4208 Langmuir 2006, 22, 4208-4214

Influence of Temperature on the Adsorption of Mellitic Acid onto

Kaolinite
Michael J. Angove,* John D. Wells, and Bruce B. Johnson
Colloid and EnVironmental Chemistry Laboratory, La Trobe UniVersity, P.O. Box 199,
Bendigo, Victoria 3552, Australia

ReceiVed December 21, 2005. In Final Form: February 23, 2006

The adsorption of mellitic acid (benzene-1,2,3,4,5,6-hexacarboxylic acid) onto kaolinite was investigated at five
temperatures between 10 and 70 °C. Mellitic acid adsorption increased with increasing temperature at low pH (below
pH 5.5), but at higher pH, the effect of increasing temperature was to reduce the amount adsorbed. Potentiometric
titrations were conducted, adsorption isotherms were measured over the same temperature range, and the data obtained
were used in conjunction with adsorption edge and ATR-FTIR spectroscopic data to develop an extended constant
capacitance surface complexation model of mellitic acid adsorption. A single set of reactions was used to model all
data at the five temperatures studied. The model indicates that mellitic acid sorbs via outer-sphere complexation to
surface hydroxyl (SOH) groups on the kaolinite surface rather than to permanent charge sites. The reactions proposed
are SOH + L6- + 2H+ h [(SOH2)+(LH)5-]4- and SOH + L6- h [(SOH)(L)6-].6- Thermodynamic parameters
calculated from the temperature dependence of the equilibrium constants for these reactions indicate that the adsorption
of mellitic acid onto kaolinite is accompanied by a large entropy increase.

Introduction Although phyllosilicate minerals are very important in natural

Humic substances are known to sorb to mineral surfaces in
soil and aquatic systems, causing a substantial change in the
surface characteristics of the minerals.1,2 Hence, before we can
hope to understand the nature of interactions between particulate
matter and solution species in soil and sediment systems we
must first understand the nature of the interaction between the
minerals themselves and organic matter.
There have been many investigations of the sorption of humic
and fulvic acids and soil minerals in recent years which have
provided useful insights into these complex systems.3-5 However,
because fulvic and humic molecules are both large and complex,
more detailed information on the chemical nature of the interaction
can be obtained by studying the sorption behavior of relatively
simple analogues of these macromolecules. The results of these
experimental studies are often combined with spectroscopic data
in order to deduce the most likely set of surface reactions generally
by use of a surface complexation model.6-8
While the set of reactions proposed depends, at least to some
extent, on the particular surface complexation model chosen,
infrared spectroscopic information can provide direct evidence
for the nature of the complex (whether it involves inner- or outer-
sphere complexation) and on the extent of protonation of the
surface complex when the adsorbing ligand contains carboxylate
groups.9
* To whom correspondence should be addressed. E-mail: m.angove@
latrobe.edu.au.
(1) Surface and Colloid Chemistry in Natural Waters and Water Treatment;
Beckett, R., Ed.; Plenum: New York, 1990.
(2) Marshall, S. J.; House, W. A.; Russell, N. J.; White, G. F. Colloids Surf.,
A 1998, 144, 127-137.
(3) Wershaw, R. L.; Leenheer, J. A.; Sperline, R. P.; Song, Y.; Noll, L. A.;
Melvin, R. L.; Rigatti, G. P. Colloids Surf., A 1995, 96, 93-104.
(4) Vermeer, A. W. P.; van Riemsdijk, W. H.; Koopal, L. K. Langmuir 1998,
14, 2810-2819.
(5) Yoon, T. H.; Johnson, S. B.; Brown, G. E. J. Langmuir 2004, 20, 5655-
5658.
(6) Boily, J.-F.; Persson, P.; Sjöberg, S. Geochim. Cosmochim. Acta 2000, 64,
3453-3470.
(7) Lackovic, K.; Johnson, B. B.; Angove, M. J.; Wells, J. D. J. Colloid Interface
Sci. 2003, 267, 49-59.
(8) Johnson, S. B.; Yoon, T. H.; Kocar, B. D.; Brown, G. E. J. Langmuir 2004,
20, 4996-5006.
systems, there have been fewer studies of the adsorption of organic
molecules to clays than to (hydr)oxide substrates, and those that
are published have generally not involved concurrent spectro-
scopic investigation. This is not surprising as phyllosilicate
minerals generally adsorb polar organic species less strongly
than (hydr)oxide minerals, and consequently, the adsorption bands
due to the adsorbed species are often weak. Where ATR-FTIR
studies of carboxylic acid adsorption to clay minerals have been
reported, these have included neither macroscopic adsorption
measurements nor surface complexation modeling.10,11
The available spectroscopic evidence for the type of binding
between organic acids and kaolinite provides no clear gener-
alizations. Kubicki et al.10 studied the adsorption of acetic, oxalic,
benzoic, citric, and salicylic acids on kaolinite and found no
evidence of inner-sphere complexation. For most of these
adsorbates, they found that simple washing with distilled water
resulted in the ATR-FTIR spectra of the adsorbed species
disappearing. In contrast, Specht and Frimmel11 proposed inner-
sphere complexation of oxalic, malonic, and succinic acids to
kaolinite based on changes in spectra on adsorption of these
molecules.
Temperature is an important variable in natural water and soil
systems, with seasonal variations commonly exceeding 30 K. In
regions with hot climates, surface soil temperatures often reach
50 °C in summer but fall below 5 °C in winter. Temperature
variations are known to effect the uptake of nutrients by plants:
for example, Schwartz et al.12 showed that variations of
temperature in the root zone significantly affect the uptake of
phosphate and zinc by barley. Since soil organic matter influences
the availability of trace nutrients, changes in its sorption can
influence the nutrient status of plants.
(9) Boily, J.-F.; Nilsson, N.; Persson, P.; Sjöberg, S. Langmuir 2000, 16, 5719-
5729.
(10) Kubicki, J. D.; Schroeter, L. M.; Itoh, M. J.; Nguyen, B. N.; Apitz, S.
E. Geochim. Cosmochim. Acta 1999, 63, 2709-2725.
(11) Specht, C. H.; Frimmel, F. H. Phys. Chem. Chem. Phys. 2001, 3, 5444-
5449.
(12) Schwartz, S. M.; Welch, R. M.; Grunes, D. L.; Cary, E. E.; Norvell, W.
A.; Gilbert, M. D.; Meredith, M. P.; Sanchirico, C. A. Soil Sci. Soc. Am. J. 1987,
51, 371-375.

