JOURNAL OF COLLOID AND INTERFACE SCIENCE 194, 154–165 (1997)
ARTICLE NO. CS975041
Equilibrium Adsorption Studies of Some Aromatic Pollutants
from Dilute Aqueous Solutions on Activated Carbon
at Different Temperatures
A. R. Khan, 1 R. Ataullah, and A. Al-Haddad
Department of Chemical Engineering, Faculty of Engineering and Petroleum, Kuwait University, P.O. Box 5969, Safat 13060, Kuwait
Received April 1, 1997; accepted June 16, 1997
Aqueous solutions of phenol, p-chlorophenol, and p-nitrophenol
have been used to determine the adsorption isotherm for single
solute systems on activated carbon at different temperatures. The
experimental program has been conducted to investigate the in-
fluence of concentration and temperature. All the reported equilib-
rium isotherm equations have been tried on present and published
experimental data. A generalized isotherm equation which was
proposed by Khan et al. (6, 10) and tested for bi-solute adsorption
data has been modified for single-solute system. The temperature
dependency has also been incorporated into a generalized equation.
It has been noticed that the Radke and Prausnitz (7) and general-
ized isotherm equations could represent the entire data with a mini-
mum average percentage error. The influence of different adsor-
bents, sorbate concentrations, and pH of aqueous solutions has also
been discussed in detail. The temperature dependency has been
investigated using both the Dubinin–Astakov (13) and the modified
generalized equation. The generalized equation describes the experi-
mental and published data adequately and provides a single value
of differential molar heat of adsorption, DHads , for a single solute
adsorption system. The Dubinin–Astakov isotherm equation has
shown an increasing trend of DHads as the loading of adsorbent has
increased. q 1997 Academic Press
Key Words: activated carbon; liquid phase adsorption; adsorbent
and sorbate properties.
INTRODUCTION
In industrial waste-water purification processes, adsorp-
tion–desorption techniques are commonly implemented. In-
creasing environmental awareness and concern are bound to
expand the role of activated carbon as a highly affective
adsorbent for the removal of organic pollutants from dilute
aqueous solutions. The pollutant content, pH, and tempera-
ture of waste-water are likely to vary with time, so design
of suitable adsorption systems is not that simple; this diffi-
culty is compounded by the fact that the influence of these
1
To whom correspondence should be addressed.
0021-9797/97 $25.00
Copyright q 1997 by Academic Press
All rights of reproduction in any form reserved.
AID JCIS 5041 / 6g30$$$641 10-16-97 13:03:11
variables on the equilibrium amount of adsorption is not
extensively studied. In order to effectively design activated
carbon adsorption units and to develop mathematical models
that can accurately describe their operation, information on
both the adsorption and the desorption of individual pollut-
ants under different operating conditions is required.
Extensive experimental and modeling studies have been
reported to date on the activated carbon adsorption of a broad
spectrum of hazardous compounds from aqueous solution (1,
2). Yenkie and Natarajan (3) have reported the adsorption of
some organic compounds (phenol, aniline, o-cresol, benzoic
acid, and p-methoxyphenol) at 357C at six different adsor-
bents. They could not represent the entire data by any of
the well-known isotherms. They have tabulated sections of
concentration range for which the Freundlich isotherm and
the Langmuir isotherm could represent experimental data
satisfactorily. Chatzopoulos et al. (4) have discussed the
mechanism of adsorption and desorption on activated carbon
from aqueous medium. They have tried various types of
empirical reported isotherms on their experimental data and
found that the Fritz and Schlunder isotherm (5) could fit
their data at high concentrations and was, therefore, selected.
Khan et al. (6) have proposed a generalized model for multi-
solute adsorption and tested it for published experimental
data (7–9). In their latest paper Khan et al. (10) have
analyzed their experimental adsorption data for phenol and
phenol derivatives in aqueous medium for single and bisolute
systems. They concluded that most of the single solute data
can be adequately represented by Langmuir or BET (Bru-
nauer, Emmett, and Teller) isotherm, with the exception of
phenol data, which can be represented best as a Freundlich
isotherm. The generalized equation has the minimum devia-
tion for all bisolute data as compared to other reported ex-
pressions.
In the present work, the influence of the adsorbents, con-
centration of sorbate, pH, and temperature of solution, on
the adsorption isotherm has been investigated fully for sin-
gle-solute systems. All reported isotherm equations have
been tested for current and published data to provide the
best expression single-solute adsorption mechanisms.
154
coida
 ADSORPTION STUDIES OF AROMATIC POLLUTANTS 155
THEORY isotherm Holl–Kirch, and Marczewski and Jaroniec type
equations which are given as
Adsorption mechanisms are so complicated that no simple
theory can adequately represent all experimental data. Many q0aC be
expressions have been published to describe the equilibrium Holl–Kirch type qe Å , [5]
1 / aC be

S D
relationship between the sorbate and the adsorbent (activated
n/b
carbon) most commonly used. Langmuir and Freundlich iso- q0aC be
Marczewski and Jaroniec type qe Å . [6]
therms are the most common, and sometimes piecewise 1 / aC be
equations are used for a selected range of concentration
where implementation of only one expression fails badly. He has also used a second-order polynomial function for
the Freundlich isotherm equation to compensate for the cur-
Adsorption Isotherms vature on the logarithmic plot. The coefficient n can be
BET is a special form of Langmuir isotherm which can written as n1 / ln(C ne 2 ) and the equilibrium isotherm is
account for multilayer adsorption. All the assumptions relat-
ing to surface uniformity, distribution of sites, and negligible qe Å kC ne Å kC ne 1/ n 2 ln C e . [7]
influence of the occupied site on neighboring sites are the
same as for Langmuir monolayer adsorption. The energy for Jossens et al. (14) have correlated equilibrium concentra-
a monolayer is different from that for the second or other tion of sorbate to the adsorbed amount on the adsorbent
layers, which is equivalent to energy of liquification. The (capacity). This equation can be reduced to Henry’s law at
BET equation for physical adsorption is given as low capacities:

q0aCe q0
qe Å . [1] Ce Å exp(a2q be ). [8]
(CS 0 Ce )(CS 0 Ce (1 0 a)) a1

Radke and Prausnitz (7) have proposed a slight modifica-
tion to the Langmuir equation, introducing another coeffi-
cient which has improved the fit to their experimental data:
q0aCe
qe Å . [2]
1 / aC be
Toth (11) has modified the Langmuir equation to reduce
the error between experimental data and predicted values of
Temperature Effects
In some of the proposed equilibrium equations, a tempera-
ture term has been introduced. These equations are most
often used to evaluate adsorption behavior at constant tem-
perature. Parker (13) has reported an nth order exponential
isotherm function for multilayer adsorption, which has n /1
coefficients. As the number of coefficients increases, the
possible degree of fitness improves:

