Journal of Colloid and Interface Science 276 (2004) 4752

www.elsevier.com/locate/jcis

Kinetic models of sorption: a theoretical analysis

Saeid Azizian
Department of Chemistry, Faculty of Science, Bu-Ali Sina University, Hamadan, Iran
Received 16 January 2004; accepted 18 March 2004
Available online 21 April 2004

Abstract
The kinetics of sorption from a solution onto an adsorbent has been explored theoretically. The general analytical solution was obtained for
two cases. It has been shown that at high initial concentration of solute (sorbate) the general equation converts to a pseudo-first-order model
and at lower initial concentration of solute it converts to a pseudo-second-order model. In other words, the sorption process obeys pseudo-
first-order kinetics at high initial concentration of solute, while it obeys pseudo-second-order kinetics model at lower initial concentration of
solute. The theoretical results (derived equations) show that the observed rate constants of pseudo-first-order and pseudo-second-order models
are combinations of adsorption and desorption rate constants and also initial concentration of solute. The obtained theoretical equations are
used to correlate experimental data for sorption kinetics of some solutes on various sorbents. The predictions of the theory are in excellent
agreement with the experimental data.
2004 Elsevier Inc. All rights reserved.

Keywords: Kinetics; Adsorption; Desorption; Sorption models

1. Introduction practical problems such as catalysis, corrosion, and removal

Adsorption and desorption reactions are known to be im-
portant in many heterogeneous systems and were investi-
gated recently in many cases [182]. Adsorption and desorp-
tion are time-dependent processes. It is necessary to know
the rate of adsorption for design and evaluation of adsorbent.
Also, the rate of desorption is important in the design and re-
generation of adsorbent. Thus, it is of interest to recognize
the adsorption and desorption kinetics and determine their
phenomenological coefficients characterizing the transport
of sorbate within sorbents [1251].
To properly understand an adsorption and desorption
process, we must understand two basic ingredients: equi-
libria and kinetics. With regards to adsorption/desorption
processes, thermodynamic data only provide information
about final state of a system, but kinetics deals with changes
in chemical properties in time and is concerned especially
with rates of changes. The growth of adsorption kinetics is
of interest for many aspects of surface chemistry, from un-
derstanding of adsorption/desorption mechanisms to more
E-mail address: sazizian@basu.ac.ir.
0021-9797/$ see front matter 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcis.2004.03.048
of pollutant components from solutions.
Various kinetics models have been suggested for adsorp-
tion, the simplest of which is Langmuirian adsorption. De-
spite of simplicity of this model, it fits surprisingly numerous
cases. Adsorption kinetics studies include theoretical studies
[111,84], adsorption of ions [1228], adsorption of organic
compounds [2857], and gas adsorptions.
The purpose of the present work is to derive pseudo-
first-order and pseudo-second-order models for sorption
processes from solutions by a general and different method
and therefore characterize the reaction conditions under
which the models must be used and then derive their in-
trinsic phenomenological rate coefficients.
2. Theory
Various models can be used to analyze the kinetics of
sorption process. Lagergren [83] suggested a rate equa-
tion for the sorption of solutes from a liquid solution. This
pseudo-first-order rate equation is
dq
= k1 (qe q). (1)
dt
48 S. Azizian / Journal of Colloid and Interface Science 276 (2004) 4752

