Column adsorption of Cu(II)by polymer-supported nano-iron oxides in the presence of sulfate:Experimental and mathematical modeling☆

Hui Qiu ,Xiaolin Zhang *,Zhe Xu

1 Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology(CICAEET),School of Environmental Science and Engineering,Nanjing University of Information Science&Technology,Nanjing 210044,China

2 State Key Laboratory of Pollution Control and Resource Reuse,School of the Environment,Nanjing University,Nanjing 210023,China

1.Introduction

As is generally known,iron(hydr)oxides can effectively sequestrate heavy metal ions through outer sphere or inner sphere complexation and thereby significantly affect the immigration of toxic metals in water and soil systems[1–3].In terms of high affinity to metals,easy availability and low cost,iron(hydr)oxides exhibitgreatpotentialin decontamination of water from heavy metals[4,5].Unfortunately,most iron(hydr)oxides are presented as fine orultra fine particles,and cannot be directly applied in flow-through systems due to the excessive pressure drop,low hydraulic conductivity,and poor mechanical rigidity[6–8].In the past decade,people have found an effective approach to overcome the above-mentioned technical challenges,i.e.,by encapsulating fine oxide particles inside porous hosts oflarge size and good geometrical shape(e.g.,spherical beads or fibers).The resultant composite materials are expected to incorporate the high reactivity of iron(hydr)oxides and the desirable hydraulic properties of host materials[9],and could be readily employed in fixed-bed columns or other flowthrough systems for heavy metals removal.Activated carbon,zeolite,alginate,and polymers are widely used to serve as hosts for nanoparticles nowadays[10–14].In particular,porous polymeric beads are extremely attractive partially because of their adjustable pore size and surface chemistry[15,16].Various functional groups could be grafted on the inner surface of the materials depending on the nature of pollutants.Cumbal and SenGupta[17]found that the Donnan effect greatly enhanced sorption efficiencies towards heavy metals,because the polymeric beads containing charged groups could pre-concentrate these metals,facilitating the formation of inner-or outer-sphere complex with the encapsulated iron oxides.Recently,a new composite material was fabricated by impregnating HFO nanoparticles into a commercial chloromethyl polystyrene(PS)resin,which exhibited good removal capacities towards Cu(II)and As(V)[18].

Commonly,considerable oxyanions like sulfate,phosphate,and nitrate also exist in the waters contaminated by heavy metals.These inorganic ligands can promote metal adsorption onto iron(hydr)oxides by changing theirsurface charge,forming stable surface-metalligand complexes,or hinder it through the formation of soluble complexes,which is greatly dependent upon the nature of ligands,solution pH,surface properties of iron(hydr)oxide,as well as ligand/metal ratio[19,20].For example,Aliand Dzombak[21]demonstrated that Cu(II)sorption on ferrihydrite would be dramatically enhanced in the presence of sulfate due to the formation of a ternary complex with stoichiometry≡FeOHMeSO4based on the generalized two-layer model.Conversely,with the aid of ATR–FTIR study,Beattieet al.[22]believed that the change of surface charge appeared to be a plausible mechanism for higher Cu(II)uptake to goethite in sulfate systems.

Nevertheless,almost all the available literatures discussed the effects of ligands on heavy metal removal based on the results of batch studies.Column adsorption has been scarcely exploited for this purpose,though it is widely used in realistic applications.The objective of this study is to evaluate the effect of sulfate on Cu(II)sorption by a polymer-supported HFO composite in column adsorption.The nanocomposite HFO-PS was obtained by immobilizing nanosized hydrated ferric oxide inside a commercial chloromethylated polystyrene resin[23,18].The effects of initial Cu(II)concentration, flow rate,and column length on column adsorption were examined,and several empirical models were applied to predict the breakthrough curves,including Adams–Bohart,Clark,Thomas and BDST models.

