Adsorption of Hg2+ onto Borassus Flabellifer: A redox mechanism

Chemical Engineering Journal 193–194 (2012) 328–338

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Adsorption of Hg2+ onto Borassus Flabellifer: A redox mechanism Shilpi Kushwaha a, B. Sreedhar b, Padmaja P. Sudhakar a,⇑ a b

Department of Chemistry, The M. S. University of Baroda, Vadodara 390 002, India Department of Inorganic & Physical Chemistry, Indian Institute of Chemical Technology, Hyderabad, India

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

" Results highlight controlled

adsorption of mercury on Palm shell powder (PSP). " Dramatic influence of phenolic, carboxyl and aldehydic groups in sorption. 2+ " Interestingly reduction of Hg to Hg0 by virgin agrowaste PSP.

a r t i c l e

i n f o

Article history: Received 24 January 2012 Received in revised form 17 April 2012 Accepted 17 April 2012 Available online 26 April 2012 Keywords: Redox mechanism Mercurous Mercuric Elemental mercury X-ray photoelectron spectroscopy X-ray diffraction Kinetics Isotherms

a b s t r a c t In this study palm shell powder (Borassus Flabellifer) has been used for mercury removal. The surface properties of palm shell powder were examined by potentiometric titrations, X-ray photoelectron spectroscopy (XPS), X-ray Diffraction and Fourier transform infrared (FTIR) spectroscopy and the possible functional groups available for mercury binding were found to be carboxyl, ether, alcoholic and amino functional groups. Interestingly it has been observed that mercury was present on PSP as Hg0, Hg+ and Hg2+. Kinetic, isotherm and column modeling studies reveal that complexation, ion exchange, and electrostatic interactions play a role in mercury adsorption on palm shell powder, but the relative predominance of each of these mechanisms varies with the pH of the medium. The isotherm thermodynamic parameters indicate the adsorption of mercury to be a spontaneous, exothermic process. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction Toxic heavy metals coming from many industrial effluents have adverse effects on our environment. Mercury has widespread applications in diverse fields like chloralkali industries, wood pulping industry, thermometer, barometers, batteries, dentistry, paints and military applications. Mercury is known to be highly toxic for aquatic life and human beings even at trace levels. ⇑ Corresponding author. Tel./fax: +91 265 2795552. E-mail address: [email protected] (P.P. Sudhakar). 1385-8947/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cej.2012.04.050

Adsorption is a cost effective technology for the decontamination of heavy metals. Several inexpensive adsorbents have been tested for the removal of Hg2+ like copper shavings [1], aspergillus versicolor [2], carica papaya [3], microorganisms [4], furfural adsorbents [5], waste tire rubber [6] and organic waste materials such as rice husk [7], coconut husks [8]. Chemically modified or impregnated adsorbents are much more expensive than their virgin precursors. Metal ion binding to lignocellulosic materials is known to occur through various functional groups like carboxyl, amines or phenolic groups. Research has to be thus focussed on the use of low cost adsorbents like virgin agro wastes with good adsorption capacity.

S. Kushwaha et al. / Chemical Engineering Journal 193–194 (2012) 328–338

As a local agrowaste we have used palm shell (Borassus Flabellifer) which is available throughout coastal Asia. In continuation of our work on use of palm shell powder as an adsorbent [9,10], in this study, emphasis is on the use of palm shell powder as cost effective adsorbent for the removal of mercury. The aim is to understand the mechanisms of mercury adsorption onto palm shell powder. Batch adsorption experiments were conducted under various solution pH values. Potentiometric titrations, Fourier transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS) spectra and kinetic, isotherm as well as column modeling was used to identify the adsorption mechanisms. The applicability of various isotherms and kinetic models were determined by their r2 value and error analysis. The results suggest that complexation, ion exchange, and electrostatic interactions may all plays role in mercury adsorption on palm shell powder. Furthermore the XPS spectra revealed the presence of Hg0 and Hg2+ on the surface of PSP after mercury adsorption suggesting that reduction of Hg2+ to Hg0 took place on PSP surface which is a novel finding in this study as there are no literature reports on reduction of Hg2+ using pristine lignocellulosic materials. 2. Materials and methods 2.1. Preparation of adsorbent material The shell of palm fruit (B. Flabellifer) was collected from coastal Andhra Pradesh, India. The collected palm shells were cut into small pieces, washed extensively with running tap water for 30– 40 min. followed by double distilled water and then dried in an air oven at 70 °C. The dried biomass was ground in a laboratory blender and sorted using standard test sieve of 40 mesh and termed as palm shell powder (PSP). Potentiometric titrations were carried out using 0.1 g adsorbent mass suspended in 50 mL of 0.1 M KNO3 solution, and the suspension was equilibrated for 24 h. The suspensions were acidified to pH 3.0 using 0.1 M HNO3 and then titrated to pH 11 using 0.1 M NaOH. All experiments were conducted in triplicate in a glass vessel with a lid as part of a Spectralab AT-38C Automatic potentiometric titrator. The temperature was recorded with a temperature sensor; the error of the temperature probe was 0.1 °C. The pH electrode was three-point calibrated with buffers (pH 4, 7 and 10) before each experiment, and the slope did not deviate more than 1% from the Nernst value. The titrator unit was programmed with a step volume dose mode for the titration, which adds 0.001 ml of titration solution according to the pH changes. 2.2. Batch uptake A stock solution of Hg2+ was prepared by dissolving 1.36 g of mercuric chloride E-Merck (Darmstadt, Germany) in slightly acidified double distilled water and making up to 1 L to give 1000 mg/L of Hg2+ solution. Working standards were prepared by diluting different volumes of the stock solution to obtain the desired concentration. Batch adsorption experiments were conducted at 30 °C by agitating 0.1 g of PSP with 25 mL of mercury ion solution of desired concentration maintained at pH 6.0 (except for pH experiments) in 100 mL stoppered conical flasks in a thermostated rotary mechanical shaker at 180 rpm for 4 h (except for the contact time experiments) at 30 °C. Experiments were done to determine the pH range at which the maximum mercury uptake would take place on PSP by varying the initial pH of the solution in the range 1–10 using 0.1 N NaOH and/or HCl. The effect of the initial concentration (1–1000 mg/L) was also studied in order to determine the effect of

