) Viscosity (mPa s) HHV (MJ kg−1 )

Bio-oil 1100.0 4916 31.41
Diesel 832.0 1.6–5.8 44.80
Atlox 4914 970.0 - 37.64
Methanol 791.0 0.533 20.40
2.2. The Procedure for Preparing the Sample

To determine the minimum volume required for the emulsifier, a fixed quantity of 20 g of diesel
was mixed with various amounts of the emulsifier (based on B/E ratio) in a 50 mL centrifuge tube.
A vortex mixer (Model: VM-2000, Digisystem Laboratory Instruments Inc., Taipei, Taiwan) was used
to mix the diesel and emulsifier for 1 min at a speed of 3000 rpm. The liquefaction bio-oil was in the
form of the solid phase at room temperature, so it was preheated to 100 ◦ C in the oven to transform it
into a liquid before the emulsification-extraction process. Thereafter, the liquid raw liquefaction bio-oil
was added into the diesel-emulsifier mixture and mixed by the vortex mixer for 2 min at a speed of
3000 rpm. Then, the emulsion samples were stored at room temperature (25 ◦ C). Samples, which
needed higher-temperature condition, were stored in an oven at 100 ◦ C. The co-surfactant could reduce
400
Energies 2018, 11, 3031
the surface tension of bio-oil and enhance the performance of the emulsions [13]. Therefore, further
emulsion experiments examined the effect of the adjunction of the co-surfactant by adding one more
step in the process of sample preparation. After the emulsions of diesel, emulsifier, and bio-oil were
prepared, methanol was added and mixed for 1 min at a speed of 3000 rpm. The detailed process is
shown in Figure 1.
Figure 1. Scheme of emulsion preparation and stability analysis.

There are two directions to investigate the performance of the extraction-emulsification
experiment: quantitative and qualitative methods. The quantitative research is to identify the amount
of bio-oil dissolved in diesel. The deposited bio-oil is in the solid phase, so it is easy to separate it from
the liquid emulsion and weigh the amount. In order to measure the extent of diesel-solvable bio-oil, a
parameter called solubility was defined as
m added bio−oil − minsolvable bio−oil

Solubility (%) = × 100% (1)
m added bio−oil
where m added bio−oil stands for the added mass of bio-oil and minsolvable bio−oil denotes the mass of
deposited bio-oil in the bottom of the tube.
Qualitative research is to evaluate the emulsion stability. Several methods can be used such as
measuring the pH value, moisture content, the viscosity of the emulsion, or droplet size distribution
in the emulsion [14,15]. However, there is no standard procedure for examining the long-term
stability [15]. In this study, Fourier transform infrared spectroscopy (FTIR) coupled with attenuated
total reflection (ATR) was used to examine the stability and homogeneity of the emulsions through the
identification of the functional groups present in the emulsions [12,13,19]. In some cases, even though
401
Energies 2018, 11, 3031
the emulsion seemed to be uniform by visual inspection, the FTIR spectra were different, revealing the
phase separation in the emulsion. For a homogeneous emulsion, the FTIR spectra at different places
would overlap. Therefore, the emulsion was divided into three regions, namely, the top, middle, and
bottom liquid levels in the sample tube and the three regions were analyzed using FTIR [20].
3.1. FTIR Spectra of Atlox, Bio-Oil, Diesel, and Methanol

FTIR analysis of the input components (Atlox, food waste liquefaction bio-oil, diesel, and
methanol) was performed and the spectra are shown in Figure 2. Diesel, methanol, and Atlox 4914 are
commercially available substances with distinct chemical compositions. To identify the stratification of
the bio-oil in the emulsions, the chemical composition of the liquefaction bio-oil was studied. From the
obtained spectra and functional groups summarized in Table 3, the liquefaction bio-oil contained
C-H-(CH2 ) (732 cm−1 ), C-O (1034 cm−1 ), C-O (1258 cm−1 ), C-H-(CH3 ) (1414 cm−1 ), CH3 (2918 cm−1 ),
and polymeric O-H (3320 cm−1 ) functional groups.
D $WOR[ E )RRG ZDVWH RLO

,QWHQVLW\ 7
,QWHQVLW\ 7

:DYHOHQJWK FP :DYHOHQJWK FP
F 'LHVHO G 0HWKDQRO

,QWHQVLW\ 7
,QWHQVLW\ 7

:DYHOHQJWK FP :DYHOHQJWK FP
Figure 2. FTIR spectra of (a) Atlox, (b) bio-oil, (c) diesel, and (d) methanol.
402
Energies 2018, 11, 3031
Table 3. Main functional groups of emulsifiers, co-surfactants and liquefaction bio-oil.
Wave Number Range (cm−1 ) Functional Group Mode of Vibration Reference

3550–3450 Dimetric O-H Stretching [21]
3400–3230 Polymetric O-H Stretching [21]
2949–2915 Aliphatic C-H Stretching [21,22]
2924–2858 Aliphatic C-H Stretching [23]
2835–2815 CH3 -O-CH3 (ethers) Stretching [21]
1740–1732 C=O (ester) Stretching [23]
1651–1640 C=C (aromatic) Skeletal [24]
1652–1579 C-C (ring) Stretching [21]
1486–1446 C-H- (CH3 )/(CH2 ) Symmetric Asymmetric [24]
1461–1377 C-H- (CH3 ) Symmetric [24]
1290–1211 C-O Stretching [21]
1230–1140 C-O (phenol) Stretching [21]
1239–1100 C-O-C (ester) Stretching [22,23]
970–960 C=C-H Bending [21]
730–710 C-H-(CH2 ) Bending [21]
3.2. Bio-Oil Solubility and Emulsion Stability at B/E Ratio = 1

By mixing various volumetric ratios of bio-oil/emulsifier and observing the emulsion’s stability,
Figure 3 examines the solubility of the bio-oil in the emulsions over time. In all cases, the B/E ratio was
1 but there were different amounts of bio-oil and emulsifier in the various samples in the emulsions.
In all instances, there was an increase in the amount of bio-oil in the sediment over time. The solubility
in all cases at 1 h duration was as high as at 70 wt %, except for the case of 1 g of bio-oil with 55.70 wt %.
The separation of bio-oil after 1 day was significant in the case of 1 g bio-oil and only 16.61 wt % of
it remained dissolved into the emulsion. With an increase in storage time, the sediment of the cases
involving 1, 2, and 3 g of bio-oil became pronounced. However, the solubility of the cases involving 4,
5, and 6 g of bio-oil with B/E = 1 remained above 55 wt % on the 7th day, implying that the solubility
of the emulsion improved greatly when the amounts of surfactant and bio-oil were increased. On the
7th day, the maximum solubility in the case of 6 g bio-oil with B/E = 1 was 66.48 wt %.
The FTIR spectra at three regions (i.e., top, middle, and bottom liquid levels in the tube) in
the emulsion under 2 g bio-oil with B/E = 1 for 1 h to 7 days are displayed in Figure 4. The FTIR
spectra appear to almost overlap and coincide with each other, confirming homogeneity and successful
emulsification [12].
403
Energies 2018, 11, 3031
D K G GD\V

6ROXELOLW\
6ROXELOLW\

$PRXQW RI ELR RLO J $PRXQW RI ELR RLO J
E GD\ H GD\V

6ROXELOLW\
6ROXELOLW\

$PRXQW RI ELR RLO J $PRXQW RI ELR RLO J
F GD\V

6ROXELOLW\

$PRXQW RI ELR RLO J

Figure 3. The solubility of the emulsions with different bio-oil amounts (B/E = 1) (a) 1 h, (b) 1 day,
(c) 3 days, (d) 5 days, and (e) 7 days.
404
Energies 2018, 11, 3031
D K G GD\V

