Chemical Engineering Journal 256 (2014) 128–136

Contents lists available at ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier .com/locate /ce j

Removal performance and the underlying mechanismsof plasma-induced CD/MWCNT/iron oxides towards Ni(II)

http://dx.doi.org/10.1016/j.cej.2014.06.1091385-8947/� 2014 Elsevier B.V. All rights reserved.

⇑ Corresponding author. Tel./fax: +86 512 65883945.E-mail address: shitongyang@suda.edu.cn (S. Yang).

Ping Dong a, Xilin Wu b, Zhuyou Sun c, Jun Hu b, Shitong Yang a,⇑a School for Radiological and Interdisciplinary Sciences, Soochow University, 215123 Suzhou, PR Chinab Key Laboratory of Novel Thin Film Solar Cells, Institute of Plasma Physics, Chinese Academy of Sciences, P.O. Box 1126, 230031 Hefei, PR Chinac Key Lab of Ministry of Education of Coast and Island Development, Nanjing University, 210093 Nanjing, PR China

h i g h l i g h t s

�Magnetic CD/MWCNT/iron oxides were synthesized by low-temperature plasma technique.� CD/MWCNT/iron oxides exhibited high physicochemical stability in aqueous solution.� CD/MWCNT/iron oxides exhibited favorable removal performance towards Ni(II) in the simulated effluent.� XAFS analysis suggested that Ni(II) was retained on the surface-coated b-CD and the inner iron oxides.� CD/MWCNT/iron oxides can be repeatedly used for cost-effective purification of Ni(II)-bearing wastewater.

a r t i c l e i n f o

Article history:Received 9 May 2014Received in revised form 24 June 2014Accepted 25 June 2014Available online 3 July 2014

Keywords:Low-temperature plasmaMagnetic CD/MWCNT/iron oxidesNi(II)XAFS analysisSimulated effluent

a b s t r a c t

Herein, a novel low-temperature plasma technique was adopted to graft b-cyclodextrin (b-CD) on thesurfaces of magnetic MWCNT/iron oxide particles. The as-prepared CD/MWCNT/iron oxides exhibitedhigh saturation magnetization and good physicochemical stability in solution. Batch experiments andX-ray absorption fine structure (XAFS) spectral technique were combined to verify the removal perfor-mance and the underlying mechanisms of CD/MWCNT/iron oxides towards Ni(II) from single-solute sys-tem and the simulated Ni(II)-bearing effluent. The sorption kinetics of Ni(II) on CD/MWCNT/iron oxidescan achieve equilibrium in a time period of 4 h. The surface-coated b-CD improves the dispersion prop-erty of CD/MWCNT/iron oxides and therefore enhances its removal performance towards Ni(II). The max-imum sorption capacity of Ni(II) on CD/MWCNT/iron oxides is higher than a series of adsorbent materials.XAFS analysis suggests that Ni(II) can bind on the hydroxyl sites on the surface-coated b-CD and also theFeO6 octahedra of iron oxides in an edge-shared mode, forming strong inner-sphere complexes with highthermodynamic stability. Considering its high physicochemical stability, high removal performance, highseparation convenience and favorable regeneration property, the prepared CD/MWCNT/iron oxides cansupport long-term use as a cost-effective material in the purification of Ni(II)-bearing effluents.

� 2014 Elsevier B.V. All rights reserved.

1. Introduction hypertension, anemia, muscle cramp, diarrhea and nephritic syn-

Nowadays, the continuous increase of various industrial andagricultural activities discharges a series of inorganic and organicpollutants into the soil and water media. Owing to their non-biodegradable property and biological accumulation effect, thetoxic heavy metal ions exhibit adverse, acute and even fatalimpacts on ecological safety, aquatic organisms and human healthover a long period [1,2]. For instance, the excess intake of heavymetal nickel (Ni(II)) would cause serious diseases such as hepatitis,

drome [3,4]. In view of this, advanced techniques and adsorbentmaterials are badly in need for the decontamination of Ni(II) fromwastewaters.

Among the current methods for sewage treatment, sorptionapproach has been widely adopted due to its easy handing, lowcost and high efficiency. A series of methods including hydrother-mal route, chemical co-precipitation, chemical vapor compositionand electrochemical composition have been adopted to synthesizevarious adsorbents. The derived composites exhibit favorableremoval performance towards various heavy metal ions in theaquatic systems [5–9]. In recent years, magnetic compositeshave been applied in sewage treatment due to their high sorptioncapacity and high separation convenience [10–12].

P. Dong et al. / Chemical Engineering Journal 256 (2014) 128–136 129

Plasma technique is a novel method to active the substratematerial surfaces without altering their bulk properties [13–15].The activated substrates with high reactivity can be easily deco-rated by a variety of functionalized polymers and further appliedin various research fields. In the present study, magneticMWCNTs/iron oxides were firstly prepared via the co-precipitationapproach. Afterwards, b-cyclodextrin (b-CD) macromolecules weregrafted on the surfaces of as-prepared MWCNTs/iron oxides with anovel low-temperature plasma technique. The physicochemicalstability of the prepared CD/MWCNT/iron oxides was tested bymeasuring the amount of Fe leaching into the solution. Batchexperiments were conducted to evaluate the removal performanceof CD/MWCNT/iron oxides towards heavy metal Ni(II) as a functionof various environmental factors. The regeneration property of CD/MWCNT/iron oxides towards Ni(II) removal was tested by multiplesorption/desorption cycles. In addition, X-ray absorption finestructure (XAFS) spectral technique was adopted to identify theunderlying mechanisms of Ni(II) sorption on CD/MWCNT/iron oxi-des. Finally, the environmental significance of the present researchfindings was further predicted on the basis of the macroscopic andmicroscopic results.

2. Materials and methods

2.1. Materials

The b-cyclodextrin (b-CD), Ni(NO3)2�6H2O, FeCl2�4H2O andFeCl3�6H2O were purchased in analytical pure from SinopharmChemical Reagent Co. Ltd. (China). The analytical pureNi(NO3)2�6H2O was dissolved in deionized water to obtain theNi(II) stock solution (100 mg/L). The prepared Ni(II) stock solutionwere diluted to achieve the desired concentrations in the followingsorption experiments.

