Journal of Environmental Management 152 (2015) 99e108

Contents lists available at ScienceDirect

Journal of Environmental Management

journal homepage: www.elsevier.com/locate/jenvman

Kinetic, isotherm and thermodynamic studies of amaranth dye

biosorption from aqueous solution onto water hyacinth leaves
Imelda Guerrero-Coronilla, Liliana Morales-Barrera, Eliseo Cristiani-Urbina*

gicas, Instituto Polit

Departamento de Ingeniería Bioquímica, Escuela Nacional de Ciencias Biolo
n de Carpio y Plan de Ayala s/n,
ecnico Nacional, Prolongacio

s, M

Colonia Santo Toma exico DF 11340, Mexico

a r t i c l e i n f o a b s t r a c t

Article history: The present study explored the kinetics, equilibrium and thermodynamics of amaranth (acid red 27)
Received 17 June 2014 anionic dye (AD) biosorption to water hyacinth leaves (LEC). The effect of LEC particle size, contact time,
Received in revised form solution pH, initial AD concentration and temperature on AD biosorption was studied in batch experi-
16 November 2014
ments. AD biosorption increased with rising contact time and initial AD concentration, and with
Accepted 17 January 2015
Available online
decreasing LEC particle size and solution pH. Pseudo-second-order chemical reaction kinetics provided
the best correlation for the experimental data. Isotherm studies showed that the biosorption of AD onto
LEC closely follows the Langmuir isotherm, with a maximum biosorption capacity of about 70 mg g
1.
Keywords:
Amaranth dye
The thermodynamic parameters confirm that AD biosorption by LEC is non-spontaneous and endo-
Biosorption thermic in nature. Results indicate that LEC is a strong biosorbent capable of effective detoxification of
Isotherm AD-laden wastewaters.
Kinetics © 2015 Elsevier Ltd. All rights reserved.
Thermodynamics
Water hyacinth leaves

1. Introduction water and thus difficult to remove from industrial wastewaters by

Amaranth dye (IUPAC name: trisodium (4E)-3-oxo-4-[(4-
sulfonato-1-naphthyl)hydrazono]naphthalene-2,7-disulfonate),
also known as FD&C Red No. 2, E123, C.I. Food Red 9, Azorubin S and
C.I. 16185, is a synthetic azo dye, extensively used in foods and
drinks such as wines, soft drinks, cake mixes, cereals, salad dress-
ings, sweets, caviar and coffee, in order to make them more
appealing (Zhang and Ma, 2013). It is also widely used for coloring
textiles, leather, paper, wood and phenol-formaldehyde resins and
during these processes excess dye enters into the wastewater
(Anjaneya et al., 2013; Guerrero-Coronilla, 2013). If discharged into
surface waters without prior treatment, the dye-colored waste-
waters affect aesthetics and water transparency, and may also block
the penetration of sunlight and oxygen, which is harmful to aquatic
life (Anjaneya et al., 2013). In addition, amaranth dye (AD) can
cause adverse health effects such as tumors, allergy, respiratory
problems, birth defects, cytostaticity, citotoxicity, mutagenicity,
genotoxicity and carcinogenicity (Gupta et al., 2012; Zhang and Ma,
2013; Zhang et al., 2013).
The amaranth anionic dye (AD) is highly soluble and stable in
* Corresponding author.
E-mail address: ecristia@encb.ipn.mx (E. Cristiani-Urbina).
common chemical and physical treatment methods; in addition,
these technologies are not significantly effective or economically
advantageous (Gong et al., 2005; Gupta et al., 2012). Furthermore,
treatment of azo dyes-containing waters with aerobic microbes
does not successfully degrade most dyes, and anaerobic microbial
degradation produces by-products such as aromatic amines, which
are more toxic, mutagenic and carcinogenic than the dyes them-
selves (Ahmad and Kumar, 2011). In this context, biosorption may
be a desirable alternative because its operation is simple, low cost,
reliable and effective and does not produce by-products.
The water hyacinth (Eichhornia crassipes) is a free-floating
aquatic plant of worldwide distribution, which has become an
acute, persistent and expensive environmental problem due to its
extremely rapid proliferation and congested growth (Ibrahim et al.,
2012; Malik, 2007). In addition, water hyacinth rapidly depletes
nutrients and oxygen from water bodies, interferes with navigation,
fishing, shipping, recreation, irrigation and hydropower generation,
favors breeding zones for disease-causing insects, quickens
evapotranspiration and reduces biodiversity, which, in turn, lead to
adverse effects on the environment, flora, fauna, human health and
economic development (Malik, 2007). Nevertheless, the water hy-
acinth is among the most productive plants on Earth (Ibrahim et al.,
2012) and can therefore be used as a valuable biomass resource for

http://dx.doi.org/10.1016/j.jenvman.2015.01.026
0301-4797/© 2015 Elsevier Ltd. All rights reserved.
100 I. Guerrero-Coronilla et al. / Journal of Environmental Management 152 (2015) 99e108

a variety of useful applications. Table 1

A potentially beneficial use for water hyacinth biomass is pre- Kinetic, isotherm and thermodynamic models.

