Selective Dye Adsorption by pH Modulation on Amine-Functionalized Reduced Graphene Oxide−Carbon Nanotube HybridChandrama Sarkar, Chandramika Bora, and Swapan K. Dolui*

Department of Chemical Science, Tezpur University, Napaam, Assam 784028, India

*S Supporting Information

ABSTRACT: Simultaneous functionalization, reduction, and hybridization of graphene oxide in graphene oxide−multiwallcarbon nanotube hybrid were carried out by p-phenylenediamine (rGO−CNT−PPD). The adsorption properties of the hybridfor selective removal of methyl violet (MV) and methyl orange (MO) from their mixture were investigated under differentexperimental conditions. The adsorption was correlated with the effects of pH, contact time, and temperature. Changing the pHof the solution modulates the selectivity of adsorption behavior of the hybrid. At pH 7.0, it can remove 99.7% MV selectivelyfrom the mixture, whereas at pH 3.0, it can remove MO selectively having 98% removal efficiency. The maximum adsorptioncapacities (qm) of the hybrid for MV and MO are 298 and 294 mg/g, respectively, at 298 K; these values are higher than severalpreviously reported qm values. The kinetics data for both dyes can be well fitted with the pseudo-second-order kinetics model,and the adsorption of both dyes followed a Langmuir adsorption isotherm. The adsorbent can be easily regenerated and reused,and its activity remains the same even after five adsorption−desorption cycles. The rGO−CNT−PPD hybrid could be apromising adsorbent for selective removal of synthetic dyes from their mixture by pH modulation.

1. INTRODUCTION

Dyes have become indispensable chemicals for a variety ofindustries such as dye synthesis, textile, printing, paper,electroplating, pulp mill, food, paints, polymers, and cosmeticsand the highly sophisticated biotechnology industry. Unfortu-nately, most of them are toxic, carcinogenic, and teratogenic tohuman beings.1 Waste dyes are the major source of waterpollution. Methyl violet (MV) and methyl orange (MO) arewidely used cationic and anionic dyes in several industries andresearch laboratories. Hence, MV and MO were selected asrepresentative target pollutants for this study. Nowadays,several physical, chemical, and biological methods are used inwater treatment, such as adsorption,2,3 membrane separation,4

photocatalytic degradation,5 ozone treatment,6 oxidation,7,8 ionexchange,9,10 and biological treatment.11 Among thesemethods, adsorption is the most widely used method becauseof its low cost, ease of operation, high efficiency, and simplicityin design and operation.Various carbon-based materials such as activated carbon,12

carbon nanotubes (CNT),13 porous carbon,14 graphene oxide(GO), graphene,15−17 and graphene sponges18 have beenstudied for adsorption of various dyes from water because oftheir excellent physiochemical and mechanical properties.Although activated carbon has high specific surface area, it isflammable and difficult to regenerate, and its weak hydrophilicproperties lead to weak interaction between the adsorbent anddyes.19 Due to hollow and layered structures, CNTs have beenutilized for the adsorption of a large number of different organiccompounds from water. The multiwall carbon nanotube hasbeen used as an efficient adsorbent for the removal of severaldyes such as alizarin red s, morin, reactive blue 4, and acid red183 from wastewater.20−23 GO nanosheets are decorated withnumbers of oxygen-containing functional groups such ashydroxyl, carboxylic, carbonyl, and epoxide groups, which

impart negative charge density to it in aqueous solution. Theirhigh water solubility and the electrostatic interaction of GOwith adsorbate make them a material of great interest inadsorption of charged species.15,23,24 Many research groupshave reported the use of GO as adsorbent material for removalof various dyes such as MV, rhodamine B, orange G, methyleneblue, and methyl green from aqueous solutions.15,25 Although alarge number of studies have been reported for effective dyeadsorption, very few studies emphasized selective dyeadsorption from their mixture.22,26 Ho et al. demonstratedordered mesoporous silica (OMS) could be used as anadsorbent for removal of acid blue 25 and methylene bluedyes selectively from their mixture.26 The amino-containingOMS-NH2 adsorbent can selectively remove acid blue 25,whereas OMS-COOH can absorb methylene blue from theirmixture.Recently, a new group of materials have been synthesized

from the hybridization of CNTs and GO (GO−CNT), whichshowed better properties than their individual counter-parts.27−31 However, the dye adsorption efficiency of theGO−CNT hybrid was inferior to that of a reduced grapheneoxide−carbon nanotube hybrid (rGO−CNT).32 Reduction ofGO required some highly toxic and harmful reducing agentssuch as hydrazine hydrate33 and sodium borohydride.34

Therefore, to improve the dye adsorption properties of GO−CNT hybrid, development of a suitable reduction process isessential by exploring simultaneous functionalization followedby reduction of GO. Recently, simultaneous reduction,functionalization, and stitching of GO have been carried out

Received: July 3, 2014Revised: September 17, 2014Accepted: September 25, 2014

Article

pubs.acs.org/IECR

© XXXX American Chemical Society A dx.doi.org/10.1021/ie502653t | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

with ethylemediamine35 and p-phenylenediamine (PPD)36 forvarious applications.In the present work, a novel chemical approach is established

to functionalize the GO−CNT hybrid by PPD followed byreduction of GO. This approach eliminates the use of toxic andharmful chemicals used for the synthesis of rGO. The aim ofthis study is to explore the selective dye adsorption efficiency ofrGO−CNT−PPD from the mixture of MV and MO solutionby pH modulation. The adsorption properties of the hybrid forselective removal of MV and MO from their mixture wereinvestigated under different experimental conditions. Theeffects of pH, contact time, and temperature on the adsorptioncapacity have been investigated. Also, the adsorption kineticsand isotherms have been estimated to evaluate the adsorptioncapacity and understand the mechanism of adsorption of rGO−CNT−PPD hybrid toward MV and MO.