2.2. Synthesis and characterization of CD/MWCNT/iron oxides

The procedures for fabricating the CD/MWCNT/iron oxides con-sisted the preparation of MWCNT/iron oxides, the activation ofMWCNT/iron oxide surfaces by using low-temperature plasmaand the grafting of b-CD on surface-activated MWCNT/iron oxides.The specific surface area of CD/MWCNT/iron oxides was measuredto be 63.9 m2/g by using the N2-BET method. The previous TGAanalysis result showed that the grafted amount of b-CD onMWCNT/iron oxides was 16.6 mg/g [16]. The saturationmagnetization (Ms) of CD/MWCNT/iron oxides was measured tobe 37.8 emu/g. The XRD analysis showed that the chemical co-precipitation procedure, plasma activation and the subsequentdecoration of b-CD did not disorganize the basic structure ofMWCNTs [17].

2.3. Material stability test

The prepared MWCNT/iron oxides and CD/MWCNT/iron oxideswere suspended in a series of solutions with pH 2.0–10.0 to testtheir physicochemical stability. The solid and aqueous phases wereseparated by using an external magnet after oscillating thesuspensions for a contact time of 24 h. The dissolved Fe contentsat different pH values were determined by using the atomicabsorption spectrometry.

2.4. Sorption experiments

All the experiments were performed in 10-mL polyethylenecentrifuge tubes by using batch technique. Briefly, the CD/MWCNT/iron oxide suspension, NaNO3 electrolyte solution and

Ni(II) stock solution were added to attain the required concentra-tions of each constituent. The suspension pH was regulated tothe desired values by adding tiny amounts of 0.01 mol/L HNO3 orNaOH solutions. The centrifuge tubes were gently oscillated for24 h to achieve sorption equilibrium and then the solid andaqueous phases were separated with a permanent magnet. Theconcentration of Ni(II) in the supernatant was measured by usingatomic absorption spectrophotometry. The sorption percentage(S% = ((C0 � Ce)/C0�100%) and sorption amount (qe = ((C0 � Ce)�V/m)) were derived from the initial Ni(II) concentration (C0), the finalNi(II) concentration (Ce), the mass of CD/MWCNT/iron oxides (m)and the suspension volume (V).

2.5. XAFS data collection and analysis

The sorption sample for XAFS analysis was prepared in 250 mLvessel by using a similar approach as the batch experiments.Briefly, the Ni(II) stock solution was tardily added in 10–50 lLincrements into the CD/MWCNT/iron oxide suspension under con-stant stirring. The mixture was held steadily at pH 6.5 and shakenon a rotating oscillator for 24 h. The solid and aqueous phases wereseparated by using an external magnet. The collected wet pasteswere then wrapped in a moist paper towel and sealed in a Ziplockbag for the subsequent XAFS data collection.

The Ni K-edge XAFS spectra were collected at the BL14W1 beamline in Shanghai Synchrotron Radiation Facility (SSRF, China) influorescence mode for Ni(NO3)2(aq) reference sample and the sorp-tion sample and in transmission mode for Ni(OH)2(s) referencesample. The electron beam energy was 3.5 GeV and the meanstored current was 300 mA. The energy of the X-ray was tunedby using a double-crystal Si (111) monochromator. The XAFS sig-nals of samples were collected with a multi-element high purity Gesolid-state detector. Each XAFS spectrum was collected in tripleand averaged to improve the signal to noise ratio. The energy cor-rection, fluorescence dead time calibration and the subsequentXAFS data fitting were performed by using the Athena and Artemissoftware. During fitting, the amplitude reduction factor (S0

2) wasfixed to 0.85 for all the samples. The total number of variableparameters did not exceed the maximum limit given by Stern’srule. The accuracies of R and CN values for the first coordinationshells were 0.02 Å and 20%, and those for the second coordinationshells were 0.03 Å and 40%.

3. Results and discussion

3.1. Material stability

Fig. 1 illustrates the effect of solution pH on the content of Feleaching from the MWCNT/iron oxides and CD/MWCNT/iron oxi-des. It is clear that the Fe leaching content decreases with increas-ing pH values for the two Fe-containing samples. The content of Feleaching from the CD/MWCNT/iron oxides is much lower than thatof Fe leaching from the MWCNT/iron oxides in the acidic pH range,which suggests that the surface-grafted b-CD prevents the dissolu-tion of Fe from the oxidized MWCNT surfaces and thereforeimproves the physicochemical stability of CD/MWCNT/iron oxidesin acidic solution. The high stability supports the use of CD/MWCNT/iron oxides as a potential material for the purification ofNi(II)-bearing wastewater.

3.2. Effect of contact time

Fig. 2 shows the kinetic data of Ni(II) sorption on CD/MWCNT/iron oxides as a function of initial Ni(II) concentration. The sorptionamount of Ni(II) increases rapidly during the initial contact time of

Fig. 1. Relative proportion of Fe leaching from MWCNT/iron oxides and CD/MWCNT/iron oxides at various solution pH values. T = 298 ± 1 K, m/V = 0.5 g/L,I = 0.01 mol/L NaNO3.