329

the parameter on the adsorption of metal from the solution. The optimum equilibrium time was determined as the contact time required for the concentration of metal in the solution to reach equilibrium and was obtained by varying the contact time in the range 30–240 min. At the end of the predetermined time intervals, the suspensions were filtered and the mercury content in the filtrate was analyzed by using AAnalyst 200 Perkin–Elmer atomic absorption spectrophotometer. The uptake of Hg2+ by the adsorbent under study (qe) was calculated from the difference between the initial and final concentration as follows:

qe ¼ ðC i  C e Þ=m;

ð1Þ

Where, Ci – initial concentration of metal ion mg/L; Ce – Equilibrium concentration of metal ion mg/L; m – Mass of adsorbent g/L; qe – Amount of metal ion adsorbed per gram of PSP. Each experimental result was obtained by averaging the data from three parallel experiments. Adsorption isotherm experiments were also performed by agitating 0.1 g of PSP with a series of 25 mL solutions at pH 6.0, containing different initial concentrations of (1–1000 mg/L) at 30 °C. After the established contact time (4 h) was attained, the suspension was filtered, and supernatant was analyzed for the metal concentration. The adherence of the equilibrium isotherm and data obtained to different adsorption isotherms models as given in Table S1 was tested. Similarly the mercury adsorption data obtained after agitating solution containing 10 mg/L of mercury for various contact times with PSP at pH 2, 5 and 8 were calculated to determine the order of reaction rate and the adherence to different kinetic models as given in Table S1 was tested. Thermodynamic parameters of the adsorption process (DG0, DH0 and DS0) could be determined from the experimental data obtained at various temperatures. Values of correlation coefficients and standard deviation were used to compare the models. SD was calculated using the equation.

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u N u1 X SD ¼ t ðxi  xÞ2 N i¼1

ð2Þ

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u N X u 1 SSE ¼ t ðxi  xÞ2 NðN  1Þ i¼1

2.3. Desorption studies A 0.1 g of mercury loaded PSP was treated with 10 mL of different desorbing solutions like 0.1 M EDTA, 0.1 M NH3 and 0.1 M HCl for a period of 30 min in a thermostated rotary mechanical shaker. After 30 min the amount of mercury desorbed from PSP was determined by AAS. After complete desorption PSP was thoroughly washed with distill water 2–3 times, dried at 70 °C in an air oven. Further adsorption–desorption experiments were repeated by following the same process for two more cycles. 2.4. Column studies Column experiments were conducted in a glass column packed with PSP having an internal diameter of 1 cm and a bed height of 5 cm. Hg2+ solution of known concentration (1000 mg/L) at pH 6 was passed through the column of adsorbent at a flow rate of 1 mL/min. Samples from the column effluent were collected at regular intervals and analyzed by atomic absorption spectrometry. The break-through time has been chosen when the ratio of final to initial concentration of the effluent is 1 mg/L. Adherence of the

330

S. Kushwaha et al. / Chemical Engineering Journal 193–194 (2012) 328–338

surface is positively charged at pH values below the zpc, since the oxygen-containing groups are undissociated and amino groups are protonated. On the other hand, at solution pH values greater than the zpc, the adsorbent surface becomes more negative due to dissociation of weakly acidic oxygen-containing groups. Thus, the adsorbent surface is able to attract and exchange cations in solution. This renders the adsorbents capable of adsorbing mercury from aqueous solution at pH values 6.

column data to three different models (Thomas, Yoon and Nelson and Wolborska models) was studied using equations as described in Table S1. 2.5. FTIR analysis Spectra for adsorbent and the metal-loaded PSP was obtained using a Perkin–Elmer RX1 model within the wave number range of (400–4000) cm1. Specimens of samples were first mixed with KBr and then ground in a mortar at an appropriate ratio of 1/100 for the preparation of the pellets. The resulting mixture was pressed at 10 tons for 5 min. and sixteen scans with 8 cm1 resolution were applied in recording spectra. The background obtained from the scan of pure KBr was automatically subtracted from the sample spectra.

3.2. Uptake studies 3.2.1. pH dependence The pH of the medium from which adsorption is taking place is an important controlling parameter in heavy metal adsorption process. Its effect on the removal of mercury was studied by varying pH from 1 to 11. Fig. 2 shows that adsorption of mercury reaches maximum at pH 6.0 and is constant for the entire range of pH studied. This high adsorption is believed to be associated with the formation of positively charged metal hydroxy species having strong affinity for surface functional groups. The observed reduced adsorption at low pH value may be attributed to (i) higher hydrated species [Hg2+ (aq) species] having low mobility and (ii) protonation of the surface functional groups as well as the competition between mercuric ions and H+ or H3O+ ions present in the solution (9). At higher pH values, more functional groups are available for metal ion binding due to deprotonation, resulting in high adsorption. Thereby, maximum adsorption of mercury from pH 6.0 may be due to partial hydrolysis forming HgCl+, HgClOH+, and Hg(OH)2 species, having strong affinity for the negatively charged functional groups of PSP [12].

2.6. X-ray photoelectron spectroscopic analysis The surface of the samples was analyzed using a KRATOS AXIS 165 X-ray photoelectron spectrometer. Peak fitting and presentation output are produced by an integrated VISION control and information system. The deconvolution process of C 1s spectra as well as the elemental composition evaluation may result in an error of up to 5%. All spectra are presented charge balanced and energy referenced to C 1s at 284.6 eV. 3. Results and discussion 3.1. Characterization Surface characterization studies [12] of Palm shell powder (PSP) shows the presence of micro-pores or blocked pore cavities with 237.984 Å diameter. Interestingly PSP was found to have high enough iodine value of 147.59 mg/g and significant pore size in spite of its small BET surface areas. We can conclude that PSP contain a number of narrow micro-pores inaccessible (within reasonable time) to N2 at 77 K. The surface acidity of PSP was estimated using potentiometric titration experiments. The pKa distribution curve (Fig. 1) shows the presence of surface groups on PSP. Four different class of ligating functional groups might be present in PSP i.e. carboxyl (pKa 4.0, 0.0036 mol/g); lactone (pKa 5.4, 0.0006 mol/g); amine (pKa 9.0, 0.0001 mol/g); hydroxyl/phenol (pKa 11.0, 0.0021 mol/g); and hence concentration of total functional groups comes to 0.0064 mol/g [11]. The phenolic and carboxyl groups dominate in PSP. Fig. 1 shows the proton isotherms and zero point of charge (zpc) values of PSP. PSP was found to exhibit low zpc value of 3.7. The