2.Materials and Method

2.1.Chemicals

All chemicals were of reagent grade and solutions were prepared by deionized water(18.25 MΩ·cm).The chloromethylated polystyrene(PS)beads were purchased from Zhengguang Resin Co.Ltd.(Hangzhou,China).Prior to use,the PS beads were sieved(0.5–0.6 mm),and rinsed with 1 mol·L?1HCl and 1 mol·L?1NaOH in sequence,then extracted with ethanol in a Soxhlet apparatus to remove the possible residual impurities.Finally,they were dried under vacuum at323 Kfor 24 h for further use.

2.2.Preparation of HFO-bearing hybrid sorbent

Encapsulation of HFO nanoparticles within PS beads was achieved by the following steps[24].First,10 g of dry PS beads was added into 300 ml ethanol–water(V:V=1:2)solution containing 2 mol·L?1FeCl3and stirred at 298 K for 12 h to ensure the sufficient permeation of Fe(III)into the inner surface of polymeric phase.Second,the Fe(III)-loaded PS beads were filtered,vacuum-desiccated,and immersed into a 200 ml solution containing 0.1 mol·L?1NaOH and 0.1 mol·L?1NaCl,and then stirred at 298 K for 12 h.Finally,the resulting PS beads were rinsed with ultrapure water till pH=5–6 and were thermally treated at 328 K for 10 h to obtain the hybrid HFO-PS.The amorphous nature of the loaded HFO was con firmed by comparison of the XRD pattern of the resulting hybrid sorbents and its parent materials(data not shown).

2.3.Batch sorption experiments

Batch sorption experiments were conducted using a traditional bottle-point method.To start the experiment,50 mg of adsorbent was added into 50 ml solution containing different amount of Cu(II)and sulfate.Na2SO4was added to study the in fluence of sulfate on Cu(II)sorption with the concentrations ranging from 0 to 30 mmol·L?1.All the experiments were carried out at a constant ionic strength of 0.01 mol·L?1NaNO3.Solution pH was maintained at 5.5 ± 0.1 by using 0.1 mol·L?1HNO3and 0.1 mol·L?1NaOH solution at intervals[25].The flasks were then transferred to an incubator shaker equipped with thermostat,and shaken at 298 K for 24 h to ensure the sorption equilibrium.The amount of Cu(II)adsorbed on HFO-PS was determined by the difference between the initial and the remaining concentrations in equilibrium.

2.4.Column experiments

Fixed-bed experiments were performed using a small polyethylene column(12 mm in diameter and 130 mm in length)equipped with a thermostated water bath to maintain constant temperature,and a certain amount of HFO-PS was packed within columns for test.The simulated wastewater containing Cu(II)and sulfate was prepared as the feeding solution and pumped using a Lange-580 pump(China)to ensure a constant flow rate.The operating conditions,initial Cu(II)or sulfate concentration, flow rate,and column length,were summarized in the related graphs.All the column experiments were performed at 298 K.

2.5.Analysis

To determine the concentration of soluble Cu(II),the filtrate was acidified and analyzed with Flame atomic adsorption spectrophotometry(Thermal Co.,US).The specific surface area and pore size distribution of the polymeric host and nanocomposite were measured by a NOVA3000e Instrument(USA)using N2adsorption and desorption test at 77 K.The loaded HFO particles were observed with a highresolution transmission electron microscope(TEM,JEOL JEM-100S).Mineralogy ofthe samples was determined by an X-ray diffraction analysis instrument(XTRA,Switzerland).

2.6.Mathematical models

Four mathematicalmodels,Adams–Bohart,Clark,Thomas,and BDST model were applied to simulate the breakthrough curves.

2.6.1.Adams–Bohart model

Adams–Bohart model was originally applied to a gas–solid system and nowadays has been expanded to describe and quantify other systems like solid–liquid system[26].It assumes that sorption rate is proportional to the residual capacity of the solid and the concentration of the adsorbate,which is appropriate to describe the initial part of the breakthrough curve.The model can be expressed as:

whereC0(mg·L?1)is the in fluent Cu(II)concentration,Ct(mg·L?1)is the ef fluent Cu(II)concentration at timet(min),kAB(L·mg?1·h?1)is the mass transfer coefficient,N0(mg·L?1)is the maximum volumetric adsorption capacity,Z(cm)is the column length andU0(m·h?1)is the super ficial velocity(the ratio between the volumetric flow rate and the section area).Its linear form is:

2.6.2.Clark model

The Clark model combines the Freundlich equation and the mass transfer concept to de fine a new relation for the breakthrough curve[27]:

wherenis the Freundlich constant,and bothAandrare Clark constants.For a particular adsorption process in fixed-bed column,the values ofAandrcan be determined from Eq.(4),thereby enabling the prediction of breakthrough curves[28].