130 P. Dong et al. / Chemical Engineering Journal 256 (2014) 128–136

1.5 h and thereafter it proceeds at a slower rising rate and finallymaintains changeless after a contact time of �4 h. Note that thetransitory time for separating the CD/MWCNT/iron oxides fromthe aqueous solution with an external magnet is ignored when cal-culating the total contact time. The rapid sorption dynamics proce-dure during the initial contact time may result from the fastmigration of Ni(II) from the aqueous solution onto the externalbinding sites on CD/MWCNT/iron oxide surfaces. With the gradualoccupying of surface sites, the adsorbed Ni(II) is expected to trans-port from the external surfaces to the internal location (i.e., inner-sphere sites of CD/MWCNT/iron oxides) [18]. The slow diffusionprocess causes the decrease of sorption kinetics rate with increas-ing contact time. One can see from Fig. 2 that the sorption mount ofNi(II) on CD/MWCNT/iron oxides increases from 7.7 to 13.1 mg/gas the initial Ni(II) concentration increases from 5 to 20 mg/L.The higher sorption amount at higher Ni(II) concentration is attrib-uted to the greater concentration gradient. On the whole, a contacttime of �4 h is enough to achieve the complete sorption equilib-rium. Such a short time period herein suggests that CD/MWCNT/iron oxides can potentially support the continuous treatment ofNi(II)-bearing wastewater. According to the foresaid experimental

Fig. 2. Effect of contact time and initial concentration on Ni(II) sorption on CD/MWCNT/iron oxides. T = 298 ± 1 K, pH = 6.5 ± 0.1, m/V = 0.5 g/L, I = 0.01 mol/LNaNO3.

kinetics data, the oscillating time in the following experiments wasselected as 24 h to ensure complete sorption equilibrium. The fastsorption dynamics process points to the occurrence of chemicalsorption rather than physical sorption [19,20]. This deduction isfurther proved by the absence of Ni(II) in the solution after sonica-tion of the Ni(II)-adsorbed CD/MWCNT/iron oxide suspension. ThepH variation during the sorption process can be used to identifythe sorption mechanism of Ni(II) on CD/MWCNT/iron oxides.According to our measurements, the finial pH after the sorptionprocedure was lower than the initial pH. Herein, the decrease ofpH may be attributed to the release of H+ from the CD/MWCNT/iron oxide surfaces due to the deprotonation reaction, i.e.,„SOH M „SO� + H+. Hence, Ni(II) can bind on the deprotonatedsurface sites via covalent bonds, leading to the formation ofinner-sphere complexes.

3.3. Effect of solid content

Fig. 3 illustrates the effect of solid content on the sorption per-centage and sorption amount of Ni(II) on CD/MWCNT/iron oxides.It is clear that the sorption percentage increases from �30% to�75% as the solid content (m/V) increases from 0.1 to 0.8 g/L.The total number of available binding sites on CD/MWCNT/ironoxide surfaces is in line with the applied solid content. The highernumber of surface sites for binding Ni(II) results in the increasedsorption percentage at higher solid dosage. Note that the removalefficiency of CD/MWCNT/iron oxides towards Ni(II) maintainsunchanged at m/V > 0.8 g/L. From the aspect of reducing effluentpurification cost, the optimal content for CD/MWCNT/iron oxidesto decontaminate Ni(II) at pH 6.5 is 0.8 g/L.

As shown in Fig. 3, the sorption amount of Ni(II) decreases grad-ually with rising CD/MWCNT/iron oxide dosage. At lower solid dos-age, the CD/MWCNT/iron oxide particles disperse well in thesolution. Hence, all of the surface sites are completely exposedfor Ni(II) binding, leading to a higher Ni(II) sorption amount. Incontrast, the CD/MWCNT/iron oxide particles tend to collide witheach other at higher solid dosage, causing the formation of aggre-gation and the decrease of their dispersion in solution. This phe-nomenon would decrease the total specific surface area and alsoincrease the diffusional path length, which are expected to reducethe availability of binding sites and further decrease the sorptionamount of Ni(II) on CD/MWCNT/iron oxides [21,22]. In addition,the collision between individual solid particles may desorb someweak-linked Ni(II) from the CD/MWCNT/iron oxide surfaces, which

Fig. 3. Effect of solid content on the sorption percentage and sorption amount ofNi(II) on CD/MWCNT/iron oxides. T = 298 ± 1 K, pH = 6.5 ± 0.1, CNi(II)initial = 10 mg/L,I = 0.01 mol/L NaNO3.

P. Dong et al. / Chemical Engineering Journal 256 (2014) 128–136 131

consequently decreases the sorption amount of Ni(II) at highersolid dosage.

The sorption percentages of Ni(II) on iron oxides, oxidizedMWCNTs, MWCNT/iron oxides and CD/MWCNT/iron oxides werecompared to verify the contribution of individual components(i.e., oxidized MWCNTs, iron oxides and the surface-coated CD)on Ni(II) removal. As shown in Fig. 4, the sorption percentage ofNi(II) on CD/MWCNT/iron oxides is higher than that on iron oxides,MWCNT/iron oxides, and oxidized MWCNTs. The specific surfaceareas of various solid materials can partly reflect their removal per-formance towards heavy metal ions. Generally, solid material withgreater specific surface area is expected to exhibit higher removalperformance. The specific surface areas of oxidized MWCNTs,MWCNT/iron oxides and CD/MWCNT/iron oxides are measuredto be 197.0, 88.5 and 63.9 m2/g, respectively [16]. In theory, CD/MWCNT/iron oxides would possess the lowest removal perfor-mance towards Ni(II) due to its lowest specific surface area. How-ever, our sorption experiments herein show an opposite result.This inconformity suggests that the removal sequence of the threematerials towards Ni(II) is not attributed to their discrepancy inspecific surface areas. Alternatively, the surface-grafted b-CD isresponsible for the enhanced sorption amount of CD/MWCNT/ironoxides towards Ni(II). The b-CD molecule possesses a characteristictruncated-cone structure that consists of primary hydroxyl siteslying outside the apolar cylindrical cavity and secondary hydroxylsites lying inside the apolar cylindrical cavity [12]. These hydro-philic hydroxyl sites can improve the dispersion of CD/MWCNT/iron oxides in solution. Consequently, the surface hydroxyl sitesof CD/MWCNT/iron oxides are more available for binding Ni(II),which contributes to the enhanced removal performance of CD/MWCNT/iron oxides towards Ni(II).

