Accepted Manuscript

Mistakes and inconsistencies regarding adsorption of contaminants from aqueous

solutions: A critical review

Hai Nguyen Tran, Sheng-Jie You, Ahmad Hosseini-Bandegharaei, Huan-Ping Chao

PII: S0043-1354(17)30269-5
DOI: 10.1016/j.watres.2017.04.014
Reference: WR 12811

To appear in: Water Research

Received Date: 29 October 2016

Revised Date: 29 March 2017
Accepted Date: 6 April 2017

Please cite this article as: Tran, H.N., You, S.-J., Hosseini-Bandegharaei, A., Chao, H.-P., Mistakes and
inconsistencies regarding adsorption of contaminants from aqueous solutions: A critical review, Water
Research (2017), doi: 10.1016/j.watres.2017.04.014.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

1 Mistakes and inconsistencies regarding adsorption of contaminants from aqueous

2 solutions: A critical review

3
4 Hai Nguyen Trana,b*, Sheng-Jie Youb, Ahmad Hosseini-Bandegharaeic,d, Huan-Ping Chaob*

PT
5
6
a
7 Department of Civil Engineering, Chung Yuan Christian University, Chungli 320, Taiwan

RI
b
8 Department of Environmental Engineering, Chung Yuan Christian University, Chungli 320,
9 Taiwan

SC
c
10 Wastewater Division, Faculty of Health, Sabzevar University of Medical Sciences, PO Box
11 319, Sabzevar, Iran
d
12 Department of Engineering, Kashmar Branch, Islamic Azad University, PO Box 161,

U
13 Kashmar, Iran
14
15 AN
16
M
17 Corresponding authors:

18 H. N. Tran ([email protected]) and H.- P. Chao ([email protected])

D

19 Department of Environmental Engineering, Chung Yuan Christian University, Chungli 320,

TE

20 Taiwan
21
EP

22
C
AC

1
 ACCEPTED MANUSCRIPT

23 Abstract

24 In recent years, adsorption science and technology for water and wastewater treatment has
25 attracted substantial attention from the scientific community. However, the number of
26 publications containing inconsistent concepts is increasing. Many publications either reiterate
27 previously discussed mistakes or create new mistakes. The inconsistencies are reflected by the

PT
28 increasing publication of certain types of article in this field, including “short
29 communications”, “discussions”, “critical reviews”, “comments”, “letters to the editor”, and
30 “correspondence (comment/rebuttal)”. This article aims to discuss (1) the inaccurate use of

RI
31 technical terms, (2) the problem associated with quantities for measuring adsorption
32 performance, (3) the important roles of the adsorbate and adsorbent pKa, (4) mistakes related to

SC
33 the study of adsorption kinetics, isotherms, and thermodynamics, (5) several problems related
34 to adsorption mechanisms, (6) inconsistent data points in experimental data and model fitting,
35 (7) mistakes in measuring the specific surface area of an adsorbent, and (8) other mistakes

U
36 found in the literature. Furthermore, correct expressions and original citations of the relevant
37 models (i.e., adsorption kinetics and isotherms) are provided. The authors hope that this work
38
39
AN
will be helpful for readers, researchers, reviewers, and editors who are interested in the field of
adsorption studies.
40
M
41 Keywords: Adsorption; mistake; comment; inconsistency; critical review
42
D
TE
C EP
AC

2
 ACCEPTED MANUSCRIPT

43
44 Table of Contents
45 1. Introduction ........................................................................................................................5
46 2. Technical terms used in the study of adsorption ..................................................................6
47 3. Quantities for measuring adsorption performance ...............................................................8
48 4. Incorrect assumptions regarding pKa ...................................................................................9

PT
49 5. Adsorption kinetics ...........................................................................................................11
50 5.1. The important role of initial time............................................................................11
51 5.2. Adsorption equilibrium time for porous materials ..................................................13

RI
52 5.3. Pseudo-first-order (PFO) equation..........................................................................14
53 5.3.1. Derivation of the PFO equation ...................................................................14

SC
54 5.3.2. Problems in the application of the PFO equation..........................................15
55 5.4. Pseudo-second-order (PSO) equation .....................................................................17
56 5.4.1. Derivation of the PSO equation ...................................................................17

U
57 5.4.2. Problems in the application of the PSO equation..........................................20
58 5.5. Elovich equation ....................................................................................................23
59
60
AN
5.5.1. Derivation of the Elovich equation ..............................................................23
5.5.2. Incorrect expression of the Elovich equation................................................24
61 5.6. Intra-particle diffusion model .................................................................................24
M
62 6. Adsorption isotherms ........................................................................................................25
63 6.1. Adsorption equilibrium ..........................................................................................25
D

64 6.2. Langmuir equation .................................................................................................29

65 6.2.1. Derivation of the Langmuir equation ...........................................................29
TE

66 6.2.2. Derivation of the separation factor ...............................................................31

67 6.2.3. Mistakes concerning the Langmuir model and the separation factor.............31
68 6.3. Freundlich equation ...............................................................................................33
EP

69 6.3.1. Derivation of the Freundlich equation ..........................................................33

70 6.3.2. Mistakes concerning the Freundlich equation ..............................................34
71 6.4. Redlich–Peterson equation .....................................................................................35
C

72 6.4.1. Derivation of the Redlich–Peterson equation ...............................................35

AC

73 6.4.2 Mistakes concerning the Redlich–Peterson equation .....................................36

74 6.5. Dubinin–Radushkevich equation ............................................................................36
75 6.5.1. Derivation of the Dubinin–Radushkevich equation ......................................36
76 6.5.2. Mistakes concerning the Dubinin–Radushkevich equation ...........................37
77 7. Adsorption thermodynamics .............................................................................................38
78 7.1. Principles of adsorption thermodynamics ...............................................................38
79 7.2. Equilibrium constant derived from the Langmuir constant (KL) ..............................40
3
 ACCEPTED MANUSCRIPT

80 7.3. Equilibrium constant derived from the Freundlich constant (KF) ............................44
81 7.4. Equilibrium constant derived from the partition coefficient (Kp).............................45
82 7.5. Equilibrium constant derived from the distribution coefficient (Kd) ........................47
83 8. Others mistakes ................................................................................................................49
84 8.1. Inconsistent data points in experimental data and model fitting ..............................49
85 8.2. Oxidation state of chromium ..................................................................................51

PT
86 8.3. Incorrect labels.......................................................................................................53
87 8.4. BET specific surface area of an adsorbent ..............................................................54
88 8.5. Maximum absorption wavelength in dye adsorption studies ...................................57

RI
89 8.6. Cπ-cation and π-π interactions................................................................................57
90 8.6.1. Cπ-cation interactions ..................................................................................57

SC
91 8.6.2. π-π interactions ............................................................................................59
92 8.7. Other miscellaneous errors .....................................................................................62
93 9. Nonlinear-optimization technique .....................................................................................66

U
94 10. Conclusions ....................................................................................................................68
95 Acknowledgements ..............................................................................................................69
96
97
AN
M
98
D
TE
C EP
AC

4
 ACCEPTED MANUSCRIPT

99 1. Introduction

100 Nowadays, the number of publications in international scientific journals has increasingly

101 become a standard for the assessment of scientists (Geckeis and Rabung, 2004). As the number of

102 scientific publications increase, mistakes and misconceptions enter scientific literature and some

PT
103 can also be repeated in subsequent publications. Once errors and mistakes enter the literature, it is

RI
104 difficult to eradicate them. The entrance of mistakes in scientific publications, as well as their

105 propagation, can result from the subjectivity of the authors and/or objective reasons.

SC
106 A typical study on the adsorption of organic and inorganic contaminants comprises several

107 sections, such as the effects of solution pH and ionic strength, studies of adsorption kinetics,

U
108 isotherms, thermodynamics, desorption, and regeneration. Adsorption processes can be conducted

109
AN
using either batch or column techniques. It has been found that many scientific articles in the field

110 of adsorption science and technology contain several mistakes and misconceptions including the
M
111 use of technical terms, application of quantities, determining the roles of some constants like pKa,
D

112 performing kinetic studies and modeling kinetic adsorption data, finding a suitable isotherm model
TE

113 to fit equilibrium data, calculating thermodynamic parameters, correct interpretation of adsorption

114 mechanisms, and measuring the specific surface area of an adsorbent. Moreover, a survey of the
EP

115 scientific publications related to the adsorption of contaminants in aqueous solutions reveals that

116 some less obvious inconsistencies and mistakes have unavoidably slipped the attention of authors
C

117 and were repeated in subsequent publications.

AC

118 The observation of these inconsistencies and mistakes, regardless of their origin, along

119 with comments and open discussions can prevent their propagation in scientific knowledge

120 transfer. Therefore, the presentation of a comprehensive review on common mistakes in

121 adsorption studies, besides introducing the most correct approaches to use and giving a source of
5
 ACCEPTED MANUSCRIPT

122 up-to-date literature on this issue, seems to be valuable to both readers and researchers in this

123 field.

124 The main purpose of this review is to identify mistakes in past publications and prevent the

125 propagation of such mistakes in future scientific literature. Herein, we present some common

PT
126 mistakes in various aspects of adsorption studies and analyze the reasons for their entrance into

RI
127 the literature. This was accomplished by discussing the published papers in the field of adsorption

128 of contaminants from aqueous solutions, commonly including “comments on”, “reply for

SC
129 comments”, “response to”, “authors’ response to comments on”, “comments on the authors’

130 response to comments on”, “comments on comment on”, “in reaction to”, “critical review”, and “a

U
131 note on the comments” published works. In addition to our comprehensive review on the mistakes

132
AN
existing in publications, some correct approaches are suggested where necessary to avoid the

133 propagation of such mistakes in the future scientific literature. This paper primarily focuses on
M
134 three basic constituents of adsorption theory: adsorption equilibria, kinetics, and thermodynamics.
D

135 2. Technical terms used in the study of adsorption

TE

136 Adsorption processes have their own vernacular. Therefore, correctly understandings the

137 technical terms used in adsorption technology can prevent the introduction of several unexpected
EP

138 ambiguities and discrepancies. Some basic adsorption terms are summarized in Figure 1.

139 Figure 1
C

140 To some extent, a thorough understanding of the properties of adsorbents might prevent
AC

141 common mistakes made in interpreting adsorption processes and adsorption mechanisms.

142 Numerous techniques can be used to characterize an adsorbent (Figure 2). More information on the

143 basic properties of adsorbents (i.e., biosorbent, biochar, hydrochar, and activated carbons) and the

144 relations between these properties has been published in our recent work (Tran et al. 2017c).
6
 ACCEPTED MANUSCRIPT

145 Figure 2

146 The appropriate use of technical words regarding the biosorption mechanism of heavy

147 metals has been highlighted by Robalds et al. (2016). To avoid reader confusion, they

148 recommended that “surface precipitation” be used instead of “microprecipitation” (also written as

PT
149 “micro-precipitation”). Similarly, it is necessary to distinguish between “chelation”,

RI
150 “complexation”, and “coordination”. The word “complexation” includes two meanings:

151 “chelation” and/or “coordination”. For example, the incorrect sentence “in another sense, it can

SC
152 also be defined as a collective term for a number of passive accumulation processes which in any

153 particular case may include ion exchange, coordination, complexation, chelation, adsorption and

U
154 microprecipitation” should be revised as “biosorption may include ion exchange, complexation

155
AN
(including coordination and/or chelation), physisorption, or microprecipitation” (Robalds et al.,

156 2016). In addition, the use of the term “electrostatic adsorption” may cause misunderstanding and
M
157 confusion; therefore, depending on the intended meaning, the appropriate terms (i.e., “electrostatic
D

158 attraction”, “electrostatic force”, or “electrostatic interaction”) are recommended. Notably, the
TE

159 authors also suggested a new classification system for the (bio)sorption mechanisms of heavy

160 metals (Figure 3).

EP

161 Figure 3

162 The electrical state of an adsorbent’s surface in solution is usually characterized by either
C

163 the point of zero charge (PZC) or the isoelectric point (IEP). Many authors have complained about
AC

164 the misuse of the terms IEP and PZC outside their normal meanings (Somasundaran, 1968;

165 Kosmulski, 2009). The PZC is defined as the solution conditions under which the surface charge

166 density equals zero. Potentiometric titration, or related methods, is used to determine the PZC by

167 finding the point at which the apparent surface charge density in the presence of an inert electrolyte
7
 ACCEPTED MANUSCRIPT

168 is not dependent on ionic strength (Figure 4a). The IEP occurs when the electrokinetic (ζ) potential

169 at the shear plane equals zero; ζ is determined from measuring electrokinetic (electrophoresis and

170 streaming potential) and electroacoustic phenomena (Figure 4b) (Kosmulski, 2009). IEP values

171 clearly represent only the external surface charge of particles in solution, whereas the PZC varies

PT
172 in response to the net total (external and internal) surface charge of the adsorbent (as the

RI
173 characteristic of amphoteric surfaces). Therefore, the difference between the PZC and IEP values

174 for a porous carbon system can be interpreted as a measure of the surface charge distribution. A

SC
175 difference (PZC − IEP) of greater than zero indicates that the external particle surface is more

176 negatively charged than the internal particle surface, and a difference of close to zero corresponds

U
177 to a more homogeneous distribution of surface charge. The PZC is only equal to the IEP in the

178
AN
absence of specific or counter ions. This means that equality between the PZC and IEP exists only

179 if no specific adsorption of counter ions from the solution occurs (Radovic, 1999; Onjia and
M
180 Milonjić, 2002).
D

181 Figure 4
TE

182 3. Quantities for measuring adsorption performance

183 In the field of adsorption, adsorption performance can be expressed as the amount of
EP

184 adsorbate adsorbed at equilibrium or the percentage of removed adsorbate. The amount of

185 adsorbate adsorbed at equilibrium (qe; mg/g) is often calculated using the material balance of an
C

186 adsorption system; the adsorbate, which has disappeared from the solution, must be in the
AC

187 adsorbent (Volesky, 2007).

(C o − C e )
qe = V1 (1)
m1

8
 ACCEPTED MANUSCRIPT

188 where C0 (mg/L) and Ce (mg/L) are the initial and equilibrium adsorbate concentration in solution,

189 respectively; m (g) is the dried mass of used adsorbent; and V (L) is the volume of the adsorbate

190 solution.

191 Depending on the purpose of the study, the parameter qe may be expressed in different units

PT
192 as follows: (1) mg/g for evaluating practical and engineering processes, (2) mmol/g or meq/g for

RI
193 examining the stoichiometry of a process and for studying functional groups and metal-binding

194 mechanisms, and (3) mmol/g for comparing the selective adsorption performance of various

SC
195 adsorbates (Volesky, 2007). Of these units, mg/g (milligrams of adsorbate adsorbed per gram of

196 dried adsorbent) is used most commonly for qe in adsorption studies.

U
197 In addition to qe, the adsorption performance can also be expressed as the percentage of

198
AN
removed adsorbate (%removal). However, the unit of %removal needs to be used cautiously, as it

199 is very approximate and can cause misleading conclusions about relative adsorption performance.
M
200 This unit is only appropriate for the purpose of crude measurements and perhaps for quick and
D

201 very approximate screening of adsorbent materials (Kratochvil and Volesky, 1998; Hai, 2017)
TE

(C o − C e )
% removal = × 100 (2)
Co
EP

202 To some extent, the use of %removal in the study of adsorption equilibria can cause

203 incorrect observations and, as a result, inaccurate conclusions (a detailed discussion is provided in
C

204 Section 6.1).

AC

205 4. Incorrect assumptions regarding pKa

206 A mistake regarding the fundamentals of chemistry was commented on by Rayne (2013).

207 The commented paper reported the adsorption of n-perfluorooctanesulfonamide (PFOSA) onto

208 multiwalled carbon nanotubes with differing oxygen content. However, PFOSA was mistakenly

9
 ACCEPTED MANUSCRIPT

209 assumed to be an “organic base”, resulting in the following invalid interpretation: “for PFOSA

210 (pKa = 6.52), when pH < pKa, protonation occurs on the amino group, and the decreased

211 protonation leads to the increased adsorption, but when pH > pKa, PFOSA exists as neutral

212 molecule in water.” In fact, PFOSA is an acid and, thus, is a neutral compound below its pKa value

PT
213 (estimated at between 6.0 and 7.0) and a dissociated anion at pH > pKa (Rayne, 2013).

RI
214 Recently, the effect of pH on the interaction of copper with the surface functional groups of

215 raw pomegranate-peel biosorbent was reported as follows:

SC
R–OH2+ ↔ R–OH + H+ (3)

R–OH ↔ R–O− + H+ (4)

U
R–O− + Cu2+ ↔ R–OCu+ (5)

R–O− + Cu(OH)+ ↔ R–OCu(OH)

AN (6)

216 where –R represents the surface adsorbent; R–OH2+, R–OH, and R–O− represent protonated,
M
217 neutral, and ionized surface hydroxyl functional groups, respectively; and R–OCu+ and R–
D

218 OCu(OH) are the formed bonding complexes.

TE

219 The authors (Ben Ali et al., 2016) explained that, “the effect of pH could be also explained

220 by considering the point of zero charge of the adsorbent (pHpzc). The adsorbent surface is
EP

221 positively charged for pH< 4.7 and it becomes negatively charged at pH value above 4.7.

222 Therefore, for pH values < 4.7, the adsorption is unfavorable because of repulsive electrostatic
C

223 interactions between metal ions and positively charged functional groups. The maximum
AC

224 adsorption of Cu(II) occurs at pH above pHpzc value when the adsorbent surface is negatively

225 highly charged.” However, the important role of the pKa values of carboxylic acid groups (–

226 COOH; pKa 1.7–4.7) and hydroxyl groups (–OH; pKa 9.5–13) was ignored (Volesky, 2007).

227 Correspondingly, Eqs. 3–6 indicate a misunderstanding of the surface chemistry of the adsorbent.
10
 ACCEPTED MANUSCRIPT

228 The authors investigated the effect of solution pH on the adsorption process of Cu(II) in solutions

229 with pH values from 2.0 to 6.0. Therefore, dissociation of the –OH groups to –O− did not occur at

230 these solutions pH values. Instead, the –COOH groups dissociated, forming negatively charged

231 carboxylate groups (–COO‒) in the aforementioned pH range, and were the dominant groups that

PT
232 interacted with copper ions (Tran, 2017).

RI
233 Moreover, Bui and Choi (2010) made a critical comment that Oleszczuk et al (2010)

234 mistyped the properties of two pharmaceuticals: oxytetracycline (logKow = 2.45, pKa = 7) and

SC
235 carbamazepine (logKow = −1.22, pKa = 3.27, 7.32, 9.11). The correct expression should be

236 oxytetracycline (Sw = 121 mg/L, logKow = −1.22, pKa = 3.27, 7.32, 9.11; density = 1.63 g/cm3) and

U
237 carbamazepine (Sw = 112 mg/L, logKow = 2.45, pKa = 7; density =1.15 g/cm3). Bui and Choi

238
AN
(2010) also emphasized that the pKa value of carbamazepine reported by Oleszczuk et al (2010) is

239 still debated. Carbamazepine (CBZ) is commonly characterized with two pKa values: pKa1 (= 1 or
M
240 2.3) for the equilibrium RCONH3+ ⇌ RCONH2 + H+ and pKa2 (= 13.9) for the equilibrium
D

241 RCONH2 ⇌ RCONH− + H+. The misreporting of the CBZ pKa can cause inaccurate conclusions.
TE

242 For example, Oleszczuk et al. (2010) suggested that an electrostatic interaction occurred between

243 CBZ and multiwalled carbon nanotubes (MWCNT) at pH < 7.0 because they opined that CBZ
EP

244 existed in its cationic form at pH < 7.0 (pKa = 7.0). However, CBZ with pKa1 < 2.3 and pKa2 ≥ 13.9

245 actually exists entirely as a neutral species in the pH range of 1.9–11.9, as calculated using
C

246 ACDLab. As a result, the electrostatic interaction was ruled out as a contributing interaction in the
AC

247 CBZ-adsorption mechanisms (Bui and Choi, 2010).

248 5. Adsorption kinetics

249 5.1. The important role of initial time

250 In the study of adsorption kinetics, Azizian (2006) proposed that experiments should be
11
 ACCEPTED MANUSCRIPT

251 started at t < 2 min (considered as the initial time). Adsorption kinetics data in the initial time

252 periods play an important role in performing accurate modeling and drawing conclusions.

253 Depending on the affinity between the adsorbent and adsorbate, the initial adsorption rate

254 may be very fast. For the adsorption of heavy metals, Tran et al. (2015) investigated the adsorption

PT
255 of Cd2+ onto orange-peel-generated biochar, and found that equilibrium can be reached rapidly in

RI
256 kinetic experiments with a removal rate of 80.6–96.9% (within 1 min). An analogous performance

257 was reported by Guo et al. (2015), with approximately 70.0–96.6% of the total Cd2+ in solution

SC
258 removed within the first minute of contact. Additionally, Müller et al. (2012) noted that

259 approximately 90% of Zn(II) adsorption onto Lewatit S1468 occurs within 1 min, and this

U
260 adsorption process reaches equilibrium in approximately 5 min.

261
AN
For the adsorption of organic contaminants, kinetic studies indicated that the adsorption

262 equilibrium of methylene green 5 was quickly established, with a low activation energy required
M
263 for adsorption (Ea; 3.30–27.8 kJ/mol), and activated carbons removed 50–73% of the dye from
D

264 solution within 1 min (Tran et al., 2017d; Tran et al., 2017e). Similarly, Canzano et al. (2012)
TE

265 noted that the adsorption process of anionic Congo-red dye onto raw and acid-treated pine cone

266 powder reached equilibrium in the first few minutes (< 5 min), during which no measurements
EP

267 were carried out.

268 As time goes on, the amount of adsorbate uptake per mass of adsorbent at any time t (qt)
C

269 approaches that of at equilibrium (qe) and, therefore, plotting the kinetic data is bound to produce a
AC

270 straight line, independent of the real kinetic order followed by the system. In other words,

271 equilibrium data does not describe the way by which the system reaches equilibrium. As a result,

272 the use of data recorded over a significant time interval after the attainment of equilibrium, or very

273 close to equilibrium, is likely to lead to erroneous conclusions regarding adsorption kinetics.
12
 ACCEPTED MANUSCRIPT

274 5.2. Adsorption equilibrium time for porous materials

275 The adsorption processes for inorganic and organic contaminants onto porous adsorbents

276 (i.e., activated carbon, biochar, activated carbon spheres, zeolite, and macroreticular resin) can

277 take longer (i.e., several days or weeks) to approach true equilibrium than nonporous adsorbents

PT
278 (i.e., biosorbent and hydrochar). This difference is attributed to the different adsorption

RI
279 mechanisms of these adsorbents. Pore filling is the most common adsorption mechanism for

280 porous adsorbents, along with several other interactions (i.e., electrostatic attraction, hydrogen

SC
281 bonding, surface precipitation, π-π interaction, n-π interaction, and cation exchange) that depend

282 on the characteristics of the porous adsorbents and adsorbates. Many studies have opted for a 24 h

U
283 equilibrating time for isotherm adsorption experiments, without providing evidence that

284
AN
adsorption in their system attained true equilibrium within the appropriate timeframe.

285 Therefore, it is necessary to provide the specified timeframe in which the adsorption
M
286 process approaches a true equilibrium. Figures 5a–c represent the influence of contact time on the
D

287 adsorption processes of various contaminants onto porous adsorbents (Hung and Lin, 2006;
TE

288 Masson et al.; 2016; Kim et al., 2016). Clearly, the amount of contaminant adsorbed onto the

289 porous adsorbents continued to increase after 24 h. For example, Kim et al. (2016) speculated on
EP

290 the process of methyl violet (MV) adsorption onto synthesized granular mesoporous carbon

291 (GMC) and commercial granular activated carbon (GAC). The results of their kinetic study
C

292 demonstrated that MV adsorption reached an equilibrium state within 20 d for GMC-polyvinyl
AC

293 alcohol and 26 d for GMC-carboxymethyl cellulose, while adsorption onto GAC had not reached

294 equilibrium even after 75 d. Furthermore, a phenomenon involving desorption from the adsorbent

295 surface after reaching adsorption equilibrium (Figure 5d) has been reported in the literature

296 (Sharma et al., 2013, Halder; Khan et al., 2016). Therefore, studies of the adsorption kinetics play
13
 ACCEPTED MANUSCRIPT

297 an important role in identifying the required equilibration time and the optimal contact time and for

298 an adsorption process.