3.2.2. Contact time variation Contact time variation shows that equilibrium is achieved was achieved within 150 min for PSP (Fig. 2). The rate of adsorption is very fast initially with about 95% of total mercury being removed within few minutes followed by a decreased rate with the approach of equilibrium. The removal rate is high initially due to the presence of free binding sites which gradually become saturated with time resulting in decreased rate of adsorption as equilibrium approaches. This indicates that the adsorption is mainly through surface binding. Similar observations were made by Das et al. [2]. 3.2.3. Amount of adsorbent variation The effect of dose of PSP on the removal of mercury is shown in Fig. 2, which illustrates the adsorption of mercury ion with change of PSP dose from 100 to 1000 mg. As inferred from Fig. 2, for a fixed metal initial concentration, increasing the PSP dose provided 0.004

0.0003

PSP

0.0000 0.003

LT (mol/g)

bj (mol/g)

-0.0003 -0.0006 -0.0009 -0.0012 PSP pHzpc=3.7

-0.0015

0.002

0.001

0.000

-0.0018 3

4

5

6

7 8 pH

9

10

11

12

4

5

6

Fig. 1. Potentiometric titrations and data analysis.

7

pKa

8

9

10

11

331

S. Kushwaha et al. / Chemical Engineering Journal 193–194 (2012) 328–338

2.5

(b)

2.5

(a)

2.4 2.0 2.3 1.5

2.2 2.1

1.0

2.0 0.5 1.9

PSP

qe (mg/g)

0.0

1.8 0

250

30

60

90 120 Time (min)

150

180

0

25

(c)

200

2

4

6 pH

8

10

12

(d)

20

150 100 50

15

2

10

1

5

0 0.00

0 0

200

400 600 800 Concentration (ppm)

1000

0.02

0.04

0.06 Dose (g)

0.08

0.10

Fig. 2. Effect of pH, time, dose and initial concentration of Hg2+.

6

7000

Kinetics

Isotherm

5250

5

3500 4

qe (mg/g)

qt (mg/g)

1750 3 2

8 6 4

1 2 0

0 0

50

qe (exp) Elovich

100

Time (min) PS2 LDM

150

200

PS1 IP Bangham

0

6

200

400

600

800

1000

Ce (mg/L) qe (exp) Freundlich Langmuir Temkin Halsey

Fig. 3. Kinetics and Isotherm studies of Hg2+ on PSP.

greater surface area and availability of more active sites, thus leading to the enhancement of metal ion uptake. Adsorption increased from 77% to 97% with increase in PSP dose.

3.3. Adsorption kinetics The kinetics of adsorption process describes the solute uptake, which, in turn governs the residence time of the adsorption reaction. Fig. 3 shows the adsorption kinetics conducted at pH 5 for Hg2+ removal by PSP. Adsorption of Hg2+ ions onto PSP was carried out for 180 min. to ensure attainment of equilibrium. The kinetic models of Pseudo First order, Pseudo Second order, Intraparticle diffusion, Bangham, Elovich and Liquid film diffusion models were

studied and the kinetic constants for the adsorption of mercury by PSP is presented in Table 1. The pseudo second order kinetics provided the best fit for the kinetic data. The qe values were very close to the experimental qe value and correlation coefficient values were 0.99–1.00 for PSP. This suggests that the rate limiting step in adsorption of mercury is chemisorption involving valence forces through the exchange of electrons between sorbent and sorbate, complexation, coordination and/or chelation [13]. In pseudo first order model the qe(exp) values were much higher than qe fitted values. The large discrepancies show that the reaction cannot be classified as first order although this plot has reasonably good correlation coefficient from the fitting process.

332

S. Kushwaha et al. / Chemical Engineering Journal 193–194 (2012) 328–338

causes a rapid increase in Hg2+ uptake. Later on slower adsorption might be due to intraparticle diffusion, and diffusion of mercury from the aqueous phase to PSP.

Table 1 Kinetics for different adsorbents. Pseudo 2nd order pH2 qe exp (mg g1) K2 (g mg min1) r2 SD SSE

PSP 2.178 0.029 0.999 0.094 0.009

Intraparticle diffusion Kip (mg g min0.5) r2 SD SSE

0.059 0.997 0.014 0.005

Elovich b (g mg1) a (mg g1 min1) r2 SD SSE

3.805 74.236 0.987 0.032 0.011

Pseudo 2nd order pH 5 qe exp (mg g1) K2 (g mg min1) r2 SD SSE

PSP 2.462 0.129 0.999 0.015 0.006

Intraparticle diffusion Kip (mg g min0.5) r2 SD SSE

0.024 0.918 0.034 0.014

Elovich b (g mg1) a (mg g1 min1) r2 SD SSE

8.917 4.15E+15 0.962 0.024 0.010

Pseudo 2nd order pH8 qe exp (mg g1) K2 (g mg min1) r2 SD SSE

PSP 2.315 0.163 0.999 0.027 0.009

Intraparticle diffusion Kip (mg g min0.5) r2 SD SSE

0.021 0.876 0.038 0.019

Elovich b (g mg1) a (mg g1 min1) r2 SD SSE

10.192 1.80E+17 0.929 0.028 0.017

The Elovich equation considers the rate-controlling step is the diffusion of the adsorbate molecules and describes chemisorptions on adsorbents and has been applied for the adsorption of solutes from a liquid solution. Regression coefficient of elovich model for adsorption of mercury onto PSP was found to be good enough (0.929) presenting that adsorption is mainly controlled by diffusion and chemisorptions. Adsorption capacity of PSP for mercury was less at low pH (90% in acidic pH implying that acidic groups are responsible for mercury binding at low pH, while a weak acidic group such as carboxyl group is dominant at neutral pH at 6. The initial rapid uptake of Hg2+ from solution may likely be due to binding of adsorbate ions on the surface of PSP through ion-exchange process. This instantaneous surface adsorption