10.1021/la0534571 CCC: $33.50 © 2006 American Chemical Society

Published on Web 03/25/2006
Adsorption of Mellitic Acid onto Kaolinite
While the effect of temperature on the adsorption of metal
ions to mineral surfaces has been the subject of several
studies,13-17 there have been few studies of the effect of
temperature on the adsorption of polar organic species to clay
minerals. Those that are available are often studies on soil and
sediment systems where the surface chemistry of the substrates
is uncertain at best. An exception is the work of Ward and Brady,18
who investigated the effect of temperature on the adsorption of
oxalate onto kaolinite and found that adsorption increased at the
higher temperature below pH 6 but decreased above pH 6.
The present study investigates the effect of temperature on the
adsorption of mellitic acid (benzene hexacarboxylic acid) onto
kaolinite at a range of temperatures between 10 and 70 °C, with
the aim of elucidating the species present on the surface. It
combines potentiometric titrations and adsorption measurements
as a function of both pH and concentration with ATR-FTIR
spectroscopy. We have used a simple extended constant
capacitance surface complexation model (ECCM) to describe
the sorption data at the various temperatures using surface species
that are consistent with the spectroscopic measurements. This
model, although simplistic in its description of electrostatics and
surface properties, is capable of predicting mellitate sorption
over a wide range of temperature, pH, and concentration
conditions. The results from this study are compared with those
of a previous study of the adsorption of mellitic acid to goethite
as a function of temperature.19
Experimental Section
Characterization of Kaolinite. Acid-washed kaolinite from Ajax
Chemicals (Sydney, Australia) was used without further treatment.
The BET surface area (measured with a Micromeritics ASAP 2000
instrument after degassing at 25 °C for 18 h) was 14.7 m2 g-1. XRD
analysis showed that the kaolinite sample was free of mineral
contaminants.
Equipment and General Procedure. Adsorption experiments
and potentiometric titrations were performed in a 300-mL jacketed
reaction vessel under an atmosphere of CO2-free nitrogen. The
reaction vessel and flat flange lid of borosilicate glass were separated
by a neoprene rubber gasket. Temperature was maintained ((0.5
°C) by passing water through the jacket (20 L min-1) from a
thermostated water bath. Experiments were performed at 10, 25, 40,
55, and 70 °C. A 2-m air condenser (5 mm diameter) minimized the
loss of water by evaporation at the higher temperatures; at 70 °C,
less than 2% of the water was lost from the vessel after 72 h, the
longest period used for any titration experiment. Ross Sure-Flow
double-junction combined electrodes (Orion Research) were used
to measure pH. They were contained completely within the
thermostated vessel and calibrated with NBS standard buffers at the
reaction temperature. Electrode calibration was checked at the end
of each experiment, and the results of the experiment discarded if
the pH varied by more than 0.05 from that expected for the buffer
at that temperature. The mass of kaolinite used was that required
to give a surface concentration of 94 m2 L-1, and the supporting
electrolyte was 5 mM KNO3..
The suspension was stirred using a PTFE-coated magnetic stirrer
and was pre-equilibrated for 18-24 h under a nitrogen atmosphere
at the natural suspension pH before the commencement of any
experiment. A longer equilibration time (48 h) was allowed for
experiments at 10 °C. All experiments were performed in triplicate,
(13) Tewari, P. H.; Lee, W. J. J. Colloid Interface Sci. 1975, 52, 77-88.
(14) Fokkink, L. G. H.; de Keizer, A.; Lyklema, J. J. Colloid Interface Sci.
1990, 135, 118-131.
(15) Johnson, B. B. EnViron. Sci. Technol. 1990, 24, 112-118.
(16) Goldberg, S.; Forster, H. S. Soil Sci. 1998, 163, 109-114.
(17) Angove, M. J.; Johnson, B. B.; Wells, J. D. J. Colloid Interface Sci. 1998,
204, 93-103.
(18) Ward, D. B.; Brady, P. V. Clays Clay Miner. 1998, 46, 453-465.
(19) Angove, M. J.; Wells, J. D.; Johnson, B. B. J. Colloid Interface Sci. 2006,
296, 30-40.
Langmuir, Vol. 22, No. 9, 2006 4209
and all experimental data points are shown in the figures (with the
exception of titrations, where, for clarity, only every third point
