Nanocrystalline cobalt oxide immobilized on titanium dioxide nanoparticles for the heterogeneous activation of peroxymonosulfate

Applied Catalysis B: Environmental 74 (2007) 170–178 www.elsevier.com/locate/apcatb

Nanocrystalline cobalt oxide immobilized on titanium dioxide nanoparticles for the heterogeneous activation of peroxymonosulfate Qiujing Yang, Hyeok Choi, Dionysios D. Dionysiou * Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, OH 45221-0071, USA Received 13 November 2006; received in revised form 26 January 2007; accepted 1 February 2007 Available online 6 February 2007

Abstract Recently, sulfate radical-based advanced oxidation technologies have shown significant implications for environmental remediation to decompose water pollutants. In this study, we evaluated the performance of heterogeneous activation of peroxymonosulfate (PMS) to generate sulfate radicals using cobalt catalyst immobilized on titanium dioxide nanoparticles (Co/TiO2). The Co/TiO2 catalyst was prepared via an incipient wetness impregnation method employing Degussa P-25 TiO2 and Co(NO3)6H2O. The activity of Co/TiO2 system was compared with those of Co(NO3)2 solution for homogeneous PMS activation and neat Co3O4 for heterogeneous PMS activation. More emphasis was given to the effect of cobalt loading and heat treatment on the physicochemical properties of Co/TiO2 and cobalt leaching. The results showed that heat treatment of Co/ TiO2 at 500 8C, where cobalt existed as Co3O4, induced negligible Co leaching and enhanced catalytic activity to decompose 2,4-dichlorophenol. The Co/TiO2 catalyst at Co/Ti molar ratio of 0.1 showed the highest activity via heterogeneous PMS activation. On the other hand, Co/TiO2 catalysts with Co/Ti molar ratio of above 0.2 exhibited rather much lower activity which was initiated predominantly via a homogeneous pathway from leached cobalt, although they contained considerable amounts of Co3O4. The formation of Co–OH complexes at the surface of Co/TiO2 nanoparticles, due to the ability of TiO2 to dissociate H2O for the formation of surface hydroxyl groups, was proposed to facilitate the heterogeneous PMS activation. However, high cobalt loading covering the TiO2 surface diminished the beneficial role of TiO2 due to the reduction in the concentration of surface hydroxyl groups and thus decreased the heterogeneous PMS activation. The activity of Co3O4 in Co/TiO2 catalysts was much higher than that of neat Co3O4 due to the presence of surface hydroxyl groups and uniform distribution of well-defined 1015 nm nanocrystalline Co3O4 particles at the surface of 3040 nm TiO2 nanoparticles. # 2007 Elsevier B.V. All rights reserved. Keywords: 2,4-Dichlorophenol; Advanced oxidation processes (AOPs); Advanced oxidation technologies (AOTs); Calcination; Co3O4; Cobalt; Co–OH complexes; Heterogeneous reaction; Homogeneous reaction; Leaching; Nanoparticles; Oxone; Peroxymonosulfate; Sulfate radicals; Surface hydroxyl groups; Titanium dioxide; Titania

1. Introduction Recently, the growing interests in sulfate radical-based advanced oxidation technologies (AOTs) are driven by the increasing demand of cost-effective and environmentally benign routes for wastewater treatment [1–3]. Sulfate radicals, generated by catalytic decomposition of peroxymonosulfate (PMS) in a homogeneous pathway with the aid of transition metals and/or UV radiation, have been proven to be strong oxidizing species which are capable of readily attacking and decomposing recalcitrant organic molecules in water into nontoxic species, such as CO2 and H2O [4–8]. Of the transition metals investigated, cobalt ions (Co2+) showed the best * Corresponding author. Tel.: +1 513 556 0724; fax: +1 513 556 2599. E-mail address: [email protected] (D.D. Dionysiou). 0926-3373/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.apcatb.2007.02.001

performance [5]. There are a few studies on the generation mechanism of sulfate radicals by cobalt-catalyzed decomposition of PMS in the homogeneous system [4,9,10]: Co2þ þ H2 O $ CoOHþ þ Hþ

(1)

CoOHþ þ HSO5  ! CoOþ þ SO4  þ H2 O

(2)

CoOþ þ 2Hþ $ Co3þ þ H2 O

(3)

Co3þ þ HSO5  ! Co2þ þ SO5  þ Hþ

(4)



SO4 þ organics ! ½stable and unstable intermediate products ! CO2 þ H2 O (5) Although this system is promising for effectively abating environmental pollution in water, health concerns associated

