You are on page 1of 11

RESEARCH ARTICLE

Modelling the adsorption of mercury onto natural


and aluminium pillared clays
Mabrouk Eloussaief & Ali Sdiri & Mourad Benzina
Received: 11 February 2012 / Accepted: 12 March 2012 / Published online: 25 April 2012
#Springer-Verlag 2012
Abstract
Introduction The removal of heavy metals by natural adsor-
bent has become one of the most attractive solutions for
environmental remediation. Natural clay collected from the
Late Cretaceous Aleg formation, Tunisia was used as a natural
adsorbent for the removal of Hg(II) in aqueous system.
Methods Physicochemical characterization of the adsorbent
was carried out with the aid of various techniques, including
chemical analysis, X-ray diffraction, Fourier transform in-
frared and scanning electron micrograph. Batch sorption
technique was selected as an appropriate technique in the
current study. Method parameters, including pH, temperature,
initial metal concentration and contact time, were varied in
order to quantitatively evaluate their effects on Hg(II) adsorp-
tion onto the original and pillared clay samples. Adsorption
kinetic was studied by fitting the experimental results to the
pseudo-first-order and pseudo-second-order kinetic models.
The adsorption data were also simulated with Langmuir,
Freundlich and Temkin isotherms.
Results Results showed that the natural clay samples are
mainly composed of silica, alumina, iron, calcium and mag-
nesiumoxides. The sorbents are mainly mesoporous materials
with specific surface area of <250 m
2
g
1
. Fromthe adsorption
of Hg(II) studies, experimental data demonstrated a high
degree of fitness to the pseudo-second-order kinetics with an
equilibration time of 240 min. The equilibriumdata showed the
best model fit to Langmuir model with the maximumadsorption
capacities of 9.70 and 49.75 mg g
1
for the original and alumin-
iumpillared clays, respectively. The maximumadsorption of Hg
(II) on the aluminium pillared clay was observed to occur at
pH 3.2. The calculated thermodynamic parameters (G, H
and S) showed an exothermic adsorption process. The entro-
py values varied between 60.77 and 117.59 Jmol
1
K
1
, and
those of enthalpy ranged from 16.31 to 30.77 kJ mol
1
. The
equilibrium parameter (R
L
) indicated that the adsorption of Hg
(II) on Tunisian smectitic clays was favourable under the exper-
imental conditions of this study.
Conclusion The clay of the Aleg formation, Tunisia was
found to be an efficient adsorbent for Hg(II) removal in
aqueous systems.
Keywords Pillared clay
.
Characterization
.
Mercury
adsorption
.
Thermodynamic
.
Modelling isotherm
1 Introduction
The development of industrial activities, including metal
plating, mining, metallurgy, fertilizer production, tanneries,
batteries manufacturing and textile dyeing etc., generated
high volume of wastewaters, especially in developing
countries (Srivastava and Majumder 2008). These waste-
waters have to be treated to remove toxic metals like zinc,
copper, nickel, mercury, cadmium, lead and chromium before
being discharged in the ecosystem. Mercury is one of the most
dangerous metals because of its serious effect on vital organs
(Rao et al. 1998). The toxicity of Hg depended strongly on its
oxidation state; the most reactive form of Hg (i.e. Hg
2+
) binds
Responsible editor: Vinod Kumar Gupta
M. Eloussaief (*)
:
M. Benzina
Laboratoire Eau, Energie et Environnement (LR3E), code:
AD-10-02, Universit de Sfax,
Sfax, Tunisia
e-mail: eloussaiefmabrouk@yahoo.fr
A. Sdiri
Faculty of Life and Environmental Sciences,
University of Tsukuba,
1-1-1 Tennodai,
Tsukuba 305-8572, Japan
e-mail: alisdiri@yahoo.fr
Environ Sci Pollut Res (2013) 20:469479
DOI 10.1007/s11356-012-0874-4
to the amino acid cysteine in cases of mercury poisoning
(Clarkson 1993). While inorganic mercury is the most preva-
lent form of mercury in aquatic systems, its biochemical
conversion to a more toxic methylmercury (i.e. organic form)
by microorganisms is feasible in water and soil (Stephen et al.
2009). Even low doses of mercury accumulation in various
organs (e.g. liver, kidneys, brain, spleen and bones) may cause
adverse effects like carcinogenic and mutagenic troubles,
serious intestinal and urinary complications and even death
in more severe cases (Bhakta et al. 2009). Mercury usually
enters food chain as methyl mercury through bacterial trans-
formation of a wide variety of food, especially fish. The
threshold limit for mercury (1 g L
1
) is the lowest among
all heavy metal ions (Bayramolu and Arica 2007). Therefore,
the removal of Hg(II) from wastewater is an important issue
that was extensively investigated (Sreedhar et al. 1999;
Mohan et al. 2001). Common techniques available for the
removal of mercury from solutions include chemical precipi-
tation (Sampaio et al. 2009), reverse osmosis (Patterson and
Fendorf 1997) and ion exchange (Gode and Pehlivan 2003).
The main shortcomings of these methods are their costs,
incomplete metal removal, high energy requirements and
generation of toxic sludge. Adsorption is the most preferred
method for removal of heavy metals from aqueous solutions
due to its simplicity and its high effectiveness (Gupta et al.
1998, 2006a, b; Gupta and Ali 2004; Shafaei et al. 2007; Sdiri
et al. 2011, 2012a, b). The applicability of adsorption in the
removal of heavy metals has been investigated by multiple
researchers (Gupta et al. 1997, 2001, 2007a, 2009, 2010;
Gupta and Ali 2008; Gupta and Rastogi 2008a, b; Gupta and
Nayak 2012). These studies confirmed the potential use of
adsorption as an appropriate technique for the removal of
various heavy metals. Moreover, the use of natural and cost-
effective adsorbents (i.e. clay and limestones) to reduce heavy
metal contamination is fundamentally important, especially
for developing countries (Mohan et al. 2001; Sdiri et al.
2011, 2012a). In this regard, this study has been undertaken
to develop a cheap adsorbent with large surface area and small
diffusion resistance. Pillared clays have emerged as potential
adsorbents for the removal of heavy metal ions from aqueous
solutions due to their large surface areas. In addition, the
preparation of pillared adsorbents with high number of active
surface sites would show high adsorptive capacities. Among
all these adsorbents, aluminium pillared clays (Al-PILCs)
have the required technical specifications and the potential
for use in environmental applications due to their physical and
chemical stability, large surface area and stable colloidal
suspension (Maira et al. 2001; Ni et al. 2007; Nabi et al.
2009). In this study, we prepared a new adsorbent (Al-
PILCs) for Hg(II) removal from aqueous solutions. Effects
of pH, temperature, contact time and initial concentration
of Hg(II) on the removal process were studied in detail.
The obtained results were fitted to Freundlich, Langmuir
and Temkin isotherms. Thermodynamic parameters were
also determined for the Hg(II) adsorption on natural and
aluminium pillared clays.
2 Experimental
2.1 Materials and methods
For the purpose of this study, a clay sample (RC) was collected
from the Late Cretaceous outcropping of the Aleg formation,
Jebel Semmama (Kasserine, Tunisia). The adsorbent was
washed with distilled water to remove soluble salts, sieved and
the desired fraction of less than 2 m was collected for subse-
quent analysis. Finally, the obtained clay was kept at 60C.
XRD patterns of the randomly oriented powders as well as
the separated clay fraction (2 m sized sub-sample) obtained
with an X-ray diffractometer (Philips PW 1710, Germany),
using CuK radiation (40 kV/40 mA) were recorded between
3245 with a counting time of 10 min. The powdered
rock and three oriented glass slides (untreated, glycolated and
heated to 520C) were used to identify clay mineral associa-
tions. The clay fraction (2 m) was separated by sedimenta-
tion and centrifugation (Sdiri et al. 2010).
Chemical composition of the collected clay sample was
performed by dissolving 1 g of dried clay in 50 mL of HNO
3
.
The mixture was evaporated to dryness; distilled water was
then added to remove the remaining traces of acid, and the
whole suspension was filtered with an ashless filter paper
(Whatman, England). Finally, the filtrate was analysed for
Hg(II) content, at a wavelength of 253.7 nm, by atomic
absorption spectrophotometer (ZEEnit 700, Analytik Jena,
Germany). The insoluble residue mainly composed of insol-
uble silica (SiO
2
) was estimated by gravimetric method (Ben-
zina 1990) after calcination to 900C for 1 h. The specific
surface area was determined by the adsorptiondesorption
isotherms of nitrogen (BETSurface Area Analyzer 2010,
ASAP, USA). The total pore volume was determined by
pycnometry. Infrared spectra were obtained using an FT-IR-
420 spectrophotometer (JASCO, Shimadzu Corp., Japan).
2.2 Preparation of aluminium pillared clay
The 2-m-sized clay sub-sample was dispersed in distilled
water and saturated with 1 M NaCl solution for three times
under continuous stirring. The obtained Na
+
homoionic clay
sample was washed for several times to remove chloride.
Pillaring procedure was performed according to Bergaya
(1995). The pillaring solution was prepared by titrating aque-
ous 0.5 M NaOH with aqueous 0.2 M AlCl
3
6 H
2
O until the
OH/Al ratio was equal to 2.4. At this hydrolysis ratio, Al
13
is
major species in solution. The pillaring solution was main-
tained for 3 h at 90C and then kept overnight at 30C. The
470 Environ Sci Pollut Res (2013) 20:469479
obtained solution was mixed with an appropriate volume of
the Na-exchanged clay suspension (1 % of clay) to prepare
20 mmol of aluminiumper gramof clay. The slurry was stirred
for 6 h at 30C, filtered, washed (to remove Cl