,QWHQVLW\ 7
,QWHQVLW\ 7

%RWWRP %RWWRP
0LGGOH 0LGGOH
7RS 7RS

:DYHQXPEHU FP :DYHQXPEHU FP
E GD\ H GD\V

,QWHQVLW\ 7
,QWHQVLW\ 7

%RWWRP %RWWRP
0LGGOH 0LGGOH
7RS 7RS

:DYHQXPEHU FP :DYHQXPEHU FP
F GD\V

,QWHQVLW\ 7

%RWWRP
0LGGOH
7RS

:DYHQXPEHU FP
Figure 4. Fourier transform infrared spectroscopy (FTIR) spectra of emulsion (2 g bio-oil, B/E = 1) at
three regions for (a) 1 h, (b) 1 day, (c) 3 days, (d) 5 days, and (e) 7 days.
405
Energies 2018, 11, 3031
3.3. Bio-Oil Solubility at B/E Ratios between 0.556 and 1

The separation of the emulsion in the cases of a lower amount of bio-oil with B/E = 1 was more
pronounced compared to those with a higher amount of bio-oil. To study the solubility of the bio-oil,
the solubility at the B/E ratios of 0.556–1 was investigated. The surfactant’s concentration was one
of the most important factors influencing the solubility of emulsions. The bio-oil and diesel contents
were fixed at 2 g and 20 g, respectively. An increase in the amount of the emulsifier or a decrease in the
B/E ratio caused more bio-oil to dissolve into the emulsion (Figure 5). The solubility at B/E = 1 on
the 7th day was only 12.02 wt %. When the B/E ratio decreased to 0.556, the solubility increased to
57.81 wt %, implying that a higher emulsifier concentration rendered a higher solubility of the bio-oil
into the emulsion.
The FTIR spectra at three regions (i.e., top, middle, and bottom liquid levels in the tube) in the
emulsion with 2 g bio-oil and B/E = 1 for 7 days are displayed in Figure 4e, showing the stability of
the emulsion. The FTIR spectra almost overlap and coincide with each other, confirming homogeneity
and higher stability of the emulsion.

J ELR RLO J HPXOVLILHU
6ROXELOLW\

%( UDWLR
Figure 5. The solubility of the emulsions with different B/E ratios (2 g bio-oil) after 7 days.
3.4. Temperature Effect

The fluidity of the bio-oil was low at room temperature. To increase the fluidity of the emulsion
during storage, the emulsion was heated at 100 ◦ C in the oven for 7 days. In general, both the viscosity
and interfacial tension decreased with increasing temperature, making the emulsification easier [15].
This resulted in an increased degree of solubility at 100 ◦ C compared to that at room temperature.
Compared to the storage of the emulsion at room temperature, the solubility in the case of 2 g bio-oil
with B/E = 1 at 100 ◦ C promoted the solubility to 59.86 wt % from 12.02 wt % at room temperature
(Figure 6). This clearly indicates that the effect of the bio-oil temperature at B/E = 1 played an
important role in improving its solubility. However, the impact of the storage temperature in the case
of B/E = 0.556 was not of the same magnitude; the solubility being merely intensified from 57.81 wt %
at room temperature to 63.45 wt % at 100 ◦ C, accounting for an increase of 6.64 wt %. Overall, the
temperature effect on the solubility of bio-oil increased when the B/E ratio increased.
406
Energies 2018, 11, 3031

6ROXELOLW\

%( UDWLR
Figure 6. The solubility of bio-oil at different B/E ratios (2 g of bio-oil, fixed amount) at 100 ◦ C for
7 days.
The FTIR spectra at three regions (i.e., top, middle, and bottom liquid levels in the tube) in the
emulsion at the storage temperature of 100 ◦ C with 2 g bio-oil and B/E = 1 for 7 days are displayed in
Figure 7. The FTIR spectra also almost overlap and coincide with each other, showing the homogeneity
and high stability of the emulsion again.

,QWHQVLW\ 7

%RWWRP
0LGGOH
7RS

:DYHQXPEHU FP
Figure 7. FTIR spectra of the emulsion with 2 g bio-oil and B/E = 1 under 100 ◦ C at 7 days.
3.5. Co-Surfactant (Methanol) Effect

In general, the amount of the emulsifier must be increased to improve the bio-oil usage in the
emulsion. The co-surfactants are surface-active compounds that can lower the surface tension of bio-oil
or the dispersed phase, making them suitable for emulsification [25]. In this study, methanol was used
as the co-surfactant, resulting from its influence on increasing the emulsion‘s stability [13]. In each
experimental run, 1 g of methanol was loaded into the emulsions. The bio-oil solubility in the case
407
Energies 2018, 11, 3031
of 2 g bio-oil with B/E = 1 at a duration of 7 days increased to 46.66 wt % (Figure 8), as opposed to
12.02 wt % without the addition of methanol. This evidence suggests that methanol can significantly
facilitate the emulsification performance. The impact of the co-surfactant in the case of B/E = 0.556
was not pronounced. The solubility with the co-surfactant (70.96 wt %) was higher than that without
methanol (57.81 wt %) by 13.15 wt %. The co-surfactants effect on the solubility of the bio-oil increased
when the B/E ratio increased.
The FTIR spectra at the three regions (i.e., top, middle, and bottom liquid levels in the tube) in the
emulsion after adding 1 g methanol with 2 g bio-oil and B/E = 1 for 7 days are displayed in Figure 9.
The FTIR spectra also show the high homogeneity and stability of the emulsion. By comparing the
influences of the two factors (storage temperature and co-surfactant) on the solubility of the bio-oil
from B/E = 0.556 to 1 (Figures 6 and 8), when the B/E ratio was between 0.714 and 1, the impact of the
storage temperature on the solubility was beyond the co-surfactant. On the other hand, when the B/E
ratio ranged from 0.556 to 0.625, a converse behavior was observed.

6ROXELOLW\

%( UDWLR
Figure 8. The solubility of bio-oil at different B/E ratios with 1 g methanol.