Q. Yang et al. / Applied Catalysis B: Environmental 74 (2007) 170–178

with the adverse effect of dissolved cobalt in water still need to be addressed. Consequently, it is beneficial to activate PMS via a heterogeneous manner that may prevent the problem with dissolved or leached cobalt present in the polished water. In a recent article, Anipsitakis et al. explored this concept using commercially available Co3O4 and demonstrated the heterogeneous PMS activation [11]. The Co3O4, however, is not fully utilized for the PMS activation since it exists as large macroparticles, which reduces dramatically the available surface area of the catalyst. Consequently, the purpose of this study is to improve the catalytic activity of cobalt for the heterogeneous PMS activation by controlling the physicochemical properties of cobalt or Co3O4 material at the nanolevel. In this study, we focused on immobilizing cobalt catalyst at the surface of TiO2. TiO2 nanoparticles were selected as the support material because: (i) the heterogeneous catalytic activity of cobalt can be maximized by evenly distributing cobalt on the TiO2 nanoparticles with high surface area, (ii) deposition of cobalt on TiO2 nanoparticles is the first step towards the development of a novel AOT where sulfate radicals and hydroxyl radicals can be generated simultaneously in the presence of UV radiation from PMS and TiO2, respectively, and (iii) TiO2 as a support material is environmental friendly because it is relatively nontoxic and it can be recovered or used in immobilized form [12,13]. Interestingly, due to the well-known ability of TiO2 to dissociate water for the formation of surface hydroxyl groups [12], TiO2 as a support material might play a crucial role in the heterogeneous PMS activation by facilitating the formation of surface Co–OH groups, as depicted in Eqs. (1) and (2). Co-based catalysts are widely used in CO hydrogenation or Fischer–Tropsch synthesis [14–18]. To the best of our knowledge, Co/TiO2 nanoparticles acting as an activator of PMS have never been reported so far. In this study, we synthesized Co/TiO2 nanoparticles via an incipient wetness impregnation method. More emphasis was given to the effect of Co loading at the TiO2 surface and calcination temperature on the physicochemical properties of Co/TiO2 and leaching behavior of cobalt. The heterogeneous PMS activation using Co/TiO2 was examined in terms of 2,4-dichlorophenol (2,4-DCP) decomposition. 2,4-DCP is a well-known intermediate during the synthesis and degradation of 2,4-dichlorophenoxyacetic acid that is a regulated and widely used herbicide [19]. The chemical has been included in the drinking water contaminant candidate list by the USEPA because it is toxic, hardly biodegradable, and difficult to remove from the environment [20]. Finally, a possible mechanism for the heterogeneous PMS activation over Co/TiO2 was proposed, based on the widely accepted homogeneous pathway, and the crucial role of TiO2 to generate surface hydroxyl groups was suggested. 2. Experimental 2.1. Preparation of Co/TiO2 catalysts In order to immobilize Co at the surface of TiO2 nanoparticles and control the properties of Co/TiO2 catalysts

171

at the nanolevel, we adapted an incipient wetness impregnation method. A desired amount of Co(NO3)6H2O (99.3%, Sigma) was dissolved in distilled water and then impregnated on Degussa P-25 TiO2 nanoparticles (anatase/rutile mixture; BET surface area: 50 m2/g; particle size: 30 nm) under vigorous stirring. This suspension was stirred for 24 h, and then dried under an infrared lamp at alleviated temperature of 50 8C to remove water. Finally, in order to control their crystal phase and enhance their mechanical stability, the catalysts were calcined in a furnace (Paragon model HT-22D, Thermcraft) at various temperatures ranging from 300 to 700 8C for 4 h with a ramp rate of 600 8C/h. The resulting Co/TiO2 catalyst was ground thoroughly and labeled as XCo/TiO2–Y, where X stands for the molar ratio of Co to Ti and Y stands for the calcination temperature. Heterogeneous PMS activation and homogeneous PMS activation are denoted as hetero-PMS-Act and homoPMS-Act, respectively. 2.2. Evaluation of catalytic activity The catalytic activities of three different systems (Co/TiO2 catalyst for hetero-PMS-Act partially combined with homoPMS-Act at neutral pH (or basic pH), commercial neat Co3O4 powder for hetero-PMS-Act, and Co(NO3)2 solution for homoPMS-Act) were evaluated in terms of 2,4-dichlorophenol degradation (2,4-DCP, 99%, Acros Organics). First, a quartz rectangular reactor (base: 10 cm  10 cm; height: 25 cm) containing 1 L of 50 mg/L (0.307 mM) 2,4-DCP solution with adjusted pH of 7.0 using 5 mM NaHCO3 and KHSO4 was placed on a magnetic stirrer plate. After the addition of 0.1 g of Co/TiO2 catalyst in the solution, the solution was allowed to reach adsorption equilibrium between the catalysts and 2,4DCP (there was no considerable 2,4-DCP adsorption on the catalyst surface). Then, Oxone (KHSO5 as active component, Aldrich, manufactured by Dupont) was added into the solution at KHSO5/2,4-DCP molar ratio of 3:1. For measuring the concentration of 2,4-DCP during 2 h of reaction, 10 mL samples were withdrawn at specified time intervals and quenched with 5 mL of 2.47 M methanol (Aldrich) to prevent further reaction. The sample was filtered with 0.1 mm filter (Magna Nylon, Fischer) and analyzed using a High Performance Liquid Chromatograph (HPLC, Agilent 1100 Series) with a QuatPump and a photo-diode-array detector. The column was an Eclipse XDB-C8 column (Agilent) and mobile phase was 70:30% (v/v) of 0.01N H2SO4:acetonitrile. An atomic absorption spectrometer (Perkin-Elmer AA-300) equipped with an HGA-800 electrothermal atomizer and an AS-72 auto sampler was used to analyze the cobalt leaching from the catalyst. For the characterization of spent Co/TiO2 nanoparticles, the catalysts were collected after 2 h reaction and analyzed. In parallel, the activities of commercially available neat Co3O4 (CoOCo2O3, Fischer) for hetero-PMS-Act and Co(NO3)2 (Sigma) solution for homo hetero-PMS-Act were tested for comparison. The amounts of neat Co3O4, equivalent to 0.1 g of 0.1Co/TiO2 and 0.5Co/TiO2 were 9.1 and 33.4 mg, respectively, assuming that cobalt in Co/TiO2 exists in the form