Integrating Eq. (1) for the boundary conditions t = 0 to t = t rate equation is

and q = 0 to q = q gives d
= va vd , (9)
(qe q) dt
ln = k1t, (2) d
qe = ka C(1 ) kd . (10)
dt
where q and qe are the grams of solute sorbed per gram of
By adsorption of solute from solution onto the surface
sorbent at any time and at equilibrium, respectively, and k1
of sorbent, the concentration of solute in solution decreases;
is the rate constant of first-order sorption. The pseudo-first-
i.e.,
order equation was used extensively to describe the sorption
kinetics [1519,33,34]. C = C0 , (11)
Another model for the analysis of sorption kinetics is
where C0 is the initial molar concentration of solute, C is
pseudo-second-order. The rate law for this system is ex-
its molar concentration at any time, is surface coverage
pressed as fraction, and is
dq mc q m
= k2 (qe q)2 . (3) = , (12)
dt Mw V
Integrating Eq. (3), for the boundary conditions t = 0 to where mc is the mass (g) of sorbent, qm is the maximum ca-
t = t and q = 0 to q = q, gives pacity of sorbent, Mw is the molar weight of solute (g/mol),
and V is the volume of solution (L). One can rewrite as
1 1
= + k2 t, (4) C0 Ce
(qe q) qe = , (13)
e
where k2 is the pseudo-second-order rate constant of sorp-
where Ce is the equilibrium molar concentration of solute
tion. Equation (4), can be rearranged to obtain a linear form,
and e is the equilibrium coverage fraction. By inserting
t 1 1 Eq. (11) into Eq. (10),
= + t. (5)
q k2 qe2 qe d
= ka (C0 )(1 ) kd . (14)
The plot of t/q versus t gives a straight line with slope of dt
1/k2 qe2 and intercept of 1/qe . So the gram of solute sorbed Equation (14) is our general equation, which is used at
per gram of sorbent at equilibrium (qe ) and sorption rate different conditions for derivation of various kinetic models
constant (k2 ) could be evaluated from the slope and inter- of sorption.
cept, respectively. The pseudo-second-order model was re-
cently applied for analysis of sorption kinetics from liquid 2.1. Derivation of pseudo-first-order model
solutions by Ho et al. [1824,27,28,3942].
Two kinetic models for sorption from liquid solution, rep- If the experimental condition be such that the initial con-
resented above (first- and second-order models), were de- centration of solute is very high compared to (i.e., C0 
rived and used in the literature without any assumption or ), then one can ignore the term in Eq. (14), and there-
specialization of process conditions. In this step we want fore
to derive these kinetic models by a general and different d
= ka C0 (1 ) kd , (15)
method. The advantage of our method is that it determines dt
the conditions for using first-order or second-order models d
= ka C0 (ka C0 + kd ). (16)
and also identifies the real meaning of their observed rate dt
coefficients. By definition of
Consider the adsorption and desorption of solute A in so-
lution, f = ka C0 , (17)
k1 = ka C0 + kd , (18)
ka
A + A(a) , (6) and substitution of Eqs. (17) and (18) into (16), we obtain
kd

where ka and kd are the adsorption and desorption rate con- d

= f k1 . (19)
stants and represents the vacant site. The adsorption and dt
desorption rates are Integrating of Eq. (19) gives

va = ka C(1 ), (7)

t
d
= dt, (20)
vd = kd , (8) f k1
0 0
 
where is the coverage fraction (0

1) and C is the k1
molar concentration of solute at any time [13]. The overall ln 1 = k1 t. (21)
f
 S. Azizian / Journal of Colloid and Interface Science 276 (2004) 4752 49

From Eqs. (17) and (18) we have By integrating of Eq. (29),

k1 ka C0 + kd

t
= . (22) d
f ka C0 = dt, (33)
By defining the equilibrium constant as K = ka /kd and a 2 + b + f
0 0
rewriting Eq. (22),  
1 2a + b b2 4af
k1 KC0 + 1  ln 
= . (23) b2 4af 2a + b + b2 4af
f KC0  
We know that at equilibrium d /dt = 0, so Eq. (16) at equi- b b2 4af
ln  = t. (34)
librium converts to b + b2 4af
1 KC0 + 1
= . (24) By definition of
e KC0 
By inserting of Eqs. (23) and (24) into Eq. (21), we obtain = b2 4af , (35)
 
= b , (36)
ln 1 = k1 t. (25)
e = b + , (37)
We know that
ln = (38)
q
= , (26)
e qe and substitution of the above equations into Eq. (34) we ob-
and therefore Eq. (25) converts to tain
 