2. EXPERIMENTAL SECTION2.1. Chemicals. Multiwall carbon nanotubes (Redex

Technologies Pvt. Ltd., Ghaziabad, UP, India; purity > 97%)and graphite powder (<20 μm, Sigma-Aldrich; 99% pure) wereused without further purification. Sodium nitrate (NaNO3),potassium permanganate (KMNO4), hydrogen peroxide (50%v/v), sodium lauryl sulfate, concentrated sulfuric acid (98%),hydrochloric acid, and ethanol (99%) were procured fromMerck. Methyl violet (Merck) and methyl orange (Merck)were recrystallized.2.2. Preparation of Graphene Oxide (GO) GO−CNT

Hybrid. GO was synthesized from natural graphite powderusing the Hummers method.37 GO−CNT hybrid wassynthesized by ultrasonication of GO and CNT. In thisprocess, MWCNT (10 mg) was dispersed in 0.5 wt % sodiumdodecyl sulfate (20 mL) by ultrasonication to form a

homogeneous dispersion. GO (30 mg) was dispersed in 20mL of water by ultrasonication. These two homogeneousdispersions were mixed and again sonicated for 30 min. Thenthe GO−MWCNT dispersion was centrifuged and washedseveral times with distilled water and finally, after drying, theGO−CNT hybrid was obtained as a black solid.

2.3. Preparation of rGO−CNT−PPD Hybrid. A simulta-neous reduction and functionalization of the GO−CNT hybridwas carried out by PPD. In a typical process, PPD−ethanol (0.3g in 100 mL) solution was mixed with equal volumes of theGO−CNT (0.1 g in 100 mL) dispersion prepared in water.The mixture was then refluxed at 80 °C for 24 h with stirring.The obtained solution was centrifuged and washed repeatedlywith a 1:1 ethanol/water solution until the centrifugal liquidbecame colorless and finally dried to obtain the rGO−CNT−PPD hybrid. Steps involved in the synthesis of rGO−CNT−PPD hybrid are shown in Scheme S1 of the SupportingInformation. For comparison, rGO−PPD hybrid and CNT−PPD hybrid were also synthesized following the sameprocedure as used in the synthesis of rGO−CNT−PPD hybrid.In the synthesis of CNT−PPD hybrid, CNT was dispersed in0.5 wt % sodium dodecyl sulfate.

2.4. Characterization. The infrared spectra were recordedwith samples on KBr pellets over a frequency range of 4000−500 cm−1 in a Nicolet model impact 410 FT-IR spectropho-tometer. The X-ray diffraction (XRD) study was carried out atroom temperature (ca. 298 K) using a Rigaku X-raydiffractometer with Cu Kα radiation (λ = 0.15418 nm) at 30kV and 15 mA using a scanning rate of 0.05 θ/s in the range of2θ = 10°−70°. The morphology of the GO−CNT hybrid andthe rGO−CNT−PPD hybrid were characterized using ascanning electron microscope (SEM) and a transmissionelectron microscope (TEM). The SEM characterizations were

Figure 1. FTIR spectrum of (a) GO, (b) MWCNT, (c) GO−CNT hybrid, (d) rGO−CNT−PPD hybrid, (e) rGO−PPD, and (f) CNT−PPD.

Industrial & Engineering Chemistry Research Article

dx.doi.org/10.1021/ie502653t | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXXB

conducted on a JSM-6390LV (Jeol, Japan) scanning electronmicrograph with an energy dispersive X-ray detector. Scanningwas done at micrometer range; images were taken at anaccelerating voltage of 15−20 kV and magnifications of ×15000and ×20000. Data were obtained by using INCA software. TheTEM measurements were carried out by using an FEI 200 kVTEM model Technai G2 F2OS-TWIN instrument.2.5. Batch Adsorption Experiments. Selective adsorption

