Desalination 282 (2011) 39–45

Contents lists available at ScienceDirect

Desalination

j ourna l homepage: www.e lsev ie r.com/ locate /desa l

Functionalized graphene sheets for arsenic removal and desalination of sea water

Ashish Kumar Mishra, S. Ramaprabhu ⁎Alternative Energy and Nanotechnology Laboratory (AENL), Nano Functional Materials Technology Centre (NFMTC), Department of Physics, Indian Institute of Technology Madras,Chennai, 600036, India

⁎ Corresponding author. Tel.: +91 44 22574862; fax:E-mail address: ramp@iitm.ac.in (S. Ramaprabhu).

0011-9164/$ – see front matter © 2011 Elsevier B.V. Aldoi:10.1016/j.desal.2011.01.038

a b s t r a c t

a r t i c l e i n f o

Article history:Received 17 October 2010Received in revised form 12 January 2011Accepted 13 January 2011Available online 11 February 2011

Keywords:Functionalized graphene sheetsWater purificationSodium removalArsenic removalSupercapacitor

Water pollution is a major problem in the global context and main cause of some diseases especially in India.Graphene has fascinated the scientific community by its different novel properties for various applications. Inthe present work, we have synthesized the graphene sheets by hydrogen induced exfoliation of graphiticoxide followed by functionalization. Graphene sheets were characterized by electron microscopy, X-raydiffraction, infra-red and Raman spectroscopy techniques. These functionalized graphene sheets were usedfor simultaneous removal of high concentration of inorganic species of arsenic (both trivalent andpentavalent) and sodium from aqueous solution using supercapacitor based water filter. In addition, thesefunctionalized graphene sheets based water filter was used for desalination of sea water. Adsorptionisotherms and kinetics were studied for the simultaneous removal of sodium and arsenic. Maximumadsorption capacities, using Langmuir isotherm, for arsenate, arsenite and sodium were found to be nearly142, 139 and 122 mg/g, respectively. High adsorption capacity for both inorganic species of arsenic andsodium along with desalination ability of graphene based supercapacitor provides a solution for commerciallyfeasible water filter.

+91 44 22570509.

l rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Industrialization and urbanization have resulted in the dischargeof a number of toxic chemicals of anthropogenic origin in naturalsurface water bodies. The presence of toxic contaminants in thesource water influences the efficiency of conventional water treat-ment plants. Advanced membrane filtration techniques (like micro-filtration, ultrafiltration, nanofiltration and reverse osmosis) exhibitsuperior performance in treatment and removal of chemical andbiological contaminants over conventional systems. Still uses of thesemethods are limited owing to their high operating cost. Among all theapproaches proposed for toxicmetal impurities removal, adsorption isfound to be economical and efficient [1–3].

Among toxic metals, a check on arsenic contamination of naturalwater sources has been recorded by World Health Organization as afirst priority issue. High concentrations of arsenic are found in groundwater in many regions around the world like more than 60% ofexisting wells in Bangladesh andWest Bengal, India due to the releaseof arsenic from arsenic-bearing sediments. Industrial waste watercontains veryhigh level of arsenic (around 1000 mg/L),which is amajorconcern, especially in the developing countries. Being a carcinogenicelement, high level of arsenic in groundwater poses threat to bothwildlife and humans. Arsenic exists in the natural environment mainly

in the forms of arsenite [As (III)] and arsenate [As (V)]. Arsenite is moremobile and toxic than arsenate. Adsorption on metal oxides (ironoxides, activated alumina),mixedmetal oxides and resinwere reportedmainly for pentavalent arsenic removal [4–10]. Activated carbon basedsorbents have been tested for arsenic removal but have a low adsorbentcapacity [11,12].

Sea water is the major source of water on Earth. High concentra-tion of salinity in the sea water makes it unfit for domestic purposes. Anumber of desalination technologies like thermal distillation,membrane separation, freezing and electro-dialysis have beendeveloped but could not be made commercially feasible due to highcosts [13–16]. Hence, commercially feasible technology is in demandfor desalinating sea water with low energy loss. Carbon basednanomaterials (activated carbon, carbon nanotubes) have attractedthe attention of scientific community due to their novel propertieslike large surface area, long range of porosity, good thermal stabilityand good mechanical strength. These properties encourage the use ofcarbon based as nanomembranes for water filtration. Nanopores ofcarbon nanotubes, modified with functional groups, act as adsorptionsites for metallic impurities [17–19]. In addition, the crystallinestructure of carbon nanotubes provide high transport rate for water.Hence, carbon nanotubes have emerged as an excellent membrane forwater filtration. In spite of these advantages, high cost of carbonnanotubes limits their commercial use. Graphene, as new class ofcarbon nanomaterials, is found to be economical and has novelproperties similar to carbon nanotubes. Use of graphene in differentapplications like supercapacitor, fuel cells, and photovoltaics is of

40 A.K. Mishra, S. Ramaprabhu / Desalination 282 (2011) 39–45

immense interest to the current scientific research [20,21]. However,the effective use and experimental demonstration of pure graphenefor metal removal and desalination of sea water still remain achallenge.

