Visible-Light-Active Iron Oxide-Modified Anatase Titanium(IV) Dioxide

Mar 16, 2011 - Hiroaki Tada , Qiliang Jin , Anna Iwaszuk , and Michael Nolan ... Sho Kitano , Naoya Murakami , Teruhisa Ohno , Yasufumi Mitani , Yoshi...
8 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/JPCC

Visible-Light-Active Iron Oxide-Modified Anatase Titanium(IV) Dioxide Qiliang Jin, Musashi Fujishima, and Hiroaki Tada* Department of Applied Chemistry, School of Science and Engineering, Kinki University, 3-4-1 Kowakae, Higashi-Osaka, Osaka 577-8502, Japan ABSTRACT: Fe(acac)3 is chemisorbed on the surfaces of anatase TiO2 via partial ligand exchange between the acetylacetonate and surface TiOH groups [Fe(acac)2/TiO2]. The postheating at 773 K in air forms iron oxide species on the TiO2 surface in a highly dispersed state at a molecular level ((FeOx)m/TiO2). As a result of the iron oxide surface modification, the band gap of TiO2 decreases, while the absorption due to the dd transition clearly observed for the usual impregnation samples is very weak. (FeOx)m/TiO2 gives rise to a noticeable visible light activity concomitantly with a significant increase in UV light activity, whereas Fe(acac)2/TiO2 hardly responds to visible light. Valence-band X-ray photoelectron spectra of (FeOx)m/TiO2 showed that the band gap narrowing results from the rise in the valence band top with surface modification. Also, photoluminescence spectroscopy indicated that the surface iron oxide species rapidly capture the excited electrons in the conduction band of TiO2 to suppress recombination via surface oxygen vacancy levels. Furthermore, the surface iron oxide species act as excellent mediators for electron transfer from TiO2 to O2.

I. INTRODUCTION Continuous vital research on TiO2 photocatalysts has realized their practical uses as ecocatalysts for environmental purification and self-cleaning materials.1,2 However, TiO2, with a band gap larger than 3 eV, responds only to UV light, occupying 34% of the incident sunlight. TiO2 has three polymorphic forms—rutile, brookite, and anatase—among which anatase has usually the highest UV light activity. The development of a general method for endowing commercial anatase TiO2 with visible light activity and increased UV light activity promises an enormous expansion of uses. To this end, doping of various transition metals and anions has been studied extensively and reviewed in several papers.39 Particularly, iron that is harmless and abundant in nature is an ideal candidate; however, iron doping usually does not induce the visible light response but also reduces UV light activity.1012 This is mainly because the doping generates impurity and/or vacancy levels in the bulk that act as recombination centers. Recently, the research groups of Ohno13 and Hashimoto14 have reported that surface modification of rutile TiO2 with Fe3þ by the impregnation method leads to high visible light activities for the decomposition of model organic pollutants. This approach is attractive in that visible light response can be caused by a simple procedure without introduction of impurity/ vacancy levels. The effect is noticeable for rutile TiO2; however, it is insufficient for anatase TiO2. More recently, we have shown that the surface modification of anatase TiO2 particles (ST-01, Ishihara Sangyo) having the highest level of photocatalytic activity among the commercial ones with molecular iron oxides by the chemisorptioncalcination cycle (CCC) technique15 [(FeOx)m/TiO2, Scheme 1) gives rise to a high level of visible light activity and a concomitant great increase in UV light activity.16 While its essential r 2011 American Chemical Society

action mechanism has been proposed, the origin of the surface modification effect is not fully understood. Here we report the spectroscopic and electrochemical properties of (FeOx)m/TiO2 to obtain detailed mechanistic information.