properties of the rGO−CNT−PPD hybrid were investigatedfor mixtures of two different organic dyes, MO and MV, using abatch adsorption experiment. The hybrid adsorbent (0.005 g)was added to 50 mL of mixture solution having a 1:1 MO/MVratio with 30 mg/L concentration of each dye, and the mixturewas stirred for 120 min. The adsorbent was separated bycentrifugation and the concentrations of each dye afteradsorption was determined by using a UV−vis spectropho-tometer at 464 and 582 nm wavelengths for MO and MV,respectively. The effect of pH on the adsorption was studiedwith an initial dye concentration of 30 mg/L for both dyes in apH range of 1.0−11.0 at 298 K. The pH value of the mixedsolution was adjusted with a 0.1 M HCl or 0.1 M NaOHsolution. All other experiments for selective adsorption of MVand MO were performed under pH values of 7.0 and 3.0,respectively. Kinetic experiments were carried out with aninitial dye concentration of 30 mg/L for both dyes at roomtemperature to determine the minimum time required foradsorption to reach steady state. The concentrations of dyesafter adsorption were measured at regular time intervals from15 to 180 min. The effect of temperature was examined todetermine the adsorption isotherm at different temperatures:298, 308, 318, 328, and 338 K. The adsorption isothermexperiments were performed by adding 0.005 g of adsorbent in50 mL of the aqueous solution with initial dye concentrationsof 2−50 mg/L. The adsorption capacity (qe, mg/g) of thehybrid and the percentage of removal of dyes by the adsorbentcan be calculated by applying eqs 1 and 2, respectively

= −q C C V m( ) /e 0 e (1)

= × −C C C% removal 100 ( )/0 e 0 (2)

where C0 and Ce represent the initial and equilibrium dyeconcentrations of aqueous solution (mg/L), V is the volume ofthe solution (L), and m is the mass of the adsorbent (g).2.6. Desorption Experiments. After adsorption experi-

ments, the adsorbent was separated by centrifugation, addedinto 50 mL of ethanol, and stirred for 120 min at pH 7.0 forMV. The adsorbent was collected by centrifugation and reusedfor another adsorption−desorption cycle. The concentration ofthe supernatant dye solution was examined by UV−visspectroscopy. The cycle was repeated five times following thesame procedure.

3. RESULTS AND DISCUSSION3.1. Characterization of rGO−CNT−PPD Hybrid.

3.1.1. FTIR Study. Figure 1 shows the FTIR spectra of GO,MWCNT, GO−CNT, rGO−CNT−PPD, rGO−PPD, andCNT−PPD. In the FTIR spectrum of GO (Figure 1a), thepeaks obtained at 1250, 1066, and 1720 cm−1 are attributed tothe CO epoxy stretching, CO alkoxy stretching, and CO carbonyl stretching vibrations, respectively. The absorptionbands at 1623 and 3420 cm−1 correspond to the aromatic CC stretching vibration and OH stretching vibration,respectively. Pristine MWCNT shows the characteristic CC

stretching at 1637 cm−1 (Figure 1b). In the FTIR spectrum ofGO−CNT hybrid (Figure 1c), the peaks related to both GOand CNT are observed with a little shifting of the peaks, whichindicates the interaction between GO and CNT. The peaks at1066, 1250, 1623, 1720, and 3420 cm−1 in the FTIR spectra ofGO are shifted to 1073, 1260, 1629, 1718, and 3363 cm−1,respectively, in the spectrum of the GO−CNT hybrid.Functionalization on the GO−CNT hybrid by PPD as wellas reduction of GO was confirmed by the FTIR spectrum of therGO−CNT−PPD hybrid as shown in Figure 1d. This spectrumshows that the peak intensity of CO, COH, and COgroups decreases significantly, indicating reduction of GO byPPD. Again, shifting of the peaks of the GO−CNT hybrid inthe FTIR spectrum of the rGO−CNT−PPD hybrid confirmsinteraction between them. Some new peaks appeared at 3410,3360, 1610, and 770 cm−1 corresponding to −NH2 doublestretching, N−H bending, and N−H wagging vibrations,respectively, confirming functionalization of the GO−CNThybrid with PPD. A strong peak arises at 1168 cm−1 due to C−N stretching vibration mode, which confirms the formation ofC−NH−C bands. Thus, FTIR analysis indicated simultaneousreduction of GO and amine functionalization on the GO−CNThybrid. For comparison, FTIR spectra of rGO-PPD and CNT−PPD are also examined and shown in Figure 1, panels e and f,respectively. Both spectra show the characteristic peaks foramine functional groups as well as for C−N stretching vibrationmode. In the spectrum of rGO−PPD, the intensities of thebands related to the oxygen functionalities are found todecrease, implying the reduction of GO by PPD. In bothspectra, peaks appeared at 1171 cm−1 (Figure 1e) and 1167cm−1 (Figure 1f) for C−N stretching vibration, which indicatessuccessful functionalization of PPD into both GO and CNTthrough C−NH−C linkage. Thus, the results indicate obviousfunctionalization of PPD in the GO−CNT hybrid as the peakfor C−N stretching vibration is present in both rGO-PPD andCNT−PPD components.

3.1.2. XRD Analysis. Figure 2 shows XRD patterns of GO,CNT, GO−CNT, and rGO−CNT−PPD hybrid. Pristine

graphite exhibits a strong crystalline peak at 26.5° for the(002) plane.35 However, GO exhibited a strong diffraction peakat 2θ = 11.45°, which can be indexed as (001) reflection due tooxygen-containing functional groups on carbon sheets. TheXRD pattern of CNT displayed a sharp (002) peak at 2θ =25.15°. The XRD of the GO−CNT hybrid exhibits a strongdiffraction peak at 2θ = 25.05° for CNT. rGO−CNT−PPD

Figure 2. XRD patterns of (a) GO, (b) CNT, (c) GO−CNT, and (d)rGO−CNT−PPD.