Recently, we have demonstrated the use of multiwalled carbonnanotubes based supercapacitor for arsenic removal as well desali-nation of sea water [22,23]. Only one volt dc power supply is requiredfor supercapacitor based water filter, which makes this techniqueadvantageous over other techniques used earlier for desalination ofsea water. In the present work, we have demonstrated the use ofgraphene sheets for simultaneous removal of sodium and arsenicfrom aqueous solution and desalination of sea water. Additionally,adsorption and kinetic studies for sodium and both type of inorganicarsenic species were performed. Cost effective technique for thepreparation of graphene sheets, used in the present work, furtherreduces the cost of supercapacitor based water filter and hence it is acommercially feasible solution for large scale arsenic removal fromindustrial waste water and desalination of sea water.

2. Materials and methods

2.1. Preparation of functionalized graphene sheets

To prepare graphene sheets, graphitic oxide was prepared first byoxidation of pure graphite using Hummers' method. In this method,the oxidation of graphite to graphitic oxide is accomplished bytreating graphite with essentially a water-free mixture of concen-trated sulfuric acid, sodium nitrate and potassium permanganate[24]. This graphitic oxide was further thermally exfoliated at 200 °Cunder hydrogen atmosphere. This thermal shock under the presenceof hydrogen gas leads to formation of graphene sheets [25]. Thesehydrogen exfoliated graphene sheets (HEG) were further treatedwith conc. HNO3, which introduces hydrophilic functional groups(\COOH,\C_O, and\OH) at the surface of HEG. The functionalizedgraphene sheets (f-HEG) were further washed several timeswith water to achieve pH=7 followed by drying. Hydrophilic natureof f-HEG provides better contact between water and graphene basedelectrodes [23].

2.2. Preparation of electrode and electrolyte

The carbon fabric was supplied by SGL, Germany. Electrodes wereprepared by using dispersed solution of f-HEG in ethanol and fewdrops (~20 μL) of nafion. Gel solution was coated on carbon fabric byusing spray coating technique. Further these f-HEG coated carbonfabrics were hot pressed at 50 °C under 1 t force for 15 min to providegood mechanical strength to the electrodes. Each electrode contains50 mg of f-HEG. These electrodes were used in supercapacitor basedwater filter for isotherm and kinetic study of sodium and arsenic aswell as for cyclic repeatability experiment. Glassy carbon electrode(GCE) was modified with f-HEG for cyclic voltammetry (CV) analysis.Functionalized graphene sheets (5 mg) were sonicated in 0.2 ml ofethanol for 15 min followed by the addition of 5 μl of nafion in eachsolution. This gel solution was again sonicated for 10 min and 2 μl ofthe gel solutionwas deposited on GCE and dried at room temperature.

Sodium arsenate and sodium arsenite (both from Across Organics)were selected as the solutes for studying the two types of arsenic ionimpurities. Solutions containing two different arsenic ions separatelyand sea water were taken as electrolytes for supercapacitor basedwater filter. Sea water sample was collected from Bay of Bengal shoresat Chennai, India. Sodium arsenate and sodium arsenite containingwaters with initial concentration of 300 mg/L for arsenic were usedfor kinetic studies. Isotherm studies were performed with differentinitial concentrations (50–300 mg/L) of arsenic with both types ofarsenic ions. Volumes of 65 ml for sodium arsenate and sodiumarsenite containing water were used to study isotherm and kinetic

characteristics of f-HEG for simultaneous removal of arsenic andsodium. To study the cyclic repeatability of electrodes, 100 ml ofarsenic containing water (sodium arsenate and sodium arseniteseparately) was used. Desalination performance of this supercapacitorbased water filter was also tested with 100 ml of sea water.

2.3. Fabrication of the apparatus

Cylindrical perspex of length 2 cm, width 0.5 cm and diameter4 cm was used for water collection between the electrodes. Carbonfabric supported f-HEG based electrodes were fixed at both the ends.Stainless steel plates were used as current collector and graphiteplates were used to provide conducting support to the electrodes. Setup used for the present study is similar as reported in our earlier work[22]. DC regulated power supply was used to apply 1 V across theelectrodes. For treating sea water as well as water containing arsenicimpurities, the amount of f-HEG was maintained at 50 mg at eachelectrode.