3.4. Adsorption isotherms For modeling of mercury uptake Freundlich, Langmuir, Temkin, Dubinin–Radushkevich (DR), Flory–Huggins, Elovich and Halsey isotherm models were employed. Adsorption isotherms of the type qe versus Ce were also used to verify the isotherm models. Model fits for all the isotherms along with experimental data for adsorption of Hg2+ on PSP are presented in Fig. 3. The values of model constants along with their correlation coefficients, r2 and SD values for all the systems studied are presented in Table 2. Freundlich equation describes adsorption (possibly multilayer in nature) on a highly heterogeneous surface consisting of non identical and energetically non uniform sites. The value of n  1 for the Freundlich model indicates favorable adsorption. The Langmuir isotherm model is basically developed for gas-phase adsorption on homogeneous surfaces of glass and metals and predicts a single maximum binding capacity. The parameters KL (equilibrium sorption constant) and qmax were calculated from the intercept and slope of the plot of Ce/qe versus Ce. The correlation coefficient r2 was found to be 0.984 for the Freundlich model, while for Langmuir it was 0.880. In the present work although both the equations are obeyed Freundlich isotherm model has a better correlation coefficient for all the adsorbents under study indicating that the surface is heterogeneous in the long range but many have short range uniformity [15]. The correlation coefficient of Temkin Model (0.997) indicates a less satisfactory fit than freundlich model to the experimental data. The adsorption energy, DQ was found to be positive (2.816) for Hg on PSP, which indicated the adsorption process to be exothermic. The correlation coefficient r2 was found to be less (0.875) for Elovich model suggesting that it was not a good fit for the adsorption of mercury onto PSP. Thus it can be concluded that the Freundlich isotherm model fits best with low SD values followed by Langmuir, Halsey, DR and Temkin. The adherence to these models suggests that PSP might have homogeneous and heterogeneous surface energy distributions which would induce single and multilayer adsorption occurring either simultaneously or one after another. 3.5. Thermodynamic parameters The thermodynamic parameters of the sorption process could be determined from the experimental data obtained at various temperatures using the equations:

Kd ¼

qe Ce

ð3Þ

DS0 DH 0  R RT

ð4Þ

DG0 ¼ DH0  T DS0

ð5Þ

ln K d ¼

The values of DH0 and DS0 can be calculated from the slope and intercept of the plots of ln Kd against 1/T (Fig. 4). Temperature studies showed (Fig. 4) that uptake decreases as temperature increases indicating that the mechanism of adsorption is exothermic in nature. The negative value of DH0 indicates that the adsorption of Hg2+ on PSP is exothermic. Generally the absolute magnitude of the change in energy for physisorption is between (20 and 0 kJ/mol); chemisorptions has a range of (400 and 80 kJ/mol). The negative value of DG0 in Table 2 indicates that sorption of mercury by PSP is physisorption which is

S. Kushwaha et al. / Chemical Engineering Journal 193–194 (2012) 328–338 Table 2 Isotherms and thermodynamic parameters for Hg2+ on PSP for batch and column mode. qe

(exp)

(mg g1)

238.842

Freundlich KF (mg g1)(dm3/mg)1/n n r2 SD SSE

2.23 0.991 0.984 0.041 0.009

Langmuir KL (dm3 mg1) qm (mg g1) DG (kJ mol1) r2 SD SSE

0.062 104.493 7.221 0.880 0.015 0.004

DR qm (mg g1) E0 (kJ) r2 SD SSE

218.11 19.334 0.993 0.090 0.027

Elovich qm (mg g1) KE (dm3 mg1) r2 SD SSE Temperature (K) 303 313 323 333 343

Temkin DH (kJ mol1) KT (dm3 mg1) r2 SD SSE

2.816 0.301 0.997 0.037 0.019

Halsey KH (mg g1)(dm3/mg)1/n nH r2 SD

0.16 0.991 0.984 0.069

SSE

0.015

Flori-Huggins KFH (mg g1)(dm3/mg)1/n NFH r2 SD SSE

1.38E14 8.918 0.976 0.055 0.021

6.553 5.508E+258 0.875 0.092 0.024 Kd 5.351 4.260 3.667 2.51 2.139

DG (kJ/mol) 34.184 34.678 35.171 35.664 36.157

DS (kJ/mol K) 0.049 r2 = 0.991 Error = 0.013 SD = 0.056

333

AOH group. The metal loaded adsorbents showed a shift in the absorption frequencies of OH group indicating the probability of binding of mercury to the AOH groups present in PSP by complex formation. In PSP, the band at 1732 cm1 band is associated with C@O stretching mode particularly in carbonyls of aldehydes while 1250 cm1 band is associated with CAO stretching and OAH bending mode [16]. The two weak bands at 2959 and 2895 cm1 in infrared spectrum can be attributed to CAH stretching vibration of aldehydes. After mercury adsorption 1732 cm1 band splits into two 1758 and 1713 cm1 corresponding to the presence of carbonyl (ketone and carboxylic acid) stretching frequency, which represents the oxidation of substrate and binding of mercury. There is a shift in the frequencies of the absorption band of both C@O and CAO in carboxyl group (ACOOH) due to metal binding. The CAO band shifts to higher frequencies 1283 cm1 probably due to high electron density induced by the adsorption of Hg2+ on the adjacent carbonyl groups [17]. Thus carboxylic groups are involved in adsorption of mercury as seen by comparison of IR spectra of the adsorbents and the mercury loaded-PSP. The interaction between mercury and carboxylic group of PSP causes a diminution of the distance between C@O and CAO stretching peaks [18]. 3.7. X-ray photoelectron spectroscopy

DH (kJ/mol) 19.241

Thomas model KTH(dm3/(mg min)) q0 (mg g1) r2 SD SSE

4E06 199.444 0.952 0.079 0.21

Yoon & Nelson model KYN (min1) t0.5 (exp) (min) t0.5 (cal) (min) r2 SD SSE

3.60E03 900 997.222 0.952 0.039 0.007

Wolborska model b (min1) N0(mg dm3) r2 SD SSE

0.826 415.106 0.89 0.057 0.010

spontaneous and thermodynamically favorable. Also the DG0 values become less negative with increase in temperature suggesting that adsorption is favored at lower temperatures and hence is exothermic. The positive values of DS0 suggest increased randomness during adsorption and a high affinity of PSP towards the adsorbate. 3.6. Fourier transform infrared spectroscopy FTIR spectra of PSP suggested that a large proportion of the organic functional groups are present (Fig. 5). The absorption bands of FTIR spectra listed in Table S2 reveal the changes in absorption bands of the surface functional groups of PSP after mercury adsorption. Due to AOH stretching vibration a broad peak appeared in PSP (3361 cm1) indicating the presence of

