Ethanol-Quenching Introduced Oxygen Vacancies in Strontium Titanate Surface and the Enhanced Photocatalytic Activity

Modification of the surface properties of SrTiO3 crystals by regulating the reaction environment in order to improve the photocatalytic activity has been widely studied. However, the development of a facile, effective, and universal method to improve the photocatalytic activity of these crystals remains an enormous challenge. We have developed a simple method to modify the surface environment of SrTiO3 by ethanol quenching, which results in enhanced UV, visible and infrared light absorption and photocatalytic performance. The SrTiO3 nanocrystals were preheated to 800 °C and immediately quenched by submersion in ethanol. X-ray diffraction patterns, electron paramagnetic resonance spectra, and X-ray photoelectron spectra indicated that upon rapid ethanol quenching, the interaction between hot SrTiO3 and ethanol led to the introduction of a high concentration of oxygen vacancies on the surface of the SrTiO3 lattice. Consequently, to maintain the regional charge balance of SrTiO3, Sr2+ could be substituted for Ti4+. Moreover, oxygen vacancies induced localized states into the band gap of the modified SrTiO3 and acted as photoinduced charge traps, thus promoting the photocatalytic activity. The improved photocatalytic performance of the modified SrTiO3 was demonstrated by using it for the decomposition of rhodamine B and production of H2 from water under visible or solar light.


Introduction
Since the discovery of photoelectrochemical water splitting on a titania electrode by Fujishima and Honda in 1972, photocatalysis using semiconductors has been widely studied [1]. Semiconductor-based photocatalysts are capable of directly converting solar energy to chemical energy, which provides a facile approach for environmental protection and H 2 production under sunlight irradiation [2][3][4]. Among these semiconductors, titanium dioxide (TiO 2 ) has attracted special interest because of its chemical stability, non-toxicity, and low cost [5,6]. In addition, Strontium titanate (SrTiO 3 ), a perovskite-type oxide, has been classified a wide-band gap (3.1-3.7 eV) semiconductor photocatalyst in the field of light energy exploitation. SrTiO 3 possesses various outstanding physical and chemical properties such as good chemical/catalytic stability, suitable band position, and susceptibility to change by other substance. Notably, the conduction band of SrTiO 3 is more negative than that of TiO 2 , which is 1 g of original SrTiO 3 was transferred to the Muffle furnace. The SrTiO 3 nanoparticles were heated at 800 • C for 20 min, and then open the Muffle furnace door, immediately took out the sintering boat and submerged the SrTiO 3 in 45 mL ethanol (AR) at room temperature for rapid quenching. Afterwards, the quenched sample was filtered and then dried at 80 • C for 3 h for further use.

Characterization
An X-ray diffractometer (XRD, D8 ADVANCE, Bruker, Karlsruhe, Germany) was used to check the phase structures of all the samples. Data were collected between 20 • and 80 • (2θ) with a 0.02 • step size using Cu Kα irradiation, at 36 kV tube voltage and 20 mA tube current. Field-emission transmission electron microscopy (FE-TEM, Talos F200S, FEI, Thermo, Hillsboro, OR, USA) and Field-emission scanning electron microscopy (FE-SEM, SU8220, Hitach, Tokyo, Japan) was used to determine the surface change, particle size and morphology of the samples. Sample specimens for the FE-TEM and the FE-SEM observations were prepared as follows: The powdered sample was dispersed in ethanol in an ultrasonic washing bath, and then, a drop of the suspension was dripped slightly onto a micro grid or a silicon slice and dried before imaging. X-ray Photoelectron Spectroscopy (XPS, Escalab 250Xi, Thermo Fisher, Bremen, Germany) was used to analyze the chemical composition and relative amount of the elements on the surface of the samples, with a reference of C1 s and the excitation source of 150 W Al Kα X-rays. An Electron paramagnetic resonance spectrometer (EPR, EMXplus-10/12, Bruker, Karlsruhe, Germany) was used to detect unpaired electrons in the samples at room temperature. A UV-Vis-NIR spectrophotometer (DRUV-vis, UV-3600 Plus, SHIMADZU, Kyoto, Japan) were used to record the UV-vis diffuse reflectance absorption in the range of 200-2000 nm.

