7634 Langmuir 2007, 23, 7634-7643

Chitosan Derivatives as Biosorbents for Basic Dyes

Nikolaos K. Lazaridis,* George Z. Kyzas, Alexandros A. Vassiliou, and
Dimitrios N. Bikiaris
DiVision of Chemical Technology, School of Chemistry, Aristotle UniVersity,
GR-541 24 Thessaloniki, Greece
ReceiVed February 13, 2007. In Final Form: April 24, 2007

The scope of this study was to prepare and evaluate chitosan derivatives as biosorbents for basic dyes. This was
achieved by grafting poly (acrylic acid) and poly (acrylamide) through persulfate induced free radical initiated
polymerization processes and covalent cross-linking of the prepared materials. Remacryl Red TGL was used as the
cationic dye. Equilibrium sorption experiments were carried out at different pH and initial dye concentration values.
The experimental equilibrium data for each adsorbent-dye system were successfully fitted to the Langmuir, Freundlich
and pH-dependent Langmuir-Freundlich sorption isotherms. Thermodynamic parameters of the adsorption process
such as ∆G°, ∆H°, and ∆S° were calculated. The negative values of free energy reflected the spontaneous nature of
adsorption. The typical dependence of dye uptake on temperature and the kinetics of adsorption indicated the process
to be chemisorption. The grafting modifications greatly enhanced the adsorption performance of the biosorbents,
especially in the case of powdered cross-linked chitosan grafted with acrylic acid, which exhibited a maximum
adsorption capacity equal to 1.068 mmol/g. Kinetic studies also revealed a significant improvement of sorption rates
by the modifications. Diffusion coefficients of the dye molecule were determined to be of the order 10-13 - 10-12
m2/s. Furthermore, desorption experiments affirmed the regenerative capability of the loaded material.

1. Introduction because the range of its applications has enormously expanded

in various fields, including biotechnology, water treatment,
Dyeing effluent, one of the largest contributors to textile effluent medicine and veterinary medicine, membranes, cosmetics, and
and colored wastewater, has a seriously destructive impact on the food industry.8
the environment. Several methods have been developed to remove Chitosan has a very high affinity for most classes of dyes such
color from dyehouse effluents, varying in effectiveness, economic as reactive, direct, and disperse, expressing a lack of affinity
cost, and environmental impact (of the treatment process itself). only for basic dyes.4,9,10 These characteristics are mainly due to
Among all of the treatments proposed, the sorption of dye its high content of amine functional groups, which give a cationic
molecules onto a substrate (sorbent) can be a very effective, nature to the biopolymer, as well as its hydroxyl groups. Therefore,
low-cost method of color removal.1-3 chitosan has much potential as an inexpensive and effective
Activated carbon is the most commonly used method of dye adsorbent for almost all types of dyes, except basic, because of
removal by sorption, although the performance is dependent on its innate cationic nature.
the type of carbon used, the characteristics of the wastewater, The object of the current study was to improve the adsorption
and the type of dye. However, the cost and regeneration difficulties capabilities of chitosan for basic dyes. To accomplish this
of the sorbent have compelled researchers to focus on alternative endeavor, chitosan was suitably modified with the aim of
low-cost sorbents.4 Among these, polysaccharides such as chitin, introducing anionic and nonanionic moieties but was also capable
starch, and their derivatives (chitosan and cyclodextrin) deserve of forming hydrogen bonds with groups on the chitosan
particular attention.5-7 macromolecule. The adsorption performance of similar chitosan
Chitosan (poly-â-(1f4)-2-amino-2-deoxy-D-glucose) is an derivatives has not been substantially studied. Also, to increase
aminopolysaccharide, which is a cationic polymer produced by the resistance of the final products to chemical and biological
the N-deacetylation of chitin. Chitin (poly-â-(1f4)-N-acetyl- degradation, decrease their solubility at extreme pH values, and
D-glucosamine) constitutes one of the most abundant natural improve their mechanical properties, with the prospect of using
biopolymers, second only to cellulose. It is mostly found in the the prepared adsorbents in packed columns, cross-linking
exoskeletons of crustaceans, in the cartilage of mollusks, in the reactions were carried out. Although heterogeneous chitosan
cuticles of insects, and in the cell walls of microorganisms. cross-linking has been thoroughly investigated in the literature,11
Because of its macromolecular structure, chitosan exhibits many especially through the use of dialdehydes, the effect of such a
characteristics that have been the cause of much recent attention modification on the adsorption performance of the prepared
chitosan derivatives has not been previously carried out and
* To whom correspondence should be addressed. E-mail: nlazarid@ evaluated. The adsorption-desorption of remacryl red TGL by
[Link]. Tel: +32310 997807. Fax: +32310 997759. the prepared biopolymers was extensively studied and evaluated.
(1) San Miguel, G.; Lambert, S. D.; Graham, N. J. Chem. Technol. Biotechnol.
2006, 81, 1685-1696. 2. Materials and Methods
(2) Crini, G. Bioresource Technol. 2006, 97, 1061-1085.
(3) Bhatnagar, A.; Minocha, A. K. Indian J. Chem. Technol. 2006, 13, 203- 2.1. Materials. High-molecular-weight chitosan (Ch) was obtained
127. from Sigma-Aldrich and purified by extraction with acetone in a
(4) Blackburn, S. R. EnViron. Sci. Technol. 2004, 38, 4905-4909.
(5) Guibal, E.; Van Vooren, M.; Dempsey, B.; Roussy, J. Sep. Sci. Technol. (8) Rinaudo, M. Prog. Polym. Sci. 2006, 31, 603-632.
2006, 41, 2487-2514. (9) Gibbs, G.; Tobin, J.; Guibal, E. Ind. Eng. Chem. Res. 2004, 43, 1-11.
(6) Crini, G. Prog. Polym. Sci. 2005, 30, 38-70. (10) Chiou, M.; Li, H. Chemosphere 2003, 50, 1095-1105.
(7) Yi, H.; Wu, L.; Bentley, W.; Ghodssi, R.; Rubloff, G.; Culver, J.; Payne, (11) Berger, J.; Reist, M.; Mayer, J. M.; Felt, O.; Peppas, N. A.; Gurny, R.
G. Biomacromolecules 2005, 6, 2881-2894. Eur. J. Pharm. Biopharm. 2004, 57, 19-34.

10.1021/la700423j CCC: $37.00 © 2007 American Chemical Society

Published on Web 05/27/2007
Chitosan DeriVatiVes as Biosorbents for Basic Dyes Langmuir, Vol. 23, No. 14, 2007 7635

Spectrum 1000) using KBr pellets containing the prepared materials.