2.4. Adsorption experiment

Supercapacitor based water filter was filled with 65 ml of arseniccontaining solutions and 1 V DC voltage was applied across thesupercapacitor electrodes during the experiment. Kinetic study wasperformed with two different arsenic solutions (arsenate and arseniteseparately) with the initial concentration of 300 ppm for arsenic. Inthe kinetic study, a small amount of water (5 ml) was collected atparticular intervals. The kinetic study shows that at around 40 min,adsorption curve starts to saturate and hence the isotherm study andremoval performancewas studied for treating each sample for 40 min.For isotherm study initial concentrations varying from 50–300 mg/Lfor arsenic were tested. Samples were collected in the end of theexperiment and concentrations of different metals were measuredusing ICP-OES technique.

2.5. Characterization

Morphology studies of HEG were characterized by QUANTA 3D FEG(FEI) scanning electron microscope (FESEM) and JEOL 3010 Highresolution transmission electron microscope (HRTEM). BET measure-ments were performed by Micromeritics ASAP 2020 analyzer. Ramananalysis was performed by using HORIBA JOBIN YVON HR800UVConfocal Raman Spectrometer, while Fourier transform infraredspectroscopy (FTIR) study was performed using PERKIN ELMERSpectrum One FT-IR spectrometer. In order to investigate the electro-chemical activity of HEG, CV was performed by CH instrument(CHI608C). To confirm the removal of sodium and arsenic, inductivelycoupled plasma optical emission spectroscopy (ICP-OES) analysis wascarried out using PERKIN-ELMER OPTIMA 5300DV ICP-OES instrument.

3. Results and discussion

3.1. Morphological study

TEM (Fig. 1a) and SEM (Fig. 1b) images of HEG reveal itsmorphological structure. TEM and SEM images clearly suggest thedisorder induced by exfoliation in graphite structure resulting in theform of sheets. The rapid removal of intercalated oxygen atoms andother functional groups in graphitic oxide, during exfoliation results ina wrinkled structure of graphene sheets.

BET surface area measurement of HEG (Fig. 1c) asserts the largehysteresis area of N2 adsorption–desorption isotherm suggesting thewide distributions of pores. The specific surface area of HEG calculatedusing BET equationwas found to be 442.87 m2/g. Large hysteresis areaindicates a near uniform distribution of pores and large surface area ofHEG, suggesting the high quality of synthesized graphene sheets. As

Fig. 1. TEM (a) and SEM (b) images of hydrogen exfoliated graphene sheets (c) BET measurement for surface area of hydrogen exfoliated graphene sheets.

0 20 40 60 800

25

50

75

0 20 40 60 800

1

2

0 20 40 60 80

0.00

0.06

0.12

Graphite

Graphite oxide

Inte

nsi

ty (

x 10

3 a.

u.)

HEG

Fig. 2. X-ray diffraction pattern of pure graphite, graphitic oxide and hydrogenexfoliated graphene sheets.

41A.K. Mishra, S. Ramaprabhu / Desalination 282 (2011) 39–45

synthesized graphene sheets possess much larger special surface area,which indicates that the average particle size of graphite has beendecreased during oxidation and rapid heating process under hydro-gen atmosphere. High surface area of HEG suggests its utility assupercapacitor electrodes.

3.2. Structural analysis

X-ray diffraction (XRD) pattern of pure graphite, graphitic oxideand HEG is shown in Fig. 2. XRD pattern of pure graphite shows a highintensity peak around 26.7° corresponding to the graphitic structure(002). After oxidation, the (002) peak of graphite powder disappearsand an additional peak at 10.6° is observed, which is corresponding tothe (001) diffraction peak of graphite oxide. Oxidation of graphiteleads to the increase in d-spacing, which can be attributed to the oxideinduced oxygen containing functional groups and inserted watermolecules. XRD pattern of HEG shows very less intense peak around24.4°, indicates the distorted graphite structure and hence suggeststhe formation of graphene sheets [26].

3.3. Raman spectrum and Fourier transform infrared spectrum analyses

Raman spectrum analysis of HEG and f-HEG is shown in the Fig. 3a.The figure shows two peaks corresponding to D-band (1346.5 cm−1)and G-band (1575.3 cm−1) for HEG. D-band corresponds to thedefects induced in the graphitic structure, while G-band correspondsto the in-plane vibrations of the graphitic structure [25]. Almost equalintensity of G-band and D-band suggests small crystalline size and

hence very few numbers of layers. In case of f-HEG, D-band(1350.6 cm−1) and G-band (1595.4 cm−1) shifts to the higher wavenumbers. It may be attributed to the addition of oxygen containing

1000 1500 2000

1000 1500 2000

1575.31346.5

HEG

1350.6 1595.4

Inte

nsi

ty (

a.u

.)

