Photophysical and Photocatalytic Properties of SrTiO3 Doped with Cr

Defa Wang,†,‡ Jinhua Ye,*,†,‡ Tetsuya Kako,† and Takashi Kimura§. Photocatalytic Materials Center and Analysis Station, National Institute ...
0 downloads 0 Views 402KB Size
Download PDF
15824

J. Phys. Chem. B 2006, 110, 15824-15830

Photophysical and Photocatalytic Properties of SrTiO3 Doped with Cr Cations on Different Sites Defa Wang,†,‡ Jinhua Ye,*,†,‡ Tetsuya Kako,† and Takashi Kimura§ Photocatalytic Materials Center and Analysis Station, National Institute for Materials Science (NIMS), 1-2-1 Sengen, Tsukuba, Ibaraki 305-0047, Japan, and Precursory Research for Embryonic Science and Technology (PREST), Japan Science and Technology Agency (JST), 4-1-8 Hon-cho, Kawaguchi, Saitama 332-0012, Japan ReceiVed: April 23, 2006; In Final Form: June 20, 2006

Usually, SrTiO3 monodoped with Cr cations at the Ti4+ site hardly shows visible light photocatalytic activity. Revealing the origin of this issue is important for us to find an alternative approach to make SrTiO3 active under visible light irradiation. In this paper, two Cr-doped SrTiO3s(Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3s were synthesized by a conventional solid-state reaction method, and their photophysical and photocatalytic properties were studied comparatively. It was found that both (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3 showed considerable absorption to visible light. However, their photocatalytic activities for H2 evolution from aqueous methanol solution under visible light irradiation were significantly different: the H2 evolution rate over (Sr0.95Cr0.05)TiO3 (∼21 µmol/h) was more than 100 times that over Sr(Ti0.95Cr0.05)O3 (∼0.2 µmol/h). X-ray photoelectron spectroscopy analysis results revealed that the Cr cations doped at the Sr2+ site were all trivalent state (Cr3+), while those doped at the Ti4+ site were mixed valent states (Cr3+ and Cr6+). The different photocatalytic activities of H2 evolution are supposed to closely relate to the different valent states of Cr doped at different sites (Sr2+ or Ti4+) in SrTiO3. Possible electronic structures of (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3 were proposed in relation to their photophysical and photocatalytic properties.

1. Introduction Generally, the typical photocatalysis process over a semiconductor involves the absorption of photons, band gap excitation, separation of the photoexcited electron/hole pairs, and redox reactions on the semiconductor surface. In view of efficient solar energy conversion, visible-light-responsive photocatalysts are important because ∼43% of the whole solar energy is visible light while only ∼4% is UV light.1-3 An ideal visible-light-responsive photocatalyst for water splitting is required to have not only a small band gap energy but also proper band edges suitable for the redox reactions of H2/O2 evolution. Another principal challenge is how to suppress the more facile energy-wasteful recombination of photoinduced electron/hole pairs, so that the subsequent steps needed for water splitting are able to take place. Among the vast majority of metal oxide photocatalysts, perovskite-type oxides are prominent for their broad diversity of properties. In an ideal perovskite-type cubic structure of ABO3 with the space group Pm3m (Figure 1), the A cation is 12-fold coordinated and the B cation is 6-fold coordinated with the oxygen anions. The corner-sharing BO6 octahedra form the skeleton of the structure, in which the center position is occupied by the A cation. Because of the susceptibility of partial substitution at both A and B sites, a wide range of cations and valences can be accommodated in this simple crystal structure, rendering perovskites to diverse and flexible chemical tailoring.4 As a typical A2+B4+O3-type perovskite, SrTiO3 has been known as a photocatalyst capable of decomposing H2O into H2 and O2 * Corresponding author. E-mail: [email protected]. † Photocatalytic Materials Center, NIMS. ‡ JST. § Analysis Station, NIMS.

Figure 1. Ideal ABO3 perovskite structure. The BO6 octahedron links through corners to form a three-dimensional cubic lattice, and the A cation is in the center of the cube.

without applying an external bias potential.5-10 Surface modifications of SrTiO3 with noble metals (e.g., Pt)9 or oxides (e.g., NiOx)10 have been attempted to decrease the overvoltage of H2 evolution and to suppress the recombination of photoinduced electron/hole pairs. However, SrTiO3 only responds to UV light due to its relatively large band gap (3.2 eV). On the other hand, doping foreign elements into a semiconductor with wide band gap to create a new donor or acceptor level in the forbidden band is known to be one of the methods for developing visible-lightdriven photocatalysts. There are many reports on doping of foreign elements, mostly transition metals, into SrTiO3 to enhance its absorption to visible light. In particular, chromium has been paid much attention because the occupied Cr3+ level is usually ∼2.2 eV lower than the conduction band bottom formed by Ti 3d (e.g.) or ∼1.0 eV higher than the valence band top formed by O 2p.11 The effects of Cr doping on the photocurrent of semiconductor electrodes and the photocatalytic decomposition of organic compounds have been widely

10.1021/jp062487p CCC: $33.50 © 2006 American Chemical Society Published on Web 07/26/2006