The resolution of each spectrum was 2 cm-1, and the number of
co-added scans was 64. The spectra presented are baseline corrected
and converted to absorbance mode. The surface areas of the beads
were determined with a Micromeritics BET surface area analyzer,
model TriStar 3000, by means of the adsorption of ultrapure nitrogen
at 77 K.
2.4. Experimental Procedure. Kinetic Experiments. Batch kinetic
experiments were performed by mixing a fixed amount of sorbent
(1 g/L) with 50 mL of an aqueous dye solution of 0.235 mmol/L
(100 mg/L) concentration. The mixture was shaken for 24 h, and
during this time, samples were collected at fixed intervals. After the
Figure 1. Three-dimensional molecular structure of remacryl red spectrophotometric analysis of each sample, the concentration of
TGL (red, oxygen atoms; dark gray, carbon atoms; light gray, dye in the aqueous solution at fixed time (Ct) was calculated. Although
hydrogen atoms; green, chloride atoms; and blue, nitrogen atoms). the pH value of dye baths when using basic dyes is acidic to neutral
so that better fixation of dyestuffs onto the fabric can be achieved,
Soxhlet apparatus for 24 h, followed by drying under vacuum at the dyewaste equilibration tank of dyehouses is nonetheless rather
room temperature. The average molecular weight was estimated to alkaline (pH ∼10).4 Thus, the pH value of the solution was adjusted
be 3.55 × 105, and the degree of deacetylation was 82 wt %, as to 10 by microadditions of 0.1 M NaOH.
determined according to the procedures described by Rinaudo.8 Equilibrium Experiments. The effect of the initial dye concentration
Acrylamide (Aam) 97% p.a. was purchased from Sigma-Aldrich was determined by placing 1 g/L of the adsorbent in contact with
and was used without further purification. Acrylic acid (Aa) received 50 mL of aqueous solutions containing different dye concentrations
from Merck was purified by distillation under vacuum. Potassium (0.047-1.175 mmol/L), and the suspensions were shaken for 24 h
persulfate (KPS) obtained from Merck was used as received. at pH 10.
Glutaraldehyde (GA, 50 wt % in water) was received and used as The influence of pH on the adsorption process was studied by
reagent grade from Sigma-Aldrich. All solvents were of analytical mixing 1 g/L of sorbent with 50 mL of an aqueous dye solution of
grade. Remacryl red TGL (C19H25Cl2N5O2) (Figure 1) with a 0.235 mmol/L concentration. The pH value, ranging from 2 to 12,
molecular weight of 426.34 g/mol, which was kindly supplied by was kept constant throughout the adsorption process. The suspension
Hochest (Germany), was selected as the sorbate. was shaken for 24 h using a bath to control the temperature at 298
2.2. Preparation of the Biosorbents. Well-documented persul- ( 1 K (Julabo SW-21C).
fate-induced free radical reaction procedures were utilized for the All the sorption experiments were conducted in the presence of
realization of the grafting reactions onto the chitosan backbone, as 0.001 mol/L NaCl as the background electrolyte because it is used
reported in the literature.12-15 Optimum reaction parameters, as a stimulator in the dyeing processes. In general, the added salts
specifically, temperature, time, solvent volume, and relative con- may have two functions: (i) they may screen the electrostatic
centrations of the monomer and the initiator, were used for the interaction of opposite charges between sorbent and dye molecules,
grafting of Aa15 and Aam.16 The exact details of the experimental and an increase in salt concentration could decrease the amount
procedures followed are described in Supporting Information. The sorbed; (ii) they may enhance the degree of dissociation of the dye
final grafting percentages, determined on the basis of the percentage molecules and facilitate the amount of dye sorbed.10,17
weight increase of the final product relative to the initial weight of Desorption Experiments. After the adsorption experiments,
chitosan (%G ) 100 × (W2 - W1)/W1 where W1 and W2 denote the samples were collected and filtered using fixed pore-size mem-
weights of the initial dry chitosan and grafted chitosan after extraction branes. Small fractions of the dye (1 to 2%) and the adsorbent (1%)
and drying, respectively), were estimated to be approximately 230% were retained on the filter membrane; these small variations due to
poly(acrylamide) for Ch-g-Aam and 300% poly(acrylic acid) for filtration were neglected. Desorption experiments were performed
Ch-g-Aa. All prepared products were ground to fine powders of by mixing the collected amount of chitosan after adsorption with
between 75 and 125 µm. aqueous solutions over the pH range of 2-12. As above, after 24
The biosorbent particles were heterogeneously chemically cross- h of shaking at a temperature of 298 K, the samples were collected,
linked with bifunctional reagent glutaraldehyde (GA), rendering the and spectrophotometric analysis revealed the optimum desorption
substances insoluble in acidic media and improving their resistance pH.
to chemical and biological degradation and their mechanical strength 2.5. Analysis. Samples were extracted using a syringe, filtered
and abrasion resistance ((Ch)c, (Ch-g-Aam)c, (Ch-g-Aa)c). This through a 50 µm pore size membrane, and then analyzed spectro-
reaction has been extensively reviewed in the literature for photometrically by monitoring the absorbance of the dyestuff using
chitosan,11,16 whereas the respective copolymers’ heterogeneous a UV-vis spectrophotometer (model U-2000, Hitachi). The deter-
reaction has not been previously studied. Whereas cross-linking mined λmax wavelength value for remacryl red TGL dye was 488
usually leads to lower adsorption capacities, it is deemed necessary nm. Prior to the adsorption experiments, the effect of pH over the
because of the unacceptable properties of the pristine material for calibration curves of each dye was studied, but no significant deviation
use in packed columns. The experimental procedure is found in was observed. The amount of dye uptake at equilibrium, Qe, was
Supporting Information. The penetration of GA into the gelled bead calculated using the mass balance equation. The amount of dye
and the fraction of amine groups actually cross-linked were not sorbed or desorbed at different pHs was expressed as a percentage
measured, nor were the mechanical properties of the gelled and of the removal.
dried particles. 2.6. Theory. Kinetic Modeling. The study of sorption is important
2.3. Characterization Techniques. FTIR spectra of the biosor- in wastewater treatment because it provides valuable information on
bents were obtained using a Perkin-Elmer FTIR spectrometer (model the reaction pathways and the mechanism of sorption. In addition,
predicting the solute uptake rate is of utmost importance in designing
(12) Jayakumar, R.; Prabaharan, M.; Reis, R. L.; Mano, J. F. Carbohydr. an appropriate wastewater treatment plant because it can control the
Polym. 2005, 62, 142-158. residence time of solute at the solid-solution interface.18-20
(13) Kumbar, G. S.; Soppimath, S. K.; Aminabhavi, M. T. J. Appl. Polym. Sci. The main issue when searching for an appropriate sorption
2003, 87, 1525-1536.
(14) Yasdani-Pedram, M.; Retuert, J.; Quijada, R. Macromol. Chem. Phys. mechanism is to select a mathematical model that not only fits the
2000, 201, 923-930.
(15) Yasdani-Pedram, M.; Lagos, A.; Retuert, P. Polym. Bull. 2002, 48, 93- (17) Inbaraj, B.; Chien, J.; Ho, G.; Yang, J.; Chen, B. Biochem. Eng. J. 2006,
98. 31, 204-215.
(16) Knaul, J. Z.; Hudson, S. M.; Creber, K. A. M. J. Polym. Sci., Part B: (18) Sontheimer, H.; Crittenden, J.; Summers, S. ActiVated Carbon for Water
Polym. Phys. 1999, 37, 1079-1094. Treatment, 2nd ed.; DVGW: Karlsruhe, Germany, 1988.
7636 Langmuir, Vol. 23, No. 14, 2007 Lazaridis et al.