(qe q) 2a +
ln = k1 t, (27) ln = t; (39)
qe 2a +
where k1 is the observed rate constant of the pseudo-first- therefore
order model. This equation is the same as Eq. (2), which was
et +
derived by Lagergren with a different method. It is interest- = . (40)
ing to note that the observed rate constant (k1 ) in Eq. (27) or 2a(1 et + )
(2) is a combination of adsorption (ka ) and desorption (kd ) Equation (40), represents the variation of surface coverage
rate constants (according to Eq. (18)) and is not the intrin- fraction ( ) with time. For justification of this equation, we
sic adsorption rate constant which was mistakenly reported apply the boundary conditions t = 0 and t = to Eq. (40).
in the literature. Equation (18) shows that the observed rate At t = 0 Eq. (40) converts to
constant (k1 ) is a linear function of the initial concentration e
of solute and a plot of k1 vs C0 gives a straight line with 0 = , (41)
2a(1 e )
ka and kd as slope and intercept, respectively. So by evalu-
ating k1 at various initial concentrations, one could find ka , where the numerator of Eq. (41) is zero, so that 0 = 0 at
kd , and therefore the equilibrium constant K = ka /kd . The t = 0.
analysis of experimental data for this case will be done latter. At t (or equilibrium), Eq. (40) becomes

2.2. Derivation of pseudo-second-order model e = (42)
2a
or
If the initial concentration of solute was not too high for
the term in Eq. (14) to be ignored, then for derivation KCe
e = , (43)
of the rate law one must integrate Eq. (14) directly. By re- 1 + KCe
arrangement of Eq. (14) we obtain which is the Langmuir adsorption isotherm.
  Equation (40) can be rearranged to the equation
d 1
= ka 2 + C0 + ka + ka C0 (28)
dt K et e
or = (44)
2a(1 et e )
d
= a 2 + b + f, (29) and by using Eq. (38)
dt
where (et 1)
= . (45)
2a( et )
a = ka , (30)
  For small values of x, we have the following mathemati-
1
b = + C0 + ka , (31) cal approximation:
K
f = ka C0 . (32) ex 1 + x. (46)
50 S. Azizian / Journal of Colloid and Interface Science 276 (2004) 4752