F H S DJ G
N i
equilibrium adsorption data. The application of his equation CS
qe Å q0 exp 0 ∑ ai RT ln . [9]
is best suited to multilayer adsorption similar to BET iso- Ce
iÅ1
therms which is a special type of Langmuir isotherm and
has very restrictive validity (12). The Toth correlation is If this equation is limited to a single term with an arbitrary
given as exponent, this equation reduces to the Dubinin–Astakov
equation (15) which has been successfully used by Parker
q0aCe (13) for his experimental adsorption data of 1,1,2-trichloroe-
qe Å . [3]
(1 / aC be ) 1 / b thene on activated carbon at 257C:

Fritz and Schlunder (5) have proposed an empirical ex-

pression which can fit a wide range of experimental data
due to large number of coefficients in their equation:
F H S DJ G
qe Å q0 exp 0a RT ln
CS
Ce
b
[10]

These above-mentioned equations are all empirical, but

q0a1C be 1
qe Å . [4]
1 / a2C be 2
Parker (13) has compared all different reported isotherms
for the adsorption of 1,1,2-trichloroethene from aqueous so-
lution. He has reported a few modified forms of Langmuir
there are two thermodynamically based theories available in
literature which are often cited. The vacancy solution theory
(VST) suggested by Suwanayuen and Danner (16) is found
to be very useful for correlating single gas adsorption iso-
therms and gas mixture adsorption equilibria. By substituting
water molecule in the vacuum space, Fukuchi et al. (17)

AID JCIS 5041 / 6g30$$$642 10-16-97 13:03:11 coida

156 KHAN, ATAULLAH, AND AL-HADDAD
have derived an adsorption isotherm for dilute aqueous solu-
tions. The Gibbs adsorption equation for dilute aqueous solu-
tions, RTd ln C Å (A/q)d p, has been introduced, and the
Wilson equation has also been adopted to describe the non-
ideality in the adsorbed phase. The adsorption isotherm ob-
tained for a dilute aqueous solution is similar to Eq. [8],
the equilibrium concentration as function of capacity, and
for a single solute the isotherm equation is given as
Ce Å F q0 uL13
H 10u S 1 0 (1 0 L31 ) u
L13 / (1 0 L13 ) u DG
F
1 exp 0
L31 (1 0 L31 ) u
0
(1 0 L13 ) u
1 0 (1 0 L31 ) u L13 / (1 0 L13 ) u
, G
where the Henry’s law constant H is defined by H Å
lim (qe /Ce ) and u Å qe /q0 .
Cer0
The second theory is the Ideal Adsorbed Solution theory
for gaseous adsorption, which has been further developed by
O’Brien and Myers (18) for calculation of multicomponent
equilibrium from pure component data. They have proposed
an expression for a single-component isotherm data by a
Langmuir equation with a first-order correction:
qe Å q0 S aCe
1 / aCe
/
s 2aCe (1 0 aCe )
2(1 / aCe ) 3 D .
It seems that adsorption data from dilute solutions and
also for gaseous adsorption can best be represented by either
Langmuir or modified Langmuir isotherms. Equations [1–
6] are all different modified forms of the Langmuir equation.
Khan et al. (6) have put forward a generalized equation
for multicomponent adsorption and have tested it on pub-
lished data for bisolute solutions. In their later work Khan
et al. (10) have compared different types of multicomponent
isotherms on their experimental data and concluded that the
generalized equation could represent the experimental data
most accurately with minimum average percentage error.
The generalized model for a single solute could be formu-
lated in the form of the simple expression
q0aCe
qÅ . [13]
(1 / aCe ) b
To incorporate the influence of temperature on adsorption,
Kapoor et al. (19) have presented a review of existing mod-
els of the concentration dependence of the surface diffusion
coefficient, DS , in gas–solid systems. The model proposed
by Gilliland et al. (20) has also been widely used to explain
the dynamics of adsorption for dilute aqueous solutions
(21, 22),
AID JCIS 5041 / 6g30$$$642 10-16-97 13:03:11
DS Å DSO exp S 0 jDHad
RT D , [14]
where j is an experimentally determined constant which is
dependent on types of bonds between the surface and the
sorbate. Al-Duri and McKay (23) have reported that the
value of surface diffusion coefficient, DS , increases with
initial sorbate concentration in the solution:
DS Å D0 exp S D
k qe
q0
. [15]
[11] By comparing Eqs. [14] and [15], qe can be related to
temperature. Furuya et al. (24) have explained intraparticle
mass transport by two parallel mechanisms, pore and surface
diffusions, which are similar direct linear functions of tem-
perature. The diffusion coefficients are related to each other
by the following equation:
DP
De Å DS / . [16]
(dqe /dCe ) rP
Integration of Temperature Term in Correlation
Equation [13] has been further developed to take into
account of the effect of the variation of temperature on equi-
[12]
librium adsorption expressions. Considering the condensa-
tion analogy, the equilibrium concentration of sorbate is re-
lated to temperature. For condensation phenomenon, the
Gibbs free energies for both phases, liquid and vapor, are
equal at equilibrium and the vapor pressure is expressed as
a function of temperature as d ln P/dT Å g /T 2 . The equilib-
rium concentration of sorbate is assumed to be represented
by a similar function of temperature, which results in the
equation
q0aCe ( 0 g /T / d )
qe Å , [13a]
[1 / aCe ( 0 g /T / d )] b
where g and d are arbitrary constants and g can be used for
the evaluation of heat of adsorption. In the present work
experimental data have been obtained by conducting experi-
ments at different temperatures for single solute adsorption
on activated carbon from aqueous solutions. Various types
of reported equilibrium isotherm equations with this general-
ized form have been tested for the present data and also for
the published data to determine the best possible equation
for the entire single adsorption data. The influence of various
variables, the adsorbent, the concentration of sorbate, the
solution pH, and the temperature, has been investigated.
EXPERIMENTAL
Hydro-Anthrasit H (activated carbon) of bulk density 540
kg m03 , particle porosity of 0.485, and surface area of 3000
coida
 ADSORPTION STUDIES OF AROMATIC POLLUTANTS 157
TABLE 1
Physical Properties of Different Adsorbents Used During the Experimental Program of Yenkie and Natarajan (3)

