Geochemistry of inorganic arsenic and selenium in a tropical soil effect of reaction time, ph, and comp

Geochemistry of inorganic arsenic and selenium
in a tropical soil: effect of reaction time, pH,
and competitive anions on arsenic and selenium adsorption
Kok-Hui Goh, Teik-Thye Lim
*
Environmental Engineering Research Centre, School of Civil and Environmental Engineering,
Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore
Received 8 May 2003; received in revised form 12 November 2003; accepted 18 November 2003
Abstract
Factors that can affect As and Se adsorption by soils influence the bioavailability and mobility of these elements in
the subsurface. This research attempted to compare the adsorption capacities of As(III), As(V), Se(IV), and Se(VI) on a
tropical soil commonly found in Singapore in a single-species system. The effect of reaction time, pH, and competitive
anions at different concentrations on the adsorption of both As and Se species were investigated. The As and Se
adsorption isotherm were also obtained under different background electrolytes. The batch adsorption experiments
showed that the sequence of the As and Se adsorption capacities in the soil was As(V) > Se(IV) > As(III) > Se(VI). The
adsorption kinetics could be best described by the Elovich equation. The adsorption of As(V), Se(IV), and Se(VI)
appeared to be influenced by the variable pH-dependent charges developed on the soil particle surfaces. Phosphate had
more profound effect than SO
2À
4
on As and Se adsorption in the soil. The competition between PO
3À
4
and As or Se
oxyanions on adsorption sites was presumably due to the formation of surface complexes and the surface accumulation
or precipitation involving PO
3À
4
. The thermodynamic adsorption data for As(V) and Se(IV) adsorption followed the
Langmuir equation, while the As(III) and Se(VI) adsorption data appeared to be best-represented by the Freundlich
equation.

Ó 2003 Elsevier Ltd. All rights reserved.
Keywords: Arsenic; Selenium; Adsorption; Kinetics; Elovich model; Competitive anions
1. Introduction
Arsenic and selenium are among the inorganic con-
taminants that have become of evolving environmental
concern lately. The accumulation of As and Se in soils,
aquifer sediments and drinking water through various
pathways has threatened the health of plants, wildlife,
and human beings. The presence of these ions in the
environment is regulated by many environmental and
public health agencies or authorities. For example,
under the Safe Drinking Water Act of United States,
Maximum Contaminant Levels (MCLs) in drinking
water established for As and Se are 10 and 50 lg/l,
respectively. The new standard for As (Arsenic Rule)
was adopted by USEPA on January 22, 2001 to replace
the old standard of 50 lg/l in drinking water. The rule
became effective on February 22, 2002. Sources of As
and Se contamination are predominately associated with
anthropogenic activities, arising from application of
agricultural pesticides, disposal of industrial wastes,
landfilling of sewage sludges, and combustion of fuels
(Grossl et al., 1997; Pezzarossa and Petruzzelli, 2001).
Public and political concerns have arisen as a result of
*
Corresponding author. Tel.: +65-6790-6933; fax: +65-6791-
0676.
E-mail address: (T T. Lim).
0045-6535/$ - see front matter Ó 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chemosphere.2003.11.041

Chemosphere 55 (2004) 849–859
www.elsevier.com/locate/chemosphere
the potential groundwater and surface water contami-
nation by these contaminants.
Arsenic can exist in inorganic form, organic form,
and gaseous state. The oxidation states of As in the
natural systems are )3, 0, +3, and +5. The main inor-
ganic forms of As in contaminated soils and sediments
are +5 and +3 (Harper and Haswell, 1988) but some-
times the oxidation states of )3 and 0 are expected to be
found in highly reducing conditions (McBride, 1994).
Oxidation states play a significant role in determining
the potential mobility and sensitivity of As toward
changes of the environmental conditions in soils. Arse-
nite is much more toxic (Ferguson and Gavis, 1972),
more soluble and therefore more mobile as compared to
As(V). Arsenite (as H
3
AsO
3
and H
2
AsO
À
3
) normally
predominate in slightly reduced soils whereas As(V) (as
H
2
AsO

À
4
and HAsO
2À
4
) occur predominantly in well-
oxidized soils.
Selenium can also be present in inorganic and or-
ganic forms, and sometimes in inorganic–organic form.
Selenium can easily form compounds with metals and
occurs in about 50 minerals (Kabata-Pendias and Pen-
dias, 2001). It is present in four different oxidation states
in aqueous and subsurface systems, namely )2, 0, +4,
and +6. The fate and transport of Se in contaminated
sites are very much influenced by its chemical form and
speciation. Selenite and Se(VI) are the predominantly Se
species found in contaminated soils. According to Neal
and Sposito (1989), Se(IV) is strongly adsorbed by soils
while Se(VI) is only weakly sorbed and leaches easily.
Selenide and elemental Se are usually found in reducing
environments and are unavailable to plants and animals.
Selenite is present in mildly oxidizing, neutral pH envi-
ronments and typical humid regions, while Se(VI) is the
predominant form under ordinary alkaline and oxidized
conditions.
Adsorption is one of the most commonly reported,
and possibly the initial reaction to occur when As or Se
interacts with soils. Adsorption of inorganic As and Se
on soil is of paramount importance because this process
regulates the fate and mobility of As and Se in soil.

Although the overall rate of adsorption relies on
numerous factors, the adsorption study of As and Se on
soils has been mainly linked to environmental factors
such as pH (Pierce and Moore, 1982; Xu et al., 1988;
Ticknor and McMurry, 1996; Kuan et al., 1997; Man-
ning and Goldberg, 1997; Garcia-Sanchez et al., 2002;
Goldberg, 2002), redox potential, reaction time (Prasad,
1994; Carbonell Barrachina et al., 1996; Lo and Chen,
1997; Su and Suarez, 2000; O’Reilly et al., 2001), and
oxidation states of As and Se. Competitive ions and soil
properties are also being emphasized due to their key
roles in controlling As and Se mobility in soils.
Adsorption processes involving As and Se are con-
sidered to be rapid. According to Prasad (1994), it was
found that the removal of As(V) from aqueous solution
was rapid in the initial stages of contact and reached a
maximum in the range of 35–60 min for soil minerals
such as hematite and feldspar. Carbonell Barrachina
et al. (1996) observed that 80% of the total amount of
As(III) adsorbed was sorbed by Spanish soil in the first
30 min. O’Reilly et al. (2001) discovered that As(V)
sorption on goethite was initially rapid, with over 93%
As(V) adsorption within 24 h. There is noticeably little
information reported on the kinetics of both As(III) and
As(V) adsorption in the same soil system. This infor-
mation is essential because both As(III) (predominantly
in the reduced condition) and As(V) (predominantly in
the oxidized condition) are often discovered in either
redox environments because of the relatively slow redox
transformation of As (Masscheleyn et al., 1991). The