Raman shift (cm-1)

f-HEG

1000 2000 3000 40000

25

50

75

100

1000 2000 3000 4000

0

25

50

75

100

2853 2922

3428

HEG(b)

1048

1384

16271741 2853

2923 3430

Tran

smit

tan

ce (

a.u

.)

Wavenumber (cm-1)

f-HEG

(a)

Fig. 3. Raman (a) and FTIR (b) spectrums of hydrogen exfoliated graphene sheets andfunctionalized hydrogen exfoliated graphene sheets.

0.0 0.2 0.4 0.6 0.8 1.0-8

-4

0

4

8

0.0 0.2 0.4 0.6 0.8 1.0-8

-4

0

4

8

0.0 0.2 0.4 0.6 0.8 1.0-10

-5

0

5

10

Arsenate solution

Arsenite solution

Cu

rren

t x1

0-4 (

A)

sea water

Voltage (V)

f-HEGHEG

Fig. 4. Cyclic voltammetry curves for hydrogen exfoliated graphene sheets andfunctionalized hydrogen exfoliated graphene sheets with the scan rate of 5 mV/s.

42 A.K. Mishra, S. Ramaprabhu / Desalination 282 (2011) 39–45

functional groups at the surface of graphene sheets which bring abouta change in the reduced mass of the harmonic oscillator [27].

FTIR spectrums of HEG and f-HEG are shown in Fig. 3b. The bandcorresponding to hydroxyl group (\OH stretching, 3431 cm−1) isquite prominent as compared to the insignificant ratios of anti-symmetric and symmetric _CH2 vibrations (2920 and 2853 cm−1)for both HEG and f-HEG. Peak at 1627 cm−1 may be attributed to O\Hbending vibrations. Intense peaks corresponding to the C_O and C\Ostretching vibrations of COOH groups at 1741, 1048 and 1384 cm−1,respectively can also be observed [25,26].

3.4. Electrochemical analysis

Cyclic Voltammetry (CV) was performed to investigate theelectrochemical activity of HEG and f-HEG towards arsenate, arseniteand sodium ions. In CV analysis, Ag/AgCl and Pt wire electrodes weretaken as reference and counter electrodes respectively, while HEG andf-HEG modified glassy carbon electrodes were taken as workingelectrodes. CV analysis was performed with aqueous solutions ofsodium arsenate and sodium arsenite and sea water as electrolytes

with a constant scan rate of 5 mV/s. Fig. 4 shows the cyclicvoltammetry curves for all the electrolytes. The almost rectangularshapes of cyclic voltammetry curves obtained with each electrolyteand both electrode materials, suggest the formation of double layer(electronic charge of electrode and ionic charges of electrolyte at theirinterface). For each electrolyte capacitance corresponding to f-HEG isfound to be higher as compared to the HEG. This suggests thatfunctionalization of graphene sheets improves the adsorption, whichmay be attributed to the better contact between electrode andelectrolyte due to the intermediate bonding betweenwater moleculesand functional groups at the surface of graphene sheets [22,23].

3.5. Adsorption isotherm studies

The quantity of the metallic impurities that could be adsorbed atf-HEG based electrodes is a function of concentration, which can beexplained by adsorption isotherms. In the present study, Langmuirand Freundlich isotherms were tested with the simultaneousremoval of sodium and arsenic [28,29]. Sodium arsenate and sodiumarsenite containing aqueous solutions with the initial arsenic concen-trationsvarying from50 mg/L to 300 mg/Lwereused for abovepurpose.Langmuir model assumes that the single adsorbate binds to a single siteon the adsorbent and that all surface sites on the adsorbents have thesameaffinity for the adsorbate. Langmuir isotherm is representedby thefollowing equation

Qe =abCe

1 + bCeð1Þ

where ‘Qe’ is the amount of metal impurity adsorbed per unit weight ofadsorbent (mg/g), ‘Ce’ is the equilibrium concentration of water (mg/L),‘b’ is the constant related to the free energy of adsorption (L/mg) and ‘a’is maximum adsorption capacity.

The Freundlich isotherm can be derived from the Langmuir iso-therm by assuming that there exists a distribution of sites on theadsorbents for different adsorbates with each site behaving accord-ingly to the Langmuir isotherm. Freundlich isotherm is represented bythe following equation

Qe = k Ceð Þ1n ð2Þ

Table 1Isotherm constants for sodium and arsenic removal.