Industrial & Engineering Chemistry Research Article

dx.doi.org/10.1021/ie502653t | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXXC

exhibits a strong peak at 2θ = 26.1° with a slightly broadeningnature. The strong peak at 26.1° is due to the presence of CNT,and the broadening is due to their interactions, confirming theformation of the rGO−CNT−PPD hybrid.3.1.3. SEM Analysis. A salient difference in surface

morphology between GO−CNT and rGO−CNT−PPD hybridis observed (Figure 3). The SEM image of GO−CNT revealedthat the GO sheets are entirely wrapped by the tubularnetworks of CNT, confirming the incorporation of CNT intoGO (Figure 3a), whereas after functionalization with PPD, arough morphology of the surface of the hybrid was observed,indicating functionalization of the hybrid (Figure 3b).3.1.4. TEM Analysis. TEM images of GO−CNT and rGO−

CNT−PPD hybrid are shown in Figure 4. The TEM image ofGO−CNT hybrid (Figure 4a) revealed that nanotubes wereattached on the GO sheet without large bundles or aggregates.However, a different morphology is observed in Figure 4b. Thewalls of the CNT become thicker due to the presence of PPD.Thus, the image confirms successful functionalization of thehybrid by PPD.3.2. Adsorption Equilibrium and Kinetics Study.

Equilibrium studies of MV and MO were performed at roomtemperature at pH 7.0 and 3.0, respectively. Kinetics study wasinvestigated to understand the mechanism of adsorption. Tofind out the minimum time required for adsorption to reach thesteady state to determine the adsorption kinetics, the effect ofcontact time was examined. It can be seen in Figure 5 that the

adsorption capacity of the rGO−CNT−PPD hybrid increasedinstantly within the first 20 min and thereafter proceededslowly until the steady state was reached within 120 min, whichindicates very fast adsorption kinetics of the dyes on the hybrid.After 120 min, the value of qe did not change. On the basis ofthe above result, a contact time of 120 min was selected forboth cases for a sure establishment of the adsorption

Figure 3. SEM images of (a) GO−CNT and (b) rGO−CNT−PPD hybrid.

Figure 4. TEM images of (a) GO−CNT and (b) rGO−CNT−PPD hybrid. Inset in (a) is the TEM image of GO−CNT hybrid at higher resolution(50 nm).

Figure 5. Effect of contact time on adsorption capacity at pH 7.0 and3.0 onto rGO−CNT−PPD hybrid.

Industrial & Engineering Chemistry Research Article

dx.doi.org/10.1021/ie502653t | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXXD

equilibrium in further adsorption studies. At pH 7.0, the valueof qe for the adsorption of MV was found to be 299 mg/g atsteady state, whereas at pH 3.0 the value was 294 mg/g due tothe adsorption of MO.To investigate the detailed characteristics of the adsorption

phenomenon, two conventional kinetics models, pseudo-first-order and pseudo-second-order kinetics models, were appliedto identify the dynamic of the adsorption process. The pseudo-first-order model can be expressed as38

− = −q q q k tlog( ) log /2.303te e 1 (3)

where qe and qt are the adsorbed dye amounts on the hybrid atequilibrium and at various times t (mg/g), respectively. k1 is therate constant of the pseudo-first-order model of adsorption(min−1). The pseudo-second-order model can be expressedas39

= +t q k q t q/ 1/ /t 2 e2

e (4)

where qe and qt are defined as in the pseudo-first-order modeland k2 is the rate constant of the pseudo-second-order model ofadsorption (g/mg·min). The values of qe and k2 can becalculated from the slope and intercept of the linear plot of t/qtagainst t.Kinetic parameters and correlation coefficients (R2) are

summarized in Table 1 (kinetic models are shown in Figure S1of the Supporting Information). The R2 values for the pseudo-second-order model at pH 7.0 and 3.0 are 1, whereas for thepseudo-first-order model the values are 0.946 and 0.923,respectively. Again, the experimental adsorption capacity (qe,exp)

values at both pH values of the pseudo-first-order model werealso very close to the calculated adsorption capacity (qe,cal).Thus, these results suggest that the pseudo-second-orderkinetics model fits the adsorption of MV as well as MO onrGO−CNT−PPD hybrid at pH 7.0 and 3.0, respectively.

3.3. Adsorption Isotherms. The adsorption isothermmodels provide information on the interaction between theabsorbent and the adsorbate when the adsorption processreaches equilibrium. Langmuir and Freundlich isotherm modelswere used to analyze adsorption equilibrium. The Langmuirisotherm theory assumes that adsorption cannot proceedbeyond monolayer coverage of adsorbate over a homogeneousadsorbent surface, and there are no interactions betweenadsorbed molecules. The Freundlich isotherm is an empiricalequation based on multilayer adsorption on a heterogeneoussurface. The equations of the linearized Langmuir isotherm (eq5) and Freundlich (eq 6) adsorption isotherm models are40,41

= +C q C q q K/ / 1/e e e m m L (5)