from each titration is shown). Thus, each figure provides an indication
of the magnitude of the variation (ca. 3-5%) in the experimental
methods and data.
Equilibration Times. Kinetic experiments were conducted to
determine the period required for adsorption to reach equilibrium.
An aliquot of adsorbate was added to a mineral suspension, and the
pH varied in steps from 3 to 9. At each pH, samples were taken
immediately and then again after 5, 10, 15, 20, 30, 60, and 120 min.
Ligand concentrations in solution remained constant after ap-
proximately 10 min at all temperatures except 10 °C, where 15 min
was required.
Adsorption Edges. After pre-equilibration, the suspension was
titrated to about pH 3 with 0.100 M HNO3, and an aliquot of stock
10 mM mellitic acid solution added to provide a total mellitic acid
concentration of 100 µM. The pH was then increased in increments
of 0.2-0.5 to about 10 by addition of 0.100-M KOH. An equilibration
period of about 25 min was allowed between each increment, after
which a 5 mL sample was taken from the suspension. Samples were
filtered immediately through 0.22-µm polycarbonate membrane filters
(Poretics), and the supernatant solutions analyzed by high-
performance liquid chromatography (HPLC) (Shimadzu LC-10AI).
A Vydac #302IC4.6 anion-exchange column was used with a 5 mM
H2SO4 mobile phase. HPLC allowed the separation of nitrate, which
has a broad UV signature, from mellitic acid. Detection was by
photo-diode-array detector (Shimadzu SPD-M10AVP), with the
greatest sensitivity achieved at 230 nm. The detection limit for mellitic
acid by this method was ca. 1 µM. Adsorption of ligand to the
membrane filter was small (<5%), and once the first 2 mL of the
filtrate was discarded, the effect could be discounted: there was no
difference between results obtained by filtration and those found by
centrifugation.
Measurements of Proton Stiochiometry at Constant pH. After
the pre-equilibration period, the pH of the suspension was adjusted
to the desired experimental pH. The pH was maintained after each
addition of ligand by automated additions of standard 0.100 M KOH
or 0.100 M HNO3 from a Metrohm 665 Dosimat. The total ligand
concentration was increased in steps from 0 to approximately 100
µM by addition of aliquots from a 1-mM stock solution whose pH
had been previously adjusted to that of the experiment. The
equilibration period between successive additions was about 25 min.
Sampling and analysis were conducted as for the adsorption edge
experiments. The number of protons or hydroxyl ions released per
mellitate ion adsorbed (the proton stoichiometry) was estimated
from the volume of KOH or HNO3 dispensed by the Dosimat to
maintain the experimental pH.
Potentiometric Titrations. After pre-equilibration, the initial pH
of the suspension was noted and the suspension titrated with 0.100
M HNO3 to about pH 3.5. Solid mellitic acid (dried at 40 °C and
stored in a desiccator) was added, and the suspension allowed to
equilibrate for 3 h. The system was then titrated in steps to pH 9.5
with 0.100 M KOH using a PC-controlled titration system (Metrohm
691 pH meter and a Metrohm 665 Dosimat). After each addition of
base, the pH was monitored until the drift was less than 0.002 min-1,
a criterion that was typically achieved within 20-30 min. Most
titrations took about 48 h. In some experiments, titration speeds
were altered by relaxing or tightening the drift criterion; the elapsed
time for these titrations varied from 6 to 72 h. The experimental
difference between 48- and 72-h titrations was negligible. In some
experiments, a back-titration was also conducted, the titrant being
0.100 M HNO3. The total volume of acid used for back-titration was
within 3% of the volume of 0.100 M KOH used for the forward
titration. The forward titration data were used for modeling purposes.
The proton dissociation constants for mellitic acid have been
previously determined from potentiometric titrations.19
Kaolinite Dissolution in the Presence of Mellitic Acid. The
dissolution of kaolinite was investigated at a number of temperatures
in the presence of 1 mM and 100 µM mellitic acid between pH 3
and 10 after an equilibration time of 24 h. ICPAES (ARL 3410) was
used to monitor the concentrations of Al and Si. Results of these
4210 Langmuir, Vol. 22, No. 9, 2006 AngoVe et al.