SrTiO3 Doped with Cr Cations on Different Sites studied.12-17 Unfortunately, there have been no reports so far on efficient photocatalytic H2/O2 evolution from water over Crdoped SrTiO3. Recently, co-doping of Cr with other pentavalent cations (e.g., Sb, Ta) into SrTiO3 have been reported for H2 production from aqueous methanol solution.18,19 The previous studies, in which the Cr cations were doped or co-doped with other cations unexceptionally at the Ti4+ site in SrTiO3, were essentially based on the fact that the main electronic properties of SrTiO3 are dominated by the network of corner-shared TiO6 octahedra: the valence band top is made up predominately of the O 2p states, and the conduction band bottom is determined by the Ti 3d states. On the other hand, the important characteristics of ABO3 perovskites are their susceptibility of partial substitution of cations at both A and B sites and the stability of mixed oxidation states or unusual oxidation states in the crystal structure. We consider that if the Cr cations are alternatively doped at the Sr2+ site rather than the Ti4+ site, different photophysical and photocatalytic properties might be obtained. Indeed, we have found that, in contrast to the case of doping Cr cations at a Ti4+ site, doping Cr cations at a Sr2+ site significantly improved the photocatalytic activity for H2 evolution under visible light irradiation. Here, we report the effects of the substitution site of Cr cations on the crystal structures, UV-vis absorption spectra, and photocatalytic H2 evolution activities of two Cr-doped SrTiO3s(Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3. In particular, the valence state of Cr cations doped at the Sr2+ site in (Sr0.95Cr0.05)TiO3 or at the Ti4+ site Sr(Ti0.95Cr0.05)O3 were analyzed by X-ray photoelectron spectroscopy (XPS). Possible electronic structures are suggested in relation to their respective photophysical and photocatalytic properties. 2. Experimental Section 2.1. Photocatalyst Preparation. Polycrystalline powder samples of SrTiO3 and Cr-doped SrTiO3s(Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3swere synthesized by solid-state reactions. First, the precursors SrCO3, TiO2, and Cr2O3 (99.9%, Wako) in the corresponding ratios of SrTiO3, (Sr0.95Cr0.05)TiO3, and Sr(Ti0.95Cr0.05)O3 were mixed carefully with addition of ethanol. Then, the mixed powders were precalcined at 900 °C for 12 h and finally calcined at 1100 °C for 24 h in an alumina crucible in air, with intermediate milling. 2.2. Photocatalyst Characterization. The crystal structure of as-prepared materials was determined by an X-ray diffractometer (XRD, JEOL JDX-3500) operated at 20 kV and 10 mA using Cu KR radiation (λ ) 1.541 78 Å). The scanned range was 2θ ) 10-100°, with a step of 2θ ) 0.02° and 0.5 s/step. The valence state of Cr was analyzed on an X-ray photoelectron spectroscope (PHI Quantera SXM). The UV-vis diffuse reflectance spectrum was measured at room temperature with a UV-vis spectrometer (UV-2500, Shimadzu) and was converted to absorbance spectrum by the Kubelka-Munk method. A field emission scanning electron microscope (FE-SEM, JEOL-JSM 6500F) and a high-resolution transmission electron microscope (HRTEM, JEOL JEM-2010, operated at 200 kV), both of which were equipped with energy dispersive spectrometer (EDS), were employed for microstructure characterization and composition analysis. The induction coupled plasma-atomic emission spectrometry (ICP-AES) was also employed for chemical composition analysis of the Cr-doped SrTiO3 samples. Surface area was measured on a Gemini 2360 surface area analyzer (Micromeritics, SHIMADZU) by nitrogen absorption at 77 K using the Brunauer-Emmett-Teller (BET) method. 2.3. Photocatalytic Activity Evaluation. Photocatalytic H2 evolution was conducted over the Pt(0.4wt%)/photocatalyst

