Steam reforming of glycerol: Hydrogen production optimization

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 6 0 9 7 e6 1 0 6 Available online at www.sciencedirect.com ScienceDirect journal homepage: www.elsevier.com/locate/he Steam reforming of glycerol: Hydrogen production optimization , C.R. Apesteguı́a M.E. Sad*, H.A. Duarte, Ch. Vignatti, C.L. Padro lisis y Petroquı́mica e Catalysis Science and Engineering Research Group (GICIC), Instituto de Investigaciones en Cata INCAPE-(UNL-CONICET), Santiago del Estero 2654, 3000 Santa Fe, Argentina article info abstract Article history: The glycerol steam reforming reaction was studied using Pt-based catalysts in order to Received 30 December 2014 selectively produce hydrogen. The global steam reforming reaction is the combination of Received in revised form two consecutive steps: i) glycerol decomposition and ii) water gas shift reaction (WGS). 5 March 2015 Pt supported over solids with markedly different physicochemical properties (SiO2, Accepted 10 March 2015 MgO, Al2O3 and TiO2) were prepared and tested in steam reforming reaction of glycerol Available online 6 April 2015 (10% wt. aqueous solution) at 573e623 K. Glycerol to gas products conversion of 100% and Keywords: from chlorine-free solution). Acidic supports favored undesirable reactions conducting to Steam reforming liquid products and coke precursors. Furthermore, WGS reaction was studied at reaction Glycerol conditions compatible with steam reforming over Pt/SiO2, Pt/TiO2 and two catalysts pre- hydrogen yield of 78.8% were obtained by using Pt over an inert support (Pt/SiO2 prepared Pt catalysts pared for that purpose: Pt/CeO2 and Pt/ZrO2. Pt/TiO2 showed the highest CO conversion at Water gas shift reaction 623 K. In order to maximize H2 formation, a double-bed catalytic system (0.5% wt. Pt/ Hydrogen yield SiO2 þ 0.5% wt. Pt/TiO2) was used achieving a 100% hydrogen yield without deactivation on stream. Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. Introduction The depletion of worldwide oil supplies together with current awareness of the need to find alternative forms of energy from renewable resources are the main motivation for the study and development of new technologies for the production of both liquid and gaseous fuels. Carbon dioxide is the main greenhouse gas associated with global warning; it is produced in all combustion processes involving fossil fuels as well as in other industrial processes such as cement production and sweetening of natural gas [1]. One-fifth of global carbon dioxide emissions are created by the transport sector, which accounts for about 60% of global oil consumption [2]. Therefore, alternate transportation fuels, such as bioethanol, biodiesel, and hydrogen, will play an important role in the world's future [3,4]. Concretely, hydrogen has a high energy yield of 122 kJ/g, which is 2.75 times greater than hydrocarbon fuels [5] and is a clean fuel with no CO2 emissions that can easily be used in fuel cells for generation of electricity. Hydrogen can be produce by using different technologies from a wide variety of primary energy sources [6]. However, approximately 90% of the hydrogen produced nowadays comes from nonrenewable carbonaceous raw material [7]. Currently, much research has * Corresponding author. INCAPE, Santiago del Estero 2654, 3000 Santa Fe, Argentina. E-mail address: [email protected] (M.E. Sad). URL: http://www.fiq.unl.edu.ar/gicic http://dx.doi.org/10.1016/j.ijhydene.2015.03.043 0360-3199/Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. 6098 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 6 0 9 7 e6 1 0 6 been focused on sustainable and environmental friendly energy from biomass to replace conventional fossil fuels; besides biomass and biomass-derived fuels can be used to sustainably produce hydrogen [8]. Glycerol is obtained on a large amounts in South American countries as a byproduct of biodiesel production; also can be obtained from fermentation of sugars such as glucose either directly or as sub-product of lignocellulose into ethanol conversion [9]. Although glycerol is a very versatile product and it can be used in food, beverages, pharmaceuticals and to produce a variety of chemicals [10,11], a large excess is sold at low prices and therefore, is interesting to find viable processes to obtain products with higher added value and/or fuels from this molecule. Fig. 1 shows several possible reactions from glycerol to get more valuable products such as dehydration, oxidation, etherification, esterification, cracking, hydrogenolysis and CeC and CeO cleavages leading to H2, CO and CO2 or alkanes/alkenes respectively. The routes showed in Fig. 1 are not exhaustive but accounts for the numerous ways to produced chemicals or fuels from an abundant and inexpensive renewable source such as glycerol. As it is remarked in Fig. 1, we will focus in the present paper on selective H2 production from this polyol via steam reforming reaction. The overall steam reforming reaction is an endothermic reaction (DH0 ¼ 123 kJ/mol, Reaction (1)) and it is the result of combination of glycerol decomposition (Reaction (2), DH0 ¼ 245 kJ/mol) and Water Gas Shift (WGS, Reaction (3)) [12]. C3O3H8 þ 3H2O 4 3CO2 þ 7H2 (Reaction 1) C3O3H8 4 3CO þ 4H2 (Reaction 2) CO þ H2O 4 CO2 þ H2 (Reaction 3) The CO/H2 ratio formed by glycerol steam reforming depends on the reaction conditions and the catalyst employed. The CO obtained from glycerol decomposition may further react in presence of water to form CO2 by the exothermic (DH0 ¼ 41 kJ/mol) WGS reaction. Additionally, methane formation may proceed from CO or CO2 and H2 by two also exothermic reactions (Reactions (4) and (5), DH0 ¼ 206 kJ/mol and DH0 ¼ 165 kJ/mol respectively) [12]. CO þ 3H2 4 CH4 þ H2O (Reaction 4) CO2 þ 4H2 4 CH4 þ 2H2O (Reaction 5) In order to favor H2 production via steam reforming reaction, the catalyst must promote the cleavage of CeC, OeH, and CeH bonds in the oxygenated hydrocarbon reactant (leading to H2 and CO), and facilitate the water gas shift reaction to remove adsorbed CO from the surface as CO2, as opposed to the cleavage of CeO bonds (leading to alkanes) [13]. Thermodynamic studies of glycerol steam reforming with the method of Gibbs free energy minimization for hydrogen and/or synthesis gas production [14e16] concluded that high temperature, low pressure, and high water/glycerol ratio favor hydrogen production. Optimal conditions for hydrogen production from glycerol were a temperature of 925e975 K and a water/glycerol ratio of 9e12 at atmospheric pressure. Under these conditions, methane production is minimized and the carbon formation is thermodynamically inhibited. However, the biomass-derived oxygenated compounds typically have low thermal stabilities, associated with their high oxygen contents, making it difficult to process them at such a high temperature. Therefore, it is an important challenge to perform the reforming of glycerol at temperatures lower than conventional gasification of biomass (>900 K). Thus, the selection of a catalyst that promotes steam reforming reaction Fig. 1 e Glycerol valorization: different reactions conducting to valuable products. i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 6 0 9 7 e6 1 0 6 at low temperature favoring the selective formation of hydrogen is a crucial task. According to literature [17e22], several transition metals (Ru, Rh, Ni, Ir, Co, Pt, Pd and Fe) dispersed on supports are suitable to catalyze steam reforming reaction of biomass-derived oxygenated. Actually, it was reported that the cleavage of CeC bonds as well as CeH and/or OeH bonds to form adsorbed species on the catalyst surface occurs readily over Group VIII metals, mainly Pd and Pt [23]. Pt has been proved to be one of the most active and selective metal for glycerol and ethylene glycol steam reforming due to its capability to break CeC bond and suitability to catalyze the WGS reaction [17,24]. It was also reported that the support plays an important role on catalytic performance of polyols steam reforming [25,26] and water gas shift reaction [27]. We studied in this paper the effect of the nature of the support (acidity, basicity and redox properties) when using Ptbased catalysts in hydrogen production from glycerol by combination of the two reactions involved: glycerol decomposition and WGS. Specifically, we compared the hydrogen yields obtained via steam reforming of glycerol at 623 K over Pt (y0.5% wt.) supported on silice, alumina, titania and magnesium oxide. We also investigated the effect of the support when using Pt catalysts in WGS reaction at the reaction conditions similar to steam reforming. In order to maximize hydrogen production, we proposed to use a double bed system including the best catalyst for glycerol decomposition and then the most suitable catalyst for improving WGS reaction. Previous results have been reported by Kunkes et al. [28], who using an integrated catalytic system with two beds (5% wt. PtRe/C and 1% wt. Pt/CeO2/ZrO2), have reached a maximum hydrogen yield of 80%. Therefore, the objective of the present study is to explore the effect of the nature of the support for Pt-based catalysts tested in the steam reforming of glycerol (10% wt. glycerol) and water gas shift reaction to maximize the hydrogen yield. Experimental Catalyst preparation Two sets of Pt catalysts were prepared to be tested in both glycerol steam reforming (Group I) and WGS reaction (Group II). The first group of Pt-supported samples (Pt nominal 0.5% wt.) were prepared by incipient-wetness impregnation at 303 K of SiO2 (Grace G62, 99.7%), g-Al2O3 (Cyanamid Ketjen CK300), TiO2 (Hombifine N, Sachtleben Chemie) and MgO. The oxide magnesium sample was prepared by hydration of commercial MgO (Carlo Erba, 99%, 27 m2/g) in order to increase the support specific area. Distilled water (50 cm3) were slowly added to 5 g of commercial MgO and stirred at room temperature; the temperature was then raised to 353 K and stirring was maintained for 3 h. Sample was drying in an oven at 358 K overnight and the resulting Mg(OH)2 was decomposed in N2 flow at 773 K to obtain MgO. Prior to impregnation with Pt salts, all the supports, excepting MgO, were treated in air at 773 K during 4 h. Four Pt-based catalysts were prepared by using H2PtCl6.6H2O solution (Aldrich 99.995%): Pt/SiO2eI, Pt/ Al2O3, Pt/TiO2 and Pt/MgO. The Pt/SiO2-II sample was prepared 6099 with the same SiO2 support but using a chloride-free impregnation aqueous solution (tetramine platinum nitrate, Pt(NH3)4(NO3)2, Aldrich, 99.99%). The second group of catalysts includes Pt supported on CeO2 (Rhodia HSA5) and ZrO2 together with Pt/SiO2-II and Pt/TiO2 prepared as descripted above. ZrO2 support was prepared by solegel method [29]; the precursor Zr(OC3H7)4 (70% in 1-propanol, Aldrich) was dissolved in 100 cm3 of isopropyl alcohol and was slowly added (1 cm3/min) at 305 K to 100 cm3 of distilled water stirred at 500 rpm. CeO2 and ZrO2 supports were dried overnight at 393 K and treated in flowing air at 873 K for 4 h. Pt-supported samples were prepared by incipient-wetness impregnation of supports at 303 K with aqueous solutions of Pt(NH3)4(NO3)2 (Aldrich, 99.99%). After impregnation, these two samples were dried overnight at 363 K and then treated in dry air at 723 K for 3 h. Catalyst