Photocatalytic Degradation of RhB
The photocatalytic degradation activities of Q-SrTiO 3 were evaluated by monitoring the rate of decomposition of RhB in an aqueous solution under visible-light or UV-light irradiation from a 500 W Xe lamp equipped with a UV cut off filter (>420 nm) or 500 W Hg lamp, the light intensity of 500 W Xe lamp equipped with a UV cut off filter and 500 W Hg lamp are 6 mW/cm 2 and 48 mW/cm 2 , respectively. A cylindrical Pyrex vessel equipped with a lamp was used as the photocatalytic reactor, with water circulation to keep the reaction temperature at about 27 • C. 40 mL of an aqueous solution of RhB (4 × 10 −5 M) and 0.02 g Q-SrTiO 3 or SrTiO 3 were placed in a quartz tube for the degradation reaction. Before the photodegradation, a dark reaction was conducted for 30 min to ensure adsorption-desorption equilibrium between the photocatalyst and the RhB solution. Continuous magnetic stirring was carried out to keep the photocatalyst suspended in the RhB solution. Next, the mixture was exposed to visible light or UV light. The samples were collected at regular intervals (1.5 h or 1 h), and the concentration of RhB in the solution was determined using a UV-Vis spectrophotometer at 553 nm. The percentage of degradation was recorded as C/C 0 , and the reaction constant (K app ) was calculated from the slope of the linear regression obtained from the plot of −ln (C/C 0 ) vs. time, where C 0 and C are the absorbance of the RhB solution initially and at a particular time, respectively.

Photocatalytic Evolution of hydrogen
The photocatalytic H 2 production experiments were conducted in a 400 mL Pyrex quartz glass reactor at normal pressure and temperature. The photocatalyst (100 mg) was dispersed in 100 mL of 10% aqueous methanol solution (methanol acting as a sacrificial agent) using a magnetic stirrer. Then the reaction mixture was dispersed in an ultrasonic washing bath for 10 min. Before the irradiation by a 300 W Xe lamp (CRL-HXF300, China) as the sunlight source, the reactor was deaerated with nitrogen gas. During the photocatalytic reaction, the reactant solution was maintained at room temperature by using a Low-temperature cooling circulating pump (CEL-CR300, China), and magnetic stirring was continually maintained to keep the photocatalyst suspended in the aqueous methanol solution.
The amount of H 2 generated was tested using an online Shimadzu GC-2014C gas chromatograph (Shimadzu, Japan) equipped with an MS-5A column. The total reaction time for each sample was 5 h, and the H 2 concentration was measured every hour.

Characterization of the Photocatalysts
The peaks in the powder XRD patterns of SrTiO 3 and Q-SrTiO 3 ( Figure 1) matched with the (100), (110), (111), (200), (210), (211), (220) and (310) planes, indicating a characteristic SrTiO 3 cubic structure (JCPDS card: 73-6001). Moreover, a small peak for SrCO 3 was observed, probably due to the coexistence of SrTiO 3 and SrCO 3 under the atmospheric operating conditions adopted in the hydrothermal method [29]. No other diffraction peak was observed in the XRD patterns. Comparison with the XRD patterns from the local enlargement of the diffraction peaks (inset of Figure 1) revealed a slight shift (2θ~0.15 • ) of the (110) peak to a lower angle for Q-SrTiO 3 . According to the Bragg equation (2d sin θ = λ. where d, θ and λ are the crystal spacing, diffraction angle, and X-ray wavelength, respectively), a shift in the diffraction peaks toward a lower angle suggests an increase in the lattice parameters. This might be attributed to the substitution of Sr 2+ (ionic radius Sr 2+ > Ti 4+ ) for Ti 4+ in Q-SrTiO 3 [30]. Nanomaterials 2019, 9, x FOR PEER REVIEW 4 of 11 chromatograph (Shimadzu, Japan) equipped with an MS-5A column. The total reaction time for each sample was 5 h, and the H2 concentration was measured every hour.