2.6.3.Thomas model

The Thomas model is one of the most generally and widely used models in column performance theory[29].It assumes a Langmuir modelofequilibrium,no axialdispersion is derived with the adsorption,and the rate driving force obeys pseudo-second order reversible reaction kinetics[30].The expression of Thomas model is given as follows:

wherekTh(L·mg?1·h?1)is the Thomas rate constant,qTh(mg·g?1)is the theoretical saturated adsorption capacity,v(ml·h?1)is the flow rate of the ef fluent,andm(g)is the mass of the adsorbent.The value ofCt/C0is the ratio of ef fluent and in fluent Cu(II)concentrations at certain time.The kinetic coefficientkThand the adsorption capacity of the columnqThcan be obtained from a plot ofCt/C0againsttat a given flow rate using linear regression.

2.6.4.BDST model

The bed depth service time(BDST)model,which was derived from Adams–Bohart model,is generally accepted as a rapid prediction method,stating thatbed heightZand service timetofa column bearsa linear relationship[31].It assumes that sorption is a continuous process wherein equilibrium is not attained instantaneously and the rate of sorption is proportional to sorption capacity that still remains on sorbent.It can be expressed by the following equation:

wherek′is the adsorption rate constant(L·mg?1·h?1),Zis the column length(cm),Uis the linear flow velocity(m·h?1),andtis the service time of column under above conditions(h).

3.Results and Discussion

3.1.Characterization of HFO-based nanocomposite

The as-obtained hybrid material HFO-PS was characterized by scanning electron microscope(SEM),transmission electron microscope(TEM),X-ray diffraction(XRD)and N2adsorption–desorption test at 77 K.As suggested by the TEM image shown in Fig.1a,the PS matrix was characterized by homogeneous nature.As depicted in Fig.1b,one can observe lots of irregular shadows of dozens of nanometers appearing in HFO-PS sample,indicating that HFO particles were encapsulated within PS matrix successfully.Also,it can be observed that HFO is well dispersed as separated nanoparticles in polystyrene matrix,which facilitates the decontamination performance of HFO-PS by providing more accessible sites than bulky HFO.Besides,the impregnation of HFO resulted in a significantincrease in the specific surface area from 38.90 of PS to 68.52 m2·g?1of HFO-PS,which is mainly ascribed to the well dispersed HFO nanoparticles within the PS host[32,33].Fig.S1 shows the X-ray diffraction spectra of the composite HFO-PS,from which one can see a broad peak ataround 20°arising from the presence of the PS matrix[18],con firming that the impregnated HFO was mainly in the form of amorphous phase.The loading content of Fe in HFO-PS was 7.67%in mass(see supplementary material).

3.2.Effects of sulfate on Cu(II)sorption

3.2.1.Batch sorption

Adsorption isotherms were conducted at 298 K to determine the effects of sulfate on Cu(II)removal by HFO-PS.As shown in Fig.2,the adsorption capacity of Cu(II)is about 22 mg·g?1when the equilibrium concentration of phosphate is 20 mg·L?1without sulfate.The amount of adsorbed Cu(II)at the same condition is significantly enhanced and up to 42,47 and 53 mg·g?1in the presence of sulfate at the concentration of 0.5,2 and 10 mmol·L?1sulfate respectively.Previous studies indicated that Cu(II)is adsorbed onto the surface of ferrihydrite by the formation of edge sharing inner-sphere sorption complexes[34].In the presence of sulfate,the formation of Cu–SO4ternary complexes is distinguished with the aid of XPS spectra,and it plays an important role for the enhanced Cu(II)adsorption on HFO-PS[24].

