Environ. Sci. Technol.
2004, 38, 5786-5796
Adsorption of Aromatic Compounds
from Water by Treated Carbon
Materials
DANIELA M. NEVSKAIA,*
EVA CASTILLEJOS-LOPEZ,
V I C E N T A M U Ñ O Z , A N D
ANTONIO GUERRERO-RUIZ
Departamento de Quimica Inorganica y Quimica Tecnica,
Facultad de Ciencias, UNED, P Senda del Rey 9,
Madrid 28040, Spain
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Carbon materials with different textural and surface
chemistry properties have been studied to analyze their
behavior in removing aromatic compounds (phenol,
Downloaded via UNIV OF TORONTO on September 23, 2020 at 02:32:17 (UTC).
o-chlorophenol, p-nitrophenol, aniline, and phenol compound
mixtures) from water. A mesoporous high surface area
graphite and a microporous activated carbon with (HSAGox
and ACox) and without (HSAGT and ACT) oxygen surface
groups, were used as adsorbents. Apparent surface
areas, surface oxygen groups, and zero points of charge
have been determined. The adsorption behavior of single
compounds on ACT depends on the relation between the
molecular and the pore sizes. The aniline, the nitrophenol,
and the chlorophenol interact with the oxygen surface
groups of oxidized graphite, while there is no evidence of
any type of interaction of the phenol with these groups.
The adsorption of the organic compound mixtures on the
thermally treated samples is determined by the acid-base
character of the adsorbate-adsorbent, whereas on the
oxidized carbons, the controlling forces are the specific
interactions between organic molecules and the oxygenated
groups. Selectivity coefficients for the different mixtures
are presented over the entire range of adsorption.
Introduction
Waters polluted by organic compounds are very frequently
treated by means of adsorption processes in which activated
carbons act as adsorbents. So, many studies have been made
in order to understand thoroughly the adsorption mecha-
nisms involved in these processes (1-8). The information
obtained in this way leads to a better design as well as to an
improved efficiency for the activated carbons used as
adsorbents. Adsorption capacity of an activated carbon
depends on the nature of the adsorbent, the nature of the
adsorbate, and the solution conditions (pH, temperature,
ionic strength). With respect to the activated carbon surface,
the four main components are the carbon basal planes, edges
and crystal defects, ash impurities (i.e. metal oxides), and
oxygen surface groups. The latter are mainly placed on the
edges of the graphitic basal planes. Surface functional groups
can be classified as acidic (carboxyl, carbonyl, phenolic,
hydroxyl, lactone, anhydride) and basic (chromene- and
pyrone-like structures) (9). Despite being a small fraction of
the overall carbon surface, the oxygen groups are however
* Corresponding author phone: +34-913987241; fax: +34-
913986697; e-mail: dmartin@ccia.uned.es.
5786 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
very active, exhibiting a significant influence on the adsorp-
tion capacity (10).
Since phenolic molecules and related compounds are
common water contaminants, many attempts have been
made to understand how the different ring substituents affect
the adsorption (11-20). Thus, for example, Kaneko et al. (20)
observed an adsorption trend with the compound’s solubility,
while Deng et al. (17) related the uptake with the electron
density of the aromatic ring and Moustafa et al. (15) proposed
a dependence on pK values. Although the nature of the
adsorption driving forces for phenolic compounds is still
under discussion, the scales seems to tip toward the π-π
argument (19). That is, phenolic compounds (in their
uncharged form) are adsorbed on the carbon surface by
dispersion forces between the π electrons of aromatic ring
and the π electrons of the graphene layers. However, as
pointed out by many authors (19, 21-25), the pH of the
solution can affect the charge of phenolic compounds and
therefore influences the electrostatic interactions between
the adsorbent and the adsorbate.
It is unusual in wastewaters to find only a sole pollutant.
Thus, a better understanding of the multicomponent ad-
sorption from aqueous solutions is needed to improve
treatment designs. Competitive adsorption on the adsorbent
surface can take place when several aromatic compounds
are present in the water (26-29), in such a way that
compounds with similar properties (i.e., size, polarity,
interaction energy) compete for a limited number of adsorp-
tion sites. Studies of multicomponent systems (30-32), even
though very important for improving the efficiency of water
treatment, are less common, possibly as a consequence of
experimental limitations and the difficulty of interpreting
isotherms.
In a previous paper (33), it has been found that the textural
characteristics, the surface inorganic impurities, and the
intrinsic oxygen surface groups (in quite low amounts) of
the adsorbents affect the selectivity and the adsorption
behavior of phenol-aniline mixtures. Here we present a more
rigorous study of the effect of oxygen surface groups on the
selectivity and adsorption behavior of aromatic compounds.
To perform this study, more oxygen groups have been
introduced on the adsorbent surfaces by means of a chemical
treatment. In addition, the study has been extended to other
systems with the purpose of finding how the surface chemistry
affects the selectivity and the adsorption behavior of several
pairs of adsorbates with different acidity and molecular sizes.
Finally, adsorption selectivities for the different systems have
been obtained over a large range of organic pollutant
concentrations.
Experimental Section
Materials. A commercial activated carbon provided by Norit
RS 0.8 (Amersfoort, The Netherlands), here denoted as carbon
AC (particle size 0.75-1 mm) was used in this study. Prior
to surface treatments, the inorganic impurities were removed
with HCl and HF treatments (7). The high surface area
graphite (HSAG-300, particle size 15 µm) was provided by