X-ray photoelectron spectroscopy was employed to study the binding energy (BE) of oxygen (O 1s), carbon (C 1s), nitrogen (N 1S) in PSP and to study the shift in binding energy after mercury adsorption. The C 1s spectrum of PSP comprised of three peaks with BE of 284.661.8, 285.984, and 287.2 eV as seen in the deconvoluted spectra. These peaks can be assigned to CAH, ether or alcoholic, and carboxylate groups respectively, which represent the three different chemical environments of carbon atoms in PSP (Fig. 6). Among them, carboxylate and CAH carbons represent the abundant species. Two O 1s peaks were identified after deconvolution. The binding energy peaks of 531.4, 532.846 can be assigned to COO and CAOAR (alcohol and ether groups). However, a relatively stronger intense peak at a BE of 398.759 eV indicates that the N-atoms existed in a more reduced state on the surfaces of PSP due to mercury adsorption. This may be due to the formation of the covalent bond of NAHg in which, Hg shared electrons with the N-atom, which decreased the electron cloud density of the nitrogen atom. Fig. 6 displays the XPS deconvoluted spectra of mercury in the Hg 4f region and the peaks at 99.779 and 103.949 eV correspond to the characteristic binding energies (4f 7/2 and 4f 5/2) of elemental mercury [19]. On the other hand, the peaks at 100.303 and 104.878 eV correspond to the binding energies (4f 7/2 and 4f 5/2) of Hg2+ [20] suggesting that part of the mercury on PSP is present in elemental form (34%) and the rest in +2 state. This could also be due to intermediate oxidation state: Hg(I) or due to adsorbed Hg(II). The XPS peaks, Hg 4f of Hg(I) and Hg(II), are close, and therefore, it is difficult to differentiate between the two [21]. Table 3 and Table S3 show the changes occurring on the surface of PSP when Hg2+ is loaded onto it. An increase in the O/C ratio is observed which can be attributed to the oxidation of lignin aromatic carbons (aldehydic groups) by Hg2+ and formation of alcoholic sites, carboxylic sites and elemental mercury. Similar observation has been made by Dupont and Guillon during their studies on the removal of hexavalent chromium with a lignocellulosic substrate extracted from wheat bran [22] which could be related to the abundance of lignin and fatty acid moieties, which allow the reduction of Hg2+ into Hg0 on carboxylic moieties.

334

S. Kushwaha et al. / Chemical Engineering Journal 193–194 (2012) 328–338

PSP 1.8

260

1.6 1.4

240

1.0

LnKd

qe (mg/g)

1.2

220

0.8 0.6

200

0.4

180

0.2

160

-0.2

0.0

30

40

50

60

70

0.00294 0.00300 0.00306 0.00312 0.00318 0.00324 0.00330

Temperature ( 0C)

1/T (1/K)

Fig. 4. Temperature variation with respect to mg of Hg2+/g of PSP and thermodynamic plot.

Fig. 5. FTIR Spectra of pristine and mercury loaded PSP.

3.8. X-ray diffraction pattern However, XRD patterns showed neat presence of Hg+, Hg2+ and Hg as presented in Fig. 7. Moreover, reducing functional groups are reported of being capable for reduction of HgCl2 to Hg2Cl2 [23,24] to further elemental mercury and that this may further promote the binding of mercury at the surface and within pores of the matrix by colloidal precipitation. 0

3.9. Column The column breakthrough curves for mercury adsorption by PSP as shown in Fig. 8 (inset). The effluent concentration is seen to have the typical ‘S’ shape. A total of 2.1 L of 1000 mg/L metal ion solution was passed through the column containing 5 g of PSP. As seen from Fig. 8 (inset) the breakthrough for mercury is seen to take place at 249.9 bed volumes for PSP. Thomas and Yoon–Nelson models were also applied to the column adsorption data at a flow rate of 1 mL/min at an initial metal ion concentration of 1 g/L and bed height 5 cm with all the adsorbents under study. From the linear plots of ln[(Co/Ct)  1] versus Veff (Fig. S1) Thomas rate constant (kTh) and bed capacity (qTh) were calculated and are presented in Table 2.

The theoretical predictions based on the model parameters are compared in Fig. 8 with the observed data. Similarly, from the plot of sampling time (t) versus ln[Ce/(C0  Ce)], the Yoon and Nelson constant KYN and s (the time necessary to reach 50% of the retention) were calculated and are shown in Table 2. The well fit of the experimental data onto the Thomas and Yoon–Nelson model indicate that external and internal diffusion will not be the limiting step. From the equations in Table S1 it is evident that the characteristic parameter associated with Thomas and Yoon and Nelson models vary but both the models predict essentially same uptake capacity and C/C0 values for a particular experimental set of data. Hence same r2 and SD values were obtained as also suggested by Baral et al. [25]. 3.10. Desorption Solutions of 0.1 M HCl, 0.1 M EDTA and 0.1 M NH3 have been studied as eluents for desorption of Hg2+. From Fig. 9 it is evident that desorption of Hg2+ from the metal-loaded PSP with 0.1 M HCl resulted with 90% recovery. This indicates that ion exchange is involved in the adsorption process [26]. However, the use of 0.1 M EDTA and 0.1 M NH3

335

S. Kushwaha et al. / Chemical Engineering Journal 193–194 (2012) 328–338

Fig. 6. XPS spectra of PSP and Hg-loaded PSP.

resulted in 81.2% and 70.2% recovery of Hg2+ respectively. It was observed that mercury was easily desorbed within 30 min, which would prove highly advantageous for metal recovery. From Fig. 9 it is evident that the removal capacity of adsorbents shows insignificant changes in the second and third cycle. Thus regeneration and reuse of the adsorbents under study is an economical and efficient method for removal of Hg2+ from water. 3.11. Adsorption mechanism

carboxylic sites, CO2 and elemental mercury. The reducing property of PSP was also observed by us during our study on removal of Cr6+ [27]. 0  Hg2þ 2 þ 2e ! 2Hg

E0 ¼ 0:796 V

2Hg2þ þ 2e ! Hg2þ 2 Hg2þ þ 2e ! Hg0

E0 ¼ 0:911 V

E0 ¼ 0:85 V 

The XPS spectra confirmed the presence of Hg0 and Hg2+/Hg+ on the surface of PSP after mercury adsorption suggesting that reduction of Hg2+/Hg+ to Hg0 took place on PSP surface. This could be due to the oxidation of lignin aromatic carbons (aldehydic, phenolic and carboxyl groups) by Hg2+/Hg+ and formation of alcoholic sites,