172

Q. Yang et al. / Applied Catalysis B: Environmental 74 (2007) 170–178

of Co3O4. In the Co(NO3)2 system, the amounts used were the same as those from cobalt leaching in 0.1 g of 0.1Co/TiO2 and 0.5Co/TiO2 systems. Moreover, the activity of Co/TiO2 at basic pH of around 11.6 adjusted using NaHCO3 was tested. Unless otherwise specified, all the experiments were carried out at neutral pH. 2.3. Catalyst characterization The crystallographic structure of Co/TiO2 catalyst was investigated with X-ray diffraction (XRD) analysis using a Kristalloflex D500 diffractometer (Siemens) with Cu Ka ˚ ) radiation. The morphology and composition of (l = 1.5406 A the catalysts were measured by a JEM-2010F (JEOL) high resolution-transmission electron microscope (HR-TEM) with field emission gun at 200 kV, equipped with Oxford Isis energy dispersive X-ray spectroscope (EDX). The samples were dispersed in methanol using an ultrasonicator for 5 min and fixed on a carbon-coated copper grid (LC200-Cu, EMS). A Tristar 3000 (Micromeritics) surface area and porosimetry analyzer was used to determine the Brunauer–Emmett–Teller (BET) surface area and pore volume of Co/TiO2. The samples were purged with nitrogen gas for 2 h at 150 8C using SmartPrep Programmable degas system (Micromeritics). In order to determine the chemical states of surface oxygen and cobalt in the catalyst, X-ray photoelectron spectroscope (XPS, Perkin-Elmer model 5300) with Mg Ka X-rays was used. During analysis, the pressure was kept between 108 and 109 Torr. Charge correction was performed by referencing the C1s peak for hydrocarbons to a binding energy of 284.6 eV. Curve fitting of the XPS spectra was accomplished using a combination of 90% Gaussian and 10% Lorentzian peak shape. Atomic concentrations were obtained from the XPS spectra based on peak areas and sensitivity factors provided by the software (RBD Enterprises, Bend, OR, USA). UV–vis spectra were obtained under atmospheric conditions using a UV–vis spectrophotometer (Shimadzu 2501 PC) mounted with an integrating sphere attachment (ISR 1200) for diffuse reflectance measurement. Since all UV–vis spectra were recorded under identical conditions, the intensities of the signals could be correlated with the relative content of elemental species or chemical bindings.

Fig. 1. Removal efficiency of 2,4-DCP and concentration of leached cobalt after 2 h of reaction using 0.1Co/TiO2 calcined at different temperatures.

required for the homo-PMS-Act to achieve the same extend of degradation [6]. Co/TiO2-500 exhibited the highest activity and negligible Co leaching. For Co/TiO2-500 catalysts, the effect of cobalt loading in the rang 0.0–0.5 of Co:Ti molar ratio on their catalytic activity and Co leaching was explored and the results are shown in Fig. 2. The cobalt leaching increased with increasing Co loading, as expected, but the activity of Co/TiO2 catalysts did not follow the same trend. The highest degradation efficiency of 2,4-DCP was obtained at 0.1–0.2 of Co/Ti molar ratio. Low Co content in 0.001Co/TiO2 exhibited lower catalytic activity while high Co loading in 0.5Co/TiO2 induced detrimental effect on its catalytic activity. For more insights on this interesting finding, comparative experiments to this Co/TiO2 system were conducted, employing Co(NO3)2 solution for homo-PMS-Act as well as neat Co3O4 for hetero-PMS-Act. Fig. 3 shows 2,4-DCP transformation in the three different systems. As illustrated in Fig. 3(a), after 2 h of reaction, over 75% of 2,4-DCP (Curve I) was removed in 0.1Co/TiO2 system with cobalt leaching of 36 mg/L acting as a homogeneous catalyst. Homo-PMS-Act by Co(NO3)2 at the same concentration as the leached cobalt showed only 20% 2,4-DCP removal (Curve II). This result strongly indicates a hetero-PMS-Act by the Co/TiO2. It is worthy to note that 0.5Co/TiO2 system

3. Results and discussion 3.1. Catalytic activity of Co/TiO2 and cobalt leaching The effect of calcination temperature in the range of 300– 700 8C on the catalytic activity of 0.1Co/TiO2 and Co leaching was investigated and the results are shown in Fig. 1. In case of Co/TiO2-uncalcined catalysts, significant Co leaching as high as 2.4 mg/L was observed and complete decomposition of 2,4DCP was achieved within 2 h predominantly via homo-PMSAct by the leached cobalt. However, Co/TiO2-calcined catalysts exhibited significantly reduced Co leaching (lower than 50 mg/ L) and thus activated PMS via the heterogeneous pathway since at least around 1 mg/L of dissolved Co was reported to be

Fig. 2. Removal efficiency of 2,4-DCP and concentration of leached cobalt after 2 h of reaction using Co/TiO2 calcined at 500 8C with different Co:Ti ratios.