It is worth noting that the coverage of iron oxides on MWCNTsurfaces improves the separation performance of the preparedMWCNT/iron oxides. Despite the removal performance ofMWCNT/iron oxides is slightly lower than that of oxidizedMWCNTs (Fig. 4), the magnetic MWCNT/iron oxides can be easilyseparated by using an external magnet. The surface-coated b-CDcan compensate the slightly decreased removal performance ofMWCNT/iron oxides. As shown in Fig. 4, the prepared CD/MWCNT/iron oxides exhibits favorable removal performancetowards heavy metal Ni(II). In addition, the surface-coated b-CDcan improve the dispersion and physicochemical stability of CD/MWCNT/iron oxides in solution. Moreover, the high separation

Fig. 4. Sorption percentages of Ni(II) on iron oxides, oxidized MWCNTs, MWCNT/iron oxides and CD/MWCNT/iron oxides as a function of solid content. T = 298 ± 1 K,pH = 6.5 ± 0.1, CNi(II)initial = 10 mg/L, I = 0.01 mol/L NaNO3.

convenience is expected to reduce the cost of using CD/MWCNT/iron oxides in real effluent treatment.

3.4. Sorption and desorption isotherms

Fig. 5 shows the sorption and desorption isotherms of Ni(II) onCD/MWCNT/iron oxides. It is clear that the sorption isothermreveals the typical L shape with an asymptote at higher equilib-rium concentration. This sorption isotherm curve is expected dueto the finite specific surface area of CD/MWCNT/iron oxides at aspecific dosage. In addition, the L-type sorption isotherm elimi-nates the formation of precipitates, which is expected to induce acontinuous increase of sorption amount with rising equilibriumconcentration. As shown in Fig. 5, the curve of desorption isothermcurve is obviously higher than that of sorption isotherm, whichpoints to the presence of sorption–desorption hysteresis due toan irreversible binding. In other words, the sorption of Ni(II) onCD/MWCNT/iron oxides is driven by inner-sphere complexationrather than electrostatic interaction/physical sorption [23–25]. Tofurther verify the underlying sorption mechanism, the sorptionisotherm experiments were performed at different levels of ionicstrength. The experimental results show that the variation of ionicstrength has negligible influence on the sorption isotherm shapeand the sorption capacity of Ni(II) (data not shown). This phenom-enon suggests that Ni(II) is retained on CD/MWCNT/iron oxides viainner-sphere complexation. The deduction herein is further sup-ported by the decrease of solution pH after the sorption procedure.

Herein, the Langmuir qe ¼ bqmaxCe1þbCe

� �and Freundlich qe ¼ KFCn

e

� �

models were adopted to simulate the sorption and desorption iso-therms (Fig. 5). In the equation, Ce is the residual concentration ofNi(II) in solution (mg/L); qe is the sorption amount of Ni(II) on CD/MWCNT/iron oxides after equilibrium (mg/g); qmax is the maxi-mum sorption amount of Ni(II) on CD/MWCNT/iron oxides at com-plete monolayer coverage (mg/g); b (L/mg) is a parameter thatrelates to the sorption heat; KF is the sorption capacity when theNi(II) equilibrium concentration equals to 1 (mg1 � n Ln/g) and nrepresents the degree of sorption dependence at equilibrium con-centration. The parameters calculated from the model fitting arelisted in Table 1. One can see from the R2 values that Langmuirmodel simulates the sorption and desorption isotherms better thanFreundlich model. This experimental phenomenon suggests a uni-form binding energy for the entire sites on CD/MWCNT/iron oxide

Fig. 5. Langmuir and Freundlich model fitting for the sorption and desorptionisotherms of Ni(II) on CD/MWCNT/iron oxides. T = 298 ± 1 K, pH = 6.5 ± 0.1,m/V = 0.5 g/L, I = 0.01 mol/L NaNO3. The solid lines represent the Langmuir modelfitting curves and the dash lines represent the Freundlich model fitting curves.

Table 1The derived parameters from Langmuir and Freundlich model fitting.

Conditions Langmuir model Freundlich model

qmax (mg/g)

b (L/mg)

R2 KF (mg1�n Ln/g)

n R2

Sorption 38.24 0.072 0.991 4.824 0.568 0.938Desorption 47.62 0.194 0.992 10.368 0.463 0.980

Fig. 6. Recycling of CD/MWCNT/iron oxides in the removal of Ni(II) from aqueoussolution. T = 298 ± 1 K, pH = 6.5 ± 0.1, m/V = 0.5 g/L, I = 0.01 mol/L NaNO3.

132 P. Dong et al. / Chemical Engineering Journal 256 (2014) 128–136

surfaces and the formation of monolayer Ni(II) coverage. In otherwords, the surface-adsorbed Ni(II) do not interact with each other,ruling out the occurrence of surface polymerization and co(precip-itation). It is worth noting that CD/MWCNT/iron oxides have afinite specific surface area and limited sorption capacity. Accord-ingly, the sorption and desorption isotherms of Ni(II) could be bet-ter simulated by Langmuir model rather than Freundlich model, asan exponential increase of sorption amount with increasing equi-librium concentration is assumed in Freundlich model. Herein,the accordance of the experimental data with the Langmuir modelalso indicates that chemosorption is the main driving force forNi(II) sorption on CD/MWCNT/iron oxides [22,26]. The n valuesderived from the Freundlich model are from unity, indicating anonlinear sorption of Ni(II) on CD/MWCNT/iron oxide surfaces.