Surface area Particle density Pore volume Void

Type of adsorbent Origin (m2 g01) (kg m03) (m3 kg01 1 1000) fraction

Adsorbent 1, RRL Coconut shell 900 749

Adsorbent 2, CAL Coal 1090 820 0.75 0.62
Adsorbent 3, KURARE 1020 722 0.926 0.67
Adsorbent 4, F200 Bituminous coal 825 858 0.724 0.53
Adsorbent 5, F400 Bituminous coal 998 795 0.825 0.65
Adsorbent 6, LCK Coconut shell 1025 753

m2 g 01 was washed with hot deionized water and was dried
at 1107C to remove excessive moisture. Double-distilled and
deionized water was used for the preparation of solutions of
different solutes, phenol, p-chlorophenol, and p-nitrophenol
of analytical reagent grade supplied by Aldrich Chemicals.
The initial concentration of these compounds for the prepara-
tion of stock solution for adsorption isotherm determination
was 1000 ppm (g m03 ) which was double-checked by volu-
metric titration (25). A specified quantity, 0.2 dm3 , of differ-
ent solutions with known initial concentration containing a
measured amount of activated carbon was taken in a 0.25
dm3 Erlenmeyer flask and sealed with a screw cover. The
flasks, including a blank one, were kept in a shaker–incuba-
tor at a controlled temperature for a minimum period of 24
h for high temperatures and a maximum period of 72 h
for low temperatures. The samples were filtered through a
Whatman filter paper number 42 next day and equilibrium
concentrations were determined by UV absorption using
8700 series Unicom UV/V spectrophotometer. No addi-
tional buffers were added; the pH of solutions varied from
4 to 6 for the selected range of concentration during the
course of the present research for phenol and phenol deriva-
tives.
RESULTS AND DISCUSSION
Yenkie and Natarajan (3) have experimentally investi-
gated the influence of different adsorbents on the adsorption
of aromatic compounds from dilute aqueous solutions. The
properties of the different adsorbents in experimental use
are listed in Table 1. It is noted from the experimental data
that phenol, benzoic acid, o-cresol, and p-methoxyphenol
have shown high selectivity for adsorbent 1 due to the sur-
face charge, which facilitated the adsorption of solutions

FIG. 1. (a) Comparison of different adsorbents for the adsorption of benzoic acid from dilute aqueous solution (3). (b) Comparison of the adsorption
of different aromatic sorbates from dilute aqueous solutions on adsorbent 4 (F200) (3).

AID JCIS 5041 / 6g30$$$643 10-16-97 13:03:11 coida

 158

AID
JCIS 5041
/
TABLE 2
Comparison of Different Isotherms for Single Organic Solutes in Aqueous Medium for Different Adsorbents (Yenkie and Natarajan, 1991)

Langmuir Radke & Prausnitz Present model

qoaCe Freundlich qoaCe VST model (1985) IAS model (1983) qoaCe
qe Å qe Å qe Å
1 / aCe qe Å kC ne 1 / aCbe Eq. [11] Eq. [12] (1 / aCe)b

6g30$$5041
Solute qo a % err k n % err qo a b % err qo H L13 L31 % err qo a s % err qo a b % err

1
Aniline 1.64 3.60 9.39 1.54 0.65 18.6 1.38 3.83 1.47 6.13 1.36 2.81 0.12 2.02 3.60 1.79 0.62 2.89 6.85 8.14 0.68 3.12 6.8
Phenol1 1.85 4.19 11.0 2.02 0.67 26.0 1.43 4.79 1.45 7.10 1.47 4.32 0.16 1.92 5.00 2.11 0.73 2.71 8.20 9.64 0.75 3.26 8.0
Benzoic acid1 2.25 4.16 2.52 2.06 0.48 8.79 2.19 3.72 1.18 1.83 2.1 4.42 0.17 1.98 2.87 3.5 0.98 1.71 1.3 4.02 2.14 1.41 1.3
p-Methoxyphenol1 1.56 38.3 6.29 1.98 0.28 8.29 1.79 51.2 0.86 1.15 1.91 10.1 0.001 6.07 10.7 1.89 19.5 1.34 1.83 0.902 89.2 0.85 1.1
o-Cresol1 1.83 17.7 4.87 2.36 0.41 15.5 1.94 18.6 0.93 4.25 2.04 13.1 0.04 3.51 11.4 2.58 5.93 1.48 4.55 1.42 24.6 0.91 4.1
Aniline2 1.57 2.75 6.16 1.32 0.66 16.4 1.54 2.67 1.14 5.6 1.74 1.61 0.04 3.36 5.60 1.57 2.75 0.001 6.20 2.47 1.71 1.33 5.7
Phenol2 1.57 2.75 6.16 1.32 0.66 16.4 1.54 2.67 1.14 5.6 1.74 1.61 0.04 3.36 5.60 1.57 2.75 0.001 6.20 2.47 1.71 1.33 5.7
Benzoic acid2 1.86 6.35 2.17 1.92 0.47 9.81 1.86 6.35 1.01 2.2 1.96 4.84 0.074 2.84 5.30 2.79 1.99 1.43 1.68 1.82 6.54 0.988 2.1
p-Methoxyphenol2 1.69 13.5 3.40 1.82 0.33 9.15 1.76 16 0.90 1.68 1.89 1.81 0.006 3.75 8.10 2.39 4.52 1.51 1.90 1.21 21.8 0.885 1.65
o-Cresol2 1.72 12.3 10.7 1.94 0.36 3.93 1.93 98.5 0.669 3.8 2.93 5.75 0.001 7.16 7.90 1.92 9.16 1.14 9.46 0.289 273 0.668 3.9

