Chemical Engineering Journal 189190 (2012) 264274

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal
j our nal homepage: www. el sevi er . com/ l ocat e/ cej
Removal of basic dyes in binary system by adsorption using
rarasaponinbentonite: Revisited of extended Langmuir model
Aln Kurniawan, Hogiartha Sutiono, Nani Indraswati, Suryadi Ismadji

Department of Chemical Engineering, Widya Mandala Surabaya Catholic University, Kalijudan 37, Surabaya 60114, Indonesia
a r t i c l e i n f o
Article history:
Received 23 January 2012
Received in revised form23 February 2012
Accepted 23 February 2012
Keywords:
Adsorption
Malachite green
Methylene blue
Binary system
Extended Langmuir
a b s t r a c t
The preparation of organo-bentonite via microwave radiation using natural surfactant (i.e. rarasaponin)
from the tropical soapfruit pericarps of Sapindus rarak was conducted in this work. The adsorption per-
formance of bentonite and its modied form was studied for the removal of two dyes (i.e. methylene
blue and malachite green) in single and binary systems. Langmuir and Freundlich models were applied
to evaluate adsorption equilibrium data in single systems. From physical meaning interpretation of the
model parameters, it was found that Langmuir model gave the best representation. A new approach
in representing binary experimental isotherms with extended Langmuir model incorporating fractional
surface coverage in 3D mesh plot was also veried. The results show that recently proposed modication
of extended Langmuir model could be used to evaluate binary adsorption data well with reasonable tted
parameter value.
2012 Elsevier B.V. All rights reserved.
1. Introduction
The increasing environmental pollution from domestic and
industrial efuents, particularly in the developing countries is of
major concern. One kind of contaminants that are frequently pol-
luting the water systems are dyes. Dyes are complex aromatic
substances that indispensable in many industries such as tex-
tile, paper, paint, leather, pigment, and pharmaceutical to impart
color. Compared with natural dyes, synthetic dyes are superior in
terms of color availability, permanent, ease to use, quick-setting
and ensured by accurate formulas [1]. More than 100,000 kinds
of synthetic dyes are commercially available with total amount
of production reach 710
5
tonnes annually [2]. Of these, mala-
chite green (MG) and methylene blue (MB) are two kinds of dyes
that are most widely used among all other dyes in basic dye
class [3,4]. Malachite green (4-[(4-dimethylaminophenyl)phenyl-
methyl]-N,N-dimethylaniline) is a triphenylcarbeniumsalts, which
composed of triarylmethane ring as the chromophore structure.
Its applications are widely found in the paper, leather, silk and
wool industries to color such products and aquaculture eld as a
biocide, fungicide and parasiticide therapeutic agent [3]. In other
hand, methylene blue (3,7-bis(dimethylamino)phenothiazin-5-
ium chloride) is a thiazine dye that used as redox indicator in the
chemistry eld, bacteria stain and indicator of viable yeast cell in
the biology eld and dyestuff for cotton, wool, paper and silk [4,5].

Corresponding author. Tel.: +62 313891264; fax: +62 313891267.

E-mail address: [email protected](S. Ismadji).
Despite of the commercial importance of MG and MB dyes, the
presence of these substances in natural environment (e.g. water
bodies) even in trace quantities are highly undesirable for both
toxicological andesthetical reasons. Moreover, the prolongedaccu-
mulation of these dyes in human body via food chain can cause
several diseases and disorders due to their carcinogenic and muta-
genic activity [6,7]. Therefore, the removal of these dyes fromwaste
efuents is of interest prior to release into natural environment.
Several conventional methods are available for the treatment
and decontamination of dyes including physical, chemical and
biological methods such as ozonation, photocatalysis, chemical
oxidation, membrane ltration, otation, coagulation, aerobic and
anaerobic microbial degradation, enzyme degradation and adsorp-
tion [8,9]. However, most of them are considered ineffective
for handling colored efuents due to chemical stability and low
biodegradabilityof most dyes. Inother hand, adsorptionhas proven
to be a promising and cost effective method for this purpose. Cost is
actually an important parameter for the selection of the adsorbent
materials in designing adsorption system. In recent years, atten-
tion has been focused on the search and development of various
low cost alternative adsorbents for dyes removal, including agri-
cultural solid wastes, industrial by-products and natural materials
[1012].
Clay has been emerged as potential natural adsorbents in the
adsorptionelddue totheir abundant reserve, lowcost, highchem-
ical andmechanical stabilities as well as highadsorptive properties.
The applications of clay-based adsorbents for the removal of dyes
have received growing attentions in the last decade [10,1315].
Bentonite is a type of natural clay mineral that abundantly found in
1385-8947/$ see front matter 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.cej.2012.02.070
A. Kurniawan et al. / Chemical Engineering Journal 189190 (2012) 264274 265
Table 1
Information on the methylene blue and malachite green dyes.
Methylene blue (MB) Malachite green (MG)
C.I number 52015 42000
C.I name Basic blue 9 Basic green 4
Chemical formula C
16
H
18
N
3
SCl C
23
H
25
N
2
Cl
Molecular weight (gmol
1
) 319.85 364.91
(nm) 664.1 617.3
Molecular 3D-structure
many countries, including Indonesia. It is originated fromthe alter-
ation of glass-rich volcanic rocks such as tuffs and ash deposits that
composedof montmorillonite mineral as the mainconstituent [16].
The surface properties of bentonite can be remarkably enhanced by
replacing inorganic metal cations (typically Na
+
and Ca
2+
) at ben-
tonite interlayer with organic cations. This substitution results in
the hydrophobic surface nature of bentonite, which is suitable for
the adsorptionof organic substances. Several studies have reported
on the modication of bentonite using various surfactants such as
CTAB, CDBAC, CPC, and TMAB for dye removal [1719]. However,
the use of synthetic surfactants often causes environmental issues
and need an extra cost for the treatment of unused chemicals. For
this reason, the utilization of natural surfactant (i.e. rarasaponin)
fromthetropical soapfruit pericarps of Sapindus rarakis highlighted
in this work.
This work deals with the removal of two basic dyes (i.e. MB
and MG) in single and binary systems using Trenggalek ben-
tonite modied with rarasaponin surfactant. To date, there is
no report on the removal of these dyes in binary system using
rarasaponinbentonite. A new approach of extended Langmuir
model for representation of experimental isotherm in binary sys-
temwas also proposed. Most of studies on the binary adsorption of
various substances usedsuchmodel toevaluate adsorptionequilib-
riumdata by introducing original K
L
and q
m
parameters fromsingle
system [2024]. However, such procedure cannot be precisely
implemented in binary system, where the competition between
adsorbate components for the adsorption sites occurs and may act
as the sorption-controlling factor. Therefore, we proposed a revis-
ited mathematic expression of q
m
and K
L
parameters in extended
Langmuir model for binary systemby inclusionof fractional surface
coverage.
2. Materials and methods
2.1. Materials
Malachite green (A.R. Grade, purity>96%) and methylene blue
(A.R. Grade, purity>95%) were purchased fromSigma Aldrich and
used as received. Some information about these dyes and their
molecular structures are summarized in Table 1. Rarasaponin sur-
factant used in this study was isolated from the tropical fruit
pericarps of S. rarak by extraction process. The detail procedure
of rarasaponin extraction can be found elsewhere [10].
2.2. Preparation of adsorbents
Bentonite used in this study was collected fromthe mining site
located in Trenggalek, East Java, Indonesia. The type of bentonite
fromthis location is Ca-bentonite. Prior to use, bentonite was puri-
ed using H
2
O
2
solution according to the procedure reported by
Rahardjoet al. [25]. Thecationexchangecapacity(CEC) of bentonite
was determined based on the ASTM C837-99 standard procedure
andfoundtobe 0.537meq/g clay. The mineral compositions of ben-
tonite were analyzed by atomic absorption spectrophotometry, in
a FAAS Shimadzu AA6200 and the results are given as follows: SiO
2
57.8%, Al
2
O
3
21.6%, Fe
2
O
3
6.4%, CaO4.3%, Na
2
O1.9%, MgO1.5%, K
2
O
0.7% and H
2
O 5.8%.
Rarasaponinbentonite was synthesized by following proce-
dure: 10g of bentonite was dispersed into 50mL distilled water
containingrarasaponinsurfactant withconcentrationequivalent to
1.5 times of the CEC of bentonite. The mixture was then subjected
to heating in a microwave oven (National NN-S327 WF) for 60s at
heating power of 700W. The resulting rarasaponinbentonite was
then repeatedly washed with distilled water and dried in an oven
at 100