Q. Yang et al. / Applied Catalysis B: Environmental 74 (2007) 170–178

Fig. 3. 2,4-DCP transformation with (a) 0.1 g of 0.1Co/TiO2 calcined at 500 8C equivalent to (I) and (I0 ) Co/TiO2 catalyst (triplicates) at neutral and basic pH, respectively, (II) 36 mg/L of Co from Co(NO3)2 solution and (III) 9.1 mg neat Co3O4. (b) 0.1 g of 0.5Co/TiO2 calcined at 500 8C equivalent to (I) and (I0 ) Co/ TiO2 catalyst (triplicates) at neutral and basic pH, respectively, (II) 100 mg/L of Co from Co(NO3)2 solution and (III) 33.4 mg neat Co3O4.

(Fig. 3(b)-I) with much more cobalt loading showed lower catalytic activity than 0.1Co/TiO2 (Fig. 3(a)-I). More interestingly, the PMS activation in 0.5Co/TiO2 exhibited homogeneous nature since the activity of 0.5Co/TiO2 system, which induced 99.7  3.7 mg/L of cobalt leaching, is very close to that of Co(NO3)2 system (Fig. 3(b)-II). Thus, too high Co loading at the TiO2 surface inhibited the hetero-PMS-Act. In ideal heterogeneous systems employing neat Co3O4 (Fig. 3(a)-III and (b)-III), high cobalt dosage slightly improved 2,4-DCP decomposition. However, the activity of neat Co3O4 heterogeneous system was significantly lower than that of 0.1Co/TiO2 heterogeneous system. This implies that TiO2 as a support material for Co might play a crucial role in the hetero-PMSAct. Moreover, the degradation efficiencies of 2,4-DCP over 0.1Co/TiO2-500 and 0.5Co/TiO2-500 at pH = 11.6 (Fig. 3(a)-I0 and (b)-I0 ) were much lower than them at neutral pH (Fig. 3(a)-I and (b)-I), respectively. 3.2. Physicochemical properties of Co/TiO2 catalyst In order to understand some interesting findings observed in this hetero-PMS-Act by Co/TiO2 catalyst and elucidate the role

173

Fig. 4. XRD patterns of (a) 0.1Co/TiO2 and (b) 0.5Co/TiO2: (^) anatase, (&) rutile, (!) cobalt nitrate hydrate, (*) Co3O4, and (~) CoTiO3).

of TiO2 in this system, the physicochemical properties of Co/ TiO2 catalyst were investigated using XRD, HR-TEM, porosimetry, UV–vis, and XPS. As presented in Fig. 4 and summarized in Table 1, in Co/ TiO2 catalysts calcined at below 500 8C, the crystal phase of TiO2 was a mixture of anatase and rutile due to the nature of P25 TiO2 (mixture of 70% anatase and 30% rutile). Upon calcination at 700 8C, anatase was completely transformed to rutile. The characteristic peaks of cobalt nitrate hydrate (JCPDS 18-0425) were observed in Co/TiO2-uncalcined. When the calcination temperature increased up to 500 8C, the peaks for Co3O4 (JCPDS 42-1467) appeared at 2u of 31.58, 37.18, and 45.18. In case of 0.1Co/TiO2, however, the peaks for cobalt nitrate hydrate and Co3O4 are too weak to be resolved due to the low cobalt loading. The Co3O4 is a decomposition product of cobalt nitrate during high temperature heat treatment in air, which involves the release of nitrogen dioxide, water and oxygen [21]. In addition, CoTiO3 (JCPDS 15-0866) was formed in Co/TiO2-700 since cobalt ions were migrated into the TiO2 lattice during heat treatment [22]. The changes in cobalt species might be responsible for the variation of catalytic activity and cobalt leaching described in Figs. 1–3. In Co/TiO2-uncalcined, weak attachment of cobalt nitrate hydrate on TiO2 led to significant Co leaching, inducing a homogeneous system which achieved complete 2,4-DCP

174

Q. Yang et al. / Applied Catalysis B: Environmental 74 (2007) 170–178

Table 1 Structural properties of Co/TiO2 catalysts Conditions

SBETa (m2/g)