experiments showed that over the time scale of the adsorption

experiments and titrations the extent of dissolution was negligible
when the acid concentration was 100 µM. At mellitic acid
concentrations of 1 mM, dissolution products were detected at
concentrations less than 2 µM. Therefore, over the time period and
typical pH range of all experiments, the concentrations of dissolution
products in the various adsorption systems were low enough that
their effects could be ignored. Higher concentrations of dissolution
products were detected when solutions were held below pH 3 for
more than 1 h. Hence, particular care was taken to ensure that
experiments were not conducted at pH values less than about 3.
ATR-FTIR Spectroscopy. A Bruker Equinox 55 FTIR spec-
trometer equipped with a 9-reflection DiComp/ZnSe Diamond
DuraSamplIR ATR system and a deuterated triglycine sulfate (DTGS)
detector was used to collect the IR spectra. Spectra of 20 mM solutions
of mellic acid at several pH values were obtained by collecting 100
scans at a resolution of 4 cm-1 between 650 and 4000 cm-1; a
concave lens on top of an O-ring gasket was pressed over the sample
to prevent drying or CO2 contamination during measurement of the
spectra. Spectra from the cell, water, and where necessary, water
vapor, were subtracted from those of mellitic acid solutions to yield
spectra of mellitic acid solution species.
Next, spectra of a kaolinite suspension (490 m2 L-1) in the absence
of mellitic acid were collected at several pH values. The suspension
was equilibrated at the desired pH for 24 h, and separate spectra
(500 scans) of the mineral paste and the supernatant solution collected
after centrifugation with the supernatant and paste applied uniformly
to the ATR crystal. The spectrum of the supernatant was subtracted
from that of the mineral paste. Figure 1. ATR-FTIR spectra of 10 mM solutions of mellitic acid
Finally, sufficient mellitic acid was added to the equilibrated and of 2 mM mellitic acid adsorbed onto kaolinite at the pH values
kaolinite suspension to give a concentration of 2 mM. Spectra were indicated.
collected at a range of pH values between 7 and 2.9. For each sample,
500 scans were collected from the supernatant and paste after calculating surface equilibrium constants. However, the effect of
centrifugation. The supernatant spectra were subtracted from the temperature on surface electrostatics can be estimated if we assume
paste spectra to remove any contribution for ligands remaining in that the capacitance is constant across the temperature range. The
solution. The spectrum of kaolinite at the appropriate pH was specific capacitance, κ, at temperature T2 can then be obtained by
subtracted from the spectrum of the mellitate/kaolinite paste in order multiplying the value at T1 (usually 298 K) by the ratio T1/T2.
to obtain the spectrum of the adsorbed mellitic acid. All calculations A more complete description of the derivation of extended constant
were accomplished with the Bruker software, Opus. capacitance models is given by Nilsson et al.22
Modeling of the Adsorption Results. The results of all The model for the kaolinite-mellitic acid system was built up in
experiments were modeled with an ECCM using the computer a stepwise fashion. Previously published model parameters for simple
programs FITEQL20 and GRFIT.21 ECCMs have been used to systems (kaolinite only and mellitic acid only)17,19 had been obtained
describe the adsorption of various benzene carboxylic acids onto a from potentiometric titration experiments involving only kaolinite
number of different surfaces.7,19,22-24 An ECCM allows adsorption or mellitic acid. These parameters were then used as constants in
to occur in two planes: ‘strongly’ bound ions adsorb at the surface the combined system and values for the additional parameters obtained
plane (0-plane) as inner-sphere complexes, and ‘weakly’ bound ions by fitting the model to the experimental data. All titration and sorption
adsorb in an outer plane (β-plane), separated from the surface by edge data were used to determine the optimal set of adsorption
at least one layer of water molecules. reactions. The lines drawn in Figures 1-4 show the fit of the model
ECCMs assume that the total capacitance (Ctot) between the surface to the experimental data. A detailed description of how surface acidity
and bulk solution is constant and can be calculated by treating the constants were obtained has been published previously.17 The model
inner and outer surface planes as capacitors in series. Thus for kaolinite surface acidity includes reactions for the protonation
and deprotonation of surface hydroxyl groups at the edges of the
1/Ctot ) 1/Cinner + 1/Couter crystals and an additional reaction

where Cinner is the specific capacitance inside the 0-plane, and Couter
is the specific capacitance between the 0- and β-planes. The value
of Ctot is obtained from acid-base constant-capacitance modeling
of the titration data of the substrate surface. Values for kaolinite
have been presented previously.17 FITEQL version 3.1 does not
allow for temperature variation of the electrostatic component when
XH + K+ h XK + H+
that accounts for the permanent-charge exchange sites present on
kaolinite due to isomorphous substitution. X represents an exchange
site, and K+ an ion from the electrolyte. The justification for this
description of the kaolinite surface can be found in a previous paper.25