HgCl2 þ 2e ! Hg0 þ 2Cl

E0 ¼ 0:410 V 

Hg2 Cl2 þ 2e ! 2Hg0 þ 2Cl 

2Hg2þ þ 2Cl þ 2e ! Hg2 Cl2

E0 ¼ 0:268 V E0 ¼ 1:440 V

336

S. Kushwaha et al. / Chemical Engineering Journal 193–194 (2012) 328–338

Table 3 Summary of binding energy and area ratios of carbons and Hg-loaded PSP. Sample surface C 1s Valence state PSP

Hg-PSP

O 1s Valence state PSP Hg-PSP N 1S PSP Hg-PSP Hg-PSP

Proposed components

Binding energy

Intensity (Counts/s)

Relative quantity

CA(C, H) graphitic C CA(O, N, H) phenolic, alcoholic, etheric C@O, OACAO, COOR -carbonyl or quinine CA(C, H) CA(OH, OR) O@CAO carboxyl or ester

284.661 285.984 287.200 284.642 286.569 288.442

2.310 1.774 2.495 2.504 2.196 2.624

0.6906 0.1871 0.1223 0.6639 0.2209 0.1152

C@O, CAO (Lactones, phenolic and etheric) Singly bonded oxygen CAO C@O, CAO (Lactones, phenolic and etheric) Singly bonded oxygen CAO

531.400 532.846 530.749 532.546

2.050 2.072 2.751 2.563

0.4424 0.5576 0.7579 0.2521

CANAC (pyrrolic nitrogen, pyridines) CANAC (pyrrolic nitrogen, pyridones) Hg-4f 7/2 (Hg0) Hg-4f 5/2 (Hg0) Hg-4f 7/2 (Hg+2) Hg-4f 5/2(Hg+2)

399.839, 398.597, 400.846 398.759 98.460 102.985 100.303 104.878

1.492, 1.512, 1.258 2.317 2.279 1.733 2.155 3.366

0.6129, 0.2141, 0.1729 1.000 0.2169 0.1359 0.4310 0.2170

* 0

* Hg

1+

Hg

2+

O

(a) 100

80

O

O

*

#

st

nd

1 2 3

(c)

% Desorption

rd st 1 2

nd

3

rd

60

* * *

# #

PSP 20

q e (mg/g)

O

#

10

(b)

% Desorption

NH3

Hg

EDTA HCl

#

P S P -H g

30

40

50

O

40

60

70

20

2 T h eta Fig. 7. XRD of metal loaded and pristine samples (# Hg0;

⁄

Hg1+;

o

Hg2+).

0 Eluents

Adsorption Cycle

Desorption Cycle

Fig. 9. Desorption and cycles of adsorption: (a) Effect of desorbents, (b) cycles of adsorption, (c) cycles of desorption.

1.0

0.8

 CH þ Hg2þ þ H2 O ! CAOH þ Hg0 þ Hþ

0.6 C/C0

1.0

PSP

 CH þ Hg2þ þ H2 O ! CAOH þ Hgþ þ Hþ

0.8 C/C0 (mg/L)

0.4

0.2

0.6

 CH þ Hgþ þ H2 O ! CAOH þ Hg0 þ Hþ

0.4

 CAOH=  CHO þ Hg2þ þ H2 O ! C@O=  COOH þ Hg0 þ H3 Oþ

0.2 0.0 0

50

100 150 200 250 300 350 400 450 Number of Bed Volumes

0.0 0 qe (exp)

500

1000 1500 Time (min) Thomas Yoon & Nelson

2000 Wolborska

Fig. 8. Column modeling study using Thomas, Yoon & Nelson, Wolborska model and compared with experimental qe values, Inset is the figure for breakthrough capacity.

 CAOH=  CHO þ Hg2þ þ H2 O ! C@O=  COOH þ Hgþ þ H3 Oþ  CAOH=  CHO þ Hgþ þ H2 O ! C@O=  COOH þ Hg0 þ H3 Oþ  COOH þ Hg2þ þ H2 O ! CO2 þ Hg0 þ H3 Oþ

337

S. Kushwaha et al. / Chemical Engineering Journal 193–194 (2012) 328–338 Table 4 Comparison of maximum adsorption capacities and cost of different agro-based adsorbents for metal-uptake. Sr. no.

Adsorbent

Hg+2 (mg g1)

Price (USD/kg)

References

1. 2. 3. 4. 5. 6. 7. 8. 9.

Treated rice husk Ptolemais lignite, raw (PSK-R) Demineralised (PSK-D) Walnut shell carbon (Carbon A) Walnut shell carbon (Carbon B) Sulfuric acid treated fruit shell of Indian almond (Terminalia catappa) carbon Chitosan ACC Palm shell powder (PSP)

9.32 209.1 176.7 151.5 100.9 94.43 51.55 65.00 104.49

– – – – – – 15.43 22.00 0.28

[36] [37] [37] [38] [38] [39] [40] [41] Present study

The elemental mercury could be adsorbed onto PSP by physisorption mechanism. However there is no leaching of elemental mercury from the adsorbents and thus the adsorbent containing elemental mercury would not prove hazardous. E.I. El-Shafey reported reduction of Hg2+ to Hg1+ using sulfuric acid treated rice husk while Cox et al. reported reduction of Hg2+ to Hg1+ and Hg0 using sulfuric acid treated flax shive [28,29]. However there have been no literature reports on reduction of Hg2+ using virgin lignocellulosic materials though there have been reports on reduction of Cu2+ to Cu0 and Cr6+ to Cr3+ [22,30]. Huang et al. have studied the adsorption of Hg2+ onto bayberry tanninimmobilised collagen fiber wherein they have reported the adsorption mechanism to be only chelation of Hg2+ [31]. Reduction of Hg2+ to Hg0 has been reported on organic matter present in soil, fungi and other microorganisms [32–34]. The FTIR and XPS spectra of PSP also showed the presence of amine, carboxyl, hydroxyl and phenolic groups. The role of the amine group on the sorption process may be considered in the light of the following reaction schemes: Hg2+ was considered to be adsorbed due to binding with amino groups by electrostatic interaction as follows:

ðANH2 Þ þ Hg2þ ! ðANHÞAHg þ 2Hþ In view of the possibility of stronger electrostatic attraction between the lone pairs of nitrogen atom and mercuric species than that of H+ ion, the binding of mercuric species to nitrogen may be considered to be stronger compared with that of hydrogen ion (H+). The sorption of mercury at low pH values will be decreased or increased according to the reaction schemes given. The carboxyl and hydroxyl groups become deprotonated as represented by equations below with concomitant increase in pH values resulting in net negative charge on the surface of PSP. Thus more electron rich binding sites are exposed and attract the metal ion to the surface of PSP [35].