II. EXPERIMENTAL SECTION A. Sample Preparation. Anatase TiO2 particles (ST-01, specific surface area = 309 m2 3 g1, Ishihara Sangyo) were used as a standard photocatalyst. Also, TiO2 particles with a mean size of 20 nm (PST-18NR, Nikki Syokubai Kasei) were coated on fluorine tin oxide (FTO) film-coated glass substrates (12 Ω/0) by a squeegee method, and the sample was heated in air at 773 K to form mp-TiO2 films. By the CCC technique utilizing this reaction, (FeOx)m/TiO2 was prepared as follows: After TiO2 particles (1 g) or mesoporous TiO2 nanocrystalline film-coated SnO2 substrates (mp-TiO2/FTO, 25 mm  50 mm) had been added to 100 mL of a Fe(acac)3 solution (solvent, ethanol/nhexane = 3:17 v/v), they were allowed to stand for 24 h at 298 K. Unless otherwise noted, the Fe(acac)3 concentration was maintained at 6.5  104 M. The resulting samples were washed repeatedly with solvent for the physisorbed complexes to be removed and dried, followed by heating in air at 773 K for 1 h. The adsorption isotherm of Fe(acac)3 on TiO2 was analyzed by use of the Langmuir model with the following equation: Γ1 = (KΓ¥Ceq)1 þ Γ¥1, where Γ and Γ¥ are the adsorption amount per unit TiO2 surface area at equilibrium Fe(acac)3 Received: February 3, 2011 Revised: February 27, 2011 Published: March 16, 2011 6478

dx.doi.org/10.1021/jp201131t | J. Phys. Chem. C 2011, 115, 6478–6483

The Journal of Physical Chemistry C

ARTICLE

Scheme 1. Schematic Representation of the CCC Technique for Preparing (FeOx)m/TiO2

Figure 1. (A) UVvis spectral change of a 6.5  104 M Fe(acac)3 solution with addition of ST-01 (0.1 g). (B) Difference DRIFT spectra: (a) complex-adsorbed ST-01  ST-01; (b) complex-adsorbed ST-01 after heating at 773 K  ST-01.

concentration (Ceq) and the saturated adsorption amount, respectively, and K is the adsorption equilibrium constant. B. Spectroscopic Characterization. UVvis diffuse reflectance spectra of FeOx/TiO2 and FeOx/mp-TiO2 were recorded on a Hitachi U-4000 spectrophotometer. The spectra were converted to the absorption spectra by using the KubelkaMunk function. X-ray photoelectron spectroscopic (XPS) measurements were performed using a Kratos Axis Nova X-ray photoelectron spectrometer with a monochromated Al KR X-ray source (hν = 1486.6 eV) operated at 15 kV and 10 mA. The takeoff angle was 90, and multiplex spectra were obtained for Fe 2p, O 1s, and Ti 2p photopeaks. All the binding energies (EB) were referenced with respect to C 1s at 284.6 eV. The photoluminescence (PL) spectra were measured with an excitation wavelength of 320 nm at 77 K on a Jasco FP-6000 spectrofluorometer.

C. Photoelectrochemical Measurements. Currentpotential curves of the (FeOx)m/mp-TiO2/FTO electrodes were measured in a 0.1 M Na2ClO4 electrolyte solution in a regular threeelectrode electrochemical cell on a galvanostat/potentiostat (HZ-5000, Hokuto Denko). Glassy carbon and an Ag/AgCl electrode (TOA-DKK) were used as counterelectrode and reference electrode, respectively. D. Photocatalytic Activity Evaluation. In the decompositions of 2-naphthol (2-NAP), the reaction cells were irradiated with a Xe lamp (Wacom XRD-501SW) through a band-pass filter (D33S, AGC Techno Glass) superposed on a piece of FTOcoated glass transmitting only the 330400 nm range for the UV light photocatalytic activity evaluation and a high pass filter (L-42, Toshiba) to cut off UV light for the visible-light-induced activity test. ST-01 or (FeOx)m/ST-01 particles (0.1 g) was 6479

dx.doi.org/10.1021/jp201131t |J. Phys. Chem. C 2011, 115, 6478–6483

The Journal of Physical Chemistry C

ARTICLE

Figure 2. (A) Concentrations of (a) Fe(acac)3 and (b) AcacH in solution and (c) acac/Fe number ratio for the adsorbed species as a function of adsorption time. (B) Adsorption isotherm of Fe(acac)3 on ST-01 at 298 K. (Inset) Langmuir plot.

Figure 3. UVvis absorption spectra of (A) Fe(acac)2/ST-01 and (B) (FeOx)m/ST-01 with varying Γ.

placed in 50 mL of 1.0  105 M solution of 2-NAP (solvent, acetonitrile:water = 1:9999 v/v) in a borosilicate glass container was irradiated. Two milliliters of the solution was sampled every 15 min and the electronic absorption spectra of the reaction solutions were measured on a spectrometer (Shimadzu, UV-1800) to determine 2-NAP concentration from the absorption peak at 224 nm.