(20) Herbelin, A. L.; Westall, J. C. In FITEQL: A computer program for Results

determination of chemical equilibrium constants from experimental data, Version
3.2; Department of Chemistry, Oregon State University: Corvallis, Oregon, 1996.
(21) Ludwig, C. In GrFit - A computer program for solVing speciation
problems: EValuation of equilibrium constants, concentrations and other physical
parameters; University of Berne: Switzerland, 1996.
(22) Nilsson, N.; Persson, P.; Lövgren, L.; Sjöberg, S. Geochim. Cosmochim.
Acta 1996, 60, 4385-4395.
(23) Nordin, J.; Persson, P.; Nordin, A.; Sjöberg, S. Langmuir 1998, 14, 3655-
3662.
(24) Persson, P.; Nordin, J.; Rosenqvist, J.; Lövgren, L.; Öhman, L.-O.; Sj
öberg, S. J. Colloid Interface Sci. 1998, 206, 252-266.
ATR-FTIR Spectra. Spectra of mellitate solution species
and mellitate adsorbed onto kaolinite are shown in Figure 1 as
a function of pH. From published Ka values,26 the predominant
species present at each pH in the solution spectra are L6- (pH
(25) Angove, M. J.; Johnson, B. B.; Wells, J. D. Colloids Surf., A 1997, 126,
137-147.
(26) Purdie, N.; Tomson, M. B.; Riemann, N. J. Solution Chem. 1972, 1,
465-476.
Adsorption of Mellitic Acid onto Kaolinite
Figure 2. Titration of 100 µM mellitic acid in kaolinite suspensions
(94 m2 L-1) at three temperatures: 10 (1), 40 (b), and 70 °C (2).
The lines were calculated from the surface complexation models:
the parameters are given in Table 1.
Figure 3. Adsorption of mellitic acid from 100 µM solution onto
kaolinite at five temperatures: 10 (1), 25 (O), 40 (b), 55 (4), and
70 °C (2). The lines show the best fit obtained from surface
complexation models: the parameters are given in Table 1.
Figure 4. Speciation of mellitic acid adsorbed from 100 µM solution
onto kaolinite (100 m2 L-1) predicted by the surface complexation
model for 25 °C: the parameters are given in Table 1.
7.7), H2L4- (pH 5.5), and H3L3- (pH 3.5). We have not included
the spectrum of HL5- (pH 6.5) because it is almost identical to
that for L6-.
Langmuir, Vol. 22, No. 9, 2006 4211
The mellitate solution spectra have been discussed in a previous
paper,27 and the bands for L6- and H2L4- have been assigned
on the basis of theoretical frequency calculations. The band at
1578 cm-1 is due principally to an asymmetric stretching vibration
of the carboxylate group (νC-Oas). The remaining strong bands
in the L6- spectrum are more complex, with that at 1426 cm-1
due to a combination of a C-C stretching vibration and a bending
motion of the carboxylate groups. The band at 1334 cm-1 is due
to a νC-Os vibration coupled with νring. As the solution pH is
decreased, the spectrum alters, with the bands at 1577 and 1334
cm-1 shifting to higher wavenumbers due to the presence of
-COH groups as the mellitate ion is protonated. Additional bands
at 1717 and 1254 cm-1 are also found at pH 5.5: their intensity
increases as the pH is decreased. The band at 1717 cm-1 is due
principally to a carbonyl stretching vibration (νCdO), while that
at 1254 cm-1 is complex, resulting from the splitting of the band
originally at 1334 cm-1 and the presence of -COH groups.
The adsorbed spectra were collected between pH 2.9 and 7.
The bands in the spectrum at pH 7 had low intensity because of
the small amount adsorbed to kaolinite at that pH, so that spectrum
is not shown in Figure 1. However, all adsorbed spectra collected
between pH 4.9 and 7 had similar characteristics, with bands
near 1584, 1433, and 1338 cm-1 which closely parallel the bands
at 1578, 1426, and 1334 found for the L6- solution species. The
relative intensities and widths of these bands also changed very
little with pH. The spectrum for adsorbed mellitate at pH 3.9 is
also similar, although the higher-wavenumber band had shifted
to 1588 cm-1 and showed a slight increase in intensity relative
to the other two bands. By pH 2.9, the adsorbed spectrum is
significantly different from those at higher pH, with additional
bands at 1716 and 1254 cm-1. The positions and relative intensities
of these bands are similar to those found in solution at pH 5.5.
The bandwidths at half-height of the adsorbed spectra were
generally similar to those of the solution spectra: the exception
is the band at about 1584 cm-1, which was significantly broader
in the adsorbed spectra especially at lower pH.
Potentiometric Titrations. Potentiometric titrations were
conducted at each temperature in suspensions that contained 100
µM mellitic acid and 94 m2 L-1 kaolinite. Figure 2 shows the
titration data at 10, 40, and 70 °C. The lines represent the best
fit of the extended constant capacitance surface complexation
model at each temperature using the parameters shown in Table
1.
Adsorption Edges. Figure 3 shows the mellitic acid adsorption
edges at five temperatures when the mellitic acid concentration
was 100 µM. The lines in the figure represent the fit of the
surface complexation modeling, with the parameters given in
Table 1. Again, the proposed reactions accurately predict the
trends in the adsorption data as a function of both pH and
temperature.
The adsorption edge data show some interesting temperature
effects, with mellitic acid adsorption increasing with increasing
temperature below pH 5.5 but decreasing with increasing
temperature at higher pH. The crossover point in the adsorption
edges, which occurs at about pH 5.5, will be discussed later.
Proton Stoichiometry. The proton stoichiometry, χ, was
measured at pH 4.50 and 8.50, which are, respectively, below
and above the pH of the temperature-dependence intersection
points found in the adsorption edge experiments. In these
experiments, the pH was maintained at the chosen value by adding
acid or base to compensate for the hydrogen ions released or
taken up upon adsorption of increasing amounts of mellitic acid.
The average number of protons, or hydroxyls, released per acid
(27) Johnson, B. B.; Sjöberg, S.; Persson, P. Langmuir 2004, 20, 823-828.
4212 Langmuir, Vol. 22, No. 9, 2006 AngoVe et al.