2ðACOOHÞ þ Hg2þ ! ðACOOÞ2 AHg þ 2Hþ The binding mechanism of Hg2+ with phenolic groups could be:

2ðAC6 H5 AOHÞ þ Hg2þ ! ðAC6 H5 AOÞ2 AHg þ 2Hþ Thus it can be postulated that mercury cations can undergo ion exchange reactions as well as interact with oxygen and nitrogen containing groups on carbons in order of decreasing acidity, i.e. carboxyl > lactone > amine > hydroxyl/phenolic. 3.12. Cost estimation The cost-effective and economic removal of toxic heavy metals from wastewaters can be done only with the low-cost and easily available adsorbents. The precursor material used in the present study, palm shell, is available free of cost as a waste product. After consideration of the expenses for transportation, handling,

chemicals, electrical energy and the final cost would be approximately 0.28 USD/kg for PSP. The cheapest variety of commercially available activated carbon available in India is E-Merck Carbon (E-Merck India), which costs 20–22 USD/kg. The costs of different adsorbents available in literature are compared in Table 4 along with their adsorption capacities. This indicates the favorable use of palm shell powder as a low-cost adsorbent for mercury which is seen to have high adsorption capacity for mercury. 4. Conclusions Adsorption of metals is a complex process that is based upon a range of mechanisms which differ according to the type of adsorbent, the degree of processing it has undergone and also the adsorbate. The different mechanisms which play a role include ion exchange, chelation and physisorption. Potentiometric, FT-IR and XPS studies reveal that carboxyl, amino, hydroxyl, lactonic and phenolic groups on PSP seem to be responsible for mercury adsorption. XPS and XRD analyses also indicate that Hg2+ was reduced to Hg+/Hg0 on PSP surface and was adsorbed onto it by physisorption. Pseudo second order kinetics describes the overall sorption process well while intraparticle diffusion, diffusion of mercury from the liquid phase to the adsorbent surface might be having some role up to variable extents in deciding the rate processes. The sorption process is exothermic, spontaneous and accompanied by decrease in entropy. Prepared carbons show potential as effective systems for the removal of trace levels of mercury from aqueous systems. Sorption isotherms of Hg2+ on PSP were studied and modeled using Freundlich, Langmuir, Temkin, Dubinin–Radushkevich (DR), Flory–Huggins, and Halsey isotherms. The DG0 values from Langmuir and thermodynamic calculations indicate physisorption as the major mechanism for adsorption of mercury. The adherence to pseudo second order kinetic model and Halsey model indicate multilayer and ion exchange as the mode of adsorption thus justifying our hypothesis that a range of mechanisms are involved in the adsorption process to different extents. PSP thus shows potential as the most cost effective system for removal of mercury from aqueous solutions. Acknowledgment This work has been funded by the Board of research in Nuclear Sciences, INDIA. The authors thank The M. S. University of Baroda and Head Department of Chemistry, The M. S. University of Baroda, for laboratory facilities. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cej.2012.04.050.

338

S. Kushwaha et al. / Chemical Engineering Journal 193–194 (2012) 328–338

References [1] P. Huttenloch, E.R. Karl, K. Czurda, Use of copper shavings to remove mercury from contaminated groundwater or wastewater by amalgamation, Environ. Sci. Technol. 37 (2003) 4269. [2] S.K. Das, A.R. Das, A.K. Guha, A study on the adsorption mechanism of mercury on Aspergillus versicolor biomass, Environ. Sci. Technol. 41 (2007) 8281. [3] B. Shaik, Z.V.P. Murthy, B. Jha, Sorption of Hg(II) from aqueous solutions onto Carica papaya: application of isotherms, Ind. Eng. Chem. Res. 47 (2008) 980. [4] W.D. Irene, V.C. Harald, L. Ying, N.T. Kenneth, D.D. Wolf, Removal of mercury from chemical waste water by microorganisms in technical scale, Environ. Sci. Technol. 34 (2000) 4628. [5] T. Budinova, D. Savova, N. Petrov, M. Razvigorova, V. Minkova, N. Ciliz, E. Apak, E. Ekinci, Mercury adsorption by different modifications of furfural adsorbents, Ind. Eng. Chem. Res. 42 (2003) 2223. [6] A.S. Gunasekara, J.A. Donovan, B. Xing, Ground discarded tires remove naphthalene, toluene and mercury from water, Chemosphere 41 (2000) 1155. [7] N. Khalid, S. Ahmad, S.N. Kiani, J. Ahmed, Removal of mercury from aqueous solutions by adsorption to rice husks, Sep. Sci. Technol. 34 (1999) 3139. [8] M.K. Sreedhar, A. Madhukumar, T.S. Anirudhan, Evaluation of an adsorbent prepared by treating coconut husk with polysulphide for the removal of mercury from wastewater, Ind. J. Eng. Mater. Sci. 6 (1999) 279. [9] S. Kushwaha, S. Sodaye, P.P. Sudhakar, Adsorption of Hg(II) from aqueous solution onto Borassus Flabellifer: equilibrium and kinetic studies, Desalination Water Treat. 12 (2009) 100. [10] S. Kushwaha, G. Sreelatha, P. Padmaja, Preparation and characterization of activated carbons from Borassus Flabellifer, J. Porous Mat. (2012) 1–16, http:// dx.doi.org/10.1007/s10934-012-9571-4. [11] T.J. Bandosz, J. Jagiello, C. Contescu, J.A. Schwarz, Characterization of the surfaces of activated carbons in terms of their acidity constant distributions, Carbon 31 (1993) 1193. [12] C.M. Castilla, M.V. Lopez-Ramon, F. Carrasco-Marin, Changes in surface chemistry of activated carbons by wet oxidation, Carbon 38 (2000) 1995. [13] R.S. Vieira, M.M. Beppu, Interaction of natural and crosslinked chitosan membranes with Hg(II) ions, Colloids Surf. A. 279 (2006) 196. [14] J. Sarma, A. Sarma, K.G. Bhattacharyya, Biosorption of commercial dyes on Azadirachta indica leaf powder: a case study with a basic dye rhodamine B, Ind. Eng. Chem. Res. 47 (2008) 5433. [15] G.S. ElShafei, I.N. Nasr, A.S.M. Hassan, S.G.M. Mohammad, Kinetics and thermodynamics of adsorption of cadusafos on soils, J. Hazard. Mater. 172 (2009) 1608. [16] M.M. Figueira, B. Volesky, H.J. Mathieu, Instrumental analysis study of iron species biosorption by Sargassum biomass, Environ. Sci. Technol. 33 (1999) 1840. [17] E. Fourest, B. Volesky, Contribution of sulfonate groups and alginate to heavy metal biosorption by the dry biomass of Sargassum fluitans, Environ. Sci. Technol. 30 (1996) 277. [18] S. Fonglim, Y. Mingzheng, S. Wenzou, J.P. Chen, Characterization of copper adsorption onto an alginate encapsulated magnetic sorbent by a combined FTIR, XPS, and mathematical modeling study, Environ. Sci. Technol. 42 (2008) 2551. [19] X. Zeng, S. Prasad, S. Bruckenstein, X-ray photoelectron spectroscopy and timeof-flight secondary ion mass spectrometry study of Hg(I) and sulfate adsorption processes accompanying the coulostatic underpotential deposition of mercury on gold, Langmuir 14 (1998) 2535. [20] K.P. Lisha, S.M. Maliyekkal, T. Pradeep, Manganese dioxide nanowhiskers: a potential adsorbent for the removal of Hg(II) from water, Chem. Eng. J. 160 (2010) 432. [21] J. Wang, B. Deng, H. Chen, X. Wang, J. Zheng, Removal of Aqueous Hg(II) by polyaniline: sorption characteristics and mechanisms, Environ. Sci. Technol. 43 (2009) 5223.