Lonza (Switzerland).
Surface Treatments. To eliminate surface groups, the
demineralized activated carbon (AC) and the as-received
graphite (HSAG) were treated at 1173 K for 60 min in a furnace
under a helium flow (100 cm3 min-1), at a heating rate of 10
K min-1. The resulting samples were denoted as ACT and
HSAGT, respectively.
To introduce oxygen, surface groups AC and HSAG were
treated with HNO3 (10 wt % concentration). The mixture
10.1021/es049902g CCC: $27.50  2004 American Chemical Society
Published on Web 09/30/2004
(carbon and acid solution) was stirred and heated at 348 K
until almost all liquid was evaporated. The ratio of carbon/
HNO3 was 1 g/10 mL. The procedure was repeated three
times. The resulting samples were washed with doubly
distilled water using a Soxhlet apparatus until there was an
absence of NO3- ions in solution. The resulting samples were
named ACox and HSAGox, respectively.
Adsorbent Characterization. The textural characteriza-
tion of samples was carried out by N2 adsorption at 77 K. An
automatic Micromeritics ASAP 2010 volumetric system was
used to obtain the gas adsorption isotherms. The accuracy
of these measurements is 5%. The BET equation was applied
to the N2 isotherms in order to determine the apparent surface
areas. The pore size distributions were obtained by applying
the DFT (Density Functional Theory) method to the N2
adsorption isotherms. To determine the zero point of charge
(zpc), the electrophoretic mobility (µ) versuss pH of the
samples was measured in a Zeta Meter 3.0+ at 298 K. As χa
> 1 (χ is the reciprocal of Debye length and a is the particle
radius), the Smoluchowski equation (34) µ ) r0ζ/η was
applied to obtain the zeta potential (ζ), where η is the viscosity,
and r and o are the permittivity of a material and of a vacuum,
respectively. The concentration of the sample in each
determination was approximately 100 ppm. The pH was
adjusted with HCl and NaOH. Suspensions were stabilized
24 h before measurements. The surface groups were deter-
mined quantitatively following the Boehm method (35). Also,
the surface functional groups of the samples were studied
by temperature programmed decomposition (TPD) under
vacuum (36). The apparatus used for these experiments
consists of a quartz bulb directly attached to a quadrupole FIGURE 1. Pore size distribution calculated by DFT method.
mass spectrometer (Balzers QMG 421-C).
Adsorption Measurements. To obtain the adsorption TABLE 1. Structure Characteristics of Activated Carbons and
isotherms, suspensions of 0.1 g of carbon and 2 mL of Graphites
bidistilled water were placed in contact with 20 mL water
solutions of different known concentrations of phenol, DFT pore vol (cm3 g-1)
aniline, o-chlorophenol, p-nitrophenol, and phenol and/or SBET
aniline/o-chlorophenol/p-nitrophenol mixtures (50/50 molar sample (m2 g-1) total <2 nm % 2-50 nm %
ratio). The temperature of the suspensions was maintained
ACT 1080 0.404 0.333 82.4 0.071 17.6
to 298 K. pH values of the adsorbent-water suspensions
ACox 873 0.486 0.280 57.6 0.206 42.4
were as follows: for thermally treated samples 5.8 (ACT) and HSAGT 306 0.198 0.057 28.8 0.140 71.2
6.1 (HSAGT) and for oxidized samples 4.2 (ACox) and 3.7 HSAGox 104 0.132 0.016 12.0 0.116 88.0
(HSAGox). When adsorbates are added, the pH values of the
suspensions shifted toward higher values (5-9), always under
the pK values of the different compounds. Adsorption and Cj and CPh are the aniline or o-chlorophenol or p-
equilibrium was reached after 24 h with ellipsoidal stirring nitrophenol and phenol concentrations, respectively. To
(Precision Scientific 360) at 298 K. The stirring rate was 100 evaluate the constants, a calibration of the bisolute systems
rpm. The solids were separated by filtration using a 0.22 µm was made using a least-squares multiple regression technique
Millex-GV Millipore filter, and the supernatants were analyzed with a correlation coefficient value not less than 0.999.
by UV spectroscopy at 269.5 nm (phenol), 273.2 nm (o- Reproducibility of adsorbed amounts of phenol and/or
chlorophenol), 227.2 nm (p-nitrophenol), and 278.5 nm aniline/o-chlorophenol/p-nitrophenol was fairly good, within
(aniline) in a Varian CARY 1 spectrophotometer. To measure an error of about 5%.
the more concentrated solution, a quartz path length cell of
10 mm was used, while the low concentrations were measured Results and Discussion
with a quartz path length cell of 50 mm. The adsorbed Adsorbent Characterization. Table 1 summarizes some of
amounts q (mol m-2) were calculated on the basis of the the textural characteristics of the activated carbons and of
concentration changes, determined by UV spectroscopy the high surface area graphites. The HNO3 treatment gives
before and after adsorption. To determine the single- rise to a decrease of the apparent surface areas of both types
component concentration from the phenol and aromatic of materials and also produces a modification of the pore
compounds mixtures, a two equation system was solved for distribution. In this way, the oxidation causes the destruction
both compounds using the two overlapping signals from the and the widening of the activated carbon micropores, giving
UV spectra. The equations system are given by: rise to an increase in the mesopore volume. Also this
treatment leads to a diminishing and to a widening of the
A1Ph ) a1Ph,jCPh + a1,jCj graphite mesopores as can be seen in the Figure 1, where the
size derivative of pore volume (dW/dp) is presented versus
A2j ) a2Ph,jCPh + a2,jCj mean pore size.
Table 2 shows the zero point of charge (zpc) values of the
where j denotes aniline, o-chlorophenol, or p-nitrophenol; adsorbates and the amounts of surface groups as determined
A1,Ph (269.5) and A2,j (j ) 279.8, 273.2, or 227.2) are their by potentiometric titrations. When the samples are oxidized,
respective absorbances; a1Ph,j, a1,j, a2Ph,j, and a2,j are constants; the zpc shifts toward lower values. This displacement signifies

VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5787
 acids (37) and lactones (38), respectively. The CO profiles,
which result from the decomposition of surface groups that
contain one oxygen atom, display only one peak at 870 K
that is assigned to phenolic groups (38). Meanwhile the ther-
mally treated samples scarcely present signals (Figure 2).
Molecular Area Calculations. Some authors (38) have
found that the adsorbed amounts were a function of the
adsorbate size and the adsorbent pore size. In this way, the
equilibrium geometry, the angles, and the chemical bond
length between atomic centers have been calculated using
the Pc Spartan Pro program Version 1.07 (Wavefunction Inc.),
to obtain the molecular areas of the adsorbates (Figure 3).
The surface area of the molecules (σt) on face-down
orientation was calculated assuming that the molecular shape
corresponds to an ellipse (area ) πab; where a is the molecular
half length and b is the molecular half width). The values
FIGURE 2. TPD profiles of graphites and activated carbons. obtained are in the range of 40-46 Å2 (Table 3), which is in
agreement with different reported values (11, 14, 39, 40). The
surface area of the molecules for the “stand up” orientation
TABLE 2. Zero Point Charge and Boehm Titration of was calculated by multiplying the width by the thickness (c)
Adsorbentsa of the molecules.
sample zpc BG AG CG LG PG Adsorption of Single Compounds. In Figure 4, the
adsorbed amounts (mol m-2) of the four compounds on
ACT 4.0 194.4 0 0 0 0
activated carbons and high surface area graphites, with and
ACox 2.5 139.7 838.1 349.2 96.0 392.8
HSAGT 4.4 36.7 0 0 0 0 without oxygen surface groups, are depicted. To eliminate
HSAGox 1.8 70.7 336.9 174.7 50.9 111.3 the influence of the affinity between the adsorbate and the
solvent, the equilibrium concentrations of the different
a Total basic groups (B ), total acidic groups (A ), carboxyl groups
G G
compounds have been divided by their respective solubilities
(CG), lactonic groups (LG), and phenolic groups (PG) in µmol g-1.
(Cs) for plotting on the x-axes. The linear graphs (the smaller
graphs) illustrate the high concentration part of the isotherms
that the carbon surface remains negatively charged at quite (to 10-5 mol L-1), while the logarithmic representations (the
low pH values due to ionization of the introduced oxygen larger graphs) emphasize the low concentration zone (from
surface groups. Boehm titrations (Table 1) do not detect 10-7 mol L-1). The determined pH values of the suspensions
oxygen surface groups on the thermally treated surfaces. range from 5 and 9; this allows us to conclude that the
When the surfaces are treated with HNO3 solution, carboxylic adsorbed compounds are in their molecular (uncharged)
and phenolic groups are mainly introduced on the carbon forms. This means that adsorption of these compounds
surface. As can be seen in Table 2, the nitric treatment should be caused by weak dispersion forces (19), and
introduces more oxygen surface groups per weight on the electrostatic forces do not participate.
AC surface than on the HSAG (but per unit area, the density Adsorption on Activated Carbons. On the thermally treated
of oxygen groups on HSAGox surface is higher than on ACox). activated carbon (Figure 4 ACT), the adsorption capacity trend
These results are in agreement with TPD experiments is Ph ∼ ClPh ∼ An > NO2Ph. Relationship between this trend
(Figure 2). The CO2 desorption profiles of the oxidized sam- and the adsorbate acidity (NO2Ph > ClPh > Ph >An) has not
ples, which result from the decomposition of surface groups been found. Likewise the aromatic ring p-electron density
that contain two oxygen atoms show a peak at 550 K and a of the compounds (An > Ph > ClPh > NO2Ph) cannot be
shoulder at 650 K. These features correspond to carboxylic related with the adsorption trend.