Characterization of the Photocatalysts
The peaks in the powder XRD patterns of SrTiO3 and Q-SrTiO3 ( Figure 1) matched with the (100), (110), (111), (200), (210), (211), (220) and (310) planes, indicating a characteristic SrTiO3 cubic structure (JCPDS card: 73-6001). Moreover, a small peak for SrCO3 was observed, probably due to the coexistence of SrTiO3 and SrCO3 under the atmospheric operating conditions adopted in the hydrothermal method [29]. No other diffraction peak was observed in the XRD patterns. Comparison with the XRD patterns from the local enlargement of the diffraction peaks (inset of Figure 1) revealed a slight shift (2θ~0.15°) of the (110) peak to a lower angle for Q-SrTiO3. According to the Bragg equation (2d sin θ = λ. where d, θ and λ are the crystal spacing, diffraction angle, and X-ray wavelength, respectively), a shift in the diffraction peaks toward a lower angle suggests an increase in the lattice parameters. This might be attributed to the substitution of Sr 2+ (ionic radius Sr 2+ > Ti 4+ ) for Ti 4+ in Q-SrTiO3 [30]. 20   The UV-visible absorption spectra of SrTiO3 and Q-SrTiO3 ( Figure 2a) exhibited an absorption onset at ~400 nm, which corresponds to a band gap of 3.1 eV. In contrast to the SrTiO3, the photoabsorption of Q-SrTiO3 was dramatically enhanced in the both UV, visible and infrared light regions, consistent with the color change of the sample from white to gray (inset of Figure 2a). The improved light absorption was attributed to the formation of surface oxygen vacancies in Q-SrTiO3. Similar results have been observed in other studies [19,31].
EPR is highly sensitive to unpaired electrons; hence, it was used for the detection of oxygen vacancies and Ti 3+ species in SrTiO3 and Q-SrTiO3 (Figure 2b) Both SrTiO3 and Q-SrTiO3 showed a distinct EPR signal at g = 1.977 and g = 2.002, which could be ascribed to Ti 3+ and oxygen vacancies, respectively. Because of its intrinsic non-stoichiometry, SrTiO3 always contains a fraction of oxygen vacancies and Ti 3+ ions [32]. EPR spectra revealed that Q-SrTiO3 exhibited a stronger signal intensity for oxygen vacancies than did SrTiO3, indicating the presence of more oxygen vacancies in Q-SrTiO3, thus favoring enhanced photocatalytic activity of the Q-SrTiO3. Takata and Domen also demonstrated that doping of a cation with a lower valence ion than that of the parent cation (such as Sr 2+ in SrTiO3) can introduce oxygen vacancies, thus effectively improving its photocatalytic activity [32]. Thus, based on the XRD results, due to the exists of abundant oxygen vacancies, in order to keep the The UV-visible absorption spectra of SrTiO 3 and Q-SrTiO 3 (Figure 2a) exhibited an absorption onset at~400 nm, which corresponds to a band gap of 3.1 eV. In contrast to the SrTiO 3 , the photoabsorption of Q-SrTiO 3 was dramatically enhanced in the both UV, visible and infrared light regions, consistent with the color change of the sample from white to gray (inset of Figure 2a). The improved light absorption was attributed to the formation of surface oxygen vacancies in Q-SrTiO 3 . Similar results have been observed in other studies [19,31].
EPR is highly sensitive to unpaired electrons; hence, it was used for the detection of oxygen vacancies and Ti 3+ species in SrTiO 3 and Q-SrTiO 3 (Figure 2b) Both SrTiO 3 and Q-SrTiO 3 showed a distinct EPR signal at g = 1.977 and g = 2.002, which could be ascribed to Ti 3+ and oxygen vacancies, respectively. Because of its intrinsic non-stoichiometry, SrTiO 3 always contains a fraction of oxygen vacancies and Ti 3+ ions [32]. EPR spectra revealed that Q-SrTiO 3 exhibited a stronger signal intensity for oxygen vacancies than did SrTiO 3 , indicating the presence of more oxygen vacancies in Q-SrTiO 3 , thus favoring enhanced photocatalytic activity of the Q-SrTiO 3 . Takata and Domen also demonstrated that doping of a cation with a lower valence ion than that of the parent cation (such as Sr 2+ in SrTiO 3 ) can introduce oxygen vacancies, thus effectively improving its photocatalytic activity [32]. Thus, based on Nanomaterials 2019, 9, 883 5 of 11 the XRD results, due to the exists of abundant oxygen vacancies, in order to keep the regional charge balance of the Q-SrTiO 3 , the Sr 2+ ions could substituted for Ti 4+ ions, so that lattice expansion occurs (Figure 2c) [32,33]. SrTiO3 + SrO = 2SrTi + Vo·· + 4OO. The high-resolution transmission electron