Table 1. Reactions and Constants for Mellitic Acid Adsorption onto Goethite and Kaolinite from Surface Complexation Modelinga
10 °C 25 °C 40 °C 55 °C 70 °C
pKw b 14.53 14.00 13.54 13.15 12.81
Mellitic Acid
LH5- h L6- + H+ (pKa6)c 6.4 ((0.1) 6.8 ((0.1) 7.1 ((0.1) 7.3 ((0.1) 7.5 ((0.1)
LH24- h LH5- + H+ (pKa5)c 6.4 ((0.1) 6.3 ((0.1) 6.2 ((0.2) 6.1 ((0.1) 6.0 ((0.2)
LH33- h LH24- + H+ (pKa4)c 4.1 ((0.2) 4.3 ((0.2) 4.6 ((0.3) 4.9 ((0.2) 5.1 ((0.2)
Kaolinite
XH + K+ h XK + H+ d -3.25 -2.88 -2.52 -2.36 -2.05
SOH h SO- + H+ d -8.54 -7.15 -7.59 -7.31 -7.19
SOH + H+ h SOH2+rd 2.93 3.24 3.81 3.65 4.90
SOH + L6- + 2H+ h [(SOH2)+(LH)5-]4- 16.6 ((0.2) 16.6 ((0.1) 17.3 ((0.2) 17.9 ((0.2) 18.1 ((0.2)
SOH + L6- h [(SOH)(L)6-]6- 4.4 ((0.2) 4.4 ((0.1) 4.0 ((0.1) 4.4 ((0.1) 4.4 ((0.1)
capacitance (F m-2)d inner, outer 3.1, 7.2 3.0, 7.0 2.9, 6.7 2.8, 6.4 2.6, 6.1
surface site density XH, SOH (µmol m-2)d 2.1, 2.9 ((0.2) 2.3, 3.2 ((0.2) 2.2, 3.2 ((0.2) 2.1, 3.4 ((0.2) 2.0, 3.7 ((0.2)
a
Equilibrium constants are listed as log K ((3σ); I ) 5 mM KNO3. b Reference 33. c Reference 19. d Reference 17.

Table 2. Average Proton Stoichiometry for Mellitic Acid Table 3. Thermodynamic Parameters for Mellitic Acid
Adsorption onto Kaolinite at pH 4.50 and 8.50 Adsorption onto Kaolinite from Surface Complexation Modeling
temperature (°C) χ ((0.2) ∆H ∆S
reaction (kJ mol-1) (J K-1 mol-1)
pH 4.50 10 0.0
25 0.1 SOH + L6- + 2H+ h [(SOH2)+(LH)5-]4- 50 ( 10 500 ( 30
40 0.2 SOH + L6- h [(SOH)(L)6-]6- 0(8 80 ( 20
55 0.0
70 -0.1 Adsorption entropies for both reactions are positive but are much
pH 8.50 10 0.0 larger for the formation of [(SOH2)+(LH)5-]4- than for
25 0.0 [(SOH)(L)6-].6-
40 0.1
55 0.0
70 0.0 Discussion

molecule adsorbed at the five temperatures studied is shown in
Table 2. Negative χ values indicate uptake of protons (or release
of hydroxyl ions) during adsorption. The table shows that the
proton stoichiometry of adsorption on kaolinite at both pH values
was near zero, with no apparent effect of temperature (P > 0.17)
at either pH.
Modeling the Data and Determination of Thermodynamic
Parameters. Many different adsorption reactions were tested.
Those proposed represent the only set that was consistent with
the spectroscopic data while fitting the titration and adsorption
edge data at all temperatures with the minimum number of
adjustable parameters. Mellitate adsorption onto kaolinite was
best represented by a model that included two outer-sphere
mellitate complexes. (Outer-sphere complexes are indicated by
square brackets [ ].) The full set of model parameters at each
temperature are shown in Table 1.
At lower pH, the predominant reaction proposed in the model
is
SOH + L6- + 2H+ h [(SOH2)+(LH)5-]4-
The value of the equilibrium constant for this reaction (Table 1)
increased by about 1.5 orders of magnitude as temperature
increased. The second reaction
SOH + L6- h [(SOH)(L)6-]6-
became more important at higher pH values. For this second
reaction, the equilibrium constant did not change significantly
with temperature.
Table 3 shows the values obtained for the thermodynamic
parameters from the temperature dependence of the equilibrium
constants derived from the surface complexation modeling. These
results indicate that the formation of [(SOH2)+(LH)5-]4- on
kaolinite is an endothermic process, while the enthalpy for the
formation of [(SOH)(L)6-]6- is indistinguishable from zero.
Effect of Temperature on Mellitic Acid Adsorption. Figure
3 indicates that adsorption increases with temperature at low pH
but decreases with increasing temperature at high pH. This
distinctive behavior has been noted previously for the oxalate-
kaolinite18 and mellitic acid-goethite19 systems.
For the mellitic acid-kaolinite system, both the increase in
adsorption at low pH and the decrease in adsorption at high pH
can be explained by considering the electrostatic interactions
between mellitate species and the surface sites. The equilibrium
constants for the surface protonation reactions (Table 1) show
that at higher pH, where protonation of the SOH sites is unlikely,
an increase in temperature gives rise to increased deprotonation
of both the SOH and the XH surface sites, resulting in a larger
net negative surface charge at higher temperatures. The decrease
in adsorption with increasing temperature at higher pH therefore
most probably results from the increase in electrostatic repulsion
between the negatively charged mellitate ion and the more
negative surface charge.
At low pH, most of the SOH sites are protonated. Table 1
indicates that the equilibrium concentration of SOH2+ sites
increases with temperature, making the electrostatic interaction
between these sites and the mellitate species in solution more
favorable. Thus, the amount adsorbed increases as the temperature
is increased. This increase in mellitic acid sorption with increasing
temperature at low pH is much more pronounced on kaolinite
than was found previously for goethite.19 The surface protonation
constants for goethite indicate that the surface concentration of
SOH2+ sites tends to decrease with increasing temperature at
low pH, reducing the electrostatic interaction with mellitate
species; Angove et al.19 attributed the slightly enhanced adsorption
to goethite at low pH to reduced solvation of the mellitate species.
Confirmation of the importance of electrostatic effects can be
seen from the fact that the mellitic acid adsorption edges (Figure
3) intersect near the point of zero charge (pHPZC) for kaolinite,
ca. 5.5. The surface is positively charged below the pHPZC but
negatively charged above it; in each case, the magnitude of the
Adsorption of Mellitic Acid onto Kaolinite
charge increases with temperature, respectively enhancing or
inhibiting the adsorption of negatively charged mellitate species.
Similarly, Ward and Brady18 found that at higher temperatures
the adsorption of oxalate to kaolinite increased below pH 6 but
decreased above pH 6. Other studies of anion adsorption onto
various oxides and soils28-31 have generally demonstrated a trend
for adsorption to decrease with increasing temperature over a
wide pH range. Since soils generally carry a net negative charge
at pH values above 4, these results are also in accord with our
study.
ATR-FTIR Spectroscopy. The band pattern in the ATR-
FTIR spectra of adsorbed mellitic acid changes little from pH
7 to 4.9, with three bands at 1584, 1433, and 1338 cm-1. The
spectrum at pH 3.9 is also similar, although the first and third
bands are shifted somewhat to higher wavenumbers. These spectra
have very similar band positions to the L6- solution spectrum,
though all are shifted slightly to higher wavenumbers (ca. 6
cm-1). Thus, the spectra suggest that mellitic acid is adsorbed
to kaolinite as the L6- ion at pH g 5. The relatively small shift
in band positions and the similarity in band shapes to those in
solution indicate that adsorption most probably involves outer-
sphere complexation. Between pH 4 and 5, the further small
shift in the positions of the outer bands most probably indicates
the presence of HL5-. Not until the pH is near 3 do the surface
spectra show evidence of the more-protonated H2L4- species,
with the appearance of bands at 1712 and 1254 cm-1. Hence,
the spectra are consistent with outer-sphere complexation of L6-
and HL5- over the pH range used in the adsorption experiments,
as proposed in the surface complexation model.
The increase in protonation of surface species compared with
solution species at the same pH, especially at lower pH values,
has been noted before for adsorption of various benzene-
carboxylates onto goethite.6,9,27,32 While this may be the result
of a higher pH adjacent to the positively charged surface, the fact
that the surface species at pH 4 are less protonated on kaolinite
than on goethite,27 although the positive charge on the kaolinite
surface is smaller than that on goethite, suggests that surface-
specific factors are also important.
While our infrared spectra indicate that outer-sphere com-
plexation is the dominant mode of adsorption for mellitic acid
on kaolinite, significant changes in adsorbed spectra led Specht
and Frimmel11 to propose that oxalic, malonic, and succinic acids
formed inner-sphere complexes with kaolinite. Their spectro-
scopic results are surprising, as they suggest significant adsorption
of both oxalic and succinic acid above pH 10, while Ward and
Brady18 found no significant adsorption of oxalate to kaolinite
above pH 10.
Surface Complexation Modeling. When modeling the
adsorption of mellitic acid onto kaolinite, we sought to minimize
the number of adjustable parameters while giving an adequate