data with satisfactory accuracy but also complies with a reasonable f(X) ) (1 - X)n (5)
sorption mechanism. Generally, several steps are involved during
the sorption process by sorbent particles: (i) diffusion of the dye Most simple reactions have integer orders, n, between 0 and 3. For
from the solution to the boundary film surrounding the particle (bulk sorption, the value of n was customarily 1 or 2. Integrating eq 2 for
diffusion); (ii) diffusion from the film to the particle surface (external boundary conditions (i) t ) 0, X ) 0 and (ii) t ) t, X ) X, and n
diffusion); (iii) diffusion from the surface to the internal sites (surface ) 1, 2 results in pseudo-first-order (eq 6) and pseudo-second-order
diffusion or pore diffusion); and (iv) uptake that involves several kinetic (eq 7) models
physicochemical mechanisms. Providing sufficient agitation to avoid
particle and solute gradients in the batch reactor allows bulk diffusion Ct ) C0 - (C0 - Ce)(1 - e-k1t) (6)
to be neglected.20

( )
To compare measurements from various experiments for a kinetic 1
investigation, it is necessary to introduce a dimensionless degree of Ct ) C0 - (C0 - Ce) 1 - (7)
1 + k2t
conversion. Thus, by normalizing the remaining dye concentration,Ct,
with respect to some reference value, an index of sorption is defined
where k1 and k2 (min-1) are the rate constants for the pseudo-first-
by eq 120
order and pseudo-second-order adsorption models, respectively.
Sorption Isotherms. For the effective design and operation of a
C0 - Ct
X) (1) wastewater treatment plant (i.e., for dye removal) processing used
C0 - Ce sorbent materials, it is necessary to quantify the ultimate capacity
of the sorbent. Once the ultimate capacity is defined from adsorption
where C0 and Ce are the initial and the equilibrium concentrations equilibria, it is possible to estimate the lowest sorbent usage. Tóth
of dye in the aqueous solution (mmol L-1). investigated the dependency of the characteristic function Ψ(Ce) )
External Diffusion. At the beginning (0-15 min) of the test, where (d lnCe/d ln Qe) - 1 on the liquid-phase concentration.18
the surface concentration is very small compared to the concentration If Ψ(Ce) is proportional to the concentration in the aqueous
in the solution and only external mass transfer resistance controls phase, then the Langmuir equation23 is, after integration,
the adsorption rate,18 the solution of the mass transfer equation is
given by eq 2 QmaxCe
Qe ) (8)
(1/b) + Ce
Ct A
ln ) -kf t (2)
C0 V If Ψ(Ce) is constant, then the Freundlich24 equation is, after
integration,
where kf is the initial external mass-transfer coefficient (m/min), A
is the sorbent exchange surface (m2), and V the volume of the solution. Qe ) KFCe1/n (9)
By plotting ln Ct/C0 versus t, the kf coefficient can be determined.
Intraparticle Diffusion. Once the adsorbate molecule accumulates The most promising extensions to the Langmuir and Freundlich
on the outer surface of the sorbent, internal diffusion commences. isotherms are based on the generalized Langmuir and generalized
In fact, it is highly likely that both external and internal diffusion exponential isotherms. Such an isotherm is the Langmuir-Freundlich
occur simultaneously. However, only one mechanism will control (L-F, eq 10) one, which is essentially the Freundlich isotherm
the whole sorption process. The high complexity of the diffusion approaching a maximum at high concentration.19
equations can be simplified to calculate the corresponding parameters.
Intraparticle diffusion D, which is the sum of pore and surface Qmax (KLFCe)â
diffusion, may be calculated from eq 321,22 Qe ) (10)
1 + (KLFCe)â

[
X ) 1 - exp - ( 4πDt
2.3d2)] 1/2
(3) To incorporate the pH dependence into eq 10, one possible form is25

where D is the intraparticle diffusion coefficient and d is the particle Qmax (KLFCe)â
Qe ) k(pH) (11)
diameter. 1 + (KLFCe)â
The intraparticle diffusion coefficient can be calculated from the
half-time for adsorption (i.e., by substituting X ) 0.50). Throughout where k(pH) is a function of pH that matches the observed pH
the adsorption literature, many simple equations are reported for the dependence. This form of the modified L-F isotherm empirically
calculation of diffusion coefficients D from kinetic adsorption data. incorporates pH into the standard isotherm as an additional
However, the results may vary from model to model.21 independent variable. The function k(pH) can take any form that is
Reaction Kinetics. Kinetic studies customarily utilize the following suitable. In this case, the following expression was used
basic conversion rate equation (eq 4)20
1
dX k(pH) ) (12)
) k(T) f(X) ) k(T)(1 - X)n (4) 1 + fe-gpH
dt
where Qe is the equilibrium dye concentration in the solid phase,
where f(X) is a conversion-dependent function and k(T) is the reaction Qmax is the maximum amount of sorption (mmol/g), b is the Langmuir
rate constant. Sorption has been accepted to follow Arrhenius kinetics. sorption equilibrium constant (L/mmol), KLF is the Langmuir-
The analysis originally developed for isothermal homogeneous Freundlich constant (L/mmol)1/â, KF is the Freundlich constant
reactions was also reported to apply to sorption. Thus, f(X) could representing the sorption capacity (mmol1 - 1/n L1/n/g), n is the constant
be represented by eq 5, assuming a constant system volume: depicting the sorption intensity, â is the Langmuir-Freundlich
heterogeneity constant, and f and g are empirical parameters of the
(19) Ho, Y. S.; Ng, J. C. Y.; McKay, G. Sep. Purif. Method 2000, 29, 189-
232. L-F equation.
(20) Lazaridis, N. K.; Karapantsios, T. D.; Georgantas, D. Water Res. 2003,
37, 3023-3033. (23) Langmuir, I. J. Am. Chem. Soc. 1918, 40, 1361-1368.
(21) Khraisheh, M. A. M.; Al-Degs, Y. S.; Allen, S. J.; Ahmad, M. N. Ind. (24) Freundlich, H. M. F. Z. Phys. Chem. 1906, 57, 384-470.
Eng. Chem. Res. 2002, 41, 1651-1657. (25) Decker, D. L.; Papelis, C.; Tyler, S. W.; Logsdon, M. J.; Šimůneck, J.
(22) Tseng, R. L.; Tseng, S. K. J. Colloid Interface Sci. 2005, 287, 428-437. Vadose Zone J. 2006, 5, 419-429.
Chitosan DeriVatiVes as Biosorbents for Basic Dyes Langmuir, Vol. 23, No. 14, 2007 7637