III. RESULTS AND DISCUSSION In the UVvis spectrum of Fe(acac)3, there is a weak and broad absorption centered at 430 nm (B1) and a strong absorption at 272 nm (B2). Acetylacetone (AcacH) has also a strong absorption at 272 nm, and the absorption coefficient at 272 nm of the complex (2.81  104 M1 3 cm1) was in good agreement with three times that of AcacH (9.44  103 M1 3 cm1). The B1 and B2 bands for the complex are assignable to the dd transition and the ππ* transition in the acetylacetonate ligand, respectively. Figure 1A shows UVvis spectral change of a 6.5  104 M Fe(acac)3 solution with adsorption on ST-01 (1 g). The absorption intensities of both B1 and B2 significantly weaken as the adsorption the progresses. To study the adsorption state of Fe(acac)3 on ST-01, diffuse reflectance Fourier transform infrared (DRIFT) spectra were measured. Figure 1B shows the difference DRIFT spectra for (a) complex-adsorbed ST-01  ST-01 and (b) complex-adsorbed ST-01 after heating at 773 K  ST-01. In

spectrum a, three absorption peaks are present at 1593, 1527, and 1374 cm1, which can be assigned to the combination of ν(CC) þ ν(CO), the combination of ν(CO) þ ν(CC), and δs(CH3), respectively.17 Also, a negative signal due to the surface OH groups of ST-01 appears at 3668 cm1.18 These results strongly suggest that the adsorption of Fe(acac)3 on ST-01 proceeds via ligand exchange between the acetylacetonate ligand and the surface TiOH groups in a similar manner as Mg(acac)2.19 The concentration of Fe(acac)3 and the sum of the concentrations of Fe(acac)3 and the free AcacH generated with the adsorption were determined from the absorbance of B1 and B2, respectively. Figure 2A shows the concentrations of (a) Fe(acac)3 and (b) AcacH in the solution and (c) the number ratio of acac/Fe for the adsorbed species as a function of adsorption time. The acac/Fe ratio is ca. 2 regardless of the adsorption time. Figure 2B shows the adsorption isotherm of Fe(acac)3 on ST-01 at 298 K. The adsorption amount steeply increases with increasing equilibrium concentration. The Langmuir plot exhibits good linearity (inset, Figure 2B). The saturated adsorption amount and the equilibrium constant were calculated to be 0.35 ion 3 nm2 and 1.5  103 M1, respectively. These results also support the presumption that Fe(acac)3 is chemisorbed on the ST-01 surface with liberation of the partial acetylacetonate ligands as AcacH. As shown in Figure 1B(b), the residual acetylacetonate ligands are completely oxidized to yield iron oxide-modified ST-01 6480

dx.doi.org/10.1021/jp201131t |J. Phys. Chem. C 2011, 115, 6478–6483

The Journal of Physical Chemistry C

ARTICLE

Figure 4. (A) Time courses for Fe(acac)2/ST-01-photocatalyzed decomposition of 2-NAP under visible-light illumination. (B) Time courses for (FeOx)m/ST-01-photocatalyzed decomposition of 2-NAP under visible-light illumination. (C) Time courses for (FeOx)m/ST-01-photocatalyzed decomposition of 2-NAP under UV-light illumination. (D) Plots of pseudo-first-order rate constants for (FeOx)m/ST-01-photocatalyzed decomposition of 2-NAP under visible-light (kvis, b) and UV-light (kUV, 2) illumination as a function of Γ.

[(FeOx)m/ST-01] by postheating at 773 K in air. Fe on the TiO2 surface was dissolved by treatment with 35% HCl, and (FeOx)m/ ST-01 solid was completely dissolved into 96% H2SO4 at 353 K. Fe amounts in both solutions were almost equal, although ST-01 manufactured from illmenite by the sulfuric acid method contains a slight amount of Fe in the bulk as an impurity. This fact indicates that most Fe exists on the surface of ST-01. The Fe loading amount is expressed by the number of Fe3þ ions per unit TiO2 surface area (Γ, ions 3 nm2). The chemisorptioncalcination cycle was repeated to control Γ. Consequently, it can be concluded that Fe(acac)3 is chemisorbed on the ST-01 surface via ligand exchange (eq 1): FeðacacÞ3 þ lðTis  OHÞ f FeðacacÞ3  l ðOTis Þl þ lAcacH ðl  1Þ

ð1Þ

where the subscript s denotes the surface atom, and iron oxide species are formed by postheating (eq 2): Δ