FIGURE 3. Compound structures.

5788 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
FIGURE 4. Adsorption isotherms of single compounds.

TABLE 3. Some Chemical Properties and Molecular Size of Adsorbates

solubility a b c σt σt
compd (mol L-1) pKa (Å)a (Å)b (Å)c (Å2)d (Å2)e
phenol (Ph) 0.998 9.8 3.96 3.19 4.11 39.7 26.2
o-chlorophenol (ClPh) 0.221 8.5 3.97 3.57 4.11 44.5 29.3
p-nitrophenol (NO2Ph) 0.122 7.2 4.59 3.19 4.11 46.0 26.2
aniline (An) 0.397 4.6f 4.05 3.19 4.11 40.5 26.2
a Molecular half length. b Molecular half width. c Thickness of the molecule. d Face-down orientation. e Stand-up orientation. f pKb.

TABLE 4. Adsorbed Amounts q (µmol m-2) of Single Aromatic Compounds and Degree of Coverage θ (%) Assuming That All
Molecules Are in a Face-down or Stand-up Orientation
compd phenol chlorophenol nitrophenol aniline
HSAGT q 3.00 3.06 3.14 2.62
θ (%) face down 71 82 87 64
HSAGox q 2.92 4.52 5.00 5.25
θ (%) face down 70 121 138 128
θ (%) stand up 46 80 79 83
ACT q 3.05 2.95 2.33 2.96
θ (%) face down 73 79 65 72
ACox q 2.00 2.12 2.04 2.30
θ (%) face down 48 57 57 56

The differences in the maximum adsorbed amounts
between Ph, ClPh, and An are in the 4% range (Table 4).
Thus, it can be said that the three compounds are adsorbed
in approximately equal amounts despite their difference in
molecular sizes or pK values. In the case of nitrophenol, the
difference in the uptake (23%) is high enough to be significant
and to merit a further explanation.
The ACT sample exhibits an important amount of
micropores of size about 5 Å (Figure 1). The greater length
of the nitrophenol (Table 3) in comparison with the other
molecules would be able to produce a greater steric hindrance
for adsorption in these micropores (or at least in a fraction
of them), and consequently, the quantity adsorbed of the
nitrophenol is small as compared to the other compounds.
This is confirmed by the fact that at lower degrees of coverage
(where the filling of smaller, greater energy pore sizes is taking
place) the difference between the adsorption isotherm of
the nitrophenol and the adsorption isotherms of the rest of
the compounds is more pronounced (Figure 4, ACT loga-
rithmic scale). This is indicative of the inaccessibility of
p-nitrophenol by the smaller micropores. This hypothesis is
also supported by the evidence that on the graphite surface
(Figure 4 HSAGT), where only mesopores are presented
(Figure 1), the adsorbed amounts of the nitrophenol are very
similar to those of the chlorophenol and the phenol
molecules. Therefore, in the case of the thermally treated
microporous activated carbon, the main factor that affects
the adsorption behavior is the relation between the porous
structure of the adsorbent and the molecular size of the used
adsorbate. However, the differences in acidity or π-electron
density of these compounds are not important enough to be
reflected in the adsorption behavior.
When the organic compounds are adsorbed on the ACox,
two main facts are observed: a change in the adsorption
trend and a decrease in the adsorbed amounts with respect
to the ACT sample. Assuming that the oxygen surface groups
produce the acidification of the carbon surface and that their
presence is better tolerated by the more basic aniline than
by the more acid-substituted phenols, the compound uptake
order (at low degrees of coverage: An > Ph > ClPh > NO2Ph)
follows closely the basicity or aromatic ring π-electron density
trend (Figure 4 ACox). At high degrees of coverage the
difference in the uptakes between the phenol, the chlo-
rophenol, and the nitrophenol is within the accuracy range

VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5789

FIGURE 5. Possible arrangement of the molecules onto oxidized
graphite.
(5%), while the aniline is adsorbed in higher amounts (13%).
Thus, in the case of the ACox sample, the influence of the
surface chemistry is shown again (i.e., higher amount of
aniline adsorption), especially at low coverage. As a conse-
quence of the oxidizing treatment, textural modifications
have been produced. The mean micropore size of the
activated carbon becomes larger (Figure 1) and the mesopore
percent also increases. These changes reduce the differences
in the uptakes and so, all phenolic compounds are adsorbed
FIGURE 6. Isotherms of adsorption of one of compounds (open symbols) with respect to the mixture on ACT. The projection on the Y-Z
axis represents the isotherm of the compound with respect to its equilibrium concentration. The projection on the X-Y axis informs about
the concentration of both compounds in the liquid phase.
5790 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
in rather similar amounts at high equilibrium concentrations.
The fact that the adsorbed amounts on the ACox decrease
in comparison with the ACT cannot be attributed to the
oxygen surface groups because the adsorbed aniline is also
decreased.
Adsorption on Graphites. On mesoporous graphites, the
surface chemistry is expected to be more important for the
adsorption behavior than are the textural properties. At high
concentrations, the thermally treated graphite shows (Figure
4 HSAGT) a similar behavior for the phenol, the chlorophenol
and the nitrophenol (5% of difference) and a lower uptake
for aniline (13% of difference), the most basic compound.
Therefore, the graphite surface has less affinity for the more
basic molecule of aniline than for the more acid-substituted
phenols. However, it seems that the difference in pK values
(Table 3) of the substituted phenols is not high enough to
result in changes in the amount adsorbed on this type of
surface.
In the case of the oxidized graphite, the adsorption trend
(An > NO2Ph > ClPh > Ph) cannot be explained by the
aromatic compounds (Figure 4 HSAGox) basicity trend. It is
also notable that the adsorbed amounts, except for the case
FIGURE 7. Isotherms of adsorption of one of compounds (open symbols) with respect to the mixture on ACox. The projection on the Y-Z
axis represents the isotherm of the compound with respect to its equilibrium concentration. The projection on the X-Y axis informs about
the concentration of both compounds in the liquid phase.