= +q n C Kln (1/ ) ln lne e F (6)

where Ce is the equilibrium concentration of the dye in solution(mg/L), qe is the adsorbed dye amount per gram of the hybrid(mg/g), qm represents the maximum adsorption capacity of theadsorbent (mg/g), and KL and KF are the Langmuir andFreundlich constants, which are related to the affinity of thebinding sites and adsorption capacity, respectively. The valuesof qm and KL can be calculated from the slope and intercept ofthe linear plot of Ce/qe versus Ce. n represents the adsorption

Table 1. Kinetic Parameters for the Adsorption of MV at pH 7.0 and of MO at pH 3.0 onto rGO−CNT−PPD Hybrid

pseudo-first-order kinetics pseudo-second-order kinetics

dye pH qe,exp (mg/g) k1 (min−1) qe,cal (mg/g) R2 k2 (g/mg·min) qe,cal (mg/g) R2

MV 7.0 298 0.046 74.7 0.945 1.86 × 10−3 333.33 1MO 3.0 294 0.048 46.77 0.923 2.89 × 10−3 333.33 1

Table 2. Isotherm Parameters for the Adsorption of MV and MO by rGO−CNT−PPD Hybrid at pH 7.0 and 3.0, Respectively

Langmuir isotherm Freundlich isotherm

dye pH T (K) qm(mg/g) KL(L/mg) R2 KF n R2

MV 7.0 298 333.33 0.67 0.992 150.66 3.86 0.713MO 3.0 298 333.33 0.33 0.999 187.35 2.39 0.69

Figure 6. (a) Effect of pH on adsorption of MV−MO mixture on rGO−CNT−PPD hybrid at 464 and 582 nm; (b) UV−visible spectra of MO +MV solution before adsorption (red), after adsorption at pH 7.0 (yellow) and after adsorption at pH 3.0 (violet).

Industrial & Engineering Chemistry Research Article

dx.doi.org/10.1021/ie502653t | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXXE

intensity; n > 1 indicates a favorable condition for adsorption.The values of KF and n can be obtained from the intercept andslope of the linear plot of ln qe against ln Ce.The isotherms of both models at pH 7.0 and 3.0 based on

the experimental data are shown in Figure S2 of the SupportingInformation, and the relative parameters and correlationcoefficients are summarized in Table 2. At both pH values, itcan be seen that the adsorption capacity increases with anincrease in the initial dye concentration. As the concentrationincreases, the driving force from the concentration gradientincreases, which accelerates the diffusion speed of dyes towardthe rGO−CNT−PPD hybrid. The better fit by the Langmuirisotherm for both cases demonstrates that a monolayeradsorption process has taken place on the surface of therGO−CNT−PPD hybrid and there is no subsequentinteraction between adsorbed species.3.4. Effect of pH. Solution pH is one of the most important

factors in determining the adsorption properties of anadsorbent because it can change the net charge of theadsorbent and adsorbate. Figure 6a shows the effect of thesolution pH on MO−MV mixture adsorption by rGO−CNT−PPD hybrid with the initial pH ranging from 1.0 to 11.0. Atdifferent pH values, the concentration of dyes after adsorptionwas measured by UV−visible spectroscopy at two differentwavelengths, 464 and 582 nm. At 464 nm, the value of qe ishighest (294 mg/g) at pH 3.0, whereas qe decreases sharplyfrom pH 7.0 onward. At 582 nm, the hybrid showed a littlechange in adsorption capacity (around 294 mg/g) in the pHrange from 6.0 to 11.0, with the maximum at pH 7.0 (298.5mg/g), whereas at pH 3.0, the adsorption capacity declinedsharply and even reached 27.5 mg/g (at 582 nm wavelength).Again, the UV−vis spectrum (Figure 5b) shows that at pH 7.0,there is no peak at 582 nm wavelength for MV, whereas at pH3.0, the peak for MO at 464 nm disappeared completely. Theseresults indicated that at pH 7.0 the adsorbent can remove MV

and that at pH 3.0 the adsorbent selectively removes MO fromthe mixture of MO−MV solution. For these reasons all otherexperiments were carried out at pH 7.0 and 3.0 to remove MVand MO, respectively.The effect of pH on selective adsorption of MO and MV

from their mixture is shown Scheme 1. At pH 7.0, rGO−CNT−PPD hybrid can remove 99.7% MV selectively from themixture, which is due to the electrostatic interaction betweenthe adsorbent and the cationic dye MV. Two types ofelectrostatic interactions occur between the adsorbent andMV. Nitrogen atoms of PPD can offer their lone pairs to form acoordinate bond with the cationic dye, and the unreducedoxygen-containing functional groups of GO can effectivelyattract the cationic dye, whereas at pH 3.0, the hybrid canremove MO selectively with 98% removal efficiency. At thispH, concentration of hydrogen ion increases in the solution,which neutralizes the negative charges of the hybrid. Due to thepresence of excess protons at pH 3.0, protonation takes place inthe amine functional group of PPD, as a result electrostaticinteractions taking place between the adsorbent and the anionicdye MO. Therefore, at pH 3.0, the rGO−CNT−PPD hybridadsorbed only MO.