299 Figure 5

300 In aqueous-phase adsorption, two kinetic reaction models (i.e., pseudo-first-order and

PT
301 pseudo-second-order equations), the Elovich or Roginsky–Zeldovich model, and the intraparticle

RI
302 diffusion model have been widely applied to mathematically describe the intrinsic kinetic

303 adsorption constant. Furthermore, most mistakes found and discussed in the literature correspond

SC
304 to the use of kinetic models. Therefore, the following sections (5.3–5.6) include a thorough

305 discussion of these models.

U
306 5.3. Pseudo-first-order (PFO) equation

307 5.3.1. Derivation of the PFO equation

AN
308 Recently, Ho published a series of critical papers (Ho and McKay, 1998b; Ho, 2005, Ho,
M
309 2014a; Ho, 2015; Ho, 2016e) discussing two main points. First, Ho (Ho and McKay, 1998b)
D

310 claimed to offer the correct reference style for citing Lagergren’s paper, with the original
TE

311 translated from “Lagergren, S. (1898), Zur theorie der sogenannten adsorption gelöster stoffe.

312 Kungliga Svenska Vetenskapsakademiens Handlingar, Band 24, No. 4, 1–39” into English as
EP

313 “Lagergren, S., About the theory of so-called adsorption of soluble substances, Kungliga Svenska

314 Vetenskapsakademiens. Handlingar, Band 24, No. 4, 1898, pp. 1–39.” Second, Ho stated that
C

315 Lagergren’s first-order rate equation has been called pseudo-first-order since 1998 (Ho, 2005) in
AC

316 order to distinguish kinetic equations based on the adsorption capacity of solids from those based

317 on the concentration of a solution. However, the responses of many authors (Özacar; 2005, Kumar,

318 2006d; Utomo et al., 2010) state that Lagergren’s first-order rate expression had been referred to as

319 a pseudo-first-order equation before Ho’s publication in 1998. For example, in 1990, Sharma et al.
14
 ACCEPTED MANUSCRIPT

320 (1990) studied the removal of Ni(II) from an aquatic environment by wollastonite. In the section

321 on adsorption kinetics, they stated: “the adsorption of Ni (II) on wollastonite follows the first order

322 adsorption rate expression of Lagergren (1898).”

323 The great contributions of Ho toward correcting the citation style format for the PFO

PT
324 equation cannot be denied. However, the PFO equation was not originally proposed or initially

RI
325 expressed by Ho. In the scientific community, the authors who first propose a theoretical model

326 should always be given credit for their contributions via correct citation (Kumar and Rattanaphani,

SC
327 2007; Wu, 2009; Utomo et al., 2010). Therefore, the original work by Lagergren (1898), who first

328 presented the first-order-rate equation for the adsorption of oxalic acid and malonic acid onto

U
329 charcoal, should be cited. The PFO equation can be expressed correctly in nonlinear (Eq. 7) and

330 linear forms (Eq. 8 or Eq. 9).

AN
qt = qe (1 − e −k1t )
M
(7)

k1 (8)
log(qe − qt ) = log(qe ) − t
D

2.303

ln(qe − qt ) = ln(qe ) − k1t

TE

(9)

331 where qe and qt are the amounts of adsorbate uptake per mass of adsorbent at equilibrium and at
EP

332 any time t (min), respectively; and k1 (1/min) is the rate constant of the PFO equation.

333 5.3.2. Problems in the application of the PFO equation

C

334 It cannot be denied that the PFO equation is widely applied in adsorption kinetics.
AC

335 However, several problems related to the application of the PFO equation have been discussed in

336 the literature. First, many incorrect expressions for its linear forms have been published. Various

337 incorrect forms are shown in Eq. 10 (Alagumuthu et al., 2010), Eq. 11 (Fu et al., 2007; Jing et al.,

338 2014; Nekouei and Nekouei, 2017), Eq. 12 (Ho, 2014e; Ho, 2014a; Jing et al., 2014; Ho, 2016e;

15
 ACCEPTED MANUSCRIPT

339 Ho, 2016b; Ho, 2016c; Ho, 2016a; Ho, 2016d; Ho, 2017), Eq. 13 (Ho, 2015), Eqs. 14 and 15 (Jing

340 et al., 2014), Eq. 16 (Kumar, 2006a), and Eq. 17 (Kumar and Porkodi, 2007b), which should be

341 avoided in future studies. As a result of these incorrect expressions, recalculation of the PFO

342 equation parameters was necessary to obtain the correct parameters (Wu et al., 2017).

PT
log(qe ) − k1
log(qe − qt ) = t (10)
2.303

RI
log(qe ) − k1
log(qe − qt ) = (11)
2.303t

SC
1 k1 1
= + (12)
qt qet qe

U
k1t
log(qe − qt ) = log(qe ) (13)

Ci kt
2.303
AN
log = 1 (14)
Ct 2.303
M
C i − Ct kt
log(1 − = 1 (15)
Ci − Ce 2.303
D
TE

qe k1
log = (16)
qe − qt 2.303t
EP

k1
log(qe − qt ) = log(qe ) − (17)
2.303

343 where Ci (mg/L), Ct (mg/L), and Ce (mg/L) are the concentrations of adsorbate at the initial time (t
C

344 = 0), any time t, and equilibrium, respectively.

AC

345 The second problem is the two unknown parameters (qe and k1). Furthermore, in most

346 cases, the PFO equation is only appropriate for the initial 20 to 30 min of contact time, not for the

347 whole range (Ho and McKay, 1998a). Thus, plots of Eq. 8 or Eq. 9 are only linear over

348 approximately the first 30 min; beyond this initial period, the experimental and theoretical data
16
 ACCEPTED MANUSCRIPT

349 will not fit adequately (McKay et al., 1999). Another important problem is the selection of an

350 appropriate qe value. Notably, the adjusted qe value cannot be lower than the maximum measured

351 value of qt. This is because mathematical errors associated with taking the logarithm of a negative

352 number will occur when using Eq. 8 or Eq. 9 (Plazinski et al., 2009). Finally, the qe value

PT
353 calculated using the PFO equation is not equal to the qe value obtained from experiments, which

RI
354 further indicates the inability of the PFO equation to fit kinetic adsorption data (Febrianto et al.,

355 2009). The presence of a boundary layer or external resistance controlling the beginning of the

SC
356 sorption process was argued to be responsible for this discrepancy (McKay et al., 1999).

357 Therefore, two methods have been recommended to obtain an accurate estimation of the

U
358 kinetic parameters in the PFO equation: (1) a trial and error process to obtain the optimal qe value

359
AN
(Ho and McKay, 1998a), or (2) application of a nonlinear optimization technique. A simple guide

360 to using the nonlinear method is introduced in Section 9.

M
361 5.4. Pseudo-second-order (PSO) equation
D

362 5.4.1. Derivation of the PSO equation

TE

363 In 1984, Blanchard et al. (1984) initially proposed a second-order rate equation for the

364 removal of heavy metals from water using natural zeolites, as follows:
EP

dn
= K (no − n)2 (18)
dt
C

Integration of Eq. 18 gives:

AC

1
− α = Kt (19)
no − n

After rearranging, Eq. 19 becomes:

Ktno + αno −1
n= (20)
Kt + α

17
 ACCEPTED MANUSCRIPT

365 where n is the amount of M2+ fixed or the amount of NH4+ released at each instant; no is an

366 exchange capacity; and K is a rate constant.

367 Plotting 1/(n0 − n) as a function of time (t) must produce a straight line, the slope of which

368 gives the rate constant K and the intercept leads to the exchange capacity. Considering the

PT
369 boundary condition of n = 0 for t = 0, it follows that α = 1/n, and Eq. 20 can be rearranged to:

RI
n2 + Kt
no = (21)
1 + Knt

SC
370 If qt = no, qe = n, and k2 = K, Eq. 21 becomes:

qe2k2t
qt = (22)

U
1+ k2qet

371

372
AN
where qe (mg/g) and qt (mg/g) are the amount of adsorbate adsorbed at equilibrium and at any t

(min), respectively; and k2 (g/mg × min) is the rate constant of the PSO equation.
M
373 The nonlinear form of the PSO equation (Eq. 22) has been applied elsewhere (Ho, 1995;

374 Ho; Wase et al., 1996). Recently, a series of comments made by Ho implied that the derivation and
D

375 first application of the PSO equation belongs to his work. Any application of the PSO equation in
TE

376 the field of adsorption studies without citation of the original work or with a misquotation (i.e., the

377 citation of a secondary reference) might suffer from his comments. However, in reaction to Ho’s
EP

378 comments, a series of notes were published by other scholars (Kumar, 2006d; Kumar, 2006c;
C

379 Kumar and Fávere, 2006; Kumar and Guha, 2006; Kumar and Rattanaphani, 2007) to clarify the
AC

380 correct citation of the PSO equation. Thus, the nonlinear form of the PSO equation for solid-liquid

381 adsorption systems was not originally reported by Ho (1995) or first applied by Ho et al. (1996),

382 but initially proposed by Blanchard et al. (1984).

383 In the literature, the PSO equation can be correctly expressed in four linear forms (Eqs. 23–

18
 ACCEPTED MANUSCRIPT

384 26), and the initial adsorption rate (h) can be described by Eq. 29.

385 Expression for the linear form (Type 1) with a plot of t/qt versus t:

t 1 1
= ( )t + (23)
qt qe k2qe2

PT
386 Expression of the linear form (Type 2) with a plot of 1/qt versus 1/t:

1 1 1 1
=( 2) +

RI
(24)
qt k2qe t qe

387 Expression of the linear form (Type 3) with a plot of qt versus qt/t:

SC
1 qt
qt = − ( ) + qe (25)
k 2 qe t

U
388 Expression of the linear form (Type 4) with a plot of qt/t versus qt:

qt
= −(k 2 qe )qt + k 2 qe2
AN (26)
t
M
389 The initial adsorption rate (h) was proposed by Ho et al. (1996). First, Eq. 22 can be

390 rearranged to:

D

1
TE

qt =
1 t (27)
2
+
k 2 qe qe
EP

391 Subsequently, when t approaches 0, Eqs. 23 and 27 can be expressed as Eq. 28. Setting h =

392 qt/t, the initial adsorption rate (h; mg/g × min) can then be determined by Eq. 29.
C

t 1 qt
= ↔ = k 2 qe2 (28)
AC

qt k 2 qe2 t

h = k 2 qe2 (29)

393 Clearly, only the linear form (Type 1, Eq. 23) might have been reported by Ho et al., (1996)

394 for the adsorption of dyes from waste streams by peat (Özacar, 2005; Kumar, 2006d; Kumar and

19
 ACCEPTED MANUSCRIPT

395 Fávere, 2006; Wu, 2009). Kumar and Fávere (2006) hypothesized that, in adsorption research, the

396 Langmuir equation (Langmuir, 1918) has been the most widely used isotherm to describe

397 adsorption processes under equilibrium conditions. In the literature, four linear types of the

398 Langmuir isotherm have been reported. Irrespective of the linear expression reported, these

PT
399 equations have been widely called the Langmuir isotherm. It is not correct to claim that the

RI
400 Langmuir model was derived or proposed by other authors who transformed the original Langmuir

401 expression into a new linear expression (Kumar and Guha, 2006). In scientific literature, the names

SC
402 of scientists (i.e., Langmuir or Freundlich) have always been presented, even after modification of

403 their equations (Arica, 2003). Similarly, Utomo et al. (2010) suggested that it might not be

U
404 necessary to cite papers that deal with mathematical modification and different mathematical

405
AN
expressions of the models (i.e., PFO and PSO, and Langmuir). However, from the scientific point

406 of view, the authors who first proposed a theoretical model should be credited because of their
M
407 great contributions to the advancement of science. Therefore, the original work by Blanchard et al.
D

408 (1984) should be cited for the expression of the PSO equation (Özacar, 2005; Kumar, 2006d;
TE

409 Kumar and Fávere, 2006; Plazinski et al., 2009; Wu, 2009; Utomo et al., 2010).

410 5.4.2. Problems in the application of the PSO equation

EP

411 First, many incorrect expressions of the linear form of PSO equation are found in the

412 literature, as shown in Eq. 30 (Fu et al., 2007; Kumar and Porkodi, 2007b), Eq. 31 (Ho, 2014f),
C

413 Eqs. 32 and 33 (Ho, 2014g), Eq. 34 (Ho, 2013), Eq. 35 (Lin and Wang, 2009), Eq. 36 (Ho, 2014c),
AC

414 Eq. 37 (Ho, 2014d), Eq. 38 (Ho, 2014), and Eq. 39 (Ho, 2015).

t 1 1
= 2
+ (30)
qt k 2 qe qe t

20
 ACCEPTED MANUSCRIPT

t 1 1
= + (31)
qt 2k 2 qe

t 1 t
= + (32)
qt k 2 q e q e

PT
t 1 1
= ( qe2 ) + t (33)
qt k2 qe

RI
t 1 1
= + t (34)
qt k2 − qe qe
2

SC
t 1 1
= 2
( )t (35)
qt (k2 qe ) qe

U
t 1 t
= k2 qe2 + (36)
qt 2 qe AN
1 1 1
= 2
+ (37)
qt k2 qe qe
M

t 1 1
= + (38)
D

2
qt k2 qe qe
TE

t kt t
= 22 + (39)
qt k2 qe qe
EP

415 Second, based on the best fit of kinetic experimental data using the PSO model, many

416 authors have drawn controversial conclusions. For example, “the adsorption process is the
C

417 chemisorption, which involve valence forces through sharing electrons between adsorbate and
AC

418 adsorbent”, “the best fit of experimental kinetic data in pseudo-second-order kinetics suggests the

419 chemisorption, which may involve valency forces through sharing of electrons between dye anion

420 and adsorbent”, and other similar conclusions. The problem with such conclusions has been

421 discussed by several researchers (Kumar, 2006a; Lima et al., 2015; Lima et al., 2016), who

21
 ACCEPTED MANUSCRIPT

422 claimed that adsorption mechanisms cannot be directly assigned based on observing simple kinetic

423 experiments or by fitting kinetic models (i.e., the PFO and PSO models).

424 Adsorption mechanisms can only be established by (1) using several analytical techniques

425 (i.e., FTIR, SEM, nitrogen adsorption-desorption isotherms, Raman spectroscopy, TGA/DTA,

PT
29 13
426 DSC, Si and C solid-state NMR, XRD, XPS, pHPZC, pHIEP, CHN element analysis, Boehm

RI
427 titration, and solution calorimetry), and (2) having a good sense of the chemical nature of the

428 adsorbate and adsorbent, adsorbent’s surface, and chemical or physical interactions between the

SC
429 adsorbent and adsorbate (Volesky, 2007; Lima et al., 2015; Lima et al., 2016). The use of

430 analytical techniques together with adsorptive thermodynamic data (i.e., changes in enthalpy and

U
431 entropy) and activation and adsorption energies, are necessary to confirm whether the adsorption

432
AN
of contaminants in aqueous solution is a chemical or physical process (Lima et al., 2015; Tran et

433 al., 2016).

M
434 Finally, in most cases, experimental data for adsorption kinetics are well fitted by the linear
D

435 form of the PSO equation. However, caution should be taken in drawing conclusions based on
TE

436 fitting to the linear form of the PSO equation. For example, Figure 6 presents a plot of t/qt versus t.

437 Clearly, the R2 values for the linear form of the PSO equation are very high (R2 > 0.99); however,
EP

438 the corresponding R2 values for the nonlinear form of the PSO equation are significantly lower (R2

439 = 0.53–0.68). This finding means that the adsorption process of methylene green 5 (MG5) onto
C

440 synthesized activated carbon and commercial activated carbon (porous materials) is not
AC

441 adequately described by the PSO equation. In fact, Tran et al. (2017a) concluded that the process

442 of MG5 adsorption onto these porous materials involved π-π interactions and pore filling, and was

443 not related to chemical adsorption. Obviously, the nonlinear method can be applied to obtain

444 kinetic model parameters that are more accurate than those obtained using the linear method.
22
 ACCEPTED MANUSCRIPT

445 Figure 6

446 5.5. Elovich equation

447 5.5.1. Derivation of the Elovich equation

448 An empirical equation was firstly proposed by Roginsky and Zeldovich (1934) for the

PT
449 adsorption of carbon monoxide onto manganese dioxide. However, this equation is now generally

RI
450 known as the Elovich equation and has been extensively applied to chemisorption data

451 (McLintock, 1967). This equation can be expressed mathematically as follows:

SC
dqt
= α exp(− βqt ) (40)
dt

U
452 where qe and qt are the amounts of adsorbate uptake per mass of adsorbent at equilibrium and at

453 AN
any time t (min), respectively; α (mg/g × min) is the initial rate constant because dqt/dt → α when

454 qt → 0; and β (mg/g) is the desorption constant during any one experiment.
M
455 By applying the boundary conditions of qt = 0 at t = 0, the integrated form of Eq. 40 will
D

456 become Eq. 41 (nonlinear).

TE

1
qt = ln(1 + αβ t ) (41)
β

457 To simplify the Elovich equation, Chien and Clayton (1980) assumed αβt >> 1. Thus, a
EP

458 linear form (Eq. 42) can be obtained:

C

1 1
qt = ln(αβ ) + ln(t ) (42)
β β
AC

459 and a plot of qt versus lnt should give a linear relationship with a slope of (1/β) and an

460 intercept of (1/β)ln(αβ).

23
 ACCEPTED MANUSCRIPT

461 5.5.2. Incorrect expression of the Elovich equation

462 Errors in the presentation of the nonlinear form of the Elovich model are common in the

463 literature. These incorrect equations (Eqs. 43–45) can be found in several previous papers (Ho and

464 McKay, 1998a; Lin and Wang, 2009; Alagumuthu et al., 2010; Ben Ali et al., 2016).

PT
qt = β ln(αβ ) + ln(t ) (43)

RI
1
qt = β ln(αβ ) + ln(t ) (44)
β

SC
qt = β ln(αβ ) + β ln(t ) (45)

465 5.6. Intra-particle diffusion model

U
466 The linearized transformation of the intra-particle diffusion model (Weber and Morris,

467 1963) is presented as follows:

AN
qt = k p t + C (46)
M

468 where kp (mg/g × min) is the rate constant of the intra-particle diffusion model and C (mg/g) is a
D

469 constant associated with the thickness of the boundary layer, where a higher value of C
TE

470 corresponds to a greater effect on the limiting boundary layer.

471 Generally, although the PSO model can adequately describe adsorption kinetic
EP

472 experimental data, this model does not reveal the adsorption mechanisms. Similarly, the Elovich
C

473 equation seems to describe various reaction mechanisms, such as bulk diffusion and surface
AC

474 diffusion. In contrast, the intra-particle diffusion model can be useful for identifying the reaction

475 pathways and adsorption mechanisms and predicting the rate-controlling step. In a solid-liquid

476 sorption process, adsorbate transfer is often characterized by film diffusion (also known as

477 external diffusion), surface diffusion, and pore diffusion, or combined surface and pore diffusion.

478 In short, if a plot of qt against t0.5 is linear and passes through the origin, the adsorption is entirely
24
 ACCEPTED MANUSCRIPT

479 governed by intra-particle diffusion. In contrast, if the intra-particle diffusion plot gives multiple

480 linear regions, then the adsorption process is controlled by a multistep mechanism.

481 Four steps associated with transport processes during adsorption by porous adsorbents

482 were originally proposed in reference (Walter, 1984). The first stage is transport in the solution

PT
483 phase (known as “bulk transport”; occurs quickly), which can occur instantaneously after the

RI
484 adsorbent is transferred into the adsorbate solution; therefore, it does not control engineering

485 design. In most cases, this stage occurs too rapidly and its contribution is considered negligible.

SC
486 The second stage is “film diffusion” (occurs slowly). In this stage, the adsorbate molecules are

487 transported from the bulk liquid phase to the adsorbent’s external surface through a hydrodynamic

U
488 boundary layer or film. The third stage involves diffusion of the adsorbate molecules from the

489
AN
exterior of the adsorbent into the pores of the adsorbent, along pore-wall surfaces, or both (known

490 as “intraparticle diffusion”; occurs slowly). The last stage, adsorptive attachment, often occurs
M
491 very quickly; therefore, it is also not significant for design. These four steps are summarized in
D

492 Figure 7 (Weber and Smith, 1987).

TE

493 Figure 7

494 6. Adsorption isotherms

EP

495 6.1. Adsorption equilibrium

496 Milonjić (2009a) and Milonjić (2010) suggested that the amount of adsorbate adsorbed onto
C

497 an adsorbent depended on the equilibrium concentrations of metal ions, equilibrium solution pH,
AC

498 and temperature. An adsorption isotherm should be given as the equilibrium adsorbed amounts

499 versus the equilibrium ion concentrations for a constant equilibrium solution pH and temperature.

500 Thus, environmental parameters in the sorption system (especially solution pH) must be carefully

501 controlled at the given value over the entire contact period until the sorption equilibrium is
25
 ACCEPTED MANUSCRIPT

502 reached. These comments were in good agreement with the suggestion that metal-ion adsorption is

503 pH dependent (Volesky, 2003; Tien, 2007; Volesky, 2007; Tien, 2008). However, the pH effect

504 has invariably been examined in terms of the initial pH of the aqueous solution. Not a single case

505 has reported data demonstrating the changes in the solution pH during the course of an adsorption

PT
506 experiment.

RI
507 In the scientific literature, adsorption equilibrium studies in batch experiments have been

508 conducted in different ways as follows: (1) solution pH values were not mentioned in the

SC
509 manuscript, (2) initial pH values were neither measured nor adjusted and final pH values were

510 measured, (3) initial and final pH values were measured, (4) initial pH values were adjusted and

U
511 final pH values were not mentioned, (5) initial pH values were adjusted and final pH values were

512
AN
measured, (6) initial and final pH values were adjusted, (7) pH was controlled several times during

513 the course of the reaction, (8) a constant pH was maintained throughout the equilibration time by
M
514 adding acidic or alkaline solutions, and (9) experiments were conducted at a constant pH using a
D

515 buffered solution (Smičiklas et al., 2009).

TE

516 Kumar (2006a) recommended that presenting a plot of qe versus Ce for the complete

517 adsorption isotherm in an adsorption study plays a key role in identifying the regions (e.g., Henry,
EP

518 Freundlich, Langmuir, and BET) in which the experimental data relating to adsorption equilibrium

519 are actually located. In contrast, plots of the Freundlich, Langmuir, or other adsorption isotherm
C

520 models do not help to identify these regions. Figures 8a and 8b provide typical complete
AC

521 adsorption isotherms that should be presented in the investigation of liquid-phase adsorption.

522 Kumar and Sivanesan (2006) also recommended that using equilibrium data covering the

523 complete isotherm was the best way to obtain the parameters in isotherm expressions; equilibrium

524 data with a partial isotherm was not sufficient, instead equilibrium data that covers the complete
26
 ACCEPTED MANUSCRIPT

525 isotherm was required.

526 Furthermore, an inconsistent definition of the adsorption isotherm plot in the literature was

527 recognized by Li and Pan (2007); the plot of qe against Ce/Wo (with Wo being a liquid/solid ratio) is

528 invisible with the traditional definition of the complete adsorption isotherm (the plot of qe versus

PT
529 Ce).

RI
530 Recently, Senthil Kumar et al. (2014) presented a plot of %removal versus C0 in an

531 adsorption equilibrium study instead of presenting the complete adsorption isotherm (Figure 8a).