of the phenol (that remains constant), are higher on the
oxidized graphite (Table 4) by up to a factor of 2 (aniline) in
comparison with their adsorption on the thermally treated
graphite. The high increase in adsorbed amounts of the
chlorophenol, the nitrophenol, and the aniline is attributed
to their specific interactions with the graphite oxygen surface
groups.
In fact, as a consequence of the introduction of oxygen
groups, the surface hydrophobicity is changed. The surface
is more hydrophilic, allowing more water molecules to be
adsorbed (41). As the water molecules have a polar nature,
they will preferably hydrogen-bond with the oxygen surface
groups rather than adsorb on the graphite surface. The
increase in the chlorophenol, the nitrophenol, and the aniline
uptakes signifies that these aromatic compounds are able to
displace the water molecules from the adsorbent surface in
this case even more than when the oxygen groups are absent.
This fact corroborates the hypothesis that some kind of
aromatic compound-oxygen surface group interactions are
originated. Nevertheless, it seems reasonable to suppose that
the specific interactions could have a different nature for the
case of the basic aniline than for the acid chlorophenol and
nitrophenol molecules.
Surface Coverage and Molecular Orientation. With the
purpose of knowing if the adsorbent surface is totally or
partially covered by the adsorbate molecules, as well as if
some changes in the molecular orientation are taking place
during the adsorption, experimental and theoretical uptakes
are compared using the following equations:
qexp
θ) (1)
qt
1026
qt (µmol m-2) ) (2)
σt N A
where qexp and qt are the experimental and theoretical
maximum adsorbed amounts (µmol m-2), respectively; NA
is the Avogadro constant; and θ is the degree of coverage.

VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5791
FIGURE 8. Isotherms of adsorption of one of compounds (open symbols) with respect to the mixture on HSAGT. The projection on the
Y-Z axis represents the isotherm of the compound with respect to its equilibrium concentration. The projection on the X-Y axis informs
about the concentration of both compounds in the liquid phase.
The degrees of coverage vary from 64 to 87% for thermally
treated adsorbents and fall to 48-57% for oxidized activated
carbon (Table 4). Furthermore, from the degrees of coverage
it is reasonable to conclude that the aromatic molecules are
adsorbed mainly in a flat orientation.
In the case of the oxidized graphite, the degrees of coverage
of the chlorophenol, the nitrophenol, and the aniline are
higher than 100% considering a stacked orientation of the
molecules. This indicates the possibility of a bilayer formation
or a change in the orientation of the adsorbed molecules.
Bertoncini et al. (41) reported that phenol adsorption on a
graphite surface on a stand-up orientation. However, bilayer
formation has not been found for phenolic compounds in
the consulted literature. So, assuming that the chlorophenol,
the nitrophenol, and the aniline are adsorbed in a stand-up
orientation, around 80% of the surface coverage is reached,
while the phenol covers only about 46% of the surface.
Nevertheless, the fact that the phenol uptake scarcely varies
when the oxygenated groups are introduced (Table 4) suggests
that interactions of the phenol with these groups do not take
place and no change of orientation for the adsorbed
molecules is produced. Also an intermediate situation can
5792 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
be considered. A 100% of the coverage could be reached,
with the adsorbed molecules linked to oxygen surface groups
in a stand up arrangement (63% for aniline, 65% for
nitrophenol, and 51% for chlorophenol) and the molecules
adsorbed on the surface free of oxygen groups in a stacked
orientation (37% for aniline, 35% for nitrophenol, and 49%
for chlorophenol).
The changes in the adsorption orientation support the
hypothesis of a specific interaction between the organic
molecules and the graphite oxygen surface groups. On this
point, the possible disposition of the organic compounds
adsorbed on the surface of the HSAGox can be inferred. As
mentioned previously, the oxidizing treatment on the surface
of the graphite results in several types of groups, and the
acidities of these functional groups are very different. Thus,
the carboxylic acids have a value of pK of about 4, lactones
close to 6, whereas the phenolic groups have a pK value of
approximately 9 (42). In the case of aromatic compounds,
the aniline is clearly a base whereas the nitrophenol and the
chlorophenol are acids of varying strength. Therefore, it is
reasonable to suppose that the aniline will tend to interact
with the most acid groups on the graphite surface (Figure 5).
FIGURE 9. Isotherms of adsorption of one of compounds (open symbols) with respect to the mixture on HSAGox. The projection on the
Y-Z axis represents the isotherm of the compound with respect to its equilibrium concentration. The projection on the X-Y axis informs
about the concentration of both compounds in the liquid phase.

In the case of phenolic compounds, two options are
possible. Either the interaction takes place through the
hydroxilyc group or by means of another substituent (nitro
or chloro groups). Since the phenol does not interact with
the oxygen surface groups, it seems possible that the OH
ring groups of the nitrophenol and the chlorophenol do not
react with either of the surface groups. Therefore, is proposed
that these two molecules interact with the graphite oxygen
groups through their chloro or nitro groups (Figure 5).
Stoeckli et al. (43, 44) have demonstrated that the
adsorption mechanism of the phenol is the same for
nonporous and microporous carbons. Therefore the decrease
of the quantities adsorbed on ACox noted previously could
be due to micropores preventing (due to steric hindrance)
the molecules of the adsorbate from standing on the surface
of the activated carbon.
Adsorption of Mixtures. Figures 6-9 show the partial
adsorption isotherms of the phenol-aromatic compound
systems with respect to the concentration (Ceq/Cs) in the
liquid phase of both components of the mixture for both the
heated and the oxidized activated carbons and graphites.
The total uptakes, the percent corresponding to each
compound of the mixtures, and the surface coverages are
given in Table 5. Although the initial mixtures are 50% in
each compound, the data show that the adsorbed amounts
of each component are very different.
Adsorption on Activated Carbons. On the thermally treated
activated carbon, the most acid component of the mixtures
is adsorbed in large amounts (see Figure 6). However, in
case of the phenol-nitrophenol system, the phenol uptakes
are higher than the nitrophenol ones at low coverage (see
also selectivities in Figure 10) due to the inaccessibility of the
last of the smaller micropores. On the oxidized activated
carbon (ACox, Figure 7), the preferential adsorption of
compounds is affected by factors such as the acid-basic
character and specific interactions. The adsorption behavior
of the phenol-aniline mixture is clearly directed by the acid-
basic character. That is, more basic molecules of the aniline
are retained to a greater extent on the acid sites of the oxidized
carbon than are the more acid phenol molecules. In the
phenol-chlorophenol mixture again the most basic com-
pound is adsorbed more at high equilibrium concentrations.

VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5793
FIGURE 10. Adsorption selectivities of phenol on adsorbent surfaces.
However, at lower degrees of coverage, more chlorophenol
is adsorbed (as shown in Figure 10) due to its specific
interactions with oxygen surface groups. In the case of the
phenol-nitrophenol mixture, the specific interactions are
again the driving force for the adsorption. That is, the more
acid nitrophenol is adsorbed in greater quantities than is the
more basic phenol on the acid activated carbon surface.
Therefore, when a mixture is adsorbed, the surface chemistry
can play a more important role in the process than when a
single compound is adsorbed.
Adsorption on Graphites. In the case of the thermally
treated graphite (Figure 8), as for the ACT carbon, higher
uptake occurs for the most acid component of the mixture.
5794 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
That is, more phenol is adsorbed from the mixtures than
aniline, although less than chlorophenol and nitrophenol.
The fact that more nitrophenol than phenol is adsorbed on
the mesoporous surface from the phenol-nitrophenol
mixture confirms the hypothesis of a steric hindrance on the
activated carbon surface.
On the oxidized graphite (HSAGox, Figure 9), more basic
aniline is adsorbed in higher amounts than is the phenol,
which can be attributed to specific bonds of the basic
molecules with the acid groups of this carbon. However, in
the case of the phenol-chlorophenol and the phenol-
nitrophenol mixtures, the more acid compounds (chloro-
phenol or nitrophenol) are adsorbed to a greater extent on
TABLE 5. Adsorbed Amounts q from Mixturesa and Degree of Coverage θ (%)
phenol + chlorophenol phenol + nitrophenol phenol + aniline
sample total Ph ClPh total Ph NO2Ph total Ph An
HSAGT q 3.02 28.0 72.0 3.01 23.7 76.3 2.86 55.9 44.1
θ (%) face down 78 20 58 81 17 64 69 38 31
HSAGox q 4.67 35.9 64.1 5.13 11.0 89.0 5.10 18.3 81.7
θ (%) face down 112 30 82 140 13 127 124 22 102
θ (%) stand up 74 20 54 81 9 72 81 15 66
ACT q 2.81 34.1 65.9 2.73 36.0 64.0 2.96 56.3 43.7
θ (%) face down 73 23 50 73 24 49 72 40 32
ACox q 2.52 59.5 40.5 2.18 34.0 66.0 2.85 40.2 59.8
θ (%) face down 70 36 34 58 18 40 69 27 42
a Expressed in µmol m-2 for total and in percent related with each compound.