[22] L. Dupont, E. Guillon, Removal of hexavalent chromium with a lignocellulosic substrate extracted from wheat bran, Environ. Sci. Technol. 37 (2003) 4235. [23] L. Gonzales, C. Moreno-Castilla, A. Guerrero-Ruiz, F. Rodriguez-Reinoso, Effect of carbon–oxygen and carbon sulphur surface complexes on the adsorption of mercury chloride in aqueous solutions by activated carbon, Chem. Tech. Biotechnol. 32 (1982) 575. [24] M.F. Emmanuel, A.G. Leyva, E.A. Gautier, M.I. Litter, Treatment of phenyl mercury salts by heterogeneous photocatalysis over TiO2, Chemosphere 69 (2007) 682. [25] S.S. Baral, N. Das, T.S. Ramulu, S.K. Sahoo, S.N. Das, G.R. Chaudhury, Removal of Cr(VI) by thermally activated weed Salviniacucullata in a fixed-bed column, J. Hazard. Mater. 161 (2009) 1427. [26] L. Xiaomin, Y. Tang, C. Xiuju, L. Dandan, L. Fang, W. Shao, Preparation and evaluation of orange peel cellulose adsorbents for effective removal of cadmium, zinc, cobalt and nickel, Colloids Surf. A: Physicochem. Eng. Aspects 317 (2008) 512. [27] S. Kushwaha, B. Sreedhar, P.P. Sudhakar, A spectroscopic study for understanding the speciation of cron palm shell based adsorbents and their application for the remediation of chrome plating effluents, Biores. Technol. doi:http://dx.doi.org/10.1016/j.biortech.2012.04.009. [28] E.I. El-Shafey, Removal of Zn(II) and Hg(II) from aqueous solution on a carbonaceous sorbent chemically prepared from rice husk, J. Hazard. Mat. 175 (2010) 319. [29] M. Cox, E.I. El-Shafey, A.A. Pichugin, Q. Appleton, Preparation and characterization of a carbon adsorbent from flax shive by dehydration with sulphuric acid, J. Chem. Technol. Biotechnol. 74 (1999) 1019. [30] J.G. Parsons, M. Hejazi, K.J. Tiemann, J. Henning, J.L. Gardea-Torresdey, An XPS study of the binding of copper(II), zinc(II), chromium(III) and chromium(VI) to hops biomass, Microchem. J. 71 (2002) 211. [31] X. Huang, X. Liao, B. Shi, Hg(II) removal from aqueous solution by bayberry tannin-immobilized collagen fiber, J. Hazard. Mat. 170 (2009) 1141. [32] N. Belzile, G.J. Wu, Y.W. Chen, V.D. Appanna, Detoxification of selenite and mercury by reduction and mutual protection in the assimilation of both elements by Pseudomonas fluorescens, Sci. the Total Environ. 367 (2006) 704. [33] K. Schluter, Review: evaporation of mercury from soils. An integration and synthesis of current knowledge, Environ. Geol. 39 (2000) 249. [34] N.D. Hutson, B.C. Attwood, K.G. Scheckel, XAS and XPS characterization of mercury binding on brominated activated carbon, Environ. Sci. Technol. 41 (2007) 1747. [35] E.L. Capel, G.D. Abbott, K.M. Thomas, D.A.C. Manning, Coupling of thermal analysis with quadrupole mass spectrometry and isotope ratio mass spectrometry for simultaneous determination of evolved gases and their carbon isotopic composition, J. Anal. Appl. Pyrolysis 75 (2006) 82. [36] Q. Feng, Q. Lin, F. Gong, S. Sugita, M. Shoya, Adsorption of lead and mercury by rice husk ash, J. Colloid and Interface Science 278 (2004) 1. [37] G. Skodras, I. Diamantopoulou, G.P. Sakellaropoulos, Role of activated carbon structural properties and surface chemistry in mercury adsorption, Desalination 210 (2007) 281. [38] M. Zabihia, A.H. Asla, A. Ahmadpour, Studies on adsorption of mercury from aqueous solution on activated carbons prepared from walnut shell, J. Hazard. Mater. 174 (2010) 251. [39] B.S. Inbaraj, N. Sulochana, Mercury adsorption on a carbon sorbent derived from fruit shell of Terminalia catappa, J. Hazard. Mater. 133 (2006) 283. [40] C.P. Huang, Y.C. Chung, M.R. Liou, Adsorption of Cu(II) and Ni(II) by pelletized biopolymer, J. Hazard. Mater. 45 (1996) 265. [41] J.R. Rangel-Mendez, M. Streat, Adsorption of cadmium by activated carbon cloth: influence of surface oxidation and solution pH, Water Res. 36 (2002) 1244.