SC
532 However, Hai (2017) made a critical note that it is clear that at every initial concentration of

533 methylene blue (MB; 50, 100, 150, 200, 250 mg/L), the removal percentage of MB by sulfuric

U
534 acid-treated orange peel (STOP) at 30 °C was always overwhelmingly higher than the

535
AN
corresponding values at 40, 50, and 60 °C. This result revealed that increasing the temperature

536 from 30 to 60 °C resulted in a decrease in the adsorption capacity (qe) of MB dye onto STOP
M
537 (Figure 8a). However, the maximum adsorption capacities of STOP from the Langmuir model
D

538 (Qomax) exhibited the following order: 60 °C (83.333 mg/g) > 50 °C (71.428 mg/g) > 40 °C (62.5
TE

539 mg/g) > 30 °C (50 mg/g). This logic problem might be caused by the adsorption process of MB

540 onto STOP not reaching equilibrium at the different temperatures (see adsorption isotherms in
EP

541 Figure 8a) owing to the low initial MB concentrations (50–250 mg/L) used at a high solid/liquid

542 ratio of 4.0 g/L (Hai, 2017). Therefore, in adsorption equilibrium studies, presenting plots of
C

543 %removal versus Co, %removal versus Ce, or qe versus Co should be avoided, and a plot of qe
AC

544 versus Ce for the complete adsorption isotherm must be presented.

545 Figure 8b provides a typical example of the comparison of adsorption performance under

546 the same operating conditions (i.e., temperature and solution pH). Clearly, “adsorbent B” exhibits

547 a higher affinity than “adsorbent A” at low initial adsorbate concentrations (Co) and low
27
 ACCEPTED MANUSCRIPT

548 equilibrium adsorbate concentrations (Ce), whereas the opposite order (“adsorbent A” >

549 “adsorbent B”) was found at higher Co and Ce values (Volesky, 2007).

550 It was recently identified that many researchers have made questionable conclusions when

551 studying the effects of agitation rate or shaking speed on the adsorption capacity (qe) or maximum

PT
552 adsorption capacity (Qomax) of an adsorbent toward an adsorbate. For example, Chu et al. (2005)

RI
553 concluded that the adsorption capacity of vitamin E onto silica varies as a function of the agitation

554 rate (120–180 rpm). The Qomax values exhibited the following order: 43.71 mg/g at 180 rpm >

SC
555 24.20 mg/g at 160 rpm > 24.20 mg/g at 140 rpm > 17.62 mg/g at 120 rpm. However, Choong and

556 Chuah (2005) made a critical remark that the agitation rate only impacts on the speed at which a

U
557 system reaches equilibrium, but not on the equilibrium itself. This means that when the adsorption

558
AN
process obtains a true equilibrium (sufficient contact time between adsorbent and adsorbate), the

559 Qomax values at different agitation rates will be insignificantly different. This is because the
M
560 decrease of the boundary layer resistance a higher agitation rate will facilitate the transport of the
D

561 adsorbate to the adsorbent.

TE

562 Figure 8

563 Although adsorption isotherms can insignificantly contribute to elucidating adsorption

EP

564 mechanisms, they are less helpful in this regard than data on adsorption kinetics and

565 thermodynamics (Volesky, 2003; Tien, 2007; Tien, 2008). However, collecting adsorption
C

566 isotherms is a useful strategy to both describe the relationship between the adsorbate concentration
AC

567 in solution (liquid phase) and the adsorbent (solid phase) at a constant temperature and design

568 adsorption systems (Tran et al., 2016).

569 A wide variety of models of adsorption isotherms have been applied in the literature. These

570 models can be classified as follows: (1) irreversible isotherms and one-parameter isotherms (i.e.,
28
 ACCEPTED MANUSCRIPT

571 Henry isotherm); (2) two-parameter isotherms (i.e., Langmuir, Freundlich, Dubinin–

572 Radushkevich, Temkin, Flory–Huggins, and Hill); (3) three-parameter isotherms (i.e., Redlich–

573 Peterson, Sips, Toth, Koble–Corrigan, Khan, Fritz–Schluender, Vieth–Sladek, and Radke–

574 Prausnitz); and (4) more than three-parameter isotherms (i.e., Weber–van Vliet, Fritz–Schlunder,

PT
575 and Baudu) (Kumar and Porkodi, 2006; Hamdaoui and Naffrechoux, 2007; Foo and Hameed,

RI
576 2010). Of these, the Langmuir and Freundlich models are the most commonly used, followed by

577 the Dubinin–Radushkevich and Redlich–Peterson models, because of the usefulness of their

SC
578 model parameters, their simplicity, and their easy interpretability.

579 6.2. Langmuir equation

U
580 6.2.1. Derivation of the Langmuir equation

581
AN
The theoretical Langmuir equation (Langmuir, 1918), which was originally applied to the

582 adsorption of gases on a solid surface, was developed using the following assumptions: (1) a fixed
M
583 number of accessible sites are available on the adsorbent surface and all active sites have the same
D

584 energy; (2) adsorption is reversible; (3) once an adsorbate occupies a site, no further adsorption
TE

585 can occur on that site; and (4) there is no interaction between adsorbate species. The nonlinear

586 form of the Langmuir model is described in Eq. 47 and its four linearized forms are shown in Eqs.
EP

587 48–51.

o
Qmax + K L Ce
qe =
C

(47)
1 + K LCe
AC

588 Hanes–Woolf linearization of the Langmuir model (Type 1):

Ce 1 1
= ( o )Ce + o (48)
qe Qmax Qmax K L

589 Lineweaver–Burk linearization of the Langmuir model (Type 2):

29
 ACCEPTED MANUSCRIPT

1 1 1 1
=( o ) + o (49)
qe Qmax K L Ce Qmax

590 Eadie–Hoffsiee linearization of the Langmuir model (Type 3):

− 1 qe
qe = ( ) + Qmax
o
(50)

PT
K L Ce

591 Scatchard linearization of the Langmuir model (Type 4):

RI
qe
= − K L qe + Qmax
o
KL (51)
Ce

SC
592 where Qomax (mg/g) is the maximum saturated monolayer adsorption capacity of an adsorbent, Ce

593 (mg/L) is the adsorbate concentration at equilibrium, qe (mg/g) is the amount of adsorbate uptake

U
594 at equilibrium, and KL (L/mg) is a constant related to the affinity between an adsorbent and

595
AN
adsorbate. For a good adsorbent, a high theoretical adsorption capacity Qomax and a steep initial
M
596 sorption isotherm slope (i.e., high KL) are generally desirable (Kratochvil and Volesky, 1998;

597 Wang, 2009).

D

598 The limitations of using the four linear forms of the Langmuir model have been
TE

599 highlighted by Bolster and Hornberger (2007). The transformation of data for linearization can

600 result in modifications of error structure, introduction of error into the independent variable, and
EP

601 alteration of the weight placed on each data point, which often leads to differences in the fitted

602 parameter values between linear and nonlinear versions of the Langmuir model. For the Type 1
C

603 linearization, because Ce and Ce/qe are not independent, the correlation between Ce and Ce/qe is
AC

604 overestimated, i.e., this equation may provide good fits to data that do not conform to the

605 Langmuir model. For the Type 2 linearization, transformation leads to clustering of data points

606 near the origin and is extremely sensitive to variability at low values of qe (high values of 1/qe).

607 For the Type 3 linearization, the abscissa is not error free, and qe/Ce and qe are not independent.
30
 ACCEPTED MANUSCRIPT

608 Thus, the correlation between qe/Ce and qe is underestimated, i.e., this equation may provide poor

609 fits to data that do conform to the Langmuir model. Finally, the variables qe and qe/Ce are also

610 not independent in the Type 4 linearization. In this case, the correlation between qe and qe/ Ce is

611 underestimated, i.e., this equation may provide a poor fit to data that do conform to the Langmuir

PT
612 model.

RI
613 6.2.2. Derivation of the separation factor

614 If the experimental data are adequately described by the Langmuir model, it is essential to

SC
615 calculate the separation factor. Hall et al. (1966) originally proposed that the essential

616 characteristics of the Langmuir isotherm model can be expressed in terms of a dimensionless

U
617 constant called the separation factor or equilibrium parameter RL, which is defined as follows:

RL =
1
AN
(52)
1 + K LCo
M
618 where RL is a constant separation factor (dimensionless) of the solid-liquid adsorption system, KL
D

619 is the Langmuir equilibrium constant, and Co (mg/L) is the initial adsorbate concentration.
TE

620 Notably, the isotherm shape can be predicted using the separation factor (RL) (Weber and

621 Chakravorti, 1974) and the Freundlich exponent n (Worch, 2012). Table 1 summarizes the various
EP

622 isotherm shapes.

623 Table 1
C

624 6.2.3. Mistakes concerning the Langmuir model and the separation factor
AC

625 Wan, et al. (2014) published a corrigendum because of a misapplication of the nonlinear

626 form of the Langmuir equation (Eq. 53). A similar mistake in presenting the nonlinear form of the

627 Langmuir equation (Eq. 54) was observed by Kumar and Porkodi, (2007b).

31
 ACCEPTED MANUSCRIPT

Ce C K
= oe + o L (53)
qe Qmax Qmax

Ce 1 1
= o
+ o (54)
qe K LQmax Qmax Ce

PT
628 Furthermore, an unexpected typographical error was made in reference (Egirani, 2004). As

629 a result, the sentence “a plot of Ce/qe versus Ce should indicate a straight line of slope 1/KLQomax

RI
630 and an intercept of 1/Qomax” was corrected to “a plot of Ce/qe versus Ce should indicate a straight

line of slope 1/Qomax and an intercept of 1/ KLQomax” by Ho (2004a).

SC
631

632 The incorrect expression and inaccurate calculation of RL were recognized and discussed

U
633 by Kumar (2006a) and Kumar (2006b). First, the separation factor RL is not a Langmuir constant,

634 AN
and is only employed for isotherms that obey the Langmuir model. Second, RL values will vary

635 with the initial adsorbate concentration, irrespective of the shape of the isotherm, and RL values at
M
636 different initial adsorbate concentrations cannot be calculated directly from the Langmuir

637 isotherm, but can be calculated from the Langmuir equilibrium constant KL. These
D

638 misunderstandings and mistakes are summarized in Table 2.

TE

639 Furthermore, a mistake related to the definition of RL compared with the original definition

640 was stated by Fu et al. (2007). The incorrect presentation is shown in Eq. 55.
EP

1
RL = (55)
1 + K LQmax
o
Co
C

641 Recently, Hai (2017) commented that “the authors’ conclusion about separation factor
AC

642 (RL) was not valid. It is noted that there is no information or value of RL mentioned in the paper’s

643 results and discussion. It is impossible for the authors to conclude that the values of RL were found

644 to be between 0 and 1, indicating the favorable adsorption of MB dye onto STOP.”

645 Clearly, the separation factor is related to the Langmuir equation in the adsorption
32
 ACCEPTED MANUSCRIPT

646 isotherm. Nevertheless, Goswami et al. (2016) recently described an incorrect relationship

647 between the separation factor and adsorption kinetic models (i.e., PFO, PSO, and intra-particle

648 diffusion) at different initial cadmium concentrations. This is a fundamental misunderstanding that

649 should be avoided.

PT
650 Table 2

RI
651 6.3. Freundlich equation

652 6.3.1. Derivation of the Freundlich equation

SC
653 The Freundlich equation is one of the earliest empirical equations used to describe

654 equilibrium data and adsorption characteristics for a heterogeneous surface (Freundlich, 1906).

U
655 Unlike the Langmuir equation, the Freundlich isotherm can describe neither the (arithmetic)

656
AN
linearity range at very low concentrations nor the saturation effect at very high concentrations.

657 Hence, the Freundlich isotherm does not describe the saturation behavior of an adsorbent (Wang,
M
658 2009). The nonlinear and linear forms of the Freundlich equation can be expressed as shown in
D

659 Eqs. 56 and 57, respectively.

TE

qe = K F Cen (56)

log qe = n log Ce + log K F (57)

EP

660 where qe (mg/g) is the amount of adsorbate uptake at equilibrium, Ce (mg/L) is the adsorbate
C

661 concentration at equilibrium, KF (mg/g)/(mg/L)n is the Freundlich constant, and n (dimensionless)

AC

662 is the Freundlich intensity parameter, which indicates the magnitude of the adsorption driving

663 force or the surface heterogeneity.

33
 ACCEPTED MANUSCRIPT

664 6.3.2. Mistakes concerning the Freundlich equation

665 Several investigators have identified an incorrect expression (Eq. 58) of the linear form of

666 the Freundlich equation (Kumar and Porkodi, 2007b; Alagumuthu et al., 2010).

PT
log qe = + log K F (58)
n log Ce

RI
667 Furthermore, several discussions that are inconsistent with the original concepts of the

668 Freundlich theory have been identified in the literature (Lu, 2008; Hai, 2017). Hai (2017)

SC
669 commented that the exponent n of the Freundlich equation must be dimensionless. Therefore, the

670 original authors incorrectly reported the units of the exponent n (g/L); and there is no validation for

U
671 the statement of the significance of the n value for the adsorption of MB dye onto STOP is as

672
AN
follows: when n = 1 the adsorption is linear, when n < 1 the adsorption is a chemical process, and

673 when n > 1 the adsorption is a physical process. According to the Freundlich theory, the adsorption
M
674 isotherm becomes linear when n = 1, favorable when n < 1, and unfavorable when n > 1. Similarly,
D

675 Harsha et al. (2015) also made a basic mistake when they reported the units of the exponent n as
TE

676 g/L.

677 Lu (2008) highlighted that in many cases KF is not the maximum adsorption capacity and
EP

678 the KF value is only equal to the maximum adsorption capacity when n approaches infinity.

679 Although KF is not defined as the maximum adsorption capacity (Qomax), KF values and Qomax
C

680 values should be of the same order. Such a mistake has been discussed by Tran (2017), where the
AC

681 reported KF values followed the order: 4.471 at 303 K > 3.439 at 313 and 323 K > 2.877 at 333 K,

682 while the Qomax values were 30.12 mg/g at 313 K > 21.367 mg/g at 323 K > 21.276 mg/g at 333 K

683 > 20.492 mg/g at 303 K. This mistake might be attributable to a miscalculation.

684 Notably, Kumar (2006a) highlighted that values for the exponent n (Eq. 59) in the range of

34
 ACCEPTED MANUSCRIPT

685 0–10 suggest favorable adsorption. If the experimental equilibrium data do not lie in the Henry

686 region, then the linear regression method will just check the hypothesis instead of verifying the

687 theory behind the Freundlich model; therefore, it is not possible in practice to obtain n > 10

688 (Kumar, 2006a). A corresponding error was also found in a published paper (Huang et al., 2014),

PT
689 in which the n values reported for the removal of aniline and Cr(VI)/aniline + Cr(VI) by an

RI
690 activated carbon/chitosan composite were 10.20 and 11.34, respectively.

qe = K F Ce1 / n (59)

SC
691 6.4. Redlich–Peterson equation

U
692 6.4.1. Derivation of the Redlich–Peterson equation

693

694
AN
A three-parameter Redlich–Peterson isotherm was proposed upon considering the

limitations of the Freundlich and Langmuir isotherms (Redlich and Peterson, 1959). This model
M
695 incorporates the features of the Freundlich and Langmuir models and might be applicable for

696 demonstrating adsorption equilibrium over a wide range of adsorbate concentrations. The
D

697 nonlinear form of this empirical model is given as follows:

TE

K RPCe
qe = (60)
1 + a RP Ceg
EP

698 where KRP (L/g) and aRP (mg/L)−g are the Redlich–Peterson constants and g (dimensionless) is an

699 exponent whose value must lie between 0 and 1.

C

700 Eq. 60 becomes a linear isotherm (Henry’s law equation) at low surface coverage (g = 0),
AC

701 reduces to the Langmuir isotherm when g = 1, and transforms into the Freundlich isotherm when

702 KRP and aKP >> 1 and g = 1. Therefore, if the g value is outside the range of 0–1, the data is not

703 adequately explained by the Redlich–Person equation. Accurate calculation of g can be helpful in

704 expatiation of where the isotherm presentations, neither in the Freundlich or Langmuir (Kumar
35
 ACCEPTED MANUSCRIPT

705 and Sivanesan, 2006).

706 6.4.2 Mistakes concerning the Redlich–Peterson equation

707 A series of comments related to mistakes in calculating the exponent g have been reported

708 elsewhere (Inbaraj, 2006; Kumar and Porkodi, 2007a; Kumar et al., 2007; Kumar et al., 2007;

PT
709 Kumar and Porkodi, 2008). These mistakes involved values of the exponent g that were outside the

RI
710 range of 0–1 (Table 3). It is not easy to predict the parameters of the Redlich–Peterson equation

711 because there are three unknown parameters. Although a new linear form of the Redlich–Peterson

SC
712 equation was published by Wu et al. (2010), the nonlinear form might be more appropriate for

713 accurately calculating the parameters of adsorption models with more than two unknown

U
714 parameters, such as the Redlich–Peterson model.

715
ANTable 3

716 6.5. Dubinin–Radushkevich equation

M
717 6.5.1. Derivation of the Dubinin–Radushkevich equation
D

718 The Dubinin–Radushkevich equation was developed to account for the effect of the porous
TE

719 structure of an adsorbent (Dubinin and Radushkevich, 1947), and is expressed as follows:

qe = q DR e − K RDε
2
(61)
EP

The linear form of the Dubinin–Radushkevich equation is:

C

2
ln q e = − K DR ε + ln q DR (62)
AC

1
ε = RT ln(1 + ) (63)
Ce

By inserting Eq. 63 into Eq. 61, Eq. 64 can be obtained:

36
 ACCEPTED MANUSCRIPT

2 2 2 1
ln q e = − K DR R T ln (1 + ) + ln q DR (64)
Ce

720 The parameters qDR and KDR in Eq. 64 can be obtained as follows: (1) a plot of lnqe against

721 ln2(1 + 1/Ce) has a slope = −KDRR2T2 and an intercept = lnqRD, and the E value can be obtained

PT
722 using Eq. 65, and (2) a plot of lnqe against R2T2ln2(1 + 1/Ce) has a slope = −KDR and an intercept =

723 lnqRD, and the E value can be obtained using Eq. 66. Notably, the E values obtained from Eqs. 65

RI
724 and 66 are the same (Tran et al., 2016).

SC
1 RT
E= = (65)
2 K DR − 2 slope

U
1 1
E= = (66)
2 K DR − 2slope

725
AN
where qRD (mg/g) is the adsorption capacity, KRD (mol2/kJ2) is a constant related to the sorption
M
726 energy, ɛ is the Polanyi potential, E (kJ/mol) is the mean adsorption energy, R is the gas constant, T

727 is the temperature in Kelvin, and qe and Ce are obtained from Eq. 1.
D

728 6.5.2. Mistakes concerning the Dubinin–Radushkevich equation

TE

729 Some authors have highlighted erroneous expressions of the Dubinin–Radushkevich

730 equation (Cavas, 2008; Fu et al., 2008). These erroneous forms are shown in Eqs. 67–70.
EP

731 Incorrect nonlinear form of the Dubinin–Radushkevich equation:

C

2
q e = q DR exp( − K DR ( RT (ln 1 + 1 / C e ) ) (67)
AC

732 Incorrect linear form of the Dubinin–Radushkevich equation:

1
ln q e = −2 K DR RT ln(1 + ) + ln q DR (68)
Ce

733 Incorrect expressions for the Polanyi potential (ɛ):

37
 ACCEPTED MANUSCRIPT

ε = RTe1/ C e (69)

1
ε = RT ln( ) (70)
Ce

PT
734 Additionally, the magnitude of E may give useful information about the type of adsorption

735 process (physical or chemical) (Fu et al., 2008; Tran et al., 2016). Fu et al. (2008) commented that

RI
736 the E values for the removal of phenolic compounds by organophilic bentonite can be categorized

SC
737 as corresponding to ion exchange (E = 8–16 kJ/mol), whereas the original authors made an

738 inconsistent conclusion that the adsorption of two phenolic compounds could be regarded as

U
739 physical adsorption.

740 AN
Recently, Tran (2017) made the following comment on the conclusions of the original

741 paper: “the obtained value of E in this work varies from 32.596 kJ/mol to 40.572 kJ/mol for studied
M
742 temperature which is higher than the adsorption energy values previously enumerated.” However,

743 these adsorption energy (E) values might not have been calculated correctly. Assuming the β
D

744 values at 303 K (β = 1.93 × 10−8 mol2/kJ2), 313 K (2.99 × 10−8), 323 K (2.00 × 10−8), and 333 K
TE

745 (2.35 × 10−8) reported by these authors are correct, the recalculated adsorption energy (E) values
EP

746 are 5090 kJ/mol at 303 K, 4089 kJ/mol at 313 K, 5000 kJ/mol at 323 K, and 4613 kJ/mol at 333 K.

747 Thus, the E values in this study ranged from 4089 to 5091 kJ/mol, which is impossible for a
C

748 heavy-metal biosorption process.

AC

749 7. Adsorption thermodynamics

750 7.1. Principles of adsorption thermodynamics

751 Thermodynamic studies are an indispensable component of predicting adsorption

752 mechanisms (e.g., physical and chemical). The key distinctions between physical and chemical

38
 ACCEPTED MANUSCRIPT

753 adsorption were summarized in detail in a recent report (Tran et al., 2016). The thermodynamic

754 parameters can be computed according to the laws of thermodynamics using the following

755 equations:

∆G o = − RT ln K C (71)

PT
756 The relationship between ∆G° and ∆H° and ∆S° is described as follows:

RI
∆Go = ∆H 0 − T∆S 0 (72)

757 The well-known van’t Hoff equation is obtained by substituting Eq. 71 into Eq. 72

SC
− ∆H o 1 ∆S o (73)
ln KC = x +
R T R

U
758 where R is the universal gas constant (8.3144 J/(mol × K)) and T is the absolute temperature in

759 Kelvin.
AN
760 The Gibbs energy change (∆G°) is directly calculated from Eq. 71, whereas the enthalpy
M
761 change (∆H°) and entropy change (∆S°) are determined from the slope and intercept, respectively,
D

762 of a plot of lnKC versus 1/T (Eq. 73). The common units of ∆G°, the gas constant, and temperature
TE

763 are J/mol, J/(mol × K), and K, respectively; therefore, the equilibrium constant KC in Eq. 73 must

764 be dimensionless (Milonjić, 2007; Milonjić, 2009b; Canzano et al., 2012; Dawood and Sen, 2012;
EP

765 Zhou et al., 2012; Zhou and Zhou, 2014; Tran et al., 2016; Ghosal and Gupta, 2017; Hai, 2017;

766 Rahmani-Sani et al., 2017; Zhou, 2017).

C

767 Eq. 72 describes the relationship of ∆G° with ∆H° and ∆S°, and it is useful for checking the
AC

768 logic of calculated values for these thermodynamic parameters. Unfortunately, associated

769 mistakes have been found in the literature. Fu et al. (2008) substituted the values of ∆H° and ∆S°

770 into Eq. 72 to obtain ∆G°, but the obtained ∆G° values were significantly different from the ∆G°

771 values obtained using Eq. 71. Fu et al. (2008) highlighted that this mistake might result from

39
 ACCEPTED MANUSCRIPT

772 incorrect application of the method for calculating the equilibrium constant.

773 Clearly, accurate estimation of thermodynamic parameters is directly dependent on

774 accurate determination of the equilibrium constant between two phases (KC; dimensionless). In the

775 literature, the thermodynamic parameters can be calculated from KC values derived from

PT
776 adsorption-isotherm constants (i.e., Langmuir, Freundlich, Frumkin, Flory–Huggins, and Henry)

RI
777 or the partition coefficient (Liu, 2009; Doke and Khan, 2013; Tran et al., 2016). However, to

778 obtain an appropriate calculation method, several factors need to be considered thoroughly. First,

SC
779 the equilibrium constant (KC) must be dimensionless. Second, the linear regression coefficient (R2)

780 of the van’t Hoff equation must be high. Third, the adsorbate in solution must have a high or low

U
781 concentration. In addition, the temperatures used for calculating the thermodynamic parameters

782
AN
must have units of Kelvin (K), not degrees Celsius (°C). The incorrect use of temperature units led

783 to the publication of an erratum by Ho (2004a), with accurate recalculations.