C for 24h.
2.3. Characterizations of adsorbents
The pore structure of the adsorbents was analyzed by nitrogen
adsorption at 77K, in a Quadrasorb SI at relative pressure (p/p
0
)
range from approximately 0.001 to 0.995. The samples were
degassed in vacuum condition at 423K for at least 24h prior to
the adsorption analysis. The surface area of the adsorbents (S
BET
)
was determined by standard BET equation applied at relative
pressure range from 0.06 to 0.3. The total pore volume (V
T
) of the
adsorbents was calculated from the gas adsorption data at the
highest relative pressure (p/p
0
=0.995). The pore size distributions
(PSDs) of the adsorbents were evaluated by Density Functional
Theory (DFT) model available within the instrument with medium
regularization.
The layer structure of bentonite and rarasaponinbentonite was
characterized by XRD method (Rigaku Miniex Goniometer). The
powder XRD patterns of the samples were collected at 30kV and
15mA, with a Ni-ltered CuK radiation (=1.542

A) at a step size
of 0.01

.
The surface functional groups of the adsorbents were analyzed
by infrared spectroscopy, in a FTIR Shimadzu 8400S spectrometer
using KBr disk pelleting technique. The spectra data of the samples
were recorded in the mid-IR wavenumber range (4000500cm
1
).
The pH
pzc
(point of zero charge) of bentonite and
rarasaponinbentonite were determined by pH-drift procedure
[26], and found to be 2.86 and 3.02, respectively.
2.4. Adsorption experiments
For single systems, the stock solutions of dyes at initial concen-
tration of 1000mg/L (2.74mmol/L MG and 2.67mmol/L MB)
266 A. Kurniawan et al. / Chemical Engineering Journal 189190 (2012) 264274
were prepared by dissolving a known amount of dye into dis-
tilled water. Adsorption isothermexperiments were carried out by
adding various mass of adsorbent into a series of conical asks con-
taining 100mL of dye solution. The asks were then shaken in a
Memmert shaking water bath at various temperatures (303.15K,
323.15K and 343.15K) within 12h until equilibrium condition
reached. Subsequently, the mixture was centrifuged at 4500rpm
for 5min to remove solid particles. The resultant supernatant was
taken and analyzed its residual concentration using a Shimadzu
UV/Vis-1700 spectrophotometer at maximum absorbance wave-
lengths () of 617.3nmfor MG and 664.1nmfor MB, respectively.
The effect of pH was also studied at pH range of 310. To adjust
the pH, an appropriate amount of 0.1N HCl or NaOH solutions
was added to the system. The calibration curves for dyes in single
systems were prepared by measuring the absorbance of different
concentrations of the dyes. The amount of dye adsorbed at equilib-
riumper unit of mass, q
e
, was calculated as follows:
q
e
=
(C
0
C
e
) V
m
(1)
where C
0
and C
e
are the initial and equilibrium concentrations
of dye in solution (mmol/L), respectively, V is the volume of dye
solution (L) and mis the mass of the adsorbents used (g).
For binary systems, the dye mixtures were prepared by adding
a known amount of MB dye (0.25g, 0.5g, 0.75g and 1g) into 1L of
MGsolutions with initial concentration of 2.74mmol/L. Adsorption
isotherm experiments were conducted based on the procedure in
single system. The residual dye concentration in binary systemwas
determined by measuring MG and MB dyes at
1
and
2
, respec-
tively, to give optical densities of d
1
and d
2
, as previously described
by Choy et al. [20]:
C
e,MG
=
k
MB,2
d
1
k
MB,1
d
2
k
MG,1
k
MB,2
k
MG,2
k
MB,1
(2)
C
e,MB
=
k
MG,1
d
2
k
MG,2
d
1
k
MG,1
k
MB,2
k
MG,2
k
MB,1
(3)
where k
MG,1
, k
MG,2
, k
MB,1
and k
MB,2
are the calibration constants for
MG and MB dyes at the two wavelengths,
1
and
2
, respectively.
The amount of i dye adsorbed at equilibriumper unit of mass, q
e,i
,
was calculated as follows:
q
e,i
=
(C
0,i
C
e,i
) V
m
(4)
where C
0,i
and C
e,i
are the initial and equilibriumconcentrations of
i dye in the mixture (mmol/L), respectively, V is the volume of dye
mixture (L) and mis the mass of the adsorbents used (g).
3. Results and discussion
3.1. Pore characteristics of adsorbents
The pore structure of the solid is a critical variable in the adsorp-
tion process since it determines the failure or success of such
process from both equilibria and kinetic viewpoints. In this work,
the pore structure of bentonite and rarasaponinbentonite was
analyzed by nitrogen adsorption at 77K (Supporting information
Fig. S1). From this gure, it can be conrmed the largely meso-
porosity on the surface of bentonite and rarasaponinbentonite,
as denoted from the adsorptiondesorption isotherm curves that
exhibit type-IV shape based on IUPAC classication. Moreover, the
mesoporous characters of the adsorbents were also observed from
the hysteresis curve that lies between adsorption and desorption
isotherm curves at relative pressure (p/p
0
) above 0.5 (Supporting
information Fig. S1).
The surface area (S
BET
) of rarasaponinbentonite was lower than
bentonite (45.3m
2
/g vs. 51.8m
2
/g) while its total pore volume (V
T
)
is higher than bentonite (0.49cm
3
/g vs. 0.41cm
3
/g). This is likely
due to the formation of larger pore as a result of intercalation of
organic modier (i.e. acetyl cation) at the interlayer spacing of ben-
tonite. Such pore characteristics are also evidenced fromDFT result
(Supporting information Fig. S2).
3.2. Layer characteristics of adsorbents
XRD analysis can be used as useful tool to characterize the layer
structure of bentonite before and after modication. The powder
XRDpatterns of bentonite and rarasaponinbentonite are shownin
Fig. 1. FromFig. 1, it canbe seenthe reectionpeak for Ca-bentonite
at 2 =5.92