,QWHQVLW\ 7

%RWWRP
0LGGOH
7RS

:DYHQXPEHU FP
Figure 9. FTIR spectra of the emulsion with 2 g bio-oil, B/E = 1, and 1 g methanol for 7 days.
408
Energies 2018, 11, 3031
4. Conclusions
The simultaneous emulsification-extraction process of liquefaction bio-oil and diesel with the aid
of a commercial emulsifier (Atlox 4914) and co-surfactant (methanol) have been explored where 20 g
of diesel was used. When the bio-oil amount increased from 1 g to 6 g with the fixed B/E ratio of 1, its
solubility increased from 55.70 wt % to 71.35 wt %. However, the solubility with 1 g bio-oil (B/E = 1)
decreased from 55.70 wt % to 4.41 wt % after the 7 days storage and decreased from 71.35 wt % to
66.48 wt % with 6 g bio-oil (B/E = 1). This suggests that performing emulsification at a larger scale may
facilitate the solubility. When a higher amount of emulsifier was used with the B/E ratio decreasing
from 1 to 0.556, the bio-oil solubility also increased from 12.02 wt % to 57.81 wt % after a 7-day duration.
To reduce the viscosity of the emulsions, the emulsions with B/E ratios ranging from 0.556 to 1 were
stored in an oven at 100 ◦ C for 7 days. The storage temperature of 100 ◦ C had more impact on the
solubility at higher B/E ratios, especially for the case of B/E = 1, which enhanced the solubility by a
factor of 47.84% as compared to room temperature (25 ◦ C). Moreover, the addition of a co-surfactant
(methanol) at B/E = 1 at room temperature intensified the bio-oil solubility from 12.02 wt % (without
addition) to 46.66 wt %. When the liquefaction bio-oil was extracted and emulsified in the emulsions,
their higher stability after 7-day storage was achieved from the FTIR analysis. The results of the study
indicate that hydrothermal liquefaction can be applied to waste-to-fuel technology and the liquefaction
bio-oil from food waste has a high potential to replace the diesel fuel used in industrial furnaces
and boilers.
Author Contributions: D.L, M.Š. and V.Č. did experiments, analyzed data and wrote the paper. Y.-Y.L. and
C.-W.W. helped with the experiments and writing the paper. W.-H.C. provided this research idea and organize
this study, developed the experiment, analyzed data, and provided facilities and instruments for research.
Funding: The authors gratefully acknowledge the financial support (MOST 106-2923-E-006-002-MY3) of the
Ministry of Science and Technology, Taiwan, R.O.C., and the financial support (TUCENOT LO1404) of the research
center for low-carbon energy technologies.
Conflicts of Interest: The authors declare no conflicts of interest
References
1. Kavitha, S.; Banu, J.R.; Priya, A.A.; Uan, D.K.; Yeom, I.T. Liquefaction of food waste and its impacts on
anaerobic biodegradability, energy ratio and economic feasibility. Appl. Energy 2017, 208, 228–238. [CrossRef]
2. Bach, Q.V.; Chen, W. A comprehensive study on pyrolysis kinetics of microalgal biomass. Energy Convers. Manag.
2017, 131, 109–116. [CrossRef]
3. Chen, W.; Chen, C.J.; Hung, C.I.; Shen, C.H.; Hsu, H.W. A comparison of gasification phenomena among raw
biomass, torrefied biomass and coal in an entrained-flow reactor. Appl. Energy 2013, 112, 421–430. [CrossRef]
4. Karthikeyan, O.P.; Wong, S.M.J.W.C. Bio-refining of food waste for fuel and value products. In Proceedings
of the 4th International conference on Energy and Environment Research (ICEER 2017), Porto, Portugal,
17–20 July 2017; pp. 14–21.
5. Lin, C.K.; Pfaltzgraff, L.A.; Mubofu, E.B.; Abderrahim, S. Food waste as a valuable resource for the production
of chemicals, materials and fuels. Current situation and global perspective. Energy Environ. Sci. 2013,
6, 426–464. [CrossRef]
6. Chen, W.H.; Lin, B.J.; Huang, M.Y.; Chang, J.S. Thermochemical conversion of microalgal biomass into
biofuels: A review. Bioresour. Technol. 2015, 184, 314–327. [CrossRef] [PubMed]
7. Dimitriadis, A.; Bezergianni, S. Hydrothermal liquefaction of various biomass and waste feedstocks for
biocrude production: A state of the art review. Renew. Sustain. Energy Rev. 2017, 68, 113–125. [CrossRef]
8. Chen, Z.; Long, J. Organosolv liquefaction of sugarcane bagasse catalyzed by acidic ionic liquids.
9. Saber, M.; Golzary, A.; Hosseinpour, M.; Takahashi, F.; Yoshikawa, K. Catalytic hydrothermal liquefaction of
microalgae using nanocatalyst. Appl. Energy 2016, 183, 566–576. [CrossRef]
10. Gollakota, A.R.K.; Kishore, N.; Gu, S. A review on hydrothermal liquefaction of biomass. Renew. Sustain.
409
Energies 2018, 11, 3031
11. Jena, U.; Das, K.C.; Kastner, J.R. Effect of operating conditions of thermochemical liquefaction on biocrude
production from Spirulina platensis. Bioresour. Technol. 2011, 102, 6221–6229. [CrossRef] [PubMed]
12. Lin, B.-J.; Chen, W.-H.; Budzianowski, W.M.; Hsieh, C.-T.; Lin, P.-H. Emulsification analysis of bio-oil and
diesel under various combinations of emulsifiers. Appl. Energy 2016, 178, 746–757. [CrossRef]
13. De Luna, M.D.G.; Cruz, L.A.D.; Chen, W.-H.; Lin, B.-J.; Hsieh, T.-H. Improving the stability of diesel
emulsions with high pyrolysis bio-oil content by alcohol co-surfactants and high shear mixing strategies.
Energy 2017, 141, 1416–1428. [CrossRef]
14. Ikura, M.; Stanciulescu, M.; Hogan, E. Emulsification of pyrolysis derived bio-oil in diesel fuel.
Biomass Bioenergy 2003, 24, 221–232. [CrossRef]
15. Chiaramonti, D.; Bonini, M.; Fratini, E.; Tondi, G.; Gartner, K.; Bridgwater, A.V.; Grimm, H.P.; Soldaini, I.;
Webster, A.; Baglioni, P. Development of emulsions from biomass pyrolysis liquid and diesel and their use in
engines—Part 1: Emulsion production. Biomass Bioenergy 2003, 25, 85–99. [CrossRef]
16. Guo, Z.; Wang, S.; Wang, X. Stability mechanism investigation of emulsion fuels from biomass pyrolysis oil
and diesel. Energy 2014, 66, 250–255. [CrossRef]
17. Leng, L.; Yuan, X.; Chen, X.; Huang, H.; Wang, H.; Li, H.; Zhu, R.; Li, S.; Zeng, G. Characterization
of liquefaction bio-oil from sewage sludge and its solubilization in diesel microemulsion. Energy 2015,
82, 218–228. [CrossRef]
18. Houghton, R.A.; Hall, F.; Goetz, S.J. Importance of biomass in the global carbon cycle. J. Geophys. Res. Biogeosci.
2009, 114, G00E03. [CrossRef]
19. Jiang, X.; Ellis, N. Upgrading Bio-oil through Emulsification with Biodiesel: Mixture Production. Energy Fuels
2010, 24, 1358–1364. [CrossRef]
20. Qian, Y.; Zuo, C.; Tan, J.; He, J. Structural analysis of bio-oils from sub-and supercritical water liquefaction of
woody biomass. Energy 2007, 32, 196–202. [CrossRef]
21. Mistry, B.D. A Handbook of Spectroscopic Data Chemistry; Oxford Book Company: Oxford, UK, 2009.
22. Assanvo, E.F.; Gogoi, P.; Dolui, S.K.; Baruah, S.D. Synthesis, characterization, and performance characteristics
of alkyd resins based on Ricinodendron heudelotii oil and their blending with epoxy resins. Ind. Crops Prod.
2015, 65 (Suppl. C), 293–302. [CrossRef]
23. Bora, M.M.; Gogoi, P.; Deka, D.C.; Kakati, D.K. Synthesis and characterization of yellow oleander (Thevetia
peruviana) seed oil-based alkyd resin. Ind. Crops Prod. 2014, 52 (Suppl. C), 721–728. [CrossRef]
24. Benavente, V.; Fullana, A. Torrefaction of olive mill waste. Biomass Bioenergy 2015, 73, 186–194. [CrossRef]
25. 1Sun, M.; Wang, Y.; Firoozabadi, A. Effectiveness of Alcohol Cosurfactants in Hydrate Antiagglomeration.
Energy Fuels 2012, 26, 5626–5632.
410
energies
Article
Variation in the Distribution of Hydrogen Producers
from the Clostridiales Order in Biogas Reactors
Depending on Different Input Substrates
Martin Černý 1 , Monika Vítězová 1 , Tomáš Vítěz 2 , Milan Bartoš 1 and Ivan Kushkevych 1, *
1 Department of Experimental Biology, Faculty of Science, Masaryk University, Kamenice 753/5,
62500 Brno, Czech Republic; [email protected] (M.Č.); [email protected] (M.V.);
[email protected] (M.B.)
2 Department of Agricultural, Food, and Environmental Engineering, Faculty of AgriSciences,
Mendel University, Brno, Zemědelska 1, 61300 Brno, Czech Republic; [email protected]
* Correspondence: [email protected] or [email protected]; Tel.: +420-549-49-5315
Received: 17 September 2018; Accepted: 10 October 2018; Published: 23 November 2018
Abstract: With growing demand for clean and cheap energy resources, biogas production is emerging
as an ideal solution, as it provides relatively cheap and clean energy, while also tackling the
problematic production of excessive organic waste from crops and animal agriculture. Behind this
process stands a variety of anaerobic microorganisms, which turn organic substrates into valuable
biogas. The biogas itself is a mixture of gases, produced mostly as metabolic byproducts of the
microorganisms, such as methane, hydrogen, or carbon dioxide. Hydrogen itself figures as a potent
bio-fuel, however in many bioreactors it serves as the main substrate of methanogenesis, thus
potentially limiting biogas yield. With help of modern sequencing techniques, we tried to evaluate
the composition in eight bioreactors using different input materials, showing shifts in the microbial
consortia depending on the substrate itself. In this paper, we provide insight on the occurrence of
potentially harmful microorganisms such as Clostridium novyi and Clostridium septicum, as well as
key genera in hydrogen production, such as Clostridium stercorarium, Mobilitalea sp., Herbinix sp.,
Herbivorax sp., and Acetivibrio sp.
Keywords: biogas; Clostridiales; hydrogen-producing bacteria; bioreactors; anaerobic fermentation;