anions) and
calcined to 250C for 2 h and subsequently at 450C for 3 h.
Finally, the obtained product (i.e. Al-PILC clay) was pow-
dered, sieved and dried at 60C. Both original (RC) and
pillared (Al-PILC) clay samples were used for the removal
experiments.
2.3 Batch adsorption
A batch sorption method was used to study the adsorption of
mercury (Hg(II)) metals onto the Late Cretaceous clays of
Aleg formation (Ali and Gupta 2007). A stock solution of
1,000 mg L
1
Hg(II) was prepared by dissolving appropriate
amount of HgCl
2
(Aldrich Corp., Germany) in distilled
water. This solution was diluted to 550 mg L
1
for adsorp-
tion experiments. Fifty milligrams of clays was added to
50 mL of Hg(II) solution at initial pH 3.2. The suspension
was mixed on a thermostated shaker bath operating at 25C
and 200 rpm during 240 and 360 min for Al-PILC and RC
samples, respectively. It is noteworthy that the removal of
mercury by the pillared and the original clay sample attained
their equilibrium within 240 and 360 min, respectively.
Therefore, different equilibrations times (240 and 360 min)
were adopted. After the reaction, the suspension was centri-
fuged at 2,500 rpm for 20 min; the supernatant was with-
drawn and stored at 4C until analysis for Hg(II) by flame
atomic absorption spectroscopy at a wavelength of
253.7 nm. The removed amount was determined from the
difference between the initial and final concentrations. All
experiments were performed in triplicate.
Effects of pH, temperature, initial metal concentration and
contact time were also investigated. To study the effect of
contact time, a set of samples were prepared as described
above but then shaken for 30, 60, 120, 180, 240, 300, 360,
420, 480 and 540 min. The initial metal concentration,
20 mg L
1
, was the same in all of the contact time experiments,
as were the temperature (25C), clay amount (1 gL
1
) and
initial metal solution pH (3.2). To investigate the effect of pH,
the initial pH of the solution was adjusted to 39. The contact
time in the pH experiments was fixed at 240 and 360 min, and
the shaking rate was 200 rpm; metal concentrations were held
at 20 mg L
1
. Temperature effects were evaluated by conduct-
ing this method at 25, 35 and 45C, while maintaining the
same initial pH, contact time and metal concentration.
Adsorption process was quantified by calculating the
adsorption amount (milligrams per gram) as defined by the
following:
q
e