The Langmuir and Freundlich models were employed to fit the equilibrium data[35].As listed in Table 1,the correlation coefficients for Langmuir model are larger than 0.97,while that for Freundlich model were higher than 0.95,suggesting that Langmuir model is more appropriate to represent the experimental data in the studied concentration range.The normalized saturation capacities from Langmuir model,qm,are 58.82,62.67,and 67.13 mg·g?1HFO in the presence of sulfate at 0.5,2,and 10 mmol·L?1respectively,much higher than the sulfatefree systems(qm,32.26 mg·g?1HFO).In particular,theKLvalues increases exponentially with the concentration of sulfate,and the conformed formula was presented as follows:

3.2.2.Column sorption

Column adsorption was employed to evaluate the effect of sulfate on the performance of the hybrid sorbent HFO-PS in a typical flow-through system[36],and the breakthrough curvesare illustrated in Fig.3.Column adsorptions were carried out at sulfate concentrations of 0,0.5,2 and 10 mmol·L?1,with the initial Cu(II)concentration of 4.86 mg·L?1,flow rate of 25 ml·h?1,and column length of 4.42 cm.From Fig.3,we can observe that the breakthrough point(set atCt/C0=0.2)occurred after 0.29 L in the sulfate-free system,while 0.79 L,1.15 L,and 1.48 L for sulfate concentration of 0.5,2 and 10 mmol·L?1,respectively.Cu(II)broke through fairly later in the sulfate systems.At sulfate concentration of 10 mmol·L?1,the treatment capacity of HFO-PS is nearly 5-folder that in sulfate-free system.It is clear that the addition ofsulfate leads to a considerable enhancement in the treatment capacity of HFO-PS,which is consistent with the batch studies.It is really attractive because sulfate is generally present with Cu(II)ions in the contaminated waters.

Fig.1.TEM images of the samples.(a)PS and(b)HFO-PS.

Fig.2.(a)Adsorption isotherms of Cu(II)sorption onto HFO-PS under variable sulfate concentrations at 298 K(adsorption dose,0.5 g·L?1;pH,5.5± 0.1).(b)The relationship between Langmuir adsorption constant K L and sulfate concentration.

Table 1Langmuir and Freundlich isotherm constants for Cu(II)sorption by HFO-PS

3.3.Effect of operating parameters

The in fluence of the operation parameters,initial Cu(II)concentration, flow rate,and column length on the breakthrough curves was evaluated and mathematically modeled.

3.3.1.Effect of the inlet concentration

The initial Cu(II)concentration was set as 4.86,7.29 and 9.72 mg·L?1,respectively[Fig.4(a)],with the flow rate of 25 ml·h?1,the bed height of 4.42 cm,and the sulfate concentration of 2 mmol·L?1.The breakthrough(set atCt/C0=0.2)occurred after 1.15 L for the initial 4.86 mg·L?1Cu(II),0.84 L for 7.29 mg·L?1Cu(II),and 0.65 L for 9.72 mg·L?1Cu(II).As expected,the breakthrough time was shortened with the increased inlet concentration because the sites of HFO-PS for Cu(II)sorption are limited.The slope of breakthrough curve becomes steeper when the inlet Cu(II)concentration increases.

3.3.2.Effect of theflow rate

The effectsof flow rate on Cu(II)removalby HFO-PSwere investigated at 12.5 ml·h?1,25 ml·h?1and 50 ml·h?1,respectively,where the initial Cu(II)concentration,the column length and the sulfate concentration were set as 4.86 mg·L?1,4.42 cm and 2 mmol·L?1.As seen from Fig.4(b),a breakthrough point(set atCt/C0=0.2)appeared at 0.64 L for the flow rate of 50 ml·h?1,1.14 L for 25 ml·h?1,and 1.72 L for 12.5 ml·h?1.It is obvious that the breakthrough point occurred at a less treatment capacity with a higher flow rate.It is attributed to the insufficient residence time in the column at a higher rate,which makes the Cu(II)diffusion into the pores of the sorbent inhibited.