10-16-97 13:03:11
Aniline3 1.92 3.51 2.1 1.74 0.52 6.32 1.81 3.2 1.18 1.92 1.35 6.62 1.57 0.01 1.20 2.81 0.75 1.93 1.86 4.29 1.43 1.68 1.70
Phenol3 1.73 5.4 3.2 1.66 0.48 12.7 1.68 4.97 1.13 2.40 1.65 5.71 0.22 1.99 5.12 2.72 1.33 1.64 2.33 2.61 3.34 1.24 2.50
Benzoic acid3 3.11 2.72 10 2.52 0.57 15.4 2.67 2.51 1.77 6.8 2.35 0.088 0.005 1.52 2.85 1.02 0.38 6.07 7.52 29.3 0.26 5.84 7.40
p-Methoxyphenol3 1.65 29.2 12.1 1.93 0.28 5.21 1.87 71.4 0.80 3.6 2.10 14.6 0.001 7.05 8.88 1.98 16.5 1.50 4.10 0.56 208.0 0.776 3.40
o-Cresol3 2.29 8.16 2.1 1.57 0.42 8.18 2.45 8.64 0.94 1.85 2.61 5.2.0 0.039 3.12 4.20 3.38 3.05 1.33 1.60 1.93 10.7 0.91 1.70

coida
Aniline4 1.76 2.57 2.76 1.58 0.71 10.3 1.77 2.58 0.98 2.85 1.27 4.67 1.67 0.034 3.64 2.47 1.12 1.12 2.85 1.74 2.61 0.991 2.76
Phenol4 1.67 3.19 9.88 1.6 .679 20 1.4 3.43 1.41 7.75 1.46 3.1 0.12 2.37 9.37 1.67 3.18 1 1 1004 10.0 6.15 0.825 2.54 8.70
Benzoic acid4 2.07 4.81 0.75 2.05 0.52 8.64 2.08 4.84 0.989 0.73 2.13 9.39 0.33 2.08 1.46 3.03 1.73 1.31 1.00 1.98 5.08 0.975 0.73
p-Methoxyphenol4 1.31 26.4 8.71 1.45 0.24 3.24 1.44 80.7 0.812 2.40 1.55 17.82 0.001 7.56 8.82 1.57 15.2 1.47 1.45 1.21 21.8 0.885 1.65
o-Cresol4 1.786 9.3 4.38 1.925 0.39 6.1 1.88 12.23 0.84 2.35 2.07 6.0 0.03 3.75 5.67 2.1 5.88 1.01 2.35 0.989 20.85 0.80 2.41
KHAN, ATAULLAH, AND AL-HADDAD

Aniline5 1.74 3.31 6.36 1.585 0.61 15.0 1.53 3.32 1.403 4.63 1.474 2.49 0.132 2.00 4.26 2.09 .607 2.512 4.75 7.43 0.71 2.82 4.81
Phenol5 1.46 5.93 7.54 1.468 0.547 22.79 1.26 3.82 1.62 5.54 1.465 3.05 0.118 2.39 9.40 2.305 1.202 1.915 4.42 3.09 2.57 1.52 4.42
Benzoic acid5 2.81 3.115 5.79 2.78 0.679 13.39 2.57 3.11 1.21 5.38 2.43 7.88 0.515 1.379 7.20 2.803 3.115 1 1 1004 5.79 6.524 1.274 1.787 5.3
p-Methoxyphenol5 1.678 25.4 6.37 1.92 0.296 7.67 1.825 34.97 0.862 3.07 1.98 9.83 0.003 5.52 12.41 2.076 11.71 1.402 3.32 0.945 59.81 0.842 3.04
o-Cresol5 1.98 16.76 9.72 2.41 0.357 6.17 2.34 49.97 0.0718 5.164 1.868 38.83 2.38 1 1 1004 17.445 2.263 11.34 1.12 7.08 1.01 44.14 0.795 5.12
Aniline6 1.448 7.251 3.24 1.59 0.478 9.54 1.52 8.31 0.874 1.98 1.701 1.255 0.012 3.62 4.32 1.73 4.57 .952 2.10 0.895 13.36 0.815 2.03
Phenol6 1.687 4.345 20.4 1.72 0.653 34.0 1.22 5.26 1.66 13.61 1.388 1.468 0.035 2.354 15.0 0.533 0.555 6.63 14.42 12.69 0.545 4.31 13.9
Benzoic acid6 3.10 3.94 12.27 3.29 0.666 25.0 2.097 4.991 1.684 7.53 2.276 7.34 0.331 1.171 3.33 0.881 0.544 6.71 9.10 21.20 0.537 4.227 8.70
p-Methoxyphenol6 1.98 17.1 3.54 2.21 0.316 9.13 2.05 19.1 0.938 2.695 2.204 14.76 0.026 4.021 12.62 2.93 4.77 1.62 1.606 1.548 24.35 0.914 2.48
o-Cresol6 1.934 36.0 13.34 2.56 0.31 6.36 2.44 93.67 0.759 4.43 1.98 80.1 2.59 0.022 26.31 2.33 19.67 1.43 6.92 98.79 0.983 0.814 4.77

Note. Superscripts 1–6 refer to adsorbent. See Table 1 for descriptions and origins.
 ADSORPTION STUDIES OF AROMATIC POLLUTANTS 159

FIG. 2. (a) Comparison of different isotherm equations for the adsorption of p-chlorophenol from aqueous solution (7). (b) Comparison of different
isotherm equations for the adsorption of toluene from dilute aqueous solution (4).