The Langmuir isotherm assumes a monolayer coverage of CSe

adsorbate over a homogeneous adsorbent surface, and the sorption Kc ) (17)
CAe
of each molecule onto the surface has equal sorption activation
energy. The Freundlich isotherm is derived by assuming a
heterogeneous surface with a nonuniform distribution of heat of where CSe and CAe are the equilibrium concentrations of dye molecules
adsorption over the surface, and multilayer reversible sorption can in the solution and on the adsorbent, respectively.
be expressed. The main characteristics of the Langmuir isotherm Modeling of equilibrium and kinetic data was performed by a
can be expressed by a dimensionless constant separation factor or nonlinear regression method that involves the Levenberg-Marquardt
equilibrium parameter RL, which is given by the following method.28
expression26
3. Results and Discussion
1
RL ) (13) 3.1. Characterization of the Prepared Biosorbents. Ho-
(1 + bC0)
mogeneous graft copolymerization of the vinyl monomers onto
where b is the Langmuir constant. The value of RL indicates the the chitosan backbone was carried out in an aqueous solution
shape of the isotherm to be either unfavorable (RL > 1) or linear (RL using KPS as the initiator. The exact mechanism of such reactions
) 1) or favorable (0 < RL < 1) or irreversible (RL ) 0). has not been well clarified. Several research groups have put
Certain properties of adsorbents are involved in all adsorber forward different mechanisms, even for the same reaction
designs. These are adsorbent densities and void fractions, isotherms, system.13,29,30 A possible mechanism is derived from the current
and data on mass-transfer kinetics and fixed-bed dynamics. Obvi- literature.12,14,30,31 Applying heat to KPS in an aqueous solution
ously, the cost of the adsorbent must also be taken into account. causes it to decompose to sulfate anionic radicals (SO•4 -). These
Capacity governs the capital cost of an adsorption unit. It not only anionic radicals, which are attracted by the amino groups of
dictates the amount of adsorbent required but also establishes the chitosan (bearing in mind that polysaccharides are predominantly
volume of the adsorber vessels.27
Mass-transfer kinetics is a catchall term related to intraparticle
oxidized through C2-C3 bond cleavage32 and that chitosan can
mass-transfer resistance. It is important because it controls the cycle act as a weak reducing agent), react together with chitosan and
time of a fixed-bed adsorption process. Fast kinetics, or a high rate form the active sites on the biomacromolecular backbone (Scheme
of diffusion, implies a desirably sharp breakthrough curve. Such a 1). The monomer is grafted onto these sites, which is subsequently
curve means that the effluent concentration remains level until the followed by propagation.
adsorbent is almost saturated and then rises sharply. Slow kinetics The chemical cross-linking of the chitosan backbone was
leads to a distended breakthrough curve in which the effluent carried out through the use of bifunctional reagent glutaraldehyde.
concentration begins fluctuating soon after an adsorption run. The Although cross-linking can also occur between the hydroxyl
effect of a distended breakthrough curve can be overcome by adding
groups of the polysaccharide and the cross-linking agent,
adsorbent at the product end or by increasing the cycle time, which
reduces the throughput per unit of adsorbent. Both of these options especially in an acid-catalyzed environment, it mainly occurs
increase the amount of adsorbent required. To compensate for slow through the amino groups because of their greater reactivity. It
diffusion, it is also possible to use small particles, but there is a is generally accepted that the cross-linking mechanism is a Schiff’s
corresponding sacrifice due to the increased pressure drop. The base reaction between the dialdehyde and the free amine on the
intraparticle diffusion D enables engineers to assess the diffusional chitosan backbone (Scheme 1).11 However, there is further
time constant d2/D. This time constant, in turn, is used with elapsed evidence suggesting that the reaction actually takes place between
time, t, to define the dimensionless time, Dt/d2. In general, the response the amino groups of chitosan and a hemiacetal.16 The cross-
of a spherical particle to a sudden change in composition is linked linking reaction rendered the products practically insoluble, even
to the dimensionless time by the function X ) f(1 - e-Dt/d2). Thus, in extremely acidic media. Their color became darker because
when searching for an effective (fast) adsorbent, it is usually safe
of the formation of chromophore groups (-CdN-). Furthermore,
to choose one having a large diffusivity, a small diameter, or both.27
Thermodynamics. By using the equilibrium constant b obtained the subsequently gelled particles were elastic and resilient to
for each temperature from the Langmuir model, the Gibbs free energy shear forces. The dry particle sizes of the cross-linked copolymers
(∆G°) can be calculated according to eq 14 remained virtually unaltered. However, because of the grafted
groups and the smaller degree of cross-linking (due to the
∆G° ) -RT ln K (14) decreased reactivity of the grafted products toward GA), the
final modified biopolymers swelled considerably more than the
where K corresponds to b in the Langmuir equation. The standard unmodified.
enthalpy change (∆H°) and standard entropy change (∆S°) of the The grafted copolymers, before the cross-linking reactions,
process could be calculated using eq 15
exhibited different solubility behaviors than chitosan itself.
Whereas chitosan dissolves at pH values lower than about 4, the
∆G° ) ∆H° - T∆S° (15)
copolymer grafted with poly(acrylamide) could dissolve at pH
If the plot of ∆G° versus T is linear, the values of ∆H° and ∆S° could values as high as 6.5. However, the copolymer grafted with
be calculated. poly(acrylic acid) was insoluble over the whole pH range. The
Desorption thermodynamics were obtained from the integrated decreased solubility of the latter copolymer can be explained by
van’t Hoff equation considering the probable inter- and/or intramolecular hydrogen
bonding or internal ammonium salt formation between the
∆H° ∆S°
ln Kc ) - + (16)
RT R (28) Marquardt, D. W. J. Soc. Ind. Appl. Math. 1963, 11, 431-441.
(29) Maddavinia, G. R.; Pourjavadi, A.; Hosseinzadeh, H.; Zohuriaan, M. J.
where Kc is the equilibrium constant for desorption calculated from Eur. Polym. J. 2004, 40, 1399-1407.
(30) Prashanth, K. V. H.; Tharanathan, R. N. Carbohyd. Polym. 2003, 54,
eq 17 343-351.
(31) Peppas, N. A. Hydrogels in Medicine and Pharmacy; CRC Press: Boca
(26) Hall, R. K.; Eagleton, L. C.; Acrivos, A.; Vermeulen, T. Ind. Eng. Chem. Raton, FL, 1987; Vol. 3, p 109.
Fundam. 1966, 5, 212-223. (32) Doba, T.; Rodehed, C.; Ranby, B. Macromolecules 1984, 17, 2512-
(27) Knaebel, K. S. Chem. Eng. 1999, April, 92-101. 2519.
7638 Langmuir, Vol. 23, No. 14, 2007 Lazaridis et al.