3.3.3.Effect of the column length

The breakthrough curves were examined at different column lengths of 2.21,4.42,6.63 cm,using an initial Cu(II)concentration of 4.86 mg·L?1, flow rate of 25 ml·h?1,and the sulfate concentration at 2 mmol·L?1.From Fig.4c,the experimental breakthrough time(set atCt/C0=0.2)was found to be 0.34 L for 2.21 cm,1.16 L for 4.42 cm and 1.74 L for 6.63 cm.Itseems that an increase in the column length causes a significant increase in the breakthrough time.It is mainly due to the increased adsorption sites and longer contact time as the bed height increased[37].Besides,the slopes of the breakthrough curve are roughly similar,as a change of the column length does not affectthe mass transfer of the process at the same concentration and flow rate.

3.4.Mathematical modeling

Four mathematical models including Adams–Bohart,Clark,Thomas,and BDST model were employed to fit the experimental data.Each model allows to determine different parameters on the fixed-bed sorption of Cu(II)onto HFO-PS.The standard error(SE)method is used to evaluate the difference between the predicted and experimentalvalues.

Fig.3.Breakthrough curves of Cu(II)removal by HFO-PS at different sulfate levels.(Initial Cu(II),4.86 mg·L?1;column length,4.42 cm; flow rate,25 ml·h?1).

Fig.5(a)presents the predicted curves according to Adams–Bohart model.The characteristic parameters such as adsorption capacityN0and the relative coefficientkABwere shown in Table S1(see supplementary material).Clearly,Adams–Bohart model could not fit the experimental data accurately and a considerable deviation betweenN0andNexpwas observed.The regression coefficientR2also showed unsatisfactory fitness.Also,the calculated adsorption capacityN0could not match the corresponding experimental ones well.Herein,Adams–Bohart model was not proper to describe Cu(II)adsorption in column runs.

Clark model gave a new simulation of breakthrough curves.Batch sorption showed that Freundlich model is valid for the sorption of Cu(II)by HFO-PS,allowing the use of the Freundlich constant(n=3.75 at sulfate of 2 mmol·L?1)to calculate the parametersAandrin Clark model.From Fig.5b,we found that Clark model fitted the experimental data much better than the Adams–Bohart model with the regression coefficientsR2around 0.90(Table S1),though the considerable difference was still observed between the calculated and experimental data.

The breakthrough curves predicted by Thomasmodelwere depicted in Fig.4,and the fit of Thomas model allows predicting the sorption capacity of the column(qTh)and the Thomas rate constant(kTh).The determined coefficients and relative constants were obtained using linear regression analysis according to Eq.(6),and the results are listed in Table S2.In all the cases,a negligible difference between the predicted and the experimental data is observed.From the regression coefficients and standard error,it can be concluded that Thomas model fits the experimental quite well(R2＞0.96;SE＜0.3).

Fig.4.Breakthrough curves of Cu(II)removal by HFO-PS in sulfate systems:Experimental and predicted by the Thomas model:(a)different initial Cu(II)concentrations(column length,4.42 cm; flow rate,25 ml·h?1);(b)different volumetric flow rates(column length,4.42 cm;in fluent Cu(II),4.86 mg·L?1);(c)different column lengths(in fluent Cu(II),4.86 mg·L?1; flow rate,25 ml·h?1).The sulfate concentration is constant at 2 mmol·L?1.

Fig.5.Breakthrough curves of Cu(II)removal by HFO-PS in 2 mmol·L?1 sulfate backgrounds:Experimental and predicted by Adams–Bohart model(a)and Clark model(b)at different initial Cu(II)levels.(Column length,4.42 cm; flow rate,25 ml·h?1).

At a low flow rate(12.5 ml·h?1),the sorption capacityqThis more than 4.5 mg·g?1,higher than that at high flow rate(25 and 50 ml·h?1).By contrast,the Thomas rate constant(kTh)shows the opposite trend.Fig.6 shows the calculated values ofkThandqThunder different volumetric flow rates.We also obtained the linear equations forkThandqTh(kThvsflow rate,qThvsflow rate).

Fig.6.The linear plot of Thomas parameters vs flow rate at 298 K in 2 mmol·L?1 sulfate backgrounds.(a)k Th vs flow rate.(b)q Th vs flow rate.