J. Phys. Chem. B, Vol. 110, No. 32, 2006 15825 using CH3OH as the sacrificial reagent. The cocatalyst Pt (0.4 wt %, relative to the amount of photocatalyst) was loaded on the catalyst surface in advance by an impregnation method: the powder samples were sunk into an aqueous solution of an equivalent molar amount of H2PtCl6 and then calcined at 500 °C for 2 h, forming the Pt(0.4wt%)/photocatalyst. For the reaction under visible light irradiation (λ g 420 nm), the Pt(0.4wt%)/photocatalyst powders (0.25 g) were dispersed with a magnetic stirrer in the aqueous CH3OH solution (50 mL of CH3OH + 220 mL of H2O) in an outer irradiation Pyrex glass cell. A 300 W xenon arc lamp was focused on the side window of the cell through a long-pass cutoff filter (λ g 420 nm, L42, HOYA). For the reaction under UV light irradiation, the Pt(0.4wt%)/photocatalyst powders (0.25 g) were dispersed with a magnetic stirrer in the aqueous CH3OH solution (50 mL of CH3OH + 320 mL of H2O) in an inner irradiation quartz cell. The light source was a 400 W high-pressure mercury lamp (400 HA, RIKO). The reaction cell was connected to a closed gas circulation system, and the gases evolved were in-situ analyzed with an on-line thermal conductivity detector (TCD) gas chromatograph (Shimadzu GC-8AIT, argon carrier). The quantum yields (QYs) of H2 evolution at various wavelengths were measured by using corresponding band-pass filters (MIF-W, KENKO). A spectroradiometer (USR-40, USHIO) was used for measuring the light intensity, from which the number of incident photons was calculated. The apparent quantum yield (%) was obtained by the following equation: QY (%) ) Ne/Np ) 2NH2/ Np, where Ne is the number of reacted electrons, Np is the number of incident photons, and NH2 is the number of evolved H2 molecules. 3. Results and Discussions 3.1. Materials Characterization. (A) Crystal Structure Analysis. The chemical compositions of the Cr-doped SrTiO3 samples analyzed by both ICP-AES and EDS were basically consistent with the nominal compositions. Hereafter, the Crdoped SrTiO3 samples are named by their nominal compositionss (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3. Figure 2a shows the X-ray diffraction patterns of SrTiO3, (Sr0.95Cr0.05)TiO3, and Sr(Ti0.95Cr0.05)O3 powder samples at room temperature. We can see that both (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3 have nearly the identical crystal structure as SrTiO3, indicating that doping up to 5 mol % of Cr cations at either Sr2+ or Ti4+ sites did not introduce any possible impurities. A careful comparison of the (110) diffraction peaks in the range of 2θ ) 32-33° (Figure 2b) showed that the peak position of Sr(Ti0.95Cr0.05)O3 was almost unchanged while the peak position of (Sr0.95Cr0.05)TiO3 shifted slightly toward a higher 2θ value. The reason is that the ionic radius of the possible Cr3+ (0.0615 nm) or Cr6+ (0.044 nm) is close to that of Ti4+ (0.0605 nm) but is much smaller than that of Sr2+ (0.118 nm).20 To some extent, these results could confirm the substitution of Cr cations for Sr2+ in (Sr0.95Cr0.05)TiO3 and for Ti4+ in Sr(Ti0.95Cr0.05)O3, respectively. According to the tolerance factor t ) (rA + rO)/(rB + rO)x2 (rA, rB, and rO are the empirical ionic radii at room temperature) as defined by Goldshmidt for keeping the structure of a perovskite ABO3,4 the t-values in SrTiO3, (Sr0.95Cr0.05)TiO3, and Sr(Ti0.95Cr0.05)O3 were calculated to be 0.903, 0.893, and 0.907, respectively, all of which are the normally allowed t-values (0.75 < t < 1.0). The difference of tolerance factors between SrTiO3 and (Sr0.95Cr0.05)TiO3 is apparently larger than that between SrTiO3 and Sr(Ti0.95Cr0.05)O3, being consistent with the XRD results. (B) XPS Analysis on the Valence States of Cr Cations in (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3. Besides the ionic radius,

15826 J. Phys. Chem. B, Vol. 110, No. 32, 2006

Wang et al.

Figure 4. UV-vis diffuse reflectance spectra of SrTiO3, (Sr0.95Cr0.05)TiO3, Sr(Ti0.95Cr0.05)O3, and mixtures of (95 mol % SrTiO3 + 5 mol % TiO2 + 2.5 mol % Cr2O3) and (95 mol % SrTiO3 + 5 mol % SrO + 2.5 mol % Cr2O3) at room temperature.

Figure 2. (a) Powder X-ray diffraction patterns of SrTiO3, (Sr0.95Cr0.05)TiO3, and Sr(Ti0.95Cr0.05)O3 measured at room temperature. (b) Diffraction peak positions of the (110) plane in the range of 2θ ) 3233°.

Figure 3. XPS spectra of Cr 2p in (Sr0.95Cr0.05)TiO3 ((a and a′) and Sr(Ti0.95Cr0.05)O3 (b and b′) before and after photoreaction by UV light irradiation. The Cr 2p spectrum (c) of standard Cr2O3 is measured as reference.

the electroneutrality is also required for an ABO3 perovskite. In (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3, the Cr cations doped at different sites (Sr2+ and Ti4+) might have different valence states, and the Sr and Ti cations might also change their valences correspondingly. The XPS results indicated that while the valence states of Sr2+ and Ti4+ were kept in both (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3, the valence states of Cr cations varied with the substitution sites (Sr2+ and Ti4+). From the Cr 2p spectrum in (Sr0.95Cr0.05)TiO3 (Figure 3a), only a sharp peak was observed at 577.6 eV, which could be assigned to the trivalent chromium (Cr3+) as that in the standard Cr2O3. In contrast, the Cr 2p spectrum in Sr(Ti0.95Cr0.05)O3 (Figure 3b) showed not only a shoulder peak assigned to trivalent chromium (Cr3+) at 577.6 eV but also a sharp peak assigned to hexavalent chromium (Cr6+) at 580.2 eV. In some other research, the Cr cations in Cr-doped SrTiO3 have also been found to be mixed valences.18,19 Recently, Meijer et al. studied the valence state of Cr and the insulator-to-metal transition in single crystals of