So by replacement of x = t and using the above approx- Table 1

imation to Eq. (45), we obtain Effect of initial concentration (C0 ) of some solutes on the correlation coef-
ficient of pseudo-second-order (r22 ) and pseudo-first-order (r12 ) models
(1 + t 1)
(47) Solute Sorbent C0 (mg/l) r22 r12 Ref.
2a( t) BB69 Peat 50 1.00 0.661 [39]
or BB69 Peat 500 0.997 0.920 [39]
AB25 Peat 20 0.999 0.907 [39]
t
. (48) AB25 Peat 200 0.924 0.972 [39]
2a(2 t) BB69 Wood 50 1.00 0.906 [40]
BB69 Wood 500 0.990 0.992 [40]
We know that at t = 0 the initial coverage fraction is zero
AB25 Wood 20 0.996 0.978 [40]
(0 = 0), which is compatible with the above equation. By AB25 Wood 200 0.917 0.998 [40]
substitution of (42) into (48), Cu2+ Moss peat 25 1.00 [20]
Cu2+ Moss peat 200 0.996 [20]
e t
. (49) AR114 Pith 25 0.997 0.968 [42]
(2 t) AR114 Pith 150 0.985 0.993 [42]
Rearrangement of Eq. (49) yields
1 2 1 1 pseudo-second-order kinetic model (Eqs. (5) or (54)) instead
= + (50) of the pseudo-first-order model (Eq. (27)).
e t e
or It is realized from our model that when the initial con-
centration of solute is too high compared to , then the
t 2 1 sorption process obeys pseudo-first-order kinetics, and when
= + t. (51)
e e the initial concentration solute is comparable to , then the
By replacement of = q/qm and e = qe /qm , we obtain process is pseudo-second-order. In this section we want to
analyze experimental data by this criterion.
t 2 1
= + t. (52) As listed in Table 1, recent experimental data for ad-
q qe qe sorption of aqueous basic blue 69 (BB69) and acid blue 25
By definition of k2 as (AB25) onto peat show that by increasing the initial con-
centration of this dye by 10 times, the correlation coefficient
k2 = (53) (r 2 ) of fitting to the pseudo-second-order model decreases
2qe
from 1.00 to 0.997 for BB69 and from 0.999 to 0.924 for
and substitution of Eq. (53) into Eq. (52), AB25. On the other hand, the correlation coefficient of fit-
t 1 1 ting to the pseudo-first-order model increases from 0.661 to
= + t. (54) 0.920 for BB69 and from 0.907 to 0.972 for AB25.
q k2 qe2 qe
In another report of adsorption of BB69 and AB25 from
This equation (Eq. (54)) is the pseudo-second-order model aqueous solutions onto wood [40], again it was observed that
(Eq. (5)) which was previously derived by others and with by increasing the initial concentration of solutes 10 times,
a different a method. k2 is the observed rate constant of the the correlation coefficient of the pseudo-second-order model
pseudo-second-order model. decreases from 1.00 to 0.990 and from 0.996 to 0.977 for
The advantage of our derivation is that it shows, when the BB69 and AB25, respectively, while the correlation coeffi-
initial concentration of solute is not too high, that the sorp- cient for the pseudo-first-order model increases from 0.906
tion process obeys the pseudo-second-order model. Equa- to 0.992 and from 0.978 to 0.998 for BB69 and AB25, re-
tion (53) shows that the observed rate constant of the second- spectively (see Table 1).
order model (k2 ) is a complex function of the initial concen- It is obvious that with increasing initial concentration
tration of solute (C0 ). of the solute, the correlation of experimental data to the
pseudo-second-order kinetics model decreases while that to
the pseudo-first-order model increases.
3. Analysis of experimental data and discussion Recent experimental data for adsorption of Cu2+ onto
sphagnum peat moss [20] (Table 1) show that by increasing
Numerous experimental sorptions from solution have the initial concentration of Cu2+ by eight times, the correla-
been investigated during recent years. Most of these have tion coefficient of fitting to the pseudo-second-order model
been reported as pseudo-first-order and some as pseudo- decreases from 1.00 to 0.996. In another report, for adsorp-
second-order kinetic processes. Some of these reports are tion of acid red 114 (AR114) onto pith [42], it was shown
listed in Ref. [28]. Most of the researchers selected the that by increasing the initial concentration of (AR114) from
pseudo-first-order kinetic model (Eq. (27)) without any cri- 25 to 150 mg/l, the correlation coefficient of the pseudo-
teria to analyze the experimental data, but recently Ho and second-order model decreases while this coefficient for the
McKay [28] showed that some previous results obey the pseudo-first-order model increases (Table 1). These experi-
 S. Azizian / Journal of Colloid and Interface Science 276 (2004) 4752
Fig. 1. Plot of apparent rate constant (k1 ) of the first-order model for ad-
sorption of O-xylene on fly ash vs initial concentration.
mental results indicate that when the initial concentration of
solute is low, then the pseudo-second-order model is suit-
able for analysis of sorption kinetics, and when the initial
concentration is high, the pseudo-first-order model becomes
suitable. This conclusion is in complete agreement with our
theoretical approach.
Recently, adsorption kinetics of some anionic dyes onto
cross-linked chitosan beads was investigated [82]. It was re-
ported that by increasing the initial dye concentration by ca.