withad(001)-spacingvalueof 14.93

A. Thepresenceof
other mineral compounds such as kaolin (K), quartz (Q) and calcite
(Ca) in bentonite were also observed in the reections relative to
the planes K(001), Q(101) and Ca(104), respectively.
After modication with rarasaponin surfactant, the reection
peak of rarasaponinbentonite was shifted to lower angle of
2 =5.77

with an increase in d(001)-spacing value to 15.87

A. The
slightly increase in the d(001)-spacing of rarasaponinbentonite
denotes that not all the interlayer sheets have been intercalated
by acetyl cations. Such phenomena result in the expansion of the
2:1 interlamellar distance and later will promote the adsorption
capacity of rarasaponinbentonite for dyes removal.
3.3. Surface chemistry characteristics of adsorbents
Infrared spectroscopy analysis (FT-IR) was performed to probe
the conformation of surfactant cation species at the interlayer
of rarasaponinbentonite and identify the vibrational frequency
changes of surface groups of the adsorbents, which are involved in
thesorptionprocess (gurenot shownhere). TheFT-IRassignments
of bentonite andrarasaponinbentonite are summarizedinTable 2.
FromTable 2, it canbe seenthe formationof rarasaponinbentonite
proceeded via ion exchange mechanism between inorganic inter-
layer cation (i.e. Ca
2+
) and acetyl cation (C
2
H
3
O
+
), as indicated
from the presence of C O stretch vibration at 1712cm
1
. Other
FT-IR assignments in bentonite, which are the characteristics of
montmorillonite clay also noticed such as Al O Si deformation
at 524cm
1
; Si O stretch of quartz at 677cm
1
; Al Al OH defor-
mation at 923cm
1
; Si O stretch (in-plane) of tetrahedral sheet at
1040cm
1
; O H bend for adsorbed H
2
O at the interlayer of clay at
1638cm
1
and O Hstretchof aluminol (Al OH) or silanol (Si OH)
at 3621cm
1
.
The FTIR assignments of bentonite loaded with dyes (i.e. MB
and MG) were also given in Table 2. From this table, it can be
seen the presence of assignments of N O asymmetric stretch
and S O stretch at 1528cm
1
and 973cm
1
, respectively. These
assignments may reveal to the binding of dye molecules onto
the bentonite surface, which occur via electrostatic dipoledipole
interaction between negatively charged oxygen atoms of the
deprotonated silanol sites and positively charged site of the dye
molecules (i.e. N
+
for MG and S
+
for MB). Moreover, the pres-
ence of other assignment at 1592cm
1
, attributed to the aromatic
C C stretch of the dye structure was also observed. The binding
of the dye moieties in the silanol sites was also conrmed from
the shifting of OH-stretch in the silanol species from3621cm
1
to
3563cm
1
and reduction in its corresponding peak energy. Sim-
ilar observations were also noticed in the FTIR assignments for
rarasaponinbentonite loaded with dyes.
3.4. Effect of pH
The pHis regarded as a crucial parameter in the adsorption pro-
cess since it may control the uptake mechanism of the adsorbate
as well as inuence the adsorption sites (e.g. the potential charge
A. Kurniawan et al. / Chemical Engineering Journal 189190 (2012) 264274 267
Table 2
FTIR assignments of pristine and dye-loaded adsorbents.
FTIR assignments Wavenumber (cm
1
)
Bentonite rarasaponin
bentonite
Bentonite +dyes Rarasaponin
bentonite +dyes
O H stretch (Si OH or Al OH) 3621 3614 3563 3548
C O stretch (C
2
H
3
O
+
) 1712 1706
O H bend (adsorbed H
2
O at the interlayer of clay) 1638 1636 1638 1638
C C stretch (aromatic ring) 1592 1581
N O stretch (asymmetric stretch) 1528 1542
Si O stretch (in-plane) for tetrahedral sheet 1040 1017 1025 1032
S O stretch 973 968
Al Al OH 923 916 916 916
Si O stretch (quartz) 677 675 690 682
Al O Si 524 522 523 524
of the site) on the solid surface. In the case of dye adsorption, the
changes in the structural and colour stabilities as well as degree of
ionization for such adsorbate is also dependent on the pH of the
solution [27]. The effect of pH on the adsorption of MB and MG
dyes onto rarasaponinbentonite in single system was studied at
pH range of 310 and the results are given in Fig. 2.
As noted in Fig. 2, a gradual increase in the percentage removal
for both dyes was observed in the pH range of 47 and reached
its maximum at around pH 8. Increasing pH more than this value
would lead to insignicant change in the percentage removal for
both dyes. Moreover, a substantial reduction in the color inten-
sity for MG was observed at pH>9 while the color intensity of MB
Fig. 1. Powder XRD patterns of bentonite (A) and rarasaponinbentonite (B).
268 A. Kurniawan et al. / Chemical Engineering Journal 189190 (2012) 264274
Fig. 2. Effect of pH on the adsorption of dyes onto rarasaponinbentonite
(C
0
=2.74mmol/L; t =2h; T =303.15K).
slightly changes from pH 8 to pH 10. Therefore pH 8 was selected
as optimum point and used for further experiments in this study.
Since both MG and MB are cationic dyes, which have similar pH
value for their highest removal, therefore pH8was alsousedduring
adsorption experiments in binary system.
In basic solution, the adsorption of cationic dyes was favored
due to the presence of negatively charged sites (i.e. deprotonated
silanol) on the adsorbent surface, resulting in the electrostatic
forces between these groups and positively charged dye molecules.
As mentioned in previous paragraph, the pH of the solution also
gives impact on the surface sites of the adsorbent, associated to the
protonation and deprotonation of silanol groups ( Si OH) as the
active sites. The protonation and deprotonation of silanol species
take place as follows:
Si OH + H
+
Si OH
2
+
, whenpH
solution
< pH
pzc
Si OH + OH