Isotherms constants Na As (V) As (III)

Langmuira (mg/g) 121.97 141.92 138.79b (L/mg) 0.0015 0.002 0.002R2 0.99591 0.9873 0.98837

Freundlichk 0.21652 0.35073 0.65341n 1.0765 1.11835 1.29319R2 0.99747 0.99692 0.98085

43A.K. Mishra, S. Ramaprabhu / Desalination 282 (2011) 39–45

where ‘k’ is the Freundlich constant indicative of the relative adsorptioncapacity of the adsorbent (mg/g) and (1/n) is the adsorption intensity.‘Qe’ is calculated by the following formula

Qe =C0−Ceð ÞV

mð3Þ

where ‘C0’ is the initial concentration of water, ‘V’ is the volume of waterand ‘m’ is themassof adsorbent. The isothermsconstants for sodiumandarsenic with both the isotherms studied were calculated. Fig. 5a showsthe comparative fit of Langmuir and Freundlich isotherms with theequilibriumdata plotted asQe vs.Ce for both arsenic and sodiumand theisotherm constants values are given in Table 1. From the experimentalresults, it canbe seen that Freundlichmodelfits better thanLangmuir forboth metallic impurities (arsenic and sodium). Maximum adsorptioncapacities for arsenate, arsenite and sodium were found to be nearly142, 139 and 122 mg/g, respectively. Value of ‘n’was found to be greater

0 50 100 150 200 250 3000

40

80

0 50 100 150 200 250 3000

40

80

0 50 100 150 200 250 3000

20

40

arsenate Langmuir Freundlich

(a)

Arsenite Langmuir Freundlich

Qe

(mg

/g)

Sodium Langmuir Freundlich

Ce (mg/L)

0 5 100

50

100

0 5 100

50

100

0 5 100

50

100

Number of C

As Na

% R

emov

al

As Na

Na Mg Ca K

(c)

Fig. 5. Adsorption isotherm (a) and kinetic (b) studies for sodium and arsenic removal (cdesalination of sea water.

than one for both impurities (sodium and arsenic), which is a favorablecondition for adsorption [30]. The maximum adsorption capacity usingLangmuir isotherm of graphene sheets based supercapacitor electrodes

Qe

(mg

/g)

0 25 50 75 1000

40

80

0 25 50 75 1000

30

60

0 25 50 75 1000

20

40

Arsenate Diffusion Elovich

(b)

Arsenite Diffusion Elovich

Sodium Diffusion Elovich

Time (min)

15 20 25

15 20 25

15 20 25

arsenate water

ycles

arsenite water

Sea water

) Percentage removal efficiency for simultaneous removal of sodium and arsenic and

Table 3Kinetic constants for sodium and arsenic removal.

Kinetic constants Na As (V) As (III)

Elovicha (mg/g min) 2.73515 6.17459 6.20276β (g/mg) 0.15726 0.07838 0.07352R2 0.97307 0.98529 0.98587

Intraparticle diffusionK (mg/g min1/2) 2.67611 5.77811 5.49347C (mg/g) 0.18916 0.15497 1.06187R2 0.99021 0.99122 0.99309

44 A.K. Mishra, S. Ramaprabhu / Desalination 282 (2011) 39–45

obtained above shows higher values than other reported values forarsenate and arsenite removal as mentioned in Table 2 [31–36].

3.6. Kinetic studies

The transient behavior of the metal adsorption process wasanalyzed by using Elovich and intra-particle diffusion kinetic models[37–39]. To study the simultaneous transient behavior of sodium andarsenic adsorption, sodium arsenate and sodium arsenite containingaqueous solutions with the initial arsenic concentration of 300 mg/Lwere tested. The linear form of Elovich equation is given by

Qt =1β

ln αβð Þ + 1β

ln t: ð4Þ

The rate constant for intra-particle diffusion (K) is given by-

Qt = K tð Þ12 + c ð5Þ

where ‘Qt’ is the amount of metal adsorbed on adsorbent at varioustime t (mg/g), ‘α’ is the initial sorption rate (mg/g min), ‘β’ is theextent of surface coverage (g/mg) and ‘K ’ is the intra-particle diffusionrate constant (mg/g min1/2) and c (mg/g) is a constant that gives anidea about the thickness of the boundary layer.

The rate constants for sodiumandarsenicwithboth themodelswerecalculated. Fig. 5b shows thecomparativefit of Elovichand intra-particlediffusion kinetic models with the equilibrium data plotted as Qt vs. t forboth arsenic and sodium and rate constants values are given in Table 3.Experimental results suggest that intra-particle diffusion model fitsbetter than Elovichmodel for bothmetallic impurities. Elovichmodel isan adsorption reaction model which originates from chemical reactionkinetics while intra-particle model is a diffusion model.