M
784 As the equilibrium constant KC can be derived using a number of approaches, considerable
D

785 variation in the thermodynamic parameters can be obtained; therefore, the most appropriate
TE

786 approach should be determined. Here, we thoroughly discuss the derivation of the equilibrium

787 constant KC from adsorption isotherms (i.e., Langmuir, Freundlich, and Henry constants) and the
EP

788 partition coefficient because of their popularity in the literature.

789 7.2. Equilibrium constant derived from the Langmuir constant (KL)
C

790 The Langmuir equation was initially derived from a kinetic study and then subsequently
AC

791 from a thermodynamic study. The derivation of the Langmuir equation from the thermodynamic

792 perspective can be described as follows (Crittenden et al., 2012; Tran et al., 2016). The

793 relationship between vacant surface sites on the surface of an adsorbent (Sv; mmol/m2), adsorbate

794 species in solution (A; mmol), and adsorbate species bound to surface sites (SA; mmol/m2) can be
40
 ACCEPTED MANUSCRIPT

795 described by the following reaction:

SV + A ↔ SA (74)

796 On the basis of the Langmuir expression, it is assumed that the reaction has a constant

PT
797 Gibbs energy change (∆G°; J/mol) for all sites, so the thermodynamic equilibrium constant (KC;

798 dimensionless) can be expressed as:

RI
− ∆G o

∆G = − RT ln K C
o
⇔ KC = e RT (75)

SC
799 Another Langmuir assumption is that each site is capable of binding at most one adsorbate

800 molecule (monolayer). According to the equilibrium condition, the thermodynamic equilibrium

U
801 constant may be written as:

SA
− ∆G o
AN
KC = =e RT (76)
SV [ A]
M
802 where [A] is the concentration of adsorbate A in solution at equilibrium (mg/L), R is the universal
D

803 gas constant (8.314 J/mol × K), and T is absolute temperature K (273 + °C).
TE

804 The critical problem with Eq. 76 is that there are two unknown parameters (i.e., SV and

805 [A]). However, this problem can be solved if the total sites available or monolayer coverage (ST;
EP

806 mol/m2) are fixed:

SA K C [ A]ST
ST = SV + SA = + SA ⇔ SA =
C

(77)
KC [ A] 1 + K C [ A]
AC

807 The expression of SA in units of mmol/m2 is not useful in mass balance, and units of mass

808 of adsorbate adsorbed per unit mass of adsorbent (qe; mg/g) are much more useful. The mass

809 loading (qe) can be obtained in units of mg/g (Eq. 1) by multiplying both sides of Eq. 71 by the

810 surface area (SBET; m2/g) of the adsorbent and the molecular weight of the adsorbate (Mw; g/mol).

41
 ACCEPTED MANUSCRIPT

811 Thus, Eq. 77 becomes:

o
K [ A]ST S BET M w Qmax K L Ce
SA.S BET .M w = C ⇔ qe = (78)
1 + K C [ A] 1 + K L Ce

812 where qe (mg/g) = SA × SBET × Mw is described by Eq. 1; Ce (mg/L) = [A] is the concentration of

PT
813 adsorbate in solution at equilibrium; Qomax (mg/g) = ST × SBET × Mw is the maximum monolayer

RI
814 adsorptive capacity of the adsorbent when the surface sites are saturated with adsorbate; the

815 relationship between KL and KC is described in the following discussion.

SC
816 However, the main problem is that the Langmuir constant KL is dimensional with common

817 units of L/mmol or L/mg, while the equilibrium constant KC is dimensionless (without units).

U
818 Thus, the direct application of KL (L/mmol or L/mg) in the calculation of thermodynamic

819
AN
parameters produces incorrect results, as discussed by many authors (Milonjić, 2007; Milonjić,

820 2009b; Canzano et al., 2012; Dawood and Sen, 2012; Zhou et al., 2012; Zhou and Zhou, 2014;
M
821 Anastopoulos and Kyzas, 2016; Tran et al., 2016; Ghosal and Gupta, 2017; Hai, 2017,
D

822 Rahmani-Sani et al., 2017). To solve this unit problem, several methods have been recommended
TE

823 (Milonjić, 2007; Zhou and Zhou, 2014; Tran et al., 2016). Depending on the units of KL, the

824 equilibrium constant KC can easily be obtained as a dimensionless constant.

EP

825 When an adsorption study is conducted in aqueous solution and KL has units of L/mmol, KC

826 can be easily obtained as a dimensionless parameter by multiplying KL by 55.5 and then by 1,000
C

827 (Eq. 79). This method was originally proposed by Milonjić (2007) and then developed by Zhou
AC

828 and Zhou (2014) and Tran et al. (2016). The values of the parameters ∆G°, ∆H°, and ∆S° can be

829 calculated using the following equations:

KC = 55.5×1,000× KL (79)

42
 ACCEPTED MANUSCRIPT

∆Go = −RT ln(55.5×1,000× KL ) (80)

− ∆H o 1 ∆S o
ln(55 .5 × 1,000 × K L ) = × + (81)
R T R

830 where the factor 55.5 is the number of moles of pure water per liter (1,000 g/L divided by 18

PT
831 g/mol) and the term 55.5 × 1,000 × KL is dimensionless.

RI
832 In the case of KL with units expressed in L/mg, Milonjić (2007 and 2009b) stated that KL

833 could be obtained as a dimensionless parameter by multiplying KL by 106 (Eq. 82). However, Zhou

SC
834 and Zhou (2014) recommended that KC could be obtained as a dimensionless parameter by

835 multiplying KL by the molecular weight of the adsorbate (Mw; g/mol), by 1000, and then by 55.5

U
836 (Eq. 85).

KC = 106 K L
AN (82)
M
∆G o = − RT ln(106 K L ) (83)

− ∆H o 1 ∆S o
D

ln(10 K L ) =
6
x + (84)
R T R
TE

837 where the factor 106 is the solution density (assuming the density of pure water is 1.0 g/mL) and

838 the term 106 × K is dimensionless.

EP

K C = M w × 55.5 ×1,000 × K L (85)

C

∆Go = −RT ln(Mw × 55.5×1,000× KL ) (86)

AC

− ∆ H o 1 ∆S o
ln( M w × 55.5 ×1,000 × K L ) = × + (87)
R T R

839 where the factor 55.5 is the number of moles of pure water per liter and the term Mw × 55.5 × 1,000

840 × KL is dimensionless.

43
 ACCEPTED MANUSCRIPT

841 In a recent paper, Tran et al. (2016) applied Eqs. 79–87 to compare thermodynamic

842 parameters calculated from the Langmuir constants derived from four linear forms of the

843 Langmuir model. They concluded that (1) both the recommendations of Milonjić (2007 and

844 2009b) and Zhou and Zhou (2014) provide the same methods for calculating the thermodynamic

PT
845 parameters, and (2) there is some deviation between these methods for the results from Types 1, 3,

RI
846 and 4; only Type 2 produces identical results.

847 7.3. Equilibrium constant derived from the Freundlich constant (KF)

SC
848 The Freundlich equation is consistent with the thermodynamics of heterogeneous

849 adsorption (Crittenden, et al., 2012). The Freundlich constant KF can be obtained as a

U
850 dimensionless value using Eq. 88 (Ghosal and Gupta, 2015; Tran et al., 2016).

K ρ 106 (1− )
KC = F ( ) n
1
AN
(88)
1000 ρ
M
K ρ 106 (1− )
1
∆G o = − RT ln( F ( ) n ) (89)
1000 ρ
D

K F ρ 10 6 (1− n ) − ∆ H o 1 ∆S o
1
)= x +
TE

ln( ( ) (90)
1000 ρ R T R

851 where ρ is the density of pure water (assumed as ~1.0 g/mL).

EP

852 It is noted that the units of KF rely on the units used for the liquid-phase concentration (C)

853 and solid-phase concentration (q). Units of mg/L or mmol/L for C and mg/g or mmol/g for q are
C

854 used most frequently to demonstrate adsorption from water solutions. The differing units of KF can
AC

855 be reciprocally converted using Eq. 91 (Worch, 2012).

K F = K F' ( M w )1−1 / n (91)

856 where the units of the Freundlich constant KF are (mg/g)/(mg/L)1/n, with Co and Ce (mg/L) and qe

44
 ACCEPTED MANUSCRIPT

857 (mg/g); the units of KF′ are (mmol/g)/(mmol/L)1/n, with Co and Ce (mmol/L) and qe (mmol/g); and

858 Mw is the molecular weight of the adsorbate. The exponent 1/n is not affected by the units of the

859 liquid-phase concentration (C) and solid-phase concentration (q)

860 Based on a comparison of the thermodynamic parameters calculated from the Freundlich

PT
861 constant (dimensionless) and Langmuir constant (dimensionless), Tran et al., (2016) concluded

RI
862 that the signs and magnitudes of ∆G°, ∆H°, and ∆S° calculated from KF were consistent with those

863 calculated from KL. Of course, the experimental data in the adsorption isotherms must be fitted

SC
864 well by the Langmuir and Freundlich models, and the R2 value of the van’t Hoff equation must be

865 higher than 0.90 (Tran et al., 2016). It is also evident that, to some extent, the application of KF to

U
866 estimate the thermodynamic parameters should be approached with caution.

867
AN
7.4. Equilibrium constant derived from the partition coefficient (Kp)

868 Changes in the thermodynamic partition coefficient with changes in temperature were first
M
869 reported by Biggar and Cheung (1973). The equilibrium constant can be defined as follows:
D

as γ s C s
Kp = = (92)
ae γ eCe
TE

870 where as is the activity of the adsorbate adsorbed onto the adsorbent, ae is the activity of the
EP

871 adsorbate in solution at equilibrium, γs is the activity coefficient of the adsorbate adsorbed onto the

872 adsorbent, γe is the activity coefficient of the adsorbate in solution at equilibrium, Cs is the
C

873 concentration of adsorbate adsorbed onto the adsorbent at equilibrium (mg/L), and Ce is the
AC

874 concentration of adsorbate in solution at equilibrium (mg/L). Cs is defined by the mass balance of

875 adsorbate that disappears from the solution, which should appear on the adsorbent.

876 When the concentration of adsorbate in the solution approaches zero, which results in Cs →

877 0 and Ce → 0, the activity of coefficient γ approaches unity, and Eq. 92 can be written as:

45
 ACCEPTED MANUSCRIPT

Cs as
Lim C = = Kp (93)
Cs →0 e ae

878 Kp values can be obtained plotting ln(Cs/Ce) versus Cs and extrapolating Cs to zero. If a

879 straight line fits the data with a high regression coefficient (R2) and its intersection with the vertical

PT
880 axis provides the value of Kp, the partition coefficient will be in unison with the equilibrium

881 constant. The ∆G° value can be directly calculated using Eq. 71, while the values of ∆H° and ∆S°

RI
882 are determined from the slope and intercept, respectively, of Eq. 73.

SC
883 This method has been widely applied in the literature. A summary of the applications of

884 this method can found in the critical review in reference (Doke and Khan, 2013). However,

U
885 incorrect applications of Biggar and Cheung’s definition have also been identified in the literature.

886 AN
In a typical experiment, Senthil Kumar et al. (2014) made a mistake in the application of Kp during

887 calculation of the thermodynamic parameters for the adsorption of MB onto STOP (Table 4). Hai
M
888 (2017) highlighted that it is impossible to obtain an equilibrium constant, Gibbs energy change

889 (∆G°), enthalpy change (∆H°), and entropy change (∆S°) for the adsorption process at every initial
D

890 MB concentration, as shown in Table 4. Assuming that the experimental data reported by Senthil
TE

891 Kuma et al. (2014) are correct, Hai (2017) recalculated the thermodynamic parameters based on
EP

892 the method of Biggar and Cheung (1973), and the corrected thermodynamic parameters are also

893 listed in Table 4. Likewise, Lima et al. (2015) also determined the value of the adsorption
C

894 equilibrium constant (KC) from the value of the best fit nonlinear isotherm equilibrium model
AC

895 instead of using the equilibrium constant (Kp) calculated from the initial and equilibrium

896 concentrations of the adsorbate.

897 Table 4

898 Importantly, according to the United States Environmental Protection Agency (USEPA,

46
 ACCEPTED MANUSCRIPT

899 1999), the partition coefficient is only appropriate for calculating the thermodynamic parameters

900 if the initial adsorbate concentration is low. Analogous conclusions can be found in the literature

901 (Liu, 2009; Salvestrini et al., 2014; Lima et al., 2015; Tran et al., 2016; Hai, 2017). It has been

902 highlighted that the partition coefficient is only equal to the thermodynamic equilibrium constant

PT
903 when the adsorbed concentrations in the solutions are very low. In this situation, the partition

RI
904 coefficient can be used to calculate the thermodynamic parameters. Recently, Tran et al. (2016),

905 who compared the thermodynamic parameters calculated from the Langmuir constant and the

SC
906 partition coefficient, concluded that although the signs of the calculated ∆G°, ∆H°, and ∆S°

907 parameters were the same, the partition coefficient might not be appropriate for calculating the

U
908 thermodynamic parameters owing to the low regression coefficient of the van’t Hoff equation (R2

909 = 0.53).
AN
910 7.5. Equilibrium constant derived from the distribution coefficient (Kd)
M
911 In 1987, Khan and Singh (1987) proposed a method for calculating thermodynamic
D

912 parameters (Eq. 94) based on the method of Biggar and Cheung (1973). The distribution
TE

913 coefficient (Kd) can be defined as follows:

qe
Kd = (94)
EP

Ce

914 In this method, Kd values are obtained by plotting ln(qe/Ce) against Ce and extrapolating to
C

915 zero Ce. If a straight line fits the data with a high regression coefficient (R2), then its intersection
AC

916 with the vertical axis provides the value of Kd. Fundamentally, Eq. 94 can be derived from the

917 Freundlich and Langmuir equations. If n = 1, the Freundlich equation (Eq. 56) will become

918 Henry’s equation for a linearized isotherm (Eq. 95), and the Freundlich constant (KF) will be equal

919 to the Henry constant (KH). For very low concentrations of adsorbate (KL and Ce << 1), the

47
 ACCEPTED MANUSCRIPT

920 Langmuir equation (Eq. 47) will become Henry’s equation (Eq. 95), and the Henry constant (KH)

921 will become QomaxKL. The Henry constant (KH) is also referred to as the distribution coefficient

922 (Kd).

qe = K H Ce (95)

PT
923 This method has been widely applied to calculating thermodynamic parameters in the

RI
924 literature. Unfortunately, the units of Kd in Eq. 94 are L/g as Ce has units of mg/L and qe has units

925 of mg/g. Therefore, it is impossible to directly employ Kd as the equilibrium constant for the

SC
926 calculation of ∆G°, ∆H°, and ∆S°. This problem associated with the units has been discussed by

U
927 several scientists (Patrickios and Yamasaki, 1997; Milonjić, 2007; Canzano et al., 2012; Dawood

928

929
AN
and Sen, 2012; Tran et al., 2016; Zhou, 2017). Canzano, Iovino et al. (2012) suggested that the

partition coefficient Kd (L/g) could be converted to KC (dimensionless) by multiplying Kd by a

M
930 factor of 1,000, as originally proposed by Milonjić (2007 and 2009b).

931 Similar to the partition coefficient, the distribution coefficient may be appropriate for
D

932 estimating the thermodynamic parameters when the initial concentration of the adsorbate is low
TE

933 (Salvestrini et al., 2014; Tran et al., 2016). Tran et al. (2016) concluded that the signs of ∆G°, ∆H°,

934 and ∆S° estimated from the distribution coefficient and the Langmuir constant are the same.
EP

935 However, the units of Kd (L/g) are different from those of KL (L/mg), and the adsorption

936 equilibrium data do not fit Henry’s adsorption isotherm. As a result, application of the distribution
C
AC

937 coefficient was not appropriate for calculating the thermodynamic parameters in this study; the

938 low regression coefficient of the van’t Hoff equation (R2 = 0.42) is further evidence of the

939 inappropriateness of this approach.

940 In summary, to determine accurate thermodynamic parameters, several points should be

941 thoroughly considered. First, the equilibrium constant (KC) must be dimensionless. Second, if the
48
 ACCEPTED MANUSCRIPT

942 concentrations of adsorbate used to obtain the adsorption isotherm are very low (diluted solutions),

943 the partition (Kp) or distribution coefficients (Kd) are appropriate for calculation of the

944 thermodynamic parameters; otherwise, the Langmuir (KL) or Freundlich (KF) constant models may

945 be more appropriate. Third, the plot of lnKC against 1/T corresponding to the van’t Hoff equation

PT
946 must be linear with a high regression coefficient (R2). Additionally, it is necessary to consider

RI
947 whether the adsorption process can reach equilibrium and consideration of the adsorption isotherm

948 shapes and the adsorption model fit are also recommended. Thus, presentation of the complete

SC
949 adsorption isotherm (plot of qe versus Ce) is strongly recommended. Finally, the determined

950 thermodynamic parameters must have a logical relationship with the experimental data.

U
951 Notably, numerous recent studies have reported methods of calculating thermodynamic

952
AN
parameters from the partition coefficient (Kp) or distribution coefficient (Kd) that differed from the

953 traditional concepts defined in Sections 7.4 and 7.5, respectively. A typical mistake is present in
M
954 the work of Nekouei et al. (2016). They studied the adsorption thermodynamics at only one initial
D

955 adsorbate concentration and different temperatures instead of using a range of adsorbate
TE

956 concentrations (i.e., from 100 mg/L to 1,000 mg/L) and different temperatures (i.e., 10 °C, 20 °C,

957 30 °C, and 50 °C). This mistake led to invalid conclusions regarding the adsorption
EP

958 thermodynamics.

959 8. Others mistakes

C

960 8.1. Inconsistent data points in experimental data and model fitting
AC

961 Significant differences between the number of data points in an experiment and those used

962 for model fitting should be avoided. For example, Azizian (2008a) made the following critical

963 comments about the paper of Karadag et al. (2007): there were six experimental data points

964 regarding the adsorption kinetics (plots of qt versus t), but only five experimental data points were
49
 ACCEPTED MANUSCRIPT

965 used for fitting of the experimental data to the PSO model (plot of t/qt versus t). This omission

966 resulted in inaccurate calculations and questionable conclusions. The R2 values of the PSO model

967 were very high (0.992–1.000), but the difference between the experimental (qe,exp) and calculated

968 (qe,cal) values was more than 100% (for example qe,exp = 6.720 mg/g and qe,cal = 15.526 mg/g).

PT
969 Clearly, the qe values calculated from the experimental data and the model were significantly

RI
970 different, but the authors (Karadag et al., 2007) concluded that “The pseudo-second-order kinetic

971 model agrees very well with the dynamic behavior for the adsorption of dyes RR239 and RB5 onto

SC
972 CTAB-zeolite under several different dye concentrations, temperatures, and pH values”; the

973 validity of this conclusion is very questionable.

U
974 Lima et al. (2015) observed a change in kinetic parameters when five experimental points

975
AN
were deleted to achieve a good R2 value. They concluded that the strategy of deleting some data

976 points to improve the R2 value caused serious errors in the values of qe and k1 in the PFO equation
M
977 or k2 in the PSO equation.
D

978 Azizian (2008) pointed out that in reference (Karadag et al., 2007), a comparison was made
TE

979 between the results of fitting four experimental equilibrium data points to the Langmuir model and

980 fitting five experimental equilibrium data points to the Freundlich model. Therefore, it is
EP

981 impossible for the authors (Karadag et al., 2007) to conclude that “the Freundlich model exhibited

982 a slightly better fit than the Langmuir model.” To compare the fits of two different models, the
C

983 same experimental data (including the same number of experimental data points) should be used
AC

984 (Azizian, 2008a). Likewise, Harsha et al. (2015) examined adsorption isotherms using seven

985 experimental data points. However, in the plots for calculating the Langmuir, Freundlich, and

986 Dubinin–Radushkevich models, only three experimental points were used. It is impossible to

50
 ACCEPTED MANUSCRIPT

987 obtain the Langmuir, Freundlich, and Dubinin–Radushkevich parameters from only three

988 experimental points. A similar mistake was found in reference (Goswami et al., 2016).

989 8.2. Oxidation state of chromium

990 Many investigators (Aoyama, 2003; Park et al., 2006a; Park et al., 2006b; Park et al.,

PT
991 2006c) have commented that mistakes in analyzing chromium species in aqueous solution have

RI
992 resulted in the incorrect elucidation of hexavalent chromium adsorption, attributing Cr(VI)

993 removal from an aqueous solution to “anionic adsorption”. Generally, when Cr(VI) comes into

SC
994 contact with organic substances or reducing agents, especially in an acidic medium, it is easily or

995 spontaneously reduced to Cr(III), as Cr(VI) has a high redox potential (above +1.3 V under

U
996 standard conditions). Therefore, there is a strong possibility that the mechanism of Cr(VI) removal

997
AN
by biomaterials or biomaterial-based activated carbon is not “anionic adsorption” but

998 “adsorption-coupled reduction”. Various authors (Aoyama, 2003; Park et al., 2006a; Park et al.,
M
999 2006b; Park et al., 2006c) also suggested that in Cr(VI) adsorption studies, it is necessary to
D

1000 analyze both the Cr(VI) and total Cr in aqueous solution, as well as determining the oxidation state
TE

1001 of chromium bound to the biomaterial or activated carbon.

1002 The authors (Aoyama, 2003; Park et al., 2006a; Park et al., 2006b; Park et al., 2006c)
EP

1003 pointed out that in most papers, only Cr(VI) in aqueous solutions was analyzed by the standard

1004 colorimetrical method; the pink colored complex formed from between 1,5-diphenycarbazide and
C

1005 Cr(VI) in an acidic solution can be spectrophotometrically analyzed at 540 nm. Similarly, the total
AC

1006 Cr in aqueous solutions has been analyzed using atomic absorption spectroscopy (AAS) or

1007 inductively coupled plasma-atomic emission spectroscopy (ICP-AES), which cannot distinguish

1008 Cr(VI) from the total Cr. Additionally, Aoyama (2003) underlined that the total Cr present in a

1009 solution can be determined by oxidizing any Cr(III) formed with KMnO4, followed by the same
51
 ACCEPTED MANUSCRIPT

1010 procedures described for the determination of Cr(VI). The adsorption density can be calculated

1011 from the difference between the initial concentration of Cr(VI) and the final concentration of total

1012 Cr. In contrast, the Cr(III) concentration is calculated from the difference between the final total Cr

1013 and final Cr(VI) concentrations.

PT
1014 Several techniques have been applied in the literature to ascertain the oxidation state of

RI
1015 chromium on an adsorbent, such as X-ray photoelectron spectroscopy (XPS) and X-ray absorption

1016 spectroscopy (XAS). For example, Dambies et al. (2001) investigated the characterization of

SC
1017 Cr(VI) interactions with chitosan using XPS. On the basis of their XPS data, they concluded that

1018 Cr(VI) was entirely reduced to Cr(III) on glutaraldehyde crosslinked beads, while only 60% of

U
1019 Cr(VI) was found in its reduced form on raw beads. Additionally, Cr(VI) binding and reduction to

1020
AN
Cr(III) by agricultural byproducts of Avena monida (oat) biomass was characterized by

1021 Gardea-Torresdey et al. (2000). Their XAS studies further corroborated that, although Cr(VI)
M
1022 could bind to oat biomass, it was easily reduced to Cr(III) by positively charged functional groups,
D

1023 and Cr(III) was subsequently adsorbed by available carboxyl groups.