3.5. Regeneration of the rGO−CNT−PPD Hybrid. Theadsorbent can be easily regenerated and reused for furtheradsorption−desorption cycles. To evaluate the reusability of therGO−CNT−PPD hybrid, further adsorption−desorption ex-periments were carried out at pH 7. Figure 7 shows that theremoval percentage of MV from the mixture remains almost thesame even after five cycles. The value of the removal percentageof MV at the first adsorption−desorption cycle was 99.76%,whereas after the fifth cycle the value was 99.6%. These resultsindicated that the adsorbent remained effective even afterextended regeneration and reuse cycles.

Scheme 1. Effect of pH on Adsorption of MV and MO

Industrial & Engineering Chemistry Research Article

dx.doi.org/10.1021/ie502653t | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXXF

4. SELECTIVE DYE ADSORPTIONTo our knowledge, there is no reported work for selectiveadsorption of MV and MO from their mixture. Ho et al.26 usedtwo different adsorbents (OMS-NH2 and OMS-COOH) forselective adsorption of two differently charged dyes from theirmixture. On the other hand, in our study, using the sameadsorbent, selective adsorption of a cationic and an anionic dyewas carried out by simply changing the pH of the solution. Theadsorbent selectively adsorbed one dye at a certain pH valuewithout changing the concentration of another dye. Theadsorption capacity of the rGO−CNT−PPD hybrid displayedsignificantly higher values for the mixture compared to earlierreported values of a single dye adsorption.42,43 Table 3 shows

the comparisons of maximum adsorption capacities of therGO−CNT−PPD hybrid obtained in this study with variousadsorbents previously reported for the adsorption of MV andMO.

5. CONCLUSIONGO−CNT hybrid was successfully functionalized withsimultaneous reduction of the hybrid with PPD. Its adsorptioncapacity was investigated systematically toward removal of MVand MO selectively from their aqueous mixture. The adsorptioncapacity was highly dependent on the pH and temperature ofthe solution. At pH 7.0, due to the electrostatic interactionbetween the nitrogen atoms of amine functional groups and theunreduced oxygen-containing functional groups of adsorbentwith the cationic dye, rGO−CNT−PPD hybrid can remove99.7% MV selectively from the mixture of MV−MO solution,whereas at pH 3.0, it can remove MO selectively having 98%removal efficiency due to electrostatic interactions between theprotonated amine functional group of PPD and the anionic dye

MO. In view of the easy method of the preparation of theadsorbent, its reusability, and its excellent selectivity toward dyeadsorption, this hybrid is an important tool in remediationtechnology.

■ ASSOCIATED CONTENT*S Supporting InformationSteps involved in the synthesis of rGO−CNT−PPD hybrid(Scheme S1), pseudo-first-order and pseudo-second-orderkinetic plots at pH 7.0 and 3.0 (Figure S1), and Langmuirand Freundlich adsorption isotherms at pH 7.0 and 3.0 (FigureS2). This material is available free of charge via the Internet athttp://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*(S .K .D.) Phone/fax : +9103712267006 . E -mai l :chandrama006@gmail.com.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSWe thank the Council of Scientific and Industrial Research(CSIR) India for their financial support of the research underContract 02(0066)/12/EMR-II. We gratefully acknowledgefinancial support received from UGC under the SAP programand DST under the FIST program.

■ REFERENCES(1) Azim, W.; Sani, R. K.; Banerjee, U. C. Biodegradation oftriphenylmethane dyes. Enzyme Microb. Technol. 1998, 22, 185−191.(2) Arami, M.; Limaee, N. Y.; Mahmoodi, N. M.; Tabrizi, N. S.Equilibrium and kinetic studies for the adsorption of direct and aciddyes from aqueous solution by soy meal hull. J. Hazard. Mater. 2006,135, 171−179.(3) Vandevivere, P.; Bianchi, C.; Verstaete, R. W. Treatment andreuse of wastewater from the textile wet processing industry: review ofemerging technologies. J. Chem. Technol. Biotechnol. 1998, 72, 289−302.(4) Zhang, Y.; Causserand, C.; Aimar, P.; Cravedi, J. P. Removal ofbisphenol A by a nanofiltration membrane in view of drinking waterproduction. Water Res. 2006, 40, 3793−3799.(5) Wang, R.; Ren, D.; Xia, S.; Zhang, Y.; Zhao, J. Photocatalyticdegradation of bisphenol A (BPA) using immobilized TiO2 and UVillumination in a horizontal circulating bed photocatalytic reactor(HCBPR). J. Hazard. Mater. 2009, 169, 926−932.(6) Selcuk, H. Decolorization and detoxification of textile wastewaterby ozonation and coagulation process. Dyes Pigm. 2005, 64, 217−222.(7) Dutta, K.; Mukhopadhyay, S.; Bhattacharjee, S.; Chaudhuri, B.Chemical oxidation of methylene blue using a Fenton-like reaction. J.Hazard. Mater. 2001, 84, 57−71.(8) Hasan, M. M.; Hawkyard, C. J. Reuse of spent dye bath followingdecolorization with ozone. Colour Technol. 2003, 118, 104−111.(9) Liu, C.-H.; Wu, J.-S.; Chiu, H.-C.; Suen, S.-Y.; Chu, K. H.Removal of anionic reactive dyes water using anion exchangemembranes as absorbers. Water Res. 2007, 41, 1491−1500.(10) Jorgenson, S. E. Examination of the applicability of cellulose ionexchangers for water and wastewater treatment. Water Res. 1979, 13,1239−1247.(11) Blackburn, S. R. Natural polysaccharides and their interactionswith dye molecules: applications in effluent treatment. Environ. Sci.Technol. 2004, 38, 4905−4909.(12) Liu, G. F.; Ma, J.; Li, X. C.; Qin, Q. D. Adsorption of bisphenolA from aqueous solution onto activated carbons with differentmodification treatments. J. Hazard. Mater. 2009, 164, 1275−1280.