FeðacacÞ2 ðOTis Þ þ ðx þ 23Þ=2O2 f FeOx =ðOTis Þl þ 7H2 O þ 10CO2

ð2Þ

Optical properties for Fe(acac)2/ST-01 and (FeOx)m/ST-01 are of primary importance in connection with the photocatalytic

activity. Figure 3A shows UVvis absorption spectra of Fe(acac)2/ST-01 with varying Γ. The chemisorption of Fe(acac)3 causes a shoulder in the 400500 nm range due to the dd transition with a small red shift of the absorption edge. Iron oxide surface-modified TiO2 particles prepared by the conventional impregnation method from a precursor such as FeCl3 [(FeOx)n/ TiO2] have commonly a weak electronic absorption around 470 nm in addition to the absorption at 410 nm.13,14,20,21 The former and latter absorption bands were attributed to the dd transition and to the electronic transition from Fe3þ levels to the conduction band (CB) of TiO2, respectively.22 Upon chemical doping of Cr and N ions into TiO2, similar weak shoulders appear in the visible region due to formation of localized impurity levels within the band gap.6 Figure 3B shows UVvis absorption spectra of (FeOx)m/ST-01 with varying Γ. Noticeably, the absorption intensity of the dd transition band is very weak at Γ < 0.36, and the band gap narrowing is further promoted as compared to that in the Fe(acac)2/ST-01 system. A similar spectral feature was observed also for TiO2 doped with Cr and N prepared by the physical methods of ion implantation and magnetron sputtering, which exhibits high visible light activity.3,23 High-resolution transmission electron microscopic observation of (FeOx)m/ST-01 confirmed no particles on the TiO2 surface at Γ < 1 ion 3 nm2. Thus, iron oxide species can be formed on the ST-01 surface in a highly dispersed state at a 6481

dx.doi.org/10.1021/jp201131t |J. Phys. Chem. C 2011, 115, 6478–6483

The Journal of Physical Chemistry C molecular level by the CCC technique. The use of hydrolysisresistant Fe(acac)3 as a precursor and the postheating after its strong chemisorption on the TiO2 surface are considered to contribute to the formation of unique surface iron oxide species. To evaluate the relative photocatalytic activities of Fe(acac)2/ ST-01 and (FeOx)m/TiO2 with respect to that of pristine ST-01, widely used as the high standard photocatalyst, the rate constants were determined under the same irradiation conditions with the same amount of photocatalysts. As a test reaction, the photocatalytic degradation of 2-naphthol (2-NAP) was carried out under illumination of visible light (λ > 400 nm, I420485nm = 1.0 mW 3 cm2) and UV light (330 < λ < 400 nm, I320400nm = 0.5 mW 3 cm2).24 2-NAP, which is the starting material of azo dyes, was used as a model water pollutant. 2-NAP has an absorption band centered at 224 nm due to the nπ* transition. Figure 4A shows the change in 2-NAP concentration under visible-light illumination in the presence of Fe(acac)2/ST-01. Surface modification with the complex is entirely ineffective in 2-NAP decomposition. In contrast, upon irradiation with visible light in the presence of (FeOx)m/ST-01, the decomposition of 2-NAP is drastically enhanced (Figure 4B), and concomitantly the UV light activity significantly increases (Figure 4C). In this manner, oxidation of the adsorbed iron complex is crucial for

Figure 5. PL spectra of (FeOx)m/ST-01 with varying Γ at 77 K; excitation wavelength = 320 nm.

