Journal of Colloid and Interface Science 420 (2014) 74–79

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier .com/locate / jc is

Magnetic mesoporous Fe/carbon aerogel structures with enhancedarsenic removal efficiency

0021-9797/$ - see front matter � 2014 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.jcis.2014.01.008

⇑ Corresponding author. Fax: +886 3 2654199.E-mail address: yflin@cycu.edu.tw (Y.-F. Lin).

Yi-Feng Lin ⇑, Jia-Ling ChenDepartment of Chemical Engineering and R&D Center for Membrane Technology, Chung Yuan Christian University, Chungli 32023, Taiwan, ROC

a r t i c l e i n f o

Article history:Received 20 October 2013Accepted 9 January 2014Available online 15 January 2014

Keywords:AdsorptionMesoporousFe/carbon aerogelsArsenic ionsSurface area

a b s t r a c t

Wastewater treatment has drawn significant research attention due to its associated environmentalissues. Adsorption is a promising method for treating wastewater. The development of an adsorbent witha high surface area is important. Therefore, we successfully developed mesoporous Fe/carbon aerogel(CA) structures with high specific surface areas of 487 m2/g via the carbonization of composite Fe3O4/phenol–formaldehyde resin structures, which were prepared using a hydrothermal process with theaddition of phenol. The mesoporous Fe/CA structures were further used for the adsorption of arsenic ionswith a maximum arsenic-ion uptake of calculated 216.9 mg/g, which is higher than that observed forother arsenic adsorbents. Ferromagnetic behavior was observed for the as-prepared mesoporous Fe/CAstructures with an excellent response to applied external magnetic fields. As a result, the adsorbentFe/CA structures can be easily separated from the solution using an external magnetic field. This studydevelops the mesoporous Fe/CA structures with high specific surface areas and an excellent responseto an applied external magnetic field to provide a feasible approach for wastewater treatment includingthe removal of arsenic ions.

� 2014 Elsevier Inc. All rights reserved.

1. Introduction

Arsenic in natural waters is a worldwide problem, and arsenicexposure from drinking water can cause skin, lung, bladder andkidney cancers [1]. As a result, arsenic removal technologies, suchas precipitation, coagulation, membrane filtration, foam flotation,solvent extraction, bioremediation and adsorption, have attractedsignificant attention [1]. Among these, adsorption is an inexpen-sive and commercially available technology. However, the devel-opment of an adsorbent with a high surface area is necessary toincrease the adsorption capacity of the material. Compared withbulk or commercial materials, nanomaterials are better candidatesfor adsorbents due to their high surface areas. However, nanoma-terials readily aggregate decreasing the nanomaterial surface area,and nanomaterials can be difficult to remove from solution requir-ing the use of additional separation steps such as centrifugation.

Magnetic nanomaterials, such as Fe3O4, c-Fe2O3, Fe and Ni, havedrawn significant research attention due to their excellent re-sponses to externally applied magnetic fields while in solution.Therefore, these magnetic nanomaterials can be easily removedfrom solution with an applied external magnetic field, avoidingthe problems associated with poor separation and the need foradditional separation steps. However, magnetic nanomaterials still

tend to form aggregates that decrease their surface area. To solvethis problem, the synthesis of composite porous magnetic/carbonmaterials with high surface areas provides an alternative methodfor avoiding the aggregation of the nanomaterials. Wang et al. [2]successfully synthesized Fe nanoparticle-functionalized multi-walled carbon nanotubes with a specific surface area of295.4 m2/g and used them to remove copper, lead and cadmiumions from solution. Wu et al. [3] prepared mesoporous c-Fe2O3/carbon encapsulates with a specific surface area of 877 m2/g forthe adsorption of arsenic (V) ions. Arsenic ion adsorption fits theLangmuir model, and the equilibrium adsorption capacity is17.9 mg/g. In addition, mesoporous Ni/graphitic carbon structureswith specific surface areas up to 918 m2/g were successfully syn-thesized by Sun et al., and their adsorption performance with re-spect to metal ions (Cd2+, Cu2+, Ag+ and Au3+) was also studied [4].

Carbon aerogel (CA), a novel carbon structure, has receivedresearch interest due to its high specific surface area(400–1200 m2/g), high porosity (>80%), high mechanical strength,mesoporous structure (2–50 nm) and low cost [5]. Carbon aerogelscan be prepared via the carbonization of phenolic resins under anitrogen atmosphere. The phenolic family of phenol [6] and resor-cinol [7] is reacted with formaldehyde to form the phenol–formaldehyde and resorcinol–formaldehyde resins, respectively.Long et al. (2009) successfully prepared carbon aerogels with spe-cific surface areas of 600 m2/g and a porosity of approximately 80%via the carbonization of phenol–formaldehyde resins [8]. Diverse

Y.-F. Lin, J.-L. Chen / Journal of Colloid and Interface Science 420 (2014) 74–79 75

applications of carbon aerogels, such as supercapacitors [9], ion ex-change resins [10], gas diffusion [11], hydrogen storage [12], electro-chemical double-layer capacitors [7,13], anodes of lithium ionbatteries [14] and adsorbents [5,15], have previously been studied.For metal ion adsorption, carbon aerogels were successfully usedto adsorb lead, mercury, cadmium, zinc, etc. [16]. However, CAadsorbents require an additional separation step to remove themfrom solution. To the best of our knowledge, no previous studieshave examined the application of carbon aerogels for arsenicadsorption.

In this study, for the first time, Fe3O4/phenol–formaldehyde res-ins are successfully synthesized via a one-pot hydrothermal reac-tion using phenol, ferrous sulfate, hexamethylenetetramine(HMTA) and de-ionized water. In this procedure, HMTA is reactedwith de-ionized water to form ammonia and formaldehyde, asshown in the reaction (1) of Scheme 1. The phenol–formaldehyderesins are then synthesized by reacting the as-prepared formalde-hyde with phenol in a one-pot hydrothermal reaction, as shown inthe reaction (3) of Scheme 1. The mesoporous Fe/CA structures areobtained by further carbonization of the as-prepared Fe3O4/phenol–formaldehyde resins under a nitrogen atmosphere. Thephenol–formaldehyde resins are successfully transformed intocarbon aerogels at 800 �C under a nitrogen atmosphere; then, theas-prepared CA serves as a reducing agent to reduce the Fe3O4 tozero-valent Fe, as shown in Scheme 2. As a result, mesoporousFe/CA structures are successfully synthesized with high specificsurface areas of 487 m2/g and pore sizes of 3.3 nm. The ferromag-netic Fe/CA mesoporous structures are further used for arsenic ionadsorption, and the maximum uptake of arsenic ions is calculatedto be 216.9 mg/g, which is higher than that observed for other ar-senic adsorbents [3,17–19]. The as-prepared mesoporous Fe/CAstructures reveal excellent response to an external magnetic fielddue to their ferromagnetic behavior, which is advantageous forthe separation of Fe/CA adsorbents from solution. This work dem-onstrates the preparation of ferromagnetic Fe/CA mesoporousstructures that performed well for arsenic adsorption and providesa feasible approach to wastewater treatment applications, such asthe removal of arsenic ions.

2. Experimental

The Fe3O4/phenol–formaldehyde resins were synthesized usinga hydrothermal process. Ferrous sulfate (FeSO4, 1.1 g), hexamethy-lenetetramine (HMTA, 0.187 g) and phenol (0.502 g) were dis-solved into 60 ml de-ionized water. After stirring for 1 min, the

Scheme 1. The reactions of the formation of phenolic resins using phenol andformaldehyde reactants.

solution mixture was transferred to a 100-ml Teflon-lined auto-clave. Next, the autoclave was heated at 180 �C for 24 h. After cool-ing to room temperature, the product was washed several timeswith ethanol and then dried at 50 �C for 24 h. The as-preparedproduct (Fe3O4/phenol–formaldehyde resins, 180P) was furtherannealed at 400 �C and 800 �C under a nitrogen atmosphere for6 h to form the mesoporous Fe3O4/CA (400P) and Fe/CA (800P)structures, respectively.

The as-prepared mesoporous Fe/CA structures were then usedfor the arsenic (V) ion adsorption studies. Arsenic (V) ion adsorp-tion was obtained by immersing 5 mg mesoporous Fe/CA struc-tures into 50 ml arsenic (V) ion solution with variable arsenic ionconcentration under gentle stirring for 24 h to achieve adsorptionequilibrium. After magnetic separation, the remainingconcentration of arsenic was determined by inductively coupledplasma-atomic emission spectrometry (ICP-AES) so that adsorp-tion isotherms could be obtained.

The surface morphology and the crystallite phases of the as-pre-pared samples were examined via scanning electron microscopy(SEM, Hitachi, S-3000) and powder X-ray diffraction (PXRD, PANa-lytical X’Pert PRO, PW3040/60), respectively. The functional groupsin the as-synthesized samples were characterized using Fourier-transform infrared spectroscopy (FTIR, PerkinElmer Inc., SpectrumOne). The pore size distribution and the surface area of the as-synthesized samples were measured with nitrogen adsorption/desorption isotherms (BET, Micromeritics ASAP 2020). The mag-netic properties of the as-prepared mesoporous Fe/CA structureswere studied using superconducting quantum interference devicemagnetometer (SQUID, Quantum Design MPMS7). The arsenicion content in the solution was measured via inductively coupledplasma atomic emission spectrometry (ICP-AES, Kontron, S-35).

3. Results and discussion

The corresponding compositions and crystalline phases of theas-synthesized sample (180P) were first studied using XRD asshown in Fig. 1(c). Reference patterns for the face-centered cubicphase of the Fe3O4 crystals (JCPDS card number: 89-4319) and phe-nol–formaldehyde resins are also included in Fig. 1(a) and (b) forcomparison. The pattern of Fig. 1(c) was in good agreement withthat of the cubic Fe3O4 reference material, which indicates thatthe as-prepared sample (180P) possesses the face-centered cubiccrystalline structure of Fe3O4. On the other hand, the broad peakaround 20� in Fig. 1(b) contributed from amorphous silica compo-sition of XRD sample holder and the pattern in Fig. 1(b) also showsthat there is no Fe3O4 material in the phenol–formaldehyde resins.

The surface morphologies of the as-prepared sample (180P)were further investigated via FESEM, as shown in Fig. 2(b). The sur-face morphologies of the phenol–formaldehyde resins shown inFig. 2(a) were also studied for comparison. The FESEM image inFig. 2(a) indicates that the as-synthesized phenol–formaldehyderesin particles had spherical shapes with smooth surfaces. Thediameters of the as-prepared phenol–formaldehyde resin particleswere approximately 1–3 lm. After adding ferrous sulfate to thereaction mixture, spherical shapes with rough surfaces wereobserved in the as-prepared samples (180P), as shown inFig. 2(b). The formation of Fe3O4 nanoparticles inside or onto thesurface of the phenol–formaldehyde resins results in the roughsurfaces on the 180P.

To verify the formation of composite Fe3O4/phenol–formaldehyderesin particles in the 180P samples, FTIR spectra were obtained forthe as-prepared phenol–formaldehyde resins and the 180P samples,as shown in Fig. 3(a) and (b), respectively. The characteristic peaks atapproximately 3500, 1630, 1480 and 1230 cm�1 in the FTIR spec-trum of the phenol–formaldehyde resins in Fig. 3(a) arise from the

Scheme 2. The scheme of the preparation of Fe3O4/CAs and Fe/CAs mesoporous structures using hydrothermal process.

Fig. 1. The XRD patterns of (a) the reference face-centered cubic phase of Fe3O4

crystals (JCPDS card number: 89-4319), (b) the phenol–formaldehyde resins and (c)the as-prepared 180P sample.

Fig. 3. FTIR spectra of (a) the phenol–formaldehyde resins and (b) the as-prepared180P sample.

76 Y.-F. Lin, J.-L. Chen / Journal of Colloid and Interface Science 420 (2014) 74–79

OAH stretching, C@C stretching, CH2 bending and CAO stretchingvibrations of the phenolic resins, respectively [20]. The characteris-tic peaks in the FTIR spectrum of the 180P samples in Fig. 3(b) all cor-

Fig. 2. SEM images of (a) the phenol–formaldehyde resins and (b) the as-prepared

respond to those of the as-prepared phenol–formaldehyde resins,indicating the existence of phenol–formaldehyde resins in the180P samples. Therefore, the XRD and FTIR results for the 180P

180P sample. The inset in (b) is the enlarged SEM image of the 180P sample.

Fig. 4. The XRD patterns of (a) the reference face-centered cubic phase of Fe3O4

crystals (JCPDS card number: 89-4319), (b) the reference body-centered cubicphase of Fe crystals (JCPDS card number: 87-0722), (c) the as-prepared 180P sampleand the 180P sample annealed at (d) 400 �C and (e) 800 �C under a nitrogenatmosphere.

Fig. 6. The nitrogen adsorption/desorption isotherms and of the mesoporous Fe/CAstructures.

Y.-F. Lin, J.-L. Chen / Journal of Colloid and Interface Science 420 (2014) 74–79 77

samples confirm that the as-prepared samples (180P) consist ofcomposite Fe3O4/phenol–formaldehyde resin particles.

As determined in previous investigations, phenolic resins can besuccessfully converted into mesoporous carbon aerogels with highspecific surface areas through a carbonization process under anitrogen atmosphere [8]. Therefore, the composite Fe3O4/phenol–formaldehyde resin particles (180P) were further annealed at 400and 800 �C under a nitrogen atmosphere. The corresponding com-positions and crystalline phases of the as-synthesized sample(180P) after annealing at 400 and 800 �C under a nitrogen atmo-sphere were investigated using XRD, as shown in Fig. 4(d) and(e). The reference patterns of the face-centered cubic phase ofthe Fe3O4 crystals (JCPDS card number: 89-4319) and the body-centered cubic phase of the Fe crystals (JCPDS card number: 87-0722) are also included in Figs. 4(a) and (b) for comparison. Thepattern in Fig. 4(d) is in good agreement with that of the cubicFe3O4 crystalline phase, indicating that the Fe3O4 crystalline phaseof the 180P sample after annealing at 400 �C remains face-centeredcubic. However, the diffraction peaks in the pattern in Fig. 4(e) canbe indexed to the body-centered cubic phase of Fe crystals (JCPDScard number: 87-0722). The crystalline phase changed from theface-centered cubic Fe3O4 to the body-centered cubic Fe phaseafter annealing at 800 �C under a nitrogen atmosphere with the fol-lowing possible explanations: The phenol–formaldehyde resins inthe 180P samples were converted to carbon aerogels while anneal-ing at 800 �C under a nitrogen atmosphere. Then, the face-centeredcubic Fe3O4 crystalline phase was reduced to the body-centeredcubic Fe crystalline phase by the carbon aerogel reducing agentat the annealing temperature of 800 �C [21]. As a result, thecomposite Fe/carbon aerogel particles were successfully obtained

Fig. 5. The SEM images of the as-prepared 180P sample anneale

by reducing the composite Fe3O4/phenol–formaldehyde resin par-ticles (180P) using the carbon aerogels (CA) from the carbonizationof phenol–formaldehyde resins. The SEM images of the 180P sam-ples after annealing at 400 and 800 �C under a nitrogen atmo-sphere (Figs. 5(a) and (b)) still exhibit spherical shapes withrough surfaces similar to those of the as-annealed samples, andthe particle sizes were both found to be approximately severalmicrometers.

The specific surface area and the pore size distribution of the as-synthesized composite Fe/CA structures were measured from theN2 adsorption/desorption isotherms using the Brunauer–Em-mett–Teller (BET) method. The as-prepared composite Fe/CA struc-tures exhibited a type IV isotherm and an H1 hysteresis loop, asshown in Fig. 6. The type IV isotherm indicated the existence ofmesopores (2–50 nm) in the composite Fe/CA structures, whichis in good agreement with the average pore diameter of approxi-mately 3.3 nm determined from the Barrett–Joyner–Halenda(BJH) desorption data. The specific surface areas of the compositemesoporous Fe/CA structures were found to be 487 m2/g. The spe-cific surface area and pore size distribution of the phenol–formal-dehyde resins after annealing at 800 �C under a nitrogenatmosphere were also measured from the N2 adsorption/desorp-tion isotherms for comparison. The specific surface areas and theaverage pore size of the CA were 572 m2/g and approximately2 nm, respectively.

The magnetic properties of the as-prepared composite mesopor-ous Fe/CA structures were analyzed using a superconducting quan-tum inference device (SQUID) magnetometer. The magnetichysteresis loop (Figs. 7(a) and (b)) for the as-prepared compositemesoporous Fe/CA structures measured at room temperature in anapplied magnetic field of up to 50,000 Oe indicates ferromagnetic

d at (a) 400 �C and (b) 800 �C under a nitrogen atmosphere.

Fig. 7. (a) The magnetization curve of the mesoporous Fe/CA structures measuredat room temperature. (b) The corresponding magnified curve in the vicinity of theorigin.

Fig. 8. Plot of the arsenic adsorption capacity (qe) versus the arsenic equilibriumconcentration (Ce). Inset provides the plot of Ce/qe versus Ce.

78 Y.-F. Lin, J.-L. Chen / Journal of Colloid and Interface Science 420 (2014) 74–79

behavior in the as-prepared mesoporous Fe/CA structures. The mag-netization saturation of the mesoporous Fe/CA structures was40.9 eum/g, which is significantly lower than that of bulk Fe materi-als (215 emu/g). The existence of CA and the reduced size of the Fenanoparticles are attributed to the lower magnetization saturation.Fig. 7(b) shows the magnetic hysteresis loop in an applied magneticfield between�300 and 300 Oe, which indicates that the remaining

magnetization and coercivity are 0.54 emu/g and 39.5 Oe, respec-tively. When the mesoporous Fe/CA structures dispersed in solutionwere subjected to a magnetic field, the mesoporous Fe/CA structureswere completely separated from the solution, as shown in the insetof Fig. 7(a). The inset indicates that the composite Fe/CA structurespossess excellent response to the external magnetic field, whichcan replace traditional centrifuge separation for the removing theadsorbents in solution, thus reducing energy consumption.

The results of the arsenic ion adsorption study in DI water solu-tion using the as-prepared composite mesoporous Fe/CA structuresare illustrated in Fig. 8. Fig. 8 shows the plots of the quantity of ar-senic ion adsorbed at equilibrium, qe (mg/g), versus the equilib-rium arsenic ion concentrations, Ce (mg/L), for the as-preparedcomposite mesoporous Fe/CA structures. The adsorption data werefurther used to fit the Langmuir isotherm model, as shown in theinset of Fig. 8, which is defined by the following equation: Ce/qe = 1/(q0b) + Ce/q0 [3,22] in which q0 (mg/g) and b (L/mg) are themonolayer capacity of the adsorbent and the adsorption constant,respectively. The plots of Ce/qe versus Ce in the Langmuir modelyield straight lines with a slope of 1/q0 and an intercept of 1/(q0b). Therefore, the plot in the inset of Fig. 8 indicates that theLangmuir model fits the adsorption data of the as-prepared com-posite mesoporous Fe/CA structures. The values of q0 and b werecalculated to be 216.9 mg/g and 0.134 L/mg, respectively. To thebest of our knowledge, the uptake of arsenic ions using the as-pre-pared Fe/CA structures was higher than that observed with otheradsorbents [3,17–19]. The increased adsorption of arsenic ionsusing the as-prepared Fe/CA structures resulted from the largerspecific surface area (487 m2/g), and the as-prepared Fe/CA struc-tures reveal a greater potential for use in wastewater treatmentapplications, such as the removal of arsenic ions.

4. Conclusions

In this study, composite Fe3O4/phenol–formaldehyde resinstructures were successfully prepared using a one-step hydrother-mal process with the addition of phenol. The phenol–formaldehyderesins were converted to mesoporous carbon aerogels (CA) througha carbonization process at an annealing temperature of 800 �C un-der a N2 atmosphere, and the formed mesoporous CAs also servedas the reducing agent for the reduction of Fe3O4 to zero-valent iron(Fe). As a result, mesoporous Fe/CA structures were obtained aftercarbonization of the composite Fe3O4/phenol–formaldehyde res-ins. The BET results indicate that the average pore size and the spe-cific surface area of the mesoporous Fe/CA structures were 3.3 nmand 487 m2/g, respectively. The mesoporous Fe/CA structures exhi-bit excellent response to an external magnetic field due to theirferromagnetic behavior, resulting in increased separation of theFe/CA adsorbents from solution. The arsenic ion adsorption behav-ior fits the Langmuir model, and the monolayer capacity (q0) of theadsorbent and the adsorption constant (b) were calculated to be216.9 mg/g and 0.134 L/mg, respectively, values that are higherthan those observed for other arsenic adsorbents. Therefore, theas-prepared mesoporous Fe/CA structures, which possess high spe-cific surface areas, have significant potential in wastewater treat-ment applications, such as the removal of arsenic ions.

Acknowledgment

The authors thank the National Science Council (NSC) (Projectnumber: NSC 101-2221-E-033-055-MY3) for financial support.

References

[1] D. Mohan, C.U. Pittman Jr., J. Hazard. Mater. 142 (2007) 1–53.[2] H. Wang, N. Yan, Y. Li, X. Zhou, J. Chen, B. Yu, M. Gong, Q. Chen, J. Mater. Chem.

22 (2012) 9230–9236.

Y.-F. Lin, J.-L. Chen / Journal of Colloid and Interface Science 420 (2014) 74–79 79

[3] Z. Wu, W. Li, P.A. Webley, D. Zhao, Adv. Mater. 24 (2012) 485–491.[4] L. Sun, C. Tian, L. Wang, J. Zou, G. Mu, H. Fu, J. Mater. Chem. 21 (2011) 7232–

7239.[5] S.A. Al-Muhtaseb, J.A. Ritter, Adv. Mater. 15 (2003) 101–114.[6] S.-R. Guo, J.-Y. Gong, P. Jiang, M. Wu, Y. Lu, S.-H. Yu, Adv. Funct. Mater. 18

(2008) 872–879.[7] Z. Zapata-Benabithe, F. Carrasco-Marin, C. Moreno-Castilla, J. Power Sources

219 (2012) 80–88.[8] D. Long, X. Liu, W. Qiao, R. Zhang, L. Zhan, L. Ling, J. Non-Cryst. Solids 355

(2009) 1252–1258.[9] H.-C. Chien, W.-Y. Cheng, Y.-H. Wang, S.-Y. Lu, Adv. Funct. Mater. 22 (2012)

5038–5043.[10] J.-B. Lee, K.-K. Park, S.-W. Yoon, P.-Y. Park, K.-I. Park, C.-W. Lee, Desalination

237 (2009) 155–161.[11] S.N. Schiffres, K.H. Kim, L. Hu, A.J.H. McGaughey, M.F. Islam, J. Malen, Adv.

Funct. Mater. 22 (2012) 5251–5258.[12] R. Gosalawit-Utke, T.K. Nielsen, K. Pranzas, I. Saldan, C. Pistidda, F. Karimi, D.

Laipple, J. Skibsted, T.R. Jensen, T. Klassen, M. Dornheim, J. Phys. Chem. C 116(2012) 1526–1534.

[13] W. Zhang, Z.-H. Huang, C. Zhou, G. Cao, F. Kang, Y. Yang, J. Mater. Chem. 22(2012) 7158–7163.

[14] S. Yang, Y. Cai, Y. Cheng, C.V. Varanasi, J. Liu, J. Power Sources 218 (2012) 140–147.

[15] X. Mi, G. Huang, W. Xie, W. Wang, Y. Liu, J. Gao, Carbon 50 (2012) 4856–4864.[16] A.K. Meena, G.K. Mishra, P.K. Rai, C. Rajagopal, P.N. Nagar, J. Hazard. Mater.

B122 (2005) 161–170.[17] L.-S. Zhong, J.-S. Hu, H.-P. Liang, A.-M. Cao, W.-G. Song, L.-J. Wan, Adv. Mater.

18 (2006) 2426–2431.[18] V. Chandra, J. Park, Y. Chun, J.W. Lee, I.-C. Hwang, K.S. Kim, ACS Nano 4 (2010)

3979–3986.[19] Z. Gu, J. Fang, B. Deng, Environ. Sci. Technol. 39 (2005) 3833–3843.[20] L. Pilato (Ed.), Phenolic Resins: A Century of Progress, Springer-Verlag, Berlin

Heidelberg, Germany, 2010, p. 109.[21] M. Biswas, E. McGlynn, M.O. Henry, M. McCann, A. Rafferty, J. Appl. Phys. 105

(2009) 094396.[22] Y.-F. Lin, J.-H. Chen, S.-H. Hsu, H.-C. Hsiao, T.-W. Chung, K.-L. Tung, J. Colloid

Interface Sci. 368 (2012) 660–662.