anaerobic digestion; microbial community composition
1. Introduction
Biogas production appears to be a brilliant solution to tackle common energetics, waste
management, and environmental pollution problems. Production of biogas provides us with clean
energy in a form of gas mixture, utilizing common organic wastes as an input substrate and thus
helping to maintain the biomass cycle without the formation of unnecessary byproducts [1,2].
The elegance of biogas production lies in the complex microbial consortia capable of anaerobic
fermentation, which ultimately leads to the final step, methanogenesis. Methanogenesis itself is
a dominant metabolic pathway typical for the group of microorganisms referred to as methanogenic
archaea. These microorganisms rely on the end-products of bacterial fermentation, such as short-chain
fatty acids (SCFA), carbon dioxide, hydrogen, and even methyl–amines and methanol [3–6]. However,
relatively little is known about the microbial consortia responsible for the production of methanogenesis
precursors. The composition of the bioreactor ecosystem has a direct impact on the overall fitness of
biogas production [7,8]. Unbalance in the microbial consortia may lead to the collapse of fermentative
processes or contamination of biogas by undesired products, such as H2 S [9].
In order to keep biogas production sustainable for a long period of time, with high yields of gas,
one must be able to identify key groups of microorganisms that help to maintain steady conditions

Energies 2018, 11, 3270
and precursors for biogas production [3,10]. Many anaerobic microorganisms were identified as
potential hydrogen producers, figuring in many natural and anthropogenic processes. This knowledge
then helps not only to ensure methanogenesis occurs in bioreactors but is also exploited in terms
of biohydrogen production [11–13]. Hydrogen itself is one of the many final products of bacterial
fermentation and also serves as a proton that is used in various microbial enzymatic pathways.
In connection with SCFAs, it is also associated with interspecies hydrogen-electron transfer, which
further serves as a connection among fermenting bacterial taxa [14–16]. In anaerobic digesters and
bioreactors, the microbial community is greatly dependent on the inoculum and the origin of used
substrate. These factors have a substantial effect on the behaviour of the microbial community and
yields of the biogas plant [7,8]. With crop waste and animal manure being used as an input substrate,
a great portion of the microbial community is connected to soil and gastrointestinal tract consortia.
Understanding the diversity of primary hydrogen producers can foretell if a correlation exists between
production, environment, and the microbial community, and if this has some effect on biogas yield.
Bearing this knowledge in mind, the Clostridiales order was chosen as the desired microbial group.
Many species belonging to this clade are considered potent hydrogen producers. These microorganisms
create a substantial portion of the Gram-positive microbial community in mesophilic bioreactors [17,18].
Results of this study illustrate the variability in the composition of the Clostridiales order depending on
the different input substrates.

Eight biogas plants throughout the Czech Republic were chosen based on known input material
with information about the substrate provided as the ratio of substrate components (w/w %). Names of
the plants, as well as information about substrate components, are summarized in Table 1. Fermenters
used in these biogas plants had the operational volume of 2500–3500 m3 . Sampling was carried out
anoxically and material was directly transferred from the fermenter into sterile sampling vessels.
Afterwards, the samples were stored in a thermo-isolating box and were transported for immediate
analysis to our laboratory.
Table 1. Characteristics of evaluated anaerobic digesters (Reproduced with permission of Kushkevych