3.7. Removal efficiency

ICP-OES analysis was performed to verify the assumption of CV,whichsuggests that f-HEG is a goodelectrodematerial for supercapacitorbased water filter. ICP-OES analysis gives the actual concentration ofmetal ions in the solution. The removal efficiency in percentage wascalculated by using the following formula

%Removal efficiency =C0−Cf

� �100

C0ð6Þ

where ‘C0’ is the initial concentration of metal impurity in water and ‘Cf’is the final concentration of metal impurity in water after treatment.

Table 2Comparison of maximum adsorption capacity with other sorbents.

Adsorbents Maximum adsorption capacity(mg/g)

References

Hydrous Fe oxidewith polyacrylamide

~43(As V) 31

Waste Fe(III)/Cr(III) hydroxide ~11.02(As V) 32Activated alumina ~11.24–23.97(As V) 33Oxisole ~3.2 (As V), ~2.6 (As III) 34Gibsite ~4.6 (As V), 3.3(As III)Geothite ~12.5 (As V), 7.5(As III)Granular TiO2 ~41.4(As V), 32.4(As III) 35Magnetite-reducedgraphitic oxide

~5.83(As V), 13.10(As III) 36

Fe3O4-MWNTs based electrodes ~53 (As V), 39(As III), 37.8(Na) 22Functionalized MWNTsbased electrodes

~109.5(As III), 50.5(Na) 23

Functionalized graphene sheetsbased electrodes

~142(As V), 139(As III), 122(Na) Presentwork

To find the cyclic repeatability of the electrode material, differentcycles were performed with high concentration (300 mg/L of As) ofarsenic solution and sea water using the same electrode in each cycle.Since industrial waste water and sea water contain high concentra-tions of metallic impurities, performance of graphene based waterfilter was checkedwith high concentration of arsenic. Arsenic solutionand sea water were treated for 40 min. in each cycle, because thekinetic study clearly shows that after 40 min. the adsorption of metalions (Na and As) approaches to saturation state.

Fig. 5c shows the removal efficiency of supercapacitor based waterfilter for sodium and both types of arsenic removal and for desalinationof seawater. Nearly 65%of Arsenic (As) and66%of sodium(Na) removalefficiency was obtained in case of sodium arsenate containing waterwith 20 number of repeated cycles and 50 mg of f-HEG loading at eachelectrode. Water containing sodium arsenite removal efficiency wasfound to be 54% and 55% for ‘As’ and ‘Na’ respectively with the sameelectrodes and same number of cycles as above. In each case the initialconcentration of arsenic was 300 mg/L. Nearly equal percentageremoval of ‘As’ and ‘Na’ in either case suggests the removal of eachmetallic impurity without getting affected by the other impurities.Linear variation of removal efficiency with number of cycles performedsuggests the good cyclic repeatability of electrodes for simultaneousremoval of arsenic and sodium. Simultaneous removal of highconcentration of arsenic and sodium suggests the utilization of f-HEGbased supercapacitor for the purification of multiple metal impuritiescontaining water like sea water. Sodium, magnesium, calcium andpotassiumare found to bemost abundantmetal impurities in seawater.Therefore removal of Na,Mg, Ca andK from seawaterwas checkedwith50 mgof f-HEGat eachelectrode. Initial concentrations ofNa,Mg, CaandK in sea water were found to be 10,000, 1920, 680 and 570 mg/L,respectively. Removal efficiency was found to be 68, 71, 60 and 56% forNa, Mg, Ca and K respectively with 20 numbers of repeated cycles.Nearly linear variation of removal efficiency with respect to differentnumbers of cycles was observed which again suggests the good cyclicrepeatability andhence reproducibility of electrodes for even removal ofhigher concentrations of multiple metals in sea water. Reduction ofarsenic concentration in sea water could not be measured due to thelimitation of ICP-OES at much lower concentration of arsenic.

In addition, initial and final pH values of the arsenate and arsenitecontaining water were measured. Arsenate containing water showspH values of 6.92 and 7.07 before and after treatment respectively,while arsenite containing water shows pH values of 6.12 and 6.71before and after treatment, respectively. In each case solution isapproaching towards neutral value (pH7) after treatment, which maybe attributed to the adsorption of ions at the surface of electrodes dueto the double layer formation and hence reduction in concentration ofions in the solution. In our technique, DC voltage of 1 V works as adriving force for ions (metallic impurities), which help them to formdouble layer at each electrode and hence showing the high removal ofionic impurities. This driving force (applied voltage) overcomes theeffect of ionic strength of electrolyte due to pH and is applicable foreven industrial waste water for removing metal impurities.