Cr-doped SrTiO3.21 In their study, the valence of Cr was determined to be approximately Cr4+ by X-ray absorption nearedge structure spectroscopy (XANES). Due to the interference of powder sample surfaces with the incident X-rays, only a qualitative but not a precisely quantitative result could be attained by the present XPS measurement. It may be argued that, in the nominal (Sr0.95Cr0.05)TiO3 sample, the possibility of partial occupation of the Cr cations at the Ti4+ sites could not be excluded definitely. Nevertheless, the fact is that the Cr cations in Sr(Ti0.95Cr0.05)O3 have higher valences than those in (Sr0.95Cr0.05)TiO3 because the electroneutrality should be satisfied. Undoubtedly, the different valences of Cr cations in (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3 could be additional effective evidence to prove the substitution of Cr for Sr2+ and for Ti4+ in the two Cr-doped SrTiO3, respectively. The Cr cations with different valence states in Sr(Ti0.95Cr0.05)O3 and (Sr0.95Cr0.05)TiO3 are supposed to play an important role in their photocatalytic properties, as will be described below in the section 3.2. (C) UV-Vis Diffuse Reflectance Spectra. Figure 4 shows the UV-vis diffuse reflectance spectra of SrTiO3, (Sr0.95Cr0.05)TiO3, and Sr(Ti0.95Cr0.05)O3 powder samples at room temperature. As a comparison, the UV-vis diffuse reflectance spectra of simply mixed (95 mol % SrTiO3 + 5 mol % TiO2 + 2.5 mol % Cr2O3) and (95 mol % SrTiO3 + 5 mol % SrO + 2.5 mol % Cr2O3) were also measured and depicted in Figure 4. We can see that the two simple mixtures have the same absorption edge at ∼385 nm (eg: ∼3.2 eV), which is almost identical to that of pure SrTiO3. Due to the addition of 2.5 mol % Cr2O3, two very small additional humps can be observed around ∼470 and ∼620 nm (indicated by arrows), respectively. In contrast to the simply mixed samples, the Cr-doped SrTiO3, either (Sr0.95Cr0.05)TiO3 or Sr(Ti0.95Cr0.05)O3, clearly showed considerable absorption to visible light in the wavelength range up to >800 nm. These results indicated the substitution of Cr cations for Sr2+ in (Sr0.95Cr0.05)TiO3 and for Ti4+ in Sr(Ti0.95Cr0.05)O3, respectively. Correspondingly, a new absorption edge around ∼650 nm (∼1.9 eV) was formed in either (Sr0.95Cr0.05)TiO3 or Sr(Ti0.95Cr0.05)O3, which could preliminarily be ascribed to the excitation from the occupied Cr 3d orbitals to Ti 3d orbitals.13 These results are basically in agreement with the previous calculations of Crdoped TiO2, in which the occupied Cr3+ level is usually ∼2.2 eV lower than the conduction band bottom formed by Ti 3d (e.g.) or ∼1.0 eV higher than the valence band top formed by O 2p.11 The broad absorption band from 650 to 800 nm could probably be ascribed to the oxygen vacancy states, which were usually located between 0.75 and 1.18 eV and below the conduction band bottom.22 Because the Cr cations in (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3 are 12-fold and 6-fold coordinated with the oxygen anions, respectively, a precise com-

SrTiO3 Doped with Cr Cations on Different Sites

J. Phys. Chem. B, Vol. 110, No. 32, 2006 15827

Figure 6. HRTEM images of (a) Pt/(Sr0.95Cr0.05)TiO3, and (b) Pt/ Sr(Ti0.95Cr0.05)O3, showing the similar configurations of Pt nanoparticles on the catalyst surfaces of (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3.

Figure 5. SEM morphologies of (a) SrTiO3, (b) (Sr0.95Cr0.05)TiO3, and (c) Sr(Ti0.95Cr0.05)O3 powder samples.

parison showed that the UV-vis absorption spectra of (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3 were slightly different; i.e., the latter had a higher visible light shoulder than the former. It should be pointed out that, the UV-vis spectra of both (Sr1-xCrx)TiO3 and Sr(Ti1-xCrx)O3 (x ) 0, 0.01, 0.03, 0.05) changed gradually with variation of the doping amount of Cr cations (see Supporting Information S1), implying the occupation of Cr in the crystal lattices of respective Cr-doped SrTiO3. Further details will be discussed in section 3.4 below in combination with the XPS results and the photocatalytic properties. (D) FE-SEM ObserVation on the Morphologies of SrTiO3, (Sr0.95Cr0.05)TiO3, and Sr(Ti0.95Cr0.05)O3. Parts a-c of Figure 5