conflicting results were also observed for the kinetic
study on Se adsorption. Lo and Chen (1997) found that
Se(IV) and Se(VI) achieved adsorption equilibrium at
different times. They reported that Se(IV) adsorption on
an iron-coated sand was attained rapidly within 10 min
while Se(VI) adsorption needed a duration of 1.5 h to
reach equilibrium. Conversely, Su and Suarez (2000)
reported that sorption of both Se(IV) and Se(VI) on iron
oxides reached equilibrium at the same time which was
less than 25 min. These contradictory observations
indicate the necessity to further elucidate the effect of
reaction time on Se(IV) and Se(VI) adsorption.
Numerous studies of pH influence on As adsorption
on oxides and hydroxides suggested that As(III) had a
sorption maximum at around pH 7 (Pierce and Moore,
1982; Xu et al., 1988; Goldberg, 2002), whereas As(V)
sorption reached a maximum sorption around pH 4–7
and then decreased with more alkaline pH (Manning
and Goldberg, 1997; Garcia-Sanchez et al., 2002;
Goldberg, 2002). On the other hand, a few researchers
studied the adsorption of Se onto the soil minerals
(oxides, hydroxides, or clays) and found that adsorption
decreased with increasing pH (Ticknor and McMurry,
1996; Kuan et al., 1997). The mobility, bioavailability,
and toxicity of As and Se in soils may also be greatly
affected by the presence of competitive anions. Anions
such as PO
3À
4
,SO

2À
4
,CO
2À
3
, and Cl
À
can compete with
As and Se for sorption sites. Xu et al. (2002) found that
the addition of AsO
3À
3
,Cl
À
,NO
À
3
,SO
2À
4
, CrO
2À
4
, and
CH
3
COO
À
hardly affected the As(V) adsorption on
zeolite. However, Qafoku et al. (1999) reported that

SO
2À
4
was able to displace some of the adsorbed As(V)
on an amended soil. Violante and Pigna (2002) studied
the competitive sorption of PO
3À
4
and As(V) on selected
clay minerals. They found that PO
3À
4
could inhibit As(V)
sorption on clay minerals such as gibbsite and kaolinite.
Smith et al. (2002) observed that the presence of PO
3À
4
greatly decreased As(V) sorption by soils containing low
amounts of Fe oxides but had little effect on the amount
of As(V) adsorbed by soils with high Fe content.
Goldberg (2002) found that there was no competitive
850 K H. Goh, T T. Lim / Chemosphere 55 (2004) 849–859
effect of the presence of equimolar As(III) on As(V)
adsorption. On the other hand, the competitive effect of
equimolar As(V) on As(III) adsorption was small and
apparent only on kaolinite and illite in the pH range 6.5–
9. Obviously, the study of competitive anion effects on
As(III) adsorption has not gained much attention as
compared to As(V) adsorption due to the assumption
that As(III) is less strongly adsorbed on the soil or

mineral surface. However, the anion competition effect
on As(III) adsorption should not be neglected because
of its high potential toxicity and environmental rele-
vance. Some studies revealed that SO
2À
4
could compete
with Se(IV) and Se(VI) for adsorption sites on oxides or
hydroxides (Saeki et al., 1995; You et al., 2001). The
presence of PO
3À
4
was also found to reduce Se sorption
by direct competition (Dhillon and Dhillon, 2000;
Monteil-Rivera et al., 2000). Although much research
has been devoted to study the effect of reaction time, pH
and competitive anions on As and Se adsorption on
individual soil minerals, comparatively few adsorption
studies have been focused on the soil itself which may
comprise various types of minerals. In addition, the
abundance of soils in a variety of geochemical environ-
ments and their influence on adsorption of contaminants
suggested the need for more characterization of As and
Se adsorption on different soils.
This study compared the adsorption capacities of
As(III), As(V), Se(IV), and Se(VI) on a tropical soil in a
single-species system. The influence of reaction time,
pH, and competitive anions on the adsorption of
As(III), As(V), Se(IV), and Se(VI) were examined. The
adsorption isotherms with different background elec-

trolytes were also obtained.
2. Materials and methods
2.1. Soil sample and characterization
An uncontaminated reddish brown tropical soil
sample was obtained from a site in the western part of
Singapore where heavy industries are located. The
mineral composition of the soil includes quartz, clay
minerals of predominantly kaolinite, montmorillonite
and chlorite (Parashar, 1998), and other fine materials
such as Fe and Al oxides (Lim et al., 2004). The geo-
logical formation of the soil is known as the sedimentary
Jurong Formation (Pitts, 1984). The soil sample was
prepared by pulverizing the soil so that all passed
through a 150-lm stainless steel sieve. There were no soil
particles retained on the sieve, and therefore it preserved
the mineral composition of the original bulk soil. The
soil powder was then thoroughly mixed to ensure
homogeneity in mineral composition throughout, and
kept in a dry condition for subsequent characterization
and As and Se adsorption studies. Pulverization of the
soil was essential as it ensured that each small sample (as
small as 0.1 g) taken from the soil for the characteriza-
tion or adsorption experiments would adequately rep-
resent the original bulk soil in term of its composition.
Particle size distribution of the original bulk soil was
carried out by sieving the soil through the sieves ranging
from the sizes of 14 mm to 425 lm. The soil fraction that
passed through the 425-lm sieve was further collected
for particle size analysis using Malvern Microplus
Mastersizer. For the pulverized soil, the pH was mea-