With the increase ofvolumetric flowrate,kThincreased whileqThdecreased,mainly because higher flow rate would result in faster mass transfer and more insufficient adsorption[38].The Eqs.(8)–(9)might be applied to predict the breakthrough curves in practical operation.The values ofqThandkThhave yet to show significant dependence on the initial Cu(II)concentration and column height,implying that both factors have negligible in fluence on the rate of mass transfer and saturated adsorption capacity.The underlying mechanism deserves to be further explored in future research.

The BDST modelis proposed to assess the relationship between time and column length.The lines oft-ZatCt/C0=0.1,0.2,0.3,0.4 are shown in Fig.7.The equations of these lines are as follows:

Fig.7.BDST fitting curves for Cu(II)sorption onto HFO-PS at 298 K in 2 mmol·L?1 sulfate backgrounds.(Initial Cu(II),4.86 mg·L?1;column length,4.42 cm; flow rate,25 ml·h?1).

ValuesofN0(g·L?1)andk′(L·mg?1·h?1)in sulfate systemsare determined according to the slopes and intercepts of the lines and are listed in Table S3.Once the relative constants are determined,the model can be used to estimate the service time for a given column length and specific solute concentration.It is found that the regression coefficientsR2all exceed 0.95,indicating that the BDST model could be applicable to represent Cu(II)sorption by HFO-PS in the presence of sulfate.

Note that the adsorption rate is proportional to the residual capacity ofthe adsorbentand the concentration ofCu(II),the value ofadsorption rate constantk′would decrease asCt/C0increased,due to the weakened driving force of mass transfer[39].However,the calculated value ofk′is shown to rise whenCt/C0≥0.3 as listed in Table S3.Since BDST model is originated from Adams–Bohart model,it is appropriate for the description of the initial part of the breakthrough curve(Ct/C0＜0.3).

4.Conclusions

Column adsorption of Cu(II)by polymer-supported nano-iron oxides in the presence of sulfate has been evaluated in the present study.A hybrid sorbent HFO-PS was fabricated by encapsulating nanosized HFO into macroporous polystyrene(PS)resin.We evaluated the effect of sulfate on the sorption towards Cu(II)of HFO-PS in batch and particularly column runs.Both Langmuir and Freundlich models show a good description forthe isothermadsorption,and the adsorption constantsKLhas been growing exponentially as the sulfate concentration rises.In column operations,the presence of sulfate significantly improved Cu(II)sorption onto HFO-PS,and the effects of flow rate,inlet concentration,and column length on column adsorption have been investigated.Mathematical models were applied to predict the theoretical breakthrough curves.Among them,Thomas model is found to be the most suitable one to describe the column adsorption.The calculated value of Thomas rate constantkThand adsorption capacityqThshows linear relation with flow rate,i.e.,kTh=6.17×10?4v?4.63×10?3andqTh=?5.62v+5.093,which could be useful for predicting the breakthrough curves for scaling up.Besides,a linear relationship between breakthrough time and column length was suggested by BDST model,which could possibly predict the breakthrough time for Cu(II)sorption onto HFO-PS especially forCt/C0＜0.3.

Acknowledgments

This work was supported by A Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions(PAPD),Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control.

Supplementary Material

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.cjche.2017.01.011.

[1]I.Ali,New generation adsorbents for water treatment,Chem.Rev.112(2012)5073–5091.

[2]D.J.Cain,M.N.Croteau,C.C.Fuller,Dietary bioavailability of Cu adsorbed to colloidal hydrous ferric oxide,Environ.Sci.Technol.47(2013)2869–2876.

[3]C.Dai,Y.D.Hu,Fe(III)hydroxide nucleation and growth on quartz in the presence of Cu(II),Pb(II),and Cr(III):Metal hydrolysis and adsorption,Environ.Sci.Technol.49(2015)292–300.

[4]S.Dulnee,A.C.Scheinost,Surface reaction of snII on goethite(α-FeOOH):Surface complexation,redox reaction,reductive dissolution,and phase transformation,Environ.Sci.Technol.48(2014)9341–9348.