Figure 7. Removal efficiency of MV at pH 7.0 on rGO−CNT−PPDhybrid in different cycles.

Table 3. Comparison of Adsorption Capacities of MV andMO onto Various Adsorbents

dye adsorbentadsorption

capacity (mg/g) ref

MV GO−CNT−PPD hybrid 298 this workCNT chemically bonded to rGO 245 42magnetic modified MWCNT 228 44phosphoric acid activated carbon 60.42 45sulfuric acid activated carbon 85.84 45granular activated carbon 95 46

MO GO−CNT−PPD hybrid 294 this workmesoporous carbon materials 294.1 47activated carbon modified by silvernanoparticles

27.48 43

Industrial & Engineering Chemistry Research Article

dx.doi.org/10.1021/ie502653t | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXXG

(13) Kuo, C. Y. Comparison with as-grown and microwave modifiedcarbon nanotubes to removal aqueous bisphenol A. Desalination 2009,249, 976−982.(14) Asada, T.; Oikawa, K.; Kawata, K.; Ishihara, S.; Iyobe, T.;Yamada, A. Study of removal effect of bisphenol A and β-estradiol byporous carbon. J. Health Sci. 2004, 50, 588−593.(15) Ramesha, G. K.; Kumara, A. V.; Muralidhara, H. B.; Sampath, S.Graphene and graphene oxide as effective adsorbents toward anionicand cationic dyes. J. Colloid Interface Sci. 2011, 361, 270−277.(16) Schlierf, A.; Yang, H.; Gebremedhn, E.; Treossi, E.; Ortolani, L.;Chen, L.; Minoia, A.; Morandi, V.; Samor, P.; Casiraghi, C.; Beljonne,D.; Palermo, V. Nanoscale insight into the exfoliation mechanism ofgraphene with organic dyes: effect of charge, dipole and molecularstructure. Nanoscale 2013, 5, 4205−4216.(17) Xu, J.; Wang, L.; Zhu, Y. Decontamination of bisphenol A fromaqueous solution by graphene adsorption. Langmuir 2012, 28, 8418−8425.(18) Zhao, J.; Ren, W.; Cheng, H.-M. Graphene sponge for efficientand repeatable adsorption and desorption of water contaminations. J.Mater. Chem. 2012, 22, 20197−20202.(19) Fletcher, A. J.; Yuzak, Y.; Thomas, K. M. Adsorption anddesorption kinetics for hydrophilic and hydrophobic vapors onactivated carbon. Carbon 2006, 44, 989−1004.(20) Ghaedi, M.; Hassanzadeh, A.; Kokhdan, S. N. Multiwalledcarbon nanotube as adsorbents for the kinetic and equilibrium study ofthe removal of Alizarin Red S and morin. J. Chem. Eng. Data 2011, 56,2511−2520.(21) Machado, F. M.; Bergmann, C. P.; Lima, E. C.; Royer, B.; Souza,F. E.; de Jouris, I. M.; Calvete, T.; Fagan, S. B. Adsorption of ReactiveBlue 4 dye from water solutions by carbon nanotubes: experimentsand theory. Phys. Chem. Chem. Phys. 2012, 14, 11139−11153.(22) Wang, S.; Ng, C. W.; Wang, W.; Li, Q.; Li, L. A comparativestudy on the adsorption of acid and reactive dyes on multiwall carbonnanotubes in single and binary dye systems. J. Chem. Eng. Data 2012,57, 1563−1569.(23) Bradder, P.; Ling, S. K.; Wang, S.; Liu, S. Dye adsorption onlayered graphite oxide. J. Chem. Eng. Data 2011, 56, 138−141.(24) Li, D.; Muller, M. B.; Gilje, S.; Kaner, R. B.; Wallance, G. G.Processable aqueous dispersions of graphene nanosheets. Nat.Nanotechnol. 2008, 3, 101−105.(25) Sharma, P.; Das, M. R. Removal of a cationic dye from aqueoussolution using graphene oxide nanosheets: investigation of adsorptionparameters. J. Chem. Eng. Data 2013, 58, 151−158.(26) Ho, K. Y.; McKay, G.; Yeung, K. L. Selective adsorbents fromordered mesoporous silica. Langmuir 2003, 19, 3019−3024.(27) Cai, D.; Song, M.; Xu, C. Highly conductive carbon nanotube/graphite-oxide hybrid films. Adv. Mater. 2008, 20, 1706−1709.(28) Aboutalebi, S. H.; Chidembo, A. T.; Salari, M.; Konstantinov,K.; Waxler, D.; Liu, H. K.; Dou, S. H. Comparison of GO, GO/MWCNTs composites and MWCNTs as potential electrode materialsfor supercapacitors. Energy Envion. Sci. 2011, 4, 1855−1865.(29) Kim, Y.-K.; Kwack, S.-J.; Ryoo, S.-R.; Lee, Y.; Hong, S.; Hong,S.; Jeong, J.; Ryoo, S.-R.; Lee, Y.; Hong, S.; Hong, S.; Jeong, Y.; Min,D.-H. Synergistic effect of graphene oxide/MWCNT in laserdesorption/ ionization mass spectroscopy of small molecules andtissue imaging. ACS Nano 2011, 5, 4550−4561.(30) Wang, R.; Sun, J.; Gao, L.; Xu, C.; Zhang, J.; Liu, Y. Effectivepost treatment for preparing highly conducting carbon nanotube/reduced graphene oxide films. Nanoscale 2011, 3, 904−906.(31) Huang, J. D.; Zhang, B.; Oh, S.-W.; Zheng, Q.-B.; Lin, X.-Y.;Yousefi, N.; Kim, J.-K. Self-assembled reduced graphene oxide/carbonnanotube thin films as electrodes for supercapacitors. J. Mater. Chem.2012, 22, 3591−3599.(32) Kemp, K. C.; Seema, H.; Saleh, M.; Le, N. H.; Mahesh, K.;Chandra, V.; Kim, K. S. Environmental applications using graphenecomposites: water remediation and gas adsorption. Nanoscale 2013, 5,3149−3171.(33) Stankovich, S.; Dikin, D. A.; Piner, R. D.; Kohlhaas, K. A.;Kleinhammes, A.; Jia, Y.; Yue, W.; Nguyen, S. T.; Ruoff, R. S. Synthesis