45A.K. Mishra, S. Ramaprabhu / Desalination 282 (2011) 39–45

4. Conclusions

To the best of our knowledge, the present work is the first study ofdesalination of sea water using graphene sheets based supercapacitor.Desalination of sea water and simultaneous removal of sodium andboth inorganic arsenic species (arsenate and arsenite) from aqueoussolution has been successfully demonstrated by carbon fabricsupported f-HEG electrodes based supercapacitor. Functionalizationof HEG provides better contact between electrodes and water,resulting in better performance of metals removal from aqueoussolution and sea water. Adsorption isotherms fit better to Freundlichmodel as compared to Langmuir model for sodium and arsenic.Adsorption kinetics was found to fit better for intra particle diffusionas compared to Elovich model. High removal efficiency and goodcyclic repeatability of electrodes has been demonstrated for highconcentration of both arsenic ions containing aqueous solution andsea water. Removal of high concentration of metals like Na, Mg, Ca, Kand As suggests the possible utilization of this filter for thepurification of industrial waste water disposal. Additionally, costeffective production of graphene sheets and better performancecompared to other adsorbents like carbon nanotubes provides aplatform for the development of commercially feasible supercapacitorbased water filter.

Acknowledgements

The authors acknowledge the supports of Office of Alumni Affairs,IIT Madras and DST, India. The authors acknowledge SAIF, IIT Madrasfor its support in ICP-OES and FTIR measurements. Authors arethankful to NCCR, IITMadras for BETmeasurement. One of the authors(Ashish) is thankful to DST India for providing financial support.

References

[1] K.M. Hosamani, R.S. Keri, S.K. Nataraj, T.M. Aminabhavi, R.S. Harisha, Arsenicremoval from drinking water using thin film composite nanofiltration membrane,Desalination 252 (2010) 75–80.

[2] C.M. Nguyen, S. Bang, J. Cho, K.W. Kim, Performance and mechanism of arsenicremoval from water by a nanofiltration membrane, Desalination 245 (2009)82–94.

[3] S.Y. Thomas, T.G. Choong, Y. Chuah, F.L. Robiah, I. Gregory Koay, Azni Arsenictoxicity, health hazards and removal techniques from water: an overview,Desalination 217 (2007) 139–166.

[4] M. Jang, W. Chen, F.S. Cannon, Preloading hydrous ferric oxide into granularactivated carbon for arsenic removal, Environ. Sci. Technol. 42 (2008) 3369–3374.

[5] D.L. Boccelli, M.J. Small, D.A. Dzombak, Enhanced coagulation for satisfying thearsenic maximum contaminant level under variable and uncertain conditions,Environ. Sci. Technol. 39 (2005) 6501–6507.

[6] B.K. Mandal, K.T. Suzuki, Arsenic round the world: a review, Talanta 58 (2002)201–235.

[7] M.M. Benjamin, R.S. Sletten, R.P. Bailey, T. Bennet, Sorption and filtration of metalsusing iron-oxide-coated sand, Water Res. 30 (1996) 2609–2620.

[8] E.O. Kartinen, C.J. Martin, An overview of arsenic removal processes, Desalination103 (1995) 79–88.

[9] S. Deng, G. Yu, S. Xie, Q. Yu, J. Huang, Y. Kuwaki, M. Iseki, Enhanced adsorption ofarsenate on the aminated fibers: sorption behavior and uptake mechanism,Langmuir 24 (2008) 10961–10967.

[10] J. Zhang, R. Stanforth, Slow adsorption reaction between arsenic species andGoethite (α-FeOOH): diffusion or heterogeneous surface reaction control,Langmuir 21 (2005) 2895–2901.

[11] Y. Wu, X. Ma, M. Feng, M. Liu, Behavior of chromium and arsenic on activatedcarbon, J. Hazard. Mater. 159 (2008) 380–384.

[12] P. Mondal, C.B. Majumder, B. Mohanty, Treatment of arsenic contaminated waterin a batch reactor by using Ralstonia eutropha MTCC 2487 and granular activatedcarbon, J. Hazard. Mater. 153 (2008) 588–599.

[13] M. Demircioglu, N. Kabay, I. Kurucaovali, E. Ersoz, Demineralization byelectrodialysis (ED) separation performance and cost comparison for monovalentsalts, Desalination 153 (2003) 329–333.

[14] A. Grundisch, B. Schneider, Optimising energy consumption in SWRO systemswith brine concentrators, Desalination 138 (2001) 223–229.

[15] R. Semiat, Y. Galperin, Effect of non-condensable gases on heat transfer in thetower MED seawater desalination plant, Desalination 140 (2001) 27–46.

[16] C. Charcosset, A review of membrane processes and renewable energies fordesalination, Desalination 245 (2009) 214–231.

[17] Y.H. Li, S. Wang, Z. Luan, J. Ding, C. Xu, D. Wu, Adsorption of cadium (II) fromaqueous solution by surface oxidized carbon nanotubes, Carbon 41 (2003)1057–1062.

[18] Y.H. Li, J. Ding, Z. Luan, Z. Di, Y. Zhu, C. Xu, D.Wu, B.Wei, Competitive adsorption ofPb2+, Cu2+ and Cd2+ ions from aqueous solutions by multiwalled carbonnanotubes, Carbon 41 (2003) 2787–2792.

[19] F. Fornasieroa, H.G. Parkb, J.K. Holta, M. Stadermanna, C.P. Grigoropoulosc, A. Noy,O. Bakajin, Ion exclusion by sub-2-nm carbon nanotube pores, PNAS 105 (2008)17250–17255.

[20] A.K. Geim, K.S. Noveselov, The rise of graphene, Nat. Mater. 6 (2006) 183–191.[21] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, I.V.

Grigorieva, A.A. Firsov, Electric field effect in atomically thin carbon films, Science306 (2004) 666–669.

[22] A.K. Mishra, S. Ramaprabhu, Magnetite decorated multiwalled carbon nanotubesbased supercapacitor for arsenic removal and desalination of sea water, J. Phys.Chem. C 114 (2010) 2583–2590.

[23] A.K. Mishra, S. Ramaprabhu, The role of functionalized multi walled carbonnanotubes based supercapacitor for arsenic removal and desalination of seawater, J. Exp. Nanosci. in press.

[24] W. Hummers Jr., R.E. Offeman, Preparation of graphitic oxide, J. Am. Chem. Soc. 80(1958) 1339.

[25] A. Kaniyoor, T.T. Baby, S. Ramaprabhu, Graphene synthesis via hydrogen inducedlow temperature exfoliation of graphite oxide, J. Mater. Chem. 20 (2010)8467–8469.

[26] P. Liana, X. Zhu, S. Liang, Z. Li, W. Yang, H. Wang, Large reversible capacity of highquality graphene sheets as an anode material for lithium-ion batteries,Electrochim. Acta 55 (2010) 3909–3914.

[27] V.A. Sinani, M.K. Gheith, A.A. Yaroslavov, A.A. Rakhnyanskaya, K. Sun, A.A.Mamedov, J.P. Wicksted, N.A. Kotov, Aqueous dispersions of single-wall andmultiwall carbon nanotubes with designed amphiphilic polycations, J. Am. Chem.Soc. 127 (2005) 3463–3472.

[28] I. Langmuir, The adsorption of gases on plane surfaces of glass, mica and platinum,J. Am. Chem. Soc. 40 (1918) 1361–1403.

[29] H.M.F. Freundlich, Over the adsorption in solution, J. Phys. Chem. 57 (1906)385–471.

[30] Q.U. Jiuhui, Research progress of novel adsorption processes in water purification:a review, J. Environ. Sci. 20 (2008) 1–13.

[31] Y. Shigetomi, Y. Hori, T. Kojima, The removal of arsenate in wastewater with anadsorbent prepared by binding hydrous iron(III) oxide with polyacrylamide, Bull.Chem. Soc. Jpn 53 (1980) 1475–1476.

[32] Namasivayam Chinnaiya, S. Senthilkumar, Removal of arsenic(V) from aqueoussolution using industrial solid waste: adsorption rates and equilibrium studies,Ind. Eng. Chem. Res. 37 (1998) 4816–4822.

[33] M.M. Ghosh, J.R. Yuan, Adsorption of inorganic arsenic and organoarsenicals onhydrous oxide, Environ. Prog. 6 (1987) 150.

[34] A.C.Q. Ladeira, V.S.T. Ciminelli, Adsorption and desorption of arsenic on an oxisoland its constituents, Water Res. 38 (2004) 2087–2094.

[35] S. Bang, M. Patel, L. Lippincott, X. Meng, Removal of arsenic from groundwater bygranular titanium dioxide adsorbent, Chemosphere 60 (2005) 389–397.

[36] V. Chandra, J. Park, Y. Chun, J.W. Lee, I.C. Hwang, K.S. Kim, Water-dispersiblemagnetite-reduced graphene oxide composites for arsenic removal, ACS Nano 4(2010) 3979–3986.

[37] H.A. Taylor, N. Thon, Kinetics of chemisorptions, J. Am. Chem. Soc. 74 (1952)4169–4173.

[38] M. Ozacar, I.A. Sengil, A kinetic studies of metal complex dye sorption onto pinesaw dust, Process Biochem. 40 (2005) 565–572.

[39] H. Qiu, B.C. Pan, Q.J. Zhang, W.M. Zhang, Q.X. Zhang, Critical review in adsorptionkinetic models, J. Zhejiang Univ. Sci. A. 10 (2009) 716–724.