cisely as biosorbent for the removal of toxic pollutants from Kinetic model Equation Nomenclature Reference
aqueous solutions. In this context, a recent study showed that water Pseudo-first- ln(qe1
qt) ¼ lnqe1
k1t qe1, biosorption Ho, 2006.
hyacinth leaves (LEC) have great potential as a simple and inex- order capacity (mg g
1) at
pensive method to remove AD from aqueous solutions. Particularly, equilibrium; qt,
the leaves exhibit much better performance for AD biosorption biosorption capacity
(mg g
1) at time t
than any other of the plant's vegetative organs or the entire aquatic
(h); k1, rate constant
plant. Furthermore, certain E. crassipes proteins play a dominant of the model (h
1)
q2 k t
role in the AD biosorption process (Guerrero-Coronilla, 2013). Pseudo-second- qt ¼ 1
q
e2 2
e2 k2 t
qe2, biosorption Ho, 2006.
The present work aimed to investigate the influence of several order capacity (mg g
1) at
equilibrium; k2, rate
environmental factors such as biosorbent particle size, solution pH,
constant of the
shaking contact time, initial AD concentration and temperature on model (g mg
1 h
1)
AD biosorption from aqueous solutions by LEC. Furthermore, the Elovich qt ¼ B1E lnðAE BE Þ þ B1E lnt AE, initial biosorption Flores-
biosorption mechanism of AD onto LEC was evaluated in terms of rate of Elovich model Garnica
kinetics, equilibrium and thermodynamics. Additionally, the best (mg g
1 h
1); BE, et al.,
desorption constant 2013.
AD desorption solutions were determined, and a study of AD
of Elovich model (g
desorption kinetics was undertaken. mg
1)
Fractional power qt¼KFPtV KFP, fractional power Flores-
2. Materials and methods model constant (mg Garnica
g
1); V, fractional et al.,
power model 2013.
2.1. Biosorbent preparation constant (h
1)
Isotherm models
qm bL ce
Fresh E. crassipes plants were collected from the Xochimilco Two-parameter models qe ¼
1 þ bL ce
channels in Mexico City, and thoroughly washed with distilled Langmuir qe, biosorption Febrianto
1 bL cO capacity (mg g
1) at et al.,
deionized water to remove dirt. The leaves were cut, separated RL ¼ ; q¼
1 þ bL c0 1 þ bL cO equilibrium; c , 2009.
from the plants and then oven-dried at 60  C until dry weight was
e
liquid phase
constant. Afterwards, they were milled using a Glen Creston mill, concentration of
and the resulting particles were screened using ASTM standard adsorbate at
sieves to obtain fractions of different particle size. The sieved equilibrium
(mg L
1); qm,
fractions were stored in airtight plastic containers until used. maximum
biosorption capacity
2.2. Amaranth anionic dye (AD) solutions for biosorption (mg g
1); bL,
experiments Langmuir constant
(L mg
1); RL, Hall
separation factor; co,
AD solutions were obtained by diluting 2 g L
1 of stock AD so- initial adsorbate
lution, which had been prepared by dissolving a weighed amount of concentration
AD (SigmaeAldrich Chemicals) in distilled deionized water. In this (mg L
1); q, surface
work, initial AD concentrations were varied from 10 to 500 mg L
1 1=n
coverage
Freundlich qe ¼ KF ce F KF, Freundlich model Febrianto
and the pH of each AD solution was adjusted to desired value with constant [(mg g
1) et al.,
0.1 M HCl or NaOH solutions. (mg L
1)
1/nF]; nF, 2009.
heterogeneity factor
2.3. Kinetic and equilibrium AD biosorption studies and analytical Temkin qe ¼ RT
BT lnðAT ce Þ R, ideal gas constant Febrianto
(8.314 J mol
1 K
1); et al.,
method
T, absolute 2009.
temperature (K); BT,
Batch kinetic studies were performed in order to investigate the constant related to
effect of LEC particle size, solution pH, initial AD concentration, heat of sorption (J
shaking contact time and temperature on AD biosorption from mol
1); AT, Temkin
isotherm constant (L
aqueous solutions by LEC. All experiments were conducted in mg
1)
500 mL Erlenmeyer flasks containing 120 mL AD solution of known Halsey qe ¼ ðKH =ce ÞnH
1
KH, Halsey isotherm Febrianto
concentration and 1 g (dry weight) L
1 of LEC. Care was taken to model constant (L et al.,
maintain constant pH in each test solution (±0.1 unit) throughout g
1)
1/nH; nH, Halsey 2009.
model exponent
the course of the experiments by periodic checking and adjusting
Dubinin- qe ¼ qm expð
B DR  2 Þ
EDR BDR, biosorption Febrianto
with 0.1 M HCl or NaOH solutions when necessary. Flasks were Radushkevich EDR ¼ RTln 1 þ ce 1 energy constant et al.,
agitated in an orbital shaker (Cole Parmer Inc.) at 140 rpm constant (mol2 J
2); 2009.
shaking speed. EDR, Polanyi potential
The effect of LEC particle size on AD biosorption was assessed in (kJ mol
1)
Three-parameter models 1=n
AD solution at 50 mg L
1 initial dye concentration, pH 2.0, with Sips
S
qe ¼ qm KS ce1=nS KS, affinity coefficient Febrianto
different particle sizes ranging from 0.15e0.18 mm to 1.68e2.0 mm, 1þKS ce
[(mg L
1)
1/ns] et al.,
at 18  C. In order to explore the influence of solution pH levels on ns, heterogeneity 2009.
kinetic performance, different pH values ranging from 1.5 to 7.0 coefficient
Redlich-Peterson qe ¼ KRP ceBRP KRP (L g
1), ARP
were assayed in AD solution at 50 mg L
1 initial dye concentration, 1þARP ce
Febrianto
[(L mg
1)BRP], et al.,
at 18  C. The effect of initial AD concentration on dye biosorption Redlich-Peterson 2009.
was studied by varying the initial dye concentration in the range model isotherm
from 10 to 500 mg L
1. To investigate the influence of temperature
 I. Guerrero-Coronilla et al. / Journal of Environmental Management 152 (2015) 99e108 101

on AD biosorption, experiments were carried out at temperatures
ranging from 18 to 50  C, and three different initial AD concen-
trations (50, 100 and 200 mg L
1) were used. For the equilibrium
biosorption studies, LEC biomass (1 g L
1) was mixed with dye
solutions of different initial AD concentrations (8.9e434 mg L
1) at
18  C, with constant agitation at 140 rpm for 72 h to ensure bio-
sorption equilibrium.
LEC-free controls were run simultaneously and under exactly
the same conditions as used for the AD biosorption experiments in
order to check for glassware sorption and other potential side-
effects such as AD photolysis and precipitation. No measurable
changes in AD concentrations were detected in the LEC-free con-
trols throughout the experiments conducted in this work, which
suggests that the observed AD removal in the LEC experiments was
exclusively due to the biosorbent.
Samples were collected at different experimental times and
centrifuged at 3000 rpm during 5 min. The resulting supernatants
were analyzed by spectrophotometry (Genesys™ 10UVeVisible,
Thermo Electron Scientific Instruments Corporation) at 520 nm
wavelength to determine residual AD concentration (Guerrero-
Coronilla, 2013).
The amount of AD removed at time t by the unit mass (dry
weight) of biosorbent (qt, mg g
1), which represents the AD bio-
sorption capacity, was calculated according to the following mass
balance equation:
2.4. Biosorption kinetics modeling
The dynamics of AD biosorption onto LEC were analyzed using
four different kinetic models: the pseudo-first-order, pseudo-sec-
ond-order, Elovich and fractional power models (Table 1).
2.5. Equilibrium modeling
The equilibrium distribution of AD between the liquid phase and
the LEC biomass was expressed in terms of an AD biosorption
isotherm. Several two-parameter (Langmuir, Freundlich, Temkin,
Halsey and Dubinin-Radushkevich) and three-parameter (Sips,
Redlich-Peterson, Radke-Prausnitz and Toth) biosorption isotherm
models (Table 1) were used to analyze the experimental equilib-
rium data of AD biosorption.
2.6. Biosorption thermodynamics
To describe the thermodynamic behavior of AD biosorption onto
LEC biomass, parameters such as Arrhenius activation energy (EA),
and the changes in activation enthalpy (DH#), entropy (DS#), and
Gibbs free energy (DG#) were calculated (Table 1).

2.7. Desorption studies

ðC0
Ct ÞV After AD biosorption experiments, the AD-loaded LEC was

qt ¼ (1) separated by centrifugation and the residual AD concentration in
W
the supernatant was measured in order to calculate biosorbed dye
where Co is initial AD concentration (mg L
1) at time to ¼ 0 h; Ct is on the biosorbent. Subsequently, AD-loaded LEC was gently washed
residual AD concentration (mg L
1) at time t ¼ t; V is solution with distilled deionized water to remove any unbiosorbed AD and
volume (L), and W is dry weight of LEC (g). oven-dried at 60  C until dry weight became constant. All batch
desorption experiments were conducted in Erlenmeyer flasks
containing an eluent solution and 1 g (dry weight) L
1 of AD-loaded
Table 1 (continued )
LEC. Flasks were agitated in an orbital shaker at 140 rpm and 18  C.
Kinetic model Equation Nomenclature Reference In order to select the best eluents for separating AD from AD-
constants; BRP, loaded LEC, different strong acid (0.1 M HCl, H2SO4 and HNO3),
Redlich-Peterson strong alkaline (0.01 M NaOH and KOH), weak acid (0.1 M NH4Cl),
model exponent weak alkaline (0.1 M Na2CO3), and neutral salt (0.1 M NaCl, NaNO3
(0  R  1.0)
AR RR cBR and Na2SO4) solutions were brought into contact with AD-loaded
Radke-Prausnitz qe ¼ A e
AR (L g
1) and RR Febrianto
R þRR ce
BR
1

(L mg
1), Radke- et al., LEC over a 24 h period. Further batch desorption kinetic studies
Prausnitz model 2009. were carried out using the best eluents. The amount of AD desorbed
constants was determined by using spectrophotometer at 520 nm. Percent-
BR, Radke-Prausnitz age desorption was calculated as the percentage release of AD
model exponent
Toth qe ¼ qm BTo ce
BTo, Toth model Febrianto
initially bound to the biosorbent.
½1þðBTo ce Þ1=nTo nTo
constant (L mg
1)
1/ et al.,
nTo
2009.
nTo, Toth model 2.8. Statistical and data analysis
exponent
Thermodynamic  
models
EA
The AD biosorption and desorption experiments conducted in
RT
Arrhenius k ¼ A0 exp A0, frequency factor Dogan this work were reproducible within 5% error at most, and mean
(g mg
1 h
1); EA, et al., values from three independent replicates are reported herein. AD
activation energy 2009. biosorption and desorption data were statistically analyzed by
(kJ mol
1)
Eyring Pol

anyi
#
ln Tk ¼ ln KhB þ DSR
DH
#
KB, Boltzmann Dogan
analysis of variance (Tukey's multiple comparison test; overall
RT
constant; h, Planck et al., confidence level ¼ 0.05) using GraphPad Prism software version
constant; DS#, 2009. 6.0c (GraphPad Software, Inc.).
activation entropy All the kinetic and isotherm parameters of the biosorption
change
models were evaluated by non-linear regression analysis of the
(kJ mol
1 K
1); DH#,
activation enthalpy experimental data using GraphPad Prism software version 6.0c. The
change (kJ mol
1) determination coefficients (r2) and the standard deviation of re-
Gibbs free DG# ¼ DH#
TDS# DG#, Gibbs free Dogan siduals (Sy.x) were used as a measure of the goodness of fit of the
energy energy change et al., sorption models. Values of r2 close to 1.0 and small Sy.x values
[kJ mol
1] 2009.
indicate better curve fitting.
102 I. Guerrero-Coronilla et al. / Journal of Environmental Management 152 (2015) 99e108

3. Results and discussion

3.1. Effect of LEC particle size on AD biosorption

Fig. 1A shows the effects of LEC particle size on AD biosorption at

solution pH of 2.0. The AD biosorption capacity of LEC increased
with decreasing LEC particle size during the first 5 h of contact time.
This behavior may be attributed to the larger external surface area
for biosorption as LEC particles became smaller.
However, at longer experimental times, the AD biosorption ca-
pacities shown by the different LEC particle sizes tended to
approach the same value (Fig. 1A). Likewise, the AD biosorption
capacity at equilibrium revealed no statistically significant differ-
ences between the assayed LEC particle sizes (p > 0.05), which
indicates that in porous materials such as LEC, the contribution of
the external surface area to the total surface area is limited;
therefore, LEC particle size reduction has a negligible effect on the
total surface area and, consequently, on the equilibrium biosorption
capacity (Pe
rez-Marín et al., 2009). In contrast, the decrease in LEC
particle size drastically reduced the time needed to reach AD bio-
sorption equilibrium (Exp te; Table S1 in Supplementary data), from
72 h for a particle size of 1.68e2.0 mm to 5 h for 0.15e0.30 mm
particle size, which indicates that the AD biosorption rate increased
as LEC particle size decreased. These results agree with those re-
ported elsewhere (Pe
rez-Marín et al., 2009; Schiewer and Patil,
2008). Since the AD biosorption rate was fastest at the smallest
LEC particle size, a LEC particle size of 0.15e0.3 mm was used in
further studies.

3.2. Influence of solution pH levels on AD biosorption

Fig. 1B shows variations in AD biosorption capacity of LEC as a

Fig. 1. Influence of LEC particle size (A), solution pH (B) and initial AD concentration
(C) on AD biosorption capacity of LEC.
function of biosorption time for the assayed solution pH values,
which ranged from 1.5 to 7.0. Biosorption of AD onto LEC showed
strong pH dependence. Virtually no AD biosorption occurred at pH
values of 5.0, 6.0 and 7.0; in contrast, starting at pH of 5.0, AD
biosorption capacity increased with decreasing pH values and
reached its highest level at pH 1.5 and 2.0, with no significant dif-
ference for these two pH values (p > 0.05).
The pH dependence of AD biosorption onto LEC can be
explained on the basis of the different activity of the surface
functional groups of the biosorbent and the dye. At low pH values,
the active LEC biosorption sites are protonated; this generates
electrostatic attraction between the positively charged LEC ligands
and the anionic AD molecules, thus leading to maximum AD bio-
sorption. This is in agreement with the findings of Guerrero-
Coronilla et al. (2014), who reported that proteins play a major
role in the biosorption of AD from aqueous solutions by LEC and
that under acidic conditions, AD bind to proteins as a result of
electrostatic attraction between the negatively charged sulfonic
groups of the dye and positively charged groups (e.g. amide groups)
of proteins. It should be noted that anionic amaranth dye forms
strong complexes with positively charged compounds through
electrostatic interactions (Tunc and Duman, 2007). Contrariwise,
the decrease in AD biosorption capacity observed at higher pH
values responds to a less positive charge of the LEC surface, which
generates stronger electrostatic repulsion towards AD molecules
(Ahmad and Kumar, 2011).
At the optimal pH values of 1.5 and 2.0, AD biosorption capacity
at equilibrium (Exp qe) was approximately 43 mg g
1 and equilib-
rium time (Exp te) was about 5 h (Table S2 in Supplementary data).
The two results were significantly different (p < 0.05) from those
obtained at other pH values for the same parameters. Based on
these results, further studies were conducted at a solution pH of 2.0.
An optimum pH value of 2.0 has also been reported for AD
I. Guerrero-Coronilla et al. / Journal of Environmental Management 152 (2015) 99e108 103

biosorption by bottom ash, de-oiled soya (Mittal et al., 2005),

peanut hull (Gong et al., 2005), hen feathers (Mittal, 2006) and
alumina-reinforced polystyrene (Ahmad and Kumar, 2011).

3.3. Effect of contact time and initial AD concentration on dye

biosorption

Fig. 1C shows the kinetic profiles of AD biosorption by LEC at

different initial AD concentrations ranging from 10 to 500 mg L
1.
Regardless of initial concentration, AD biosorption capacity by LEC
progressively increased as experimental time passed, until bio-
sorption capacity reached a maximum constant value, which cor-
responded to the AD biosorption capacity at equilibrium. Also, the
rate of AD biosorption was fast in the initial stages of the process,
but it gradually slowed down approaching equilibrium.
As initial AD concentration increased from 10 to 200 mg L
1, AD
biosorption capacity increased. This effect may be attributed to
greater availability of AD molecules in solution, which favors the
interactions between AD molecules and LEC. In addition, as initial
AD concentration rises, the dye concentration gradient also in-
creases, which in turn strengthens the thermodynamic driving
force that overcomes the mass transfer resistance of dye molecules
from the aqueous solution to the LEC biomass. As a result, the
probability of collision between AD molecules and active bio-
sorbent sites increases, and, consequently, so does the biosorption
capacity (Flores-Garnica et al., 2013).
The time needed to reach equilibrium (Exp te) increased from
0.17 h to 7 h, as initial AD concentration rose from 10 to 200 mg L
1
(Table S3 in Supplementary data); this may be explained on the
basis that LEC biosorbed larger amounts of AD as initial AD con-
centration increased. In contrast, at initial AD concentrations higher
than 200 mg L
1, the equilibrium biosorption capacities and times
were very similar (p > 0.05), which was probably due to the satu-
ration of the LEC binding sites. The above results also indicate that
saturation of the biosorbent surface is dependent on initial AD
concentration.

3.4. Effect of temperature on AD biosorption

Fig. 2 displays the kinetic profiles of AD biosorption by LEC at

different initial AD concentrations and temperatures. During the
first hours of experimentation, AD biosorption capacity and rate
increased with rising temperature; consequently, the time needed
to reach biosorption equilibrium (Exp te) was shorter at higher
temperatures (Table S4 in Supplementary data). These results
indicate the endothermic nature of AD biosorption onto LEC. The
build-up in AD biosorption rate with rising temperature may be
due to an intensification of the surface activity and an increase in
the kinetic energy of AD molecules (Suazo-Madrid et al., 2011).
Interestingly, at a given initial AD concentration (50, 100 or
200 mg L
1), the effect of temperature on equilibrium AD bio-
sorption capacity was small compared to other influencing factors
such as LEC particle size, solution pH, and initial AD concentration.
These results agree with those reported by other authors who
found that the dependence of equilibrium biosorption capacity of
biosorbents on temperature change can be negligible (Mittal et al.,
2005; Akar et al., 2008).

3.5. Kinetic modeling of AD biosorption process

The pseudo-first-order, pseudo-second-order, Elovich and

Fig. 2. Effect of temperature on AD biosorption capacity of LEC at initial AD concen-

trations of 50 mg L
1 (A), 100 mg L
1 (B) and 200 mg L
1 (C).
104 I. Guerrero-Coronilla et al. / Journal of Environmental Management 152 (2015) 99e108

fractional power models were used in the present work to model concentration (Ce) rose, until it reached a maximum constant value
the kinetic process of AD biosorption onto LEC at different LEC of approximately 69.1 mg g
1. This behavior indicates that as the
particle sizes, solution pH levels, initial AD concentrations and biosorption sites were gradually filled up, it became more difficult
temperatures. for AD molecules to find available vacant sites, which suggests a
At pH values of 5, 6 and 7.0, none of the tested kinetic models progressive saturation of the biosorbent (Limousin et al., 2007). The
showed good fit of the experimental results, probably because at AD biosorption isotherm resembles the type L isotherm of the Giles
these pH values, AD biosorption levels were very low. et al. (1974) classification, which is indicative of high affinity be-
Tables S1eS4 in Supplementary data show the experimental tween LEC and AD molecules, and that no strong competition exists
equilibrium biosorption capacity (Exp qe), the kinetic parameter between the AD molecules and the solvent to occupy the bio-
values of the pseudo-first-order, pseudo-second-order, Elovich and sorption active sites. The L-type isotherm generally reflects the
fractional power models for AD biosorption at LEC particle sizes sorption of a solute monolayer and the occurrence of chemisorp-
ranging from 0.15-0.3 mm to 1.68e2.0 mm, at pH values from 1.5 to tion (Limousin et al., 2007).
4.0, initial AD concentrations from 10 to 500 mg L
1, and temper-
atures from 18 to 50  C, along with the corresponding r2 and Sy.x
values.
The pseudo-second-order model clearly provided the highest r2
and the lowest Sy.x values of the four assayed kinetic models at all
LEC particle sizes, solution pH values, initial AD concentrations and
temperatures. Furthermore, the predicted equilibrium biosorption
capacity (qe2) from the pseudo-second-order model matched
experimental values closely (Exp. qe, Tables S1eS4 in Supplemen-
tary data). Additionally, the pseudo-second-order model
adequately described variations in AD biosorption capacity at the
different contact times, LEC particle sizes, pH levels, initial AD
concentrations and temperatures assayed (continuous lines in
Figs. 1 and 2).
Therefore, the pseudo-second-order model was chosen as the
most suitable to describe AD biosorption kinetics by LEC, which
suggests that a chemisorption mechanism is probably the rate-
controlling step of the overall rate of AD biosorption (Febrianto
et al., 2009). A similar result has been reported for the adsorption
of dyes from aqueous solution by activated carbon cloth (Ayranci
and Duman, 2009).
Moreover, the pseudo-second-order model adequately repre-
sents the experimental kinetic data of AD biosorption onto alumina
reinforced polystyrene (Ahmad and Kumar, 2011), as well as the
biosorption of other anionic azo dyes such as reactive brilliant red
K-2BP onto Aspergillus fumigatus (Wang et al., 2008), reactive yel-
low 42 and reactive red 45 onto Citrus sinensis (Asgher and Bhatti,
2010), brilliant yellow onto hen feathers (Mittal et al., 2012), and
direct brown onto Spirogyra sp. (Venkata-Mohan et al., 2008;
Sivarajasekar et al., 2009), among others.
The present results revealed that the higher the time required to
reach the AD biosorption equilibrium (Exp te) at the assayed LEC
particle sizes, solution pH values, initial AD concentrations and
temperatures (Tables S1eS4 in Supplementary data), the lower the
value of the pseudo-second-order rate constant (k2). This may be
explained by the fact that the k2 constant plays a role as time-
scaling factor, i.e. the k2 value decreases with increasing time
needed to reach an equilibrium state by a biosorption system
(Plazinski et al., 2013).
Results obtained here showed that the k2 constant increased
with a rise in temperature of the biosorption system at all initial AD
concentrations assayed (Table S4 in Supplementary data). This
behavior may be explained by an increase in the interactions be-
tween AD molecules and LEC with rising temperature, and cor-
roborates a faster AD biosorption rate at higher temperatures, as
well as the endothermic nature of the biosorption process (Suazo-
Madrid et al., 2011).

3.6. Biosorption isotherm modeling

The experimental biosorption isotherm obtained for AD removal Fig. 3. Comparison between the experimental biosorption isotherm and the predicted
by LEC at pH 2.0 and 18  C is shown in Fig. 3. Clearly, the equilib- isotherm data derived from two-parameter (A) and three-parameter (B) models for AD
rium capacity (qe) increased gradually as the equilibrium AD biosorption onto LEC.
 I. Guerrero-Coronilla et al. / Journal of Environmental Management 152 (2015) 99e108 105

In this work, different biosorption equilibrium models of two
(Langmuir, Freundlich, Temkin, Halsey and Dubinin-Radushkevich)
and three (Sips, Redlich-Peterson, Radke-Prausnitz, and Toth) pa-
rameters were evaluated to fit the experimental AD biosorption
isotherm. The isotherm parameters of all the above-mentioned
models, along with their corresponding r2 and Sy.x values are
presented in Table 2. The equilibrium data predicted by the
isotherm models are shown in Fig. 3. It is clear from the results that
the Langmuir model and all the three-parameter models tested
showed good fit of the behavior of experimental equilibrium data.
Furthermore, according to the r2 and Sy.x values, neither the
Langmuir model nor the three-parameter models showed any
added advantage (Table 2). However, the exponent value of the Sips
(ns ¼ 0.988), Redlich-Peterson (ВRP ¼ 0.999) and Toth (nTo ¼ 1.026)
models was extremely close to 1.0, for which value these isotherm
models are effectively reduced to the Langmuir model
(Vijayaraghavan et al., 2006). In addition, from a practical point of
view, the Langmuir model is simpler than the three-parameter
models and can consequently be applied and interpreted easier
and is likely to be more helpful. Thus, experimental equilibrium
data for the biosorption of AD by LEC should preferably be fitted
with the Langmuir model, which has practical importance for en-
gineering design and scale-up.
The maximum AD biosorption capacity determined from the
Langmuir isotherm model was 70.61 mg g
1 (Table 2), which is
consistent with the experimental value (69.1 mg g
1). The
maximum experimental biosorption capacity of AD obtained in this
work was significantly higher than that reported in the few studies
published so far on AD sorption, such as those using peanut hull
(Gong et al., 2005), bottom-ash, de-oiled soya (Mittal et al., 2005)
and alumina reinforced polystyrene (Ahmad and Kumar, 2011).
These results indicate that LEC is a strong biosorbent with high
potential for purification of AD-contaminated water and industrial
wastewater.
The Langmuir model predicted an equilibrium constant (bL) of
0.229 L mg
1 (Table 2), and this was used to calculate the Hall
separation factor (RL) and the surface coverage (q) (Table 1). The
first is indicative of the biosorption isotherm shape that predicts
whether a biosorption isotherm is “favorable” or “unfavorable”,
while the latter indicates the fraction of the biosorption sites
occupied by AD at equilibrium.
The separation factor fell from 0.329 to 0.01 as the initial AD
concentration increased from 8.9 to 434 (Fig. 4A), which indicates
that the biosorption of AD onto LEC increased as the initial AD
concentration rose. Furthermore, the RL values are between 0 and 1,
indicating that the AD biosorption by LEC is favorable at all assayed
AD concentrations and confirming the suitability of the biosorbent
for the sorbate (Flores-Garnica et al., 2013). The surface coverage (q)
values were approaching unity (from 0 to 0.99) with increasing

Table 2
Isotherm constants of two- and three-parameter isotherm models for AD bio-
sorption onto LEC.

Two-parameter isotherms Three-parameter isotherms

Langmuir Sips
bL [L mg
1] 0.2291 ± 0.02 KS [(mg L
1)
1/ 0.232 ± 0.029
ns
]
qm [mg g
1] 70.61 ± 1.363 qm [mg g
1] 70.760 ± 1.849
r2 0.991 nS 0.988 ± 0.087
Sy.x 2.459 r2 0.991
Sy.x 2.589
Freundlich
KF [(mg g
1) (mg 26.33 ± 3.84 Redlich-Peterson
L
1)
1/nF]
nF 5.327 ± 0.933 KRP [L g
1] 16.230 ± 1.917
r2 0.859 ARP 0.231 ± 0.050
[(mg L
1)
BRP]
Sy.x 9.773 BRP 0.999 ± 0.021
r2 0.9911
Temkin Sy.x 2.592
AT [L mg
1] 6.389 ± 3.304
BT [J mol
1] 105.5 ± 11.17 Radke-Prausnitz
r2 0.909 AR [L g
1] 16.23 ± 2.034
Sy.x 7.02 RR [L mg
1] 70.36 ± 8.094
BR 0.001 ± 0.02251
Halsey r2 0.988
KH [(L g
1)
1/nH] 2.718E-08 ± 1.08E- Sy.x 2.749
07
nH
0.188 ± 0.035
r2 0.803 Toth
Sy.x 10.3 qm [mg g
1] 70.87 ± 2.027
BTo [(L mg
1)
1/ 0.236 ± 0.04209
nT
]
Dubinin-Radushkevich nTo 1.026 ± 0.1382
qm [mg g
1] 59.92 ± 3.549 r2 0.991
BDR [mol2 J
2] 4.381E- Sy.x 2.586
06 ± 1.275E-06
r2 0.8419
Fig. 4. Dependence of Hall separation factor (A) and surface coverage (B) on initial AD
Sy.x 9.225
concentration.
106 I. Guerrero-Coronilla et al. / Journal of Environmental Management 152 (2015) 99e108

initial AD concentration (Fig. 4B), which indicates that the LEC

surface was almost completely covered by a monomolecular layer
of AD molecules at high AD concentrations. It is also apparent from
Fig. 4B that the surface coverage ceased to vary significantly at
higher AD concentrations and that the reaction rate became almost
independent from the AD concentration. The q values indicated
effective biosorption of AD from aqueous solutions by LEC at all the
initial AD concentrations assayed.

3.7. Biosorption thermodynamics

Thermodynamic parameters such as the Arrhenius activation

energy (EA) as well as the changes in activation enthalpy (DH#),
entropy (DS#) and Gibbs free energy (DG#) were calculated (Table 1)
to evaluate the thermodynamic behavior of AD biosorption onto
the LEC biomass and to gain insight into the AD biosorption
mechanism. The thermodynamic parameters values are summa-
rized in Table 3. The EA values ranged from 19.8 to 34.8 kJ mol
1 for
the tested concentration range of the AD solution, which are within
the interval of EA values (8.4e83.7 kJ mol
1) for chemisorption
processes (Aksu et al., 2008). These results indicate that LEC bio-
sorbs AD molecules by a chemical sorption reaction, which is
consistent with the results from the kinetic and equilibrium studies
described above.
At all the initial AD concentrations tested, the DH# values were
positive (17.26e32.26 kJ mol
1), which further confirms the
endothermic nature of the AD biosorption process, while the DS#
values were negative (from
0.14 to
0.10 kJ mol
1 K
1), indicating
that AD molecules are stable on the LEC surface. The association,
fixation or immobilization of AD molecules as a result of bio-
sorption is attributed to a decrease in the degree of freedom of AD
molecules which gives rise to a negative DS# (Pandey et al., 2010). In
addition, negative DS# values indicate that no significant change
occurs in the internal structure of the biosorbent during the bio-
sorption process (Dog an et al., 2009). Furthermore, at all initial AD
concentrations and temperatures assayed, the values of DG# were
positive (58.58e64.8 kJ mol
1), indicating that the AD biosorption Fig. 5. Effect of eluents on AD desorption (A) and desorption kinetics of AD (B).
reaction is not spontaneous and therefore requires some energy
from an external source to take place (Dog an et al., 2009; Alkan
desorption studies were conducted, using different eluent solu-
et al., 2008). The biosorption processes of methyl violet onto
tions. As illustrated in Fig. 5A, the decrease in AD desorption per-
palm kernel fiber (Ofomaja et al., 2011), maxilon blue 5G onto
centage in relation to the type of eluent manifested the following
sepiolite (Alkan et al., 2008) and methylene blue onto hazelnut
an et al., 2009) were also found to be endothermic order: strong alkaline (pH z 12) > weak alkaline
shell (Dog
(pH z 11.5) > neutral salt (pH z 7) > weak acid (pH z 5.1) > strong
(DH# > 0), non-spontaneous (DG# > 0) and occur without changes
acid (pH z 1) solutions. These results indicate that AD desorption
in the structure of the sorbents (DS# < 0).
percentage depended on the eluent pH: the higher the eluent pH,
the greater the AD desorption percentage. The potential of the
3.8. AD desorption studies alkaline eluents to elute AD from AD-loaded LEC can be explained
on the basis that the high eluent pH increased the number of
In order to explore the possibility of recovery of LEC and AD,

Table 3
Thermodynamic parameters for the biosorption of AD onto LEC.

C0 (mg L
1) T ( C) DG# (kJ mol
1) EA (kJ mol
1) A0 (g mg
1 h
1) DH# (kJ mol
1) DS# (kJ mol
1 K
1)
50 18 58.575 19.808 186.830 17.260
0.1419
28 59.994
40 61.697
50 63.116
100 18 61.006 23.533 318.939 20.985
0.1375
28 62.380
40 63.620
50 64.584
200 18 61.588 34.807 26396.851 32.258
0.1007
28 62.595
40 63.804
50 64.812
 I. Guerrero-Coronilla et al. / Journal of Environmental Management 152 (2015) 99e108
negatively charged groups on the LEC surface, which favored
desorption of AD molecules due to electrostatic repulsion (Çelekli
and Bozkurt, 2013). Contrarily, the strong mineral acids were only
able to elute negligible amounts (1.5e6.4%) of biosorbed AD mol-
ecules. Based on these results, further AD desorption kinetic studies
were conducted using alkaline eluents (0.01 M NaOH and KOH, and
0.1 M Na2CO3).
Fig. 5B shows that with all assayed alkaline eluents, AD
desorption was very fast initially, although subsequently the rate
gradually diminished. All AD was desorbed from AD-loaded LEC
with 0.01 M NaOH and KOH at only 15 min, whereas about 80% of
AD was desorbed with 0.1 M Na2CO3 in 10 min. A two-way analysis
of variance revealed no significant difference in the extent, rate and
time of AD desorption (p > 0.05), when 0.01 M NaOH and KOH
solutions were used as eluents. These results clearly indicate that
NaOH and KOH are the best eluents for AD desorption from AD-
loaded LEC. However, as NaOH is cheaper than KOH, the cost of
the AD desorption process would be lower if NaOH was used as the
eluent. The effective performance on the part of NaOH for AD
desorption has been reported previously (Ahmad and Kumar, 2011;
Re^go et al., 2013).
4. Conclusions
The possibility of using LEC for the biosorption of AD was
examined in this work. The kinetic, equilibrium and thermody-
namic parameters were also analyzed. LEC particle size, solution
pH, initial AD concentration and temperature affected the capacity
and/or rate of AD biosorption. AD biosorption kinetics were found
to follow a pseudo-second-order rate expression. Equilibrium bio-
sorption data for AD onto LEC were best represented by the Lang-
muir isotherm. The AD biosorption process was endothermic and
non-spontaneous. Desorption study revealed that 0.01 M NaOH
and KOH can elute all AD from the AD-loaded LEC in about 15 min.
This study demonstrates that LEC can be used as a low cost, effec-
tive and eco-friendly biosorbent for the treatment of AD-polluted
wastewaters.
Acknowledgments
The authors gratefully acknowledge the support provided by the
scientific team of the Central Laboratory of Biotechnology and
Molecular Biology at Escuela Nacional de Ciencias Biolo
gicas,

cnico Nacional (IPN), as well as the financial support
Instituto Polite
provided by the Secretaría de Investigacio
n y Posgrado, IPN (Nos.
20141472 and 20151245). The CONACyT awarded a graduate
scholarship to one of the co-authors (I.G.-C.). E.C.-U. is a holder of
grants from COFAA-IPN, EDI-IPN, and SNI-CONACyT. L.M.-B. is a
holder of a grant from SNI-CONACyT.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.jenvman.2015.01.026.
References
Ahmad, R., Kumar, R., 2011. Adsorption of amaranth dye onto alumina reinforced
polystyrene. Clean-Soil Air Water 39, 74e82.
Akar, T., Safa Ozcan, A., Tunali, S., Ozcan, A., 2008. Biosorption of a textile dye (acid
blue 40) by cone biomass of Thuja orientalis: estimation of equilibrium, ther-
modynamic and kinetic parameters. Bioresour. Technol. 99, 3057e3065.
_ Tunç, O., 2008. A comparative adsorption/biosorption study of
Aksu, Z., Tatli, A.I.,
acid blue 161: effect of temperature on equilibrium and kinetic parameters.
Chem. Eng. J. 142, 23e39.
Alkan, M., Dog an, M., Turhan, Y., Demirbaş, O., Turan, P., 2008. Adsorption kinetics
and mechanism of maxilon blue 5G dye on sepiolite from aqueous solutions.
107
Chem. Eng. J. 139, 213e223.
Anjaneya, O., Shrishailnath, S.S., Guruprasad, K., Nayak, A.S., Mashetty, S.B.,
Karegoudar, T.B., 2013. Decolourization of amaranth dye by bacterial biofilm in
batch and continuous packed bed bioreactor. Int. Biodeter. Biodegr 79, 64e72.
Asgher, M., Bhatti, H.N., 2010. Mechanistic and kinetic evaluation of biosorption of
reactive azo dyes by free, immobilized and chemically treated Citrus sinensis
waste biomass. Ecol. Eng. 36, 1660e1665.
Ayranci, E., Duman, O., 2009. In-situ UV-visible spectroscopic study on the
adsorption of some dyes onto activated carbon cloth. Sep. Sci. Technol. 44,
3735e3752.
Çelekli, A., Bozkurt, H., 2013. Sorption and desorption studies of a reactive azo dye
on effective disposal of redundant material. Environ. Sci. Pollut. Res. 20,
4647e4658.
Dogan, M., Abak, H., Alkan, M., 2009. Adsorption of methylene blue onto hazelnut
shell: kinetics, mechanism and activation parameters. J. Hazard. Mater. 164,
172e181.
Febrianto, J., Kosasih, A.N., Sunarso, J., Ju, Y.H., Indraswati, N., Ismadji, S., 2009.
Equilibrium and kinetic studies in adsorption of heavy metals using biosorbent:
a summary of recent studies. J. Hazard. Mater. 162, 616e645.
Flores-Garnica, J.G., Morales-Barrera, L., Pineda-Camacho, G., Cristiani-Urbina, E.,
2013. Biosorption of Ni(II) from aqueous solutions by Litchi chinensis seeds.
Bioresour. Technol. 136, 635e643.
Giles, C.H., Smith, D., Huitson, A., 1974. A general treatment and classification of the
solute adsorption isotherm. I. Theoretical. J. Colloid Interface Sci. 47, 755e765.
Gong, R., Ding, Y., Li, M., Yang, C., Liu, H., Sun, Y., 2005. Utilization of powdered
peanut hull as biosorbent for removal of anionic dyes from aqueous solution.
Dyes Pigments 64, 187e192.
Guerrero-Coronilla, I., 2013. Amaranth (Acid Red 27) Dye Biosorption by Water
Hyacinth. M.Sc. Thesis. National School of Biological Sciences, National Poly-
technic Institute, Mexico City, Mexico.
Guerrero-Coronilla, I., Morales-Barrera, L., Villegas-Garrido, T.L., Cristiani-Urbina, E.,
2014. Biosorption of amaranth dye from aqueous solution by the root, leaf, stem
and entire plant of E. crassipes. Environ. Eng. Manag. J. 13 (8), 1917e1926.
Gupta, V.K., Jain, R., Mittal, A., Saleh, T.A., Nayak, A., Agarwal, S., Sikarwar, S., 2012.
Photo-catalytic degradation of toxic dye amaranth on TiO2/UV in aqueous
suspensions. Mater. Sci. Eng. C 32, 12e17.
Ho, Y.S., 2006. Review of second-order models for adsorption systems. J. Hazard.
Mater. B136, 681e689.
Ibrahim, H.S., Ammar, N.S., Soylak, M., Ibrahim, M., 2012. Removal of Cd(II) and
Pb(II) from aqueous solution using dried water hyacinth as a biosorbent.
Spectrochim. Acta A 96, 413e420.
Limousin, G., Gaudet, J.P., Charlet, L., Szenknect, S., Barthe s, V., Krimissa, M., 2007.
Sorption isotherms: a review on physical bases, modeling and measurement.
Appl. Geochem. 22, 249e275.
Malik, A., 2007. Environmental challenge vis a vis opportunity: the case of water
hyacinth. Environ. Int. 33, 122e138.
Mittal, A., Kurup, L., Gupta, V.K., 2005. Use of waste materials ebottom ash and de-
oiled soya, as potential adsorbents for the removal of amaranth from aqueous
solutions. J. Hazard. Mater. B117, 171e178.
Mittal, A., 2006. Removal of the dye, amaranth from waste water using hen feathers
as potential adsorbent. Electron. J. Environ. Agric. Food Chem. 5, 1296e1305.
Mittal, A., Thakur, V., Gajbe, V., 2012. Evaluation of adsorption characteristics of an
anionic azo dye brilliant yellow onto hen feathers in aqueous solutions. Environ.
Sci. Pollut. Res. 19, 2438e2447.
Ofomaja, A.E., Ukpebor, E.E., Uzoekwe, S.A., 2011. Biosorption of methyl violet onto
palm kernel fiber: diffusion studies and multistage process design to minimize
biosorbent mass and contact time. Biomass Bioenerg. 35, 4112e4123.
Pandey, P.K., Sharma, S.K., Sambi, S.S., 2010. Kinetics and equilibrium study of
chromium adsorption on zeoliteNaX. Int. J. Environ. Sci. Technol. 7, 395e404.

rez-Marín, A.B., Aguilar, M.I., Meseguer, V.F., Ortun
Pe ~ o, J.F., Sa

ez, J., Llore

ns, M., 2009.
Biosorption of chromium(III) by orange (Citrus cinensis) waste: batch and
continuous studies. Chem. Eng. J. 155, 199e206.
Plazinski, W., Dziuba, J., Rudzinski, W., 2013. Modeling of sorption kinetics: the
pseudo-second order equation and the sorbate intraparticle diffusivity.
Adsorption 19, 1055e1064.
Re^go, T.V., Cadaval Jr., T.R.S., Dotto, G.L., Pinto, L.A.A., 2013. Statistical optimization,
interaction analysis and desorption studies for the azo dyes adsorption onto
chitosan films. J. Colloid Interface Sci. 411, 27e33.
Schiewer, S., Patil, S.B., 2008. Pectin-rich fruit wastes as biosorbents for heavy metal
removal: equilibrium and kinetics. Bioresour. Technol. 99, 1896e1903.
Sivarajasekar, N., Baskar, R., Balakrishnan, V., 2009. Biosorption of an azo dye from
aqueous solutions onto Spirogyra. J. Univ. Chem. Technol. Metall. 44, 157e164.
Suazo-Madrid, A., Morales-Barrera, L., Aranda-García, E., Cristiani-Urbina, E., 2011.
Nickel(II) biosorption by Rhodotorula glutinis. J. Ind. Microbiol. Biotechnol. 38,
51e64.
Tunc, S., Duman, O., 2007. Investigation of interactions between some anionic dyes
and cationic surfactants by conductometric method. Fluid Phase Equilib. 251,
1e7.
Venkata-Mohan, S., Ramanaiah, S.V., Sarma, P.N., 2008. Biosorption of direct azo dye
from aqueous phase onto Spirogyra sp. I02: evaluation of kinetics and mecha-
nistic aspects. Biochem. Eng. J. 38, 61e69.
Vijayaraghavan, K., Palanivelu, K., Velan, M., 2006. Treatment of nickel containing
electroplating effluents with Sargassum wightii biomass. Process Biochem. 41,
853e859.
Wang, B.E., Hu, Y.Y., Xie, L., Peng, K., 2008. Biosorption behavior of azo dye by
108 I. Guerrero-Coronilla et al. / Journal of Environmental Management 152 (2015) 99e108

inactive CMC immobilized Aspergillus fumigatus beads. Bioresour. Technol. 99, methods. Food Chem. 136, 442e449.
794e800. Zhang, Y., Gan, T., Wan, C., Wu, K., 2013. Morphology-controlled electrochemical
Zhang, G., Ma, Y., 2013. Mechanistic and conformational studies on the interaction sensing amaranth at nanomolar levels using alumina. Anal. Chim. Acta 764,
of food dye amaranth with human serum albumin by multispectroscopic 53e58.