having pH values lower than 7. Aniline solution has pH
value higher than 7 and has shown minimum adsorbence for
the above reported adsorbent 1, while on adsorbent 3, it was
highly adsorbed due to its surface structure.
The surface heterogeneity and chemical nature of the ad-
sorbent surface are important factors for the determination
of the influence of pH on the adsorption of single solute
systems. It has been discussed (26) that high pH depresses
the equilibrium amount adsorbed but neutral salt addition
increases adsorption. The buffers have unnoticeable influ-
ence or adsorption. Ionization of the sorbates at high pH
values and changes in the nature of the solid surface with
pH are responsible for the above reported results. The change
in selectivity with pH value indicates that electrostatic inter-
action between surface and sorbate ions are very important
at high values of pH but specific interaction becomes more
significant with decreasing pH.
In Fig. 1a, the adsorption data for benzoic acid on six
different adsorbents are shown for comparison and in Fig.
1b the selectivity of various sorbates for a specific granular
activated carbon (adsorbent 4) is given. Phenol and aniline
sorbates have less adsorption than benzoic acid, o-cresol,
and p-methoxyphenol due to polarity, molecule structure,
mass, and size. Yenkie and Natarajan (3) have presented
their data using Langmuir and Freundlich isotherms for se-
lected concentration range. A FORTRAN program has been
used for the estimation of coefficients for the reported corre-
lations based on nonlinear optimization technique (27). All
the isotherm equations with the generalized expression (Eq.
[13]) have been tried and the results are given in Table 2.
The value of the mean absolute percentage error É[qe (experi-
mental) 0 qe (predicted)]/qe (experimental) 1 100É has
been selected as a test criterion for the fit of the correlation.
Equations [2] and [13] which are not very different from
each other, represent the data for the whole reported experi-
mental concentration range with minimum average percent-
age error. The physical significance for each constant in
these equations has been explained by Martinez and Basmad-
jian (28). Equation [13] is a simplification of a generalized
equation put forward by Khan et al. (6) and successively
used for bisolute systems due to its unique characteristic that
it can cover both extreme cases, Langmuir and Freundlich
type isotherms. The amount adsorbed is a binomial function
of equilibrium concentration and can accommodate all dif-
ferent types of variations where other correlations fail.
The comparison of different isotherms for adsorption of
p-chlorophenol (7) and toluene (4) from aqueous solutions
on activated carbon over a wide concentration range is
shown in Figs. 2a and 2b and the coefficients for all reported
isotherms are tabulated in Table 3. Figure 2a has an enlarged

AID JCIS 5041 / 6g30$$$643 10-16-97 13:03:11 coida

160 KHAN, ATAULLAH, AND AL-HADDAD

insert to show the low concentration. Over a wide concentra-

% err

2.6
3.7
1.5
1.6
7.0
13
3.2
8.23
tion range there is a large average percentage error for almost

(1 / aCe)b
all isotherms with the exception of Eqs. [2] and [13], where

Present model

0.8
0.85
0.55
0.84
0.86
1.71

0.88
1.01
qoaCe

b
the average percentage error is within an acceptable limit.
For almost all systems the numerically evaluated values

155.0
258.0

482.0
99.0
3.53

3.96
40.7
1427.0
qe Å

a
of the saturated amount adsorbed q0 , using Eqs. [1–6] and

Note. Sorbates and references are as follows: toluene (4); acetone, p-cresol, and p-chlorophenol (7); benzene, n-pentanol, 4n-propylphenol, and 4-methylphenol (33).
[13], have the same order of magnitude. The values of the

0.9
1.65
0.11
0.82

5.93
0.28
0.94
0.89
Langmuir coefficient a also have behavior similar to that

qo
observed for q0 with very few exceptions. In Table 5, the
% err

2.3

2.5
7.3
17.0
6.43
6.58
6.87
13.1
coefficients for Dubinin–Astakov equation (Eq. [10a]) and
the generalized equilibrium expression (Eq. [13a]) for dif-
IAS model (1983)

1.25
1.16
1.73
1.81
1.83
0.84
1.59
1.87
ferent sorbates at specified experimental temperature range
s
Eq. [12]

are tabulated. The Dubinin–Astakov equation (Eq. [10a])

293.0
108.0
26.7
0.16
12.2
15.5
2.28
14.1
does not represent experimental data adequately with an av-
Comparison of Different Isotherms for Single Organic Solutes in Aqueous Solution

erage percentage error more than 10% (Figs. 3, 4, and 5a)

while the generalized equation (Eq. [13a]) has minimum
1.1
3.24

2.33
2.64
3.84
0.55
1.82
1.32
qo

average percentage error between the predicted and experi-

mental values of the adsorbed amount as functions of equi-
% err

5.0
11

45.0
16.0
7.3
24.3
75.0
25.7

librium concentrations and temperatures. Therefore, there is

a noticeable difference between the numerically computed
5.1

8.1

9.0
0.9
0.03

0.19

2.57

3.69
L31
VST model (1985)

value of saturated amount adsorbed q0 , using Eqs. [10a] and

[13a]. The Dubinin-Astakov constant, the affinity coefficient
Eq. [11]

1.95
0.19
2.19

0.02
6135.0

78.4
0.01
0.01
L13

a, is an inverse function of adsorption energy and can be

compared with the constant in Eq. [13a], a( 0 g /T / d ).
135.0
0.39
49.7

2.98
2055.0

17.3
86.0
35.1

The discrepancy is due to the poor fit of Eq. [10a] to the

H

experimental data for different temperatures.

TABLE 3

The influence of temperature has been investigated for

2.1
3.27
2.25

2.39
5.16

14.2
39.0
0.98
qo

gaseous adsorption data (29, 30). All different types of iso-

therms have been attempted and results are given in Table
% err

7.0

3.6
8.5
2.58
1.94
1.35
1.42

12.5

4. Figures 3a and 3b are plots of equilibrium equations for

Radke & Prausnitz

ammonia and isobutane gas adsorption on activated carbon

1 / aCbe

0.6

0.9
0.86

0.85

1.27
0.85

0.67

1.01
qoaCe

as a function of temperature. Equations [13] and [13a],

which have been suggested for aqueous solutions, have
148.0

670.0
503.0
60.4
1.95
83.6

12.2
63.8
qe Å

shown minimal average percentage error for the case of

a

gaseous equilibrium adsorption phenomenon (partial pres-

3.32
0.26
1.87
2.02
1.83
0.68
2.11
0.84

sure of gas representing equilibrium concentration). The en-

qo

thalpy of adsorption has also been calculated based on exper-

% err

5.4

5.4
13.0
6.56

16.5
12.3
15.3
32.1

imental data and is reported in Table 6.