show the SEM micrographs of SrTiO3, (Sr0.95Cr0.05)TiO3, and Sr(Ti0.95Cr0.05)O3 powder samples synthesized under the same condition. We can notice that the three samples have almost the same mean particle size but different surface morphologies. In contrast to the smooth surface of the SrTiO3 particle, plenty of microsteps and edges have been formed on the particle surfaces of Cr-doped SrTiO3. It indicates that, on one hand, the surface areas of Cr-doped samples are probably larger than that of SrTiO3; on the other hand, the microsteps and edges can be expected as the effective reaction sites. Using the BET method, the surface areas of SrTiO3, (Sr0.95Cr0.05)TiO3, and Sr(Ti0.95Cr0.05)O3 powder samples were measured to be 1.61, 1.92, and 2.13 m2/g, respectively, being consistent with the SEM observations. The EDS element mapping confirmed that all elements, especially the doped Cr, were homogeneously distributed in both Sr(Ti0.95Cr0.05)O3 and (Sr0.95Cr0.05)TiO3 powder samples. (E) HRTEM ObserVation on the Configurations of Pt Nanopaticles Deposited on the Surfaces of (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3. Nanocomposite structures of semiconductor/ noble metal have been found useful in improving the efficiency of photocatalytic and photoelectrochemical conversion of light energy because of enhancement of the interfacial charge-transfer kinetics. For instance, platinum has been frequently used in photoreactions for the purpose of enhancing the reaction rates of water splitting,23,24 hydrogen evolution from aqueous alcohol solution,25,26 and oxidization of organic compounds.27-29 To clarify whether the Pt particles have played a role in the different H2 evolution activities of the two Cr-doped SrTiO3, as will be described below in section 3.2, the Pt particles on the surfaces of (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3 have been extensively observed with HRTEM. From the typical micrographs as shown in Figure 6a,b, we can see that the Pt particles have almost the same mean size (3-5 nm) and dispersion configuration.

15828 J. Phys. Chem. B, Vol. 110, No. 32, 2006

Figure 7. Photocatalytic H2 evolution from the aqueous CH3OH solution (50 mL CH3OH + 220 mL H2O) over Pt(0.4wt%)-loaded (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3 powder (0.25 g) photocatalyst under visible light irradiation (λ g 420 nm). Light source: a 300 W Xe lamp (operated at 200 W).

Figure 8. Long-term course of photocatalytic H2 evolution from the aqueous CH3OH solution (50 mL of CH3OH + 220 mL of H2O) over Pt(0.4wt%)/(Sr0.95Cr0.05)TiO3 powder (0.25 g) photocatalyst under visible light irradiation (λ g 420 nm). Before starting the next cycle of reaction, the H2 gas evolved in the last cycle was evacuated from the reaction cell. Light source: a 300 W Xe lamp (operated at 200 W).

Furthermore, XPS analysis revealed that the Pt particles loaded on the surfaces of both (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3 were all the metallic state after light irradiation. The abovementioned results indicated that neither the configuration nor the valence state of Pt nanoparticles should be responsible for the different photocatalytic activities of H2 evolution over Sr(Ti0.95Cr0.05)O3 and (Sr0.95Cr0.05)TiO3. 3.2. Photocatalytic Activities. (A) H2 EVolution ActiVities of (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3. H2 evolution from aqueous CH3OH solution over Pt(0.4wt%)-loaded (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3 powder catalysts under visible light irradiation (λ g 420 nm) is shown in Figure 7. Clearly, the H2 evolution activities were quite different: the average H2 evolution rate over Pt(0.4wt%)/(Sr0.95Cr0.05)TiO3 was ∼21 µmol/ h, while that over Pt(0.4wt%)/Sr(Ti0.95Cr0.05)O3 was only ∼0.2 µmol/h. Figure 8 shows a long-term course of H2 evolution over Pt(0.4wt%)/(Sr0.95Cr0.05)TiO3 under visible light irradiation (λ g 420 nm). With an increase of irradiation time, the amount of evolved H2 increased, whereas almost no H2 was evolved when the light was turned off (“dark test”). Probably due to the suppressing effect caused by the evolved H2, the evolution rate was gradually decreased in each reaction cycle of ∼50 h. Nevertheless, the initial evolution rate could be recovered in the next reaction cycle after evacuating the H2 gas evolved in the last cycle. The molar amount of photocatalyst used in the reaction was ∼1.38 mM, among which the doped Cr was ∼68.8 µM. The overall molar amount (5.52 mM) of H2 evolved during the long-term reaction of 10 cycles (>500 h) was nearly 4 times that of catalyst used in the reaction. In terms of the reacted electrons to the amount of Cr cations doped in (Sr0.95Cr0.05)-

Wang et al.

Figure 9. Wavelength dependences of H2 evolution rate and quantum yield over Pt(0.4wt%)/(Sr0.95Cr0.05)TiO3 (0.25 g) powder photocatalyst. Light source: a 300 W Xe lamp (operated at 200 W). Reaction time: 2 h.