Scheme 1. Mechanism of the Grafting Reaction of Poly(acrylamide) and Poly(acrylic acid) onto the Chitosan Backbone and the
Cross-Linking Reaction with GA

carboxyl groups of the grafted poly(acrylic acid) chains and the spectra of the grafted chitosan copolymers are drastically altered
amino groups of chitosan. compared with the spectra of chitosan, exhibiting a similarity
The FTIR spectra of pure chitosan and the cross-linked chitosan with the pure homopolymersspoly(acrylamide) and poly(acrylic
are presented in Figure 2. The most typical absorption bands of acid). This is due to the presence of the homopolymers (230 and
chitosan are situated at 1663 and 1556 cm-1, corresponding to 300% grafting, respectively) and the strong absorption bands of
amide I and II, respectively, and a broad band appearing at ∼3300 the grafting macromolecules that mask the absorbance bands of
cm-1 as a result of the stretching vibration of O-H, the extension chitosan. In the spectra of chitosan grafted with poly(acrylamide),
vibration of N-H, and interhydrogen bonds of the polysaccharide. the amide carbonyl absorption band from the grafted chains
The amide II band may be overlapped by the NH2 bending band appears at 1672 cm-1. The absorption bands at 1728 cm-1 in the
at ∼1590 cm-1, with the presence of this group being confirmed spectra of chitosan grafted with poly(acrylic acid) corresponds
by the absorption occurring at ∼1420 cm-1 (which may be related to the carbonyl absorption from the grafted poly(acrylic acid),
to the presence of NH2 groups). The absorption bands at ∼1150 whereas the peaks at 799 and 611 cm-1 are characteristic of
(asymmetric stretching of the COC bridge), ∼1080, and ∼1030 poly(acrylic acid). The formation of the imine moiety by the
cm-1 (skeletal vibrations involving CO stretching) are charac- cross-linking reaction in the copolymers is not visible in the
teristic of chitosan’s saccharide structure.33,34 The cross-linking FTIR spectra as a result of masking by the strong carbonyl
reaction can be confirmed by the appearance of a strong absorption absorption bands of the grafted polymers.
peak at 1665 cm-1 in the spectra of the modified chitosan,
corresponding to the imine moiety formed as a result of the To elucidate the effect of the various modifications over the
reaction between the free amino groups of the chitosan backbone crystallinity of the prepared adsorbents, X-ray diffraction patterns
and the aldehyde groups of GA. As was expected, the FTIR were obtained for each sample. From the X-ray diffractogram
of chitosan (Figure 3), two characteristic reflection falls were
(33) Fernades, A. L. P.; Morais, W. A.; Santos, A. I. P.; de Araujo, A. M. L.; observed at 2θ ) 11 and 20°, typical of chitosan, which have
dos Santos, D. E. S.; dos Santos, D. S.; Pavinatto, F. J.; Oliveira, O. N.; Dantas, been assigned to crystal forms I and II, respectively. The pattern
T. N. C.; Pereira, M. R.; Fonseca, J. L. C. Colloid Polym. Sci. 2005, 284, 1- 9.
(34) de Vasconcelos, C. L.; Bezerril, B. M.; dos Santos, D. E. S.; Dantas, T. is similar for cross-linked chitosan, revealing an unaffected
N. C.; Pereira, M. R.; Fonseca, J. L. C. Biomacromolecules 2006, 7, 1245-1252. crystallinity because the reaction was carried out in a hetero-
Chitosan DeriVatiVes as Biosorbents for Basic Dyes Langmuir, Vol. 23, No. 14, 2007 7639

Figure 2. FTIR spectra of (A) Ch, (B) (Ch)c loaded with dye, (C)
(Ch)c, (D) (Ch-g-Aam)c loaded with dye, (E) (Ch-g-Aam)c, (F)
(Ch-g-Aa)c loaded with dye, (G) (Ch-g-Aa)c, and (H) remacryl red
TGL.

Figure 4. SEM micrographs of (a) (Ch)c, (b) (Ch-g-Aam)c, and

Figure 3. X-ray diffraction patterns of the prepared materials. (c) (Ch-g-Aa)c.
Table 1. Kinetic Constants for the Sorption of Remacryl Red
TGL onto Chitosan Derivatives
pseudo-first- pseudo-second- intraparticle
order model order model diffusion model
k1 k2 kfx104 Dx1013
sorbent min-1 R2 min-1 R2 m/min m2/s R2
(Ch)c 0.029 0.924 0.048 0.986 25.94 3.25 0.979
(Ch-g-Aam)c 0.065 0.834 0.117 0.955 13.99 7.98 0.931
(Ch-g-Aa)c 0.154 0.949 0.303 0.996 5.35 19.60 0.978

geneous environment. However, both the grafted and cross-linked

copolymers exhibit one broad peak at around 2θ ) 21°, whereas
the crystalline peak of chitosan at 2θ ) 11° disappears. This can
probably be attributed to the weakening of hydrogen bonding
interactions between the hydroxyl and amino groups of the Figure 5. Sorption kinetic data for chitosan derivatives, fitted to
chitosan macromolecules as a result of stereochemical hindrance the pseudo-second-order equation. The inset presents data, as X
of the side chains or the diminished concentrations of those groups. values, fitted to the intraparticle model.
From this, it is believed that the grafting reaction of the a minimum equal to 0.3 m2/g, whereas for (Ch-g-Aam)c and
macromolecular chains onto the chitosan backbone is initiated. (Ch)c, the BET surface area values are 0.44 and 0.87 m2/g,
Scanning electron micrographs of the prepared particles are respectively. These values are typical of nonporous materials.
presented in Figure 4. It is readily observed that the drying method 3.2. Sorption Kinetics for Remacryl Red TGL. The kinetics
caused the collapse of any porous microstructure of the particles. of dye sorption by chitosan derivatives is depicted in Figure 5.
This is possibly due to hydrophilic interactions between the water The plots are characterized by a monotonous decreasing trend
molecules and the carboxylic, hydroxyl, and amino groups on with the steep descent at the beginning of sorption (within 90
the macromolecular chains of the prepared materials. This loss min) being succeeded by a more gradual decay (90-300 min),
of porous structure has a significant effect on the surface area reaching equilibrium at 300 min. Chitosan modifications had a
of the particles. For (Ch-g-Aa)c particles, the surface area reaches strong effect on both the sorption rate and loading in the order
7640 Langmuir, Vol. 23, No. 14, 2007 Lazaridis et al.

Table 2. Equilibrium Constants for the Sorption of Remacryl Red TGL onto Chitosan Derivatives
Freundlich model pH-dependent L-F
T KF Qmax KLF
sorbent K mmol1 - 1/n L1/n/g n R2 mmol/g (L/mmol)1/â â R2
(Ch)c 298 0.559 2.213 0.979 0.575 4.634 0.869 0.998
318 0.643 2.160 0.989 0.639 3.338 0.759 0.999
338 0.732 2.065 0.993 0.656 4.790 0.825 0.998
(Ch-g-Aam)c 298 0.998 1.919 0.983 0.858 5.958 0.950 0.999
318 0.996 2.120 0.986 0.902 5.679 0.798 0.999
338 0.994 2.159 0.988 0.934 5.730 0.758 0.999
(Ch-g-Aa)c 298 0.996 1.948 0.972 1.253 14.477 0.958 0.997
318 0.998 1.905 0.977 1.309 19.084 0.955 0.998
338 0.997 1.858 0.983 1.431 25.934 0.905 0.998