Crystallite size (nm)b Anatase

Rutile

Co3O4

CoTiO3

P-25 TiO2

55.3

28.2

40.0

–

–

0.1Co/TiO2 Uncalcined 500 8C 700 8C

32.9 42.2 7.46

21.5 21.9 –

37.3 34.3 39.5

– 14.0 –

– – 52.3

0.5Co/TiO2 500 8C 700 8C Neat Co3O4 c

37.9 8.48 11.5

22.2 – –

39.1 34.9 –

23.1 – 55.4

– 57.4 –

a

SBET: BET specific surface area. (1 0 1) for anatase TiO2, (1 1 0) for rutile TiO2, (2 2 0) for Co3O4, and (1 0 4) for CoTiO3. The crystallite size of TiO2 in Co/TiO2 catalysts was underestimated due to the scattering effect by Co. c Commercially available neat Co3O4, which was used for the comparative heterogeneous PMS activation experiments for Co3O4 in Co/TiO2 catalysts synthesized in this study. b

decomposition. The 0.1Co/TiO2-500, where most of Co existed as Co3O4, exhibited the highest activity among the Co/TiO2calcined catalysts, while Co/TiO2-300 showed lower activity due to low content of Co3O4. Furthermore, transformation of Co3O4 to CoTiO3 can be a possible reason for the low activity of

Co/TiO2-700. As shown in Fig. 1, heat treatment significantly reduced cobalt dissolution. But the cobalt leaching slightly increased with increasing calcination temperature most probably due to increase in the grain size of cobalt oxide. It is easily expected that the large cobalt oxide particles have reduced contact area to the surface of TiO2 support, compared to small cobalt oxide particles. As a result, the large cobalt oxide particles are susceptible to leaching from the catalyst. Fig. 5 shows HR-TEM images of Co/TiO2-500. Clear lattice fringes revealed crystal characteristics of TiO2 and Co3O4. When comparing the dspace values from the images, Co3O4 particles were differentiated from TiO2 particles. Fig. 5(b) shows such an example, suggesting dspace of 0.46 nm between the lattice fringes in small particles corresponds to the (1 1 1) plane of Co3O4 while dspace of 0.36 nm in large particles corresponds to the (1 0 1) plane of anatase crystal TiO2 [23–25]. The small 10–15 nm Co3O4 particles were well-incorporated into the edge of large 30– 40 nm TiO2 particles, forming a composite of Co3O4 and TiO2 with heterojunction structure. Although there was no pronounced difference in the texture properties between 0.1Co/TiO2 and 0.5Co/TiO2, more Co3O4 particles covering the TiO2 surface were frequently observed during HR-TEM analysis for 0.5Co/ TiO2 (Fig. 5(d)). This observation was further supported by the slightly lower BET specific surface area of 0.5Co/TiO2 (37.9 m2/ g) than that of 0.1Co/TiO2 (42.2 m2/g) (Table 1). The morphology of 0.1Co/TiO2-300 and 0.1Co/TiO2-700 is shown

Fig. 5. HR-TEM images of: (a and b) 0.1Co/TiO2 and (c and d) 0.5Co/TiO2 calcined at 500 8C. Note different magnifications.

Q. Yang et al. / Applied Catalysis B: Environmental 74 (2007) 170–178

175

Fig. 6. HR-TEM images of 0.1Co/TiO2 calcined at: (a) 300 8C and (b) 700 8C.

in Fig. 6. The grain size of 0.1Co/TiO2 catalysts was significantly increased from 20–25 nm for Co/TiO2-300 to larger than 70 nm for Co/TiO2-700. Fig. 7 shows UV–vis absorption spectra of Co/TiO2 catalysts. The absorption bands at 200–400 nm are due to the charge transfer from the valence band to the conduction band of TiO2 [26]. A broad band at 510 nm for Co/TiO2uncalcined corresponds to the 4T1g(F) ! 4T1g (P) transition in octahedral high-spin Co2+ complexes [21]. The absorption

bands at 740 nm for Co/TiO2-300 and Co/TiO2-500 are due to the presence of Co3O4 phase [27,28]. A broad band between 500–700 nm for Co/TiO2-700 can be assigned to CoTiO3 [29]. It is worthy to note that the relatively low absorption intensity at 200–400 nm for 0.5Co/TiO2-500 compared with that for 0.1Co/ TiO2-500 demonstrated that more Co3O4 covers the surface of TiO2 as discussed previously. Fig. 8 shows the chemical state of Co species in Co/TiO2500. The bands at 780.1 eV for Co 2p3/2, 796.1 eV for Co 2p1/2 were identical to those of Co3O4 [21,30–32]. Shake-up satellites II centered at 789.0 and 805.8 eV are also characteristic of the Co3O4. The peaks at 782.1 and 797.5 eV resulted from a chemical shift of the main spinorbit components because the Co cations on the nanocrystal surface are chemically interacted with surface hydroxyls which correspond to the shake-up satellites I at +5.6 eV from the Co 2p bands, suggesting a strong possibility of the presence of many hydroxyl groups on the surface of Co/TiO2 [31,32]. 3.3. Heterogeneous reaction mechanism and role of TiO2 During the 2,4-DCP decomposition over the Co/TiO2 catalysts, surface properties such as the chemical environment

Fig. 7. UV–vis spectra of: (a) 0.1Co/TiO2 and (b) 0.5Co/TiO2.

Fig. 8. XPS spectra of Co 2p for 0.1Co/TiO2 calcined at 500 8C.