2.5 times the correlation coefficient of fitting to the pseudo-
first-order model increased.
As we mentioned in the theoretical section, for the sys-
tems, whose sorption kinetics obey the pseudo-first-order
model, the plot of ln(1 q/qe ) vs time gives a straight line
with a slope of k1 (Eq. (27)). In our new theoretical ap-
proach k1 is the apparent sorption rate constant (not intrinsic
adsorption rate constant) and is related to the adsorption and
desorption rate constants (ka and kd ) by Eq. (18). This equa-
tion shows that the apparent sorption rate constant (k1 ) is a
linear function of the solute initial concentration (C0 ). For
justification of this equation, we used reported experimental
data. Banerjee and co-workers reported that adsorption of
O-xylene by fly ash obeys a first-order kinetics model [33].
In their report, the initial concentrations of solute (O-xylene)
are 50, 100, 125, and 150 mg/l and the derived sorption
rate constant (k1 ) at each concentration is 0.0660, 0.0905,
0.1103, and 0.1160 min1 , respectively. They also derived
the adsorption and desorption rate constants (ka , kd ) at each
concentration. Calculation of the equilibrium constant of
this system from their adsorption and desorption rate con-
stants (K = ka /kd ) at each concentration shows that the
equilibrium constant is a function of the initial concentra-
tion of solute! This result is not true because the equilibrium
constant (K) is independent of initial concentrations. This
means that their model and their analysis of the apparent
rate constant (k1 ) is not correct. We examine their experi-
mental data with our model (Eq. (18)) and plot the diagram
of k1 vs C0 in Fig. 1. This figure shows that the plot is lin-
ear; its slope and intercept are the adsorption rate constant
(ka = 5.24 109 l/mg min) and desorption rate constant
(kd = 0.04 min1 ), respectively. So the equilibrium con-
stant (K = ka /kd ) of this process is equal to 0.013 l/mg. It
must be noted that this equilibrium constant, obtained based
51
on our model, is independent of the initial concentration of
solute.
Another interesting result of the present theoretical ap-
proach is Eq. (53). This equation shows that the observed
rate constant of the pseudo-second-order model (k2 ) is a
complex function of the initial concentration of solute. This
conclusion is in complete agreement with experimental data
[19,20,42]. Recent experimental results show that k2 is a
function of initial concentration of solute (C0 ). The fitted
equations are k2 = C0 /(Ak C0 + Bk ) for adsorption of metal
ions [20] and dye [42] with Ak and Bk as fitting constant
and k2 = 1.81 104(C0 )2.75 for adsorption of Pb(II) on
peat [19]. In the previous published papers there is no rea-
son or mathematical derived equation for the dependency of
k2 on the initial concentration of solute.
In summary, according to our new theoretical approach,
when the initial concentration of solute (C0 ) is high, its sorp-
tion kinetics fits better to the pseudo-first-order model, while
for values of C0 that are not too high, the fit of sorption
to the pseudo-second-order kinetics is better. It is also con-
cluded that for systems whose sorption kinetics obeys the
pseudo-first-order model, the observed rate constant is a lin-
ear function of the initial concentration of solute; its slope
and intercept are adsorption and desorption rate constants
respectively. For the systems which obey the pseudo-second-
order model their observed rate constant is a complex func-
tion of initial concentration of solute. We finish our discus-
sion by emphasizing the advantages of using our theoretical
approach to describe the kinetics of adsorption.
References
[1] T. Panczyk, W. Rudzinski, J. Phys. Chem. B 106 (2002) 7846.
[2] W.J. Ceyrolles, P. Viot, J. Talbot, Langmuir 18 (2002) 1112.
[3] W. Rudzinski, T. Panczyk, Langmuir 18 (2002) 439.
[4] I. Vinokurov, J. Kankare, Langmuir 18 (2002) 6789.
[5] W.H. Kuan, S.L. Lo, C.M. Chang, M.K. Wang, Chemosphere 44
(2000) 174.
[6] D.M. Ruthven, Adsorption 6 (2000) 287.
[7] C. Nguyen, D.D. Do, Langmuir 16 (2000) 1868.
[8] M.A. Brusatori, P.R. Van Tassel, J. Colloid Interface Sci. 219 (1999)
333.
[9] H.M. Selim, L. Ma, H. Zhu, Soil Sci. Soc. Am. J. 63 (1999) 768.
[10] J. Talbot, X. Jin, N.-H.L. Wang, Langmuir 10 (1994) 1663.
[11] M. Kocirik, A. Zikanova, J. Dubsky, J. Ind. Eng. Chem. Fundam. 12
(1973) 440.
[12] D.A. Dzombak, F.M.M. Morel, J. Colloid Interface Sci. 112 (1986)
588.
[13] Y. Yin, H.E. Allen, C.P. Huang, Environ. Sci. Technol. 31 (1997) 496.
[14] M. Prasad, S. Saxena, S.S. Amritphale, N. Chandra, Ind. Eng. Chem.
Res. 239 (2000) 3034.
[15] C.W. Chenug, J.F. Porter, G. McKay, Sep. Purif. Technol. 19 (2000)
55.
[16] M. Ajmal, R.A.K. Rao, R. Ahmad, J. Ahmad, J. Hazard. Mater. 79
(2000) 117.
[17] Z. Reddad, C. Gerente, Y. Andres, P.L. Cloirec, Environ. Sci. Tech-
nol. 36 (2002) 2067.
[18] Y.S. Ho, G. McKay, Trans. IchemE 76 (1998) 332.
[19] Y.S. Ho, G. McKay, Trans. IchemE 77 (1999) 165.
[20] Y.S. Ho, G. McKay, Water Res. 34 (2000) 735.
52 S. Azizian / Journal of Colloid and Interface Science 276 (2004) 4752

[21] Y.S. Ho, D.A.J. Wase, C.F. Forster, Environ. Technol. 17 (1996) 71. [55] Y.S. Ho, D.A. John Wase, C.F. Forster, Water Res. 29 (1995) 1327.
[22] Y.S. Ho, G. McKay, Ads. Sci. Technol. 16 (1998) 243. [56] A.R. Khan, T.A. Al-Bahry, A. Al-Haddad, Water Res. 31 (1997) 2102.
[23] Y.S. Ho, G. McKay, Water Res. 33 (1999) 578. [57] H.-T. Shu, D. Li, A.A. Scala, Y.H. Ma, Sep. Purif. Technol. 11 (1997)
[24] Y.S. Ho, J.C.Y. Ng, G. McKay, Sep. Sci. Technol. 36 (2001) 241. 27.
[25] M.A. Hashim, K.H. Chu, J. Ind. Eng. Chem. 36 (2001) 241. [58] A. Meghea, H.H. Rehner, I. Peleanu, R. Mihalache, J. Radioanal. Nucl.
[26] M. Ajmal, R.A.K. Rao, S. Anwar, J. Ahmad, R. Ahmad, Bioresource Chem. 229 (1998) 105.
Technol. 86 (2003) 147. [59] C. Nguyen, C.J. Sonwane, S.K. Bhatia, D.D. Do, Langmuir 14 (1998)
[27] Y.S. Ho, J.C.Y. Ng, G. McKay, Sep. Purif. Methods 29 (2000) 189. 4950.
[28] Y.S. Ho, G. McKay, Process Biochem. 34 (1999) 451. [60] E. Longhinotti, F. Pozza, L. Furlan, M.N.M. Sanchez, M. Klug,
[29] N.A. Fernandez, E. Chacin, E. Gutierrez, N. Alastre, B. Llamoza, C.F. M.C.M. Laranjeira, V.T. Farere, J. Braz. Chem. Soc. 9 (1998) 435.
Forster, Bioresource Technol. 54 (1995) 111. [61] C. Kaewprasit, E. Hequet, N. Abidi, J.P. Gourlout, J. Cotton Sci. 2
[30] A.H. Mollah, G.W. Robinson, Water Res. 30 (1996) 2907. (1998) 164.
[31] G. Cornelissen, P.C.M. VanNoort, J.R. Parsons, H.A. Govers, J. Envi- [62] W.-H. Kuan, S.-L. Lo, M.K. Wang, C.-F. Lin, Water Res. 32 (1998)
ron. Sci. Technol. 31 (1997) 454. 915.
[32] A. Karimi-Jashni, R.M. Narbaitz, Water Res. 31 (1997) 3039. [63] R. Mihalache, I. Peleanu, I. Meghea, A. Tudorache, J. Radioanal. Nucl.
[33] K. Banerjee, P.N. Cheremisinoff, S.L. Cheng, Water Res. 31 (1997) Chem. 229 (1998) 133.
249. [64] S. Duran, D. Solpan, O. Guven, Nucl. Instrum. Methods Phys. B 151
[34] S.H. Lin, Y.W. Chen, Ind. Eng. Chem. Res. 36 (1997) 4347. (1999) 196.
[35] R. Sivaraj, C. Namasivayam, K. Kadirrelu, Waste Manage. 21 (2001) [65] F.H. Frimmel, M. Assenmacher, M. Sorensen, G. Abbt-Braun, G.
105. Grabe, Chem. Eng. Process. 38 (1999) 601.
[36] C. Namasivayam, N. Muniasamy, K. Gayatri, M. Rani, K. Ran- [66] G. Sugihara, D.-S. Shigematsu, S. Nagadome, S. Lee, Y. Sasaki, H.
ganathan, Bioresource Technol. 57 (1996) 37. Igimi, Langmuir 16 (2000) 1825.
[37] C. Namasivayam, D. Prabha, M. Kumutha, Bioresource Technol. 64 [67] T.A. Albanis, D.G. Hela, T.M. Sakellarides, T.G. Danis, Global Nest
(1998) 77. Int. J. 2 (2000) 237.