TE

1024 A study on using fermentation waste for detoxification of Cr(VI) contaminated aqueous

1025 solutions was published by Park et al. (2008). In this study, the colorimetric method was used in
EP

1026 combination with excess potassium permanganate to analyze both the Cr(VI) and total Cr in

1027 aqueous solution, and XPS was used to ascertain the oxidation state of the chromium bound on the
C

1028 biomass. Park et al. (2008) concluded that the mechanism of Cr(VI) removal by the fermentation
AC

1029 waste involved reduction of Cr(VI) to Cr(III) (redox reaction). As a result, kinetic and equilibrium

1030 models based on the “reduction” mechanism were used to describe the behavior of Cr(VI) and total

1031 Cr in aqueous solution.

1032 Two main mechanisms for removal of Cr(VI) from aqueous solution by nonliving
52
 ACCEPTED MANUSCRIPT

1033 biomass were proposed by Park et al. (2005). In the first mechanism (direct reduction), Cr(VI)

1034 was directly reduced to Cr(III) in the aqueous phase by contact with electron-donor groups of the

1035 biomass, i.e., groups with lower reduction potentials than that of Cr(VI) (+1.3 V). However, the

1036 second mechanism comprised three steps: (1) binding of anionic Cr(VI) ion species to positively

PT
1037 charged groups on the biomass surface, (2) reduction of Cr(VI) to Cr(III) by adjacent

RI
1038 electron-donor groups, and (3) release of the Cr(III) ions into the aqueous phase, owing to

1039 electronic repulsion between the positively charged groups and the Cr(III) ions, or complexation

SC
1040 of Cr(III) with adjacent groups capable of binding chromium. If there are a small number of

1041 electron-donor groups in the biomass or a low concentration of protons in the aqueous phase,

U
1042 chromium bound on the biomass can remain in the hexavalent state. Therefore, the extent to

1043
AN
which mechanisms I and II operate depends on the nature of the biosorption system, including

1044 solution pH, temperature, functional groups on the biomass, and biomass and Cr(VI)
M
1045 concentrations (Park et al., 2006b; Park et al., 2006a). These two main mechanisms are
D

1046 summarized in Figure 9.

TE

1047 Figure 9

1048 8.3. Incorrect labels

EP

1049 Incorrect labeling of compounds should also be avoided. A publication entitled “Selenite

1050 adsorption and desorption in main Chinese soils with their characteristics and physicochemical
C

1051 properties” included such a mistake, as pointed out by Goldberg (2016). The purpose of this work
AC

1052 was to study the adsorption of Se(IV) (selenite), not Se(VI) (selenate); however, the paper reported

1053 that the authors prepared Se solutions using Na2SeO4 (hexavalent Se(VI)). This mistake was

1054 explained by the incorrect labeling of Na2SeO3 as Na2SeO4 in the Materials and Methods section.

53
 ACCEPTED MANUSCRIPT

1055 Similarly, the author of the paper entitled “Adsorption sequence of toxic inorganic anions

1056 on a soil” prepared solutions from Na2HAsO4·7H2O as As(V) species. However, they mislabeled

1057 Na2HAsO4·7H2O as NaAsO2, a trivalent As(III) species (Goldberg, 2009).

1058 8.4. BET specific surface area of an adsorbent

PT
1059 Using the Brunauer–Emmett–Teller (BET) surface area of pahokee peat (0.88 m2/g) and

RI
1060 assuming that the atrazine molecule is a sphere with a radius (r) of 4.16 × 10−10 m, Borisover and

1061 Graber (1997) estimated the maximum adsorption capacity of pahokee peat for atrazine. First, the

SC
1062 cross-section area (A) of atrazine can be obtained as 3.14 × (4.16 × 10−10)2 = 54.4 × 10−20 m2;

1063 subsequently, the surface covered by 1 mol of atrazine (MA) can be calculated as (54.4 × 10−20 m2)

U
1064 × (6.023 × 1023/mol) = 3.3 × 105 m2/mol. Second, the maximum possible concentration of atrazine

1065
AN
on the surface area (Qomax,covered; maximum adsorption capacity) was obtained by dividing the BET

1066 surface area (0.88 m2/g) by the surface covered by 1 mol of atrazine (3.3 × 105 m2/mol) to give 2.7
M
1067 × 10−6 mol/g. Finally, Qomax,covered was converted into weight units using (2.7 × 10−6 mol/g) ×
D

1068 (215.7 g/mol) to give 582 µg/g. This value (582 µg/g) is substantially higher than the maximum
TE

1069 adsorption capacity obtained from the Langmuir equation (161 µg/g). Although these calculations

1070 are correct, these inconsistent values might originate from a misunderstanding by the authors
EP

1071 (Borisover and Graber, 1997) who state that “N2 BET measured external surface area”, and thus

1072 attributed the value of 582 µg/g to the maximal possible atrazine concentration on the external
C

1073 surface of pahokee peat.

AC

1074 According to the 2011 guide for the accelerated surface area and porosimetry (ASAP)

1075 system, the definition of the BET specific surface area (m2/g) includes both the t-Plot external

1076 surface area (m2/g) and the t-Plot micropore area (m2/g). Table 5 gives a typical example of correct

1077 and incorrect use of the Micromeritics report (experimental data measured using a Micromeritics
54
 ACCEPTED MANUSCRIPT

1078 ASAP 2020 sorptometer at 77 K). Although there were no problems found in the N2

1079 adsorption/desorption isotherms, the report results for BET analysis of the prepared activated

1080 carbon in Table 5 contain discrepancies, as the BET specific surface area value is less than the

1081 external surface area. This error resulted from calculation of the specific surface area (SBET) in the

PT
1082 range 0.38 < p/p0 < 0.57, instead of 0.05 < p/p0 < 0.3.

RI
1083 Table 5

1084 Recently, Ben Ali et al. (2016) used the “iodine number” to determine the surface area of a

SC
1085 biosorbent without any treatment (pomegranate peel; PGP) and concluded that “the specific

1086 surface area obtained is equal to 598.78 m2/g. Iodine number is generally used as an

U
1087 approximation for surface area and microporosity of active carbons with good precision”. Tran

1088
AN
(2017) commented that there are two serious misconceptions in this work that need to be

1089 discussed.
M
1090 First, the iodine number of activated carbon is often determined following the internal
D

1091 method, ASTM D4607-14 (D4607-14 2014). According to the ASTM, the iodine number (mg/g;
TE

1092 amount of iodine adsorbed (mg) by 1.0 g of activated carbon) is a relative indicator of porosity in

1093 activated carbon. Iodine molecules (≈0.27 nm) can be adsorbed into the micropores (pore width >
EP

1094 1 nm) of porous materials. However, the iodine number does not necessarily provide a measure of

1095 the ability of activated carbon to adsorb other species. Although the iodine number may be used to
C

1096 approximate the surface area for several types of activated carbon, it must be realized that the
AC

1097 relationship between surface area and iodine number cannot be generalized, as it varies with

1098 changes in the raw carbon material, processing conditions, and pore volume distribution.

1099 Biosorbents that have not undergone any treatment have never been defined as porous materials;

1100 thus, the porosity of a biosorbent is negligible.

55
 ACCEPTED MANUSCRIPT

1101 Therefore, the iodine number cannot be applied for the determination of porosity for any

1102 biosorbent. Moreover, it is impossible to obtain the very high iodine number reported for PGP

1103 (602 mg/g). The high iodine number for PGP can be attributed to the high sulfur content (0.89%)

1104 and prolonged contact time (>30 s) (Tran, 2017). According to the ASTM (D4607-14 2014), after

PT
1105 transferring activated carbon (g) into hydrochloric acid solution (5 wt%), the mixture should be

RI
1106 boiled gently for 30 s to remove any sulfur that may interfere with the test results. Therefore, it is

1107 clear that the iodine number is greatly affected by high sulfur contents. Notably, the contact time

SC
1108 between activated carbon and iodine solution (0.1 N) strongly impacts the determined iodine

1109 number. The contact time recommended by the ASTM was approximately 30 s, while the contact

U
1110 time used by Ben Ali et al. was approximately 4 min. Recently, Tran et al. (2017c) prepared

1111
AN
activated carbons from golden shower using different chemical activation methods with K2CO3

1112 (GSAC, GSBAC, and GSHAC). Their results indicated that the iodine numbers (mg/g) of the
M
1113 prepared activated carbons at a contact time of 30 s were significantly lower than those at 5 min;
D

1114 GSAC (2,604 mg/g < 2,883 mg/g), GSBAC (1,568 mg/g < 2,296 mg/g), GSHAC (2,695 mg/g <
TE

1115 4,842 mg/g).

1116 Second, the textural properties of an adsorbent (i.e., specific surface area, total pore
EP

1117 volume, and micropore volume) can be obtained from conventional analysis of nitrogen

1118 adsorption-desorption isotherms, which are measured at 77 K using a sorptometer (e.g.,

C

1119 Micromeritics ASAP 2020). The BET method is widely employed to compute specific surface
AC

1120 area (SBET) (Marsh and Reinoso, 2006). Therefore, it is impossible to measure the surface area of

1121 an adsorbent using the iodine number method. This mistaken assumption will lead to incorrect

1122 result, such as the extremely high surface area of PGP (598.78 m2/g). The BET surface areas of

1123 various biosorbents are approximately 40 m2/g for yellow passion-fruit shell, 20 m2/g for orange
56
 ACCEPTED MANUSCRIPT

1124 peel, 8.17 m2/g for wheat straw, 8.13 m2/g for sargassum, 4.01 m2/g for Moringa oleifera lamarck

1125 seed powder, 1.31 m2/g Spirogyra species, 1.21 m2/g for waste pomace from an olive oil factory,

1126 0.76 m2/g for soy meal shells, 0.48 m2/g for rubber tree leaves, and 0.48 m2/g for rice bran (Farooq

1127 et al., 2010; Tran et al., 2016). Thus, the discussions and conclusions regarding the surface area of

PT
1128 PGP made by Ben Ali et al. (2016) are not valid, as they are inconsistent with the ASTM

RI
1129 definitions and fundamental understanding of the porosities of materials.

1130 8.5. Maximum absorption wavelength in dye adsorption studies

SC
1131 A comment on the method of using the maximum absorption wavelength (λmax) to calculate

1132 Congo red (CR) solution concentrations was published by Zhou et al. (2011). They pointed out that

U
1133 CR is very sensitive to pH and changes from red to blue owing to a π-π* transition of the azo group

1134
AN
following protonation (Figure 10a). At lower pH values, CR is protonated and cationic CR shows

1135 two tautomeric forms: ammonium rich variety and azonium variety. Therefore, the maximum
M
1136 absorption wavelength used to calculate the concentration of CR in solution is strongly dependent
D

1137 on the solution pH (both initial and final) (Figure 10b); for example, λmax was 576 nm at pH 2.18–
TE

1138 3.16, 567 nm at pH 3.86, and 496 nm at pH ≥ 4.71.

1139 Figure 10
EP

1140 Tien (2007) and Tien (2008) highlighted that commercial dyes are often mixtures of active

1141 ingredients and filler materials are often not recognized. Therefore, the possibility that
C

1142 experiments may involve bi-solute adsorption is often not considered.

AC

1143 8.6. Cπ-cation and π-π interactions

1144 8.6.1. Cπ-cation interactions

1145 Recently, Morosanu et al. (2016) investigated the biosorption of lead ions onto rapeseed

1146 biomass that was used as a biosorbent without any treatments. On the basis of FTIR measurements
57
 ACCEPTED MANUSCRIPT

1147 before and after adsorption, they concluded that the disappearance of the peak at 1710 cm−1, which

1148 is characteristic of C=O stretching, indicated surface complexion through Cπ-cation interactions.

1149 Similarly, Medellin-Castillo et al. (2017) examined the biosorption of Pb2+ and Cd2+ onto a

1150 biosorbent derived from industrial chili seeds. They also found changes in the intensity of the

PT
1151 absorbance peak at 1657 cm−1 in heavy-metal-loaded chili seeds, suggesting that the adsorption of

RI
1152 Cd2+ and Pb2+ involved π-cation interactions between the aromatic rings of the lignin and the

1153 Cd2+ and Pb2+ cations in the solution.

SC
1154 In general, Cπ-cation interactions are attributed to electrostatic interactions between the

1155 aromatic rings of basic carbonaceous materials (i.e., biochar, carbon foam, carbon nanotubes,

U
1156 activated carbon, and graphene) and metallic cations. For example, Tran et al. (2015) used

1157
AN
orange-peel-derived biochar, which was produced through a carbonization process under

1158 limited-oxygen conditions (also known as pyrolysis) at a high temperature (>400 °C). Biochar is
M
1159 known as a carbon-enriched porous material, similar to activated carbon. Thus, biochar also
D

1160 possesses a graphitic structure (C=C bonds; π-electrons). As a result, Cπ-cation interactions played
TE

1161 a primary role in the adsorption of heavy metals onto biochar. Analogous conclusions were

1162 reached by other scholars (Swiatkowski et al., 2004; Uchimiya et al., 2010; Rivera-Utrilla and
EP

1163 Sánchez-Polo, 2011).

1164 The contribution of Cπ-cation interactions to the mechanism of heavy-metal adsorption

C

1165 onto carbonaceous material cannot be denied. However, unlike biochar and activated carbon,
AC

1166 biosorbents (commonly derived from lignocellulose materials) do not possess a graphitic

1167 structure. In addition, although hydorchar prepared through hydrothermal carbonization (i.e., 150–

1168 250 °C controlled temperature for 2–72 h at a specific pressure) exhibits an aromatic carbon

1169 network, its properties are similar to those of biosorbents (Tran et al., 2017c). Therefore, the
58
 ACCEPTED MANUSCRIPT

1170 presence of Cπ-cation interactions in heavy-metal adsorption onto biosorbents and hydrochar is

1171 still controversial and needs to be further investigated.

1172 Bui and Choi (2010) considered the role of cation-π bonding between the protonated

1173 amino group of oxytetracycline (OTC) and the graphitic π electrons of multiwalled carbon

PT
1174 nanotubes (MWCNTs). Essentially, the cation-π bonding is dominated by the electrostatic forces

RI
1175 between the cation and the permanent quadrupole of the π-electron-rich aromatic structure and

1176 cation-induced polarization (Ji et al., 2009). Like π-π interactions, cation-π bonding would be

SC
1177 suppressed with increasing pH. This is because at a higher pH, deprotonation of the charged amino

1178 group and protonation of the enol groups will be enhanced and so the electron-acceptor ability of

U
1179 the oxytetracycline molecule is weaker (Ji et al., 2009). Bui and Choi (2010) opined that

1180
AN
Oleszczuk et al. (2010) ignored the important roles of Ca2+ in the solution; for example,

1181 Ca2+ possibly forms a complex with OTC in the solution or Ca2+ can act as a cation bridge between
M
1182 the negatively charged OTC and the negative charges of the MWCNTs. Therefore, Bui and Choi
D

1183 (2010) suggested that Ca2+ may simultaneously bind with the negatively charged OTC and interact
TE

1184 either with the negative charges or the graphitic π electrons of the MWCNTs via cation-π bonding.

1185 As pH is increased, both MWCNTs and OTC become more negatively charged and, consequently,
EP

1186 probably interact more strongly with Ca2+, leading to lower desorption of OTC. This hypothesis

1187 could be further verified by studying the desorption of OTC in the presence of NaCl, instead of
C

1188 CaCl2.
AC

1189 8.6.2. π-π interactions

1190 π-π interactions (also known as π-π electron donor-acceptor interactions) between

1191 the π-electrons in a carbonaceous adsorbent and the π-electron in the aromatic ring of an adsorbate

1192 were initially proposed by Coughlin and Ezra (1968). In short, the addition of
59
 ACCEPTED MANUSCRIPT

1193 electron-withdrawing groups (i.e., oxygen-containing and nitrogen-containing functional groups)

1194 at the edges of the individual graphene layers within a carbonaceous solid causes a considerable

1195 drop in the π-electron density. Positive holes are consequently created in the conduction π-band of

1196 the individual graphene layers, and the interactions between the π-electrons of a carbonaceous

PT
1197 adsorbent and the π-electrons of the adsorbate aromatic rings become weaker.

RI
1198 Recently, Tran et al. (2017b), and Tran et al. (2017d) investigated the mechanisms of

1199 methylene green 5 (MG5) adsorption by commercial activated charcoal, synthesized activated

SC
1200 carbons, and prepared biochars. They proposed that the primary mechanisms in MG5 adsorption

1201 were π-π interactions and pore filling, while hydrogen bonding and n-π interactions were minor

U
1202 contributors (Figure 11). To identify the existence of π-π interactions, they used two pieces of

1203
AN
experimental evidence. First, FTIR analysis showed that a peak corresponding to the skeletal

1204 vibration of aromatic C=C bonds decreased in intensity and upshifted after MG5
M
1205 adsorption. Secondly, oxygenation of the surface of the carbonaceous solids (i.e., biochar and
D

1206 activated carbon) through a hydrothermal process with acrylic acid resulted in a decrease in MG5
TE

1207 adsorption and indicated the importance of π-π interactions to the adsorption process.

1208 As discussed in Section 8.6.1, hydrochar also possesses an aromatic carbon network;
EP

1209 therefore, π-π interactions might be expected between the π-electrons of the aromatic C=C bonds

1210 in hydrochar and the π-electrons of an adsorbate. However, a recent study (Tran et al., 2017e)
C

1211 demonstrated that oxygenation of a hydrochar surface through a hydrothermal process with acrylic
AC

1212 acid contributed to increasing MG5 adsorption and indicated the negligible role of π-π interaction

1213 in the adsorption process. FTIR analysis demonstrated that the aromatic C=C peak did not

1214 significantly decrease in intensity or shift toward higher/lower wavenumbers after adsorption,

1215 which further confirms the insignificant contribution of π-π interactions. Electrostatic attraction
60
 ACCEPTED MANUSCRIPT

1216 played a major role in the adsorption mechanism, while hydrogen bonding and n-π interactions

1217 were minor contributors. The primary adsorption mechanisms of MG5 onto hydrochar were

1218 similar to those on biosorbents, but dissimilar to biochar and activated carbon (i.e., π-π interaction

1219 and pore filling). An identical result was obtained in the investigation of phenol, MG5, and acid

PT
1220 read 1 adsorption onto a commercial glucose-derived spherical hydrochar functionalized with

RI
1221 triethylenetetramine (Tran, et al., 2017a). Therefore, it can be concluded that the addition of

1222 oxygen- and nitrogen-containing functional groups (electron-withdrawing groups) on the surface

SC
1223 of hydrochar does not cause a considerable drop in the π-electron density, but this process provides

1224 abundant adsorption sites on the hydrochar surface.

U
1225 Jiang et al. (2009) and Chen et al. (2009) argued over the primary adsorption mechanisms

1226
AN
of hydroxyl- and amino-substituted aromatics on carbon nanotubes and graphite (CNTs/graphite).

1227 Chen et al. (2009) proposed that π-π interactions played a major role in the adsorption mechanism.
M
1228 However, Jiang et al. (2009) remarked that the experiments of Chen et al. (2009) were
D

1229 insufficient to support the presence of π-π electron-donor−acceptor (EDA) interactions between
TE

1230 hydroxyl-substituted aromatics and CNTs/graphite. According to Jiang et al. (2009), phenolic

1231 compounds were not appropriate candidates to assess the presence of π-π EDA interactions
EP

1232 between π-electron-donating aromatics and CNTs because of the significant effects of oxygen.

1233 Therefore, the oxygen effect should be considered in more detail in order to gain in-depth insights
C

1234 into the adsorption mechanisms of environmentally relevant phenolic compounds onto
AC

1235 carbonaceous materials.

61
 ACCEPTED MANUSCRIPT

1236 Figure 11

1237 8.7. Other miscellaneous errors

1238 Azizian (2007) commented that the theoretical development of empirical PFO and PSO

1239 kinetic adsorption models using statistical rate theory (SRT) was first reported by Azizian (2004),

PT
1240 not by Rudzinski and Plazinski (2007). Furthermore, the idea that the PFO model can only be

RI
1241 derived theoretically at a nearly constant bulk concentration was proposed by Azizian (2004). In

1242 addition, the conclusion that the PFO and PSO models are simplified forms of a more general

SC
1243 equation was also proposed by Azizian (2004), not Rudzinski and Plazinski (2007). Similarly,

1244 mistakes related to either no citation of the original paper or incorrect citation were pointed out by

U
1245 Azizian (2008b).

1246
AN
A series of mistakes related to incorrectly digitized data, typos, inaccurate calculations,

1247 incorrect use of conditional equilibrium constants, incorrect units, incorrect references, incorrect
M
1248 application of modeling procedures, and other miscellaneous errors have been highlighted by
D

1249 Gustafsson and Lumsdon (2014).

TE

1250 The distinction between heat of adsorption, adsorption energy, and activation energy in

1251 adsorption was reported elsewhere (Inglezakis and Zorpas, 2012). The authors concluded that it is
EP

1252 important to distinguish the differential enthalpy at zero coverage (∆Hdo) from the differential

1253 enthalpy or isosteric heat (∆Hd) and the standard heat of complete coverage or integral enthalpy
C

1254 (∆H°), which expresses the total heat generation for the adsorption and is related to the
AC

1255 thermodynamic equilibrium constant.

1256 Li et al. (2010) pointed out that (1) although the cited documents were solid and sound in

1257 their scientific aspects, the reviewers seemed to fail to take full advantage of the intrinsically

1258 contained information; (2) there were some critical errors in the main text; and (3) the writing style
62
 ACCEPTED MANUSCRIPT

1259 of the paper was misleading and, to some extent, lacked cautiousness. They focused on 11

1260 comments and concluded that the original paper’s conclusions were not optimistic. However, most

1261 of these comments were rejected by the original authors (Mohan and Pittman, 2011).

1262 Regarding the correct citation for determining pHPZC using the batch equilibration method,

PT
1263 Milonjić (2009a) commented that this method was originally proposed by Milonjić et al. (1975).

RI
1264 Furthermore, Milonjić (2009a) noted that using correct and updated citations were very important

1265 for researchers to find relevant information, pioneer ideas, and make progress in a particular

SC
1266 subject. Additionally, from a scientific point of view, it is always necessary to give credit to the

1267 authors who first proposed a method or theoretical model. Recently, Tran et al. (2017b) applied the

U
1268 “drift method” to determine the pHPZC of commercial activated charcoal (CAC). The effects of

1269
AN
various operation conditions (i.e., different degassing times with N2, background electrolytes,

1270 concentrations of an electrolyte, solid/liquid ratios, and contact times) on the pHPZC were
M
1271 investigated. The results demonstrated that the pHPZC (9.81 ± 0.07) of CAC was insignificantly
D

1272 dependent on the operation conditions.

TE

1273 Geckeis and Rabung (2004) pointed out that photon correlation spectroscopy (PCS) is a

1274 method for colloid size determination, not for zeta-potential analysis. Methods that have been
EP

1275 correctly applied for zeta-potential analysis include electrophoretic light scattering and laser

1276 Doppler velocimetry (LDV), and the units for the measured zeta potentials are mV, not eV.
C

1277 Lee (2017) considered the potential effects of flow velocity on the concentration and
AC

1278 distribution of triphenyltin chloride (TPT) on nano zinc oxide (nZnO). The author remarked that

1279 TPT was possibly re-suspended by the hydraulic flow that occurs during hydrant flushing in a

1280 reactor because TPT can be loosely deposited on the nZnO surface. As reported in the literature,

1281 the flow velocity in a wastewater treatment system is one of the factors affecting the removal
63
 ACCEPTED MANUSCRIPT

1282 efficiency for a tested pollutant. The distribution of TPT on nZnO might vary during the

1283 adsorption process according to the time of hydrant flushing; therefore, the process of TPT

1284 adsorption onto nZnO might be affected by the operating conditions of a reactor. However, the

1285 authors of the commented paper replied that the application of nZnO for the removal of TPT from

PT
1286 dockyard wastewater was examined using a batch adsorption technique. As a result, the effects of

RI
1287 flow velocity on the adsorption process conducted in the batch experiments were negligible

1288 compared to that in pilot studies and/or column experiments.