Si O

+H
2
O, whenpH
solution
> pH
pzc
In acidic solution (e.g. pH=3), the adsorption of dyes was
retarded due to sorption competition between hydrogen ions (H
+
)
and dye molecules for occupancy of the active sites. Moreover, high
concentration of H
+
ions causes the protonation of silanol groups,
which resulted in ionic repulsion between cationic dye molecules
and protonated silanol sites. As the pH of the system increases,
the concentration of H
+
ions in bulk solution decreases and the
silanol sites become deprotonated, therefore increasing the neg-
ative charge density on the adsorbent surface and facilitating the
adsorption of dye molecules.
3.5. Adsorption isotherms
Adsorption isotherm is the most valuable information in the
adsorption studies for the practical and operational design of the
sorption system since it describes the distribution of the adsor-
bate molecules within the solid and liquid phase at equilibrium. By
evaluating this aspect, the information on the adsorption kinetics
of single component system, adsorption equilibria of multicompo-
nent systems and adsorption kinetics of multicomponent systems
can be obtained [28]. In the beginning, the adsorption models
such as Langmuir, Freundlich, Sips, Toth, DubininRaduskevich,
RedlichPeterson, Temkin, andHill are extensively usedtodescribe
adsorption in gas phase, but in later, most of them can in prin-
ciple be extended to liquid phase adsorption. Several studies have
reportedonthe applications of different adsorptionmodels to eval-
uate various liquid phase adsorption systems [2933].
3.5.1. Adsorption isotherms for single system
In this study, the two-well known adsorption models namely
Langmuir and Freundlich were used to evaluate adsorption equi-
libriumdatainsinglesystem. Langmuir model (1918) was proposed
based on the kinetic principle, that is the adsorption rate is equal to
the desorption rate fromthe surface to maintain zero rate of accu-
mulation when equilibrium is approached [28]. This model also
assumed that the adsorption on surface occurs at a nite or de-
nite localized site with no lateral interaction and steric hindrance
between the adsorbed species, even on the adjacent sites [34]. The
Langmuir model can be expressed mathematically as follows:
q
e
=
q
m
K
L
C
e
1 +K
L
C
e
(5)
where q
m
is the maximum adsorption capacity for the adsorbent,
corresponds to the monolayer surface coverage (mmol/g) and K
L
is
Langmuir constant represented adsorption afnity (L/mmol). The
characteristic of Langmuir model can be essentially expressed by
dimensionless constant calledequilibriumparameter, R
L
, as follows
[35]:
R
L
=
1
1 +K
L
C
0
(6)
The values of R
L
indicate various nature of the adsorption systems
to be unfavorable (R
L
>1), favorable (0<R
L
<1), linear (R
L
=1) or
irreversible (R
L
=0).
The second adsorption model used in this study was Freundlich.
Freundlich (1932) proposed an adsorption on surface with non-
ideal and reversible characteristics, which is not restricted to the
monolayer adsorption [36]. To date, this model has been success-
fully applied to describe various adsorption systems of organics
onto activated carbon and other adsorbents. However, the applica-
tion of Freundlich model to evaluate adsorption data is generally
valid in the narrow range of concentration since this model does
not have Henry lawlimit at lowconcentration and saturation limit
at high-end concentration. The Freundlich model has the following
form:
q
e
= K
F
C
1/n
e
(7)
where K
F
and n are Freundlich constants for adsorptivity of solute
[(mmol/g)(mmol/L)
n
] and system heterogeneity, respectively.
Higher value of n denotes the more heterogeneous the system is
and the systemapproaches a rectangular isotherm(or irreversible
isotherm) when n value is getting larger than about 10 [28]. The K
F
constant can also be regarded as the maximumadsorption capacity
of the adsorbent only when the value of n approaches innite [37].
The plots of the applied models against adsorption equilibrium
data at various temperatures are depicted in Figs. 3 and 4. The
values of the model parameters were obtained by nonlinear least
square tting andthe results are giveninTable 3. FromFigs. 3and4,
it seems that bothLangmuir andFreundlichmodels cant all exper-
imental data well, as indicated by good values of correlation factor
(R
2
) that lies between 0.95 and 0.99. The values of K
L
and K
F
param-
eters seemto increase at higher temperature, denoting that higher
temperature favors the adsorption process due to greater adsorp-
tivity of dye molecules toward the adsorbent surface. The increase
in afnity parameter can also be explained fromkinetics viewpoint
that is higher temperature provides more energy for the solutes to
undergo an interaction with the active sites on the solid surface.
Similar observation is also noticed in the q
m
parameter, in which
this value increases by the increase of temperature. Higher value