sured in soil slurries with soil:water ratios of 1:1 and 1:20
using Corning pH meter model 145. HORIBA conduc-
tivity meter was used to determine the electrical con-
ductivity (EC) of the soil. The organic content in the soil
was determined as chemical oxygen demand (COD)
using the method modified from the APHA Standard
Method Part 5220c (APHA, 1998). The content of Si,
Al, Fe, Mn, As, and Se in the soil was determined by
microwave-assisted acid digestion (EPA Method 3052)
and subsequent analysis using Inductively Coupled
Plasma-Optical Emission Spectrometer (ICP–OES)
of Perkin-Elmer Optima 2000. The cation exchange
capacity (CEC) and anion exchange capacity (AEC) of
the soil at various pH values were evaluated using a
buffer salt extraction method adopted by Lim et al.
(1997). The soil surface charge density was determined
in a batch system with a constant soil:water ratio of
1:20, using a method modified from alkalimetric and
acidimetric titration methods described by Stumm
(1992). The BET surface area of the soil was determined
using Micropore System QUANTA CHROME
(AUTOSORB-1).
2.2. Adsorption experiments
All the glassware used was soaked in 5% HNO
3
for
overnight and then rinsed with Milli-Q (18.2 MX cm
resistivity) water before being used in this study. The
plastic labware was washed several times with RO/DI
water before use. Analytical reagent grade chemicals and

RO/DI water were used for preparation of all solutions
throughout the study. All experiments were performed
using the batch adsorption technique at room tempera-
ture in the 90-ml polypropylene (PP) centrifuge bottles.
The batch adsorption experiments were performed in
triplicate and carried out by using the pulverized soil as
the adsorbent and As(III), As(V), Se(IV), and Se(VI) as
the adsorbates. NaAsO
2
(Merck, FW 129.92, Assay min
98.5%), Na
2
HAsO
4
Æ 7H
2
O (Sigma, FW 312, Assay
99.4%), Na
2
SeO
3
(Fluka Chemika, FW 172.94, Assay
>95%), and Na
2
SeO
4
(Sigma, FW 188.9) were used as
the As(III), As(V), Se(IV), and Se(VI) sources, respec-
tively. A concentration of 200 lM of the adsorbate in
single-species system was used in all the adsorption

experiments, except in the experiment investigating
adsorption isotherms. The soil-to-solution ratio used in
K H. Goh, T T. Lim / Chemosphere 55 (2004) 849–859 851
this study was 1:20. The natural pH (pH around 4.7) was
selected during the adsorption experiments with the aim
of minimizing the number of dissolved As or Se species
(to only one or two predominant species), reduce the
number of possible adsorbing species and eventually
simplify the overall system. pK
a
values of arsenious acid
(H
3
AsO
3
) and arsenic acid (H
3
AsO
4
) are as follows:
pK
1
¼ 9:22, pK
2
¼ 12:13, and pK
3
¼ 13:4; pK
1
¼ 2:20,
pK

2
¼ 6:97, and pK
3
¼ 11:53, respectively (O’Neill,
1995). While pK
a
values of selenious acid (H
2
SeO
3
) and
selenic acid (H
2
SeO
4
) are as follows: pK
1
¼ 2:55 and
pK
2
¼ 8:15; pK
1
¼À3:0 and pK
2
¼ 1:66, respectively
(Faust and Aly, 1999). Hence, at pH 4.7, H
3
AsO
0
3

and
H
2
AsO
À
4
are the respective predominant dissolved
As(III) and As(V) species. While HSeO
À
3
and SeO
2À
4
are
the predominant dissolved Se(IV) and Se(VI) species at
the same pH.
2.2.1. Effect of reaction time, pH and competitive anions
The purpose of this kinetic adsorption experiment
was to identify the reaction time required for adsorption
to reach equilibrium. This experiment was carried out by
determining the amount of the adsorbates adsorbed at
various reaction times, ranging from 10 min to 24 h.
Initially, solutions containing the adsorbates in single-
species system were prepared with 0.01 M of NaCl as
background electrolyte. The soils and the solutions were
then added into each PP bottle to make up a soil:solu-
tion ratio of 1:20. Soil-less blanks (with addition of the
adsorbate only) were also prepared for determination of
initial concentration of the adsorbates (C
0

). The sus-
pensions in the bottles were continuously agitated at
35 oscillations per minute in a reciprocating shaker
and their pH values were periodically measured and
adjusted back by adding 0.01 M HCl or 0.01 M NaOH
to the natural pH of the soil. This was to eliminate the
effect of changes in pH taking place throughout the
adsorption process. At the end of the desired reaction
times, the bottles were centrifuged and the supernatants
were filtered through the 0.45-lm Whatman membrane
filters. The filtrates were subsequently acidified with
HNO
3
and kept in the refrigerator before ICP–OES
analysis.
The effect of pH on adsorption was studied by
determining the amount of the adsorbate adsorbed
within the pH range of 3–7. The adsorption at more
basic pH values was not studied because soil properties
would be damaged at this pH values and influence the
adsorption behavior of the soil. The soil was first added
into each PP bottle and followed by the addition of 1 M
sodium acetate for pH stabilization purpose. The
adsorbate with 0.01 M NaCl was then added into each
PP bottle. Soil-less blanks were also kept for C
0
deter-
mination. The suspensions in the bottles were then ad-
justed to the desired pH values by adding 0.01 M HCl or
0.01 M NaOH. The suspensions were continuously

agitated for 24 h. The pH values of the suspensions were
periodically measured and adjusted if necessary to en-
sure that adsorption had taken place consistently at the
desired pH values. The 24-h adsorption equilibration
period was chosen based on the results of the earlier
kinetic adsorption experiment. At the end of the 24-h
pH equilibration, the pH values of the suspensions were
recorded again. The bottles were subsequently centri-
fuged and the supernatants filtered. The filtrates were
then acidified and kept for ICP–OES analysis.
The competitive anions that were employed in this
study were SO
2À
4
and PO
3À
4
. The effect of these compet-
itive anions on adsorption was investigated by deter-
mining the amount of the As or Se adsorption at
different concentrations of SO
2À
4
and PO
3À
4
ranging from
0.002 to 0.05 M. The soils and the solutions were
added into each PP bottle. Soil-less blanks were also
prepared to measure C