Table 3. Thermodynamic Constants for the Sorption and

(Ch-g-Aa)c > (Ch-g-Aam)c > (Ch)c. The residual dye content
Desorption of Remacryl Red TGL onto Chitosan Derivatives
was 0.064 mmol/L for (Ch)c but decreased to 0.037 and 0.010
mmol/L for (Ch-g-Aam)c and (Ch-g-Aa)c, respectively. sorption desorption
Table 1 presents the initial mass-transfer coefficient, the T -∆G° ∆H° ∆S° -∆G° -∆H° ∆S°
sorbent K kJ/mol kJ/mol kJ/mol K kJ/mol kJ/mol kJ/mol K
parameters for the two kinetic models, and the intraparticle
diffusion coefficients. A high correlation obtained between the (Ch)c 298 21.77 3.40
318 23.46 3.01 0.083 3.44 2.59 0.003
experimental data and the kinetic models, although more precise, 338 25.10 3.50
was obtained for the pseudo-second-order model. The foregoing (Ch-g-Aam)c 298 21.86 5.75
kinetic models were used to describe successful experimental 318 24.41 12.22 0.115 5.64 10.50 0.016
data for the sorption of dyes onto chitosan10 and activated 338 26.45 5.12
(Ch-g-Aa)c 298 23.96 6.10
carbons.22,35-37
318 26.33 16.02 0.134 5.61 14.04 0.027
However, the intraparticle model, presented as the inset in 338 29.31 5.03
Figure 5, succeeded in predicting the experimental data. The
resulting diffusion coefficients for (Ch)c, (Ch-g-Aam)c, and (Ch- edges). Figure 6b depicts the RL values at 298 K plotted versus
g-Aa)c were 3.25 × 10-13, 7.98 × 10-13, and 1.96 × 10-12 m2/s, initial dye concentration, suggesting favorable sorption (0 < RL
respectively. The calculated Biot number was significantly higher < 1).
than 100, giving evidence that sorption is mainly controlled by After the determination of the Freundlich parameters (KF, n)
intraparticle difusion.23 The rate of adsorption and the magnitude and kinetic parameters (kf, D), the design characteristics and
of D are dependent upon the nature of the adsorption process. breakthrough curve can be easily estimated. Hand, Crittenden,
For physisorption processes, the magnitude of D ranges from and Thacker18 presented constant pattern solutions, in a form
10-6 to 10-9 m2/s, and for chemisorption systems, from 10-9 to that is user-oriented, to the plug-flow homogeneous surface
10-17 m2/s. This may be explained by the stronger bonds holding diffusion model for fixed bed.
the molecules more tightly to the sorbent walls, thus lowering
Additionally, Table 3 presents thermodynamic parameters.
the rate of molecular migration.38
The negative values of ∆G° show that the sorption of dye onto
3.3. Sorption Isotherms for Remacryl Red TGL. Figure 6a chitosan derivatives was spontaneous under the experimental
presents the concentration of dye in the solid phase versus its conditions, without requiring an induction period. An increasing
concentration in the liquid phase after 24 h of mixing for chitosan trend in the ∆G° values revealed an increasing trend in the degree
derivatives at three different temperatures (298, 318, and 338 of spontaneity. The positive values of ∆H° of sorption reflect
K). The data showed an increase in the amount of dye sorbed its endothermic nature, which corresponds to the results of two
when the initial dye concentration was increased. The amount processes: (a) desorption of the water molecules previously
of dye adsorbed increased following an increase from 298 to 338 adsorbed on the dye and (b) adsorption of dye molecules on the
K, indicating the endothermic nature of the process. This was adsorbent. Each dye molecule has to displace more than one
more significant for (Ch)c and (Ch-g-Aa)c. The shapes of the molecule of solvent. The net result corresponds to the endothermic
curves clearly indicate that the isotherms for all three systems process.40 The positive values of ∆S° confirm that the adsorption
are L type according to the classification of the sorption isotherm of the dyes is a combination of the two aforementioned simple
in solution by Giles et al.39 These isotherm types, commonly processes. However, another possible interpretation of the positive
referred to as Langmuir-type isotherms, are characterized by a changes in enthalpy and entropy could be the release of numerous
high degree of adsorption at very low concentrations. The data water molecules from the adsorbent: the adsorption of hydrated
for all the of prepared materials were fitted by the three isotherm anions onto a hydrophilic polymer network inevitably disturbs
models, and the fitting results are presented in Figure 6a (Langmuir the order of water molecules in the closest environment and
model) and in Table 2. The correlation coefficient, R2, reveals releases them to the external liquid. Therefore, adsorbed molecules
that Freundlich and pH-dependent L-F isotherms provide are probably attracted because of long-distance electrostatic
adequate theoretical correlation. However, the latter should be interactions between oppositely charge groups. During the
evaluated by pH-dependent equilibrium experiments (sorption formation of ionic bonds, counterions should gain a higher degree
of freedom and increase the entropy. The sorption of the dyes
(35) Preethi, S.; Sivasamy, A.; Ramamurthi, V.; Swaminathan, G. Ind. Eng. increasing as the temperature increases may also suggest that the
Chem. Res. 2006, 45, 7627-7632.
(36) Kavitha, D.; Namasivayam, C. Bioresource Technol. 2007, 98, 14-21. number of active surface centers available for sorption increases
(37) Senthilkumaar, S.; Kalaamani, P.; Subburaam, C. V. J. Hazard. Mater. with temperature.41 Furthermore, the increase in the uptake of
2006, B136, 800-808.
(38) Walker, G. M.; Weatherley, L. R. Water Res. 1999, 33, 1895-1899.
(39) Giles, C.; MacEwan, T.; Nakhawa, S.; Smith, D. J. Chem. Soc. 1960, (40) Singh, B. K.; Rawat, N. S. J. Chem. Technol. Biotechnol. 1994, 61, 57-
3973-3993. 65.
Chitosan DeriVatiVes as Biosorbents for Basic Dyes Langmuir, Vol. 23, No. 14, 2007 7641

Figure 7. Effect of pH on the sorption and desorption of remacryl

red TGL onto chitosan derivatives.
The adsorption mechanism for each adsorbent was resolved
through the analysis of the FTIR spectra (Figure 2) of the prepared
materials loaded with remacryl red TGL after batch adsorption
in an aqueous dye solution of a 1.175 mmol/L (500 mg/L)
concentration at pH 10 for 24 h. The intensities of the respective
dye absorption bands of the adsorbed dye quantity are too low
for them to be discernible in each adsorbent’s spectra. Nonetheless,
significant shifts of the corresponding adsorbents’ peaks reveal
the interaction with the dye molecules. The adsorption of the dye
on (Ch)c incurs a clear shift of the broad hydroxyl band around
3400 cm-1 to a lower wavenumber, whereas the amide I band
is shifted from 1665 to 1659 cm-1. This is substantial evidence
Figure 6. (a) Sorption isotherms, fitted to the Langmuir model of hydrogen bond formation between the hydroxyl and amide
(R2 > 0.993), for chitosan derivatives at different temperatures. groups of chitosan with the dye molecules (Figure 8a). The
(b) Separation factor (RL) versus the initial dye concentration at respective spectra of (Ch-g-Aam)c exhibits a similar shift of the
298 K. hydroxyl band, although this is not clear enough to lead to safe
conclusions. However, a significant shift of the amide carbonyl
dyes with temperature could also be a result of the enhanced rate peak from 1679 to 1672 cm-1 confirms the interaction with the
of intraparticle diffusion of sorbate because the diffusion is an dye molecule. This shift to lower wavenumber is also probably
endothermic process.42 In general, the enthalpy change due to due to the formation of intermolecular hydrogen bonds with the
chemical adsorption (>20 kJ/mol) is considerably larger than dye molecules. The degree of protonation of the amine group
that due to physical adsorption (<20 kJ/mol). However, depends on both the degree of acetylation and the pKa of
electrostatic repulsion forces, as a result of the hydrophilic and chitosan.43,44 Considering the pKa of the chitosan’s amine group
ampholytic nature of the adsorbent, weaken the adsorption forces (ca 6.5) and the reported zero point of charge (4.24-8.6), the
and increase the degree of freedom of the molecules. Thus, protonated amino groups of chitosan in an acidic environment
although the adsorption enthalpies suggest that the adsorption exert electrostatic repulsive forces on the positively charged dye
process might be considered to be physical adsorption in nature, molecules (Figure 8b). As a result, low levels of adsorption are
this should be considered to be misleading. observed, and desorption is increased in an acidic environment.
Usually, dyes are supplied as a mixture of pure dyes with inert Increasing the pH value causes these amino groups to decrease
materials. Taking into account the actual fraction of the dye in in number, resulting in improved adsorption behavior. Adsorption
the powder (86%), the Qmax values derived from the Langmuir proceeds mainly through the formation of hydrogen bonds
model at 298 K were 0.479, 0.727, and 1.068 mmol/g (204.22, between the dye and OH- groups of the chitosan macromolecule.
309.82, and 510.74 mg/g) for (Ch)c, (Ch-g-Aam)c, and (Ch- The grafting of poly(acrylamide) causes a decrease in amino
g-Aa)c, respectively. group concentration in the prepared adsorbents. As a result, less-
3.4. Effect of pH on the Sorption/Desorption-Potential electrostatic repulsive forces are exerted as a result of protonated
Sorption Mechanism. pH is an important parameter that affects amino groups. Also, the grafting reaction inserts additional
dye sorption; it could influence the properties of the sorbent as sorption sites because it introduces groups capable of forming
well as sorbate speciation, especially in the case of metal ion hydrogen bonds. The overall result of the aforementioned
sorption. Figure 7 shows the experimental results of dye removal processes is enhanced adsorption capacity and behavior.
by chitosan derivatives under various constant pH values and
equilibrium conditions. The pH significantly affected the sorption (43) Jeon, C.; Holl, W. H. Water Res. 2003, 37, 4770-4780.
performance. In general, uptakes were significantly higher in (44) Hasan, S.; Krishnaiah, A.; Ghosh, T.; Viswanath, D.; Boddu, V.; Smith,
E. Ind. Eng. Chem. Res. 2006, 45, 5066-5077.
basic solutions compared to those in acidic solutions. Complete (45) Meshko, V.; Markovska, L.; Mincheva, M.; Rodrigues, A. E. Water Res.
dye removal was achieved for all sorbents at pH 12. However, 2001, 35, 3357-3366.
(46) Ko, D. C. K.; Tsang, D. H. K.; Porter, J. F.; McKay, G. D. Langmuir 2003,
(Ch-g-Aa)c performed sufficiently even at lower pH values. 19, 722-730.
(47) Porkodi, K.; Kumar, K. V. J. Hazard. Mater. 2006, 138, 633-635.
(41) McKay, G.; Otterburn, M. S.; Sweeney. A. G. Water Res. 1980, 14, (48) Singh, K.; Mohan, D.; Sinha, S.; Tondon, G.; Gosh, D. Ind. Eng. Chem.
21-37. Res. 2003, 42, 1965-1976.
(42) Tiwari, P. H.; Cambell, A. B.; Woon, L. Can. J. Chem. 1972, 50, 1642- (49) Choy, K. K. H.; Porter, J. F.; McKay, G. Chem. Eng. J. 2004, 103, 133-
1650. 145.
7642 Langmuir, Vol. 23, No. 14, 2007 Lazaridis et al.

Figure 9. Desorption mechanism for (Ch-g-Aa)c with R representing

the rest of the copolymer’s structure.

the desorption of dye from chitosan derivates.43 The regenerative

and reusability potential for the prepared sorbents was examined
by performing desorption experiments over the whole pH range,
the results of which are presented in Figure 7. An almost linear
curve was obtained, giving high desorption percentages in acidic
environments (pH 2, 80-92%) and low desorption percentages
in alkaline environments (pH 12, 30-61%). Thus, the exact
opposite trend of the sorption process with pH was followed, as
is usually the case, and the potential desorption mechanism is
given in Figure 9.
Additionally, thermodynamic parameters of desorption were
evaluated for chitosan derivatives at three different temperatures
(298, 318, and 338 K), and the results are presented in Table 3.
The negative values of ∆G° and ∆H° show that the desorption
of dye was spontaneous and exothermic. The positive values of
∆S° show an increased randomness at the solid/liquid interface.
The absolute values of desorption enthalpies were lower than the
respective values of sorption but followed the same trend as in
sorptionshigher absolute values for (Ch-g-Aa)c (16.02 and 14.04
kJ/mol for sorption and desorption respectively), followed by
(Ch-g-Aam)c (12.22 and 10.50 kJ/mol) and (Ch)c (3.01 and 2.59
kJ/mol). Therefore, decreasing the temperature of desorption
may cause an increase in dye desorption, making the process
fully reversible.
The molar ratio values between the dye and the binding sites
Figure 8. Main interactions between dyes and (a) (Ch)c, (b) (Ch-
of the adsorbents at the saturation point, when modeled by the
g-Aam)c, and (c) (Ch-g-Aa)c with R representing the rest of the
copolymer’s structure. Langmuir equation, were 3.2 × 10-2, 3.4 × 10-2, and 8.4 × 10-2
for (Ch)c, (Ch-g-Aam)c, and (Ch-g-Aa)c, respectively. This
The most significant spectra alterations are exhibited in (Ch- means that about 31, 29, and 12 moles of possible sorption sites
g-Aa)c. A new strong peak appears at 1580 cm-1 and is attributed are required for the sorption of 1 mole of dye by each adsorbent,
to the asymmetric deformation of the carboxylate ion. Strong respectively. This difference clearly demonstrates that not all of
ionic interactions between the negatively charged carboxylate the binding sites are actually active/available or at least accessible.
ions and the positively charged dye molecules are the principle The hydrogen bonds linking monomer units of the same chain
reason leading to the high adsorption performance of (Ch-g- (intramolecular bonds) or monomer units of different chains
Aa)c. It should also be noted that the carbonyl absorption band (intermolecular bonds) decrease their reactivity. The nonporous
at 1728 cm-1 is shifted toward a lower wavenumber and appears structure of the polymer and its residual crystallinity are also
as a shoulder at 1718 cm-1. Thus, a further hydrogen bond critical factors in the hydration and accessibility to sorption sites.
interaction is probably revealed, although it is unsafe to gather Furthermore, the sorption sites of (Ch-g-Aa)c are much more
any reliable information from the absorption band of the carboxyl active that those of other adsorbents, whereas those of (Ch-g-
group’s O-H bond around 3160 cm-1. The substantial adsorp- Aam)c are slightly stronger than the sorption sites of (Ch)c. This
tion increase of the adsorbent bearing carboxylic groups can is in agreement with the proposed sorption mechanism because
be attributed to the interaction between the protonated car- in the case of (Ch-g-Aa)c, electrostatic interactions are much
boxyl groups (R-COO-) and the cationic groups of the dye stronger than hydrogen bonds, whereas hydrogen-bond-forming
(R1R2N+R3R4) molecules. Increasing the pH of the solution causes groups are introduced into (Ch-g-Aam)c and the small concen-
a concurrent increase in the deprotonation of the adsorbent’s tration of protonated amino groups, which exert repulsive forces
carboxyl groups, leading to the formation of a greater number on the positively charged dye molecules, is further reduced.
of anionic carboxylic groups. The increase in the anionic carboxyl Chitosan is a well-known low-porosity specific surface area
groups is the cause of the significant observed improvement of material. Activated carbon, the most common sorbent material,
dye adsorption. In addition, the reaction that best represents the can have a specific surface area of 1500 m2/g (and even higher)
desorption mechanism for (Ch-g-Aa)c is presented in Figure 8c. and a porosity of 47%. It is characterized by an extensive porous
All of the previous data mean that the uptake capacity of dye structure; consequently, a greater fraction of the total exchange
molecules was affected by solution pH, which is called the pH area is accessible to the dye molecules compared to the fraction
effect. Therefore, the decrease of solution pH can be used for available to chitosan. Therefore, a difference in sorption behavior
Chitosan DeriVatiVes as Biosorbents for Basic Dyes Langmuir, Vol. 23, No. 14, 2007 7643