Herein, the maximum sorption capacity (i.e., qmax derived fromthe Langmuir model fitting) of CD/MWCNT/iron oxides towardsNi(II) was carefully compared with other adsorbents to further ver-ify the potential of using this material in effluent purification. Asshown in Table 2, the qmax value of Ni(II) on CD/MWCNT/ironoxides is higher than a series of adsorbent materials such asTBA-kaolinite [27], CMC-bentonite [28], PAA-MWCNTs [29],Na-rectorite [30], NH2-MCM-41 [31], chitosan-BPMAMF [32],chitosan-RB2 [33], coir pith, ZrO-montmorillonite [34] and NaO-Cl-granular activated carbon (NaOCl-GAC) [35], while lower thanthose of NaOCl-MWCNTs and NaOCl-SWCNTs [35]. Herein, thehigher sorption capacities of NaOCl-MWCNTs and NaOCl-SWCNTsmay be partly attributed to the higher solution pH and wider con-centration range used in the sorption isotherm experiments. How-ever, it is rather discommodious to separate these materials fromthe aqueous phase. The conventional ultracentrifugation and filtra-tion methods are unsatisfactory due to the fact that high speedcentrifugation consumes vast electric energy and filtration is proneto filter blockages. The disadvantage in the separation limits theirfurther application potential in effluent purification and contami-nated environment remediation. In contrast, the CD/MWCNT/ironoxides can be easily separated from the aqueous phase by usingan external magnet. In addition, the reusability test shows thatthe sorption amount of Ni(II) on CD/MWCNT/iron oxides keeps

Table 2Comparison of Ni(II) sorption capacity of CD/MWCNT/iron oxides with other sorbents.

Materials Experimentalconditions

qmax

(mg/g)References

TBA-kaolinite pH = 5.7, T = 303 K 2.75 [27]CMC-bentonite pH = 6.0, T = 303 K 2.89 [28]PAA-MWCNTs pH = 5.4, T = 293 K 4.17 [29]Na-rectorite pH = 6.5, T = 308 K 7.18 [30]NH2-MCM-41 pH = 5.0, T = 295 K 8.80 [31]Chitosan-BPMAMF pH = 3.0, T = 298 K 9.60 [32]Chitosan-RB2 pH = 8.5, T = 298 K 11.20 [6]Coir pith pH = 5.3, T = 300 K 15.95 [33]ZrO-montmorillonite pH = 5.8, T = 303 K 22.00 [34]NaOCl-GAC pH = 7.0, T = 298 K 26.39 [35]CD/MWCNT/iron oxides pH = 6.5, T = 298 K 38.24 The present studyNaOCl-MWCNTs pH = 7.0, T = 298 K 38.46 [35]NaOCl-SWCNTs pH = 7.0, T = 298 K 47.85 [35]

almost unchanged after ten cycles of sorption–desorption experi-ments at the initial Ni(II) concentration of 10 mg/L and 20 mg/L(Fig. 6). The results suggest that CD/MWCNT/iron oxides can berepeatedly used for the removal of Ni(II) from aqueous phases.Herein, the high removal performance, high separation conve-nience and excellent regeneration property make CD/MWCNT/ironoxides to be a cost-effective material for sewage disposal withminimum cost, which will further improve the environmentaland economic benefits.

3.5. Simulated Ni(II)-bearing effluent purification

The actual aquatic environment is a complicated and heteroge-neous system due to the simultaneous presence of different com-ponents such as heavy metal ions, organic pollutants, electrolyteions, microorganisms, natural minerals, etc. In view of this, theremoval performance of CD/MWCNT/iron oxides towards a simu-lated effluent was tested so as to verify the application potentialof our research findings in the real water systems. Herein, the sim-ulated effluent contains 10 mg/L of Cd(II), 10 mg/L of Cu(II), 10 mg/L of Ni(II), 6 mg/L of As(V), 20 mg/L of rhodamine B (RB), 50 mg/L

Fig. 7. Removal efficacy of CD/MWCNT/iron oxides towards simulated Ni(II)-bearing effluent. T = 298 ± 1 K, pH = 6.5 ± 0.1, m/V = 0.5 g/L, I = 0.01 mol/L NaNO3.

P. Dong et al. / Chemical Engineering Journal 256 (2014) 128–136 133

1-naphenol, 10 mg/L of humic acid (HA) and 0.01 mol/L NaNO3

electrolyte solution. Typical experiment was carried out by adding200 mL of the simulated effluent into a 500 mL beaker. The subse-quent procedure was performed at pH 6.5 via a similar approach asthat in batch experiments. As shown in Fig. 7, the removal percent-ages of CD/MWCNT/iron oxides towards Cd(II), Cu(II), Ni(II), As(V),RB, 1-naphthol and HA are 46%, 78%, 55%, 41%, 50%, 53% and 34%,respectively. Herein, the sorption percentage of Ni(II) in the simu-lated effluent is slightly lower than that in the single-solute sys-tem, which may be induced by the influence of coexisted

Fig. 8. Normalized K-edge XANES spectra (A) and the corresponding first derivatives (B)CNi(II)initial = 10 mg/L, I = 0.01 mol/L NaNO3.

Fig. 9. The k3-weighted EXAFS spectra and the RSF magnitudes (uncorrected for phexperimental k3-weighted EXAFS spectra; (B) symbols represent the experimental RSFm/V = 0.5 g/L, CNi(II)initial = 10 mg/L, I = 0.01 mol/L NaNO3.

components. Nevertheless, CD/MWCNT/iron oxides still exhibitfavorable removal performance towards heavy metal Ni(II) in thesimulated effluent.

3.6. Sorption mechanisms

To further identify the sorption mechanisms of Ni(II) on CD/MWCNT/iron oxides, the XAFS spectra of Ni(II)-containing refer-ence and sorption sample prepared at pH 6.5 were collected andanalyzed in detail. A weak absorption feature appears before the

of Ni(II) reference and sorption samples. T = 298 ± 1 K, pH = 6.5 ± 0.1, m/V = 0.5 g/L,

ase shift) of Ni(II) reference and sorption samples. (A) Solid lines represent themagnitudes and solid lines represent the spectral fits. T = 298 ± 1 K, pH = 6.5 ± 0.1,

134 P. Dong et al. / Chemical Engineering Journal 256 (2014) 128–136

edge jump of the normalized K-edge XANES spectra (dash line inFig. 8A), which is commonly for the other transition metals suchas Co(II) and Zn(II) [36]. The absence of shoulder structure in thenormalized XANES spectra (Fig. 8A) and the first derivative curves(Fig. 8B) indicates that the adsorbed Ni(II) is present in an octahe-dral coordination environment. The specific binding type and thecorresponding microstructure of Ni(II) on CD/MWCNT/iron oxideswill be identified in the EXAFS analysis procedures to furtherdetermine the underlying sorption mechanisms.