[38] M. Belmouden, A. Assabbane, Y. Ait Ichou, Ann. Chim. Sci. Mater. 26 [68] M. Belmouden, A. Assabbane, Y.A. Ichou, J. Environ. Monitoring 2
(2001) 79. (2000) 257.
[39] Y.S. Ho, G. McKay, Chem. Eng. J. 70 (1998) 115. [69] K.-S. Chou, J.-C. Tsai, C.-T. Lo, Bioresource Technol. 78 (2001) 217.
[40] Y.S. Ho, G. McKay, Trans. IChemE B 76 (1998) 183. [70] H.M. Anasthas, V.G. Gaikar, React. Funct. Polym. 47 (2000) 23.
[41] Y.S. Ho, G. McKay, Can. J. Chem. Eng. 76 (1998) 822. [71] F. Stoeckli, M.V. Lopez-Ramon, C. Moreno-Castilla, Langmuir 17
[42] Y.S. Ho, G. McKay, Resource Conserv. Recycling 25 (1999) 171. (2001) 3301.
[43] A. Carton, G. Gonzalez Benito, J.A. Rey, M. dea Fuente, Bioresource [72] P.X. Wu, Z.W. Liao, H.F. Zhang, J.G. Guo, Environ. Int. 26 (2001)
Technol. 66 (1998) 75. 401.
[44] M. Carabasa, A. Ibarz, S. Garza, G.V. Barbosa-Canovas, J. Food [73] J.M. Chern, Y.W. Chien, Ind. Eng. Chem. Res. 40 (2001) 3775.
Eng. 37 (1998) 25. [74] C.-H. Wu, C.-Y. Kuo, C.-F. Lin, S.-L. Lo, Chemosphere 47 (2002) 283.
[45] D.M. Nerskaia, A. Santianes, V. Munoz, A. Guerrero-Ruiz, Carbon 37 [75] S. Nouri, J. Sci. Islamic Rep. Iran 13 (2002) 241.
(1999) 1065. [76] E. Demirbas, M. Kobya, S. Oncel, S. Sencan, Bioresource Technol. 84
[46] M.C. Annesini, F. Gironi, B. Monticelli, Water Res. 34 (2000) 2989. (2002) 291.
[47] J.-M. Chem, Y.-W. Chien, Ind. Eng. Chem. Res. 40 (2001) 3775. [77] A. Goswami, A.K. Singh, Anal. Chim. Acta 454 (2002) 229.
[48] T. Sismanoglu, S. Pura, Colloids Surf. A Physicochem. Eng. As- [78] D.N. Mbui, P.M. Shiundu, R.M. Ndonye, G.N. Kamau, J. Environ.
pects 180 (2001) 1. Monitoring 4 (2002) 978.
[49] A.G. Gayubo, A.M. Tarrio, A.T. Aguayo, M. Olazar, J. Bilbao, Ind. [79] P. Monneyron, C. Four-Brasquet, A. Sakoda, M. Suzuka, P. LeCloirec,
Eng. Chem. Res. 40 (2001) 3467. Langmuir 18 (2002) 5163.
[50] V. Gokmen, A. Serpen, J. Food Eng. 53 (2002) 221. [80] F. Haghseresht, J.J. Finnerty, S. Nouri, G.Q. Lu, Langmuir 18 (2002)
[51] M.S. Chiou, H.Y. Li, J. Hazard. Mater. B 93 (2002) 233. 6193.
[52] S.W. Wang, A.L. Hines, D.S. Farrier, J. Chem. Eng. Data 24 (1979) [81] M.A. Zolfigol, T. Madrakian, E. Ghaemi, A. Afkhami, S. Azizian, S.
345. Afshar, Green Chem. 4 (2002) 611.
[53] M. Hodi, K. Polyak, J. Hlaray, Environ. Int. 21 (1995) 325. [82] M.S. Chiou, P.Y. Ho, H.Y. Li, Dyes Pigments 60 (2004) 69.
[54] N.I. Ivanova, I.L. Volchkova, E.D. Shchukin, Colloids Surf. A Physic- [83] S. Lagergren, Handlingar 24 (1898) 1.
ochem. Eng. Aspects 101 (1995) 239. [84] H.C. Thomas, J. Am. Chem. Soc. 66 (1944) 1664.