0
. The suspensions were contin-
uously agitated for 24 h and their pH values were peri-
odically measured and adjusted back to the natural pH
of the soil. After 24 h, the samples were treated as those
carried out in the aforementioned adsorption experi-
ments.
The amount of the adsorbate adsorption was calcu-
lated as follows:
Sð%Þ¼
ðC
0
À C
f
Þ
C
0
 100% ð1Þ
where S is the adsorption of the adsorbate (%); C
0
is the
initial concentration of the adsorbate in the soil-less
blank (mg/l); C
f
is the final concentration of the adsor-
bate in the filtrate (mg/l).
Sðmg=kg soilÞ¼
ðC
0
À C

f
ÞV
M
ð2Þ
where S is the adsorption of the adsorbate (mg/kg soil),
V is the volume of the solution added (l), and M is the
soil added into each bottle (kg).
2.2.2. Adsorption isotherm
Adsorption isotherms were obtained by carrying out
adsorption experiments with fixed amount of soil but
varying initial adsorbate concentrations prepared by
adding various volumes of stock solutions of As(III),
As(V), Se(IV), and Se(VI) and making up with solution
containing Cl
À
,SO
2À
4
or PO
3À
4
and RO/DI water to the
desired volume. The concentration of Cl
À
,SO
2À
4
or PO
3À
4

in the adsorbate was 0.01 M, which served as the
background electrolyte. Soil-less blanks were also pre-
pared to determine C
0
for each adsorption determina-
tion. The suspensions in the bottles were continuously
agitated for 24 h and their pH values were periodically
measured and adjusted back to the natural pH of the
soil. At the end of the agitation, the samples were cen-
trifuged, filtered, and acidified before ICP–OES analysis.
852 K H. Goh, T T. Lim / Chemosphere 55 (2004) 849–859
3. Results and discussion
3.1. Geochemical properties of tropical soil
The tropical soil used in this study consisted of 38%
gravel, 22% sand, 21% silt, and the remaining 19% clay
and other colloidal particles. This soil is grouped as
gravelly clay loam in accordance with the USDA soil
classification system. The characteristics of the tropical
soil used in this study are shown in Table 1. It can be
seen from Table 1 that the soil was very acidic with pH
values of about 4.5 for a soil:water ratio of 1:1 and
about 4.7 for a soil:water ratio of 1:20. The parent
material, weathering processes, and absence of carbon-
ate content were factors that contribute to the acidity of
the soil. The EC value of the soil was low. Organic
matter was insignificant in the soil. Significant amount
of Fe in the soil was possibly due to the forms of Fe
oxides, Fe oxyhydroxides, or hydrous Fe oxides. The
amount of Mn in the soil was insignificant. The native
As and Se contents in the soil were less than 30 mg/kg.

The CEC value of the soil was 34 cmol/kg soil at natural
pH (pH 4.7) whereas the AEC value of the soil was 21
cmol/kg soil at the same pH value. The CEC value of the
soil was higher than the CEC of a typical clay loam
texture, which ranged from 15 to 30 cmol/kg soil (Do-
nahue et al., 1977). According to Evangelou (1998), this
might be due to the presence of montmorillonite and
oxides of Fe and Al that have high CEC values. The soil
has a point of zero charge (pzc) at pH of about 4.6 (Fig.
1). The low pzc could be due to kaolinite clay mineral in
the soil which typically has a pzc ranging from 2 to 4.6
(Evangelou, 1998). The BET surface area of the soil was
9.17 m
2
/g.
3.2. Time-dependent As and Se adsorption
The kinetics of As and Se adsorption on the tropical
soil in 0.01 M NaCl medium are depicted in Fig. 2. In
general, the adsorption capacities of the adsorbates
in decreasing order were: As(V) < Se(IV) < As(III) <
Se(VI). It was observed that the adsorption rates of
As(III), As(V), Se(IV), and Se(VI) were rapid in the first
hour and then decelerated noticeably as the reaction
plateaued after 8 h. The increases in adsorption beyond
8 h were marginal and seemed to approach equilibrium
at about 24 h, in which the percentages of As(III),
As(V), Se(IV), and Se(VI) adsorbed at this time were
58%, 92%, 72%, and 25%, respectively. It could be de-
duced from the results that As(III), As(V), Se(IV), and
Se(VI) virtually attained adsorption equilibrium at

Table 1
Geochemical properties of the tropical soil used in this study
Soil property
Particle size distribution (%)
Sand 22
Silt 21
Clay 19
Soil color Reddish
brown
pH at room temperature
(1:1 soil:water) 4.5
(1:20 soil:water) 4.7
Electrical conductivity at room temperature
(1:1 soil:water) (mS/cm) 0.08
Organic content (COD, %) BDL
Chemical content (mg/kg soil)
Si 309 000
Al 15 400
Fe 26 500
Mn 30
As <30
Se <30
CEC (cmol/kg soil) at natural pH 34
AEC (cmol/kg soil) at natural pH 21
BET surface area (m
2
/g) 9.17
BDL ¼ below detection limit.
-3.00E-01
-2.00E-01

-1.00E-01
0.00E+00
1.00E-01
2.00E-01
3.00E-01
02468101214
pH
Surface Charge Density
(mmol/g soil)
Fig. 1. Surface charge density of the tropical soil as a function
of pH.
Fig. 2. Changes of As(III), As(V), Se(IV), and Se(VI) adsorp-
tion with time.
K H. Goh, T T. Lim / Chemosphere 55 (2004) 849–859 853
about 24 h in a single-species system under comparable
operating condition. The 24-h adsorption period was
therefore chosen as the reaction time for the subsequent
adsorption experiments. The time-dependent adsorption
results of As and Se were also analyzed using several
kinetic models including zero-order model, Lagergren
(first-order) model, squared-driving force mass transfer
model, Elovich model, Ritchie (second-order) model etc.
The Elovich model was found to best describe the
adsorption kinetics of both As and Se species on the
tropical soil within a 24 h time-frame. The linear form of
the Elovich equation can be expressed as
S ¼
1
a
lnðaaÞþ

1
a
ln t

þ
1
aa

ð3Þ
or it can be simplified as (if t ) 1=aa):
S ¼
1
a
lnðaaÞþ
1
a
ln t ð4Þ
where S is the amount of adsorbate adsorbed per unit
mass of adsorbent (mg/kg) at time t while a and a are the
constants. The time-dependent adsorption results that
fitted to the Elovich kinetic model are shown in Fig. 3.
The testing of the Elovich equation is carried out by
plotting S versus ln t, in which a straight line can be
obtained. The constant a can be calculated from the
slope of the straight line. The constant a can be obtained
from the a value and the intercept on the y-axis (i.e.,
ln t ¼ 0) plot. The respective values of a, a, and the
coefficient of correlation (r) of the adsorbates are given
in Table 2.
According to Elovich rate law, the rate of As and Se

adsorption decreased exponentially with the increasing
As and Se coverage on the soil surface:
dS
dt
¼ a Á e
ÀaS
ð5Þ
where a is the reaction rate at zero coverage (initial
condition of S ¼ 0att ¼ 0 which gives Elovich equation
in the form of Eq. (3) when solving Eq. (5)). The cov-
erage scale factor a is the reciprocal of the coverage S
1=e
at which the adsorption rate has fallen to 1=e of its
initial value (Wang et al., 2000):
dS
dt

1=e
¼ a Á e
ÀaS
1=e
¼ a=e ð6Þ
It is also worthwhile noted from the literature that
the Elovich model was used by Carbonell Barrachina
et al. (1996) to adequately describe the adsorption
kinetics of As(III) on several Spanish soils.
3.3. Effect of pH on As and Se adsorption
The adsorption of As(III), As(V), Se(IV), and Se(VI)
on the tropical soil as a function of pH is shown in Fig.
4. The adsorption of As(V) increased with pH up to a

maximum of 92% or equivalent to 337 mg/kg soil at pH
4.5 and then gradually decreased with further increase
in pH. However, the adsorption rate of As(III) contin-
ued to increase when the pH values increased from 3 to
7. In the case of Se, both Se(IV) and Se(VI) decreased in
adsorption when pH increased. The adsorption of
Se(IV) fell from 83% at pH 3 to 59% at pH 7, while for
Se(VI) the percentages of adsorption dropped from 46%
at pH 3 to 15% at pH 7. It was found that the tropical
Fig. 3. Kinetic study of As(III), As(V), Se(IV), and Se(VI)
adsorption using Elovich kinetic model.
Table 2
Regression analysis results of the Elovich kinetic model for the
adsorption of As(III), As(V), Se(IV), and Se(VI)
Adsorbate a a r
As(III) 4.78 · 10
2
0.0516 0.985
As(V) 4.30 · 10
3
0.0366 0.998
Se(IV) 2.72 · 10
4
0.0566 0.988
Se(VI) 2.01 · 10
7
0.1906 0.938
Fig. 4. Effect of pH on the adsorption of As(III), As(V), Se(IV),
and Se(VI).
854 K H. Goh, T T. Lim / Chemosphere 55 (2004) 849–859

soil had greater adsorption of As(V) than the adsorption
of As(III), Se(IV), and Se(VI) from pH 3–7. According
to thermodynamic calculations reported by O’Neill
(1995), when pH < pzc (pH 4.6), the positively charged
surfaces on the soil would likely prefer the adsorption of
the monovalent As(V) oxyanion, H
2
AsO
À
4
, to the un-
charged As(III) molecule H
3
AsO
0
3
. The cause of greater
adsorption of As(V) than As(III) when pH > pzc was
believed to be partly due to the high affinity of Fe oxides
to As(V), in which pzc for Fe oxide is in the pH range of
6–8 according to Evangelou (1998). On the other hand,
Se(IV) was more favorably adsorbed than Se(VI) by the
soil from pH 3 to 7, and this observation suggested that
Se(IV) was adsorbed more than Se(VI) over a rather
wide pH range.
The declining trend in the adsorption of As(V),
Se(IV), and Se(VI) with increasing pH showed evidence
of similar behavior to that of surface charge density of
the soil (Fig. 1). The soil surface charge density de-
creased as the pH of the system rose, with which the

adsorption results showed agreement. This implied that
the decreasing adsorption as pH increased could be due
to the increasing amount of the OH
À
. This was because
OH
À
could result in electrostatic repulsion between
As(V), Se(IV), or Se(VI) oxyanions and hydroxylic
functional groups (generated by accumulation of OH
À
)
on the soil surface. Since As(III) has less negative charge
character as compared to As(V), Se(IV), and Se(VI) in
the investigated pH range, it did not exhibit repulsion
and conversely its adsorption increased with increasing
pH which matched the findings reported by earlier
researchers (Frost and Griffin, 1977; Goldberg and
Glaubig, 1988; Goldberg, 2002). On the other hand, the
adsorption of As(III), As(V), Se(IV), and Se(VI) oc-
curred at pH below pzc could be attributed to attraction
between opposite charges by electrostatic force. In short,
it could be deduced that adsorption of As and Se by the
soil was dependent on the variable charge developed on
the soil particle surface.
The decline of As(V) adsorption with decreasing pH
at pH less than 4.5 could be attributed to (i) dissolution
of the soil minerals (mainly amorphous Fe and clay
minerals) which would release the adsorbed As(V) or
reduce its adsorption and (ii) high solubility of Fe ars-

enates at acidic condition (Matera and H

echo, 2001).
Iron arsenate was a potential sink for As(V) in the
aqueous solution. At low pH, the high solubility of this
compound inhibited its precipitation and therefore re-
duced As(V) retention by the soil. The shift from anionic
As species from H
2
AsO
À
4
to neutral As species H
3
AsO
4
when the pH decreased could also contribute to the
decrease in As(V) adsorption at extremely acidic con-
dition. It was also observed that the maximum As(V)
adsorption occurred at pH around 5 according to pre-
vious works (Manning and Goldberg, 1997; Garcia-
Sanchez et al., 2002; Goldberg, 2002).
3.4. Effect of competitive anions on As and Se adsorption
Arsenite and Se(VI) adsorption decreased signifi-
cantly (P < 0:01) as the concentration of SO
2À
4
increased
as illustrated by Fig. 5. However, as the SO
2À

4
concen-
tration increased beyond 0.01 M, the amount of As(III)
and Se(VI) adsorbed was not affected. For the case of
As(V), the adsorption fluctuated marginally between
257 and 286 mg/kg soil for SO
2À
4
concentration up to
0.05 M, while Se(IV) adsorption only experienced a
minor reduction when SO
2À
4
concentration increased.
Arsenate and Se(IV) consistently showed higher adsorp-
tion values than As(III) and Se(VI) and they were less
affected by SO
2À
4
competition for adsorption sites.
Without SO
2À
4
, the maximum adsorption for As(III),
As(V), Se(IV), and Se(VI) was 45%, 70%, 76%, and 44%,
respectively.
These observations suggested that SO
2À
4
could com-

pete with As(III) and Se(VI) oxyanions for adsorption
sites. Nonetheless, the presence of SO
2À
4
could hardly af-
fect the As(V) and Se(IV) adsorption. On the other hand,
the influence of SO
2À
4
on As(V) and Se(IV) adsorption
was found insignificant when its ionic strength was fur-
ther increased. The adsorption behavior of As(V) and
Se(IV) which was unaffected by the changes of SO
2À
4
ionic strength is macroscopic evidence for strong specific
binding mechanism (inner-sphere complex) between
As(V) or Se(IV) and soil minerals (Hayes et al., 1988).
Conversely, As(III) and Se(VI) adsorbed were more easily
displaced by SO
2À
4
because both of them might only form
weak bond (outer-sphere complex) with soil minerals
according to Zhang and Sparks (1990).
In the presence of PO
3À
4
, there were significant de-
creases in As(III), As(V), Se(IV), and Se(VI) adsorption

as the concentration of PO
3À
4
increased, as illustrated in
Fig. 6. The adsorption of all As and Se oxyanions
showed a sharp decrease over a low PO
3À
4
concentration
range, and then gradually reached a plateau phase when
PO
3À
4
concentration was further increased. In general,
Fig. 5. Effect of SO
2À
4
at various concentrations on the
adsorption of As(III), As(V), Se(IV), and Se(VI).
K H. Goh, T T. Lim / Chemosphere 55 (2004) 849–859 855
the effect of competitive anion on As and Se adsorption
on the tropical soil was more evident in the case of PO
3À
4
as compared to SO
2À
4
. Arsenate and Se(IV) adsorption
were less influenced by the presence of PO
3À

4
than As(III)
and Se(VI) mainly due to the greater adsorption of
As(V) and Se(IV) by the soil. The results suggested that
the competition between PO
3À
4
and As or Se oxya-
nions on adsorption sites possibly involved two different
reactions on soil surface. The reactions were the for-
mation of surface complex and the surface accumulation
or precipitation. The increased competition due to the
presence of PO
3À
4
for soil adsorption sites might occur
through high affinity of PO
3À
4
or the effect of mass action
of the increasing concentration of PO
3À
4
in solution as
suggested by Smith et al. (2002). Phosphate could
associate with surface functional groups on the surfaces
of soil minerals such as Fe oxides, forming strong inner-
sphere surface complexes and reducing surface potential.
As a result, reduction in available adsorption sites oc-
curred and this inhibited the adsorption of As or Se.

Accumulation or precipitation of PO
3À
4
on the soil sur-
face might also promote the formation of negatively
charged surface sites and reduce surface potential.
Therefore, the electrostatic repulsion between As or Se
and the negatively charged surface sites increased and
thus reduced the adsorption of As and Se.
3.5. Adsorption isotherm
Various adsorption isotherms were used to describe
the thermodynamics of As(III), As(V), Se(IV), and
Se(VI) adsorption. In this study, As(V) and Se(IV)
adsorption isotherms could be described by Langmuir
equation while As(III) and Se(VI) adsorption isotherms
could be best represented by Freundlich equation. The
Langmuir equation can be expressed as follows:
C
e
S
¼
1
bS
m
þ
C
e
S
m
ð7Þ

where C
e
is the equilibrium concentration of As or Se
oxyanions in the solution (mg/l), S is the amount of As
or Se adsorbed at equilibrium (mg/kg soil), b is constant,
and S
m
is the maximum possible adsorption at equilib-
rium in the Langmuir equation (mg/kg soil).
The linear form of Freundlich equation is expressed
as
log S ¼ log K
p
þ n log C
e
ð8Þ
where K
p
and n are constants. The sets of best-fitted
adsorption isotherms are plotted in Fig. 7, while Table 3
shows the types of best-fitted adsorption isotherm and
the values of their parameters as well as the correlation
of determinations (r
2
).
Arsenate and Se(IV) adsorption increased signifi-
cantly with increased adsorbate loading initially and
then increased gradually when higher adsorbate loadings
were added due to surface sites saturation. As the sites
on the soil were occupied, it was increasingly more dif-

ficult for As(V) and Se(IV) to find available adsorption
sites on the soil. In contrast, the adsorption of As(III)
and Se(VI) tended to increase more steadily as the
equilibrium concentration (C
e
) increased, as compared
to As(V) and Se(IV) adsorption. Therefore, it could be
deduced that both As(III) and Se(IV) adsorption iso-
therms did not reach the maximum adsorption under the
experimental condition in this study and thus they
indicated the higher binding sites and low surface cov-
erage. Either at the lower or higher equilibrium con-
centrations, the adsorption of As(V) was the greatest
and followed by the adsorption of Se(IV), As(III), and
Se(VI).
4. Conclusions
The adsorption capacities of As and Se on the trop-
ical soil investigated generally followed the order:
As(V) > Se(IV) > As(III) > Se(VI). In addition, the
studies of pH effect, competitive anion effect as well as
the adsorption isotherm further verified this sequence of
As and Se adsorption capacity on the tropical soil. The
adsorption kinetics could be best described with Elovich
kinetic model. The adsorption of As(V), Se(IV), and
Se(VI) was dependent on the variable charge developed
on the soil surface, except for the case of As(III). The
reduction in As(V), Se(IV), and Se(VI) adsorption with
increasing pH demonstrated evidence of identical
behavior to that of the soil surface charge density. While
As(III) adsorption increased with increasing pH because

it has lesser negative charge character that prohibited
electrostatic repulsion between As(III) oxyanion and
hydroxylic functional group.
From the investigation on the effect of competitive
anions, it was found SO
2À
4
could hardly affect As(V) and
0
20
40
60
80
100
0 0.01 0.02 0.03 0.04 0.05 0.06
Adsorption (%)
0
50
100
150
200
250
300
350
Adsorption (mg/kg soil)
As(III) As(V) Se(IV) Se(VI)
0
20
40
0 0.005 0.01

Fig. 6. Effect of PO
3À
4
at various concentrations on the
adsorption of As(III), As(V), Se(IV), and Se(VI).
856 K H. Goh, T T. Lim / Chemosphere 55 (2004) 849–859
Se(IV) adsorption but could compete with As(III) and
Se(VI) for adsorption sites. Phosphate had more pro-
found effect on As(III), As(V), Se(IV), and Se(VI)
adsorption than SO
2À
4
, which could be partly attributed
to its higher negative charge compared to that of SO
2À
4
.
The competition between PO
3À
4
and As or Se oxyanions
on adsorption sites was probably caused by two different
reactions on surfaces of soil minerals. The reactions were
believed to be the formation of surface complex and
the surface accumulation or precipitation. In term of
adsorption isotherm, Langmuir isotherm could better
represent equilibrium adsorption of As(V) and Se(IV)
0.0
0.5
1.0

1.5
2.0
2.5
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0
log C
e
(mg/l)
log S (mg/kg soil)
0.01 M Chloride Ion
0.01 M Sulfate Ion
0.01 M Phosphate Ion
`
As(III)
0.0
0.1
0.2
0.3
0 1020304050
C
e
(mg/l)
C
e
/S (kg/l)
0.01 M Chloride Ion
0.01 M Sulfate Ion
0.01 M Phosphate Ion
As(V)
(a) (b)
0.0

0.1
0.2
0.3
0.4
0.5
0 102030405060
C
e
(mg/l)
C
e
/S (kg/l)
0.01 M Chloride Ion
0.01 M Sulfate Ion
0.01 M Phosphate Ion
Se(IV)
(c)
0.0
0.5
1.0
1.5
2.0
2.5
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0
log C
e
(mg/l)
log S (mg/kg soil)
0.01 M Chloride Ion
0.01 M Sulfate Ion

0.01 M Phosphate Ion
Se(VI)
(d)
Fig. 7. Best-fitted adsorption isotherms of different adsorbates on the tropical soil: (a) Freundlich isotherms of As(III) adsorption; (b)
Langmuir isotherms of As(V) adsorption; (c) Langmuir isotherms of Se(IV) adsorption; (d) Freundlich isotherms of Se(VI) adsorption.
Table 3
The parameters of best-fitted adsorption isotherms calculated for As(III), As(V), Se(IV), and Se(VI) adsorption
Adsorbate Background
electrolyte
Langmuir Freundlich
S
m
br
2
K
p
nr
2
As(III) Cl
À
32.5 0.47 0.976
SO
2À
4
16.6 0.70 0.963
PO
3À
4
6.35 0.87 0.980
As(V) Cl

À
217 0.45 0.987
SO
2À
4
208 0.40 0.996
PO
3À
4
154 0.16 0.881
Se(IV) Cl
À
145 17.3 0.999
SO
2À
4
139 2.77 0.996
PO
3À
4
81.3 0.48 0.977
Se(VI) Cl
À
35.9 0.34 0.929
SO
2À
4
18.7 0.34 0.798
PO
3À

4
24.0 0.36 0.933
K H. Goh, T T. Lim / Chemosphere 55 (2004) 849–859 857
while Freundlich isotherm was more suitable in
describing As(III) and Se(VI) equilibrium adsorption.
References
APHA, 1998. Part 5220c: Chemical oxygen demand (COD),
closed reflux, titrimetric method. In: Standard Methods for
the Examination of Water and Wastewater. American
Public Health Association, NW, Washington, DC.
Carbonell Barrachina, A.A., Burl

o Carbonell, F.M., Mataix
Beneyto, J.J., 1996. Kinetics of arsenite sorption and
desorption in Spanish soils. Commun. Soil Sci. Plant Anal.
27 (18–20), 3101–3117.
Dhillon, S.K., Dhillon, K.S., 2000. Selenium adsorption in soils
as influenced by different anions. J. Plant Nutr. Soil Sci.
163 (3), 577–582.
Donahue, R.L., Miller, R.W., Shickluna, J.C., 1977. Soils: An
Introduction to Soils and Plant Growth. Prentice-Hall,
Englewood Cliffs, NJ.
Evangelou, V.P., 1998. Environmental Soil and Water Chem-
istry. John Wiley, New York.
Faust, S.D., Aly, O.M., 1999. Chemistry of Water Treatment.
Lewis Publishers, Boca Raton, FL.
Ferguson, J.F., Gavis, J., 1972. A review of arsenic cycle in
natural waters. Water Res. 6, 1259–1274.
Frost, R.R., Griffin, R.A., 1977. Effect of pH on adsorption of
arsenic and selenium from landfill leachate by clay minerals.

Soil Sci. Am. J. 41, 53–57.
Garcia-Sanchez, A., Alvarez-Ayuso, E., Rodriguez-Martin, F.,
2002. Sorption of As(V) by some oxyhydroxides and clay
minerals. Application to its immobilization in two polluted
mining soils. Clay Minerals 37 (1), 187–194.
Goldberg, S., 2002. Competitive adsorption of arsenate and
arsenite on oxides and clay minerals. Soil Sci. Soc. Am. J. 66
(2), 413–421.
Goldberg, S., Glaubig, R.A., 1988. Anion sorption on a
calcareous, montmorillonitic soil-arsenic. Soil Sci. Soc.
Am. J. 2, 1297–1300.
Grossl, P.R., Eick, M., Sparks, D.L., Goldberg, S., Ainsworth,
C.C., 1997. Arsenate and chromate retention mechanisms
on goethite. 2. Kinetic evaluation using a pressure–jump
relaxation technique. Environ. Sci. Technol. 31, 321–326.
Harper, M., Haswell, S.J., 1988. A comparison of copper, lead
and arsenic extraction from polluted and unpolluted soils.
Environ. Technol. Lett. 9, 1271–1280.
Hayes, K.F., Papelis, L., Leckie, J.O., 1988. Modeling ionic
strength effects on anion adsorption at hydrous oxide/
solution interfaces. J. Colloid Interface Sci. 125, 717–726.
Kabata-Pendias, A., Pendias, H., 2001. Trace Elements in Soils
and Plants. CRC Press, Boca Raron, FL.
Kuan, W.H., Lo, S.L., Wang, M.K., Lin, C.F., 1997. Removal
of Se(IV) and Se(VI) from water by aluminum-oxide-coated
sans. Water Res. 32 (3), 915–923.
Lim, T.T., Tay, J.H., Teh, C.I., 1997. Sorption and speciation
of heavy metals from incinerator fly ash in a marine clay.
J. Environ. Eng. ASCE 123 (11), 1107–1114.
Lim, T.T., Tay, J.H., Wang, J.Y., 2004. Chelating agent-

enhanced heavy metal extraction from a contaminated
acidic soil. J. Environ. Eng. ASCE 130 (1), 59–66.
Lo, S.L., Chen, T.Y., 1997. Adsorption of Se(IV) and Se(VI) on
an iron-coated sand from water. Chemosphere 35 (5), 919–
930.
Manning, B.A., Goldberg, S., 1997. Arsenic(III) and arsenic(V)
adsorption on three California soils. Soil Sci. 162 (12), 886–
895.
Masscheleyn, P.H., Delaune, R.D., Patrick, W.H., 1991.
Arsenic and selenium chemistry as affected by sediment
redox potential and pH. J. Environ. Qual. 20 (3), 522–
527.
Matera, V., H

echo, I.L., 2001. Arsenic behavior in contami-
nated soils: mobility and speciation. In: Selim, H.M.,
Sparks, D.L. (Eds.), Heavy Metals Release in Soils. Lewis
Publishers, Boca Raton, FL, pp. 207–235.
McBride, M.B., 1994. Environmental Chemistry of Soils.
Oxford University Press, New York.
Monteil-Rivera, F., Fedoroff, M., Jeanjean, J., Minel, L.,
Barthes, M.G., Dumonceau, J., 2000. Sorption of selenite
(SeO
2À
3
) on hydroxyapatite: an exchange process. J. Colloid
Interface Sci. 221, 291–300.
Neal, R.H., Sposito, G., 1989. Selenate adsorption on alluvial
soils. Soil Sci. Soc. Am. J. 53, 70–74.
O’Neill, P., 1995. Arsenic. In: Alloway, B.J. (Ed.), Heavy metals

in Soils. Blackie Academic and Professional, Glasgow,
pp. 83–99.
O’Reilly, S.E., Strawn, D.G., Sparks, D.L., 2001. Residence
time effects on arsenate adsorption/desorption mechanisms
on goethite. Soil Sci. Soc. Am. J. 65, 67–77.
Parashar, S.P., 1998. Behavior of geosynthetic reinforced wall
with clay backfill. Ph.D. Thesis, Nanyang Technological
University, Singapore.
Pezzarossa, B., Petruzzelli, G., 2001. Selenium in soil: sorption
and desorption processes. In: Selim, H.M., Sparks, D.L.
(Eds.), Heavy Metals Release in Soils. Lewis Publishers,
Boca Raton, FL, pp. 191–206.
Pierce, M.L., Moore, C.B., 1982. Adsorption of arsenite and
arsenate on amorphous iron hydroxide. Water Res. 16,
1247–1253.
Pitts, J., 1984. A review of geology and engineering geology in
Singapore. Q. J. Eng. Geol. 17, 93–101.
Prasad, G., 1994. Removal of arsenic (V) from aqueous systems
by adsorption onto some geological materials, Department
of Applied Chemistry, Banaras Hindu University, India.
Qafoku, N.P., Kukier, U., Sumner, M.E., Miller, W.P.,
Radcliffe, D.E., 1999. Arsenate displacement from fly ash
in amended soils. Water Air Soil Pollut. 114 (1–2), 185–198.
Saeki, K., Matsumoto, S., Tatsukawa, R., 1995. Selenite
adsorption by manganese oxides. Soil Sci. 160 (4), 265–
272.
Smith, E., Naidu, R., Alston, A.M., 2002. Chemistry of
inorganic arsenic in soils: II. Effect of phosphorus, sodium,
and calcium on arsenic sorption. J. Environ. Qual. 31, 557–
563.

Stumm, 1992. Chemistry of the Solid–Water Interface: Pro-
cesses at the Mineral–Water and Particle–Water Interface in
Natural Systems. John Wiley, New York.
Su, C.M., Suarez, D.L., 2000. Selenate and selenite sorption on
iron oxides. Soil Sci. Soc. Am. J. 64, 101–111.
Ticknor, K.V., McMurry, J., 1996. A study of selenium and tin
sorption on granite and goethite. Radiochim. Acta 73 (3),
149–156.
858 K H. Goh, T T. Lim / Chemosphere 55 (2004) 849–859
Violante, A., Pigna, M., 2002. Competitive sorption of arsenate
and phosphate on different clay minerals and soils. Soil Sci.
Soc. Am. J. 66, 1788–1796.
Wang, Z., Ainsworth, C.C., Friedrich, D.M., Gassman, P.L.,
Joly, A.G., 2000. Kinetics and mechanism of surface
reaction of salicylate on alumina in colloidal aqueous
suspension. Geochim. Cosmochim. Acta 64 (7), 1159–1172.
Xu, H., Allard, B., Grimvall, A., 1988. Influence of pH and
organic substance on the adsorption of As(V) on geologic
materials. Water Air Soil Pollut. 40, 293–305.
Xu, Y.H., Nakajima, T., Ohki, A., 2002. Adsorption and
removal of arsenic(V) from drinking water by aluminum-
loaded Shirasu-zeolite. J. Hazard. Mat. 92 (3), 275–
287.
You, Y.W., Vance, G.F., Zhao, H.T., 2001. Selenium adsorp-
tion on Mg–Al and Zn–Al layered double hydroxides. Appl.
Clay Sci. 20 (1–2), 13–25.
Zhang, P.C., Sparks, D.L., 1990. Kinetics and mechanisms of
sulfate adsorption/desorption on goethite using pressure-
jump relaxation. Soil Sci. Soc. Am. J. 54, 1266–1273.
K H. Goh, T T. Lim / Chemosphere 55 (2004) 849–859 859