[5]H.Foerstendorf,N.Jordan,K.Heim,Probing the surface speciation of uranium(VI)on iron(hydr)oxides by in situ ATR FT-IR spectroscopy,J.Colloid Interface Sci.416(2014)133–138.

[6]A.V.Vitela-Rodriguez,J.R.Rangel-Mendez,Arsenic removal by modified activated carbons with iron hydro(oxide)nanoparticles,J.Environ.Manag.114(2013)225–231.

[7]R.Y.Li,L.B.Zhang,P.Wang,Rational design of nanomaterials for water treatment,Nano7(2015)17167–17194.

[8]H.Qiu,C.Liang,X.L.Zhang,M.D.Chen,Y.X.Zhao,T.Tao,Z.W.Xu,G.Liu,Fabrication of a biomass-based hydrous zirconium oxide nanocomposite for preferable phosphate removal and recovery,ACS Appl.Mater.Interfaces7(2015)20835–20844.

[9]S.J.Tesh,T.B.Scott,Nano-composites for water remediation:A review,Adv.Mater.26(2014)6056–6068.

[10]L.C.Oliveira,D.I.Petkowicz,A.Smaniotto,S.B.Pergher,Magnetic zeolites:A new adsorbent for removal of metallic contaminants from water,Water Res.38(2004)3699–3704.

[11]S.F.Lim,Y.M.Zheng,S.W.Zou,J.P.Chen,Characterization of copper adsorption onto an alginate encapsulated magnetic sorbent by a combined FT-IR,XPS,and mathematical modeling study,Environ.Sci.Technol.42(2008)2551–2556.

[12]J.A.Arcibar-Orozco,M.Avalos-Borja,J.R.Rangel-Mendez,Effect of phosphate on the particle size offerric oxyhydroxides anchored onto activated carbon:As(V)removal from water,Environ.Sci.Technol.46(2012)9577–9583.

[13]O.Hakami,Y.Zhang,C.J.Banks,Thiol-functionalised mesoporous silica-coated magnetite nanoparticles for high efficiency removal and recovery of Hg from water,Water Res.46(2012)3913–3922.

[14]H.Qiu,S.J.Zhang,B.C.Pan,W.M.Zhang,L.Lv,Oxalate-promoted dissolution of hydrous ferric oxide immobilized within nanoporous polymers:Effect ofionic strength and visible light irradiation,Chem.Eng.J.232(2013)167–173.

[15]M.Hua,Y.N.Jiang,B.Wu,B.C.Pan,X.Zhao,Q.X.Zhang,Oxalate-promoted dissolution of hydrous ferric oxide immobilized within nanoporous polymers:Effect of ionic strength and visible light irradiation,ACS Appl.Mater.Interfaces5(2013)12135–12142.

[16]B.C.Pan,F.C.Han,G.Z.Nie,B.Wu,K.He,L.Lv,New strategy to enhance phosphate removal from water by hydrous manganese oxide,Environ.Sci.Technol.48(2014)5101–5107.

[17]L.Cumbal,A.K.SenGupta,Arsenic removal using polymer-supported hydrated iron(III)oxide nanoparticles:Role of Donnan membrane effect,Environ.Sci.Technol.39(2005)6508–6515.

[18]G.Z.Nie,B.C.Pan,S.J.Zhang,B.J.Pan,Surface chemistry of nanosized hydrated ferric oxide encapsulated inside porous polymer:Modeling and experimental studies,J.Phys.Chem.C117(2013)6201–6209.

[19]M.Komárek,A.Vaněk,V.Ettler,Chemical stabilization of metals and arsenic in contaminated soils using oxides—A review,Environ.Pollut.172(2013)9–22.

[20]M.A.G.Hinkle,Z.M.Wang,D.E.Giammar,J.G.Catalano,Interaction of Fe(II)with phosphate and sulfate on iron oxide surfaces,Geochim.Cosmochim.Acta158(2015)130–146.

[21]M.A.Ali,D.A.Dzombak,Interactions of copper,organic acids,and sulfate in goethite suspensions,Geochim.Cosmochim.Acta60(1996)5045–5053.