fit to the experimental data. We do not claim that the model
presented here provides a complete description of all the surface
processes that occur. In fact, the spectroscopic results indicate
that it is almost certain that more highly protonated mellitic acid
species would be involved in sorption below pH 3.5. Our aim
was to find a set of surface reactions that adequately described
the potentiometric titration and adsorption data at the five
temperatures studied, while being consistent with the spectro-
(28) Barrow, N. J.; Shaw, T. C. Soil Sci. 1977, 124, 347-354.
(29) Barrow, N. J. J. Soil Sci. 1986, 37, 267-275.
(30) Barrow, N. J. J. Soil Sci. 1992, 43, 37-45.
(31) Goldberg, S.; Forster, H. S.; Heick, E. L. Soil Sci. 1993, 156, 316-321.
(32) Boily, J.-F.; Persson, P.; Sjöberg, S. J. Colloid Interface Sci. 2000, 227,
132-140.
(33) CRC Handbook of Chemistry and Physics, 72nd ed.; Lide, D. R., Ed.;
CRC Press: Boca Raton, FL, 1991.
Langmuir, Vol. 22, No. 9, 2006 4213
scopic data. The reactions and equilibrium constants generated
by fitting the model to the data were then used to generate
thermodynamic parameters which provide additional information
about the adsorption processes.
The proton stoichiometries measured during adsorption at
constant pH support the reaction mechanisms proposed in the
modeling. Speciation curves proposed by the model for mellitic
acid adsorption on kaolinite are shown in Figure 4. We can
estimate the proton stoichiometry predicted by the model for the
kaolinite system at each pH from the surface and solution species
present. The main surface reactions at pH 4.5 are
H2L4- + SOH h [(SOH2+)(LH5-)]4- (65%)
H3L3- + SOH h [(SOH2+)(LH5-)]4- + H+ (35%)
while, at pH 8.5, the reactions are
L6- + SOH h [(SOH)(L6-)]6- (50%)
L6- + SO- + H+ h [(SOH)(L6-)]6- (50%)
The predicted proton stoichiometries are therefore quite small
(χ ≈ +0.3 at pH 4.5 and χ ≈ -0.5 at pH 8.5). The experimental
values obtained from the isotherms show values of χ near zero
at both pH values, consistent with the proposed reaction scheme.
The dominant surface reaction at pH 8.5 suggests that mellitic
acid is held to the kaolinite surface through hydrogen bonding
to neutral SOH sites. This mechanism corresponds with that
proposed by Ward and Brady18 for the adsorption of oxalic acid
onto kaolinite. While at first sight this reaction may appear to
contradict the importance of electrostatic interactions in deter-
mining the temperature dependence of adsorption, it need not.
As the surface becomes more negatively charged at higher pH,
there will be fewer SOH sites and the net negative charge will
repel negatively charged mellitate ions, thereby decreasing
adsorption.
Previous models for the adsorption of benzenecarboxylates
onto goethite have proposed outer-sphere complexation with
SOH2+ sites only.6,9,19 For goethite, the relatively high pHPZC
(∼8.5) means that significant concentrations of SOH2+ sites are
available for adsorption even at relatively high pH.
Thermodynamic Parameters. The thermodynamic param-
eters obtained from the surface complexation modeling of mellitic
acid adsorption on kaolinite indicate that adsorption enthalpies
are relatively small. On the other hand, the entropy change for
the formation of [(SOH2+)(HL5-)]4-, estimated from the variation
with temperature of the equilibrium constants, is relatively large.
This most probably reflects a marked change in the hydration
of both the mellitate ions and the charged surface upon adsorption.
The reduction in the net charge of the complex compared with
the separate reactants will result in a substantial liberation of
water molecules. Not surprisingly, the entropy change for the
formation of [(SOH)(L6-)]6- is much smaller, as there is no
change in net charge on formation of that complex.
It is difficult to compare the thermodynamic parameters
determined in this study with those found in other work since
the values depend both on the model used and, to some extent,
on the experimental procedure employed. With these reservations,
we make the following comments. In a study of organic acid
adsorption onto kaolinite at two different temperatures, Ward
and Brady18 deduced that the enthalpy change for the formation
of an outer-sphere oxalate-kaolinite surface complex was small,
in agreement with our results for the mellitate-kaolinite system.
This small enthalpy change is consistent with the notion that
4214 Langmuir, Vol. 22, No. 9, 2006
mellitate-kaolinite bonds are the result of electrostatic and
hydrogen-bonding interactions (outer-sphere complexation). The
large entropy change found on formation of [(SOH2+)(LH5-)]4-
is consistent with the values found for the similar complexes
formed on adsorption of mellitate onto goethite.19 This indicates
that solvent effects are likely to play an important role in the
adsorption of organic anions.
Conclusion
Mellitic acid adsorption increases at higher temperature when
the pH is below pHPZC, while at pH values above the pHPZC,
adsorption decreases as temperature increases. This behavior
most probably results from the effect of temperature on the surface
charge of kaolinite: the net positive surface charge increases
with increased temperature at low pH, but the net negative charge
increases with increasing temperature at higher pH. Thus, the
relative concentration of surface sites on which adsorption occurs
is the major determinant of the adsorption of mellitic acid onto
kaolinite.
ATR-FTIR spectroscopy and surface complexation modeling
suggest that mellitic acid is bound to kaolinite as outer-sphere
complexes in the pH range from 3 to 8. The principal adsorbed
species at lower pH, [(SOH2+)(HL5-)]4-, occurs over a range
of pH where solution species are protonated to a much greater
AngoVe et al.
extent. While electrostatic attraction is involved in sorption at
lower pH, the species proposed at higher pH, [(SOH)(L6-)]6-,
involves hydrogen bonding between the mellitate ion and neutral
surface species. The enthalpies of adsorption of both species are
small, but the entropy of adsorption of the lower-pH species is
large and positive, suggesting that solvent effects play an important
role. Not surprisingly, the permanent negative-charge sites on
kaolinite play no role in sorption.
These results indicate that polar organic species adsorb to
SOH groups on clay minerals principally below pH 8. Thus, at
higher pH values, both surface hydroxyl and permanent-charge
sites are available to interact with metal cations. Since polar
organic species bind by outer-sphere complexation, their car-
boxylate groups are largely free to complex with metal ions.
Sorbed humic and fulvic acids are therefore expected to increase
metal ion adsorption at lower pH, while causing little change to
sorption at higher pH values.
Acknowledgment. Financial support was provided by the
Australian Research Council Small Grants Scheme. Professor
Staffan Sjöberg is gratefully acknowledged for his contribution
and suggestions.
LA0534571