of graphene-based nanosheets via chemical reduction of exfoliatedgraphite oxide. Carbon 2007, 45, 1558−1565.(34) Shin, H. J.; Kim, K. K.; Benayad, A.; Yoon, S. M.; Park, H. K.;Jung, I. S.; Jin, M. H.; Jeong, H. K.; Kim, J. M.; Choi, J. Y.; Lee, Y. H.Efficient reduction of graphite oxide by sodium borohydride and itseffect on electrical conductance. Adv. Funct. Mater. 2009, 19, 1987−1992.(35) Kim, N. H.; Kulia, T.; Lee, J. H. Simultaneous reduction,functionalisation and stitching of graphene oxide with ethylenediaminefor composites application. J. Mater. Chem. A 2003, 1, 1349−1358.(36) Wang, B.; Luo, B.; Liang, M.; Wang, A.; Wang, J.; Fang, Y.;Chang, Y.; Zhi, L. Chemical amination of graphene oxides and theirextraordinary properties in the detection if lead ions. Nanoscale 2011,3, 5059−5066.(37) Hummers, W. S.; Offeman, R. E. Preparation of graphitic oxide.J. Am. Chem. Soc. 1958, 80, 1339−1339.(38) Ho, Y. S.; McKay, G. Sorption of dye from aqueous solution bypeat. Chem. Eng. J. 1998, 70, 115−124.(39) Weber, W. J.; Morris, J. C. Kinetics of adsorption on carbonfrom solution. J. Sanit. Eng. Div. 1963, 89, 31−60.(40) Langmuir, I. The constitution and fundamental properties ofsolids and liquids. Part I. Solids. J. Am. Chem. Soc. 1916, 38, 2221−2295.(41) Freundlich, H. Concerning adsorption in solutions. J. Phys.Electrochem. 1906, 57, 385−470.(42) Kotal, M.; Bhowmick, A. K. Multifunctional hybrid materialsbased on carbon nanotube chemically bonded to reduced grapheneoxide. J. Phys. Chem. C 2013, 117, 25865−25875.(43) Pal, J.; Deb, M. K.; Deshmukh, D. K.; Verma, D. Removal ofmethyl orange by activated carbon modified by silver nanoparticles.Appl. Water Sci. 2013, 3, 367−374.(44) Madrakian, T.; Afkhami, A.; Ahmadi, M.; Bagheri, H. Removalof some cationic dyes from aqueous solution using magnetic modifiedMWCNT. J. Hazard. Mater. 2011, 196, 109−114.(45) Senthilkumaar, S.; Kalaamani, P.; Subburaam, C. V. Liquidphase adsorption of crystal violet onto activated carbon derived frommale flowers of coconut tree. J. Hazard. Mater. 2006, 136, 800−808.(46) Azizian, S.; Haerifar, M.; Bashiri, H. Adsorption of methyl violetonto granular activated carbon: equilibrium, kinetics and modelling.Chem. Eng. J. 2009, 46, 36−41.(47) Mohammadi, N.; Khani, H.; Gupta, V. K.; Amereh, E.; Agarwal,S. Adsorption process of methyl orange dye onto mesoporous carbonmaterial − kinetic and thermodynamic studies. J. Colloid Interface Sci.2011, 362, 457−462.

Industrial & Engineering Chemistry Research Article

dx.doi.org/10.1021/ie502653t | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXXH