Removal and recovery of mercury from aqueous

Applied Surface Science 257 (2011) 4754–4759

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Removal and recovery of mercury from aqueous solution using magnetic silica nanocomposites Bo Yune Song, Yujin Eom, Tai Gyu Lee ∗ Department of Chemical and Biomolecular Engineering, Yonsei University, 262 Seongsanno, Seodaemun-gu, Seoul 120-749, Republic of Korea

a r t i c l e

i n f o

Article history: Received 8 November 2010 Received in revised form 31 December 2010 Accepted 31 December 2010 Available online 4 January 2011 PACS: 61.66.Fn 75.50.Tt 81.07.Pr 96.30.Dz Keywords: Cobalt ferrite Silica-CoFe2 O4 Magnetic silica nanocomposite Thiol Mercury

a b s t r a c t Thiol-functionalized magnetic silica nanocomposite was synthesized and tested for its mercury pick-up capability in aqueous solution. Magnetic property was to be utilized upon the collection of the adsorbents and the recovery adsorbed Hg by subsequent separation process. Cobalt ferrite nanoparticle, the core of magnetic silica nanocomposite, was synthesized using a thermal decomposition method and grown to a particle having an average size of 13 nm. The dispersed nanoparticles were then further arranged into spherical groups using a nanoemulsion method to enhance the reactivity toward magnets followed by tetraethyl orthosilicate coating using a modified Stöber method. The pore structure was modified by an additional coating of cetyltrimethylammonium bromide and tetraethyl orthosilicate. Finally, the surface of the magnetic silica nanocomposite was functionalized with thiol group. When tested for mercury adsorption capacity, a sufficiently high Hg adsorption capacity of 19.79 mg per g of adsorbent was obtained at room temperature and a pH of 5.5. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Recently, silica based mesoporous materials have been studied in many areas such as catalysts, absorbents, sensor technologies and drug delivery systems [1,2] as a result of its excellent characteristics including high surface area, large pore volume, and narrow and tunable distribution of pore sizes and is amenable to surface modification with various organic materials such as alkyls, thiols, amines, sulfonic and carboxylic acids, alkoxy groups, or aromatic groups. Organic functionalization improves the thermal, chemical, and physical stabilities of mesoporous material. Several studies on the use of functionalized silica particles for the removal of mercury ions from contaminated aqueous streams have been reported [3–10]. Among them, functionalized mesoporous silica with thiol functional group showed exceptionally high selectivity and removability for Hg(II) [3,4,11–13]. However, unlike the Hg pick up, not much attention has been given to the recovery of the removed Hg from the aqueous solution until now.

∗ Corresponding author. Tel.: +82 2 2123 5751; fax: +82 2 312 0560. E-mail address: [email protected] (T.G. Lee). 0169-4332/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2010.12.156

In addition, magnetic nanoparticles have been extensively studied due to their numerous applications in catalysis, magnetic fluids, biotechnology/biomedicine, magnetic resonance imaging, and drug delivery systems, etc. [14–19]. Among many magnetic nanoparticles, cobalt ferrite has attracted much attention because of its high cubic magnetocrystalline anisotropy, coercivity, and saturation magnetisation [20]. The obvious advantage of using magnetic nanoparticles resides in its controllability and subsequent recoverability in the aqueous solution upon using appropriate magnet. Once desirable amount of Hg ions are removed from the wastewater stream, the silica magnetic nanoparticles are collected using magnet and the adsorbed Hg is further separated and recovered. In this study, magnetized silica nanocomposites are functionalized and tested for their ability to adsorb Hg in aqueous solution. First, CoFe2 O4 nanoparticle, the core of magnetic silica nanocomposite, is synthesized using a thermal decomposition method and grown further having an average size of 13 nm. The dispersed nanoparticles are then further arranged into a spherical group followed by tetraethyl orthosilicate (TEOS) coating. The pore structure is secured by additional coatings of cetyltrimethylammonium bromide (CTMABr) and TEOS. Finally, the surface of the magnetic silica nanocomposite is modified with thiol group. The prepared sam-

B.Y. Song et al. / Applied Surface Science 257 (2011) 4754–4759

4755

Fig. 1. A mechanistic illustration of the magnetic silica nanocomposite synthesis.

ples were analyzed by X-ray diffractometer (XRD), vibrating sample magnetometer (VSM), Fourier transform infrared spectroscopy (FT-IR), N2 adsorption–desorption isotherms, transmission electron microscopy (TEM), and atomic force microscopy (AFM). Also, Hg adsorption capacities of CoFe2 O4 nanoparticles, silica-coated CoFe2 O4 nanoparticles, and magnetic silica nanocomposites are evaluated and discussed. A cold vapour atomic absorption (CVAA) spectrophotometer was used for Hg measurement. 2. Experimental 2.1. Synthesis of CoFe2 O4 nanoparticles Fig. 1 is a mechanistic illustration of the magnetic silica nanocomposite synthesis. Cobalt ferrite nanoparticles were synthesized by the thermal decomposition method [21], in which 0.71 g of iron(III) acetylacetonate, 0.26 g of cobalt(II) acetylacetonate, and 2.58 g of 1,2-hexadecanediol were dissolved in 1.9 ml of oleic acid, 2.87 ml of oleylamine, and 20 ml of benzyl ether under a nitrogen atmosphere. The mixture was heated to 200 ◦ C for 120 min and then heated to 300 ◦ C for 60 min under a nitrogen blanket. At this point, the mixture was a black solution, which was cooled at room temperature and purified with 40 ml of ethanol. The synthesized cobalt ferrite particles were dispersed in 30 ml of hexane. 2.2. Synthesis of silica-coated CoFe2 O4 nanoparticles In order to synthesize silica-coated cobalt ferrite nanoparticles, 5 mg of CoFe2 O4 nanoparticles was dissolved in 10 ml of chloroform, followed by vaporisation of the solvent. After the chloroform was completely vaporised, the remaining particles were added to 20 ml of amphipathic solution containing 600 mg of polyvinyl acetate [19]. The reactants were emulsified with an ultrasonicator and homogeniser for 20 min. The emulsified solution was purified by centrifugation at 15,000 rpm for 30 min. The resulting particles were then coated with TEOS using a modified Stöber method [22–24], in which 60 ␮l of TEOS was added to a solution

of de-ionized (DI) water and ammonia containing the synthesized particles. The reaction was complete after 10 h. 2.3. Synthesis of magnetic silica nanocomposites Once cobalt ferrite particles were coated with TEOS, they were centrifuged at 15,000 rpm for 30 min, dispersed in 5 ml of DI water, and heated to 80 ◦ C for 30 min. Then, 0.015 g of CTMABr, 0.04 ml of NaOH (2 M), and 50 ␮l of TEOS were added and allowed to react for 2 h at 80 ◦ C. The resulting solution was centrifuged at 14,000 rpm for 20 min and dried in a vacuum oven at 80 ◦ C for 7 h. Next, 40 ml of ethanol was added to the dried particles and allowed to react at 80 ◦ C for 24 h. The solution was once again centrifuged at 14,000 rpm for 20 min and rinsed with 5 ml of DI water. The solution containing dispersed mesoporous magnetic silica nanoparticles were heated to 80 ◦ C for 30 min, and 10 ␮l of 3-mercaptopropyl-trimethoxysilane was quickly added. 2.4. Measurements The structural characteristics of CoFe2 O4 nanoparticles, silicacoated CoFe2 O4 nanoparticles, and magnetic silica nanocomposite were analyzed by XRD (MiniFlex, Rigaku) using nickel-filtered ˚ Magnetic properties were meaCu K␣ radiation ( = 1.5406 A). sured by VSM (Model 7304, LakeShore Cryotronics, Inc., USA) with a maximum magnetic field of 10,000 Oe. For identification of chemical functional groups, FT-IR (Nicolet Magna 550 Series II, Midac, USA) were obtained with a KBr pellet in the range 4000–400 cm−1 . The surface area, pore diameter and pore volume were acquired by BET and Barrett–Joyner–Halenda’s equations using N2 adsorption–desorption isotherms (ASAP2020, Micromeritics, USA). The characterization by images was carried out using AFM (Dimension 3100 SPM, Digital Instruments, USA) and TEM (JEM-2100, JEOL, Japan). As for AFM, first, solutions of DI water/silica-coated CoFe2 O4 nanoparticles and DI water/magnetic silica nanocomposites were individually prepared. Each solution was then sprayed onto the surface of gold-coated silicon wafer. After letting the solution react with gold-coated silicon wafer, the

4756


Fig. 2. XRD patterns of (a) CoFe2 O4 nanoparticles and (b) silica-coated CoFe2 O4 nanoparticles.

Fig. 3. Saturation magnetization measurement of (a) CoFe2 O4 nanoparticles and (b) silica-coated CoFe2 O4 nanoparticles.

wafer was washed with DI water 5 times. Finally, the wafer surface was analyzed using AFM.

slightly less than that of uncoated sample, whereas the coercivity of coated sample is significantly larger than that of uncoated one. This is probably due to the formation of non-crystalline silica on the surface of CoFe2 O4 nanoparticle as well as the restriction of domain wall motion by the magnetic dilution effect of inert silica. Fig. 4 shows the FT-IR spectra of (a) mesoporous magnetic silica nanoparticles and (b) magnetic silica nanocomposites. The absorption band around 3400 cm−1 corresponds to O–H vibrations and is clearly observed in both spectra [26]. After the attachment of thiol groups, two strong-intensity peaks were observed at 2900 cm−1 and 2850 cm−1 in Fig. 4I(b) as well as a weak-intensity peak at 2600–2550 cm−1 in Fig. 4II(b). Peaks at 2900 cm−1 and 2850 cm−1 were attributed to the C–H stretch of the methylenes of the alkyl chain [27] and the weak-intensity peak at around 2600–2550 cm−1 was attributed to the S–H stretching vibration [26,27]. The FT-IR spectra indicate that the thiol groups are present in the magnetic silica nanocomposites prepared in this study. Fig. 5 shows N2 adsorption–desorption isotherms of (a) mesoporous magnetic silica nanoparticles and (b) magnetic silica nanocomposites measured at 77 K. The pore size, specific surface area, and pore volume of mesoporous magnetic silica nanoparticles are 5.52 nm, 231.79 m2 /g, and 0.32 cm3 /g, respectively—typical values for mesoporous silica materials [28]. As for magnetic silica nanocomposites, the pore size, specific surface area, and pore vol-

2.5. Removal of mercury in aqueous solution To test for their mercury adsorption capacity, the magnetic silica nanocomposites were added into Hg solutions with various concentrations and remained in the solutions for 60 min. The Hg solutions were prepared using a standard Hg solution (1641d, NIST). After adsorption of Hg, the magnetic silica nanocomposites were separated from the solution using a strong magnet, and the remaining solution was analyzed for its Hg concentration. The amount absorbed by the magnetic silica nanocomposite was determined by measuring Hg concentrations of the solution before and after the adsorption. Hg concentration was analyzed by US EPA method 7470A and CVAA Hg analyzer (RA-915+ , Lumex Ltd., Russia) [25]. To prepare the samples for Hg analysis, an aliquot of Hg solution was diluted to 100 ml followed by the addition of 5 ml of sulphuric acid, 2.5 ml of nitric acid, and 15 ml of potassium permanganate solution. After making sure the purple colour persisted for 15 min, 8 ml of potassium persulfate was added, and the mixture was heated for 2 h at 95 ◦ C. The sample was then cooled at room temperature, and 6 ml of sodium chloride–hydroxylamine sulphate solution was added to reduce the excess permanganate. Once the sample preparation steps were completed, 5 ml of stannous sulphate was added to reduce Hg to its elemental form. Finally, the amount of Hg in the sample was measured using the CVAA Hg analyzer. 3. Results and discussion 3.1. Physical and chemical characterizations The XRD images of (a) CoFe2 O4 nanoparticles and (b) silicacoated CoFe2 O4 nanoparticles are shown in Fig. 2. The samples were scanned from 20o to 80o (2) in steps of 0.2o /min. The peaks at 2 = 30.2, 35.6, 43.0, 53.4, 57.1, and 62.5 were well matched to spinel ferrite d(2 2 0), d(3 1 1), d(4 0 0), d(4 2 2), d(5 1 1), and d(4 4 0) peaks, respectively. The broad band at 2 = 20–30o in Fig. 2(b) corresponds to amorphous silica spheres. These results indicate the successful coating of silica on the cobalt ferrite nanoparticles. The magnetic properties of the sample were measured by VSM. As seen in Fig. 3, the saturation magnetization of coated sample is

Fig. 4. FT-IR spectra of (a) mesoporous magnetic silica nanoparticles.and (b) magnetic silica nanocomposites (I: 4100–2000 nm, II: 2700–2400 nm).


4757

Fig. 6 shows the TEM images of (a) CoFe2 O4 nanoparticles, (b) fully grown CoFe2 O4 nanoparticles, (c) PVA-treated CoFe2 O4 nanoparticles, and (d) silica-coated CoFe2 O4 nanoparticles. CoFe2 O4 nanoparticles shown in Fig. 6(a) were further grown having an average particle size of 13 nm (Fig. 6(b)). Fig. 6(c) shows that dispersed nanoparticles are arranged into circular groups upon PVA treatment. The resultant silica-coated CoFe2 O4 nanoparticles are shown in Fig. 6(d). These images all confirmed that each synthesis step was successfully completed. 3.2. Removal of mercury in aqueous solution

Fig. 5. N2 adsorption–desorption isotherms of (a) mesoporous magnetic silica nanoparticles and (b) magnetic silica nanocomposites.

ume are 10.32 nm, 61.75 m2 /g, and 0.16 cm3 /g, respectively. The decreases in both specific surface area and pore volume of the magnetic silica nanocomposites compared to those of mesoporous magnetic silica nanoparticles are probably due to the presence of thiol functional groups.

Prepared CoFe2 O4 nanoparticles, silica-coated CoFe2 O4 nanoparticles, and magnetic silica nanocomposites were then tested for their ability to pick up Hg in aqueous solution. Table 1 shows much higher Hg removal capacity for magnetic silica nanocomposites than CoFe2 O4 nanoparticles or silica-coated CoFe2 O4 nanoparticles. The highest Hg adsorption achieved in this study was 19.79 mg Hg per 1 g of adsorbent which is higher than the value (14 mg Hg per 1 g of adsorbent) previously reported by Dong et al. [29]. These results suggest that thiol groups, only present in magnetic silica nanocomposites, must play an important role in Hg adsorption. This was further confirmed by the AFM images of (a) silica-coated CoFe2 O4 nanoparticles and (b) magnetic silica nanocomposites obtained at room temperature (Fig. 7). The images show that only magnetic silica nanocomposites are present

Fig. 6. TEM images of nanoparticles: (a) CoFe2 O4 , (b) fully grown CoFe2 O4 , (c) PVA-treated CoFe2 O4 , and (d) silica-coated CoFe2 O4 .

4758


Table 1 Comparison of the Hg adsorption ability of synthesised nanoparticles and nanocomposites at a Hg concentration of 560 ␮g/l.

Amount of adsorbent [mg] Amount of Hg adsorption per gram of adsorbent [mg/g]

Silica-coated CoFe2 O4 nanoparticles

Mesoporous magnetic silica nanoparticles

Magnetic silica nanocomposites

0.9 0.79

0.9 1.2

0.9 19.79

Fig. 7. AFM images of (a) as synthesis non-reacted silica-coated CoFe2 O4 nanoparticles and (b) magnetic silica nanocomposites on gold-coated silicon wafer. Table 2 The mercury adsorption capacity of magnetic silica nanocomposites at various concentrations of Hg in the solution. Test no.

Amount of Hg adsorption per gram of adsorbent [mg/g]

1 2 3

on the surface of the gold-coated silicon wafer by chemical reaction between thiol groups of the magnetic silica nanocomposite and Au atoms of the wafer [30]. Magnetic silica nanocomposites were further tested at various initial Hg concentrations and the results are shown in Table 2. 4. Conclusions First, CoFe2 O4 nanoparticles were synthesized using a thermal decomposition method followed by silica-coating using a modified Stöber method. Next, additional silica coating was carried out to secure the pores of the CoFe2 O4 nanoparticles. Finally, the surface of silica layer was chemically modified by a thiol group. Since prepared nanocomposites have cobalt–ferrite cores, their movement can be controlled by external magnetic forces. Magnetic silica nanocomposites were also tested for their Hg adsorption capacity and the results showed that they can be effectively used as Hg adsorbent with the highest adsorption capacity of 19.79 mg Hg per 1 g of adsorbent. Therefore, once the magnetic silica nanocomposites remove Hg from the aqueous solution, Hg recovery can be easily achieved by collection of adsorbents using appropriate magnets and by subsequent separation process. Acknowledgements This work was supported by Ministry of Environment as the EcoTechnopia 21 project (No. 2007-01003-0059-0) and by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (No. 2009-0079977).

Hg concentration in the solution [␮g/l] 6.65

73

560

13.1 12.9 13.1

16.86 16.94 16.97

19.79 17.36 16.95

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.apsusc.2010.12.156.

References [1] Y.L.F. Giraldo, B.L. López, L. Pérez, S. Urrego, L. Sierra, M. Mesa, Macromol. Symp. 258 (2007) 129. [2] E.J. Kwon, T.G. Lee, Appl. Surf. Sci. 254 (2008) 4732. [3] X. Feng, G.E. Fryxell, L.Q. Wang, A.Y. Kim, J. Liu, K.M. Kemmer, Science 276 (1997) 923. [4] L. Mercier, T.J. Pinnavaia, Environ. Sci. Technol. 32 (1998) 2749. [5] X. Chen, X. Feng, J. Liu, G.E. Fryxell, M. Gong, Sep. Sci. Technol. 34 (1999) 1121. [6] S.V. Mattigod, X. Feng, G.E. Fryxell, J. Liu, M. Gong, Sep. Sci. Technol. 34 (1999) 2329. [7] A.M. Liu, K. Hidajat, S. Kawi, D.Y. Zhao, Chem. Commun. (2000) 1145. [8] V. Antochshuk, M. Jaroniec, Chem. Commun. (2002) 258. [9] D. Perez-Quintanilla, I. Del Hierro, M. Fajardo, I. Sierra, Micropor. Mesopor. Mater. 89 (2006) 58–68. [10] D. Perez-Quintanilla, I. Del Hierro, M. Fajardo, I. Sierra, J. Hazard. Mater. 134 (2006) 245–256. [11] J. Aguado, J.M. Arsuaga, A. Arencibia, Micropor. Mesopor. Mater. 109 (2008) 513–524. [12] A. Walcarius, C. Delacote, Anal. Chim. Acta 547 (2005) 3. [13] A. Arencibia, J. Aguado, J.M. Arsuaga, Appl. Surf. Sci. 256 (2010) 5453. [14] D. Zins, V. Cabuil, R. Massart, J. Mol. Liq. 83 (1999) 217. [15] E. Manova, T. Tsoncheva, D. Paneva, I. Mitov, K. Tenchev, L. Petrov, Appl. Catal. Gen. 277 (2004) 119. [16] H.X. Jin, L. Li, N.J. Chu, Y.P. Liu, L.Y. Wang, Q. Lu, J. Qian, L.N. Sun, Q. Tang, H.L. Ge, X.Q. Wang, Mater. Chem. Phys. 112 (2008) 112. [17] H.M. Reinl, M. Peller, M. Hagmann, P. Turner, R.D. Isselsb, M. Reisera, Magn. Reson. Imag. 23 (2005) 1017. [18] M.P. Sharrock, L.W. Carlson, IEEE Trans. Magn. 31 (1995) 2871. [19] J. Yang, S.-B. Park, H.-G. Yoon, Y.-M. Huh, S. Haam, Int. J. Pharm. 324 (2006) 185. [20] L. Zhao, H. Yang, Y. Cui, X. Zhao, S. Feng, J. Mater. Sci. 42 (2007) 4110.

B.Y. Song et al. / Applied Surface Science 257 (2011) 4754–4759 [21] S. Sun, H. Zeng, D.B. Robinson, S. Raoux, P.M. Rice, S.X. Wang, G. Li, J. Am. Chem. Soc. 126 (2004) 273. [22] W. Stöber, A. Fink, E. Bohn, J. Colloid Interface Sci. 26 (1968) 62. [23] Y. Kobayashi, S. Saeki, M. Yoshida, D. Nagao, M. Konno, J. Sol–Gel Sci. Technol. 45 (2008) 35. [24] S. Liu, L. Lu, Z. Yang, P. Cool, E.F. Vansant, Mater. Chem. Phys. 97 (2006) 203. [25] US EPA Method 7470A, Mercury in liquid waste (Manual Cold-Vapor Technique) 1994.

4759

[26] G. Socrates, Infrared Characteristic Group Frequencies, 2nd ed., John Wiley & Sons, Chishester, 1994. [27] W. Yantaseem, Y. Lin, X. Li, G.E. Fryxell, T.S. Zemanian, V.V. Viswanathan, Analyst 128 (2003) 899. [28] L.S. Li, Y. Wang, D.J. Young, S.C. Ng, T.T.Y. Tan, Electrophoresis 31 (2010) 378. [29] J. Dong, Z. Xu, F. Wang, Appl. Surf. Sci. 254 (2008) 3522. [30] G. Schmid, S. Peschel, T. Sawitowski, Z. Anorg. Allg. Chem. 623 (1997) 719.