As shown in Fig. 9A, the k3-weighted EXAFS spectrum of theprepared sorption sample is obviously distinct from that of theNi(NO3)2(aq) and Ni(OH)2(s) reference samples, implying thatNi(II) is not present in the fully-hydrated or precipitation form.By contrast, the spectrum exhibits two beak features at �5.3 and�7.5 �1 (marked in dash circles), which suggests the presence

Table 3EXAFS-derived structural parameters for Ni(II) reference and sorption samples.

Sample conditions First shell (Ni–O)

R (Å) CN r2 (Å2)

Ni(NO3)2(aq) 2.05 5.9 0.004Ni(OH)2(s) 2.04 5.8 0.005Sorption sample pH 6.5 2.03 5.9 0.004

R–bond distance, CN–coordination number, r2–Debye–Waller factor, Res–a measure of thin italic were restricted to be constant during the EXAFS data analysis.

Fig. 10. Schematic illustration of the preconcentration and subsequ

of heavy backscattering from higher atomic shells [4,37]. Fouriertransformation was performed to gain the corresponding RSFs forthe typical oscillation features in the k3-weighted EXAFS spectra.As shown in Fig. 9B, the RSFs for all the Ni(II)-containing samplesexhibit a single peak with high intensity at �1.60 Å (uncorrectedfor phase shift), which arises from the signal of O atoms in the firstshell. An additional peak appears in the R range of 2.20–3.20 Å forthe Ni(OH)2(s) reference sample and the sorption sample preparedat pH 6.5, pointing to the presence of higher coordination shells.

The RSFs were fitted by using the least-square approach basedon the theoretically derived phase and amplitude functions andthe obtained structural parameters are listed in Table 3. TheNi(NO3)2(aq) reference sample contains �5.9 O atoms at a Ni–Obond distance (RNi–O) of �2.05 Å, suggesting that the centralNi atom is located in a Ni(II)(H2O)6 octahedral coordination

Second shells (Ni–M) %Res

Bond R (Å) CN r2 (Å2)

4.2NiANi 3.11 6.0 0.007 6.8NiAC 2.85 0.6 0.007 7.9NiAFe 3.21 1.1 0.007

e agreement between the theoretical and experimental EXAFS curves. The r2 values

ent treatment procedures of Ni(II) by CD/MWCNT/iron oxides.

P. Dong et al. / Chemical Engineering Journal 256 (2014) 128–136 135

environment [38,39]. The Ni(OH)2(s) reference sample contains�5.8 O atoms at RNi–O �2.04 Å in the first shell and �6.0 Ni atomsat RNi–Ni �3.11 Å in the second shell. Spectral analysis of thesorption sample prepared at pH 6.5 shows �5.9 O atoms atRNi–O � 2.03 Å in the first shell and �0.6 C at RNi–C � 2.85 Å as wellas �1.1 Fe at RNi–Fe � 3.21 Å in the second coordination shell.Herein, the absence of Ni–Ni bond in the second coordination shelleliminates the occurrence of surface polymerization or co(precipi-tation). The presence of Ni–C bond in the second coordination shellsuggests that Ni(II) is retained on the binding sites of the oxidizedMWCNTs and/or the surface-grafted b-CD molecules. However, thepowder EXAFS technique cannot differentiate these two bindingmodes from each other. Specifically, a CNNi–C value of �0.6 impliesthe formation of 2:1 complexes between Ni(II) and the bindingsites [40]. In addition, the presence of Ni–Fe bond in the secondcoordination shell indicates that some of the Ni(II) tends to directlybind on the FeO6 octahedra of the iron oxides coated on theoxidized MWCNT surfaces. The derived RNi–Fe value of �3.21 Åcorresponds with the edged-shared binding of Ni(II) on the FeO6

octahedra [41]. This binding mode will result in the formation ofstrong inner-sphere complexes with high thermodynamic stability.The underlying sorption mechanisms obtained herein can provideimportant parameter guidance for the preparation of high-efficientadsorbent materials.

4. Environmental significance

The present study provides a novel material synthesis approachfor the removal of Ni(II) from wastewaters. The low-temperatureplasma technique can effectively active the MWCNT/iron oxidesubstrate and further introduce b-CD macromolecules on the acti-vated MWCNT/iron oxide surfaces. The surface-grafted b-CD mac-romolecules greatly improve the dispersion property and also thephysicochemical stability of the prepared CD/MWCNT/iron oxidesin solution. Hence, the surface hydroxyl sites are much more avail-able for binding Ni(II), which contributes to the favorable removalperformance of CD/MWCNT/iron oxides towards Ni(II) in both thesingle-solute system and the simulated effluent. The preparationand the treatment procedures of CD/MWCNT/iron oxides towardsNi(II) are schematically illustrated in Fig. 10. The XAFS-derivedsorption mechanisms can provide important parameter guidancefor the development of new environmental remediation methodand preparation of high-efficient adsorbent materials. Herein, XAFSanalysis results suggest that Ni(II) can bind on the hydroxyl sites ofsurface-coated b-CD molecules. Considering its low-priced, envi-ronmental-friendly and nice biocompatible properties, b-CD canbe grafted on other substrates to form new composites and findtheir application potential in environmental and biomedical fields.

The research findings derived from the combination of batchexperiment and XAFS technique point out that the CD/MWCNT/iron oxides act as a potential material for the purification ofNi(II)-bearing effluent. Note that the levels of environmental fac-tors such as the solution pH, salinity, redox potential and temper-ature are obviously different for multifarious aquatic systems. Inview of this, more studies are required to supplement the currentresearch findings so as to further check the application potential ofCD/MWCNT/iron oxides in the decontamination of Ni(II).Moreover, additional approaches such as the molecule- and ion-imprinted techniques are being considered to improve the selec-tive removal performance of CD/MWCNT/iron oxides towardsNi(II) from large volumes of wastewaters.