The present experimental data and the reported data have
Freundlich
qe Å kC ne

0.2

0.6
0.27
0.58

0.17

0.38
0.32
0.35

been collected for adsorption from dilute aqueous solutions

n

at different temperatures. The coefficients of different equi-

3.76
0.14
1.78
1.94
4.33
0.71
4.63
2.96

librium equations are listed in Table 5. The enthalpy values

k

of adsorption for dilute aqueous solutions are tabulated in

% err

Table 6.
7.3

8.3
5.81
20.9
13.1
17.5

13.3
6.36
1 / aCe

The equilibrium equations for adsorption of cobalt phtha-

qoaCe
Langmuir

locyanine (dye) (12) and for propanol (7) from aqueous

745.0
487.0
41.4
0.34
31.8
39.5
9.63
18.1
a

solutions for different temperatures are plotted in Figs. 4a

qe Å

and 4b, respectively. The present experimental data have

2.0
2.79
0.83
0.18

2.58
0.49
1.47
0.88
qo

also been analyzed in the similar manner for different tem-

peratures. In Fig. 5a, the adsorption data for phenol has been
4n-Propylphenol
p-Methylphenol
p-Chlorophenol

plotted with generalized and Dubinin–Astakov equations.

Sorbate

n-Pentanol

The enthalpy of adsorption for different organic compounds

p-Cresol

Benzene
Acetone
Toluene

from aqueous solutions have been worked out. These values

are comparable to the reported enthalpy of adsorption (31).

AID JCIS 5041 / 6g30$$$643 10-16-97 13:03:11 coida

 ADSORPTION STUDIES OF AROMATIC POLLUTANTS 161
TABLE 4
Comparison of Different Isotherms for Gaseous Adsorption

Langmuir Radke & Prausnitz Present model

qoaPe Freundlich qoaPe qoaPe
qe Å qe Å qe Å
1 / aPe qe Å kP en 1 / aPbe (1 / aPe)b
T
Sorbate (K) qo a % err k n % err a qo b % err qo a b % err

Ammoniaa 249 0.224 6.998 5.554 1.267 0.404 18.26 0.397 4.65 0.9 3.78 0.388 4.538 0.87 3.8
Ammoniaa 273 0.068 6.878 2.278 .881 0.441 9.26 0.111 4.72 0.916 1.888 0.098 5.136 0.884 1.9
Ammoniaa 303 0.024 6.045 6.525 .23 0.659 5.68 0.271 0.844 0.585 2.69 0.21 0.903 0.459 2.9
Ammoniaa 353 0.020 2.070 12.13 .068 0.683 3.98 23.31 0.07 0.323 4.029 1.87 0.047 0.331 4.02
Ammoniaa 424 0.018 0.660 14.6 0.031 .58 9.92 98.7 0.031 0.421 9.93 1.186 0.028 0.419 9.81
Isobutaneb 298 2.833 1.724 5.258 1.08 0.131 7.01 4.79 1.44 0.945 2.385 4.924 1.3 0.943 2.37
Isobutaneb 323 1.750 1.520 7.57 .914 0.132 7.02 6.169 1.091 0.911 5.29 6.504 0.902 0.908 5.4
Isobutaneb 373 0.312 1.352 7.76 .363 0.318 17.85 0.466 1.052 0.936 8.36 0.391 1.149 0.949 7.86

a
(29).
b
(30).

Parker (13) used the Dubinin–Astakov equation (Eq. [10]) differential molar adsorption enthalpy,
for the evaluation of adsorption potential, adsorption poten-
tial distribution, adsorption enthalpy, and adsorption entropy bA b / aaeT
(Fig. 6). The equations for the following functionalities DHads Å 0 ,
ab
were developed by Jaroniec et al. (15) and are given as
adsorption potential,
and differential molar adsorption entropy
A Å RT ln
CS
CeS D or RT ln S
fT / c
Ce D ,
DSads Å 0
aae A 10 b
, [10a]
b
adsorption potential distribution,
x(A) Å abA b 01 exp( 0aA b ), where ae is the thermal coefficient of the maximum capacity

FIG. 3. (a) Comparison of different isotherm equations for the adsorption of isobutane (gas) at different temperatures (30). (b) Comparison of
different isotherm equations for the adsorption of ammonia (gas) for different temperatures on activated carbon (29).

AID JCIS 5041 / 6g30$$$644 10-16-97 13:03:11 coida

162 KHAN, ATAULLAH, AND AL-HADDAD

TABLE 5
Comparison of Different Absorption Correlations for Different Temperatures

Dubinin–Astakov (Eq. [10a]) Generalized equation (Eq. [13a])

Sorbate T (K) qo a C f b % error qo a g d b % error

Isobutane1 298–373 4.659 2.13 1 1008 2512.7 04506.4 1.741 13.48 1.018 632.18 04747.9 019.91 0.915 6.4
Ammonia2 249–424 176.345 3.88 1 1005 1316.4 8235.3 1.155 14.37 4.13 210.16 03365.7 019.81 0.839 12.4
Phenol3 293–353 3.302 1.01 1 10025 264.92 94.22 5.59 11.54 2.756 2523.4 01358.8 012.91 0.962 11.57
Phenol4 303–328 23.762 4.93 1 1009 152.17 27921 1.947 14.79 4.5 1 1005 40.88 0193.99 17.543 0.515 6.4
p-Chlorophenol4 303–328 3.258 9.79 1 10021 38.69 3368.9 4.47 6.09 0.842 3.623 0285.6 2.217 0.802 3.54
p-Nitrophenol4 298–328 6.55 8.67 1 1002 0.131 09.71 0.272 8.17 1.83 135.94 0209.03 01.989 0.92 7.26
Propanol5 273–343 3 5.27 1 1004 0.528 2.839 0.938 15.21 0.155 45.44 01690.8 09.557 0.524 9.07
Propionitrile5 273–343 2.996 1.61 1 1007 5.527 34.842 1.702 7.4 0.143 38.651 02488.1 011.313 0.505 9.47
Cobalt phthalocyanin6 293–333 0.647 1.67 1 1003 3308.7 35519 0.5 15.47 0.046 34.231 056.787 21.49 0.895 10.54