TiO3, the turnover number reached more than 80. Deactivation of the photocatalyst was not observed throughout the whole reaction process. The above-mentioned results indicated that the synthesized photocatalyst was stable upon light irradiation and that the H2 evolution inherently resulted from the photocatalytic reaction process. (B) WaVelength Dependences of H2 EVolution Rate and Quantum Yield oVer Pt(0.4wt%)/(Sr0.95Cr0.05)TiO3. The wavelength dependences of the H2 evolution rate and the quantum yield over Pt(0.4wt%)/(Sr0.95Cr0.05)TiO3 were performed by using various long-pass cutoff filters and band-pass filters, respectively. As shown in Figure 9, both the evolution rate and quantum yield of H2 were increased with a decrease of the incident light wavelength, being consistent with the UV-vis absorption spectrum as shown in Figure 3. Considerable activity was still obtained even under the irradiation of visible light of as long as 580 nm. The quantum yield at λ ) 420.4 nm (λ0 ) 420.4 nm, Tmax ) 44.8%, ∆λ/2 ) 14.7 nm) reached 0.86%. All these results confirmed that the H2 evolution was inherently the result of photocatalytic reaction over Pt(0.4wt%)/(Sr0.95Cr0.05)TiO3. (C) Effect of Photoreduction by UV Light Irradiation on the H2 EVolution ActiVities of (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3. As mentioned above, the Sr(Ti0.95Cr0.05)O3 showed almost no activity of H2 evolution under visible light irradiation due to the existence of hexavalent Cr cations (Cr6+). It is known that the Cr6+ level in Sr(Ti0.95Cr0.05)O3 is usually located below the Ti 3d level,18 and the potential level of Ti 3d in Sr(Ti0.95Cr0.05)O3 is so critical that it is just above the potential level for H2 evolution. We can thus deduce that the potential level of the empty Cr6+ is lower than that for H2 evolution. As a result, H2 could not be evolved over Sr(Ti0.95Cr0.05)O3 since the empty Cr6+ level behaved as the trapping center for photoinduced electrons. Concerning the charge and distortion compensations, the Cr cations with different valence states (Cr6+ and Cr3+) in Sr(Ti0.95Cr0.05)O3 should neighbor each other, with the oxygen ions being shifted slightly toward the more charged Cr6+.30 We think that if the hexavalent chromium (Cr6+) is reduced to trivalent chromium (Cr3+) by photogenerated electrons, the Sr(Ti0.95Cr0.05)O3 might be rendered to show a high activity for H2 evolution. For this consideration, a purposely designed experiment was carried out, and the result confirmed our hypothesis. As shown in Figure 10, almost no H2 was evolved over Sr(Ti0.95Cr0.05)O3 under visible light irradiation for more than 90 h. The solution containing Sr(Ti0.95Cr0.05)O3 powders after a long time of visible light irradiation was then transferred to a UV light reaction cell with addition of an extra 100 mL of H2O. In the

SrTiO3 Doped with Cr Cations on Different Sites

Figure 10. Photocatalytic H2 evolution from the aqueous CH3OH solution over Pt(0.4wt%)/Sr(Ti0.95Cr0.05)O3 powder photocatalyst (0.25 g) in a cycle of visible light-UV light-visible light irradiation. UV light source: a 400 W high-pressure Hg lamp. Visible light source: a 300 W Xe lamp (operated at 200 W) + cutoff filter (λ g 420 nm).

initial stage (∼15 h) of reaction under UV light, the H2 evolution activity was still very low. After this stage, a much higher H2 evolution rate was obtained. The powder samples after the above-mentioned UV light irradiation were filtered and transferred to an outer irradiation reaction cell containing an aqueous methanol solution (220 mL of H2O and 50 mL of CH3OH), and then H2 evolution was performed under visible light. It was very interesting to note that the H2 evolution activity of Sr(Ti0.95Cr0.05)O3 under visible light was dramatically improved. Preliminarily, we assumed that the hexavalent chromium (Cr6+) had been reduced to the trivalent chromium (Cr3+) by photogenerated electrons during UV light irradiation. To confirm this assumption, the sample Sr(Ti0.95Cr0.05)O3 after UV light irradiation was analyzed by XPS. In contrast to the Cr 2p spectrum in Sr(Ti0.95Cr0.05)O3 before UV light irradiation (Figure 3b), which contained not only the peak assigned to trivalent chromium at 577.6 eV but also the peak assigned to hexavalent chromium at 581.2 eV, the Cr 2p spectrum in Sr(Ti0.95Cr0.05)O3 after UV light irradiation (Figure 3b′) showed only one peak assigned to the trivalent chromium at 577.6 eV, which was identical to that in (Sr0.95Cr0.05)TiO3. A similar reaction with UV light irradiation was also performed on (Sr0.95Cr0.05)TiO3. It showed that UV light irradiation did not impose any effect on the H2 evolution activity of (Sr0.95Cr0.05)TiO3 under visible light irradiation, because no change of the valence state of Cr3+ was observed from the Cr 2p spectra in (Sr0.95Cr0.05)TiO3 before and after UV light irradiation (Figure 3a,a′). In general, all of the above-mentioned results further confirmed that the different photocatalytic H2 evolution activities of Sr(Ti0.95Cr0.05)O3 and (Sr0.95Cr0.05)TiO3 resulted from the different valence states of Cr cations doped at different positions. 3.3. Electronic Structures of (Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3. Previous calculations on the electronic structure of SrTiO3 31-35 suggest that the valence band top is made up predominately of the O 2p states, and the conduction band bottom is determined by the Ti 3d states. Doping Cr cations at either the Sr2+ site or the Ti4+ site in SrTiO3 will certainly change the electronic structures, which can be judged simply from the UV-vis absorption spectra as shown in Figure 4. XPS analysis revealed that the Cr cations doped in (Sr0.95Cr0.05)TiO3 were only one state, Cr3+, while those doped in Sr(Ti0.95Cr0.05)O3 were two states: Cr6+ and Cr3+. Since the Cr6+ level in Sr(Ti0.95Cr0.05)O3 is usually located below the Ti 3d level and is also lower than the potential level for H2 evolution, it can only behave as the trapping center for photoinduced electrons. As a consequence, H2 cannot be evolved over Sr(Ti0.95Cr0.05)O3. For (Sr0.95Cr0.05)TiO3 containing only Cr3+, its valence band

J. Phys. Chem. B, Vol. 110, No. 32, 2006 15829

Figure 11. Schematic band structures of (a) SrTiO3, (b) (Sr0.95Cr0.05)TiO3, and (c) Sr(Ti0.95Cr0.05)O3 before and after photoreduction. The hexavalent Cr6+ was reduced to Cr3+ upon UV light irradiation (see text for details).

bottom is lifted up by the occupied Cr3+ level, while the original conduction band bottom of SrTiO3 determined by Ti 3d is not affected. This is similar to the case of Cr-doped TiO2, in which the occupied Cr3+ level has previously been calculated to be ∼2.2 eV lower than the conduction band bottom formed by Ti 3d (e.g.) or ∼1.0 eV higher than the valence band top formed by O 2p.11 Therefore, the (Sr0.95Cr0.05)TiO3 shows a higher activity of H2 evolution than Sr(Ti0.95Cr0.05)O3 under visible light irradiation. Upon UV light irradiation, the hexavalent Cr cations (Cr6+) in Sr(Ti.95Cr.05)O3 were reduced to the trivalent Cr cations (Cr3+) by photogenerated electrons. The trivalent Cr cations (Cr3+), as in the case of (Sr0.95Cr0.05)TiO3, were located above O 2p and acted as electron donors, thus rendering the Sr(Ti0.95Cr0.05)O3 to show a high activity for H2 evolution. There was almost no change of the UV-vis absorption spectra of Sr(Ti0.95Cr0.05)O3 before and after UV light irradiation. Preliminarily, we speculated that the potential difference between Cr6+ and Ti 3d might be close to that between Cr3+ and O 2p. Upon UV light irradiation, the Cr6+ level disappeared, and, meanwhile, the Cr3+ reduced from Cr6+ overlapped with those already existing, lifting up the total potential level of Cr3+ above O 2p. Figure 11 schematically illustrates the electronic structures of SrTiO3, (Sr0.95Cr0.05)TiO3, and Sr(Ti0.95Cr0.05)O3 before and after photoreduction. The photoinduced electrons in the conduction band of (Sr.95Cr.05)TiO3 and in that of photoreduced Sr(Ti0.95Cr0.05)O3 were undoubtedly able to reduce water to form H2 in the presence of sacrificial reagent CH3OH. Regarding the donor level of occupied Cr 3d, which possesses both thermodynamic and kinetic potentials for oxidation of methanol, we are not sure whether it is also thermodynamically deep enough for oxidation of H2O to form O2. The fact was that O2 was not able to evolve over either (Sr0.95Cr0.05)TiO3 or Sr(Ti0.95Cr0.05)O3. It appears that, as a process of four-electron oxidation, O2 evolution is much harder than oxidation of methanol. 4. Conclusions The photophysical and photocatalytic properties of Cr-doped SrTiO3s(Sr0.95Cr0.05)TiO3 and Sr(Ti0.95Cr0.05)O3swere investigated systematically. Extremely different photocatalytic activities for H2 evolution were observed over (Sr0.95Cr0.05)TiO3 and Sr(Ti.95Cr.05)O3 because of the different valent states of Cr cations doped at different sites (Sr2+ or Ti4+) in these two materials. The Cr cations doped at Sr2+ site in (Sr0.95Cr0.05)TiO3 were trivalent state (Cr3+), while those doped at Ti4+ site in Sr(Ti0.95Cr0.05)O3 were mixed trivalent states (Cr3+ and Cr6+). The empty Cr6+ level usually behaves as the trapping center for photoinduced electrons for its potential is lower than that

15830 J. Phys. Chem. B, Vol. 110, No. 32, 2006 for H2 evolution. Thus, H2 is not able to evolve over Sr(Ti0.95Cr0.05)O3. In contrast, the occupied Cr3+ level is located at a definite position (∼1.0 eV) above O 2p, rendering the (Sr0.95Cr0.05)TiO3 to evolve H2 under visible light irradiation. The present study has proven for the first time that properly doping trivalent Cr cations at the Sr2+ site is an alternative approach for SrTiO3 to show photocatalytic activity of H2 evolution under visible light irradiation. Furthermore, it is suggested that the hexavalent Cr cations should be avoided in Cr-containing visible-light-driven photocatalysts for water splitting. Acknowledgment. The authors thank Dr. Sei Fukushima and Mr. Satoshi Ota for their help in the XPS measurements and valuable discussions on the experimental results. This work was partially supported by a Grant-in-Aid for Scientific Research on Priority Areas (417) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of the Japanese Government. Supporting Information Available: The UV-vis spectra of Sr(Ti1-xCrx)O3 and (Sr1-xCrx)TiO3 (x ) 0, 0.01, 0.03, 0.05) and the dependence of H2 evolution activity on the doping amount of Cr in (Sr1-xCrx)TiO3 (x ) 0, 0.01, 0.03, 0.05). These materials are available free of charge via the Internet at http:// pubs.acs.org. References and Notes (1) Fujishima, A.; Honda, K. Nature 1972, 238, 37. (2) Zou, Z.; Ye, J.; Sayama, K.; Arakawa, H. Nature 2001, 414, 6254. (3) Kiwi, J. In Homogeneous and Heterogeneous Photocatalysis; Pelizzetti E., Serpone, N., Reidol, D., Eds.; Kluwer: Dordrecht, The Netherlands, 1986; p 275. (4) Pena, M. A.; Fierro, J. L. G. Chem. ReV. 2001, 101, 1981. (5) Kutty, T. R. N.; Avudaithai, M. In Properties and Applications of PeroVskite-type Oxides; Tejuca, L. G., Fierro, J. L. G., Eds.; Dekker: New York, 1993; p 307. (6) Wrighton, M. S.; Ellis, A. B.; Wolczanski, P. T.; Morse, D. L.; Abrahamson, H. B.; Ginley, D. S. J. Am. Chem. Soc. 1976, 98, 277. (7) Lehn, J. M.; Sauvage, J. P.; Ziessel, R. NouV. J. Chim. 1980, 4, 62.