characterization BET surface areas (SBET) were measured by N2 physisorption at its normal boiling point in a Quantochrome Corporation NOVA-1000 sorptometer. The Pt loadings were measured by atomic absorption spectroscopy. The platinum dispersion (DPt) of Pt supported on SiO2, Al2O3, MgO, TiO2 and ZrO2 was determined by irreversible H2 chemisorption at 298 K in a conventional vacuum instrument equipped with an MKS Baratron pressure gauge and using the double isotherm method [30]. Catalysts (0.1 g) were reduced in H2 at 573 K for 2 h and then outgassed for 2 h at 623 K except Pt/TiO2 that was outgassed at 573 K to avoid SMSI (strong metal-support interaction) prior to performing gas chemisorption experiments. The hydrogen uptake on Pt/CeO2 was measured by performing H2 pulses at 223 K in order to minimize the atomic hydrogen migration to the support [31] using a Micromeritics AutoChem II 2920 unit. Sample (0.15 g) was reduced in H2 at 673 K, flushed with Ar at 673 K for 30 min and then cooled to 223 K in Ar. The H2 uptake measurements were performed at 223 K by injecting consecutive pulses containing 0.025 cm3 of H2 in a H2/Ar stream. In all the cases, an atomic H/ Pts ¼ 1 ratio, where Pts implies a Pt atom on surface, was used to calculate DPt. Sample acidity was characterized by temperatureprogrammed desorption (TPD) of NH3 in order to estimate the total amount of acid sites. Samples (0.15 g) were treated at 723 K for 2 h in He (60 cm3/min) and then exposed to a 1% NH3/ He stream at 373 K for 40 min. Weakly adsorbed NH3 was removed by flushing with He at 373 K (2 h). The temperature was then increased at 10 K/min and the NH3 concentration in the effluent was measured by using mass spectrometry (Baltzers Omnistar unit). Sample basicity was determined by temperature-programmed desorption (TPD) of CO2 preadsorbed at 298 K. Samples (0.15 g) were treated in N2 at 673 K for 1 h and then exposed to a 3% CO2/N2 stream until saturation coverages were reached. Weakly adsorbed CO2 was removed by flushing with N2 at room temperature for 1 h. The temperature was then increased up to 673 K (10 K/min). The desorbed CO2 was converted to methane by means of a methanation catalyst (Ni/Kieselghur) operating at 673 K and monitored using an SRI 8610C gas cromatograph with a flame ionization detector. 6100 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 6 0 9 7 e6 1 0 6 Coke formed on the catalysts during reaction was measured by temperature programmed oxidation (TPO) using a 2% O2/N2 molar stream as descripted elsewhere [32]. Samples (0.05 g) used in reaction and stabilized at reaction temperature for 45 min in He flow, were heated from 298 K to 1073 K (10 K/min). The evolved CO2 by oxidation of carbonaceous deposits was converted to methane passing through a methanation catalyst (Ni/Kieselghur) operating at 673 K. Methane was detected and quantified in an SRI 8610C gas cromatograph equipped with a flame ionization detector. Catalytic activity The steam reforming reaction of glycerol was carried out in a fixed bed reactor at atmospheric pressure and moderates temperatures (573e623 K). Samples (0.2 g, particles size 0.35e0.42 mm) were treated with pure H2 (75 cm3/min) insitu, at 623 K for 1 h before reaction in order to reduce all Pt. A 10% wt. glycerol (99.5þ%; Sigma Aldrich) aqueous solution was introduced into the reactor using a syringe pump (Cole Palmer, 74900) and vaporized into flowing He (6 cm3/min). A typical reaction was conducted at a contact time of 46 g h/mol glycerol and partial pressures of PH2O ¼ 94.6 kPa, PG ¼ 2 kPa and PHe ¼ 4.7 kPa. The effluent from reactor was cooled by passing through a condensation system and then conducted to a gaseliquid separator where condensable products were drained periodically and quantified by using an Agilent 6850 gas chromatograph equipped with a flame ionization detector and a 30 m Innowax column (inner diameter: 0.32 mm, film thickness: 0.5 m). An aqueous solution of 2-propanol (Sigma-Aldrich, 99.5%) was used as external standard to calculate glycerol conversion. The gas products (H2, CO, CO2 and CH4) were analyzed on-line in an HP 5890 gas chromatograph equipped with a thermal conductivity detector and a Hayesep D 100-120 column (5 m  1/8 in  2.1 mm) using He as gas carrier. Both liquid and gas samples were collected and analyzed every 15 min during 3 h. The carbon based total conversion of glycerol (XtG ) was calculated according to Equation (1) and accounts for the moles of glycerol (in carbon basis) converted to both gaseous and liquid products. g The conversion of glycerol to gaseous products (XG ) accounts for the amount of glycerol transformed into gaseous products containing carbon atoms of the feed, i.e. CO, CO2 and CH4 (Equation (2)). Hydrogen yield (hH2) is defined in Equation (3) where R is the H2/CO2 reforming ratio of 7/3 for glycerol. This hH2 definition takes into account that 4H2 molecules come from one molecule of glycerol whereas 3 molecules of H2 proceed from water. XtG ¼ FIG -FG  100 FIG P ai $Fi g  100 XG ¼ aG $FIG hH2 ¼ FH2 1   100 FIG R (1) the reactor, respectevely, ai are the number of C atoms in the product i molecule, Fi is the molar flow of gaseous product i formed from glycerol and FH2 is the molar flow of H2. The WGS reaction was carried out in a pyrex plug-flow fixed-bed reactor (0.8 cm ID) at 623 K and 101.3 kPa feeding a mixture H2O and CO (PCO ¼ 3 kPa, PH2O ¼ 9 kPa) balanced with He. Prior to catalytic tests, samples were reduced in pure H2 at 623 K for 1 h. On-line chromatographic analysis was performed using a gas chromatograph SRI 310C equipped with a flame ionization detector and a silica gel column. Before gas chromatographic