[6]).
Geographic
Main Substrate Input Ratio Temperature (◦ C) pH ORP (mV)
Location
Modřice primary sludge, biological sludge 50:50 34 7 −3.1
maize silage, whole crop silage, poultry
Bratčice 63:31:6 43 8.3 −75
litter
Pánov maize silage, poultry litter 92:8 49 8 −58
maize silage, sugar beet pulp, whole
Úvalno 44:44:6:6 48 7.69 −38.5
crop silage, cattle manure
maize silage, sugar beet pulp, whole
Horní Benešov 29:39:12:15:5 49 7.85 −47.4
crop silage, cattle manure, grass silage
Rusín maize silage, sugar beet pulp 70:30 48 7.63 −34.7
Loděnice maize silage, sugar beet pulp 75:25 44 7.65 −36
Čejč pig manure, maize silage 75:25 – – –
Prior to DNA analysis, chemical and physical parameters were determined. The pH, redox
potential, and temperature were measured, as well as total solids content, volatile solids content,
and biogas composition. Measurements were conducted for each anaerobic digester of a biogas plant
with data summarized in Table 1. For pH and redox, potential measurement was used pH/Cond meter
3320 (WTW GmbH, Dinslaken, Germany) in accordance with standard procedures [19]. The sample
temperature was assessed by high accuracy PT100 RTD thermometer HH804U (OMEGA Engineering,
412
Energies 2018, 11, 3270
Stamford, CT 06907-0047, USA). Total solids (TS) content was determined by drying at 105 ± 5 ◦ C,
using EcoCELL 111 (BMT Medical Technology Ltd., Brno, Czech Republic) according to Czech Standard
Method [20]. After drying, samples were cooled in a desiccator, until they reached constant weight
and the value could be determined. Volatile solids content (VS) was assessed by combustion of the
samples in a muffle furnace LMH 11/12 (LAC, Ltd., Rajhrad, Czech Republic) at 550 ± 5 ◦ C according
to Czech Standard Method [21].
According to our previous studies, we used the QIAamp Fast DNA Stool Mini Kit (QIAGEN
GmbH, Hilden, Germany). This was appropriate for a fast and easy purification process of total DNA
from fresh or frozen samples and was used for DNA extraction from samples of anaerobic fermenters in
previous studies [9]. DNA extractions were conducted according to manufacturer protocol, only with
minor adjustments, which we describe below. We extracted 100 mg of each sample, which we mixed
with 1.4 mL of ASL buffer (QIAGEN GmbH, Hilden, Germany) and incubated at 95 ◦ C for 10 min.
After centrifugation, we added InhibitEX tablets to the supernatant, according to protocol, to remove
impurities and possible PCR inhibitors. After another centrifugation, 200 μL of supernatant was mixed
with 15 μL of proteinase K solution and 200 μL of buffer AL (QIAGEN GmbH, Hilden, Germany).
Then, the solution was incubated at 70 ◦ C for 10 min, cooled down and 200 μL of ethanol (96–100%)
was added to the mixture. Following the procedure, the supernatant was centrifuged through the
QIAamp kit column and was washed twice by buffers AW1 and AW2 (QIAGEN GmbH, Hilden,
Germany). Finally, the DNA elution was conducted by 200 μL of elution buffer.
For amplification of the desired product, V3 and V4 variable regions of the 16S rRNA gene were
targeted by universal primers [22]. Primers were marked for sample identification by molecular
barcoding. According to our previous research, the Maxima™ Probe qPCR Master Mix (Thermo Fisher
Scientific, Waltham, MA, USA) was used for PCR reaction. Initial denaturation was conducted at
95 ◦ C for 10 min, followed by 30 cycles of incubation at 94 ◦ C for 30 s, 60 ◦ C for 30 s, and 72 ◦ C
for 120 s, with a final extension step at 72 ◦ C with a 2 min duration. Visualization of PCR products
was performed using 1.5% agarose gel. After the reaction, DNA was extracted from the gel using
the QIAquick Gel Extraction Kit (Qiagen GmbH, Hilden, Germany). DNA was quantified using the
Quant-iTPicoGreen dsDNA Assay (Thermo Fisher Scientific, Waltham, MA, USA) and equimolar
amounts of PCR products were pooled together. Paired-end amplicons were sequenced via Illumina
Mi-Seq platform. Data analysis of 16S rRNA sequences was carried out using QIIME data analysis
package [23].
According to base quality score distributions, average base content per reading and the
guanosine-cytosine pairs (GC) distribution in the reads, quality filtering on raw sequences was
conducted. Chimeras and reads that did not cluster with other sequences were removed. The obtained
sequences with quality scores higher than 20 were shortened to the same length of 350 bp and
classified with RDP Seqmatch with an operational taxonomic unit (OTU) discrimination level set to
97%. The relative abundance of taxonomic groups was calculated for microorganisms detected in this
study. Sequences were compared using the BLAST feature of the National Center for Biotechnology
Information (NCBI) [24]. The sequences were uploaded to Geneious (Geneious 7.1.9) for comparative
genomic analyses [25]. Alignments of sequences were performed in Geneious 7.1.9 using Muscle
(Clustal W) with the BLOSUM cost matrix and clustering was performed by the neighbor-joining
method [26].
Sequences of selected microorganisms were then deposited in GenBank under accession numbers:
MH045949, MH045950, MH045951, MH045952, MH045953, MH045954, MH045955, MH045956,
MH045973, MH045974, MH045975, MH045976, MH045977. Origin7.0 software (www.origin-lab.com)
was used for further processing data and analysis of the obtained results.
413
Energies 2018, 11, 3270
3. Results
3.1. Commenting on the Control

After the analysis of raw data, the most abundant genera were investigated thoroughly. First,
the overall abundance of the Clostridiales order was compared against the total microbial background
(Figure 1A). The mean ratio was 5.84% of the whole microbial background, with a minimum of
0.64% in the case of the Modřice bioreactor, which serves as a control, and a maximum of 9.51%
in the case of the Čejč bioreactor. This discrepancy is even more visible on the graph visualizing
the ratios of detected OTUs in the manner of the whole set of raw data (Figure 1B). Variation in
the structure of the microbial community of the Modřice digester is probably a consequence of a
significant difference in input substrate. The input material, in this case, consists of biological and
waste sludge, which has a slightly less alkaline pH and a higher redox potential compared to other
reactors, with crop silage or manure being used as the main input substrate. Thus, the explanation
probably lies in the different substrate, its properties, and microbial load [27,28]. Biological sludge is
a complex microbial ecosystem, comprised of many bacterial, archaeal, and eukaryotic taxa such as
bacteria Acinetobacter sp., Flavobacterium sp., Cloacibacterium sp., Desulfovibrio sp., Desulfomicrobium sp.
or archaea such as Methanobrevibacter sp. and Methanosarcina sp. [8,29–32]. Microorganisms coming
from this environment may favour these conditions, leading to a significant shift in the Clostridiales or
total bacterial ratio due to different conditions (Figure 1A).
Figure 1. Ratio of Clostridiales representative composition: ratio between Clostridiales order and total
microbial amount (A); pie-graph indicating portions of each biogas plant to Clostridiales order ratio (B).
414
Energies 2018, 11, 3270
3.2. Comparison of the Clostridiales Order Communities

Comparing the rest of the bioreactors, it seems for the Figure 1A that there is a shift from the
Clostridiales order microorganisms towards other bacterial species. The lower ratios detected in Pánov,
Horní Benešov, Úvalno, and Rusín (Figure 1B) may be connected to the optimal growth temperature
and temperature threshold for many mesophilic microorganisms. Many Clostridiales order organisms
manifest poor growth above 44 ◦ C [33]. Judging by this, it is probable that higher temperatures may
lead to changes in the taxa, moving from Clostridiales candidates to more thermophilic organisms
from different taxonomic groups. Analysis of the most abundant genera revealed domination of the
Clostridiales group by the genus Clostridium. The mean value for the Clostridium genus is 58.98%.
In addition to Figure 1, Figure 2 shows that in the case of extreme results, the microbial consortium
in the Čejč bioreactor was dominated by this genus with the ratio of 89.44%. In the case of the Čejč
bioreactor, the Clostridium genus was dominated by the Clostridium novyi species, belonging to cluster
I, which comprised 65.63% of all OTUs retrieved from this bioreactor sample, deposited in GenBank
under accession number MH045969.
Figure 2. Visualization of the genera composition in each biogas plant and mean of its abundance to
total microbial count.
Analysis by BLAST gives 99% similarity to the sequence with GenBank accession number
LC193834.1 (Figure 3). C. novyi is characterized as 0.6–1.4 × 1.6–17 μm long gram-positive rods, which
usually occur singly or in pairs with good gas production. Among the common substrates belong
carbohydrates such as raffinose, glucose, melibiose, and cellobiose, which in turn produce products
such as acetate, propionate, isobutyrate, butyrate, isovalerate, and more importantly hydrogen,
and CO2 . C. novyi is a sporulating rod with sub-terminally to terminally localized spores, which
may cause swelling of the cell. Optimal growth temperatures range from 30–45 ◦ C, with an alkaline pH
threshold of 8.5. However, C. novyi is regarded as a soil-dwelling microorganism and a known pathogen
of animals and even humans. It is the first time that C. novyi was detected in such a high ratio compared
to other data. C. novyi was formerly detected in bioreactors only in low abundances as described by
Fröschle et al., 2015. Its potential to pose a threat to biogas plant operators remains unclear [33–36].
Another notable species is Clostridium septicum from cluster I and Clostridium stercorarium belonging to
cluster II. As in the case of C. novyi, C. septicum may also be considered a human pathogen. The optimal
temperature is 44 ◦ C. However, its growth is inhibited by temperatures above 46 ◦ C. It forms straight
415
Energies 2018, 11, 3270
to curved cells that are motile. This species creates oval spores, located sub-terminally within the cell,
which may swell if present. It is considered a strong gas producer [33].
Figure 3. Phylogenetic relationship of detected species in terms of their source.
416
Energies 2018, 11, 3270
C. stercorarium is a representative of thermophilic clostridia. Its temperature optimum lies around

65 ◦ C and under given conditions forms relatively small rods 0.3–0.4 × 2–4 μm and produces oval
terminal spores. This species is able to ferment cellulose, cellobiose, melibiose, raffinose, fructose,
and sucrose in turn for the production of acetate, lactate, ethanol, hydrogen, and CO2 . A possible
explanation for the occurrence of pathogenic clostridia may be their connection with the gastrointestinal
tract of farm animals and soil. In the case of C. stercorarium, it naturally occurs in compost, thus
probably promoting anaerobic digestion of silage in the bioreactor [18]. Other notable species detected
in this study are C. clariflavum, C. propionicum, C. neopropionicum, and C. sulfidigenes (Figure 3).
The importance of these species lies mainly in the formation of hydrogen and SCFAs, which are
connected with methanogenesis and linked to interspecies hydrogen transfer. Their metabolism
ensures the thermodynamic stability of anaerobic fermentation, providing a favorable environment for
deeper and fine digestion of the given substrate [33].
Apart from the genus Clostridium, minor clades such as Acetivibrio sp. (10.64%), Herbinix sp.
(8.26%), Herbivorax sp. (8.38%), and Mobilitalea sp. (8.40%) were detected in the bioreactors, comprising
10–20% of all retrieved OTUs (Figures 2 and 3). Acetivibrio cellulolyticus was the sole member of this
genus detected in this study, which was similar from 98% to GenBank sequence NR_025917.1 and was
deposited under accession number MH045953 (Figure 3). Its morphology consists of slightly curved
gram-negative rods with a single flagellum having dimensions of 0.5–0.8 × 4–10 μm. Major substrates
consisted of cellulose, cellobiose, and salicin, producing acetate, hydrogen, and CO2 in return [36].
Similar to Acetivibrio, the genus Herbinix is represented by the sole species H. luporum (NR_152095.1),
with a sequence similarity of 99% (MH045949). This thermophilic cellulose-degrading bacterium can
be found in bioreactors, where it is connected with plant biomass degradation. H. luporum represents a
novel genus in the family Lachnospiraceae. It is a non-motile rod, growing to 2.0–6.0 × 0.5 μm, with an
optimal growth temperature of 55 ◦ C and main fermentation products such as ethanol, acetate, butyrate,
and hydrogen [37]. Herbivorax saccinicola is a key player in plant biomass remineralization during
anaerobic digestion processes and a member of the family Ruminococcaceae. These non-motile long
and thin rods with dimensions of 5–10 × 0.2 μm grow between 45–65 ◦ C. However, the growth of this
microorganism was observed even in lower temperatures (40–44 ◦ C), one may deduce that temperature
specificity is not so strict. The main fermentation products are acetate, ethanol, and hydrogen [38].
This organism (NR_152684.1) was similar to the sequence deposited in GenBank under the accession
number MH045957 (Figure 3). Mobilitalea sibirica is the last of the moderately abundant members of
curved rod-shaped anaerobic bacteria with the main products of fermentation being ethanol, acetate,
hydrogen, and CO2 [39]. Our sequence (MH045951) possessed 99% similarity with type strain P3M-3
sequence under accession number NR_134091.1 (Figure 3). Although many genera were detected in
this study, the remaining clades created only a negligible minority which abundance ranged mostly
in order of tenths of percentages compared to other taxa (Figure 2). Thus, we didn´t evaluated their
occurrence in a deeper context.
4. Discussion
As for methanogenesis and methanogenic archaea, understanding fine arrangements in microbial
community relationships are key to maintaining steady and profitable biogas production. As it may
seem from some studies, the amount of hydrogen may be misjudged as a smaller portion of gas
production. However, it is essential for CO2 to be reduced into methane by methanogenic archaea and
naturally turned into biogas. The smaller portion of produced hydrogen also serves as a mediator
in interspecies hydrogen transfer, resulting in a high level of well-being of the anaerobic microbial
community [3,4,6]. Our data suggests that the composition of anaerobic fermenters depends on the
input substrate and may vary greatly because of the given condition. The impact of these findings
remains to be evaluated, also whether this knowledge can be exploited to the artificially potent
microenvironment, as well as to design inoculum, and knowing possible impacts of using different
input substrates. The stable and diverse microbiota of the bioreactor may prevent colonization
417
Energies 2018, 11, 3270
and succession of undesired microorganisms such as sulfate reducing bacteria. Their metabolic
products may damage not only the microbial consortia by strong competition for hydrogen, originally
utilized by methanogens, but may also lead to corrosion of the bioreactor itself [40–42]. Additionally,
this study confirmed the presence of potentially harmful microorganism such as C. septicum and
C. novyi, which are connected to myonecrosis and other dangerous illnesses, this raises questions
about aseptic and sanitation protocols when operating bioreactor processes. Their presence agrees
with previous findings, however, we detected C. novyi in a much higher ratio, comprising 65.63% of
the whole Clostridiales microbiota in the bioreactor for the first time [35]. Apart from the genus
Clostridium, families Lachnospiraceae and Ruminococcaceae, also have their place in the anaerobic
consortia. These microorganisms have similar physiological and biochemical properties as the
Clostridium genus, thus also significantly contributing to the methane production process. On the
contrary, the input substrate has a direct impact on the microbial community, as well as the well-being
of the reactor. It is not only modulating accessibility to essential nutrients, but also bears an additional
microbial load with which native microbes must compete for resources.
Author Contributions: M.Č. and I.K. conceived and contributed to the writing of the paper. Together with, M.V.,
T.V., and M.B. who further designed and performed the experiments and measurements described in this paper.
M.Č. together with T.V. analyzed the data and optimized process parameters. T.V. performed analysis of the data
and validation of the results.
References
1. Hobson, P.N. Biogas production from agricultural wastes. Experientia 1982, 38, 206–209. [CrossRef]
2. Papurello, D.; Silvestri, S.; Tomasi, L.; Belcari, I.; Biasioli, F.; Santarelli, M. Biowaste for SOFCs. Energy Procedia
2016, 101, 424–431. [CrossRef]
3. Oppermann, R.A.; Nelson, W.O.; Brown, R.E. In vivo studies of methanogenesis in the bovine rumen:
Dissimilation of acetate. J. Gen. Microbiol. 1961, 25, 103–111. [CrossRef] [PubMed]
4. Miller, T.L.; Wolin, M.J. Methanosphaera stadtmaniae gen. Nov., sp. nov.: A species that forms methane by
reducing methanol with hydrogen. Arch. Microbiol. 1985, 141, 116–122. [CrossRef] [PubMed]
5. Papurello, D.; Soukoulis, C.; Schuhfried, E.; Cappellin, L.; Gasperi, F.; Silvestri, S.; Santarelli, M.; Biasioli, F.
Monitoring of volatile compound emissions during dry anaerobic digestion of the Organic Fraction of
Municipal Solid Waste by Proton Transfer Reaction Time-of-Flight Mass Spectrometry. Bioresour. Technol.
2012, 126, 254–265. [CrossRef] [PubMed]
6. Dridi, B.; Fardeau, M.-L.; Ollivier, B.; Raoult, D.; Drancourt, M. Methanomassiliicoccus luminyensis gen. nov., sp.
nov., a methanogenic archaeon isolated from human faeces. Int. J. Syst. Evol. Microbiol. 2012, 62, 1902–1907.
[CrossRef] [PubMed]
7. Kushkevych, I.; Vítězová, M.; Vítěz, T.; Kováč, J.; Kaucká, P.; Jesionek, W.; Bartoš, M.; Barton, L.
A new combination of substrates: Biogas production and diversity of the methanogenic microorganisms.
Open Life Sci. 2018, 13, 119–128. [CrossRef]
8. Kushkevych, I.; Vítězová, M.; Vítěz, T.; Bartoš, M. Production of biogas: Relationship between methanogenic
and sulfate-reducing microorganisms. Open Life Sci. 2017, 12, 82–91. [CrossRef]
9. Kushkevych, I.; Kováč, J.; Vítězová, M.; Vítěz, T.; Bartoš, M. The diversity of sulfate-reducing bacteria in the
seven bioreactors. Arch. Microbiol. 2018. [CrossRef] [PubMed]
10. Nath, K.; Das, D. Improvement of fermentative hydrogen production: Various approaches. Appl. Microbiol.
Biotechnol. 2004, 65, 520–529. [CrossRef] [PubMed]
11. Hawkes, F.R.; Dinsdale, R.; Hawkes, D.L.; Hussy, I. Sustainable fermentative hydrogen production:
Challenges for process optimisation. Int. J. Hydrogen Energy. 2002, 27, 1339–1347. [CrossRef]
12. Levin, D.B.; Pitt, L.; Love, M. Biohydrogen production: Prospects and limitations to practical application.
Int. J. Hydrogen Energy. 2004, 29, 173–185. [CrossRef]
418
Energies 2018, 11, 3270
13. Nanqi, R.; Wanqian, G.; Bingfeng, L.; Guangli, C.; Jie, D. Biological hydrogen production by dark
fermentation: Challenges and prospects towards scaled-up production. Curr. Opin. Biotechnol. 2011, 22,
14. Baek, G.; Kim, J.; Kim, J.; Lee, C. Role and potential of direct interspecies electron transfer in anaerobic
digestion. Energies 2018, 11, 107. [CrossRef]
15. Ozturk, S.S.; Palsson, B.O.; Thiele, J.H. Control of interspecies electron transfer flow during anaerobic
digestion: Dynamic diffusion reaction models for hydrogen gas transfer in microbial flocs. Biotechnol. Bioeng.
1989, 33, 745–757. [CrossRef] [PubMed]
16. Cord-Ruwisch, R.; Seitz, H.-J.; Conrad, R. The capacity of hydrogenotrophic anaerobic bacteria to compete
for traces of hydrogen depends on the redox potential of the terminal electron acceptor. Arch. Microbiol. 1988,
149, 350–357. [CrossRef]
17. Kong, X.; Yu, S.; Fang, W.; Liu, J.; Li, H. Enhancing syntrophic associations among Clostridium butyricum,
Syntrophomonas and two types of methanogen by zero valent iron in an anaerobic assay with high organic
loading. Bioresour. Technol. 2018, 257, 181–191. [CrossRef] [PubMed]
18. Fang, H.H.P.; Zhang, T.; Liu, H. Microbial diversity of a mesophilic hydrogen-producing sludge.
Appl. Microbiol. Biotechnol. 2002, 58, 112–118. [CrossRef] [PubMed]
19. CSN EN 12176, Characterization of Sludge—Determination of pH-Value; Czech Standards Institute: Prague,
Czech Republic, 1999.
20. CSN EN 14346, Characterization of Waste—Calculation of Dry Matter by Determination of Dry Residue or Water
Content; Czech Standards Institute: Prague, Czech Republic, 2007.
21. CSN EN 15169, Characterization of Waste—Determination of Loss on Ignition in Waste, Sludge and Sediments;
Czech Standards Institute: Prague, Czech Republic, 2007.
22. Nossa, C.W.; Oberdorf, W.E.; Yang, L.; Aas, J.A.; Paster, B.J.; DeSantis, T.Z.; Brodie, E.L.; Malamud, D.;
Poles, M.A.; Pei, Z. Design of 16S rRNA gene primers for 454 pyrosequencing of the human foregut
microbiome. World. J. Gastroenterol. 2010, 16, 4135–4144. [CrossRef] [PubMed]
23. Caporaso, J.G.; Kuczynski, J.; Stombaugh, J.; Bittinger, K.; Bushman, F.D.; Costello, E.K.; Fierer, N.; Peña, A.G.;
Goodrich, J.K.; Gordon, J.I.; et al. QIIME allows analysis of high-throughput community sequencing data.
Nat. Methods 2010, 7, 335–336. [CrossRef] [PubMed]
24. Altschul, S.F.; Gish, W.; Mille, W.; Myers, E.W.; Lipman, D.J. Basic local alignment search tool. J. Mol. Biol.
1990, 215, 403–410. [CrossRef]
25. Kearse, M.; Moir, R.; Wilson, A.; Stones-Havas, S.; Cheung, M.; Sturrock, S. Geneious Basic: An integrated
and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics
2012, 28, 1647–1649. [CrossRef] [PubMed]
26. Larkin, M.A.; Blackshields, G.; Brown, N.P.; Chenna, R.; McGettigan, P.A.; McWilliam, H. Clustal W and
Clustal X version 2.0. Bioinformatics 2007, 23, 2947–2948. [CrossRef] [PubMed]
27. Venkiteshwaran, K.; Bocher, B.; Maki, J.; Zitomer, D. Relating anaerobic digestion microbial community and
process function. Microbiol. Insights 2015, 8, 37–44. [CrossRef] [PubMed]
28. Moestedt, J.; Nilsson Påledal, S.; Schnürer, A. The effect of substrate and operational parameters on the
abundance of sulphate-reducing bacteria in industrial anaerobic biogas digesters. Bioresour. Technol. 2013,
29. Shchegolkova, N.M.; Krasnov, G.S.; Belova, A.A.; Dmitriev, A.A.; Kharitonov, S.L.; Klimina, K.M.;
Melnikova, N.V.; Kudryavtseva, A.V. Microbial community structure of activated sludge in treatment
plants with different wastewater compositions. Front. Microbiol. 2016, 7, 90. [CrossRef] [PubMed]
30. Kushkevych, I.V. Activity and kinetic properties of phosphotransacetylase from intestinal sulfate-reducing
bacteria. Acta Biochim. Polonica 2015, 62, 103–108. [CrossRef]
31. Kushkevych, I.; Kollar, P.; Suchy, P.; Parak, K.; Pauk, K.; Imramovsky, A. Activity of selected salicylamides
against intestinal sulfate-reducing bacteria. Neuroendocrinol. Lett. 2015, 36, 106–113. [PubMed]
32. Kushkevych, I.; Vítězová, M.; Kos, J.; Kollár, P.; Jampílek, J. Effect of selected 8-hydroxyquinoline-
2-carboxanilides on viability and sulfate metabolism of Desulfovibrio piger. J. Appl. Biomed. 2018, 16,
33. Rainey, F.A. Clostridiales. In Bergey’s Manual of Systematics of Archaea and Bacteria; John Wiley & Sons, Inc.:
Hoboken, NJ, USA, 2015; pp. 1–5. [CrossRef]
419
Energies 2018, 11, 3270
34. Kim, M.D.; Song, M.; Jo, M.; Shin, S.G.; Khim, J.H.; Hwang, S. Growth condition and bacterial community
for maximum hydrolysis of suspended organic materials in anaerobic digestion of food waste-recycling
wastewater. Appl. Microbiol. Biotechnol. 2010, 85, 1611–1618. [CrossRef] [PubMed]
35. Fröschle, B.; Messelhäusser, U.; Höller, C.; Lebuhn, M. Fate of Clostridium botulinum and incidence of
pathogenic clostridia in biogas processes. J. Appl. Microbiol. 2015, 119, 936–947. [CrossRef] [PubMed]
36. Patel, G.B.; Khan, A.W.; Agnew, B.J.; Colvin, J.R. Isolation and characterization of an anaerobic, cellulolytic
microorganism, Acetivibrio cellulolyticus gen. nov., sp. nov. Int. J. Syst. Bacteriol. 1980, 30, 179–185. [CrossRef]
37. Koeck, D.E.; Hahnke, S.; Zverlov, V.V. Herbinix luporum sp. nov., a thermophilic cellulose-degrading
bacterium isolated from thermophilic biogas reactor. Int. J. Syst. Evol. Microbiol. 2016, 66, 4132–4137.
[CrossRef] [PubMed]
38. Koeck, D.E.; Mechelke, M.; Zverlov, V.V.; Liebl, W.; Schwarz, W.H. Herbivorax saccincola gen. nov., sp. nov.,
a cellulolytic, anaerobic, thermophilic bacterium isolated via in sacco enrichments from a lab-scale biogas
reactor. Int. J. Syst. Evol. Microbiol. 2016, 66, 4458–4463. [CrossRef] [PubMed]
39. Podosokorskaya, O.A.; Bonch-Osmolovskaya, E.A.; Beskorovaynyy, A.V.; Toshchakov, S.V.; Kolganova, T.V.;
Kublanov, I.V. Mobilitalea sibirica gen. nov., sp. nov., a halotolerant polysaccharide-degrading bacterium.
Int. J. Syst. Evol. Microbiol. 2014, 64, 2657–2661. [CrossRef] [PubMed]
40. Kushkevych, I.V. Kinetic Properties of Pyruvate Ferredoxin Oxidoreductase of Intestinal Sulfate-Reducing
Bacteria Desulfovibrio piger Vib-7 and Desulfomicrobium sp. Rod-9. Polish J. Microbiol. 2015, 64, 107–114.
41. Kushkevych, I.; Fafula, R.; Parak, T.; Bartos, M. Activity of Na+ /K+ -activated Mg2+ -dependent ATP hydrolase
in the cell-free extracts of the sulfate-reducing bacteria Desulfovibrio piger Vib-7 and Desulfomicrobium sp.
Rod-9. Acta Vet Brno 2015, 84, 3–12. [CrossRef]
42. Kushkevych, I.; Vítězová, M.; Fedrová, M.; Vochyanová, Z.; Paráková, L.; Hošek, J. Kinetic properties of
growth of intestinal sulphate-reducing bacteria isolated from healthy mice and mice with ulcerative colitis.
Acta Vet Brno 2017, 86, 405–411. [CrossRef]
420
MDPI
St. Alban-Anlage 66
4052 Basel
Switzerland
Tel. +41 61 683 77 34
Fax +41 61 302 89 18
www.mdpi.com
Energies Editorial Office

E-mail: [email protected]
www.mdpi.com/journal/energies
MDPI
St. Alban-Anlage 66
4052 Basel
Switzerland
Tel: +41 61 683 77 34
Fax: +41 61 302 89 18
www.mdpi.com ISBN 978-3-03897-597-7

Biofuel and Bioenergy Technology

Uploaded by

Copyright:

Available Formats

Biofuel and Bioenergy Technology

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Biofuel and Bioenergy Technology

Uploaded by

Copyright:

Available Formats

Biofuel and

Special Issue Editors

MDPI • Basel • Beijing • Wuhan • Barcelona • Belgrade

ISBN 978-3-03897-596-0 (Pbk)

About the Special Issue Editors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

Wei-Hsin Chen, Keat Teong Lee and Hwai Chyuan Ong

Qitai Eri, Wenzhen Wu and Xinjun Zhao

Yan-Ming Chen, Chin-Tsan Wang and Yung-Chin Yang

I-Ching Kuan, Wei-Chen Kao, Chun-Ling Chen and Chi-Yang Yu

Rei-Yu Chein and Wen-Hwai Hsu

Jeeban Poudel, Sujeeta Karki and Sea Cheon Oh

David Valero, Carlos Rico, Blondy Canto-Canché, Jorge Arturo Domı́nguez-Maldonado,

Urszula Dzieko ńska-Kubczak, Joanna Berłowska, Piotr Dziugan, Piotr Patelski,

Anna Brunerová, Hynek Roubı́k and Milan Brožek

Chen Li, Ashanti M. Sallee, Xiaoyu Zhang and Sandeep Kumar

Mohammad Anwar, Mohammad G. Rasul and Nanjappa Ashwath

Received: 24 December 2018; Accepted: 10 January 2019; Published: 18 January 2019

2. Statistics of the Special Issue

Energies 2019, 12, 290; doi:10.3390/en12020290 1 www.mdpi.com/journal/energies

• Article types: research articles (25); review articles (2).

The authors’ geographical distribution (published papers) is:

3. Brief Overview of the Contributions to This Special Issue

Research Research Type Technology/

Biodiesel • Reduction in brake power, torque and brake

Anwar et al., blend of thermal efﬁciency.

Australian • Optimisation with response surface methodology.

• Oxidizer and fuel Lewis number were between 0.4 and 1,

• Low ash content for bamboo ﬁbre (1.16%) and sugarcane

Research Research Type Technology/

• Occurrence of potentially harmful microorganisms such as

• First-Law Efﬁciency increased with increased reaction

• Optimisation with Taguchi Method.

• Hydrodynamic boundary layer of 1.6 cm (thin layer)

• Food wastes (corn stover, prairie cordgrass and unbleached

Research Research Type Technology/

Nitric acid • Jerusalem artichoke stalks were converted into bioethanol

Dziekońska- pretreatment for with nitric acid as catalyst.

Transesteriﬁcation • Ultrasonic accelerated rate of biodiesel

Equilibrium • Simulations were performed with two different models

• M082 strain showed maximum lipid content (397 mg fatty

Research Research Type Technology/

• Combined pretreatment on giant miscanthus, corn stover

and wheat straw show varied results (increased of glucan

• Pretreatment was performed using a 30 L bench-scale ball

• Acid-catalysed synthesis by 0.6 M sulphuric acid at 70 ◦ C

Research Research Type Technology/

• Emulsiﬁed ratio (bio-oil to emulsiﬁer, B/E ratio) at 1

Simultaneous showed higher solubility of 66.48 wt %.

Electrochemical • Polymer electrolyte membrane fuel cell was used to

• Optimisation with Response Surface Methodology

Research Research Type Technology/

• Wood waste was torreﬁed at 200–400 ◦ C and 0–50 min.

• Bioenergy products (alcohols and n-alkene hydrocarbons

• Biogas composition was measured for 107 small-scale

• Soursop (Annona muricata) seeds were used to produce

Research Research Type Technology/

• Biochemical Methane Potential (BMP) showed that the

addition of granular activated carbon (GAC) improved the

• Different calcination temperatures (500–900 ◦ C) of iron

Conﬂicts of Interest: The authors declare no conﬂict of interest.