Wang et al. (8) Maglizzo, R. S.; Krasna, A. I. Photochem. Photobiol. 1983, 38, 15. (9) Lehn, J. M.; Sauvage, J. P.; Ziessel, R.; Hiraire, L. Isr. J. Chem. 1982, 22, 168. (10) Domen, K.; Kudo, A.; Onishi, T.; Kosugi, N.; Kuroda, K. J. Phys. Chem. 1986, 90, 292. (11) Umebayashi, T.; Yamaki, T.; Itoh, H.; Asai, K. J. Phys. Chem. Solids 2002, 63, 1909. (12) Herrmann, J. M.; Disdier, J.; Pichat, P. Chem. Phys. Lett. 1984, 106, 618. (13) Serpone, N.; Lawless, D. Langmuir 1994, 10, 643. (14) Sakata, Y.; Yamamoto, T.; Okazaki, T.; Imamura, H.; Tsuchida, S. Chem. Lett. 1998, 1253. (15) Campet, G.; Dare-Edwards, M. P.; Hamnett, A.; Goodenough, J. B. NouV. J. Chim. 1980, 4, 501. (16) Mackor, A.; Blasse, G. Chem. Phys. Lett. 1981, 77, 6. (17) Lam, R. U. E. T.; Haart L. G. J. D.; Wiersma, A. W.; Blasse, G.; Tinnemans, A. H. A.; Mackor, A. Mater. Res. Bull. 1981, 16, 1593. (18) Kato, H.; Kudo, A. J. Phys. Chem. B 2002, 106, 5029. (19) Ishii T.; Kato, H.; Kudo, A., J. Photochem. Photobiol., A 2004, 163, 181. (20) Shannon, R. D. Acta Crystallogr. 1976, A32, 751. (21) Meijer, G. I.; Sdaub, U.; Janousch, M.; Johnson, S. L.; Delley, B.; Neisius, T. Phys. ReV. B 2005, 72, 155102. (22) Cronemeyer, D. D. Phys. ReV. 1959, 113, 1222. (23) Sato, S.; White, J. M. Chem. Phys. Lett. 1980, 72, 83. (24) Sato, S.; White, J. M. J. Phys. Chem. 1981, 85, 592. (25) Kawai, T.; Sakata, T. J. Chem. Soc., Chem. Commun. 1980, 694. (26) Pichat, P.; Herrmann, J. M.; Disdier, J.; Courbon, H.; Mozzanega, M. N. J. Chim. 1981, 5, 627. (27) Izumi, I.; Fan, F. R.; Bard, A. J. J. Phys. Chem. 1981, 85, 218. (28) St. John, M. R.; Furgala, A. J.; Sammells, A. F. J. Phys. Chem. 1983, 87, 801. (29) Izumi, I.; Dunn, W. W.; Wilbourn, K. O.; Fan, F. R.; Bard, A. J. J. Phys. Chem. 1980, 84, 3207. (30) Betsuyaku, K.; Tanaka, H.; Katayama-Yoshida, H.; Kawai, T. Jpn. J. Appl. Phys. 2001, 40, 6911. (31) Guo, X. G.; Chen, X. S.; Sun, Y. L.; Sun, L. Z.; Zhou, X. H.; Lu, W. Phys. Lett. A 2003, 317, 501. (32) Van Benthem, K.; Elsa¨sser, C.; French, R. H. J. Appl. Phys. 2001, 90, 6156. (33) De Groot, F. M. F.; Faber, J.; Michiels, J. J. M.; Czyzyk, M. T.; Abbate, M.; Fuggle, J. C. Phys. ReV. B 1993, 48, 2074. (34) Cappellini, G.; Bouette-Russo, S.; Amadon, B.; Noguera, C.; Finocchi, F. J. Phys.: Condens. Matter 2000, 12, 3671. (35) Mo, S. D.; Ching, W. Y.; Chisholm, M. F.; Duscher, G. Phys. ReV. B 1999, 60, 2416.