Acknowledgements

Financial supports from the National Natural Science Founda-tion of China (No. 41203086), the National Basic Research Program

of China (No. 2011CB933700), the Jiangsu Provincial Key Labora-tory of Radiation Medicine and Protection, the Collaborative Inno-vation Center of Radiation Medicine of Jiangsu Higher EducationInstitutions and the Priority Academic Program Development ofJiangsu Higher Education Institutions are acknowledged.

References

[1] O. Ozay, S. Ekici, Y. Baran, N. Aktas, N. Sahiner, Removal of toxic metal ionswith magnetic hydrogels, Water Res. 43 (2009) 4403–4411.

[2] L. Semerjian, Equilibrium and kinetic of cadmium adsorption from aqueoussolutions using untreated Pinus halepensis sawdust, J. Hazard. Mater. 173(2010) 236–242.

[3] X.S. Wang, J. Huang, H.Q. Hu, J. Wang, Y. Qin, Determination of kinetic andequilibrium parameters of the batch adsorption of Ni(II) from aqueoussolutions by Na-mordenite, J. Hazard. Mater. 142 (2007) 468–476.

[4] S.T. Yang, G.D. Sheng, X.L. Tan, J. Hu, J.Z. Du, G. Montavon, X.K. Wang,Determination of Ni(II) uptake mechanisms on mordenite surfaces: acombined macroscopic and microscopic approach, Geochim. Cosmochim.Acta 75 (2011) 6520–6534.

[5] S.R. Shukla, R.S. Pai, A.D. Shendarkar, Adsorption of Ni(II), Zn(II) and Fe(II) onmodified coir fibers, Sep. Purif. Technol. 47 (2006) 141–147.

[6] H.L. Vasconcelos, V.T. Fávere, N.S. Gonçalves, M.C.M. Laranjeira, Chitosanmodified with Reactive Blue 2 dye on adsorption equilibrium of Cu(II) andNi(II) ions, React. Funct. Polym. 67 (2007) 1052–1060.

[7] Q.H. Fan, D.D. Shao, J. Hu, W.S. Wu, X.K. Wang, Comparison of Ni2+ sorption tobare and ACT-graft attapulgites: effect of pH, temperature and foreign ions,Surf. Sci. 602 (2008) 778–785.

[8] L. Ducoroy, M. Bacquet, B. Martel, M. Morcellet, Removal of heavy metals fromaqueous media by cation exchange nonwoven PET coated with b-cyclodextrin-polycarboxylic moieties, React. Funct. Polym. 68 (2008) 594–600.

[9] S.B. Yang, J. Hu, C.L. Chen, D.D. Shao, X.K. Wang, Mutual effects of Pb(II) andhumic acid adsorption on multiwalled carbon nanotubes/polyacrylamidecomposites from aqueous solutions, Environ. Sci. Technol. 45 (2011) 3621–3627.

[10] C.L. Chen, J. Hu, D.D. Shao, J.X. Li, X.K. Wang, Adsorption behavior of multiwallcarbon nanotube/iron oxide magnetic composites for Ni(II) and Sr(II), J.Hazard. Mater. 164 (2009) 923–928.

[11] B.C. Kim, J. Lee, W. Um, J. Kim, J. Joo, J.H. Lee, J.H. Kwak, J.H. Kim, C. Lee, H. Lee,R.S. Addleman, T. Hyeon, M.B. Gu, J. Kim, Magnetic mesoporous materials forremoval of environmental wastes, J. Hazard. Mater. 192 (2011) 1140–1147.

[12] A.Z.M. Badruddoza, A.S.H. Tay, P.Y. Tan, K. Hidajat, M.S. Uddin, Carboxymethyl-b-cyclodextrin conjugated magnetic nanoparticles as nano-adsorbents forremoval of copper ions: synthesis and adsorption studies, J. Hazard. Mater. 185(2011) 1177–1186.

[13] P. Chevallier, M. Castonguay, S. Turgeon, N. Dubrulle, D. Mantovani, P.H.McBreen, Ammonia RF-plasma on PTFE surfaces: chemical characterization ofthe species created on the surface by vapor-phase chemical derivatization, J.Phys. Chem. B 105 (2001) 12490–12497.

[14] D.D. Shao, Z.Q. Jiang, X.K. Wang, J.X. Li, Y.D. Meng, Plasma induced graftingcarboxymethyl cellulose on multiwalled carbon nanotubes for the removal ofUO2

2+ from aqueous solution, J. Phys. Chem. B 113 (2009) 860–864.[15] D.D. Shao, G.D. Sheng, C.L. Chen, X.K. Wang, M. Nagatsu, Removal of

polychlorinated biphenyls from aqueous solutions using b-cyclodextringrafted multiwalled carbon nanotubes, Chemosphere 79 (2010) 679–685.

[16] J. Hu, D.D. Shao, C.L. Chen, G.D. Sheng, J.X. Li, X.K. Wang, M. Nagatsu, Plasmainduced grafting of cyclodextrin onto multiwall carbon nanotube/iron oxidesfor adsorbent application, J. Phys. Chem. B 114 (2010) 6779–6785.

[17] S.T. Yang, Z.Q. Guo, G.D. Sheng, X.K. Wang, Application of a novel plasma-induced CD/MWCNT/iron oxide composite in zinc decontamination,Carbohydr. Polym. 90 (2012) 1100–1105.

[18] M.H. Al-Qunaibit, W.K. Mekhemer, A.A. Zaghloul, The sorption of Cu(II) ions onbentonite-a kinetic study, J. Colloid Interface Sci. 283 (2005) 316–321.

[19] S.W. Wang, Y.H. Dong, M.L. He, L. Chen, X.J. Yu, Characterization of GMZbentonite and its application in the adsorption of Pb(II) from aqueoussolutions, Appl. Clay Sci. 43 (2009) 164–171.