analysis, the reaction products were separated and CO and CO2 completely converted to CH4 by means of a methanation catalyst (Ni/Kieselghur) operating at 673 K. Carbon monoxide conversion (XCO) was calculated as XCO ¼ YCO2/ (YCO þ YCO2)  100 where YCO and YCO2 are the molar fraction of CO and CO2 at the exit of the reactor, respectively. Results and discussion Catalysts characterization In order to investigate the effect of the nature of the support over glycerol steam reforming reaction, we prepared Pt catalysts using several solids with different physical and chemical properties: SiO2, Al2O3, TiO2 and MgO. Pt/CeO2 and Pt/ZrO2 were additionally prepared and characterized to be tested in WGS reaction. The physicochemical and acid/basic properties of these samples are shown in Table 1. The surface area (SBET) of the supports remained almost invariant after impregnation with Pt solution. The surface areas for all the catalysts, except Pt/MgO and Pt/ZrO2, were higher than 175 m2/g. The Pt dispersions (DPt) determined by H2 chemisorption at 298 K were about 35e50 % for all the catalysts, excepting Pt/TiO2 and Pt/ ZrO2. The acidic properties of the samples were analyzed by temperature programmed desorption of previously adsorbed NH3. Pt/TiO2 showed the highest acid site density followed by Pt/Al2O3 (173 and 118 mmol NH3/g respectively). Although the silica employed as support is not acid at all, Pt/SiO2eI sample showed a very small NH3 desorption peak attributed to the acidity caused by the residual chlorine which comes from the solution used to perform the impregnation [33]. Pt/SiO2-II, prepared by using a chlorine-free solution, did not exhibit any peak in NH3 TPD profile. The basicity of samples were studied by temperature programmed desorption of CO2 preadsorbed at room temperature. The total basic site densities were determined by integration of TPD curves and reported as mmol/g in Table 1. As expected, Pt/MgO displayed the highest basicity (479 mmol CO2/g). Pt/TiO2 and Pt/Al2O3 showed small CO2 desorption peaks (35e18 mmol CO2/g). Glycerol steam reforming (2) Effect of reaction temperature over glycerol conversion and product selectivities (3) Glycerol and water may react to produce H2, CO, CO2, CH4 and some condensable products formed by reactions of dehydration, dehydrogenation, hydrogenolysis, among others, of glycerol or derivatives thereof such as acetol, acrolein, where aG are the number of C atoms in the glycerol molecule, FIG and FG are the glycerol molar flow at the inlet and the exit of 6101 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 6 0 9 7 e6 1 0 6 Table 1 e Catalysts characterization. Catalyst Pt loading (% wt.) SBET support (m2/g) SBET catalyst (m2/g) Pt dispersion (%) NH3 TPD (mmol/g) CO2 TPD (mmol/g) Pt/SiO2-I Pt/Al2O3 Pt/TiO2 Pt/MgO Pt/SiO2-II Pt/CeO2 Pt/ZrO2 0.48 0.47 0.50 0.49 0.50 0.40 0.48 230 180 186 106 230 260 51 225 175 180 108 227 242 45 acetaldehyde and 1,2-propanediol. With the aim of selecting the most appropriate reaction temperature for the selective formation of H2, the influence of this parameter on the glycerol total conversion, XtG , glycerol to gaseous products cong version, XG , and gaseous products distribution was studied using Pt/SiO2-I; results at 573, 598 and 623 K are shown in Fig. 2. The initial XtG was about 90e100% for the range of temperature tested here and decreased with time on stream (XtG ¼ 30, 39 and 50% for 573, 598 and 623 K respectively after 3 h of reaction). The conversion of glycerol to gaseous products showed a similar tendency: it was about 40e60% initially but after 3 h reaction decay up to 25e39 %. The molar gas phase composition remained invariant during reaction in spite of the deactivation noticed. H2 was the main component within gas-phase products (55e60 % molar) and CH4 formation was very low (1e2% molar) in all the cases. It is important to remark here that the maximum H2 molar fraction possible is 70% according to the stoichiometry of Reactions (1)e(3). CO concentration for the lowest reaction temperature (573 K) was 40% molar whereas the CO2 formation was low (4% molar) at such temperature. However, after increasing 50 K, CO2 concentration increased to 20% at the expense of CO, whose concentration decreased to 22%; in addition, the highest H2 concentration was obtained at 623 K. Thus, the highest temperature tested here (623 K) seems to be the most suitable for reaching high H2 yields by improving both the glycerol conversion to gas phase products and H2 selectivity. We chose 623 K as the most appropriate temperature and we study the effect of the nature of the support as follow. 42 45 11 35 38 49 73 8 118 173 18 0 e e e 18 35 479 e e e Effect of the support on H2 production We compare glycerol conversion and gas-phase composition at identical reaction conditions (Table 2, rows 1e4) for the four catalysts of Pt prepared from H2PtCl6 aqueous solution and using different supports (SiO2, Al2O3, TiO2 and MgO). MgO is a basic solid whereas Al2O3 and TiO2 were the most acidic supports used here (Table 1). In addition, TiO2 has well-known redox properties which can be useful to catalyze WGS reaction and it was actually selected to be tested in steam reforming of glycerol because of this capability. The four catalysts (Pt/SiO2I, Pt/Al2O3, Pt/TiO2 and Pt/MgO) showed initial XtG y 95e100%. Pt/SiO2-I and Pt/MgO significantly favored the conversion of g glycerol to gaseous products (XG ¼ 60%) and consequently the production of H2 (hH2 ¼ 38.6 and 41.9% respectively). On the other hand, when Pt is supported on acid solids such as Al2O3 and TiO2, significant amounts of glycerol were consumed in undesirable side reactions such as dehydrations and dehydrogenations catalyzed by the presence of acid sites leading g to low XG (10 and 20%, respectively). Acrolein, acetol, 1,2propanediol, acetic acid and acetaldehyde were the main liquid products formed when using Pt/Al2O3 and Pt/TiO2 in good agreement with literature [25]. In all the cases, H2 was the main product among gas products followed by CO; CO2 was formed as a product of WGS reaction while insignificant amounts of methane were detected proving that methanation reaction is not favored in the conditions of the present research. Pt/TiO2 sample was the most active catalyst for the WGS reaction, as it is shown in Section Study of water gas shift reaction using Pt catalysts; however, the glycerol conversion Fig. 2 e Glycerol conversion and gas phase molar composition as a function of time-on-stream for different reaction temperatures on 0.48% wt. Pt/SiO2-I [46 g h/mol glycerol, 10% wt. glycerol, PT ¼ 101.3 kPa, PH2O ¼ 94.6 kPa, PG ¼ 2 kPa, g PHe ¼ 4.7 kPa]. Glycerol conversions: XtG (-), XG (C), molar gas phase composition: H2 (▵), CO (◊), CO2 (,), CH4 (B). 6102 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 6 0 9 7 e6 1 0 6 Table 2 e Glycerol conversion and H2 yield for Pt on different supports. g Catalyst XtG (t ¼ 0) XG (t ¼ 0) Pt/SiO2-I Pt/Al2O3 Pt/TiO2 Pt/MgO Pt/SiO2-II 100 100 100 95 100 60 10 20 60 99 % Molar gas phase H2 CO CO2 CH4 60 59 62 62 65 22 23 20 23 7 16 16 18 14 26 2 2 0 1 2 ƞH2 38.6 6.17 14.0 41.9 78.8 623 K, 46 g h/mol glycerol, 10% wt. glycerol, PT ¼ 101.3 kPa, PH2O ¼ 94.6 kPa, PG ¼ 2 kPa, PHe ¼ 4.7 kPa. g to gas products (XG ) on this solid was low because the acidity of this support promotes the formation of undesirable liquid products. Thus, Pt/TiO2 catalyst was not useful for the selective formation of H2 from glycerol because did not selectively promote the decomposition reaction of glycerol (CeC bond cleavage) which is a previous step to the WGS reaction. The relationship between the acidity of the support and the production of undesirable liquid compounds and the previous knowledge that the presence of residual chlorine from the solution used for impregnating the catalyst confer acidic properties to the catalyst [33], motivated us to synthesize a chlorine-free catalyst (Pt/SiO2-II) by employing Pt(NH3)4(NO3)2 as precursor; the absence of acidity was corroborated by TPD of NH3 (Table 1). When testing Pt/SiO2-II in steam reforming reaction, both the initial total glycerol conversion and initial conversion to gaseous products were almost 100% (Table 2, last row). The H2 yield was 78.8%, the highest value reported in Table 2. These results showed that Pt/SiO2-II sample was the most suitable catalyst to produce H2 from glycerol and water. Pt supported over an inert solid improves the steam reforming reaction by catalyzing the CeC, CeH, and OeH bonds cleavage and preventing CeO scissions that conduct to undesirable liquid products and also the hydrogenation of CO or CO2 that produce light alkanes. It is important to remark that the SiO2 used here did not show neither basic nor acid properties and was not active for steam reforming reaction either. Actually, we have tested the glycerol reforming reaction on SiO2 and no conversion of glycerol was noticed. Our results are in good agreement with the information available in literature suggesting that the selection of a non-acidic support is appropriate to syn-gas and/or hydrogen production via steam reforming reaction of several oxygenates compounds derived from biomass [25,34,17]. calculated the d0 parameter as d0 ¼ ½daG =dtŠt¼0 accounting for initial deactivation rate. As it can be observed in Fig. 3, Pt/ SiO2-II was the most stable catalyst during steam reforming reaction. In fact, the lowest d0 value was obtained on this catalyst, while Pt/Al2O3 and Pt/TiO2 catalysts presented the highest initial deactivation, more than one order of magnitude higher than those for Pt supported over non-acidic solid (d0 values were 0.0168, 0.0118 and 5  10 4 min 1 for Pt/Al2O3, Pt/ TiO2 and Pt/SiO2-II respectively). Sample deactivation could be caused by blockage of the active sites due to coke formation. Actually, it was reported [17] that catalyst deactivation would be caused by dehydration on the oxide catalyst supports, which leads to the formation of unsaturated hydrocarbon species that form carbonaceous deposits on the Pt surface, thereby decreasing the rate of H2 production. Therefore, the samples tested in reaction were recovered and analyzed by TPO technique. The amount of coke formed (mol C/g catalyst) was determined by integration of the oxidation profiles. We also calculated the mol of carbon per m2 of surface area (mol C/m2) as the mol C/g catalyst divided by the SBET showed in Table 1. In Fig. 4 we plotted the initial deactivation (d0 ) determined from Fig. 3 as a function of the carbon formation (reported as mol C/m2). A linear tendency was achieved indicating that catalyst deactivation is caused by coke formation. Indeed, the carbon content for Pt/ Al2O3 and Pt/TiO2 samples were 49.5 and 29.8 mmol C/m2, respectively, and they presented the highest initial deactivation rate. On the other hand, the less acidic sample, Pt/SiO2-II displayed almost no deactivation during reaction and formed the lowest amount of coke (2.8 mmol C/m2). Consequently, the Catalysts deactivation and coke formation We have observed in our experiments that glycerol to gasg phase products conversion (XG ) decreased with time. Other authors [17], that have also reported deactivation during steam reforming of glycerol, have indicated that the support plays an important role in this deactivation process. In order to compare the activity decay of the different Pt-based catalysts during glycerol steam reforming, we showed in Fig. 3 the activity aG as a function of time on stream; aG is defined as aG ¼ rG /r0G , where r0G and rG are the glycerol to gas-phase products conversion rates at t ¼ 0 and t ¼ t, respectively. From the initial slopes of activity versus time curves it was Fig. 3 e Time evolution of the activity for glycerol to gas phase products conversion (aG ) on Pt-based catalysts [Pt/ SiO2-II (C), Pt/MgO (,), Pt/SiO2eI (:), Pt/TiO2 (7) and Pt/ Al2O3 (A), 623 K, 46 g h/mol glycerol, 10% wt. glycerol, PT ¼ 101.3 kPa, PH2O ¼ 94.6 kPa, PG ¼ 2 kPa, PHe ¼ 4.7 kPa]. i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 6 0 9 7 e6 1 0 6 Fig. 4 e Initial catalyst deactivation (d0 ) as a function of the amount of coke formed during reaction. 6103 Accordingly, we studied the WGS performance of platinum supported catalysts at the steam reforming temperature (623 K). We choose simple oxides as supports (non-reducible: SiO2 and reducible: TiO2, CeO2 and ZrO2) and prepared Pt/ support catalysts by incipient-wetness impregnation as reported in Section Catalyst preparation. It is well known that Pt/CeO2, Pt/TiO2 and Pt/ZrO2 are suitable to catalyze water gas shift reaction at moderate reaction temperatures (473e623 K) [41e44]. Moreover, Pt supported on TiO2 has shown better activity towards the low-temperature WGS reaction (523 K) compared to other reducible oxides such as CeO2, ZrO2, and their mixed oxides [27]. We also tested Pt/SiO2-II due to its very good performance in glycerol decomposition as reported in Table 2. Catalytic results of WGS reaction are shown in Fig. 5 as CO conversion (XCO) versus time on stream feeding an excess of water (PCO ¼ 3 kPa, PH2O ¼ 9 kPa, balance He). CO conversion for all the catalysts slightly decreased during the first 60 min of reaction and then reached a stationary state. As expected, Pt/SiO2-II was the less active catalyst tested here [27,45]. CO conversion rate followed the order: Pt/TiO2 > Pt/ZrO2 > Pt/ CeO2 » Pt/SiO2-II. According to literature, the WGS reaction is promoted on Pt-based catalysts via metallic monofunctional or metal-support bifunctional mechanisms depending on the reducibility of support [27]. Platinum supported on nonreducible SiO2 catalyzes the WGS reaction through a monofunctional redox mechanism involving the adsorption and activation of reactants, CO and water, on the metallic fraction. Therefore, the activity for the WGS reaction on Pt-based catalysts is highly influenced by the nature of the support, acidity of the support seems to be related to the catalyst deactivation and the coke formation during reaction in good agreement with the widely accepted concept that acid sites catalyze reactions conducting to the formation of coke precursors. Thus, we conclude that the catalyst deactivation is mostly caused by blockage of the active sites by coke precursors formed on surface acid sites. Moreover, an important difference on deactivation and coke formation between the Pt/ SiO2-I and Pt/SiO2-II was observed, showing that the election of Pt-precursor is very important to get a more active and stable catalyst. In summary, Pt/SiO2-II efficiently improves glycerol conversion to gaseous products and hydrogen yield while minimizes the coke formation. Study of water gas shift reaction using Pt catalysts Significant amounts of CO (23-7 % molar in gas-phase products, Table 2) were formed during glycerol steam reforming experiments, indicating that WGS reaction was not operating eq at the equilibrium condition (XCO ¼ 100% [35,36]). This result motivated the idea of adding a second bed to integrate the glycerol steam reforming and water gas shift processes in a single reactor system operating at the same temperature. Therefore, the WGS catalyst must be active and stable at intermediate temperatures, at which Cu-based WGS catalysts tend to sinter and Fe-based WGS catalysts display low activity. Numerous investigators have observed that oxide-supported noble metal catalysts may offer significant advantages to Cu-based catalysts, including operation at higher temperatures and greater resistance to sintering [17,37e40]. Fig. 5 e CO conversion during water gas shift reaction over Pt-based catalysts [Pt/TiO2 (-), Pt/ZrO2 (C), Pt/CeO2 (:) and Pt/SiO2-II (A), 623 K, 2.17 g h/mol CO, PT ¼ 101.3 kPa, PCO ¼ 3 kPa, PH2O ¼ 9 kPa, PHe ¼ 89.3 kPa]. 6104 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 6 0 9 7 e6 1 0 6 Fig. 6 e Double bed system for maximizing the hydrogen production. essentially because the reaction intermediate and pathways occurring on the support in the bifunctional metal-support mechanism greatly depend on the support redox properties. The high CO conversion obtained on Pt/TiO2 (Fig. 5) is in good agreement with results previously informed in literature at lower temperature (303e573 K) [27]. Therefore, we selected the 0.5% wt. Pt/TiO2 catalyst as the most active sample to improve hydrogen formation from CO and water at 623 K. Integrated glycerol steam reforming and water gas shift reaction Results from Table 2 and Fig. 3 suggest that Pt/SiO2-II is the most active and stable catalyst for glycerol decomposition to gas products. In order to increase the H2 selectivity and reduce the CO content, we decided to add a second catalyst that favors WGS reaction. According to results informed in Fig. 5, Pt/TiO2 is the most suitable catalyst for this step. Thus, a two beds system (Fig. 6) formed by a first catalyst that favors glycerol decomposition (Pt/SiO2-II) and a second catalyst that promotes WGS reaction (Pt/TiO2), was tested in the glycerol to H2 conversion. Both catalysts were placed consecutively in the same reactor, operating at the same reaction temperature (623 K). g Fig. 7 shows the conversion of glycerol (XtG and XG ), hydrogen yield and molar composition of the gas phase when using 0.5% wt. Pt/SiO2-II (Fig. 7A) and the double bed system (Fig. 7B). In both cases, neither glycerol nor liquid products were detected in the exit of the reactor indicating that glycerol g was totally converted to gas phase products (XG ¼ 100%). Although glycerol is selectively transformed to gas products when using only Pt/SiO2-II, considerable amounts of CO is still present indicating that WGS reaction did not reach the equilibrium. After adding the second catalyst (Pt/TiO2), we showed in Fig. 7 that the amount of H2 formed was the maximum possible (70% molar fraction) according to the stoichiometry of the Reactions (1) and (2) conducting at a 100% of hydrogen yield. Additionally, the dual system showed no deactivation during 3 h reaction. Kunkes et al. [28] have previously reported that hydrogen yields of about 80% are achieved using using a two-bed system formed by a first bed of 5% wt. Pt/C or 5% wt. PteRe/C and a second bed of 1% wt. Pt/CeO2eZrO2. The results shown in this paper prove that Pt supported on SiO2 and TiO2 in amounts as low as 0.5% wt. may efficiently catalyze the glycerol steam reforming and WGS to get 100% H2 yield at 623 K. Conclusions Glycerol steam reforming reaction is an interesting way to produce eco-friendly hydrogen. The catalyst involved must be able to cleavage of CeC, OeH, and CeH bonds in the oxygenated reactant and prevent CeO scissions conducting to undesirable liquid products. In this sense, Pt is a good alternative but the election of the support is a crucial task. This research shows that catalysts based on Pt, even containing as low amount of noble metal as 0.4e0.5% wt., are suitable to reach 100% hydrogen yield when feeding a 10% wt. glycerol aqueous solution at 623 K. Fig. 7 e Glycerol conversion, gas phase composition and hydrogen yield for a single-bed and double-bed system. A: [Pt/SiO2II, 46 g h/mol glycerol 623 K, 10% wt. glycerol, PT ¼ 101.3 kPa, PH2O ¼ 94.6 kPa, PG ¼ 2 kPa, PHe ¼ 4.7 kPa]. B: [Pt/SiO2-II þ Pt/ TiO2, 46 g h/mol glycerol for each catalyst, 623 K, 10% wt. glycerol, PT ¼ 101.3 kPa, PH2O ¼ 94.6 kPa, PG ¼ 2 kPa, PHe ¼ 4.7 kPa]. i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 6 0 9 7 e6 1 0 6 Glycerol decomposes over Pt catalyst to form CO and H2; then in a consecutive step, the CO formed may react with water present in excess to render more H2 and CO2. In order to maximize the hydrogen production, we studied separately the two reactions involved: i) glycerol decomposition using Pt supported over solid with different physicochemical and acidebasic properties (i.e. SiO2, MgO, Al2O3 and TiO2), and ii) WGS reaction testing Pt supported over simple oxides (SiO2, TiO2, CeO2 and ZrO2). Pt/SiO2-II prepared from a chlorine-free precursor solution showed the highest H2 yield within the first series of catalyst (ƞH2 ¼ 78.8%). Acid supports greatly favor parallel reactions such as dehydrogenation, dehydration, etc. from glycerol conducting to undesirable products and causing catalyst deactivation. Studies of WGS reaction at temperatures compatible with glycerol steam reforming, i.e. 623 K, shows that Pt/TiO2 efficiently catalyze this step, even without extra H2 in the feed. In order to maximize the H2 production, both catalysts (0.5% wt. Pt/SiO2-II and 0.5% wt. Pt/TiO2) were placed in the reactor separated from each other and operating at the same temperature. This double-bed system allows to get the maximum hydrogen yield possible (ƞH2 ¼ 100%) without deactivation on stream. Acknowledgments We thank the Universidad Nacional del Litoral (UNL), the  n Cientı́fica y Tecnolo  gica Agencia Nacional de Promocio (ANPCyT) and the Consejo Nacional de Investigaciones Cien cnicas (CONICET), Argentina, for the financial tı́ficas y Te support of this work. references [1] Keskin A, Emiroglu AO. Catalytic reduction techniques for post-combustion diesel engine exhaust emissions. Energy Educ Sci Technol Part A 2010;25:87e103. [2] Balat M, Balat H. Recent trends in global production and utilization of bio-ethanol fuel. Appl Energy 2009;86:2273e82. [3] Huber GW, Iborra S, Corma A. Synthesis of transportation fuels from biomass: chemistry, catalysts, and engineering. Chem Rev 2006;106:4044e98. [4] Cortright RD, Davda RR, Dumesic JA. Hydrogen from catalytic reforming of biomass-derived hydrocarbons in liquid water. Nature 2002;418:964e7. [5] Kapdan IK, Kargi F. Bio-hydrogen production from waste materials. Enzym Micro Tech 2006;38:569e82. [6] Dincer I, Acar C. Review and evaluation of hydrogen production methods for better sustainability. Int J Hydrogen Energy 2014. http://dx.doi.org/10.1016/j.ijhydene.2014.12.035. [7] Ewan BCR, Allen RWK. A figure of merit assessment of the routes to hydrogen. Int J Hydrogen Energy 2005;30:809e19. [8] Balat H, Kırtay E. Hydrogen from biomass e present scenario and future prospects. Int J Hydrogen Energy 2010;35:7416e26. [9] Gong CS, Du JX, Cao NJ, Tsao GT. Coproduction of ethanol and glycerol. Appl Biochem Biotechnol 2000;84:543e60. [10] Pagliaro M, Ciriminna R, Kimura H, Rossi M, Pina CD. From glycerol to value added products. Angew Chem Int Ed 2007;46:4434e40. 6105 [11] Johnson DT, Taconi KA. The glycerin glut: options for the value-added conversion of crude glycerol resulting from biodiesel production. Environ Prog 2007;26(4):338e48. [12] Dauenhauer PJ, Salge JR, Schmidt LD. Renewable hydrogen by autothermal steam reforming of volatile carbohydrates. J Catal 2006;244:238e47.  R, Mavrikakis M, Dumesic JA. DFT studies for cleavage [13] Alcala of CeC and CeO bonds in surface species derived from ethanol on Pt(111). J Catal 2003;218:178e90. [14] Wang XD, Li SR, Wang H, Liu B, Ma XB. Thermodynamic analysis of glycerin steam reforming. Energ Fuel 2008;22:4285e91. [15] Adhikari S, Fernando S, Haryanto A. A comparative thermodynamic and experimental analysis on hydrogen production by steam reforming of glicerina. Energ Fuel 2007;21:2306e10. [16] Freitas ACD, Guirardello R. Comparison of several glycerol reforming methods for hydrogen and syngas production using Gibbs energy minimization. Int J Hydrogen Energy 2014;39:17969e84. [17] Soares RR, Simonetti DA, Dumesic JA. Glycerol as a source for fuels and chemicals by low-temperature catalytic processing. Angew Chem Int Ed 2006;45:3982e5. [18] Simonetti DA, Kunkes EL, Dumesic JA. Gas-phase conversion of glycerol to synthesis gas over carbon-supported platinum and platinumerhenium catalysts. J Catal 2007;247:298e306. [19] Adhikari S, Fernando SD, Haryanto A. Hydrogen production from glycerol: an update. Energy Convers Manag 2009;50:2600e4. [20] Iriondo A, Barrio VL, Cambra JF, Arias PL, Guemez MB, Sanchez-Sanchez MC, et al. Glycerol steam reforming over Ni catalysts supported on ceria and ceria-promoted alumina. Int J Hydrogen Energy 2010;35:11622e33. [21] Profeti LPR, Ticianelli EA, Assaf EM. Production of hydrogen via steam reforming of biofuels on Ni/CeO2eAl2O3 catalysts promoted by noble metals. Int J Hydrogen Energy 2009;34:5049e60. [22] LeValley TL, Richard AR, Fan M. The progress in water gas shift and steam reforming hydrogen production technologies e a review. Int J Hydrogen Energy 2014;39:16983e7000. [23] Mavrikakis M, Barteau MA. Oxygenate reaction pathways on transition metal surfaces. J Mol Catal A e Chem 1998;131:135e47. [24] Shabaker JW, Huber GW, Davda RR, Cortright RD, Dumesic JA. Aqueous-phase reforming of ethylene glycol over supported platinum catalysts. Catal Lett 2003;88:1e8. [25] Pompeo F, Santori G, Nichio NN. Hydrogen and/or syngas from steam reforming of glycerol: study of platinum catalysts. Int J Hydrogen Energy 2010;35:8912e20. [26] Ciambelli P, Palma V, Ruggiero A. Low temperature catalytic steam reforming of ethanol. 1.The effect of the support on the activity and stability of Pt catalysts. Appl Catal B e Environ 2010;96:18e27. [27] Ch Vignatti, Avila MS, Apesteguıı́a CR, Garetto TF. Catalytic and DRIFTS study of the WGS reaction on Pt-based catalysts. Int J Hydrogen Energy 2010;35:7302e12. [28] Kunkes EL, Soares RR, Simonetti DA, Dumesic JA. An integrated catalytic approach for the production of hydrogen by glycerol reforming coupled with water-gas shift. Appl Catal B e Environ 2009;90:693e8. [29] Duarte de Farias AM, Nguyen-Thanh D, Fraga MA. Discussing the use of modified ceria as support for Pt catalysts on wateregas shift reaction. Appl Catal B e Environ 2010;93:250e8. [30] Sinfelt JM, Carter JL, Yates DJC. Catalytic hydrogenolysis and dehydrogenation over copper-nickel alloys. J Catal 1972;24:283e96. [31] Perrichon V, Retailleau L, Bazin P, Daturi M, Lavalley JC. Metal dispersion of CeO2eZrO2 supported platinum catalysts 6106 [32] [33] [34] [35] [36] [37] [38] i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 0 ( 2 0 1 5 ) 6 0 9 7 e6 1 0 6 measured by H2 or CO chemisorption. Appl Catal A Gen 2004;260:1e8.  CL, Apesteguı́a CR. Phenol methylation on Sad ME, Padro acid catalysts: study of the catalyst deactivation kinetics and mechanism. Appl Catal A e Gen 2014;475:305e13. Garetto TF, Apesteguı́a CR. Oxidative catalytic removal of hydrocarbons over Pt/Al2O3 catalysts. Catal Today 2000;62:189e99. Ni M, Leung DYC, Leung MKH. A review on reforming bioethanol for hydrogen production. Int J Hydrogen Energy 2007;32:3238e47. Panagiotopoulou P, Papavasiliou J, Avgouropoulos G, Ioannides T, Kondarides DI. Wateregas shift activity of doped Pt/CeO2 catalysts. Chem Eng J 2007;134:16e22. Panagiotopoulou P, Kondarides DI. A comparative study of the water-gas shift activity of Pt catalysts supported on single (MOx) and composite (MOx/Al2O3, MOx/TiO2) metal oxide carriers. Catal Today 2007;127:319e29. Panagiotopoulou P, Kondarides DI. Effect of morphological characteristics of TiO2-supported noble metal catalysts on their activity for the wateregas shift reaction. J Catal 2004;225:327e36. Gorte RJ, Zhao S. Studies of the water-gas-shift reaction with ceria-supported precious metals. Catal Today 2005;104:18e24. [39] Bunluesin T, Gorte RJ, Graham GW. Studies of the water-gasshift reaction on ceria-supported Pt, Pd, and Rh: implications for oxygen-storage properties. Appl Catal B 1998;15:107e14. [40] Sato Y, Terada K, Hasegawa S, Miyao T, Naito S. Mechanistic study of wateregas-shift reaction over TiO2 supported PteRe and PdeRe catalysts. Appl Catal A e Gen 2005;296:80e9. [41] Panagiotopoulou P, Kondarides DI. Effect of the nature of the support on the catalytic performance of noble metal catalysts for the water gas shift reaction. Catal Today 2006;112:49e52. [42] Hwang Kyung-Ran, Ihm Son-Ki, Park Soon-Chul, Park JongSoo. Pt/ZrO2 catalyst for a single-stage water-gas shift reaction: Ti addition effect. Int J Hydrogen Energy 2013;38:6044e51. [43] Ammal SC, Heyden A. Origin of the unique activity of Pt/TiO2 catalysts for the wateregas shift reaction. J Catal 2013;306:78e90. [44] Kalamaras CM, Panagiotopoulou P, Kondarides DI, Efstathiou AM. Kinetic and mechanistic studies of the water gas shift reaction on Pt/TiO2 catalyst. J Catal 2009;264:117e29. [45] Wang Y, Zhai Y, Pierre D, Flytzani-Stephanopoulos M. Silicaencapsulated platinum catalysts for the low-temperature water-gas shift reaction. Appl Catal B e Env 2012;127:342e50.