of q
m
for rarasaponinbentonite is likely due to the expansion of
2:1 interlamellar distance as a result of insertion of acetyl cation,
which promotes the adsorption capacity of the adsorbent for dyes
removal.
A. Kurniawan et al. / Chemical Engineering Journal 189190 (2012) 264274 269
Table 3
Fitted parameters for the adsorption of MB and MG in single systemat various temperatures.
Adsorbent Adsorbate T (K) Langmuir parameters Freundlich parameters
qm (mmol/g) KL (L/mmol) R
2
RL KF (mmol.g
1
)
(mmol/L)
1/n
n R
2
Bentonite MG 303.15 0.51 1.56 0.98 0.19 0.31 2.73 0.98
323.15 0.53 2.13 0.97 0.15 0.35 3.00 0.98
343.15 0.56 3.23 0.99 0.10 0.41 3.39 0.99
MB 303.15 0.54 1.67 0.98 0.18 0.34 2.72 0.97
323.15 0.57 2.46 0.97 0.13 0.37 3.15 0.97
343.15 0.59 3.34 0.97 0.10 0.42 3.22 0.98
RarasaponinBentonite MG 303.15 0.66 1.71 0.98 0.18 0.36 2.15 0.98
323.15 0.87 2.31 0.98 0.14 0.47 2.34 0.97
343.15 0.93 3.66 0.97 0.09 0.73 2.72 0.95
MB 303.15 0.71 1.75 0.99 0.18 0.41 2.19 0.98
323.15 0.89 2.54 0.98 0.13 0.49 2.27 0.98
343.15 0.96 4.13 0.98 0.08 0.76 2.81 0.94
By comparing the tted parameters obtained, it is known that
the adsorptivity of MB dye towards the adsorbent surface is higher
than MG dye. This result can be explained from molecular view-
point that is small molecule gains easier access to the internal
pore-network of the adsorbent than large molecule. MB has lower
molecular weight therefore its molecular size is smaller than MG.
Moreover, higher afnityof MB dyeis alsocausedduetomorestabi-
lizedpositivechargeat thesulfur atomas aresult of theI inductive
effect of the benzene ring [38], resulting in stronger dipoledipole
interaction between negatively charged oxygen atom and posi-
tively charged sulfur atomcompared to negatively charged oxygen
atomand positively charged nitrogen atom.
Thevalueof nparameter inFreundlichmodel alsoincreases with
increase of temperature. This is due to faster mobility of the adsor-
bate molecules in bulk solution, which rise the randomness degree
of the system. The value of n parameter for rarasaponinbentonite
Fig. 3. Equilibriumplot and the models t for the adsorption of MB and MG in single systemonto bentonite at various temperatures.
270 A. Kurniawan et al. / Chemical Engineering Journal 189190 (2012) 264274
Fig. 4. Equilibriumplot and the models t for the adsorption of MB and MG in single systemonto rarasaponinbentonite at various temperatures.
is considered to be higher than bentonite due to the present of
organic modier in the 2:1 interlayer space, causing the solid
system to be more complex and heterogeneous. However, incon-
sistency results are noticed in this study that is the n values for
rarasaponinbentonite systems are lower than bentonite systems.
This result reveals that Freundlichmodel fails togive best represen-
tation from parameters justication. Therefore, its extended form
(extended Freundlich) is not further used to evaluate adsorption
equilibriumdata in binary system.
3.5.2. Adsorption isotherms for binary system
In this study, the extended Langmuir model was used to evalu-
ate adsorption equilibriumdata in binary system. The mathematic
expression of extended Langmuir model is given below:
q
e,i
=
q
m,i
K
L,i
C
e,i
1 +

n
j=1
K
L,i
C
e,i
(8)
For binary system, the extended Langmuir equation above
becomes:
q
e,1
=
q
m,1
K
L,1
C
e,1
1 +K
L,1
C
e,1
+K
L,2
C
e,2
(9)
q
e,2
=
q
m,2
K
L,2
C
e,2
1 +K
L,1
C
e,1
+K
L,2
C
e,2
(10)
where q
m
and K
L
in Eqs. (9) and (10) are Langmuir constants for
maximum adsorption capacity of the adsorbents (mmol/g) and
adsorptivity of the adsorbates (L/mmol) in single system, respec-
tively.
The applications of extended Langmuir model to represent
adsorption equilibrium data in binary system have been reported
in several studies [2024]. Fromthe previous studies, it was known
that the evaluation of binary adsorption equilibrium data using
extendedLangmuir model was performedbyinsertingthe values of
q
m
andK
L
parameters insingle systemtocalculate theoretical value
of q
e
and compared it with experimental results. However, such
procedure cannot be precisely implemented for binary adsorption
systemalthough it may give good correlation result. In binary sys-
tem, the competitions (total or partial) between adsorbate species
for the adsorption sites on the solid surface occur and may act as
the sorption-controlling factor. Such phenomena result in the cov-
erage onthe solidsurface by bothadsorbates withcertainfractional
loadings. Moreover, the adsorption potential on the surface is also
affected by lateral interaction or competition between adsorbate
species in the system. Fromthe statements above, it is obvious that
both K
L
and q
m
parameters fromsingle Langmuir model cannot be
adequately used to describe adsorptionbehaviors inbinary system.
Therefore, we proposeda revisitedmathematic equationfor q
m
and
K
L
parameters for binary adsorption systemas follows:
q
m(bin)
= q
m,1(sin)

1
+q
m,2(sin)

2
(11)
K
L,1(bin)
= K
L,1(sin)
exp

(12)
K
L,2(bin)
= K
L,2(sin)
exp

(13)
A. Kurniawan et al. / Chemical Engineering Journal 189190 (2012) 264274 271
Fig. 5. Comparison plotting between original-extended Langmuir (A and B) and proposed-extended Langmuir (C) models against binary experimental data (2.74mmol/L
MG +1.34mmol/L MB bentonite 303.15K).
where
1
and
2
are constants representing fractional loading of
each adsorbate species on the adsorbent surface. The mathematic
relationship between q
m
and is clearly understand that is the sur-
face of the adsorbent was occupied by both adsorbates with certain
loadings. In other hand, the mathematic relationship between K
L
and can be explained principally fromthe surface coverage view-
point that is the adsorption potential on the surface is weakened
when the adsorbate species competed each other for the adsorp-
tion sites. Since the competition between adsorbate species occurs
in binary system, therefore the values of adsorption potential (or
afnity) for each adsorbate should be lower than in single system.
Substitute Eqs. (11)(13) into extended Langmuir model give:
q
e,1(bin)
=
(q
m,1(sin)

1
+q
m,2(sin)

2
)K
L,1(sin)
exp(
2
/
1
)C
e,1(bin)
1 +K
L,1(sin)
exp(
2
/
1
)C
e,1(bin)
+K
L,2(sin)
exp(
1
/
2
)C
e,2(bin)
(14)
q
e,2(bin)
=
(q
m,1(sin)

1
+q
m,2(sin)

2
)K
L,2(sin)
exp(
1
/
2
)C
e,2(bin)
1 +K
L,1(sin)
exp(
2
/
1
)C
e,1(bin)
+K
L,2(sin)
exp(
1
/
2
)C
e,2(bin)
(15)
Fig. 6. Model t of proposed-extended Langmuir at different viewpoints against binary adsorption data (2.74mmol/L MG +1.34mmol/L MB bentonite 303.15K).
272 A. Kurniawan et al. / Chemical Engineering Journal 189190 (2012) 264274
Fig. 7. Model t of proposed-extended Langmuir at different viewpoints against binary adsorption data (2.74mmol/L MG +1.34mmol/L MB rarasaponinbentonite
303.15K).
Both equations above will return to single Langmuir model if only
one adsorbate species is presented in solution (i.e.
2
=0 in Eq.
(14) and
1
=0 in Eq. (15)). To get better understanding whether
our model is applicable to represent binary adsorption experi-
mental data, we compare the plot of experimental isotherm data
(2.74mmol/L MG +1.34mmol/L MB bentonite 303.15K) against
original-extended Langmuir (Eqs. (9) and (10)) and proposed-
extended Langmuir models (Eqs. (14) and (15)) in 3D mesh plot,
as presented in Fig. 5.
In Fig. 5, the blank data points represent the plot of C
e,MB
as x-
axis vs. C
e,MG
as y-axis vs. q
e,MB
as z-axis while the lled data points
denote the plot of C
e,MG
as x-axis vs. C
e,MB
as y-axis vs. q
e,MG
as z-
axis. For binary system, the plot of the model should t both data
points since the concentration of both adsorbate species in bulk
solution is in a dynamic balance with the interface concentration at
equilibriumcondition. Fig. 5Aclearly shows a very poor t between
the mesh plot and experimental data points since all data points lie
below the mesh plot. This result may conrm that evaluation of
binary adsorption data by comparing theoretical value of q
e
with
experimental results is not valid. Inother hand, Fig. 5BandCdenote
a very good tting between the models and experimental data.
However, the values of tted parameters obtained from original-
extended Langmuir model (Fig. 5B) are deviate away as those for
single system therefore this tting is not valid from parameters
interpretation viewpoint. To this end, the value of tted parameter
obtained (
1
) fromproposed-extended Langmuir model (Fig. 5C) is
reasonable with good correlation factor (R
2
=0.95), therefore this
proposed model is further used for evaluation of binary adsorption
equilibriumdata in this study.
The plotting of Eqs. (14) and (15) against binary experimen-
tal data (e.g. 2.74mmol/L MG +1.34mmol/L MB and 2.74mmol/L
MG +1.34mmol/L MB) onto bentonite and rarasaponinbentonite
at 303.15K was performed by SigmaPlot software and the results
are displayed in 3D-mesh plot (Figs. 6 and 7). The tted parameter
(
1
) was obtained by nonlinear least square tting and the results
are summarizedinTable 4. The values of q
m(bin)
, K
L,1(bin)
andK
L,2(bin)
parameters are calculated by inserting the value of
1
parameter
into Eq. (11), Eq. (12) and Eq. (13), respectively.
FromTable 4, it can be seen the value of
1
, which refers to the
fractional surface coverage by MG, decreases as the concentration
of MB in the dye mixture increases. At the same concentration (i.e.
2.74mmol/L (1000mg/L) MG +2.67mmol/L (1000mg/L) MB), the
fractional loading of MB (
2
) is larger than MG, denoting that MB
is preferentially adsorbed on the adsorbent surface than MG. This
result is consistent with adsorption characteristics in single system
that is the adsorptivity of MB is larger than MG, causing more cov-
erage on the adsorbent surface. Increasing temperature also leads
to the increase in afnity constants for both dyes, denoting that
higher temperature favors the adsorption process. Higher value of
q
m
for rarasaponinbentonite is due to larger interlamellar spaces
sothat more dye molecules canoccupythe internal pore inthe solid
matrix.
A. Kurniawan et al. / Chemical Engineering Journal 189190 (2012) 264274 273
Table 4
Fitted and calculated parameters for MB and MG adsorption in binary systemat various dye mixtures and temperatures.
Adsorbent Dye mixture T (K) Fitted
parame-
ter
Calculated parameters R
2

1

2
K
L,1(bin)
K
L,2(bin)
q
m(bin)
Bentonite 2.74mmol/L MG +0.67mmol/L MB 303.15 0.57 0.43 0.74 0.44 0.51 0.88
2.74 mmol/L MG +1.34mmol/L MB 0.51 0.49 0.60 0.59 0.51 0.95
2.74 mmol/L MG +2.01mmol/L MB 0.42 0.58 0.39 0.81 0.52 0.94
2.74 mmol/L MG +2.67mmol/L MB 0.36 0.64 0.26 0.95 0.52 0.91
2.74 mmol/L MG +0.67mmol/L MB 323.15 0.55 0.45 0.94 0.73 0.53 0.91
2.74 mmol/L MG +1.34mmol/L MB 0.49 0.51 0.75 0.94 0.54 0.89
2.74 mmol/L MG +2.01mmol/L MB 0.38 0.62 0.42 1.15 0.54 0.88
2.74 mmol/L MG +2.67mmol/L MB 0.33 0.67 0.28 1.30 0.54 0.93
2.74 mmol/L MG +0.67mmol/L MB 343.15 0.51 0.49 1.24 1.18 0.56 0.92
2.74 mmol/L MG +1.34mmol/L MB 0.45 0.55 0.95 1.47 0.57 0.91
2.74 mmol/L MG +2.01mmol/L MB 0.34 0.66 0.46 1.99 0.57 0.93
2.74 mmol/L MG +2.67mmol/L MB 0.31 0.69 0.35 2.13 0.57 0.94
Rarasaponinbentonite 2.74 mmol/L MG +0.67mmol/L MB 303.15 0.53 0.47 0.70 0.57 0.66 0.95
2.74 mmol/L MG +1.34mmol/L MB 0.46 0.54 0.53 0.75 0.66 0.91
2.74 mmol/L MG +2.01mmol/L MB 0.38 0.62 0.34 0.95 0.67 0.89
2.74 mmol/L MG +2.67mmol/L MB 0.32 0.68 0.20 1.09 0.67 0.91
2.74 mmol/L MG +0.67mmol/L MB 323.15 0.49 0.51 0.82 0.97 0.87 0.94
2.74 mmol/L MG +1.34mmol/L MB 0.44 0.56 0.65 1.16 0.87 0.93
2.74 mmol/L MG +2.01mmol/L MB 0.35 0.65 0.36 1.48 0.88 0.91
2.74 mmol/L MG +2.67mmol/L MB 0.30 0.70 0.22 1.66 0.88 0.95
2.74 mmol/L MG +0.67mmol/L MB 343.15 0.47 0.53 1.19 1.70 0.96 0.89
2.74 mmol/L MG +1.34mmol/L MB 0.41 0.59 0.87 2.06 0.97 0.92
2.74 mmol/L MG +2.01mmol/L MB 0.31 0.69 0.40 2.64 0.97 0.94
2.74 mmol/L MG +2.67mmol/L MB 0.27 0.73 0.25 2.85 0.97 0.91
4. Conclusions
Preparation of rarasaponinbentonite via microwave radiation
has been demonstrated in this work. Batch adsorption experi-
ments at three different temperatures were performed to test
adsorption performance of bentonite and rarasaponinbentonite
for the removal of methylene blue and malachite green in sin-
gle and binary systems. Experimental isotherms in single systems
were analyzed by Langmuir and Freundlich models and their t-
ting results were contrasted. Fromphysical meaning interpretation
of the model parameters, it was found that Langmuir model gave
the best representation. For binary system, experimental isotherms
were presented in the 3D mesh plot and were interpreted using
extended Langmuir model incorporating fractional surface cover-
age (). The results show that recently proposed modication of
extended Langmuir model can be used to describe binary adsorp-
tionexperimental datawell withreasonablettedparameter value.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.cej.2012.02.070.
References
[1] S. Chowdhury, P. Saha, Sea shell powder as a new adsorbent to remove basic
green 4 (malachite green) from aqueous solutions: equilibrium, kinetic and
thermodynamic studies, Chem. Eng. J. 164 (2010) 168177.
[2] C.I. Pearce, J.R. Lloyd, J.T. Guthrie, Theremoval of colour fromtextilewastewater
using whole bacterial cells: a review, Dyes Pigments 58 (2003) 179196.
[3] S.D. Khattri, M.K. Singh, Colour removal fromdye wastewater using sugar cane
dust as an adsorbent, Adsorpt. Sci. Technol. 17 (1999) 269282.
[4] O. Hamdaoui, Dynamic sorption of methylene blue by cedar sawdust and
crushed brick in xed bed columns, J. Hazard. Mater. 138 (2006) 293303.
[5] A.R. DiSanto, J.G. Wagner, Pharmacokinetics of highly ionized drugs. II. Methy-
lene blue absorption, metabolism, and excretion in man and dog after oral
administration, J. Pharm. Sci. 61 (1972) 10861090.
[6] J.W. Harvey, A.S. Keitt, Studies of the efcacy and potential hazards of methy-
lene blue therapy in aniline-induced methaemoglobinaemia, Br. J. Haematol.
54 (1983) 2941.
[7] E. Sudova, J. Machova, Z. Svobodova, T. Vesely, Negative effects of malachite
green and possibilities of its replacement in the treatment of sh eggs and sh:
a review, Vet. Med. 52 (2007) 527539.
[8] C. Fernandez, M.S. Larrechi, M.P. Callao, An analytical overview of processes
for removing organic dyes from wastewater efuents, Trends Anal. Chem. 29
(2010) 12021211.
[9] V.K. Gupta, Suhas, Application of low-cost adsorbents for dye removal a
review, J. Environ. Manage. 90 (2009) 23132342.
[10] A. Kurniawan, H. Sutiono, H.-Y. Ju, F.E. Soetaredjo, A. Ayucitra, A. Yudha, S.
Ismadji, Utilization of rarasaponin natural surfactant for organo-bentonite
preparation: application for methylene blue removal from aqueous efuent,
Micropor. Mesopor. Mater. 142 (2011) 184193.
[11] S. Netpradit, P. Thiravetyan, S. Towprayoon, Evaluation of metal hydroxide
sludge for reactive dye adsorption in a xed-bed column system, Water Res.
38 (2004) 7178.
[12] G. Annadurai, R.S. Juang, D.J. Lee, Use of cellulose-based wastes for
adsorption of dyes from aqueous solution, J. Hazard. Mater. 92 (2002)
263274.
[13] Y.S. Ho, C.C. Chiang, Y.C. Hsu, Sorption kinetics for dye removal fromaqueous
solution using activated clay, Sep. Purif. Technol. 36 (2001) 24732488.
[14] Q. Li, Q.-Y. Yue, H.-J. Sun, Y. Su, B.-Y. Gao, A comparative study on the prop-
erties, mechanisms and process designs for the adsorption of non-ionic or
anionic dyes onto cationic-polymer/bentonite, J. Environ. Manage. 91 (2010)
16011611.
[15] O. Gok, A.S. Ozcan, A. Ozcan, Adsorption behavior of a textile dye of Reactive
Blue 19 from aqueous solutions onto modied bentonite, Appl. Surf. Sci. 256
(2010) 54395443.
[16] H.A. Patel, R.S. Somani, H.C. Bajaj, R.V. Jasra, Synthesis and characterization
of organic bentonite using Gujarat and Rajasthan clays, Curr. Sci. 92 (2007)
10041009.
[17] B. Zohra, K. Aicha, S. Fatima, B. Nourredine, D. Zoubir, Adsorption of Direct Red
2 on bentonite modied by cetyltrimethylammonium bromide, Chem. Eng. J.
136 (2008) 295305.
[18] P. Baskaralingam, M. Pulikesi, V. Ramamurthi, S. Sivanesan, Modied hec-
torites and adsorption studies of a reactive dye, Appl. Clay Sci. 37 (2007)
207214.
[19] J. Li, L. Zhu, W. Cai, Microwave enhanced-sorption of dyestuffs to
dual-cation organobentonites from water, J. Hazard. Mater. 136 (2006)
251257.
[20] K.K.H. Choy, J.F. Porter, G. McKay, Langmuir isotherm models applied to the
multicomponent sorption of acid dyes from efuent onto activated carbon, J.
Chem. Eng. Data 45 (2000) 575584.
[21] K.K.H. Choy, S.J. Allen, G. McKay, Multicomponent equilibrium studies for
the adsorption of basic dyes from solution on lignite, Adsorption 11 (2005)
255259.
[22] K.S. Baig, H.D. Doan, Ju Wu, Multicomponent isotherms for biosorption of Ni
2+
and Zn
2+
, Desalination 249 (2009) 429439.
[23] G. Atun, E.T. Acar, Competitive adsorption of basic dyes onto calcite in single
and binary component systems, Sep. Sci. Technol. 45 (2010) 14711481.
274 A. Kurniawan et al. / Chemical Engineering Journal 189190 (2012) 264274
[24] S. Chakraborty, J.K. Basu, S. De, S.D. Gupta, Adsorption of reactive dyes from
textile efuent using sawdust as the adsorbent, Ind. Eng. Chem. Res. 45 (2006)
47324741.
[25] A.K. Rahardjo, M.J.J. Susanto, A. Kurniawan, N. Indraswati, S. Ismadji, Modied
ponorogo bentonite for the removal of ampicillin fromwastewater, J. Hazard.
Mater. 190 (2011) 10011009.
[26] P.C.C. Faria, J.J.M. Orfao, M.F.R. Pereira, Adsorption of anionic and cationic dyes
on activated carbons with different surface chemistries, Water Res. 38 (2004)
20432052.
[27] A. Saeed, M. Sharif, M. Iqbal, Application potential of grapefruit peel as dye
sorbent: kinetics, equilibrium and mechanism of crystal violet adsorption, J.
Hazard. Mater. 179 (2010) 564572.
[28] D.D. Do, Adsorption Analysis: Equilibria and Kinetics, Imperial College Press,
London, 1998.
[29] A. Seker, T. Shahwan, A.E. Eroglu, S. Yilmaz, Z. Demirel, M.C. Dalay, Equilibrium,
thermodynamic andkinetic studies for the biosorptionof aqueous lead(II), cad-
mium(II) and nickel(II) ions on Spirulina platensis, J. Hazard. Mater. 154 (2008)
973980.
[30] A. Kurniawan, A.N. Kosasih, J. Febrianto, H.-Y. Ju, J. Sunarso, N. Indraswati,
S. Ismadji, Evaluation of cassava peel waste as lowcost biosorbent for Ni-
sorption: equilibrium, kinetics, thermodynamics and mechanism, Chem. Eng.
J. 172 (2011) 158166.
[31] S. Kalyani, P.S. Rao, A. Krishnaiah, Removal of nickel(II) from aqueous solu-
tions using marine macroalgae as the sorbing biomass, Chemosphere 57 (2004)
12251229.
[32] A. Kheni, B. Zohra, B. Kahina, H. Houari, D. Zoubir, Removal of 2,4-DCP from
wastewater by CTAB/bentonite using one-step and two-steps methods: a com-
parative study, Chem. Eng. J. 146 (2009) 345354.
[33] E.N. El Qada, S.J. Allen, G.M. Walker, Adsorption of basic dyes from
aqueous solution onto activated carbons, Chem. Eng. J. 135 (2008)
174184.
[34] I. Langmuir, The constitution and fundamental properties of solids and liquids,
J. Am. Chem. Soc. 38 (1916) 22212295.
[35] T.W. Weber, R.K. Chakravorti, Pore and solid diffusion models for xed-bed
adsorbers, AIChE J. 20 (1974) 228238.
[36] H.M.F. Freundlich, Over the adsorption in solution, J. Phys. Chem. 57 (1906)
385471.
[37] X. Lu, Comment on thermodynamic and isotherm studies of the biosorption
of Cu(II), Pb(II), and Zn(II) by leaves of saltbush (Atriplex canescens), J. Chem.
Thermodyn. 40 (2008) 739740.
[38] A.E. Ofomaja, Kinetic study and sorption mechanism of methylene blue and
methyl violet onto Mansonia (Mansonia altissima) wood sawdust, Chem. Eng. J.
143 (2008) 8595.