Note. Superscripts 1–6 refer to the following references, respectively: 1, Cochran et al. (30); 2, Brunauer (29); 3, Bhatia et al. (34); 4, Present work.
5, Radke and Prausnitz (7); 6, Leitao et al. (12).

and CS is saturated solubility and the characteristic energy
is related to the coefficient a as equal to a 01 / b .
The experimental data for different temperatures have
been analyzed using the Dubinin–Astakov equation (Eq.
[10a]) and the generalized equation (Eq. [13a]) for the eval-
uation of differential molar enthalpy, DHads , and entropy,
DSads . The accuracy of the values is dependent on the degree
of fitness, which can be judged by comparing average per-
centage error. The value of ae has been calculated from
liquid density obtained from ‘‘Physical and Thermodynami-
cal Properties of Pure Chemicals’’ (32) as a function of
temperature, and CS has been assumed to be a linear function
of temperature, (CS Å cT / f ), for a narrow range of
experimental temperature. The values of differential molar
enthalpy of adsorption, DHads , for different sorbate–adsor-
bent systems have been plotted in Fig. 5b. The calculated
value of DHads using Eq. [10a], increases with increased
loading of adsorbent. This behavior is not in accordance
with commonly used isotherms except for BET multilayer
adsorption, where DHads is higher for the first monolayer
and gradually decreases as the second or multilayers are
formed. All the experimental and published data have fitted
well to the generalized isotherm correlation and have shown
minimum deviation between predicted and measured values.
From this correlation a unique value of DHads has been evalu-
ated for a given sorbate–adsorbent system, confirming
monolayer surface coverage.
CONCLUSION
Surface heterogeneity, pore structure, and surface charge
are important factors in determining the adsorption charac-
teristics of an adsorbent for a particular sorbate–adsorbent
system. For a given adsorbent, the adsorption phenomenon
for different sorbates is a strong function of solubility, activ-
ity (intermolecular forces), polarity, and molecule structure,
mass, and size. The hydrophilic and hydrophobic nature of
the molecules of chemicals (sorbates) is also important when
aqueous solutions are used for adsorption.
In general, the high values of pH suppress adsorption
on activated carbon, but addition of neutral salts expedites
adsorption. The influence of pH is attributed to the electro-
static interaction between the adsorbent surface and the sor-
bate molecule or ion.

TABLE 6
Differential Molar Enthalpy and Entropy of Adsorption

DHads DSads DHads

Sorbate (Eq. [10a], J mol01) (Eq. [10a], J mol01 K01) (Eq. [13a], J mol01) References

Isobutane 027792 093.26 039474 Cochran et al. (30)

Ammonia2 026257 088.11 027982 Brunauer (29)
Phenol3 022236 074.62 011297 Bhatia et al. (34)
Phenol4 022048 073.98 01613 Present work
p-Chlorophenol4 018102 060.75 02375 Present work
p-Nitrophenol4 02665 08.95 01738 Present work
Propanol5 06872 023.06 014075 Radke and Prausnitz (7)
Propionitrile5 012698 042.61 020686 Radke and Prausnitz (7)
Cobalt phthalocyanin6 0472 Leitao et al. (12)

Note. See Table 1 for explanation of superscripts 1–6.

AID JCIS 5041 / 6g30$$$644 10-16-97 13:03:11 coida

 ADSORPTION STUDIES OF AROMATIC POLLUTANTS 163

FIG. 4. (a) Comparison of different isotherm equations for the adsorption of propanol at different temperatures from the aqueous solution on activated
carbon. (b) Comparison of different isotherm equations for the adsorption of cobalt phthalocyanine (dye) for different temperatures from the dilute
aqueous solution on activated carbon.

Most of the reported data cover a very limited concentra- fit for the entire data. Equation [13] can cover both funda-
tion range; for high equilibrium concentrations, the pub- mental type isotherms, Freundlich for aCe & 1 and Langmuir
lished isotherm equations have large deviations, with the for b approaching unity.
exception of Eq. [2] and [13], which have shown the best The temperature influence for gaseous adsorption is given

FIG. 5. (a) Comparison of different isotherm equations for the adsorption of phenol for different temperatures from the aqueous solution on activated
carbon. (b) Comparison of different differential molar enthalpies of adsorption of various sorbates as function of relative adsorption.

AID JCIS 5041 / 6g30$$$644 10-16-97 13:03:11 coida

164 KHAN, ATAULLAH, AND AL-HADDAD

PS Sorbate gas pressure to [kPa]

form a monolayer
qe Moles of solute adsorbed [mol kg 01 ]
q0 Constant in Eq. [14]
R Gas constant [kJ kmol 01K 01 ]
T Temperature [K]

Greek Letters
ae Thermal coefficient of [K 01 ]
maximum adsorbed
amount
g Constant in Eq. [13a] [K 01 ]
d Constant in Eq. [13a]
FIG. 6. The chemical structure of cobalt phthalocyanine (dye) adsorbed
DHads Differential molar heat of [kJ kmol 01 ]
from the dilute aqueous solution on activated carbon. adsorption
DSads Differential molar entropy [kJ kmol 01 K]
of adsorption
in terms of equilibrium pressure as d ln Pe Å ( DHads / k Constant in Eq. [15]
RT 2 )dT. This equation has been introduced in the general- L13 , L31 Wilson parameter for
ized isotherm equation and successfully used in relating the interaction between
experimental data at different temperatures for adsorption in solute and solvent
both gaseous and dilute aqueous medium with minimal error. u Relative adsorption [qe /q0 ]
The other equation which has temperature dependency (Eq. s Constant in Eq. [12]
[10a]) has also been tried to represent the experimental data. n Constant in Eq. [6]
This adsorption equation for the value of constant, b equal to j Constant in Eq. [14]
unity, would relate temperature to equilibrium concentration, f Constant in Eq. [10a] [mol m03 ]
Ce , and amount adsorbed, qe , as qe /q0 Å (Ce / cT / f ) aRT . c Constant in Eq. [10a] [mol m03K 01 ]
The exponential equation has a poor fit and predicted value x Adsorption potential
of differential heat of adsorption has showed an increase distribution (Eq. [10a])
with the increase of loading of adsorbent.

NOMENCLATURE REFERENCES

1. Crittenden, J. C., Hand, H. B., Arora, H., and Lykins, B. W., Jr., J.
Am. Water Works Assn. 79, 74 (1987).
a, a1, a2 Constants in Eqs. [1–12]. 2. Noll, K. E., Gounaris, V., and Hou, W., ‘‘Adsorption Technology for
Air and Water Pollution Control,’’ Lewis Publishers, Chelsea, MI,
A Adsorption potential RT ln(Ce /CS ) 1992.
b Constant in Eqs. [2–13] 3. Yenkie, M. K. N., and Natarajan, G. S., Sep. Sci. Technol. 28, 661
Ce Equilibrium concentration [mol m03 ] (1991).
CS Concentration to form a [mol m03 ] 4. Chatzopoulos, D., Varma, A., and Irvine, R. L., AIChE J. 39, 2027
monolayer on adsorbent (1993).
5. Fritz, W., and Schlunder, E. U., Chem. Eng. Sci. 29, 1279 (1974).
surface 6. Khan, A. R., Al-Waheab, I. R., and Al-Haddad, A., Environ. Tech. 17,
De Effective diffusional [m2 s 01 ] 13 (1996).
coefficient for adsorbent. 7. Radke, C. J., and Prausnitz, J. M., Ind. Eng. Chem. Fund. 11, 445
D0 Constant in Eq. [15] (1972).
DP Pore diffusional coefficient [m2 s 01 ] 8. Fritz, W., and Schlunder, E. U., Chem. Eng. Sci. 36, 721 (1981).
9. Sheindrof, C., Rebhun, M., and Sheintuch, M., J. Colloid Interface Sci.
DS Surface diffusional [m2 s 01 ] 79, 136 (1981).
coefficient 10. Khan, A. R., Al-Bahri, T. A., and Al-Haddad, A., J. Water Res. 31,
DSO Constant in Eq. [14] 2102 (1997).
H Henry’s law constant lim (qe /Ce ) 11. Toth, J., Acta Chim. Acad. Sci. Hung. 69, 311 (1971).
k Constant in Eq. [7] Cer0 12. Leitao, A., Euridice da Conceicao, Santos, R., and Rodrigues, A., Can.
J. Chem. Eng. 70, 690 (1992).
n, n1, n2 Constant in Eq. [7] 13. Parker, G. R., Jr., Adsorption 1, 113 (1995).
Pe Equilibrium pressure of [kPa] 14. Jossens, L., Prausnitz, J. M., Fritz, W., Schlunder, E. U., and Myers,
sorbate gas A. L., Chem. Eng. Sci. 33, 1097 (1978).

AID JCIS 5041 / 6g30$$$644 10-16-97 13:03:11 coida

 ADSORPTION STUDIES OF AROMATIC POLLUTANTS
15. Jaroniec et al., J. Colloid Interface Sci. 129, 69 (1988).
16. Suwanayuen, S., and Danner, R. P., AIChE J. 26, 76 (1980).
17. Fukuchi, K., Kobuchi, S., and Arai, Y., J. Chem. Eng. Jpn. 15, 316
(1982).
18. O’Brien, J. A., and Mayers, A. I., Ind. Eng. Chem. Process Des. Dev.
24, 1181 (1985).
19. Kapoor, A., Yang, R. T., and Wong, C., Catal. Rev. Sci. Eng. 31, 129
(1989).
20. Gilliland, E. R., Baddour, R. F., Perkinson, G. P., and Saldek, K. J.,
Ind. Eng. Chem. Fundam. 13, 95 (1974).
21. Muraki, M., Iwashima, Y., and Hayakawa, T., J. Chem. Eng. Jpn. 15,
34 (1982).
22. Seidel, A., and Carl, P. S., Chem. Eng. Sci. 44, 189 (1989).
23. Al-Duri, B., and Mckay, G., Chem. Eng. Sci. 46, 193 (1991).
24. Furuya, E. G., Chang, H. T., Miura, Y., Yokomura, H., Tajima, S.,
Yamashita, S., and Noll, K. E., J. Envir. Eng. ASCE, 909 (1996).
25. Vogel, A. I., ‘‘Quantitative Chemical Analysis,’’ 5th ed., p. 408. New
York, 1989.
AID JCIS 5041 / 6g30$$$645 10-16-97 13:03:11
165
26. Myers, A. L., and Zolandz, R. R., in ‘‘Activated Carbon Adsorption of
Organics from the Aqueous Phase’’ (M. J. Mcguire and I. H. Suffet,
Eds.), Vol. 1, Chap. 12, p. 243. Ann Arbor Science, Ann Arbor, MI,
1980.
27. Rosenbrock, H. H., Computer J. 3, 175 (1960).
28. Martinez, G. M., and Basmadjain, D., Chem. Eng. Sci. 51, 1043 (1996).
29. Brunauer, S., ‘‘The Adsorption of Gases and Vapours,’’ Oxford Univer-
sity Press, Oxford, 1945.
30. Cochran, T. W., Kabel, R. L., and Danner, R. P., AIChE J. 31, 2075
(1985).
31. Mostafa, M. R., Samra, S. E., and Youssef, A. M., Indian J. Chem. 28,
946 (1989).
32. Daubert, T. E., and Danner, R. P., ‘‘Physical and Thermodynamical
Properties of Pure Chemicals Data Compilation,’’ Design Institute of
Physical Properties Data, AIChE, Hemisphere Publishing Corporation,
New York, 1989.
33. Lee, T-C., and Belfort, G., Ind. Eng. Chem. Res. Dev. 27, 951 (1988).
34. Bhatia, S. K., Kalam, A., Joglekar, H. S., and Joshi, J. B., Chem. Eng.
Comm. 98, 139 (1990).
coida