C
i
C
f
V
M
1
where q
e
is the amount of Hg
2+
ions adsorbed on the clay
(milligrams per gram), C
i
the initial Hg
2+
ions concentration in
solution (milligrams per litre), C
f
the final Hg
2+
ions concen-
tration in solution (milligrams per litre), V the volume of the
medium (litres) and M the amount of the clay used in the
reaction mixture (grams).
3 Results and discussions
3.1 Characterization of materials
Mineralogical analysis by X-ray diffraction indicated that
the raw clay was mainly composed of smectite (montmoril-
lonite) associated with kaolinite, illite, quartz, calcite and
feldspar. In the aluminium pillared sample, the untreated
oriented clay (2 m) showed a remarkable peak near
18.4 indicating an increase in d-spacing when compared
to the original clay (Fig. 1). This was expected due to the
expansion of the interlayer spacing in 2:1 clays after alu-
minium treatment (i.e. pillaring). The loss on ignition, de-
termined by calcination at 1,000C for 2 h, was attributed to
the evaporation of physically bound water near 100C, the
dehydroxylation of clay minerals near 500C and the de-
composition of calcite around 750C. The specific surface
area of the original sample, estimated to be 109 m
2
g
1
,
suggested that the studied smectitic clay may exhibit an
enhanced removal of heavy metals. Table 1 summarizes
the main physicochemical and mineralogical properties of
the studied clay materials. Our results indicated that the
studied clay samples were mainly composed of silica, alu-
minium and iron oxides. The collected clay sample
contained about 52.5 % of silica, 18.2 % alumina, but only
3 % iron oxides (Table 1). The original clay sample (RC)
contained higher amounts of iron and calcite than the alu-
minium pillared clay sample (Al-PILC). This was expected
since the treatment of the raw clay sample may cause the
dissolution of various constituents. The Al-PILC material
showed better textural proprieties when compared to the
Fig. 1 X-ray diffraction patterns of the untreated oriented clay fraction
(2 m) of the Al-PILC and the RC (Al-PILC aluminium pillared clay,
RC raw clay)
Environ Sci Pollut Res (2013) 20:469479 471
original sample (i.e. higher porosity and specific surface
area). This indicates that the exchange of the interlayer
cations (i.e. Ca
2+
and Mg
2+
) with Na
+
cations during the
saturation process enhanced the number of active sites due
to the difference in the atomic radii between the exchanged
cations. Figure 2 shows the FT-IR spectra of the RC and the
Al-PILC indicating the characteristic bands of surface hy-
droxyl group (OH) vibrating near 3,625 cm
1
. The addition-
al band near 1,100 cm
1
was attributed to the presence of
silica, while the stretching of C0O and MgOH groups
occurred at 720 and 500 cm
1
, respectively. After the alu-
minium pillaring, the spectrum of the Al-PILC changed
greatly. The broad band at 3,428 shifted to 3,625 cm
1
. This
band corresponded to the stretching vibration of the OH
groups of natural clay; the change indicates the introduction
of the OH groups onto the Al-PILC. The new band identi-
fied near 1,490 cm
1
was attributed to the stretching
vibration of the Al-OH groups (Fig. 2 b). These structural
modifications were attributed to the pillaring effects of clay
materials on their textural characteristics.
3.2 Effect of pH
The initial solution pH was adjusted to the interval 39 with
appropriate volume of solutions; all other parameters were
kept constant. The results are shown in Fig. 3. Adsorption
studies indicated that the solution pH was an important ad-
sorption parameter that strongly affects heavy metal removal
because of the decrease of positive charges on the adsorbent
surface with increasing the solution pH (Eloussaief et al.
2009). The adsorbed amounts of Hg(II) onto the pillared clay
(i.e. Al-PILC) decreased from 17 to 2 mg g
1
when pH of the
solution was increased from 3 to 9. The possible explanation
of the decreased removal capacity could be assigned to the
mercury species present at higher pH, as discussed later.
Positively charged species (Hg
2+
ions) was dominant in the
solution at pH<3.0, while Hg(OH)
2
became dominant at pH>
5. For higher pH (>8.0), the adsorption of Hg(II) became
slightly lower because of the formation of hydroxyl com-
plexes like [Hg(OH)]
+
or Hg(OH)
2
, as described by Eq. 2
(Ngah and Fatinathan 2010). The optimumof pHwas found at
3.2 values; therefore, subsequent adsorption experiments were
carried out at pH 3.2.
Hg
2
2OH

, Hg OH
2
2
3.3 Effect of contact time
Fifty milligrams of adsorbent was placed in a polypropylene
tube containing 50 mL of metal solution and then shaken for
30, 60, 120, 180, 240, 300, 360, 420, 480 and 540 min at the
optimum pH and temperature. The results show that adsorp-
tion process is clearly time dependent (Fig. 4). From this
figure, it was observed that more than 95 % of the total
Table 1 Chemical and physical characteristics of the raw and pillared
clays (% by weight)
Oxides (%) RC Al-PILC
SiO
2
52.501 57.301
Al
2
O
3
18.202 29.003
Fe
2
O
3
3.001 0.503
CaO 2.810 1.002
MgO 2.450 0.740
K
2
O 1.503 1.002
Na
2
O 1.780 0.060
LOI 16.001 9.003
V
p
(cm
3
g
1
) 0.183 0.421
Pore size (nm) 2.233 5.222
(%) 32.011 75.003
S
BET
(m
2
g
1
) 109.020 251.004
LOI loss on ignition, V
p
volume of pore, porosity, S
BET
BET specific
surface area
1490
3625
Fig. 2 FT-IR spectra of RC (a)
and Al-PILC (b)
472 Environ Sci Pollut Res (2013) 20:469479
adsorptive capacity occurred within 240 min for the Al-
PILC and 360 min for the RC sample, after what the remov-
al further increased but to a much lower extent. From
theoretical point of view, the adsorption process requires
long equilibration time while the practical approach needs
a short contact time (Sdiri et al. 2011). Based on the kinetic
results, an equilibration time of 240 and 360 min was
selected as compromises between theoretical and practical
approach. It was also observed that pillared clay sample
demonstrated a faster adsorption due to its enhanced specific
surface area (>250 m
2
g
1
) which provides more active sites
for Hg(II) adsorption, therefore allows easier diffusion of
metal cations. Several studies reported much lower equili-
bration time. For instance, Gupta and Rastogi (2008a) men-
tioned that raw biomass (i.e. Oedogonium sp.) needed much
shorter time to remove Cd(II) from aqueous solutions. The
difference in equilibration time could be attributed to the
fact that the presence of amino, carboxyl, hydroxyl and
carbonyl groups in algal biomasses enhanced their adsorp-
tion capacity despite the low specific surface area. Further-
more, the affinity of natural adsorbent to heavy metals
seemed to be lower than that of algal biomasses.
3.4 Effect of temperature
The effect of temperature on adsorption of Hg(II) ions onto
Al-PILC and RC was investigated by conducting this method
at 298, 308 and 318 K. The experimental results (Fig. 5)
showed that the Al-PILC adsorption capacity (q
e
) decreased
from 17 to 8.5 mg g
1
as temperature increased from 298 to
318 K at 20 mg L
1
. This slight change could be related to the
higher collision frequencies at higher temperature and there-
fore lower adsorption capacity. With increased temperature,
the adsorption of mercury ions decreased (Eloussaief et al.
2009; Gupta and Rastogi 2008b) confirming that the process
was exothermic. The observed change may be related to the
saturation of surface sites by the Hg(II) ions (nl and Ersoz
2006).
3.5 Effect of initial concentration
The effect of initial concentration of Hg(II) solution was
varied from 5 to 50 mg L
1
. The equilibration times were set
to 240 and 360 min for Al-PILC and RC, respectively. The
adsorbent concentration was fixed at 1 gL
1
, and a metal
solution of pH3.2 was used. The amount of mercury adsorbed
for different initial concentrations onto the samples is shown
in Fig. 5. Our results showed that the removal of mercury
increased with the increase of the initial Hg(II) concentration
due to the free reactive sites available for the adsorption of Hg
(II) ions until saturation. However, the adsorbed amount (q
e
)
was higher at high concentrations (Fig. 5) due to the larger
driving forces for mass transfer at higher concentrations.
These forces increased the loading capacity of the studied clay
samples (Mahvi et al. 2004). For instance, the initial concen-
tration, 20 mg L
1
, decreased substantially after treatment to
reach 3.29 and 15.15 mg L
1
for Al-PILC and RC samples,
respectively. Furthermore, the saturation of the surface adsorp-
tion sites was favoured due to the higher concentration of
Hg(II) ions (Herrero et al. 2005).
0
5
10
15
20
0 2 4 6 8 10
A
d
s
o
r
b
e
d


a
m
o
u
n
t
s


(
m
g

g
-
1
)
pH
Al-PILC
RC
Fig. 3 Effect of pH on the Hg(II) removal on Al-PILC and RC
0
5
10
15
20
0 100 200 300 400 500
A
d
s
o
r
b
e
d


a
m
o
u
n
t


(
m
g

g
-
1
)
Time (min)
Al-PILC
RC
Fig. 4 Effect of agitation time on the adsorption efficiency of Hg(II)
on Al-PILC and RC
0
5
10
15
20
0 10 20 30 40 50 60
A
d
s
o
r
b
e
d


a
m
o
u
n
t
s


(
m
g

g
-
1
)
Initial concentration (mg L
-1
)
25 C
35 C
45 C
25 C
35 C
45 C
Al-PILC
RC
Fig. 5 Effect of temperature on the adsorption efficiency of Hg(II) on
Al-PILC and RC
Environ Sci Pollut Res (2013) 20:469479 473
3.6 Adsorption isotherms
An isotherm describes the equilibrium relationship be-
tween the amount of the adsorbed metal ions and the
remaining concentration in the liquid phase. It gives im-
portant information about the main mechanisms involved
in the removal of heavy metal (i.e. Hg(II)) by natural and
aluminium pillared clay samples (Eloussaief et al. 2011;
Erdem et al. 2004; Naiya et al. 2009). In this study, the
isotherm data were analysed using the Langmuir, Freund-
lich and Temkin equations. Isotherm constants were cal-
culated by using Software CurveExpert 1.3.
3.6.1 Langmuir isotherm
Langmuir isotherm model is based on the assumption of
monolayer adsorption assuming that all surface sites are
energetically identical and surface itself is homogeneous
(Ngah et al. 2004). It also assumes that intermolecular forces
decrease rapidly with the distance from the adsorption sur-
face. Langmuir isotherm is expressed as:
q
e

K
L
q
m
C
e
1 K
L
C
e
3
The above equation can be rearranged to its linear form:
C
e
q
e

1
K
L

C
e
q
m
4
both q
m
and K
L
could be determined from the slope and
intercept of the linear plot C
e
/q
e
against C
e
, respectively. C
e
is the equilibrium concentration of Hg(II) ion (milligrams
per litre), q
e
is the adsorbed amount of Hg(II) (milligrams
per gram), q
m
(milligrams per gram) is the maximum ad-
sorption capacity and K
L
is the Langmuir constant related to
the adsorption energy, especially the adsorption enthalpy
(Jiang et al. 2009). Our experimental data were fitted to
the Langmuir isotherm to calculate the maximum adsorption
capacity of the studied clay samples. The results indicated
that higher removal efficiency was achieved by Al-PILC
sample (Fig. 6). The maximum adsorption capacity (q
m
)
was 49.75 mg g
1
and 9.70 mg g
1
for Al-PILC and RC,
respectively. This may indicate that the aluminium treatment
of the collected clay sample enhanced its textural properties
that would contribute to the higher adsorptive capacities.
Furthermore, the high coefficients of determinations (R
2
0
0.982 and 0.992) further confirmed that our experimental
data better fit to the Langmuir model. This is indicative of
the homogenous clay surface as supported by K
L
values (K
L
was 0.029 and 0.012 Lmg
1
for Al-PILC and RC,
respectively). The Langmuir isotherm can be expressed in
terms of the dimensionless constant (R
L
) defined as (Yong-
Mei et al. 2010):
R
L

1
1 K
L
C
0
5
where C
0
is the initial metal concentration. The value of R
L
indicates whether the adsorption process is favourable as
follows (Eloussaief and Benzina 2010):
R
L
>1 unfavourable adsorption
R
L
01 linear
0<R
L
<1 favourable
R
L
00 irreversible
Our results indicated that R
L
ranged between 0 and 1 for the
adsorption of mercury on Al-PILC and RC, indicating a
favourable adsorption (Table 2).
0
5
10
15
20
25
0 10 20 30 40 50 60
Equilibrium concentration Hg (II), (mg L
-1
)
A
m
o
u
n
t

a
d
s
o
r
b
e
d

(
m
g

g
-
1
)
Exp
Freundlich
Langmuir
Temkin
Fig. 6 Plot of lnK
d
versus 1/T for the Hg(II) adsorption on raw and
pillared clays (amount of adsorbent1 gL
1
, 25C, pH 3.2 and
200 rpm)
Table 2 Constants of isotherm models for Hg(II) adsorption onto RC
and Al-PILC at 298 K
RC Al-PILC
Langmuir isotherm k
L
(L mg
1
) 0.029 0.012
q
m
(mg g
1
) 9.701 49.752
R
L
0.040 0.051
R
2
0.992 0.982

2
2.425 22.195
Freundlich isotherm k
F
0.562 0.896
n 1.668 1.282
R
2
0.976 0.971

2
0.931 95.631
Temkin isotherm k
Te
(L mg
1
) 0.0301 0.208
b (J mol
1
) 1207.001 334.002
R
2
0.980 0.960

2
0.492 80.586
474 Environ Sci Pollut Res (2013) 20:469479
3.6.2 Freundlich isotherm
Freundlich isotherm describes multilayer adsorption on ener-
getically heterogeneous surfaces. It is an empirical equation
suitable for high and middle range of solute concentration, but
not for low concentrations. Freundlich isotherm is usually
described by the following equation (Hu et al. 2005):
q
e
K
F
C
1
n
e
6
This equation can be arranged in its linear form:
log q
e
log K
F

1
n
log C
e
7
where K
F
and n are Freundlich constants related to the adsorp-
tion capacity and intensity of adsorption, respectively. K
F
and n
were determined from the linear plot of logq
e
versus log C
e
.
Our data showed a good fitting to the Freundlich model
(Fig. 6). The high correlation coefficient (R
2
0.97) may con-
firm this hypothesis. In addition, the calculated n values were
1.28 and 1.68 for Al-PILCand RC, respectively. This indicates
favourable adsorption since the calculated adsorption intensity
ranged between 1 and 10 (1<n<10). The order of magnitude
of the calculated Freundlich constants (i.e. K
F
and n) followed
the same trends as was found for metal ions (e.g. Cu
2+
, Cd
2+
,
Cr
6+
and Pb
2+
) adsorption onto different oxides (Lin and Juang
2002).
3.6.3 Temkin isotherm
The derivation of the Temkin isotherm assumes that due to
adsorbate/adsorbent interaction, the heat of adsorption
decreases linearly rather than logarithmically, as implied in
the Freundlich equation (Temkin and Pyzhev 1940). The
Temkin isotherm is usually applicable for a heterogeneous
liquid and solid interface. It is given in the following:
q
e

RT
b
ln K
Te
C
e
8
q
e
A Bln C
e
9
A
RT
b
ln K
Te
10
where B0RT/b, R is gas constant (8.314 Jmol
1
K
1
), T is the
temperature (kelvin), K
Te
is equilibrium constant (litres per
gram), b is related to heat of adsorption (joules per mole). The
adsorption data can be analysed by plotting q
e
versus lnC
e
.
The calculated constants for Temkin isotherm are shown in
Fig. 6. The high correlation coefficient (R
2
00.96 and 0.98 for
Al-PILC and RC, respectively) showed that the adsorption of
Hg
2+
on Al-PILC and RC followed Temkin model.
3.7 Kinetic models
The well-known pseudo-first-order and pseudo-second-order
kinetic models were applied to the experimental data in order
to examine the adsorption process. The pseudo-first-order
equation can be written as follows (Mohan et al. 2001; Gupta
et al. 2006a, b, 2010; Gupta and Rastogi 2008a):
ln q
e
q
t
ln q
e
k
1
t 11
where q
e
and q
t
are the adsorbed quantity (milligrams per
gram) at equilibrium and at time t, respectively; k
1
(per min-
ute) is the adsorption rate constant. The value of k
1
and q
e
can
be calculated from the slope and intercept of the linear plot of
ln(q
e
q
t
) versus t, respectively.
The pseudo-second-order kinetic model was also used to
describe the sorption of metal ions (Gupta et al. 2007a, b).
The differential equation for the reaction is expressed as:
t
q
t

1
k
2
q
2
e

1
q
e
t 12
k
2
is the pseudo-second order rate constant (grams per milli-
gram minute). Both constants, k
2
and q
e
, were calculated from
the intercept and slope of the linear plot t/q
t
against t,
respectively.
The adsorption of Hg(II) onto (RC) and Al-PILC was
investigated kinetically using pseudo-first-order and
pseudo-second-order models. It was clearly observed that
the adsorption process does not follow the pseudo-first-
order model because of the big difference between experi-
mentally measured and calculated amount (Table 3).
The pseudo-second-order kinetic model was also used to
describe the sorption process. Both constants, k
2
and q
e
,
were calculated from the intercept and slope of the linear
plot t/q
t
against t. The calculated and measured amounts of
sorbed solute at equilibrium suggested that the process of
Hg(II) removal better fits the pseudo-second-order kinetic
model than the pseudo-first-order model (Table 3). More-
over, the coefficients of determination R
2
are higher than
those obtained from the pseudo-first-order model. Similar
results were found by Mohan et al. (2001) who studied the
kinetics of mercury adsorption from wastewater using acti-
vated carbon derived from fertilizer waste.
Table 3 Thermodynamic parameters of Hg(II) adsorption on RC and
Al-PILC
G (kJ mol
1
) H
(kJ mol
1
)
S
(J mol K
1
)
298 K 308 K 318 K
RC 0.39 0.30 0.20 30.77 117.59
Al-PILC 0.95 0.89 0.71 16.31 60.77
Environ Sci Pollut Res (2013) 20:469479 475
3.8 Model validation tests
To evaluate the fit of the above-mentioned theoretical mod-
els to the experimental data, linear coefficients of determi-
nation and non-linear chi-square test were used. The linear
coefficient of determination, R
2
, represents the percentage of
variability in the dependent variable that has been explained
by the regression line.
The chi-square test statistic was also used to calculate the
sum of the squares of the differences between the experimen-
tal data and those calculated from the model. The equivalent
mathematical statement could be described as following:
c
2

X
q
e
q
e;m

2
q
e;m
13
where q
e
and q
e
,
m
are the experimental and calculated the
equilibrium capacity (milligrams per gram), respectively. If
data from model are similar to the experimental data,
2
will
be a small number; if not,
2
will be a bigger number.
The fitness of the sorption data were further analysed by
calculating the sum of squared error (SSE) as given by the
following equation:
SSE
X
q
t;e
q
t;m

2
q
2
t;m
14
where q
t,e
and q
t
,
m
are the experimental and calculated ad-
sorption capacity (milligrams per gram) a time t, respectively.
In addition, the sum of squared error (SSE) test was also
done to support the best-fit adsorption model (Table 3). SSE
values were lower for pseudo-second-order model than the
pseudo-first-order model. Based on R
2
and SSE values, it
was confirmed that pseudo-second-order model best fits the
experimental data.
By comparing the results, the values of
2
and the cor-
relation coefficients (R
2
) (Table 3), it was found that the
Langmuir isotherm best represented the equilibrium adsorp-
tion of Hg(II) onto natural and aluminium pillared clays. It is
necessary to analyse the data set using the chi-square test to
confirm the best-fit isotherm for the adsorption system
(Meenakshi and Viswanathan 2007). If the data from the
model are similar to the experimental data,
2
will be a
small number, while if they differ,
2
will be a bigger
number. The
2
values of the isotherms are comparable
and hence the adsorption of Hg(II) follows in the order as:
Langmuir>Freundlich>Temkin isotherms.
4 Thermodynamic study
To evaluate the nature of adsorption of mercury onto natural
and aluminiumpillared clays, the removal process was analysed
in termof thermodynamic behaviour. To achieve this goal, three
thermodynamic parameters, including free energy (G), en-
thalpy (H) and entropy change (S), were determined by the
following equations (zcan et al. 2006; Akcay 2004).
K
d

q
e
C
e
15
G

RT ln K
d
16
ln K
d

S

R

H

RT
17
where K
d
is the distribution coefficient for the adsorption; S,
H and G are the changes of entropy, enthalpy and the
Gibbs energy; T (kelvin) is the temperature and R (8.314 J
mol
1
K
1
) is the gas constant. The values of H and S
were determined from the slopes and intercepts of the plots of
ln K
d
versus 1/T. The calculated thermodynamic parameters
indicated that H values were 16.31 and 30.77 kJ mol
1
for Al-PILCand RC, respectively (Table 4). The high values of
H indicated strong interactions between Hg(II) ions and the
studied clay samples. Similar finding was mentioned by
Eloussaief and Benzina (2010) when studying the efficiency
of natural and acid activated clay in the removal of Pb(II) from
aqueous solution by Tunisian clays. The negative values
obtained for G indicated the spontaneous nature of adsorp-
tion (Table 4). Moreover, the increase of G with temperature
indicated that adsorption was unfavourable at higher temper-
atures (Eloussaief and Benzina 2010). Exothermic adsorption
of Hg
2+
ion was also observed on china clay and lemna minor
powder (Shun-Xing et al. 2011). To follow the evolution of
H versus amount adsorbed of Hg(II) adsorption on studied
samples, the lnC
e
versus 1/T at different amounts (m08, 6, 4
and 2 mg g
1
) adsorbed was plotted for Al-PILC (Fig. 7a) and
Table 4 Kinetic models and (SSE) parameters for the Hg(II) adsorption onto natural and aluminium pillared clays
Measured q
e
(mg/g) Pseudo-first order Pseudo-second order
k
1
(1/min) q
e
(mg/g) R
2
SSE k
2
(g/mg min) q
e
(mg/g) R
2
SSE
RC 4.851 0.005 0.997 0.8911 7.394 0.003 5.185 0.9179 4.497
Al-PILC 16.520 0.008 0.985 0.7768 7.242 0.002 17.906 0.9854 0.110
476 Environ Sci Pollut Res (2013) 20:469479
(m02.5, 2, 1.5 and 1 mg g
1
) for RC (Fig. 7b). Then, H
versus amount adsorbed was illustrated in Fig. 7c, and it was
noted that the amount adsorbed increased with the increasing
of the released energy or H. During the adsorption of Hg(II)
onto the raw clay, the released energy was higher because of
the existence of more impurities that would prevent the diffu-
sion of the positively charged ion (i.e. Hg(II)). The adsorbed
amount by the original RC clay was much lower than that
recorded for Al-PILC clay.
5 Comparison to other studies
Based on previous relevant studies, the amount of heavy metals
removed by various clay materials is highly variable (Table 5).
In the current study, pillared clay sample (Al-PILC) demon-
strated an enhanced removal rate when compared with the
original form (RC). Sdiri et al. (2012a) mentioned that the
removal efficiency was dependent upon the physicochemical
characteristics of the clay and the metal removed, which was the
case of the present clay samples. The calculated Langmuir
capacity jumped from 9.70 mg/g for mercury removal by the
original RC sample to 49.75 mg/g, confirming the tight rela-
tionship between the physicochemical properties of the adsor-
bent and its removal capacity. These results indicate much
higher removal efficiency for the present clay samples than
was shown by Bhattacharyya and Gupta (2006), who reviewed
the removal of metals like lead, cadmiumand copper by various
kinds of clay (Table 5). However, the removal capacity occurred
to a much lower extent in comparison to the study of Mohan et
al. (2001). We found that aluminium pillared clay from western
Tunisia exhibited greater removal efficiency than those reported
in literature. Therefore, it is plausible to confirm that the Late
Cretaceous clays outcroppings of Jebel Semmama, Tunisia are
suitable for the removal Hg(II) from aqueous solutions.
Fig. 7 a Plot of lnC
e
versus 1/T for Hg(II) adsorption on Al-PILC in
different amount adsorbed (metres); b plot of lnC
e
versus 1/T for Hg(II)
adsorption on RC in different amount adsorbed (metres) and c H
versus amount adsorbed for Hg(II) adsorption on Al-PILC and RC
Table 5 Comparison of adsorption capacity with those of previous
removal studies with natural clays
Metal Sorbent q
max
(mg/g)
k
L
(L/mg) Literature
Mercury RC 9.701 0.029 Present study
Al-PILC 49.750 0.012
Activated
carbon
724.000 208.936 Mohan et al.
(2001)
Fe
2
O
4
nanoparticles
50.001 5.000 Gupta and
Nayak (2012)
Magnetic
nano-
adsorbent
71.430 4.670
Orange peel
powder
35.710 5.600
Lead Smectitic clay 50.761 0.552 Sdiri et al. (2011)
Illitic clay 25.441 0.041 Eloussaief and
Benzina (2010)
Kaolinite 11.520 20.701 Bhattacharyya and
Gupta (2006)
Cadmium Bentonite 9.272 22.702 Ulmanu et al.
(2003)
Kaolinite 6.781 32.300 Gupta and
Bhattacharyya
(2006)
Copper Illitic clay 17.983 0.212 Eloussaief et al.
(2009)
Kaolinite 4.303 19.901 Bhattacharyya and
Gupta (2008)
Environ Sci Pollut Res (2013) 20:469479 477
6 Conclusions
Clay deposits of the Late Cretaceous Aleg formation of
western Tunisia were found to be potential adsorbents for
Hg(II) adsorption from aqueous solutions. In this study, raw
and aluminium pillared clay samples achieved high removal
of Hg
2+
ions from aqueous solutions. Our study showed that
good adsorption could be achieved under the operating
conditions of pH 3.2, contact time (240 and 360 min for
Al-PILC and RC, respectively) and 1 gL
1
of clay powder
under the controlled temperature of 298 K. The experimen-
tal data demonstrated a high degree of fitness to Langmuir
model, indicating a monolayer adsorption and homogenous
surface. From thermodynamic point of view, it was conclud-
ed that the adsorption process was spontaneous and exother-
mic in nature. These results indicated that natural clays of
western Tunisia are promising adsorbents for Hg(II) remov-
al in aqueous systems. Further study is needed to evaluate
the adsorption behaviour of raw and pillared clays under
various conditions using other toxic heavy metals.
Acknowledgments We would like to thank Mr. Nidhal Baccar,
Technician in the Biotechnology Research Center of Sfax, for his help
in the analysis of our samples by atomic absorption spectrometer and
his assistance in the laboratory work. The authors would like to extend
their thanks to Professor Vinod Kumar Gupta and his team for their
prompt reviews and the constructive comments and suggestions.
References
Akcay M (2004) Characterization and determination of the thermody-
namic and kinetic properties of p-CP adsorption onto organophilic
bentonite from aqueous solution. J Colloid Interface Sci 208:299
304
Ali I, Gupta VK (2007) Advances in water treatment by adsorption
technology. Nat Protoc 1:26612667
Bayramolu G, Arica MY (2007) Kinetics of mercury ions removal from
synthetic aqueous solutions using by novel magnetic p(GMA-
MMA-EGDMA) beads. J Hazard Mater 144:449457
Benzina M (1990) Contribution ltude cintique et thermodynamique
de ladsorption des vapeurs organiques sur des argiles locales,
Modlisation dun adsorbeur lit fixe. Thse de Doctorat es-
science physique, Facult des Sciences de Tunis, p. 25
Bergaya F (1995) The meaning surface area and porosity measure-
ments of clays and pillared clays. J Porous Mater 2:9196
Bhakta JN, Salim M, Yamasaki K, Munekage Y (2009) Mercury
adsorption stoichiometry of ceramic and activated carbon from
aqueous phase under different pH and temperature. ARPN J Eng
Appl Sci 4:5259
Bhattacharyya KG, Gupta SS (2008) Adsorption of a few heavy metals
on natural and modified kaolinite and montmorillonite: a review.
Adv Colloid Interface Sci 140:114131
Clarkson TW (1993) Mercury: major issues in environmental health.
Environ Health Perspect 100:3138
Eloussaief M, Benzina M (2010) Efficiency of natural and acid-
activated clays in the removal of Pb(II) from aqueous solutions.
J Hazard Mater 178:753757
Eloussaief M, Jarraya I, Benzina M (2009) Adsorption of copper ions
on two clays from Tunisia: pH and temperature effects. Appl Clay
Sci 46:409413
Eloussaief M, Kallel N, Yaacoubi A, Benzina M (2011) Mineralogical
identification, spectroscopic characterization, and potential envi-
ronmental use of natural clay materials on chromate removal from
aqueous solutions. Chem Eng J 168:10241031
Erdem E, Karpinar N, Donat R (2004) The removal of heavy metal
cations by natural zeolites. J Colloid Interface Sci 280:309314
Gode F, Pehlivan E (2003) A comparative study of two chelating ion-
exchange resins for the removal of chromium(III) from aqueous
solution. J Hazard Mater 100:231243
Gupta VK, Ali I (2004) Removal of lead and chromium from wastewater
using bagasse fly asha sugar industry waste. J Colloid Interface
Sci 271:321328
Gupta VK, Ali I (2008) Removal of endosulfan and methoxychlor
from water on carbon slurry. Environ Sci Technol 42:766770
Gupta SS, Bhattacharyya KG (2006) Removal of Cd(II) from aqueous
solution by kaolinite, montmorillonite and their poly(oxo zirconium)
and tetrabutylammonium derivatives. J Hazard Mater 128:247257
Gupta VK, Nayak A (2012) Cadmium removal and recovery from
aqueous solutions by novel adsorbents prepared from orange peel
and Fe
2
O
3
nanoparticles. Chem Eng J 180:8190
Gupta VK, Rastogi A(2008a) Biosorption of lead fromaqueous solutions
by green algae Spirogyra species: kinetics and equilibriumstudies. J
Hazard Mater 152:407414
Gupta VK, Rastogi A (2008b) Equilibrium and kinetic modelling of
cadmium(II) biosorption by nonliving algal biomass Oedogonium
sp. from aqueous phase. J Hazard Mater 153:759766
Gupta VK, Srivastava SK, Mohan D, Sharma S (1997) Design param-
eters for fixed bed reactors of activated carbon developed from
fertilizer waste for the removal of some heavy metal ions. Waste
Manage 17:517522
Gupta VK, Mohan D, Sharma S (1998) Removal of lead from waste-
water using bagasse fly asha sugar industry waste material. Sep
Sci Technol 33:13311343
Gupta VK, Gupta M, Sharma S (2001) Process development for the
removal of lead and chromium from aqueous solutions using red
mudan aluminium industry waste. Water Res 35:11251134
Gupta VK, Mittal A, Gajbe V, Mittal J (2006a) Removal and recovery
of the hazardous azo dye acid orange 7 through adsorption over
waste materials: bottom ash and de-oiled soya. Ind Eng Chem Res
45:14461453
Gupta VK, Mittal A, Jain R, Mathur M, Sikarwar S (2006b) Adsorption
of Safranin-T from wastewater using waste materials activated
carbon and activated rice husks. J Colloid Interface Sci 303:8086
Gupta VK, Ali I, Saini VK (2007a) Adsorption studies on the removal
of Vertigo Blue 49 and Orange DNA13 from aqueous solutions
using carbon slurry developed from a waste material. J Colloid
Interface Sci 315:8793
Gupta VK, Ali I, Saini VK (2007b) Defluoridation of wastewaters
using waste carbon slurry. Water Res 41:33073316
Gupta VK, Carrott PJM, Ribeiro Carrott MML, Suhas (2009) Low-cost
adsorbents: growing approach to wastewater treatmenta review.
Crit Rev Environ Sci Technol 39:783842
Gupta VK, Rastogi A, Nayak A (2010) Adsorption studies on the
removal of hexavalent chromium from aqueous solution using a
low cost fertilizer industry waste material. J Colloid Interface Sci
342:135141
Herrero R, Lodeiro P, Rey-Castro C, Vilarin T, Manuel ESV (2005)
Removal of inorganic mercury from aqueous solutions by biomass
of the marine macroalga Cystoseira baccata. Wat Res 39:3199
3210
Hu J, Chen G, Lo IMC (2005) Cr(VI) on megahemite removal and
recovery of Cr (VI) from wastewater by megahemite nanoparticles.
Wat Res 39:45284536
478 Environ Sci Pollut Res (2013) 20:469479
Jiang Y, Pang H, Liao B (2009) Removal of copper(II) ions from
aqueous solution by modified bagasse. J Hazard Mater 164:19
Lin S, Juang R (2002) Heavy metal removal from water by sorption
using surfactant-modified montmorillonite. J Hazard Mater
92:315326
Mahvi A, Maleki H, Eslami A (2004) A potential of rice husk and rice
husk ash for phenol removal in aqueous system. American Appl
Sci 1:321326
Maira AJ, Yeung KL, Soria J, Coronado JM, Belver C, Lee CY,
Augugliaro V (2001) Gas-phase photo-oxidation of toluene using
nanometer-size TiO
2
catalysts. Appl Catal B-Environ 29:327336
Meenakshi S, Viswanathan N (2007) Identification of selective ion-
exchange resin for fluoride sorption. J Colloid Interface Sci
308:438450
Mohan D, Gupta VK, Srivastava SK, Chander S (2001) Kinetics of
mercury adsorption from wastewater using activated carbon de-
rived from fertilizer waste. Colloids Surf A 177:169181
Nabi D, Aslam I, Qazi IA (2009) Evaluation of the adsorption potential
of titanium dioxide nanoparticles for arsenic removal. J Environ
Sci 21:402408
Naiya TK, Bhattacharya AK, Das SK (2009) Adsorption of Cd(II) and
Pb(II) from aqueous solutions on activated alumina. J Colloid
Interface Sci 333:1426
Ngah WSW, Fatinathan S (2010) Adsorption characterization of Pb(II)
and Cu(II) ions onto chitosantripolyphosphate beads: kinetic,
equilibrium and thermodynamic studies. J Environ Manage
91:958969
Ngah WSW, Kamari A, Koay YJ (2004) Equilibrium and kinetics
studies of adsorption of copper(II) on chitosan and chitosan/
PVA beads. Int J Biol Macromol 34:155161
Ni M, Leung MKH, Leung DYC, Sumathy K (2007) A review and
recent developments in photocatalytic water-splitting using TiO
2
for hydrogen production. Renew Sustain Energy Rev 11:401425
zcan A, nc EM, zcan AS (2006) Kinetics, isotherm and thermo-
dynamic studies of adsorption of Acid Blue 193 from aqueous
solutions onto natural sepiolite. Colloids Surf A 277:9097
Patterson RR, Fendorf S (1997) Reduction of hexavalent chromium by
amorphous iron surface. Environ Sci Technol 31:20392044
Rao MV, Patil GR, Borole LV (1998) Effect of mercury and selenite
interaction on the mouse vital organs. J Environ Biol 19:215220
Sampaio RMM, Timmers RA, Xu Y, Keesman KJ, Lens PNL (2009)
Selective precipitation of Cu from Zn in a pS controlled contin-
uously stirred tank reactor. J Hazard Mater 165:256265
Sdiri A, Higashi T, Hatta T, Jamoussi F, Norio T (2010) Mineralogical
and spectroscopic characterization, and potential environmental
use of limestone from the Abiod formation, Tunisia. Environ
Earth Sci 61:12751287
Sdiri A, Higashi T, Hatta T, Jamoussi F, Norio T (2011) Evaluating the
adsorptive capacity of calcareous and montmorillonitic clays on
the removal of several heavy metals in aqueous systems. Chem
Eng J 172:3746
Sdiri A, Higashi T, Chaabouni R, Jamoussi F (2012a) Competitive
removal of heavy metals from aqueous solutions by montmoril-
lonitic and calcareous clays. Water Air Soil Pollut 223:11911204
Sdiri A, Higashi T, Jamoussi F, Bouaziz S (2012b) Effects of impurities
on the removal of heavy metals by natural limestones in aqueous
systems. J Environ Manage 93:245253
Shafaei A, Ashtiani FZ, Kaghazchi T (2007) Equilibrium studies
of the sorption of Hg(II) ions onto chitosan. Chem Eng J
133:311316
Shun-Xing L, Feng-Ying Z, Huang Y, Jian-Cong N (2011) Thorough
removal of inorganic and organic mercury from aqueous solutions
by adsorption on Lemna minor powder. J Hazard Mater 186:423429
Sreedhar MK, Madhukumar A, Anirudhan TS (1999) Evaluation of an
adsorbent prepared by treating coconut husk with polysulphide
for the removal of mercury from wastewater. Ind J Eng Mater Sci
6:279285
Srivastava NK, Majumder CB (2008) Novel biofiltration methods for
the treatment of heavy metals from industrial wastewater. J Haz-
ard Mater 151:18
Stephen IB, Wang JS, Lu JF, Siao FY, Chen BH (2009) Adsorption of
toxic mercury(II) by an extracellular biopolymer poly(-glutamic
acid). J Biores Technol 100:2002007
Temkin MI, Pyzhev V (1940) Kinetics of ammonia synthesis on
promoted iron catalysts. Acta Physiochim URSS 12:327
356
Ulmanu M, Maran E, Fernndez Y, Castrilln L, Anger I, Dumitriu
D (2003) Removal of copper and cadmium ions from diluted
aqueous solutions by low cost and waste material adsorbents.
Water Air Soil Pollut 142:357373
nl N, Ersoz M (2006) Adsorption characteristics of heavy metal ions
onto a low cost biopolymeric sorbent from aqueous solutions. J
Hazard Mater 136:272280
Yong-Mei H, Man C, Zhong-Bo H (2010) Effective removal of Cu(II)
ions from aqueous solution by amino-functionalized magnetic
nanoparticles. J Hazard Mater 184(392):399
Environ Sci Pollut Res (2013) 20:469479 479