[22]D.A.Beattie,J.K.Chapelet,M.Grafe,W.M.Skinner,E.Smith,Interactions of copper,organic acids,and sulfate in goethite suspensions,Environ.Sci.Technol.42(2008)9191–9196.

[23]Z.M.Jiang,L.Lv,W.M.Zhang,Q.Du,B.C.Pan,L.Yang,Q.X.Zhang,Nitrate reduction using nanosized zero-valent iron supported by polystyrene resins:Role of surface functional groups,Water Res.45(2011)2191–2198.

[24]H.Qiu,S.J.Zhang,B.C.Pan,W.M.Zhang,L.Lv,Effect of sulfate on Cu(II)sorption to polymer-supported nano-iron oxides:Behavior and XPS study,J.ColloidInterface Sci.366(2012)37–43.

[25]C.F.Baes,R.E.Mesmer,The Hydrolysis of Cations,Wiley,New York,1976.

[26]G.Bohart,E.Adams,Some aspects of the behavior of charcoal with respect to chlorine.1,J.Am.Chem.Soc.42(1920)523–544.

[27]R.M.Clark,Evaluating the cost and performance of field-scale granular activated carbon systems,Environ.Sci.Technol.21(1987)573–580.

[28]V.C.Srivastava,B.Prasad,I.M.Mishra,I.D.Mall,M.M.Swamy,Prediction of breakthrough curves for sorptive removal of phenol by bagasse fly ash packed bed,Ind.Eng.Chem.Res.47(2008)1603–1613.

[29]H.C.Thomas,Heterogeneous ion exchange in a flowing system,J.Am.Chem.Soc.66(1944)1664–1666.

[30]C.B.Lopes,E.Pereira,Z.Lin,P.Pato,M.Otero,C.M.Silva,J.Rocha,A.C.Duarte,Fixedbed removal of Hg2+from contaminated water by microporous titanosilicate ETS-4:Experimental and theoretical breakthrough curves,Microporous Mesoporous Mater.145(2011)32–40.

[31]G.M.Walker,L.R.Weatherley,Adsorption of acid dyes on to granular activated carbon in fixed beds,Water Res.31(1997)2093–2101.

[32]N.Savage,M.S.Diallo,Nanomaterials and water purification:Opportunities and challenges,J.Nanopart.Res.7(2005)331–342.

[33]B.J.Pan,J.Wu,B.C.Pan,L.Lv,W.M.Zhang,L.L.Xiao,X.S.Wang,X.C.Tao,S.R.Zheng,Development of polymer-based nanosized hydrated ferric oxides(HFOs)for enhanced phosphate removal from waste ef fluents,Water Res.43(2009)4421–4429.

[34]A.C.Scheinost,S.Abend,K.I.Pandya,D.L.Sparks,Kinetic controls on Cu and Pb sorption by ferrihydrite,Environ.Sci.Technol.35(2001)1090–1096.

[35]K.Yang,B.S.Xing,Adsorption of organic compounds by carbon nanomaterials in aqueous phase:Polanyi theory and its application,Chem.Rev.110(2010)5989–6008.

[36]A.Sperlich,A.Werner,A.Genz,G.Amy,E.Worch,M.Jekel,Breakthrough behavior of granular ferric hydroxide(GFH) fixed-bed adsorption filters:Modeling and experimental approaches,Water Res.39(2005)1190–1198.

[37]J.Y.Song,W.H.Zou,Y.Y.Bian,F.Y.Su,R.P.Han,Adsorption characteristics of methylene blue by peanut husk in batch and column modes,Desalination265(2011)119–125.

[38]J.L.Sotelo,G.Ovejero,A.Rodríguez,S.álvarez,J.García,Removal of atenolol and isoproturon in aqueous solutions by adsorption in a fixed-bed column,Ind.Eng.Chem.Res.51(2012)5045–5055.

[39]W.X.Zhang,L.Dong,H.Yan,H.J.Li,Z.W.Jiang,X.W.Kan,H.Yang,A.M.Li,R.S.Cheng,Removal of methylene blue from aqueous solutions by straw based adsorbent in a fixed-bed column,Chem.Eng.J.173(2011)429–436.