176

Q. Yang et al. / Applied Catalysis B: Environmental 74 (2007) 170–178

Fig. 9. UV–vis spectra of (a) 0.1Co/TiO2 and (b) 0.5Co/TiO2 calcined at 500 8C.

of cobalt are substantially changed as a result of the cobaltmediated PMS activation. UV–vis absorption spectra of fresh and spent Co/TiO2-500 catalysts were compared and the results are shown in Fig. 9. A distinct change in UV–vis spectra at 600– 740 nm was observed in 0.5Co/TiO2. In this 600–700 nm range, Vakros et al. observed a shoulder peak at 640 nm which was assigned to the exchange of aqua ligands with Co–OH ligands, and the relative magnitude of this peak increased with increasing the Co(II) surface concentration [33]. However, in our study, the peak cannot be clearly seen since it is too close to be distinguished from the peak of Co3O4. The band at 600– 740 nm is ascribed to Co–OH complexes and Co3O4, and thus the reduced intensity of this absorption band for spent 0.5Co/ TiO2 suggests that Co–OH complexes might be consumed during the PMS activation [33,34]. In contrast, 0.1Co/TiO2 catalyst exhibited quite similar spectra before and after use. In order to determine the amount of surface hydroxyl groups, O 1s XPS spectra for fresh and spent Co/TiO2-500 were studied. As shown in Fig. 10, all O 1s spectra of Co/TiO2 are clearly asymmetric, indicative of the existence of different oxygen species at the surface. Besides the main peak at 530.0 eV (Peak I) corresponding to lattice oxygen species from TiO2 and Co3O4, a shoulder at higher binding energy of

531.8 eV (Peak II) is identified to surface hydroxyl groups (i.e. Ti–OH and Co–OH) that are ubiquitous in air-exposed cobalt oxide materials [31,32,35–38]. The content (%) of surface hydroxyl oxygen in the total oxygen for 0.5Co/TiO2 reduced more than two times, from 18% before use to 8% after use, whereas that for 0.1Co/TiO2 slightly decreased from 16% to 14%, which is consistent with the UV–vis results. In addition, spent Co/TiO2 had a resolved peak at around 533.2 eV (Peak III), which is attributed to the adsorbed oxygen species such as C–O bonds resulting from 2,4-DCP transformation and surface bound water [37,39]. 2,4-DCP transformation over Co/TiO2 catalysts demonstrated the different natures for PMS activation: heterogeneous pathway by 0.1Co/TiO2 versus homogeneous pathway by 0.5Co/TiO2 with rather more Co3O4. The principle of heterogeneous cobalt-mediated PMS activation is supposed to be analogous to that of homogeneous system. In principle, a homogeneous cobalt-mediated PMS activation involves one electron transfer between Co(II) and Co(III) as a result of PMS decomposition, as expressed in Eqs. (1) and (2). The first step of homogeneous cobalt-mediated PMS activation is obviously to form CoOH+ through the dissociation of H2O with Co2+. The formation of CoOH+ complexes is an essential step that significantly affects the subsequent behavior of the cobaltmediated PMS activation [5]. Similarly, the dissociation of H2O to generate Co–OH complexes should be the initial step to activate PMS in the heterogeneous system of Co/TiO2. In this study, the aforementioned homogeneous pathway was modified to explain a possible pathway for heterogeneous cobalt-mediated PMS activation in the Co/TiO2 system. TiO2 is well known to possess a strong ability to dissociate H2O for the formation surface hydroxyl groups [12,40–42]. This is accomplished through both basic sites (O2 sites) to abstract protons from H2O, and acidic sites (Ti4+ sites) to combine hydroxyl groups [41]. In 0.1Co/TiO2, H2O is readily dissociated to generate hydroxyl groups on the TiO2 surfaces exposed to H2O. As a result, hetero-PMS-Act by 0.1Co/TiO2 is achieved through the direct interaction of Co species with the nearby hydroxyl groups, rather than with H2O, to form Co–OH complexes. Thus, Co–OH complexes are continuously generated along with PMS activation and ensure the subsequent rapid 2,4-DCP transformation. This can be supported by the XPS and UV–vis results mentioned previously: the binding energy peak of surface hydroxyl groups in XPS and absorption band of Co– OH in UV–vis were not diminished after use of 0.1Co/TiO2. On the other hand, in case of 0.5Co/TiO2, the TiO2 surface was predominantly covered by cobalt oxide so that the area of exposed TiO2 surface available for H2O dissociation was limited, resulting in the remarkable inhibition of the interaction of Co species with hydroxyl groups to form Co–OH complexes. Thus, Co–OH complexes were gradually consumed upon 2,4DCP transformation and not regenerated promptly. As a result, PMS was activated mainly by low concentration of cobalt ions leached from the catalyst via a homogeneous pathway, resulting in the low overall catalytic activity. It has been reported that the activity of Co-based catalyst for 2,4-DCP degradation at neutral pH was much higher than that at

Q. Yang et al. / Applied Catalysis B: Environmental 74 (2007) 170–178

177

Fig. 10. XPS spectra of O 1s for: (a) fresh and (b) spent 0.1Co/TiO2 catalysts and (c) fresh and (d) spent 0.5Co/TiO2 catalysts. (I) Lattice oxygen, (II) surface hydroxyl groups and (III) adsorbed oxygen-containing species.

acidic pH both in homogeneous [4] and heterogeneous [11] systems. The activity of Co/TiO2 at basic pH was also tested here. As seen from Fig. 3, the degradation efficiencies of 2,4DCP over 0.1Co/TiO2-500 and 0.5Co/TiO2-500 at pH 11.6 were much lower than those at neutral pH. The point of zero charge of TiO2 is around 6.9. At neutral pH, TiOH species cover the surface of TiO2, whereas at pH 9, TiO is the dominant species at the surface of TiO2 [43]. Under basic conditions, the negatively charged species (TiO and OH) hinder the adsorption of water molecules (neutral charge) on the surface TiO2 particles, and thus the subsequent water dissociation by TiO2 is reduced. As a result, the surface hydroxyl groups cannot be generated effectively through the water dissociation at the surface of Co/TiO2 particles, leading to the lower activity for 2,4-DCP transformation. Overall, basic conditions might be favorable to the generation of CoOH+ (Eq. (1)), but retard the transformation of Co3+ to Co2+ (Eqs. (3) and (4)), and thus inhibit the subsequent reactions. In a previous study on hetero-PMS-Act by neat Co3O4 [11], it was hypothesized that the Co–OH complexes generated from interaction of cobalt species with water adsorbed on the surface of Co3O4 initiated a series of chain reactions. Although no direct study on the mechanism of hetero-PMS-Act by cobaltmediated catalysts has been reported so far, a similar reaction has been proposed for iron oxide. Lin et al. reported that the heterogeneous catalytic decomposition of hydrogen peroxide on iron oxide is initiated by the formation of a precursor surface complex of H2O2 with the oxide surface, FeIII–OH [44]. However, it is expected that the hydroxyl groups at the surface of Co3O4 cannot be regenerated rapidly from the interaction of

cobalt species with water molecules, especially at basic pH because the surface of Co3O4 is negatively charged due to its point of zero charge of around 7.5 [45]. The effect of Co3O4 on the catalytic activity should also be mentioned. The formation of Co3O4 in Co/TiO2-calcined catalysts was identified and its activity was much higher than that of neat Co3O4. This is mainly due to the surface hydroxyl groups generated from TiO2 and uniform distribution of welldefined 10–15 nm nanocrystalline Co3O4 particles on 30– 40 nm TiO2 nanoparticle surface, resulting in increase in catalytic surface area of Co–OH species to activate PMS. 4. Conclusions This study investigated the heterogeneous activation of peroxymonosulfate (PMS) over Co/TiO2 nanoparticles. Heat treatment of Co/TiO2 at 500 8C, where cobalt existed as Co3O4, induced negligible Co leaching and enhanced catalytic activity. The Co/TiO2 catalyst at Co/Ti molar ratio of 0.1 showed the highest catalytic activity via heterogeneous PMS activation. On the other hand, catalysts with high Co/Ti molar ratio of above 0.2 (although they contained considerable amounts of Co3O4) exhibited rather lower catalytic activity which was initiated predominantly via a homogeneous pathway from leached cobalt. The formation of Co–OH complexes at the surface of Co/TiO2 nanoparticles, due to the ability of TiO2 to dissociate H2O for the formation of surface hydroxyl groups, was proposed to facilitate the heterogeneous PMS activation. In addition, the Co–OH complexes could not be generated by OH in the solution since the contributions of surface hydroxyl

178

Q. Yang et al. / Applied Catalysis B: Environmental 74 (2007) 170–178

groups and OH to H2O dissociation are different due to their different charge properties. The activity of Co3O4 in Co/TiO2 catalysts was much higher than that of neat Co3O4 most probably due to the surface hydroxyl groups and the uniform distribution of well-defined 10–15 nm nanocrystalline Co3O4 particles at the surface of TiO2 nanoparticles with size of 30– 40 nm. However, high cobalt loading (i.e. Co/Ti molar ratio of 0.5) covering TiO2 surface diminished the beneficial role of TiO2 and thus decreased the heterogeneous PMS activation. Acknowledgments The authors are grateful to the National Science Foundation through a CAREER award (BES-0448117) and to DuPont through a Young Professor Award to D.D. Dionysiou. Q. Yang also thanks Mr. Yongjun Chen (University of Cincinnati) for his assistance with the XRD analysis. References [1] S. Malato, J. Blanco, C. Richter, B. Braun, M.I. Maldonado, Appl. Catal. B: Environ. 17 (1998) 347. [2] J. Fernandez, P. Maruthamuthu, J. Kiwi, J. Photochem. Photobiol. A: Chem. 161 (2004) 185. [3] Z.Y. Yu, L. Kiwi-Minsker, A. Renken, J. Kiwi, J. Mol. Catal. A: Chem. 252 (2006) 113. [4] G.P. Anipsitakis, D.D. Dionysiou, Environ. Sci. Technol. 37 (2003) 4790. [5] G.P. Anipsitakis, D.D. Dionysiou, Environ. Sci. Technol. 38 (2004) 3705. [6] G.P. Anipsitakis, D.D. Dionysiou, M.A. Gonzalez, Environ. Sci. Technol. 40 (2006) 1000. [7] G.P. Anipsitakis, D.D. Dionysiou, Appl. Catal. B: Environ. 54 (2004) 155. [8] J. Fernandez, P. Maruthamuthu, A. Renken, J. Kiwi, Appl. Catal. B: Environ. 49 (2004) 207. [9] J. Kim, J.O. Edwards, Inorg. Chim. Acta 235 (1995) 9. [10] Z. Zhang, J.O. Edwards, Inorg. Chem. 31 (1992) 3514. [11] G.P. Anipsitakis, E. Stathatos, D.D. Dionysiou, J. Phys. Chem. B 109 (2005) 13052. [12] A.L. Linsebigler, G. Lu, J.T. Yates Jr., Chem Rev. 95 (1995) 735. [13] M.R. Hoffman, S. Martin, W. Choi, D.W. Bahnemann, Chem. Rev. 95 (1995) 69. [14] A. Martı´nez, C. Lo´pez, F. Ma´rquez, I. Dı´az, J. Catal. 220 (2003) 486.

[15] M.E. Dry, Catal. Today 71 (2002) 227. [16] T.C. Bromfield, A.C. Vosloo, Macromol. Symp. 193 (2003) 29. [17] B. Jongsomjit, T. Wongsalee, P. Praserthdam, Mater. Chem. Phys. 97 (2006) 343. [18] A.C. Vosloo, Fuel Proc. Technol. 71 (2001) 149. [19] R. Zona, S. Solar, P. Gehringer, Water Res. 36 (2002) 1369. [20] M. Pera-Titus, V. Garcı´a-Molina, M.A. Ban˜os, J. Gime´nez, S. Esplugas, Appl. Catal. B: Environ. 47 (2004) 219. [21] J. Girardon, A.S. Lermontov, L. Gengembre, P. Chernavskii, A. GribovalConstant, A.Y. Khodakov, J. Catal. 230 (2005) 339. [22] M. Kraum, M. Baerns, Appl. Catal. A: Gen. 186 (1999) 189. [23] C. Nethravathi, S. Sen, N. Ravishankar, M. Rajamathi, C. Pietzonka, B. Harbrecht, J. Phys. Chem. B 109 (2005) 11468. [24] B. Wen, C. Liu, Y. Liu, J. Phys. Chem. B 109 (2005) 12372. [25] I. Park, S. Jang, J. Hong, R. Vittal, K. Kim, Chem. Mater. 15 (2003) 4633. [26] D. Chatterjee, S. Dasgupta, J. Photochem. Photobiol. C 6 (2005) 186. [27] J. Vakros, C. Kordulis, A. Lycourghiotis, Langmuir 18 (2002) 417. [28] D. Barreca, C. Massignan, S. Daolio, M. Fabrizio, C. Piccirillo, L. Armelao, E. Tondello, Chem. Mater. 13 (2001) 588. [29] M. Iwasaki, M. Hara, H. Kawada, H. Tada, S. Ito, J. Colloid Interf. Sci. 224 (2000) 202. [30] M. Long, W. Cai, J. Cai, B. Zhou, X. Chai, Y. Wu, J. Phys. Chem. B 110 (2006) 20211. [31] R. Xu, H. Zeng, Langmuir 20 (2004) 9780. [32] T. He, D. Chen, X. Jiao, Y. Wang, Y. Duan, Chem. Mater. 17 (2005) 4023. [33] J. Vakros, K. Bourikas, S. Perlepes, C. Kordulis, A. Lycourghiotis, Langmuir 20 (2004) 10542. [34] L.G.A. Van De Water, G.L. Bezemer, J.A. Bergwerff, M. Versluijs-Helder, B.M. Weckhuysen, K.P. Jong, J. Catal. 242 (2006) 287. [35] B. Xin, L. Jing, Z. Ren, B. Wang, H. Fu, J. Phys. Chem. B 109 (2005) 2805. [36] L. Jing, H. Fu, B. Wang, D. Wang, B. Xin, S. Li, J. Sun, Appl. Catal. B: Environ. 62 (2006) 282. [37] T. Choudhury, S. Saied, J.L. Sullivan, A.M. Abbot, J. Phys. D: Appl. Phys. 22 (1989) 1185. [38] A. Gulino, P. Dapporto, P. Rossi, Chem. Mater. 15 (2003) 3748. [39] J. Shang, Y.F. Zhu, Y.G. Du, Z.L. Xu, J. Solid State Chem. 166 (2002) 395. [40] M.A. Henderson, Surf. Sci. Rep. 46 (2002) 1. [41] M.A. Henderson, Langmuir 12 (1996) 5093. [42] U. Diebold, Surf. Sci. Rep. 48 (2003) 53. [43] P.K. Dutta, A.K. Ray, V.K. Sharma, F.J. Millero, J. Colloid Interf. Sci. 278 (2004) 270. [44] S. Lin, M. Gurol, Environ. Sci. Technol. 32 (1998) 1417. [45] C. Pirovano, S. Trasatti, J. Electroanal. Chem. 180 (1984) 171.