ARTICLE

improvement of photocatalytic activities. Apparently, the 2-NAP decomposition follows first-order kinetics. Figure 4D shows the pseudo-first-order rate constants for illumination by visible light (kvis) and UV light (kUV) as a function of Γ. The plots exhibits volcano-type curves with maxima at Γ ≈ 0.1 ion 3 nm2. Under optimum conditions, kvis and kUV values increase compared to those for ST-01 by factors up to 8.1 and 2.5, respectively. Evidently, surface modification by iron oxide species causes noticeable visible light activity and a concomitant significant increase in UV light activity. By use of valence-band (VB) X-ray photoelectron spectroscopy (XPS), we have recently shown that the band gap narrowing of (FeOx)m/TiO2 results from the rise in the VB top with the iron oxide surface modification.16 On the other hand, direct information about empty levels can be obtained by PL spectroscopy. Figure 5 shows PL spectra of (FeOx)m/ST-01 with varying Γ at 77 K; excitation wavelength = 320 nm. ST-01 has a broad emission band centered at 538 nm (E1). The E1 signal intensity remarkably weakens when ST-01 is heated at 773 K for 1 h in air. This PL band is assignable to emission from the surface oxygen vacancy levels of anatase TiO2.25 Upon modification of ST-01 with the iron oxide species, the intensity further decreases to disappear at Γ > 0.044 ion 3 nm2, while two new emissions appear at 423 (E2) and 468 nm (E3). The E2 and E3 signals can be attributed to emissions from extrinsic levels due to surface iron oxide species. These findings suggest that excited electrons in the conduction band (CB) of TiO2 are transferred to the empty surface iron oxide levels in preference to the surface oxygen vacancy levels. Additional surface oxygen vacancy is not induced by the surface modification, in contrast to the iron doping, which is probably because the local charge balance can be conserved on the surface. Clearly, the filled and empty d levels exist within the intragap of TiO2, which is also consistent with the previous Fe 2p XPS result that the surface iron species has a mixed valence state of 2þ and 3þ. In the oxidative decomposition of organic pollutants, the key to increasing photocatalytic activity of TiO2 is to enhance the transfer of the excited electrons to O2.26,27 Figure 6 shows the current (I, milliamps)potential (E, volts vs Ag/AgCl) curves of the mp-TiO2/FTO electrodes (A) without and (B) with the iron oxide surface modification in a 0.1 M NaClO4 aqueous solution under irradiation at λ > 300 nm (I320400nm = 1.9 mW 3 cm2). For the mp-TiO2 electrode, the current hardly flows at the

Figure 6. (A) Currentpotential curves of mp-TiO2/FTO electrodes in 0.1 M NaClO4 aqueous solution, (a) without and (b) with O2, under irradiation at λ > 300 nm (I320400nm = 1.9 mW 3 cm2). (B) Currentpotential curves of (FeOx)m/mp-TiO2/FTO electrodes in 0.1 M NaClO4 aqueous solution, (a) without and (b) with O2, under UV-light irradiation. 6482

dx.doi.org/10.1021/jp201131t |J. Phys. Chem. C 2011, 115, 6478–6483

The Journal of Physical Chemistry C Scheme 2. Proposed Energy Band Diagram Scheme of (FeOx)m/TiO2

ARTICLE

promote electron transfer to O2. This simple and inexpensive technique can easily be applied to highly active TiO2 particles and films hitherto developed to expand their applications.

’ AUTHOR INFORMATION Corresponding Author

*Tel þ81-6-6721-2332; fax þ81-6-6727-2024; e-mail [email protected].

’ REFERENCES

potential range between 0 and 2 V in the absence of O2 (curve a), while the current due to the O2 reduction is observed at E ≈ 1 V in the presence of O2 (curve b). On the other hand, for the (FeOx)m(Γ = 0.58)/mp-TiO2/FTO electrodes, no current is observed in the potential range between 0 and 1.5 V without O2 (curve a). It is worth noting that the O2 reduction potential shifts toward positive direction ca. 0.5 V with O2 (curve b). Importantly, the surface iron oxide species drastically promotes electron transfer from TiO2 to O2. An energy band diagram scheme of (FeOx)m/TiO2, presumed on the basis of the results above, is shown in Scheme 2. Visiblelight absorption leads to electronic excitation from the surface d sub band to the CB(TiO2) (p1). Upon UV-light illumination (p2), the electrons in the VB(TiO2) are excited to the CB(TiO2). In both cases, without the iron oxide surface modification, the excited electrons in CB(TiO2) are rapidly trapped at the surface oxygen vacancy levels to undergo recombination with the holes (p3). On the other hand, the iron oxide surface modification permits preferential electron transfer from CB(TiO2) to the surface iron oxide levels (p4). The electrons effectively reduce adsorbed O2 by the action of the surface iron oxide species as an excellent mediator (p5). As a result of the effective suppression of recombination via surface oxygen vacancy levels, the photocatalytic activities drastically increase under visible- and UV-light illumination. A feature of the surface modification in the anodic process should also be stressed: the holes generated in the surface d sub-band take part in the oxidation process without diffusion (p6).28 Meanwhile, excess loading of surface iron oxide species would cause a drop in Fermi energy to lower the reducing power of the excited electrons or a rise in the top of the surface d subband to reduce the hole oxidation power,16 and thus the photocatalytic activity is lowered. Consequently, an optimum loading amount is present.

IV. CONCLUSIONS The CCC technique enables formation of extremely highly dispersed surface iron oxide species on the TiO2 surface. This surface modification has given rise to significant visible light activity and a concomitant increase in UV light activity. These remarkable effects have been shown to originate from the actions of the surface iron oxide species to suppress recombination and

(1) Fujishima, A.; Zhang, X.; Tryk, D. A. Surf. Sci. Rep. 2008, 63, 515. (2) Hashimoto, K.; Irie, H.; Fujishima, A. Jpn. J. Appl. Phys. 2005, 44, 8269. (3) Anpo, M.; Takeuchi, M. J. Catal. 2003, 216, 505. (4) Serpone, N. J. Phys. Chem. B 2006, 110, 24287. (5) Zhang, H.; Chen, G.; Bahnemann, D. W. J. Mater. Chem. 2009, 19, 5089. (6) Liu, G.; Wang, L.; Yang, H. G.; Cheng, H.-M.; Lu, G. Q. J. Mater. Chem. 2010, 20, 831. (7) Rajeshwar, K.; de Tacconi, N. R. Chem. Soc. Rev. 2009, 38, 1984. (8) Kudo, A.; Miseki, Y. Chem. Soc. Rev. 2009, 38, 253. (9) Chen, X.; Shen, S.; Guo, L.; Mao, S. S. Chem. Rev. 2010, 110, 6503. (10) Litter, M. I.; Navio, J. A. J. Photochem. Photobiol. A: Chem. 1996, 98, 171. (11) Nagaveni, K.; Hedge, M. S.; Madras, G. J. Phys. Chem. B 2004, 108, 20204. (12) Serpone, N.; Lawless, D.; Disdier, J.; Hermann, J.-M. Langmuir 1994, 10, 643. (13) Murakami, N.; Chiyoya, T.; Tsubota, T.; Ohno, T. Appl. Catal., A 2008, 348, 148. (14) Yu, H.; Irie, H.; Shimodaira, Y.; Hosogi, Y.; Kuroda, Y.; Miyauchi, M.; Hashimoto, K. J. Phys. Chem. C 2010, 114, 16481. (15) Tada, H. Encyclopedia of Surface and Colloid Science; Hubbard, A. T., Ed.; Marcel Dekker: New York, 2002. (16) Tada, H.; Jin, Q.; Nishijima, H.; Yamamoto, H.; Fujishima, M.; Okuoka, S.-i.; Hattori, T.; Sumida, Y.; Kobayashi, H. Angew. Chem., Int. Ed. 2011in press. (17) Nakamoto, K. Infrared and Raman Spectra of Inorganic and Coordination Compounds, 4th ed.; Wiley-Interscience: New York, 1986. (18) Tada, H. Langmuir 1996, 12, 966. (19) Tada, H.; Yamamoto, M.; Ito, S. Langmuir 1999, 15, 3699. (20) Navio, J. A.; Colon, G.; Litter, M. I.; Bianco, G. N. J. Mol. Catal. A: Chem. 1996, 106, 267. (21) Wang, C.-Y.; Bahnemann, D. W.; Dohrmann, J. K. Chem. Commun. 2000, 1539. (22) Serpone, N.; Lawless, D.; Disdier, J.; Hermann, J.-M. Langmuir 1994, 10, 643. (23) Kitano, M.; Funatsu, K.; Matsuoka, M.; Ueshima, M.; Anpo, M. J. Phys. Chem. B 2006, 110, 25266. (24) Yamada, K.-i.; Mukaihata, N.; Kawahara, T.; Tada, H. Langmuir 2007, 23, 8593. (25) Shi, J.; Chen, J.; Feng, Z.; Chen, T.; Lian, Y.; Wang, X.; Li, C. J. Phys. Chem. 2007, 111, 693. (26) Wang, C. M.; Heller, A.; Gerischer, H. J. Am. Chem. Soc. 1992, 114, 5230. (27) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Chem. Rev. 1995, 95, 69. (28) Irie, H.; Shibanuma, T.; Kamiya, K.; Miura, S.; Yokoyama, T.; Hashimoto, K. Appl. Catal., B 2010, 96, 142.

6483

dx.doi.org/10.1021/jp201131t |J. Phys. Chem. C 2011, 115, 6478–6483