this acidic surface. This behavior again indicates that both
molecules interact specifically with the graphite oxygen
surface groups. Tentatively, it can be assumed that the
adsorbed species are aligned as represented in Figure 5.
Surface Coverage and Molecular Orientation. As in the
case of single-compound adsorption on activated carbons
and on the thermally treated graphite, both mixture com-
pounds are adsorbed in a face-down orientation. The range
of degree of coverage of the mixtures (from 58 to 81%) is
situated between those of the corresponding single com-
pounds (Tables 4 and 5). In the case of the oxidized graphite,
if stacked orientation is admitted, the degree of coverage
exceeds 100% in all mixtures. Considering a stand-up
orientation of the adsorbed compounds, the surfaces cover-
age ranges between 74 and 81%.
Selectivity. The adsorption selectivity in solution is
characterized by the selectivity coefficient k (45, 46):
molar fraction of component i
in the carbon adsorbed mixture qi/(qi + qj)
k) )
molar fraction of component i CLi/(CLi + CLj)
in the aqueous solution mixture
(3)
Figure 10 represents the selectivity coefficients of the phenol
in the three mixtures on activated carbons and graphites.
Phenol-Aniline Mixture. The phenol-aniline system
behaves as expected. That is, on the more basic surface (ACT
or HSAGT) the phenol is selectively adsorbed, while on the
more acid surface (ACox and HSAGox), the aniline is preferred
instead of phenol. Furthermore, the selectivity is higher
toward the favored compound at lower equilibrium con-
centrations. This observation is reasonable because the
surface chemistry has more influence on the adsorption
behavior at a low degree of surface coverage.
Phenol-Chlorophenol Mixture. This system has less
difference in selectivity values among the compounds in the
mixture. Notice that these two compounds have nearly the
same pK values. Over practically the entire range of con-
centrations used, the chlorophenol is slightly preferred by
the adsorption sites as compared to the phenol.
Phenol-Nitrophenol Mixture. On the ACT, the preference
for phenol adsorption could be attributed to the relationship
between the size of the molecules and the size of the pores.
This hypothesis is confirmed by the fact that on the HSAGT
surface, where only mesopores are present, the nitrophenol
is more selectively adsorbed. On the ACox and HSAGox
surfaces, the selectivity to the nitrophenol molecule is
explained by the specific interactions produced between the
nitrophenol compound and the oxygen surface groups. This
selectivity is greater for the HSAGox because it has a greater
amount of oxygenated groups per unit area.
This study has shown that, by unit weight, the ACT is the
adsorbent that removes the most aromatic compounds from
the water. But by unit area, the best adsorbent could be the
oxidized graphite. The activated carbons have more advan-
tages when the organic molecules have a sufficiently small
size so that they can avoid the steric hindrance during the
adsorption. However, the HSAGox allows molecules of greater
size to adsorb and also in a faster way. In fact, we have verified
in our laboratory that the adsorption kinetic of graphites is
40 times faster than that of the activated carbon.
Acknowledgments
The authors acknowledge the financial support from the
UNED under project 2000 I + D no. 46 and Spanish Ministry
of Science and Technology under project MAT2002-04189-
C02-02. Also E.C.-L. acknowledges a predoctoral research
grant from UNED.
Literature Cited
(1) Cookson, J. T., Cheremishinoff, P. N., Eclerbusch, F., Eds. Carbon
Adsorption Handbook; Ann Arbor Science: Ann Arbor, MI, 1978.
(2) Suffet, I. H., McGuire, M. J., Eds. Activated Carbon Adsorption
of Organics from the Aqueous Phase; Ann Arbor Science: Ann
Arbor, MI, 1980; Vols. 1 and 2.
(3) Slejko, F. L. Adsorption Technology: A Step-by-Step Approach to
Process Evaluation and Application; Marcel Dekker: New York,
1985.
(4) Faust, S. D.; Aly, O. M. Adsorption Processes for Water Treatment;
Butterworth Publishers: London, 1987.
(5) Perrich, J. R. Carbon Adsorption for Wastewater Treatment; CRC
Press: Boca Raton, FL, 1981.
(6) Cheremishinoff, N. P.; Cheremishinoff, P. N. Carbon Adsorption
for Pollution Control; Prentice Hall: Upper Saddle River, NJ,
1993.
(7) Nevskaia, D. M.; Santianes, A.; Muñoz, V.; Guerrero-Ruiz, A.
Interaction of aqueous solutions of phenol with commercial
activated carbons: An adsorption and kinetic study. Carbon
1999, 37, 1065-1074.
(8) Nevskaia, D. M.; Guerrero-Ruiz, A. Comparative study of the
adsorption from aqueous solutions and the desorption of phenol
and nonylphenol substrates on activated carbons. J. Colloid
Interface Sci. 2001, 234, 316-321.
(9) Boehm, H. P. Some aspects of the surface chemistry of carbon
blacks and other carbons. Carbon 1994, 32, 759-769.
(10) Leon y Leon, C.; Radovic, L. In Chemistry and Physics of Carbon;
Thrower, P., Ed.; Marcel Dekker: New York, 1994; Vol. 24.
(11) Coughlin, R W.; Ezra, F. S.; Tan R N. Influence of chemisorbed
oxygen in adsorption onto carbon from aqueous solution. J.
Colloid Interface Sci. 1968, 28, 386-396.
(12) Snoeyink, V. L.; Weber, W. J., Jr.; Mark, H. B., Jr. Sorption of
phenol and nitrophenol by active carbon. Environ. Sci. Technol.
1969, 3, 918-926.
(13) Yonge, D. R.; Keinath, T. M.; Poznanska, K.; Jiang, Z. P. Single-
solute irreversible adsorption on granular activated carbon.
Environ. Sci. Technol. 1985, 19, 690-694.
(14) Caturla, F.; Martı́n-Martinez, J. M.; Molina-Sabio, M.; Rodriguez-
Reinoso, F.; Torregrosa, R. Adsorption of substituted phenols
on activated carbon. J. Colloid Interface Sci. 1988, 124, 528-
534.
(15) Mostafa, M. R.; Samra, S. E.; Youssef, A. M. Removal of organic
pollutants from aqueous solution. Part 1. Adsorption of phenols
by activated carbons. Ind. J. Chem. 1989, 28A, 946-948.

VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5795
(16) Shirgaonkar, I. Z.; Joglekar, H. S.; Mundale, V. D.; Joshi, J. B.
Adsorption equilibrium data for substituted phenols on activated
carbon. J. Chem. Eng. Data 1992, 37, 175-179.
(17) Deng, X.; Yue, Y.; Gao, Z. Preparation and characterization of
active carbon adsorbents for wastewater treatment from
elutrilithe. J. Colloid Interface Sci. 1997, 192, 475-480.
(18) Vidic, R. D.; Suidan, M. T.; Sorial, G. A.; Brenner, R. C. Molecular
oxygen and the adsorption of phenolsseffect of functional
groups. Water Environ. Res. 1993, 65, 53-57.
(19) Radovic, L. R.; Moreno-Castilla, C.; Rivera-Utrilla, J. Chem. Phys.
Carbon 2001, 27, 227.
(20) Kaneko, Y.; Abe, M.; Ogino, K. Adsorption characteristics of
organic compounds dissolved in water on surface-improved
activated carbon fibers. Colloids Surf. 1989, 37, 211-222.
(21) Moreno-Castilla, C.; Rivera-Utrilla, J.; López-Ramón, M. V.;
Carrasco-Marı́n, F. Adsorption of some substituted phenols on
activated carbons from a bituminous coal. Carbon 1995, 33,
845-851.
(22) Sheveleva, I. V.; Zryanina, N. V.; Voit, A. V. Effect of the media
acidity on the adsorption of organic substances from aqueous
solutions by carbon fibers. Russ. J. Phys. Chem. 1991, 65, 596-
599.
(23) Müller, G.; Radke, C. J.; Prausnitz, J. M. Adsorption of weak
organic electrolytes from dilute aqueous solution onto activated
carbon. Part I. Single-solute systems. J. Colloid Interface Sci.
1985, 103, 466-483.
(24) Cooney, D. O.; Wijaya, J. In Fundamental of Adsorption; Liapis
A. I,, Ed.; American Institute of Chemical Engineers: New York,
1987; pp 185-194.
(25) Mazet, M.; Farkhani, B.; Baudu, M. Influence of heat or chemical
treatment of activated carbon onto the adsorption of organic
compounds. Water Res. 1994, 28, 1609-1617.
(26) Fritz, W.; Schlunder, E. U. Competitive adsorption of two
dissolved organics onto activated carbon. I. Adsorption equi-
libriums. Chem. Eng. Sci. 1981, 36, 721-730.
(27) Haghseresht, F.; Nouri, S.; Max Lu, G. Q. Effects of carbon surface
chemistry and solution pH on the adsorption of binary aromatic
solutes. Carbon 2003, 41, 881-892.
(28) Meghea, A.; Peleanu, I.; Mihalache, R. Competitive adsorption
of some aromatic derivatives from waste waters on activated
carbon. Sci. Technol. Environ. Prot. 1996, 3, 15-20.
(29) Ying, W. C.; Dietz, E. A.; Woehr, G. C. Adsorptive capacities of
activated carbon for organic constituents of wastewaters.
Environ. Prog. 1990, 9, 1-9.
(30) Khan, A. R.; Al-Bahri, T. A.; Al-Haddad, A. Adsorption of phenol
based organic pollutants on activated carbon from multi-
component dilute aqueous solutions. Water Res. 1997, 31, 2102-
2112.
(31) Sheindorf, C.; Rebhun, M.; Sheintuch, M. A Freundlich-type
multicomponent isotherm. J. Colloid Interface Sci. 1981, 79,
136-142.
(32) Fritz, W.; Schluender, E. U. Simultaneous adsorption equilib-
riums of organic solutes in dilute aqueous solutions on activated
carbon. Chem. Eng. Sci. 1974, 29, 1279-1282.
(33) Nevskaia, D. M.; Castillejos-Lopez, E.; Guerrero-Ruiz, A.; Muñoz,
V. Effects of the surface chemistry of carbon materials on the
5796 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
adsorption of phenol-aniline mixtures from water. Carbon 2004,
42, 653-665.
(34) Smoluchowski, M. The thermodynamics of molecular phe-
nomena. Phys. Z. 1913, 13, 1064-1088.
(35) Boehm, H. P. Chemical identification of surface groups. Adv.
Catal. 1966, 16, 179-274.
(36) Badenes, P.; Daza, L.; Rodrı́guez-Ramos, I.; Guerrero-Ruiz, A.
In Spillover and Migration of Surface Species on Catalysts; Li,
C., Xin, Q., Eds.; Studies in Surface Science and Catalysis 112;
Elsevier: Amsterdam, 1997; p 241.
(37) Cazorla-Amoros, D.; Linares-Solano, A.; Joly, J. P.; Salinas-
Martinez de Lecea, C. In situ methods used to characterize
calcium as a catalyst of carbon gasification reactions. Catal.
Today 1991, 9, 219-226.
(38) Zielke, U.; Hüttinger, K. J.; Hoffman, W. P. Surface-oxidized
carbon fibers: I. Surface structure and chemistry. Carbon 1996,
34, 983-998.
(39) Furuya, E. G.; Chang, H. T.; Miura, Y.; Noll, K. E. A fundamental
analysis of the isotherm for the adsorption of phenolic
compounds on activated carbon. Sep. Purif. Technol. 1997, 11,
69-78.
(40) Mattson, J. S.; Mark, H. B. J.; Malbin, M. Weber, W. J., Jr.;
Crittenden, J. C. Surface chemistry of active carbon: Specific
adsorption of phenols. J. Colloid Interface Sci. 1969, 31, 116-
130.
(41) Fernandez, E.; Hugi-Cleary, D.; Lopez-Ramon, V.; Stoeckli, F.
Adsorption of phenol from dilute and concentrated aqueous
solutions by activated carbons. Langmuir 2003, 19, 9719-9723.
(42) Bertoncini, C.; Odetti, H.; Bottani, E. J. Computer simulation of
phenol physisorption on graphite. Langmuir 2000, 16, 7457-
7463.
(43) Bandosz, T. J.; Jagiello, J.; Contescu, C.; Schwartz, J. A.
Characterization of the surfaces of activated carbons in terms
of their acidity constant distributions. Carbon 1993, 31, 1193-
1202.
(44) Stoeckli, F.; Lopez-Ramon, M. V.; Moreno-Castilla, C. Adsorption
of phenolic compounds from aqueous solutions, by activated
carbons, described by the Dubinin-Astakhov equation. Lang-
muir 2001, 17, 3301-3306.
(45) Stoeckli, F.; Lopez-Ramon, M. V.; Hugi-Cleary, D.; Guillot, A.
Micropore sizes in activated carbons determined from the
Dubinin-Radushkevich equation. Carbon 2001, 39, 1115-1116.
(46) Khan, A. R.; Al-Waheab, I. R.; Al-Haddad, A. A generalized
equation for adsorption isotherms for multi-component organic
pollutants in dilute aqueous solution. Environ. Technol. 1996,
17, 13-23.
(47) Everett, D. H. Thermodynamics of adsorption from solution. II.
Imperfect systems. Trans. Faraday Soc. 1965, 61, 2478-2495.
Received for review January 20, 2004. Revised manuscript
received July 30, 2004. Accepted August 17, 2004.
ES049902G