SC
1289 Kamagaté et al. (2016) and Komárek et al. (2016) considered the influence of Si species

1290 originating from the partial dissolution of quartz on the competitive adsorption of Cd(II), Cr(VI),

U
1291 and Pb(II) onto nanomaghemite- and nanomaghemite-coated quartz. They recommended that to

1292
AN
accurately assess the interfacial reaction mechanisms occurring at the Fe-oxide/water interface in

1293 the presence of Si-complex mineral assemblages or Fe-coated sand systems, more attention should
M
1294 be paid to the possible release of silicate from Si-bearing minerals and its subsequent adsorption on
D

1295 reactive phases.

TE

1296 A critical paper entitled “comments on “zirconium-carbon hybrid sorbent for removal of

1297 fluoride from water: oxalic acid mediated Zr(IV) assembly and adsorption mechanism”” was
EP

1298 published by Zhao (2015). The authors of the original paper concluded that the fluoride adsorption

1299 occurs on the surface sites of ZrOx-AC (commercial activated carbon modified with Zr together
C

1300 with oxalic acid) with –OH displacement and/or interacts with the positive charge of zirconium to
AC

1301 form –COOH groups in the oxalate acid; however, Zhao (2015) suggested that if the fluoride

1302 concentration is high enough, it would completely desorb the oxalate acid from the ZrOx-AC

1303 surface. According to the results of previous experiments, Zhao (2015) suggested that it is

64
 ACCEPTED MANUSCRIPT

1304 necessary to further consider or add the exchangeable role of chloride and oxalic acid in the

1305 proposed fluoride-adsorption mechanism.

1306 Fan et al. (2007) additionally discussed the mechanisms of fluoride adsorption by calcined

1307 Mg-Al-CO3 layered double hydroxides (CLDH). Based on the experimental data published in the

PT
1308 original paper and analysis of other literature results, they concluded that XRD measurements

RI
1309 indicated that mixed Mg-Al oxides, as well as partially restored LDH, was present in

1310 fluoride-loaded CLDH after freeze-drying treatment. Furthermore, fluoride can be effectively

SC
1311 adsorbed onto MgO, demonstrating that fluoride adsorption over the mixed oxide along with a

1312 memory effect accounts for the effective removal of fluoride by CLDH.

U
1313 An analytical article on the magnitude of estimated maximum surface-area-normalized

1314
AN
adsorption capacities (Qmax*) for the adsorption of polycyclic aromatic hydrocarbons (PAHs) and

1315 polychlorobiphenyls (PCBs) on soot and soot-like carbon materials was published by Werner and
M
1316 Karapanagioti (2005). The original authors (van Noort et al., 2005) used the Langmuir isotherm to
D

1317 extrapolate Qmax* at the solubility limit from solid-water distribution coefficients (Ks) measured in
TE

1318 the pg/L to ng/L range along with estimated Langmuir affinities for adsorption (b) on a

1319 carbonaceous surface. After comparison with estimated surface-area-normalized monolayer

EP

1320 adsorption capacities (Qmono*) and with empirical sorption data from the literature, Werner and

1321 Karapanagioti (2005) concluded that the Langmuir isotherm cannot be used to extrapolate
C

1322 maximum sorption capacities (Qmax*) near the aqueous-sorbate saturation limit from distribution
AC

1323 coefficients measured at extremely low aqueous-sorbate concentrations. Adsorption sites in

1324 carbon materials are not uniform and other processes, such as multilayer adsorption, condensation

1325 in capillary pores, and absorption into the polymeric matrix, may be relevant near the solubility

1326 limit.
65
 ACCEPTED MANUSCRIPT

1327 Müller et al. (2017) analyzed strong conflicts between the XPS and XRD data and their

1328 interpretation by the original author in a study on the synthesis of Yb doped CuFe2O4 nanoferrite

1329 adsorbents (i.e., CuYb0.5Fe1.5O4) and the application of such adsorbents in the removal of different

1330 pollutants (e.g. methyl orange, safranin, Cr3 +, and Pb2 +). After a polemical discussion, Müller et

PT
1331 al. (2017) concluded that because any experimental study stands and falls on the reliability of the

RI
1332 investigated specimen the lack of any significance therefore holds for this study as a whole.

1333 Interestingly, a series of recommendations and suggestions were written by the editor of

SC
1334 Carbon for improving manuscript quality and avoiding mistakes (Thrower, 2007; Thrower,

1335 2008b; Thrower, 2008a; Thrower, 2010; Thrower, 2011). The author focused on five main topics:

U
1336 titles and abstracts, introduction and references, experimental, results and discussion, and

1337 language.
AN
1338 9. Nonlinear-optimization technique
M
1339 To calculate the parameters of kinetic and isotherm models accurately in both batch and
D

1340 column experiments, application of the nonlinear method instead of the linear method has been
TE

1341 recommended by many researchers (Porter, 1985; Ho, 2004b; Ho et al., 2005; Bolster and

1342 Hornberger, 2007; Han et al., 2007; Kumar, 2007; El-Khaiary et al., 2010; Chowdhury and Das
EP

1343 Saha, 2011; Lima et al., 2015; Tran et al., 2015; Tran et al., 2017d).

1344 Furthermore, to identify the best-fit model, calculation of the chi-squared (χ2, Eq. 96) value
C

1345 is recommended in addition to calculating the coefficient of determination (R2, Eq. 97) for the
AC

1346 nonlinear method. In the chi-squared test, the squares of the differences between the experimental

1347 data and data calculated using the models are divided by the corresponding data obtained and then

1348 summed. If the data obtained using a model are similar to the experimental data, χ2 is close to zero.

1349 High χ2 values indicate high bias between the experiment and model. Therefore, analyzing the data
66
 ACCEPTED MANUSCRIPT

1350 set using the chi-squared test is necessary to confirm the best-fit isotherm for a given sorption

1351 system (Ho, 2004b; Ho et al., 2005; Chowdhury and Das Saha, 2011; Lima et al., 2015; Tran et al.,

1352 2015; Tran et al., 2017b; Tran et al., 2017d; Tran et al., 2017e).

(qe,exp − qe,cal ) 2

PT
χ =∑2
(96)
qe,cal

− qe,cal )2

RI
R 2
=1 −
∑(q e, exp
=
∑ (q e, cal - qe,mean)2
∑ (q e,exp − qe, mean)2 ∑ (q
e, cal -q e, mean) + ∑(qe, cal - qe,exp )2
2
(97)

SC
1353 where qe,exp (mg/g) is the amount of adsorbate uptake at equilibrium obtained from Eq. 1, qe,cal

1354 (mg/g) is the amount of adsorbate uptake achieved from the model using the Solver add-in, and

U
1355 qe,mean (mg/g) is the mean of the qe,exp values.

1356
AN
Lima et al. (2015) confirmed that the use of nonlinear equilibrium and kinetic adsorption

1357 models gives values that are reliable and statistically relevant for modeling the isotherm and
M
1358 kinetics of adsorption. Moreover, they did not suggest using linearized equilibrium and kinetic
D

1359 adsorption models, even in the cases of the linearized Langmuir Type 1 and linearized PSO Type 1
TE

1360 equations that provide R2 values close to unity. The linearization of equilibrium and kinetic

1361 adsorption models could make the parameters determined from these models meaningless.
EP

1362 Owing to the wide application of the nonlinear method in the study of adsorption

1363 processes, we provide here a short introduction to the application of this method using the “Solver
C

1364 add-in” in Microsoft Office Excel (2013 version). First, it is necessary to load the Solver add-in in
AC

1365 Microsoft Excel: File → Excel options → Add-ins (select Excel Add-ins in Manage box) → select

1366 the Solver Add-in check box. Second, all Ce and qe values obtained experimentally are considered

1367 as input data, and the qe values from the Langmuir equation are calculated based on two variables

1368 (Qomax and KL ≠ 0). Using the Solver Add-in (Figure 12), the Qomax, KL, and R2 values can be

67
 ACCEPTED MANUSCRIPT

1369 obtained. Notably, other programs (i.e., Origin) also provide similar results and can accurately

1370 calculate the parameters of kinetic and isotherm models.

1371 Figure 12

1372 10. Conclusions

PT
1373 In scientific publications, the correct use of technical terms and accurate calculations are

RI
1374 essential. According to the viewpoints discussed in this review, we put forward the following key

1375 conclusions and perspectives:

SC
1376 i. Adsorption performance should be expressed as qe (mg/g), while the use of
1377 the unit of %removal for qe should be avoided or used cautiously.

U
1378 ii. The use of accurate citations and correct mathematical expressions in
1379 original works is also recommended.
AN
1380 iii. Measurements of adsorption kinetics should be started at an initial time of
M
1381 less than 2 min and finish when the adsorption process reaches true equilibrium.
D

1382 iv. A complete adsorption isotherm (plotting qe versus Ce) should be presented
1383 when investigating liquid-phase adsorption to identify the regions in which the
TE

1384 experimental data relating to adsorption equilibrium are actually located.

1385 v. Differences between the number of data points in an experiment and those
EP

1386 used for model fitting (i.e., the pseudo-second-order or Langmuir models) should be
1387 avoided.
C

1388 vi. For adsorption thermodynamics, the equilibrium constant (KC) must be
AC

1389 dimensionless. The optimal method for calculating thermodynamic parameters should
1390 be appropriately selected according to the data from the adsorption equilibrium
1391 experiment. Notably, it is impossible to calculate thermodynamic parameters from an
1392 experiment using one initial adsorbate concentration at different temperatures.

68
 ACCEPTED MANUSCRIPT

1393 vii. To accurately estimate the parameters of adsorption kinetic and isotherm
1394 models, the nonlinear optimization technique should be applied to decrease the bias
1395 between the qe values calculated from the experimental data and those estimated from
1396 the models. The chi-squared (χ2) test should be used alongside the nonlinear
determination coefficient (R2) to obtain the best-fit models for adsorption kinetics and

PT
1397
1398 isotherms.

RI
1399 viii. An in-depth understanding of the models applied to adsorption kinetics and
1400 isotherms is necessary to avoid misapplication of these models, as well as inaccurate

SC
1401 discussions, calculations, and conclusions.

1402 ix. More attention should be paid to various common misunderstandings about

U
1403 fundamental adsorption and chemistry concepts. Understanding the unique properties
1404
1405 adsorption mechanisms.
AN
of adsorbents and target adsorbates can help avoid mistakes made when explaining
M
1406 Although the comments and recommendations of some authors are not always correct, their

1407 observations and suggestions should be acknowledged because of their great contributions in
D

1408 transferring scientific knowledge.

TE

1409 Acknowledgements

1410 This current work was financially supported by Chung Yuan Christian University (CYCU) in
EP

1411 Taiwan. The first author would like to thank CYCU for the Distinguished International Graduate
C

1412 Students (DIGS) scholarship to pursue his doctoral studies. The authors gratefully acknowledge
AC

1413 the editor and anonymous reviewer for their invaluable insight and helpful suggestion.

1414 Reference

1415 Alagumuthu, G., V. Veeraputhiran and M. Rajan (2010). "Comments on “Fluoride removal from
1416 water using activated and MnO2-coated Tamarind Fruit (Tamarindus indica) shell:
1417 Batch and column studies”." Journal of Hazardous Materials 183(1–3): 956-957.
69
 ACCEPTED MANUSCRIPT

1418 Anastopoulos, I. and G. Z. Kyzas (2016). "Are the thermodynamic parameters correctly
1419 estimated in liquid-phase adsorption phenomena?" Journal of Molecular Liquids 218:
1420 174-185.

1421 Aoyama, M. (2003). "Comment on “Biosorption of chromium(VI) from aqueous solution by

PT
1422 cone biomass of Pinus sylvestris”." Bioresource Technology 89(3): 317-318.

1423 Arica, M. Y. (2003). "In reaction to the comment by Dr. Y.-S. Ho on our publication “Affinity

RI
1424 dye-ligand poly(hydroxyethylmethacrylate/chitosan composite membrane for
1425 adsorption lysozyme and kinetic properties, Biochemical Engineering Journal 13

SC
1426 (2003) 35–45”." Biochemical Engineering Journal 15(1): 79-80.

1427 Azizian, S. (2004). "Kinetic models of sorption: a theoretical analysis." Journal of Colloid and
1428 Interface Science 276(1): 47-52.

U
1429 Azizian, S. (2006). "A novel and simple method for finding the heterogeneity of adsorbents on
1430
AN
the basis of adsorption kinetic data." Journal of Colloid and Interface Science 302(1):
1431 76-81.
M
1432 Azizian, S. (2007). "Comment on “Kinetics of Solute Adsorption at Solid/Solution Interfaces:  A
1433 Theoretical Development of the Empirical Pseudo-First and Pseudo-Second Order
D

1434 Kinetic Rate Equations, Based on Applying the Statistical Rate Theory of Interfacial
TE

1435 Transport”." The Journal of Physical Chemistry B 111(1): 318-318.

1436 Azizian, S. (2008a). "Comments on “Adsorption Equilibrium and Kinetics of Reactive Black 5
EP

1437 and Reactive Red 239 in Aqueous Solution onto Surfactant-Modified Zeolite”
1438 (Karadag, D.; Turan, M.; Akgul, E.; Tok, S.; Faki, A. J. Chem. Eng. Data 2007, 52,
1439 1615−1620)." Journal of Chemical & Engineering Data 53(1): 322-323.
C

1440 Azizian, S. (2008b). "Comments on “Biosorption isotherms, kinetics and thermodynamics”

AC

1441 review." Separation and Purification Technology 63(2): 249-250.

1442 Ben Ali, S., I. Jaouali, S. Souissi-Najar and A. Ouederni (2016). "Characterization and adsorption
1443 capacity of raw pomegranate peel biosorbent for copper removal." Journal of Cleaner
1444 Production.

1445 Biggar, J. W. and M. W. Cheung (1973). "Adsorption of Picloram

70
 ACCEPTED MANUSCRIPT

1446 (4-Amino-3,5,6-Trichloropicolinic Acid) on Panoche, Ephrata, and Palouse Soils: A

1447 Thermodynamic Approach to the Adsorption Mechanism1." Soil Science Society of
1448 America Journal 37(6).

1449 Blanchard, G., M. Maunaye and G. Martin (1984). "Removal of heavy metals from waters by

PT
1450 means of natural zeolites." Water Research 18(12): 1501-1507.

1451 Bolster, C. H. and G. M. Hornberger (2007). "On the Use of Linearized Langmuir equation." Soil

RI
1452 Science Society of America Journal 71(6): 1796-1806.

1453 Borisover, M. D. and E. R. Graber (1997). "Comment on “Competitive Sorption between

SC
1454 Atrazine and Other Organic Compounds in Soils and Model Sorbents”."
1455 Environmental Science & Technology 31(5): 1577-1577.

U
1456 Bui, T. X. and H. Choi (2010). "Comment on “Adsorption and Desorption of Oxytetracycline
1457 and Carbamazepine by Multiwalled Carbon Nanotubes”." Environmental Science &
1458 Technology 44(12): 4828-4828.
AN
1459 Canzano, S., P. Iovino, S. Salvestrini and S. Capasso (2012). "Comment on “Removal of anionic
M
1460 dye Congo red from aqueous solution by raw pine and acid-treated pine cone powder
1461 as adsorbent: Equilibrium, thermodynamic, kinetics, mechanism and process design”."
D

1462 Water Research 46(13): 4314-4315.

TE

1463 Cavas, L. (2008). Comment on Equilibrium sorption isotherm studies of Cd(II), Pb(II) and Zn (II)
1464 ions detoxification from waste water using unmodified and EDTA-modified maize
EP

1465 husk.

1466 Chen, W., L. Duan, L. Wang and D. Zhu (2009). "Response to Comment on “Adsorption of
C

1467 Hydroxyl- and Amino-Substituted Aromatics to Carbon Nanotubes”." Environmental

1468 Science & Technology 43(9): 3400-3401.
AC

1469 Chien, S. H. and W. R. Clayton (1980). "Application of Elovich Equation to the Kinetics of
1470 Phosphate Release and Sorption in Soils." Soil Science Society of America Journal
1471 44(2): 265-268.

1472 Choong, T. S. Y. and T. G. Chuah (2005). "Comment on “Separation of vitamin E from palm
1473 fatty acid distillate using silica: I. Equilibrium of batch adsorption by B.S. Chu et al.
71
 ACCEPTED MANUSCRIPT

1474 [Journal of Food Engineering 62 (2004) 97–103]”." Journal of Food Engineering 67(3):
1475 379.

1476 Chowdhury, S. and P. Das Saha (2011). "Comparative Analysis of Linear and Nonlinear Methods
1477 of Estimating the Pseudo-Second-Order Kinetic Parameters for Sorption of Malachite

PT
1478 Green onto Pretreated Rice Husk." Bioremediation Journal 15(4): 181-188.

1479 Chu, B. S., B. S. Baharin, Y. B. C. Man and S. Y. Quek (2005). "Reply from the authors to the

RI
1480 comments by Choong and Chuah in a recent Letter to the Editor on “Separation of
1481 vitamin E from palm fatty acid distillate using silica: I. Equilibrium of batch

SC
1482 adsorption by B.S. Chu et al. [Journal of Food Engineering 62 (2004) 97–103]”."
1483 Journal of Food Engineering 69(1): 131-132.

1484 Coughlin, R. W. and F. S. Ezra (1968). "Role of surface acidity in the adsorption of organic

U
1485 pollutants on the surface of carbon." Environmental Science & Technology 2(4):
1486 291-297. AN
1487 Crittenden, J. C., R. R. Trussell, D. W. Hand, K. J. Howe and G. Tchobanoglous (2012).
M
1488 Adsorption. MWH's Water Treatment: Principles and Design, Third Edition, John
1489 Wiley & Sons, Inc.: 1117-1262.
D

1490 D4607-14, A. (2014). "Standard Test Method for Determination of Iodine Number of Activated
TE

1491 Carbon." American Society for Testing and Materials, Philadelphia, U.S.A.

1492 Dambies, L., C. Guimon, S. Yiacoumi and E. Guibal (2001). "Characterization of metal ion
EP

1493 interactions with chitosan by X-ray photoelectron spectroscopy." Colloids and

1494 Surfaces A: Physicochemical and Engineering Aspects 177(2–3): 203-214.
C

1495 Dawood, S. and T. K. Sen (2012). "Author's Responses to the comment by Canzano et al and
1496 also corrigendum to “Removal of anionic dye Congo red from aqueous solution by
AC

1497 raw pine and acid-treated pine cone powder as adsorbent: Equilibrium, thermodynamic,
1498 kinetics, mechanism and process design” published in Water Research, Vol. 46, pp.
1499 1933–1946, 2012." Water Research 46(13): 4316-4317.

1500 Doke, K. M. and E. M. Khan (2013). "Adsorption thermodynamics to clean up wastewater;

1501 critical review." Reviews in Environmental Science and Bio/Technology 12(1): 25-44.
72
 ACCEPTED MANUSCRIPT

1502 Dubinin, M. and L. Radushkevich (1947). "Equation of the characteristic curve of activated
1503 charcoal." Chem. Zentr 1(1): 875.

1504 Egirani, D. (2004). "Comments on: Removal of copper ions from aqueous solution by tree
1505 ferns." Water Research 38(20): 4535.

PT
1506 El-Khaiary, M. I., G. F. Malash and Y.-S. Ho (2010). "On the use of linearized
1507 pseudo-second-order kinetic equations for modeling adsorption systems." Desalination

RI
1508 257(1–3): 93-101.

1509 Fan, J., Z. Xu and S. Zheng (2007). "Comment on “Factors influencing the removal of fluoride

SC
1510 from aqueous solution by calcined Mg–Al–CO3 layered double hydroxides”." Journal
1511 of Hazardous Materials 139(1): 175-177.

U
1512 Farooq, U., J. A. Kozinski, M. A. Khan and M. Athar (2010). "Biosorption of heavy metal ions
1513 using wheat based biosorbents – A review of the recent literature." Bioresource
1514 Technology 101(14): 5043-5053.
AN
1515 Febrianto, J., A. N. Kosasih, J. Sunarso, Y.-H. Ju, N. Indraswati and S. Ismadji (2009).
M
1516 "Equilibrium and kinetic studies in adsorption of heavy metals using biosorbent: A
1517 summary of recent studies." Journal of Hazardous Materials 162(2–3): 616-645.
D

1518 Foo, K. Y. and B. H. Hameed (2010). "Insights into the modeling of adsorption isotherm
TE

1519 systems." Chemical Engineering Journal 156(1): 2-10.

1520 Freundlich, H. (1906). "Uber die adsorption in lo sungen." Z Physics Chemistry 57: 385-471.
EP

1521 Fu, L., X. Huang and J. Wang (2007). "Comment on “Removal of Supranol Yellow 4GL by
1522 adsorption onto Cr-intercalated montmorillonite”." Journal of Hazardous Materials
C

1523 148(3): 771-772.

AC

1524 Fu, L., J. Wang, H. Lu, Y. Su and A. Ren (2008). "Comment on “The removal of phenolic
1525 compounds from aqueous solutions by organophilic bentonite”." Journal of Hazardous
1526 Materials 151(2–3): 851-854.

1527 Gardea-Torresdey, J. L., K. J. Tiemann, V. Armendariz, L. Bess-Oberto, R. R. Chianelli, J. Rios,

1528 J. G. Parsons and G. Gamez (2000). "Characterization of Cr(VI) binding and reduction

73
 ACCEPTED MANUSCRIPT

1529 to Cr(III) by the agricultural byproducts of Avena monida (Oat) biomass." Journal of
1530 Hazardous Materials 80(1–3): 175-188.

1531 Geckeis, H. and T. Rabung (2004). "Comment on the paper: X. Wang, Th. Rabung, H. Geckeis,
1532 Effect of pH and humic acid on the adsorption of cesium onto γ-Al2O3." Journal of

PT
1533 Radioanalytical and Nuclear Chemistry 260(2): 437-438.

1534 Ghosal, P. S. and A. K. Gupta (2015). "An insight into thermodynamics of adsorptive removal of

RI
1535 fluoride by calcined Ca-Al-(NO3) layered double hydroxide." RSC Advances 5(128):
1536 105889-105900.

SC
1537 Ghosal, P. S. and A. K. Gupta (2017). "Determination of thermodynamic parameters from
1538 Langmuir isotherm constant-revisited." Journal of Molecular Liquids 225: 137-146.

U
1539 Goldberg, S. (2009). "Comments on “Adsorption Sequence of Toxic Inorganic Anions on a Soil”
1540 by K. Saeki Bull Environ Contam Toxicol (2008) 81:508–512." Bulletin of
1541
AN
Environmental Contamination and Toxicology 83(1): 1-2.

1542 Goldberg, S. (2016). "A comment on “Selenite adsorption and desorption in main Chinese soils
M
1543 with their characteristics and physicochemical properties.” J Soils Sediments
1544 15(5):1150–1158 (2015), doi:10.1007/s11368-015-1085-7." Journal of Soils and
D

1545 Sediments 16(1): 324-324.

TE

1546 Goswami, R., J. Shim, S. Deka, D. Kumari, R. Kataki and M. Kumar (2016). "Characterization
1547 of cadmium removal from aqueous solution by biochar produced from Ipomoea
EP

1548 fistulosa at different pyrolytic temperatures." Ecological Engineering 97: 444-451.

1549 Guo, H., S. Zhang, Z. Kou, S. Zhai, W. Ma and Y. Yang (2015). "Removal of cadmium(II) from
C

1550 aqueous solutions by chemically modified maize straw." Carbohydrate Polymers

1551 115(0): 177-185.
AC

1552 Gustafsson, J. P. and D. G. Lumsdon (2014). "Comment on “Citrate adsorption can decrease
1553 soluble phosphate concentration in soils: Results of theoretical modelling” by Marek
1554 Duputel, Nicolas Devau, Michel Brossard, Benoît Jaillard, Davey L. Jones, Philippe
1555 Hinsinger and Frédéric Gérard (2013)." Applied Geochemistry 46: 85-89.

1556 Hai, T. N. (2017). "Comments on “Effect of Temperature on the Adsorption of Methylene Blue
74
 ACCEPTED MANUSCRIPT

1557 Dye onto Sulfuric Acid–Treated Orange Peel”." Chemical Engineering

1558 Communications 204(1): 134-139.

1559 Halder, G., A. A. Khan and S. Dhawane (2016). "Fluoride Sorption Onto a Steam-Activated
1560 Biochar Derived From Cocos nucifera Shell." CLEAN – Soil, Air, Water 44(2):

PT
1561 124-133.

1562 Hall, K. R., L. C. Eagleton, A. Acrivos and T. Vermeulen (1966). "Pore- and Solid-Diffusion

RI
1563 Kinetics in Fixed-Bed Adsorption under Constant-Pattern Conditions." Industrial &
1564 Engineering Chemistry Fundamentals 5(2): 212-223.

SC
1565 Hamdaoui, O. and E. Naffrechoux (2007). "Modeling of adsorption isotherms of phenol and
1566 chlorophenols onto granular activated carbon: Part II. Models with more than two
1567 parameters." Journal of Hazardous Materials 147(1–2): 401-411.

U
1568 Han, R., Y. Wang, W. Zou, Y. Wang and J. Shi (2007). "Comparison of linear and nonlinear
1569
AN
analysis in estimating the Thomas model parameters for methylene blue adsorption
1570 onto natural zeolite in fixed-bed column." Journal of Hazardous Materials 145(1–2):
M
1571 331-335.

1572 Harsha, N., K. V. S. Krishna, N. K. Renuka and S. Shukla (2015). "Facile synthesis of
D

1573 [gamma]-Fe2O3 nanoparticles integrated H2Ti3O7 nanotubes structure as a

TE

1574 magnetically recyclable dye-removal catalyst." RSC Advances 5(38): 30354-30362.

1575 Ho, Y.-S. (1995). "Adsorption of heavy metals from waste streams by peat." Ph.D. Thesis, The
EP

1576 University of Birmingham, Birmingham, U.K.

1577 Ho, Y.-S. (2004a). "Erratum to “Removal of copper ions from aqueous solution by tree fern”:
C

1578 [Water Res. 37 (2003) 2323–2330]." Water Research 38(20): 4536-4537.

AC

1579 Ho, Y.-S. (2004b). "Selection of optimum sorption isotherm." Carbon 42(10): 2115-2116.

1580 Ho, Y.-S. (2005). "Comment on “Two-stage batch sorber design using second-order kinetic
1581 model for the sorption of metal complex dyes onto pine sawdust” by Özacar, M. and
1582 Şengýl, İ.A." Biochemical Engineering Journal 23(3): 291-292.

1583 Ho, Y.-S. (2013). "Comments on “Removal of zirconium(IV) from aqueous solution by Coriolus

75
 ACCEPTED MANUSCRIPT

1584 versicolor: Equilibrium and thermodynamic study”." Ecological Engineering 58: 1-2.

1585 Ho, Y.-S. (2014a). "Comments on “Adsorption of 2-mercaptobenzothiazole from aqueous

1586 solution by organo-bentonite” by P. Jing, M.H. Hou, P. Zhao, X.Y. Tang, H.F. Wan."
1587 Journal of Environmental Sciences 26(12): 2571-2572.

PT
1588 Ho, Y.-S. (2014b). "Comment on “New Calix[4]arene Appended Amberlite XAD-4 Resin with
1589 Versatile Perchlorate Removal Efficiency”." Journal of Chemical & Engineering Data

RI
1590 59(6): 2131-2131.

1591 Ho, Y.-S. (2014c). "Comments on “Adsorption characteristics and behaviors of graphene oxide

SC
1592 for Zn(II) removal from aqueous solution”." Applied Surface Science 301: 584.

1593 Ho, Y.-S. (2014d). "Comments on “Simultaneous adsorption of aniline and Cr(VI) Ion by

U
1594 activated carbon/chitosan composite”." Journal of Applied Polymer Science 131(22):
1595 n/a-n/a.

1596
AN
Ho, Y.-S. (2014e). "Comments on the paper “Removal of Cu (II) by loofah fibers as a natural and
1597 low-cost adsorbent from aqueous solutions”." Journal of Molecular Liquids 198: 322.
M
1598 Ho, Y.-S. (2014f). "The real pseudo-second-order rate equation." Industrial Crops and Products
1599 52: 17.
D

1600 Ho, Y.-S. (2014g). "Using of “pseudo-second-order model” in adsorption." Environmental

TE

1601 Science and Pollution Research 21(11): 7234-7235.

1602 Ho, Y.-S. (2015). "Comment on “Genetic characterization, nickel tolerance, biosorption, kinetics,
EP

1603 and uptake mechanism of a bacterium isolated from electroplating industrial

1604 effluent”." Canadian Journal of Microbiology 61(11): 881-882.
C

1605 Ho, Y.-S. (2016a). "Comments on the paper ‘Synthesis and application of ion-imprinted resin
AC

1606 based on modified melamine-thiourea for selective removal of Hg(II)’." Polymer

1607 International 65(7): 845-846.

1608 Ho, Y.-S. (2016b). "Comments on the paper “Removal of aqueous Hg(II) and Cr(VI) using
1609 phytic acid doped polyaniline/cellulose acetate composite membrane”." Journal of
1610 Hazardous Materials 311: 273-274.

76
 ACCEPTED MANUSCRIPT

1611 Ho, Y.-S. (2016c). "Comments on the paper “Simultaneous removal of mixed contaminants by
1612 organoclays — Amoxicillin and Cu(II) from aqueous solution”." Applied Clay Science
1613 126: 297-298.

1614 Ho, Y.-S. (2016d). "Comments on using of “pseudo-first-order model” [J. Taiwan Inst. Chem.

PT
1615 Eng. Vol. 59]." Journal of the Taiwan Institute of Chemical Engineers 68: 136-137.

1616 Ho, Y.-S. (2016e). "Comments on using of “pseudo-first-order model” in adsorption [Int. J. Biol.

RI
1617 Macromol., vol. 81]." International Journal of Biological Macromolecules 88:
1618 505-506.

SC
1619 Ho, Y.-S. (2017). "Comments on the paper “Application of Mn/MCM-41 as an adsorbent to
1620 remove Methyl Blue from aqueous solution”." Journal of Colloid and Interface
1621 Science.

U
1622 Ho, Y.-S., W.-T. Chiu and C.-C. Wang (2005). "Regression analysis for the sorption isotherms of
1623
AN
basic dyes on sugarcane dust." Bioresource Technology 96(11): 1285-1291.

1624 Ho, Y. S. and G. McKay (1998a). "A Comparison of Chemisorption Kinetic Models Applied to
M
1625 Pollutant Removal on Various Sorbents." Process Safety and Environmental Protection
1626 76(4): 332-340.
D

1627 Ho, Y. S. and G. McKay (1998b). "Sorption of dye from aqueous solution by peat." Chemical
TE

1628 Engineering Journal 70(2): 115-124.

1629 Ho, Y. S., D. A. J. Wase and C. F. Forster (1996). "Kinetic Studies of Competitive Heavy Metal
EP

1630 Adsorption by Sphagnum Moss Peat." Environmental Technology 17(1): 71-77.

1631 Huang, R., B. Yang, Q. Liu and Y. Liu (2014). "Simultaneous adsorption of aniline and Cr(VI)
C

1632 ion by activated carbon/chitosan composite." Journal of Applied Polymer Science

AC

1633 131(4): n/a-n/a.

1634 Hung, H.-W. and T.-F. Lin (2006). "Adsorption of MTBE from contaminated water by
1635 carbonaceous resins and mordenite zeolite." Journal of Hazardous Materials 135(1–3):
1636 210-217.

1637 Inbaraj, B. S. (2006). "Comment on “Adsorption of Reactive Dyes from a Textile Effluent Using

77
 ACCEPTED MANUSCRIPT

1638 Sawdust as the Adsorbent”." Industrial & Engineering Chemistry Research 45(21):
1639 7362-7362.

1640 Inglezakis, V. J. and A. A. Zorpas (2012). "Heat of adsorption, adsorption energy and activation
1641 energy in adsorption and ion exchange systems." Desalination and Water Treatment

PT
1642 39(1-3): 149-157.

1643 Ji, L., W. Chen, L. Duan and D. Zhu (2009). "Mechanisms for strong adsorption of tetracycline

RI
1644 to carbon nanotubes: A comparative study using activated carbon and graphite as
1645 adsorbents." Environmental Science & Technology 43(7): 2322-2327.

SC
1646 Jiang, J., S.-Y. Pang and J. Ma (2009). "Comment on “Adsorption of Hydroxyl- and
1647 Amino-Substituted Aromatics to Carbon Nanotubes”." Environmental Science &
1648 Technology 43(9): 3398-3399.

U
1649 Jing, P., M. Hou, P. Zhao, X. Tang and H. Wan (2014). "Reply to comments on “Adsorption of
1650
AN
2-mercaptobenzothiazole from aqueous solution by organo-bentonite” by Yuhshan
1651 Ho." Journal of Environmental Sciences 26(12): 2573-2574.
M
1652 Kamagaté, M., J. Lützenkirchen, F. Huber and K. Hanna (2016). "Comment on “Competitive
1653 Adsorption of Cd(II), Cr(VI), and Pb(II) onto Nanomaghemite: A Spectroscopic and
D

1654 Modeling Approach”." Environmental Science & Technology 50(3): 1632-1633.

TE

1655 Karadag, D., M. Turan, E. Akgul, S. Tok and A. Faki (2007). "Adsorption Equilibrium and
1656 Kinetics of Reactive Black 5 and Reactive Red 239 in Aqueous Solution onto
EP

1657 Surfactant-Modified Zeolite." Journal of Chemical & Engineering Data 52(5):

1658 1615-1620.
C

1659 Khan, A. A. and R. P. Singh (1987). "Adsorption thermodynamics of carbofuran on Sn (IV)

1660 arsenosilicate in H+, Na+ and Ca2+ forms." Colloids and Surfaces 24(1): 33-42.
AC

1661 Komárek, M., C. M. Koretsky, K. J. Stephen, D. S. Alessi and V. Chrastný (2016). "Response to
1662 Comment on “Competitive Adsorption of Cd(II), Cr(VI), and Pb(II) onto
1663 Nanomaghemite: A Spectroscopic and Modeling Approach”." Environmental Science
1664 & Technology 50(3): 1634-1635.

1665 Kim, Y., J. Bae, H. Park, J.-K. Suh, Y.-W. You and H. Choi (2016). "Adsorption dynamics of
78
 ACCEPTED MANUSCRIPT

1666 methyl violet onto granulated mesoporous carbon: Facile synthesis and adsorption
1667 kinetics." Water Research 101: 187-194.
1668 Kosmulski, M. (2009). Surface charging and points of zero charge, CRC Press.

1669 Kratochvil, D. and B. Volesky (1998). "Advances in the biosorption of heavy metals." Trends in

PT
1670 Biotechnology 16(7): 291-300.

1671 Kumar, K. V. (2006a). "Comments on “Adsorption of acid dye onto organobentonite”." Journal

RI
1672 of hazardous materials 137(1): 638-639.

1673 Kumar, K. V. (2006b). "Comments on “Equilibrium studies for the adsorption of Acid dye onto

SC
1674 modified hectorite”." Journal of Hazardous Materials 137(2): 1252-1253.

1675 Kumar, K. V. (2006c). "A note on the comments by Dr. Y.S. Ho on “Nitrate removal from

U
1676 aqueous solution by adsorption onto various materials”." Journal of Hazardous
1677

1678
Materials 136(3): 995-996.
AN
Kumar, K. V. (2006d). "A note on the comments by Dr. Y.S. Ho on “Remediation of soil
1679 contaminated with the heavy metal (Cd2+)”." Journal of Hazardous Materials 136(3):
M
1680 993-994.
D

1681 Kumar, K. V. (2007). "Optimum sorption isotherm by linear and non-linear methods for
1682 malachite green onto lemon peel." Dyes and Pigments 74(3): 595-597.
TE

1683 Kumar, K. V. and V. T. Fávere (2006). "Reply to Comments on ‘Chitosan functionalized with
1684 2[-bis-(pyridylmethyl) aminomethyl]4-methyl-6-formyl-phenol: Equilibrium and
EP

1685 kinetics of copper(II) adsorption’ by Yuh-Shan Ho: Discussion on pseudo second order
1686 kinetic expression." Polymer 47(6): 1772-1773.
C

1687 Kumar, K. V. and B. Guha (2006). "Reply to the comments on “Study on biosorption of Cr (VI)
AC

1688 by Mucor hiemalis” by Y.-S. Ho, Biochem. Eng. J. 26 (2005) 82–83." Biochemical
1689 Engineering Journal 30(2): 222-223.

1690 Kumar, K. V. and K. Porkodi (2006). "Relation between some two- and three-parameter isotherm
1691 models for the sorption of methylene blue onto lemon peel." Journal of Hazardous
1692 Materials 138(3): 633-635.

79
 ACCEPTED MANUSCRIPT

1693 Kumar, K. V. and K. Porkodi (2007a). "Comments on “adsorption of 4-chlorophenol from

1694 aqueous solutions by xad-4 resin: Isotherm, kinetic, and thermodynamic analysis”."
1695 Journal of Hazardous Materials 143(1–2): 598-599.

1696 Kumar, K. V. and K. Porkodi (2007b). "Comments on “Biosorption of nickel from protonated

PT
1697 rice bran”." Journal of Hazardous Materials 147(1–2): 679.

1698 Kumar, K. V. and K. Porkodi (2008). "Comments on “Removal of Congo red from aqueous

RI
1699 solution by anilinepropylsilica xerogel” by Pavan FA, Dias SLP, Lima EC, Benvenutti
1700 EV. Dyes and Pigments 2008;76:64–9." Dyes and Pigments 77(2): 481-482.

SC
1701 Kumar, K. V., K. Porkodi and F. Rocha (2007). "Comments on “Removal of lead from aqueous
1702 solution using Syzygium cumini L.: Equilibrium and kinetic studies”." Journal of
1703 Hazardous Materials 147(1–2): 677-678.

U
1704 Lagergren, S. (1898). "About the theory of so-called adsorption of soluble substances." Kungliga
1705
AN
Svenska Vetenskapsakademiens Handlingar 24(4): 1-39.

1706 Langmuir, I. (1918). "THE ADSORPTION OF GASES ON PLANE SURFACES OF GLASS,

M
1707 MICA AND PLATINUM." Journal of the American Chemical Society 40(9):
1708 1361-1403.
D

1709 Lee, D.-G. (2017). "Comment on: Olushola S, Olalekan S, Folahan A, Leslie F and Ximba BJ
TE

1710 (2014) Application of nano zinc oxide (nZnO) for the removal of triphenyltin chloride
1711 (TPT) from dockyard wastewater." Water SA (Online) 43(1): 175.
EP

1712 Li, W. and G. Pan (2007). "Comments on “Ion adsorption components in liquid/solid systems”."
1713 Journal of Environmental Sciences 19(6): 768.
C

1714 Li, Z., S. Imaizumi, T. Katsumi, X. Tang and Y. Chen (2010). "Comment on JHM 142 (2007) 1–
AC

1715 53 ‘Arsenic removal from water–wastewater using adsorbents—A critical review’ by

1716 D Mohan and CU Pittman Jr." Journal of Hazardous Materials 175(1–3): 1116-1117.

1717 Lima, É. C., M. A. Adebayo and F. M. Machado (2015). Kinetic and equilibrium models of
1718 adsorption. Carbon Nanomaterials as Adsorbents for Environmental and Biological
1719 Applications, Springer: 33-69.

80
 ACCEPTED MANUSCRIPT

1720 Lima, E. C., A. R. Cestari and M. A. Adebayo (2016). "Comments on the paper: a critical review
1721 of the applicability of Avrami fractional kinetic equation in adsorption-based water
1722 treatment studies." Desalination and Water Treatment 57(41): 19566-19571.

1723 Lin, J. X. and L. Wang (2009). "Comment on “Adsorption of Supranol Yellow 4 GL from

PT
1724 aqueous solution by surfactant-treated aluminum/chromium-intercalated bentonite”."
1725 Journal of Hazardous Materials 172(1): 516-517.

RI
1726 Liu, Y. (2009). "Is the Free Energy Change of Adsorption Correctly Calculated?" Journal of
1727 Chemical & Engineering Data 54(7): 1981-1985.

SC
1728 Lu, X. (2008). "Comment on “Thermodynamic and isotherm studies of the biosorption of Cu(II),
1729 Pb(II), and Zn(II) by leaves of saltbush (Atriplex canescens)”." The Journal of
1730 Chemical Thermodynamics 40(4): 739-740.

U
1731 Marsh, H. and F. R. Reinoso (2006). Activated carbon, Elsevier.

1732
AN
Masson, S., M. Gineys, S. Delpeux-Ouldriane, L. Reinert, S. Guittonneau, F. Béguin and L.
1733 Duclaux (2016). "Single, binary, and mixture adsorption of nine organic contaminants
M
1734 onto a microporous and a microporous/mesoporous activated carbon cloth."
1735 Microporous and Mesoporous Materials 234: 24-34.
D

1736 McKay, G., Y. S. Ho and J. C. Y. Ng (1999). "Biosorption of Copper from Waste Waters: A
TE

1737 Review." Separation & Purification Reviews 28(1): 87-125.

1738 McLintock, I. (1967). "The Elovich equation in chemisorption kinetics." Nature 216: 1204-1205.
EP

1739 Medellin-Castillo, N. A., E. Padilla-Ortega, M. C. Regules-Martínez, R. Leyva-Ramos, R.

1740 Ocampo-Pérez and C. Carranza-Alvarez (2017). "Single and competitive adsorption of
C

1741 Cd(II) and Pb(II) ions from aqueous solutions onto industrial chili seeds (Capsicum
AC

1742 annuum) waste." Sustainable Environment Research.

1743 Milonjić, S. K. (2007). "A consideration of the correct calculation of thermodynamic parameters
1744 of adsorption." Journal of the Serbian chemical society 72(12): 1363-1367.

1745 Milonjić, S. K. (2009a). "Comments on “factors influencing the removal of divalent cations by
1746 hydroxyapatite”." Journal of Hazardous Materials 162(2–3): 1588-1589.

81
 ACCEPTED MANUSCRIPT

1747 Milonjić, S. K. (2009b). "Comments on “removal of uranium (VI) from aqueous solution by
1748 adsorption of hematite”, by X. Shuibo, Z. Chun, Z. Xinghuo, Y. Jing, Z. Xiaojian, W.
1749 Jingsong." Journal of Environmental Radioactivity 100(10): 921-922.

1750 Milonjić, S. K. (2010). "Comments on the authors’ response to the comments on “Factors

PT
1751 influencing the removal of divalent cations by hydroxyapatite”, by Smiciklas et al."
1752 Journal of Hazardous Materials 176(1–3): 1126-1127.

RI
1753 Milonjić, S. K., A. L. Ruvarac and M. V. Šušić (1975). "The heat of immersion of natural
1754 magnetite in aqueous solutions." Thermochimica Acta 11(3): 261-266.

SC
1755 Mohan, D. and C. U. Pittman Jr (2011). "Reply to the comments on HAZMAT 142 (2007) 1–53
1756 ‘Arsenic removal from water/wastewater using adsorbents – A critical review’ by D.
1757 Mohan and C.U. Pittman Jr. made by Zhenze Li et al. [HAZMAT 175 (2010) 1116–

U
1758 1117]." Journal of Hazardous Materials 185(2–3): 1614-1617.

1759
AN
Müller, G., K. Janošková, T. Bakalár, J. Cakl and H. Jiránková (2012). "Removal of Zn(II) from
1760 aqueous solutions using Lewatit S1468." Desalination and Water Treatment 37(1-3):
M
1761 146-151.

1762 Müller, F., T. Faidt, S. Grandthyll and H. Natter (2017). "Comment on “CuYb0.5Fe1.5O4
D

1763 nanoferrite adsorbent structural, morphological and functionalization characteristics

TE

1764 for multiple pollutant removal by response surface methodology” by M.A. Rehman et
1765 al., J. Mol. Liq. 224 (2016) 1256–1265." Journal of Molecular Liquids.
EP

1766 Nekouei, F. and S. Nekouei (2017). "Comments on the paper “Adsorptive removal of methylene
1767 blue by rhamnolipid-functionalized graphene oxide from wastewater”." Water Research
1768 108: 462-463.
C

1769 Nekouei, F., H. Kargarzadeh, S. Nekouei, I. Tyagi, S. Agarwal and V. Kumar Gupta (2016).
AC

1770 "Preparation of Nickel hydroxide nanoplates modified activated carbon for Malachite
1771 Green removal from solutions: Kinetic, thermodynamic, isotherm and antibacterial
1772 studies." Process Safety and Environmental Protection 102: 85-97.

1773 Oleszczuk, P., B. Pan and B. Xing (2010). "Response to Comment on “Adsorption and
1774 Desorption of Oxytetracycline and Carbamazepine by Multiwalled Carbon
82
 ACCEPTED MANUSCRIPT

1775 Nanotubes”." Environmental Science & Technology 44(12): 4829-4829.

1776 Onjia, A. E. and S. K. Milonjić (2002). Influence of the background electrolyte on the point of
1777 zero charge of chromium(III)-oxide. Materials Science Forum. 413: 87-92.

1778 Özacar, M. (2005). "Response to comment on “Two-stage batch sorber design using

PT
1779 second-order kinetic model for the sorption of metal complex dyes onto pine sawdust”
1780 by Dr. Y.-S. Ho, Biochem. Eng. J. 21 (2004) 39–45." Biochemical Engineering Journal

RI
1781 23(3): 293.

1782 Park, D., Y.-S. Yun, J. Y. Kim and J. M. Park (2008). "How to study Cr(VI) biosorption: Use of

SC
1783 fermentation waste for detoxifying Cr(VI) in aqueous solution." Chemical Engineering
1784 Journal 136(2–3): 173-179.

U
1785 Park, D., Y.-S. Yun and J. M. Park (2005). "Studies on hexavalent chromium biosorption by
1786 chemically-treated biomass of Ecklonia sp." Chemosphere 60(10): 1356-1364.

1787
AN
Park, D., Y.-S. Yun and J. M. Park (2006a). "Comment on “Chromate ion adsorption by
1788 agricultural by-products modified with dimethyloldihydroxylethylene urea and choline
M
1789 chloride” by Wartelle and Marshall." Water Research 40(7): 1501-1504.

1790 Park, D., Y.-S. Yun and J. M. Park (2006b). "Comment on the Removal Mechanism of
D

1791 Hexavalent Chromium by Biomaterials or Biomaterial-Based Activated Carbons."

TE

1792 Industrial & Engineering Chemistry Research 45(7): 2405-2407.

1793 Park, D., Y.-S. Yun and J. M. Park (2006c). "Mechanisms of the removal of hexavalent
EP

1794 chromium by biomaterials or biomaterial-based activated carbons." Journal of

1795 Hazardous Materials 137(2): 1254-1257.
C

1796 Patrickios, C. S. and E. N. Yamasaki (1997). "A correction to the calculation of the Gibbs free
AC

1797 energy of adsorption for biomolecules in ion-exchange systems." Biophysical

1798 Chemistry 69(2–3): 219-220.

1799 Plazinski, W., W. Rudzinski and A. Plazinska (2009). "Theoretical models of sorption kinetics
1800 including a surface reaction mechanism: A review." Advances in Colloid and Interface
1801 Science 152(1–2): 2-13.

83
 ACCEPTED MANUSCRIPT

1802 Porter, W. R. (1985). "Comment on: "Equilibrium adsorption of polycyclic aromatic

1803 hydrocarbons from water onto activated carbon"." Environmental Science &
1804 Technology 19(9): 869-870.

1805 Radovic, L. R. (1999). "Surface chemistry of activated carbon materials: state of the art and

PT
1806 implications for adsorption." Surfactant science series 78: 529-565.

1807 Rahmani-Sani, A., R.-r. Shan, L.-g. Yan and A. Hosseini-Bandegharaei (2017). "Response to

RI
1808 “Letter to Editor: Minor correction to the thermodynamic calculation using the
1809 distribution constant by Shan et al. and Rahmani-Sani et al.”." Journal of Hazardous

SC
1810 Materials 325(5).

1811 Rayne, S. (2013). "Comment on “Adsorption of ionizable organic contaminants on multi-walled

1812 carbon nanotubes with different oxygen contents [Li et al., J. Hazard. Mater. 186 (2011)

U
1813 407–415]”." Journal of Hazardous Materials 252–253: 1.

1814
AN
Redlich, O. and D. L. Peterson (1959). "A Useful Adsorption Isotherm." The Journal of Physical
1815 Chemistry 63(6): 1024-1024.
M
1816 Rivera-Utrilla, J. and M. Sánchez-Polo (2011). "Adsorbent-adsorbate interactions in the
1817 adsorption of organic and inorganic species on ozonized activated carbons:
D

1818 a short review." Adsorption 17(3): 611-620.

TE

1819 Robalds, A., G. M. Naja and M. Klavins (2016). "Highlighting inconsistencies regarding metal
1820 biosorption." Journal of Hazardous Materials 304: 553-556.
EP

1821 Roginsky, S. and Y. B. Zeldovich (1934). "The catalytic oxidation of carbon monoxide on
1822 manganese dioxide." Acta Phys. Chem. USSR 1: 554.
C

1823 Rudzinski, W. and W. Plazinski (2007). "Reply to “Comment on “Kinetics of Solute Adsorption
AC

1824 at Solid/Solution Interfaces:  A Theoretical Development of the Empirical Pseudo-First

1825 and Pseudo-Second Order Kinetic Rate Equations, Based on Applying the Statistical
1826 Rate Theory of Interfacial Transport”." The Journal of Physical Chemistry B 111(1):
1827 319-319.

1828 Salvestrini, S., V. Leone, P. Iovino, S. Canzano and S. Capasso (2014). "Considerations about the
1829 correct evaluation of sorption thermodynamic parameters from equilibrium
84
 ACCEPTED MANUSCRIPT

1830 isotherms." The Journal of Chemical Thermodynamics 68: 310-316.

1831 Senthil Kumar, P., P. S. A. Fernando, R. T. Ahmed, R. Srinath, M. Priyadharshini, A. M. Vignesh

1832 and A. Thanjiappan (2014). "Effect of temperature on the adsorption of methylene blue
1833 dye onto sulfuric acid–treated orange peel " Chemical Engineering Communications

PT
1834 201(11): 1526-1547.

1835 Sharma, P., N. Hussain, D. J. Borah and M. R. Das (2013). "Kinetics and Adsorption Behavior of

RI
1836 the Methyl Blue at the Graphene Oxide/Reduced Graphene Oxide Nanosheet–Water
1837 Interface: A Comparative Study." Journal of Chemical & Engineering Data 58(12):

SC
1838 3477-3488.

1839 Sharma, Y. C., G. S. Gupta, G. Prasad and D. C. Rupainwar (1990). "Use of wollastonite in the
1840 removal of Ni(II) from aqueous solutions." Water, Air, and Soil Pollution 49(1): 69-79.

U
1841 Smičiklas, I., A. Onjia, S. Raičević and D. Janaćković (2009). "Authors’ response to comments
1842
AN
on “Factors influencing the removal of divalent cations by hydroxyapatite”." Journal
1843 of Hazardous Materials 168(1): 560-562.
M
1844 Somasundaran, P. (1968). "Zeta potential of apatite in aqueous solutions and its change during
1845 equilibration." Journal of Colloid and Interface Science 27(4): 659-666.
D

1846 Swiatkowski, A., M. Pakula, S. Biniak and M. Walczyk (2004). "Influence of the surface
TE

1847 chemistry of modified activated carbon on its electrochemical behaviour in the

1848 presence of lead(II) ions." Carbon 42(15): 3057-3069.
EP

1849 Thrower, P. (2011). "Writing a scientific paper V – Language." Carbon 49(15): 4957-4960.

1850 Thrower, P. A. (2007). "Writing a Scientific Paper: I. Titles and Abstracts." Carbon 45(11):
C

1851 2143-2144.
AC

1852 Thrower, P. A. (2008a). "Writing a scientific paper: II. Introduction and references." Carbon
1853 46(2): 183-184.

1854 Thrower, P. A. (2008b). "Writing a scientific paper: III. Experimental." Carbon 46(8): 1113-1114.

1855 Thrower, P. A. (2010). "Writing a scientific paper: IV. Results and discussion." Carbon 48(10):
1856 2675-2676.

85
 ACCEPTED MANUSCRIPT

1857 Tien, C. (2007). "Remarks on adsorption manuscripts received and declined: An editorial."
1858 Separation and Purification Technology 54(3): 277-278.

1859 Tien, C. (2008). "Remarks on adsorption manuscripts revised and declined: An editorial." Journal
1860 of Hazardous Materials 150(1): 2-3.

PT
1861 Tran, H. N. (2017). "Comments on “Characterization and adsorption capacity of raw
1862 pomegranate peel biosorbent for copper removal”." Journal of Cleaner Production 144:

RI
1863 553-558.

1864 Tran, H. N., F.-C. Huang, C.-K. Lee and H.-P. Chao (2017a). Activated carbon derived from

SC
1865 spherical hydrochar functionalized with Triethylenetetramine: synthesis,
1866 characterizations, and adsorption application. Green Processing and Synthesis.

U
1867 Tran, H. N., Y.-F. Wang, S.-J. You and H.-P. Chao (2017b). "Insights into the mechanism of
1868 cationic dye adsorption on activated charcoal: the importance of π-π interactions."
1869
AN
Process Safety and Environmental Protection.

1870 Tran, H. N., S.-J. You and H.-P. Chao (2015). "Effect of pyrolysis temperatures and times on the
M
1871 adsorption of cadmium onto orange peel derived biochar." Waste Management &
1872 Research.
D

1873 Tran, H. N., S.-J. You and H.-P. Chao (2016). "Thermodynamic parameters of cadmium
TE

1874 adsorption onto orange peel calculated from various methods: A comparison study."
1875 Journal of Environmental Chemical Engineering 4(3): 2671-2682.
EP

1876 Tran, H. N., S.-J. You and H.-P. Chao (2017c). "Activated Carbons from Golden Shower upon
1877 Different Chemical Activation Methods: Synthesis and Characterizations." Adsorption
C

1878 Science & Technology.

AC

1879 Tran, H. N., S.-J. You and H.-P. Chao (2017d). "Fast and efficient adsorption of methylene green
1880 5 on activated carbon prepared from new chemical activation method." Journal of
1881 Environmental Management 188: 322-336.

1882 Tran, H. N., S.-J. You and H.-P. Chao (2017e). "Insight into adsorption mechanism of cationic
1883 dye onto agricultural residues-derived hyrochars: negligible role of π-π interaction."
1884 Korean Journal of Chemical Engineering.
86
 ACCEPTED MANUSCRIPT

1885 Uchimiya, M., I. M. Lima, K. Thomas Klasson, S. Chang, L. H. Wartelle and J. E. Rodgers
1886 (2010). "Immobilization of Heavy Metal Ions (CuII, CdII, NiII, and PbII) by Broiler
1887 Litter-Derived Biochars in Water and Soil." Journal of Agricultural and Food
1888 Chemistry 58(9): 5538-5544.

PT
1889 USEPA (1999). Understanding variation in partition coefficient, Kd, values. United States Office
1890 of Air and Radiation. EPA 402-R-99-004A. Washington, USA, Environmental
1891 Protection Agency. 1.

RI
1892 Utomo, R. P., R. Koswojo, Y.-H. Ju, A. Ayucitra, F. E. Soetaredjo, J. Sunarso and S. Ismadji

SC
1893 (2010). "Reply to the comment on "Acid Green 25 removal from wastewater by
1894 organo-bentonite from Pacitan" by R. Koswojo, R. P. Utomo, Y.-H. Ju, A. Ayucitra, F.
1895 E. Soetaredjo, J. Sunarso, S. Ismadji [Applied Clay Science 48 (2010) 81–86]."

U
1896 Applied Clay Science 50(1): 165-166.

1897 AN
van Noort, P. C. M., M. T. O. Jonker and A. A. Koelmans (2005). "Response to Comment on
1898 “Modeling Maximum Adsorption Capacities of Soot and Soot-like Materials for PAHs
M
1899 and PCBs”." Environmental Science & Technology 39(1): 383-384.

1900 Vasanth Kumar, K., K. Porkodi and F. Rocha (2007). "Comments on “Equilibrium and kinetic
D

1901 studies for the biosorption system of copper(II) ion from aqueous solution using
1902 Tectona grandis L.f. leaves powder”." Journal of Hazardous Materials 146(1–2):
TE

1903 428-429.

1904 Vasanth Kumar, K. and S. Rattanaphani (2007). "Reply to ‘Comments on “An adsorption and
EP

1905 kinetic study of lac dyeing on silk”’ by Yuh-Shan Ho: Discussion on pseudo second
1906 order kinetic expression." Dyes and Pigments 75(1): 253-254.
C

1907 Vasanth Kumar, K. and S. Sivanesan (2006). "Equilibrium data, isotherm parameters and process
AC

1908 design for partial and complete isotherm of methylene blue onto activated carbon."
1909 Journal of Hazardous Materials 134(1–3): 237-244.

1910 Volesky, B. (2003). Sorption and biosorption, BV Sorbex.

1911 Volesky, B. (2007). "Biosorption and me." Water Research 41(18): 4017-4029.

1912 Walter, W. J. (1984). "Evolution of a Technology." Journal of Environmental Engineering 110(5):

87
 ACCEPTED MANUSCRIPT

1913 899-917.

1914 Wang, H., X. Yuan, Y. Wu, H. Huang, G. Zeng, Y. Liu, X. Wang, N. Lin and Y. Qi (2014).
1915 "Corrigendum to “Adsorption characteristics and behaviors of graphene oxide for
1916 Zn(II) removal from aqueous solution” [Appl. Surf. Sci. 279 (2013) 432–440]."

PT
1917 Applied Surface Science 301: 585.

1918 Wang, L. K. (2009). Heavy metals in the environment. Boca Raton, CRC Press.

RI
1919 Weber, T. W. and R. K. Chakravorti (1974). "Pore and solid diffusion models for fixed-bed
1920 adsorbers." AIChE Journal 20(2): 228-238.

SC
1921 Weber, W. J. and J. C. Morris (1963). "Kinetics of adsorption on carbon from solution." Journal
1922 of the Sanitary Engineering Division 89(2): 31-60.

U
1923 Weber, W. J. and E. H. Smith (1987). "Simulation and design models for adsorption processes."
1924
1925
AN
Environmental Science & Technology 21(11): 1040-1050.
Werner, D. and H. K. Karapanagioti (2005). "Comment on “Modeling Maximum Adsorption
1926 Capacities of Soot and Soot-like Materials for PAHs and PCBs”." Environmental
M
1927 Science & Technology 39(1): 381-382.
D

1928 Worch, E. (2012). Adsorption technology in water treatment: fundamentals, processes, and
1929 modeling, Walter de Gruyter.
TE

1930 Wu, C.-H. (2009). "Response to comment on “Adsorption of direct dyes from aqueous solutions
1931 by carbon nanotubes: Determination of equilibrium, kinetics and thermodynamics
EP

1932 parameter” by Dr. Ho." Journal of Colloid and Interface Science 333(1): 413-414.

1933 Wu, F.-C., B.-L. Liu, K.-T. Wu and R.-L. Tseng (2010). "A new linear form analysis of Redlich–
C

1934 Peterson isotherm equation for the adsorptions of dyes." Chemical Engineering
AC

1935 Journal 162(1): 21-27.

1936 Wu, Z., H. Zhong, X. Yuan, H. Wang, L. Wang, X. Chen, G. Zeng and Y. Wu (2017). "Reply for
1937 comment on “Adsorptive removal of methylene blue by rhamnolipid-functionalized
1938 graphene oxide from wastewater”." Water Research 108: 464-465.

1939 Zhao, Y. (2015). "Comment on “Zirconium–Carbon Hybrid Sorbent for Removal of Fluoride

88
 ACCEPTED MANUSCRIPT

1940 from Water: Oxalic Acid Mediated Zr(IV) Assembly and Adsorption Mechanism”."
1941 Environmental Science & Technology 49(19): 11982-11983.

1942 Zhou, Q., C. Xie, W. Gong, N. Xu and W. Zhou (2011). "Comments on the method of using
1943 maximum absorption wavelength to calculate Congo Red solution concentration

PT
1944 published in J. Hazard. Mater." Journal of Hazardous Materials 198: 381-382.

1945 Zhou, X. (2017). "Minor correction to the thermodynamic calculation using the distribution

RI
1946 constant by Shan et al. and Rahmani-Sani et al." Journal of Hazardous Materials 323,
1947 Part B: 735-736.

SC
1948 Zhou, X., H. Liu and J. Hao (2012). "Letters to the Editor." Adsorption Science & Technology
1949 30(7): 647-649.

U
1950 Zhou, X. and X. Zhou (2014). "The unit problem in the thermodynamic calculation of adsorption
1951 using the Langmuir eqaution." Chemical Engineering Communications 201(11):
1952 1459-1467.
AN
1953
M
1954
D
TE
C EP
AC

89
 ACCEPTED MANUSCRIPT

Table captions

Table 1. Relationship between isotherm parameters and isotherm shape

Table 2. Mistakes in the presentation and calculation of the separation factor (RL)

PT
Table 3. Incorrect values of the exponent g in the Redlich–Peterson equation

Table 4. Comparison of incorrect and correct thermodynamic parameters for methylene blue

RI
adsorption onto sulfuric acid-treated orange peel

SC
Table 5. A typical example of correct and incorrect use of a Micromeritics report (data not

published)

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Table 1. Relationship between isotherm parameters and isotherm shape

Freundlich exponent Separation factor Isotherm shapes Remarks

n=0 RL = 0 Irreversible Horizontal
n<1 RL < 1 Favorable Concave
n=1 RL = 1 Linear Linear

PT
n>1 RL > 1 Unfavorable Convex
Note: Information was adapted from Worch (2012). Copyright (2012), with permission from De
Gruyter. RL and n were calculated from equations 52 and 56, respectively.

RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Table 2. Mistakes in the presentation and calculation of the separation factor (RL)

Typical examples Remakes

Langmuir parameters
o The RL is not the Langmuir
Q max KL RL
constant, so it is mistake for
Orange peel 54.5 0.101 0.01–0.17

PT
presentation.
OP Biochar 115 0.019 0.08–0.57
Langmuir parameters The RL values must be calculated
o RL

RI
Q max KL at different range of initial
Orange peel 54.5 0.101 0.17 adsorbate concentration, so it is
OP Biochar 115 0.019 0.57 mistake for calculation.

SC
Langmuir parameters
RL
Qomax KL Correction presentation and
Orange peel 54.5 0.101 0.01–0.17 calculation.

U
OP Biochar 115 0.019 0.08–0.57
AN
Note: The data were modified from our previous publication (Tran, You et al. 2015).
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Table 3. Incorrect values of the exponent g in the Redlich–Peterson equation

Adsorbent Adsorbate g Reference

Sawdust Reactive black 1.10 (Inbaraj, 2006)
Sawdust Reactive red 1.04

PT
(Inbaraj, 2006)
Anilinepropylsilica
Congo red 2.91 (Kumar and Porkodi, 2008)
xerogel

RI
Xad-4 resin 4-chlorophenol at 303 K 1.07 (Kumar and Porkodi, 2007)
Xad-4 resin 4-chlorophenol at 318 K 1.19 (Kumar and Porkodi, 2007)

SC
2+
Tectona grandis Cu –0.35 (Vasanth Kumar et al., 2007)
2+
Syzygium cumini Pb –1.23 (Kumar et al., 2007

U
Activated carbon/
Aniline/aniline + Cr(VI) 1.04 (Huang et al., 2014)
chitoshan composite
Activated carbon/
AN
Cr(VI)/aniline + Cr(VI) 1.04 (Huang et al. 2014)
chitoshan composite
M
Activated carbon
Cr(VI) 1.02 (Huang et al. 2014)
/chitoshan composite
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Table 4. Comparison of incorrect and correct thermodynamic parameters for methylene blue
adsorption onto sulfuric acid-treated orange peel

Incorrect thermodynamic parameters reported by Senthil Kumar et al. (2014)

∆Go (kJ/mol)

PT
Co ∆Ho ∆So
(mg/L) (kJ/mol) (J/mol/K)a 30 °C 40 °C 50 °C 60 °C
50 –53.7 –137 –12.5 –10.4 –8.87 –8.49

RI
100 –37.4 –90.3 –10.1 –9.15 –8.04 –7.46
150 –32.9 –79.4 –8.86 –8.19 –7.27 –6.52

SC
200 –27.5 –64.2 –8.03 –7.54 –6.64 –6.19
250 –23.7 –54.4 –7.29 –6.61 –6.17 –5.64
Co KC values

U
(mg/L) 30 °C 40 °C 50 °C 60 °C
50 142 54.9 27.2 21.5
100
150
AN
54.8
33.7
33.6
23.3
19.9
14.9
14.8
10.5
200 24.2 18.1 11.9 9.35
M
250 18.1 12.7 9.97 7.66
Correct thermodynamic parameters suggested by Hai (2017)
∆Go ∆Ho ∆So
D

Temperature KC van't Hoff equation

(kJ/mol) (kJ/mol) (J/mol)
TE

30 °C 5.25 -4.18 13.65 -31.39

40 °C 4.33 -3.81 y = 1642 – 3.78
50 °C 3.54 -3.40 R2 = 0.978
EP

60 °C 3.27 -3.28
a
The unit of J/mol/K was expressed as the original paper; the correction unit might be kJ/mol.
C
AC
 ACCEPTED MANUSCRIPT

Table 5. A typical example of correct and incorrect use of a Micromeritics report

Incorrect data Correct data

Prepared activated carbon Prepared Biochar
BET surface area 613 536

PT
Langmuir surface area 1684 660
Micropore area NA 348

RI
External surface area 698 188
NA: The micropore area is not reported because either the micropore volume is negative or the

SC
calculation of external surface area is larger than the total surface area (BET surface area). The
unit is m2/g. Data is not published.

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Figure captions

Figure 1. Some basic terms used in adsorption science and technology

Figure 2. Basic properties of an adsorbent determined by various common techniques

PT
Figure 3. Schematic illustration of a new classification system for metal biosorption

mechanisms

RI
Figure 4. A typical example of difference between (a) point of zero charge and (b) zeta

SC
potential of glucose-derived spherical biochar prepared from 800 °C

Figure 5. Effect of contact time on the adsorption process of (a) methyl violet onto granulated

U
mesoporous carbon (GMC) and granular activated carbon (GAC); (b) methyl
AN
tert-butyl ether onto carbonaceous resin (Ambersorb 563); (c) ibuprofen (IBP),

carbamazepine (CBZ), ofloxacin (OFX), bisphenol-A (BPA), diclofenac (DFN),

M
mecoprop (MCP), pentachlorophenol (PCP), benzotriazol (BZT), and caffeine

(CAF) onto activated carbon cloths (resin phenolic precursor); and (d) methyl blue
D

onto graphene oxide

TE

Figure 6. Fits of kinetic adsorption data to the linear pseudo-second-order equation (the study

of methylene green 5 adsorption onto commercial activated carbon (CAC) and

EP

synthesized activated carbons prepared from golden shower through different

chemical activation methods (GSAC, GSBAC, GSHAC, and GSHBAC))

C
AC

Figure 7. Transport processes during adsorption by a porous adsorbent

Figure 8. (a) non-equilibrium adsorption isotherms of methylene blue onto sulfuric

acid-treated orange peel (STOP) at various temperatures, and (b) comparison of

adsorption isotherm curves; adsorbent B performs better (higher qe at q10) than

adsorbent A at lower equilibrium concentrations (e.g., Ce = 10 mg/L)

 ACCEPTED MANUSCRIPT

Figure 9. Proposed mechanisms for Cr(VI) biosorption by nonliving biomass

Figure 10. (a) Chemical structure of Congo Red and (b) UV-vis spectra of Congo Red

solutions at different solution pH values

PT
Figure 11. An example of a graphene layer and proposed mechanisms of methylene

adsorption green 5 onto biochar, synthesized activated carbon, and commercial

RI
activated charcoal

SC
Figure 12. A simple guide to the nonlinear method of calculating the Langmuir parameters

(RSS: residual sum of squares, TSS: total sum of squares)

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D

Figure 1. Some basic terms used in adsorption science and technology

TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M

Figure 2. Basic properties of an adsorbent determined by various common techniques

D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Figure 3. Schematic illustration of a new classification system for metal (bio)sorption mechanisms
M

(Adapted from Robalds et al., 2016. Copyright (2016), with permission from Elsevier)
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE

Figure 4. A typical example of difference between (a) point of zero charge and (b) zeta

potential of glucose-derived spherical biochar prepared from 800 °C (Data unpublished)

C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D

Figure 5. Effect of contact time on the adsorption process of (a) methyl violet onto granulated
TE

mesoporous carbon (GMC) and granular activated carbon (GAC) (Reprinted from Kim et al.
2016. Copyright (2016), with permission from Elsevier); (b) methyl tert-butyl ether onto
EP

carbonaceous resin (Ambersorb 563) (Reprinted from Hung and Lin 2006. Copyright (2006),
with permission from Elsevier); (c) ibuprofen (IBP), carbamazepine (CBZ), ofloxacin (OFX),
bisphenol-A (BPA), diclofenac (DFN), mecoprop (MCP), pentachlorophenol (PCP),
C

benzotriazol (BZT), and caffeine (CAF) onto activated carbon cloths (resin phenolic
AC

precursor) (Reprinted from Masson et al. 2016. Copyright (2016) with permission from
Elsevier); and (d) methyl blue onto graphene oxide (Reprinted with permission from Sharma
et al. 2013. Copyright (2013) American Chemical Society)
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D

Figure 6. Fits of kinetic adsorption data to the linear pseudo-second-order equation (the study
TE

of methylene green 5 adsorption onto commercial activated carbon (CAC) and synthesized

activated carbons prepared from golden shower through different chemical activation
EP

methods (GSAC, GSBAC, GSHAC, and GSHBAC)) (Adapetd from Tran et al., 2017d.

Copyright (2017), with permission from Elsevier)

C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D

Figure 7. Transport processes during adsorption by a porous adsorbent (Adapted with

TE

permission from Weber and Smith, 1987. Copyright (1987) American Chemical Society)
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP

Figure 8. (a) non-equilibrium adsorption isotherms of methylene blue onto sulfuric

AC

acid-treated orange peel (STOP) at various temperatures (Adapted from Hai 2016. Copyright
(2016) Taylor and Francis), and (b) comparison of adsorption isotherm curves: adsorbent B
performs better (higher qe at q10) than adsorbent A at lower equilibrium concentrations (e.g.,
Ce = 10 mg/L) (Adapted from Volesky 2007. Copyright (2007), with permission from
Elsevier)
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
Figure 9. Proposed mechanisms for Cr(VI) biosorption by nonliving biomass
D

(Adapted from Park et al., 2005. Copyright (2005), with permission from Elsevier).
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

Figure 10. (a) Chemical structure of Congo Red and (b) UV-vis spectra of Congo Red

solutions at different solution pH values (Adapted from Zhou et al., 2011. Copyright

(2011), with permission from Elsevier)

 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE

Figure 11. An example of a graphene layer and proposed mechanisms of methylene

adsorption green 5 onto biochar, synthesized activated carbon, and commercial activated
EP

charcoal (Adapetd from Tran et al., 2017 b,d. Copyright (2017), with permission from

Elsevier)
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Figure 12. A simple guide to the nonlinear method of calculating the Langmuir parameters (RSS:

residual sum of squares, TSS: total sum of squares)

M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Highlights

• This study summarizes common literature mistakes in the field of

adsorption

PT
• Reiteration and propagation of such mistakes in future publications

should be avoided.

RI
• Correct expression and citation of the models used in adsorption studies

SC
are provided

• We also highlight some problems that need to be thoroughly discussed or

U
further investigated
AN
M
D
TE
C EP
AC