Table 4. Sorption Capacity and Diffusion Coefficient of Activated Carbons and Chitosan for Dyes
sorbent sorbate Qmax (mg/g) D (m2/s) ref sorbent sorbate Qmax (mg/g) D (m2/s) ref
FS400 RB 278 6.03 × 10-11 21 Pith AR-114 22 7.06 × 10 -13
46
FS400 RY 714 4.61 × 10-11 21 Pith BR-22 75.4 1.37 × 10-12 46
FS400 RR 213 12.1 × 10-11 21 Pith BB-69 160 1.07 × 10-12 46
AC AB-74 177.2 7.87 × 10-12 22 JAC eosin yellow 31.49 1.12 × 10-12 47
AC BB-1 314.5 10.60 × 10-12 22 JAC malachite green 136.58 8.46 × 10-13 47
AC methylene blue 363.9 14.40 × 10-12 22 JAC crystal violet 27.99 1.22 × 10-10 47
AC safranin 1428.57 35 AC MB 19.59 1.56 × 10-8 48
AC methylene blue 5.87 1.0 × 10-16 36 AC MO 9.43 1.54 × 10-8 48
AC1 crystal violet 60.42 1.40 × 10-10 37 AC AB-80 158.4 1.05 × 10-14 49
AC2 crystal violet 85.84 1.18 × 10-10 37 AC AR-114 103.7 1.00 × 10-14 49
FS400 AR-2B 1.5 × 10-13 38 AC AY-117 185.7 1.43 × 10-14 49
AC MG-400 159 31.6 × 10-14 45 chitosan RB-5 550 1.0 × 10-13 9
AC MS-300 309.2 8.17 × 10-14 45 chitosan RR-189 1840 10
FS400 AR-114 106.43 4.73 × 10-14 46 (Ch)c remacryl red TGL 204.22 3.25 × 10-13 this study
FS400 AB-80 128.5 4.42 × 10-14 46 (Ch-g-Aam)c remacryl red TGL 309.82 7.98 × 10-13 this study
FS400 BB-69 878.3 2.96 × 10-14 46 (Ch-g-Aa)c remacryl red TGL 510.74 1.96 × 10-12 this study

between them can be interpreted by the restricted diffusion of Aa)c exhibiting much faster rates for all dyes in comparison to
the dye molecules through the polymer network.9,22 An indicative the rest. Kinetic data followed a pseudo-second-order equation,
comparison between chitosan and activated carbon sorbents for and the rate was controlled by intraparticle diffusion. The diffusion
dye removal is presented in Table 4. As far as sorption loadings coefficients were determined to be in the range of 3.25 × 10-13
are concerned, the chitosan derivatives exhibit among the highest to 1.96 × 10-12 m2/s.
capacities, which are much higher than those of most activated The desorption of the loaded material improved as the pH
carbons, whereas as far as diffusivity is concerned, in some cases value of the desorption eluent decreased. (Ch-g-Aa)c showed a
it is higher or equal to that of activated carbon but in most cases greater degree of desorption for all pH values, followed by (Ch-
it is lower, in agreement with the aforementioned lower porosity g-Aam)c and (Ch)c. When in an acidic environment, all adsorbents
concept. The diffusivity rates could be further enhanced by desorbed over 80% of the adsorbed dye.
increasing the porosity of the material, either though freeze drying The negative ∆G° values show that the sorption of dye onto
or some other process.44 each chitosan derivative was spontaneous under the experi-
mental conditions, whereas the positive values of the enthalpy
4. Conclusions change indicate that the sorption is endothermic. Positive values
The adsorption performance of chitosan for cationic dyes was of the entropy change show the affinity of the sorbent for the
substantially increased by grafting poly(acrylamide) and poly- dyes. Increased randomness at the biomass/solution surface
(acrylic acid) through persulfate-induced free radical copolym- occurred.
erization reactions and subsequent heterogeneous covalent cross- The prepared adsorbents exhibited impressive adsorption
linking with glutaraldehyde. The adsorption of remacryl red TGL capacities, kinetics, and regenerative abilities. However, their
was evaluated, and equilibrium uptake, adsorption kinetics, and potential applicability as sorbents in large-scale implementations
desorption of the loaded dye were studied. The experimental is hindered by their substantial swelling, which ultimately limits
observations are summarized below. their practical use in commercial packed columns. Therefore,
Increasing the pH value of the dye solution led to a concurrent the next logical step would be to devise a feasible solution to
increase in the level of adsorption for all adsorbent-dye systems. overcome this drawback, which at the same time does not
(Ch-g-Aa)c exhibited the best adsorption performance by far significantly compromise the overall adsorption performance.
(100% dye removal for pH g4), followed by (Ch-g-Aa)c and
(Ch)c. Acknowledgment. The financial support received for this
The equilibrium data were analyzed reasonably well by study from the Greek Ministry of Education through the research
Langmuir, Freundlich, and pH-dependent L-F isotherm models, program Pythagoras II is gratefully acknowledged.
with the latter giving the highest correlation (R2 > 0.990). The
maximum adsorption capacities, on the basis of the Langmuir Supporting Information Available: Experimental prepara-
analysis, for remacryl red TGL were 0.479 for (Ch)c, 0.727 for tion. This material is available free of charge via the Internet at
(Ch-g-Aam)c, and 1.068 mmol/L for (Ch-g-Aa)c. [Link]
The grafting modifications improved the sorption rates of the
adsorbents, with adsorbent-bearing carboxylic groups (Ch-g- LA700423J