[20] Q.H. Fan, D.D. Shao, W.S. Wu, X.K. Wang, Effect of pH, ionic strength,temperature and humic substances on the sorption of Ni(II) to Na-attapulgite,Chem. Eng. J. 150 (2009) 188–195.

[21] J.H. Huang, Y.F. Liu, X.G. Wang, Selective sorption of tannin from flavonoids byorganically modified attapulgite clay, J. Hazard. Mater. 160 (2008) 382–387.

[22] S.T. Yang, D.L. Zhao, H. Zhang, S.S. Lu, L. Chen, X.J. Yu, Impact of environmentalconditions on the sorption behavior of Pb(II) in Na-bentonite suspensions, J.Hazard. Mater. 183 (2010) 632–640.

[23] L.M. Zuo, S.M. Yu, H. Zhou, J. Jiang, X. Tian, Adsorption of Eu(III) from aqueoussolution using mesoporous molecular sieve, J. Radioanal. Nucl. Chem. 288(2011) 579–586.

[24] S.T. Yang, P.F. Zong, X.M. Ren, Q. Wang, X.K. Wang, Rapid and high-efficientpreconcentration of Eu(III) by core-shell structured Fe3O4@humic acidmagnetic nanoparticles, ACS Appl. Mater. Interfaces 4 (2012) 6891–6900.

[25] S.T. Yang, P.F. Zong, G.D. Sheng, Q. Wang, X.K. Wang, Fabrication of b-cyclodextrin conjugated magnetic HNT/iron oxide composite for high-efficientdecontamination of U(VI), Chem. Eng. J. 214 (2013) 376–385.

136 P. Dong et al. / Chemical Engineering Journal 256 (2014) 128–136

[26] Y.T. Zhou, H.L. Nie, C. Branford-White, Z.Y. He, L.M. Zhu, Removal of Cu2+ fromaqueous solution by chitosan-coated magnetic nanoparticles modified with a-ketoglutaric acid, J. Colloid Interface Sci. 330 (2009) 29–37.

[27] S.S. Gupta, K.G. Bhattacharyya, Adsorption of Ni(II) on clays, J. Colloid InterfaceSci. 295 (2006) 21–32.

[28] Z.J. Liu, J.W. Yang, Z.C. Zhang, L. Chen, Y.H. Dong, Study of 63Ni(II) sorption onCMC-bound bentonite from aqueous solutions, J. Radioanal. Nucl. Chem. 291(2012) 801–809.

[29] S.T. Yang, J.X. Li, D.D. Shao, J. Hu, X.K. Wang, Adsorption of Ni(II) on oxidizedmulti-walled carbon nanotubes: effect of contact time, pH, foreign ions andPAA, J. Hazard. Mater. 166 (2009) 109–116.

[30] X.L. Tan, C.L. Chen, S.M. Yu, X.K. Wang, Sorption of Ni2+ on Na-rectorite studiedby batch and spectroscopy methods, Appl. Geochem. 23 (2008) 2767–2777.

[31] K.F. Lam, K. Yeung, G. Mckay, Efficient approach for Cd2+ and Ni2+ removal andrecovery using mesoporous adsorbent with tunable selectivity, Environ. Sci.Technol. 41 (2007) 3329–3334.

[32] K.C. Justi, V.T. Fávere, M.C.M. Laranjeira, A. Neves, R.A. Peralta, Kinetics andequilibrium adsorption of Cu(II), Cd(II), and Ni(II) ions by chitosanfunctionalized with 2[-bis-(pyridylmethyl)aminomethyl]-4-methyl-6-formylphenol, J. Colloid Interface Sci. 291 (2005) 369–374.

[33] H. Parab, S. Joshi, N. Shenoy, A. Lali, U.S. Sarma, M. Sudersanan, Determinationof kinetic and equilibrium parameters of the batch adsorption of Co(II), Cr(III)and Ni(II) onto coir pith, Process Biochem. 41 (2006) 609–615.

[34] K.G. Bhattacharyya, S.S. Gupta, Adsorption of Fe(III), Co(II) and Ni(II) on ZrO-kaolinite and ZrO-montmorillonite surfaces in aqueous medium, Colloids Surf.A 317 (2008) 71–79.

[35] C. Lu, C. Liu, G.P. Rao, Comparisons of sorbent cost for the removal of Ni2+ fromaqueous solution by carbon nanotubes and granular activated carbon, J.Hazard. Mater. 151 (2008) 239–246.

[36] K. Xia, W. Bleam, P.A. Helmke, Studies of the nature of binding sites of first rowtransition elements bound to aquatic and soil humic substances using X-rayabsorption spectroscopy, Geochim. Cosmochim. Acta 61 (1997) 2223–2235.

[37] R. Dähn, A.M. Scheidegger, A. Manceau, M.L. Schlegel, B. Baeyens, M.H.Bradbury, M. Morales, Neoformation of Ni phyllosilicate upon Ni uptake onmontmorillonite: a kinetics study by powder and polarized extended X-rayabsorption fine structure spectroscopy, Geochim. Cosmochim. Acta 66 (2002)2335–2347.

[38] M. Nachtegaal, D.L. Sparks, Nickel sequestration in a kaolinite-humic acidcomplex, Environ. Sci. Technol. 37 (2003) 529–534.

[39] R. Dähn, A.M. Scheidegger, A. Manceau, M.L. Schlegel, B. Baeyens, M.H.Bradbury, D. Chateigner, Structural evidence for the sorption of Ni(II) atoms onthe edges of montmorillonite clay minerals: a polarized X-ray absorption finestructure study, Geochim. Cosmochim. Acta 67 (2003) 1–15.

[40] T.J. Strathmann, S.C.B. Myneni, Effect of soil fulvic acid on nickel(II) sorptionand bonding at the aqueous boehmite (c-AlOOH) interface, Environ. Sci.Technol. 39 (2005) 4027–4034.

[41] Y. Arai, Spectroscopic evidence for Ni(II) surface speciation at the ironoxyhydroxides–water interface, Environ. Sci. Technol. 42 (2008) 1151–1156.