microscopy (HR-TEM) images (Figure 3a,b) revealed the interplanar spacing of SrTiO3 and Q-SrTiO3 crystals to be ~0.27 nm, which is consistent with the d-spacings of the (110) crystallographic planes of cubic SrTiO3. However, Liu et al. have reported that an ice-water quenching TiO2 had introduced a disordered surface layer surrounding the crystalline core, and the surface lattice distortion is related to the generation of oxygen vacancies during the icewater quenching [34]. In contrast, Q-SrTiO3 prepared in this work used 800 °C ethanol quenching that did not lead to specific disordered surface layer, therefore, the generation of more oxygen vacancies in Q-SrTiO3 have not introduced disordered surface layer. The SEM images (Figure 3c,d) revealed the particle size and particle morphology of Q-SrTiO3 and SrTiO3 show no change, and the average diameters of Q-SrTiO3 and SrTiO3 nanocrystals are ~100 nm, hence, ethanol-quenching can not change the particle size and particle morphology of samples. The high-resolution transmission electron microscopy (HR-TEM) images (Figure 3a,b) revealed the interplanar spacing of SrTiO 3 and Q-SrTiO 3 crystals to be~0.27 nm, which is consistent with the d-spacings of the (110) crystallographic planes of cubic SrTiO 3 . However, Liu et al. have reported that an ice-water quenching TiO 2 had introduced a disordered surface layer surrounding the crystalline core, and the surface lattice distortion is related to the generation of oxygen vacancies during the ice-water quenching [34]. In contrast, Q-SrTiO 3 prepared in this work used 800 • C ethanol quenching that did not lead to specific disordered surface layer, therefore, the generation of more oxygen vacancies in Q-SrTiO 3 have not introduced disordered surface layer. The SEM images (Figure 3c,d) revealed the particle size and particle morphology of Q-SrTiO 3 and SrTiO 3 show no change, and the average diameters of Q-SrTiO 3 and SrTiO 3 nanocrystals are~100 nm, hence, ethanol-quenching can not change the particle size and particle morphology of samples.
XPS was used to investigate the surface chemical composition and VB position of SrTiO 3 and Q-SrTiO 3 . The Sr 3d 5/2 , Sr 3d 3/2 , Ti 2p 1/2 , and Ti 2p 3/2 binding energies were 133.2, 134.6, 458.0, and 464.0 eV, respectively, in accordance with the literature values (Figure 4a,b) [35]. The Sr 3d and Ti 2p spectra showed no obvious variation between SrTiO 3 and Q-SrTiO 3 . The Sr/Ti ratio of the samples was estimated according to the peak area and sensitivity factor of Sr 3d and Ti 2p ( Table 1). The atomic ratio of Sr to Ti on the surface was about 1.64 and 2.15 for SrTiO 3 and Q-SrTiO 3 , respectively. The larger Nanomaterials 2019, 9, 883 6 of 11 atomic ratio of Sr/Ti in Q-SrTiO 3 than that in SrTiO 3 might be due to the substitution of Sr 2+ for Ti 4+ on the surface of the former. XPS was used to investigate the surface chemical composition and VB position of SrTiO3 and Q-SrTiO3. The Sr 3d5/2, Sr 3d3/2, Ti 2p1/2, and Ti 2p3/2 binding energies were 133.2, 134.6, 458.0, and 464.0 eV, respectively, in accordance with the literature values (Figure 4a,b) [35]. The Sr 3d and Ti 2p spectra showed no obvious variation between SrTiO3 and Q-SrTiO3. The Sr/Ti ratio of the samples was estimated according to the peak area and sensitivity factor of Sr 3d and Ti 2p ( Table 1). The atomic ratio of Sr to Ti on the surface was about 1.64 and 2.15 for SrTiO3 and Q-SrTiO3, respectively. The larger atomic ratio of Sr/Ti in Q-SrTiO3 than that in SrTiO3 might be due to the substitution of Sr 2+ for Ti 4+ on the surface of the former.   The O 1s high-resolution X-ray photoelectron spectra of SrTiO 3 and Q-SrTiO 3 showed two typical components of SrTiO 3 (Figure 4c). The two peaks located at 529.2 and 531.5 eV were assigned to bulk oxygen and surface oxygen in the samples, respectively. Based on previous research, it was assumed that the peak intensity of surface oxygen was related to the concentration of oxygen vacancies on the surface of SrTiO 3 and Q-SrTiO 3 [19,36]. The peak intensity at 531.5 eV become stronger for Q-SrTiO 3 , indicating that the concentration of oxygen vacancies on the surface of Q-SrTiO 3 increased after the ethanol-quenching process. The introduction of more oxygen vacancies on the surface of Q-SrTiO 3 lattice resulted in impure/defect states in the band gap, enhancing the visible and near infrared-light absorption of Q-SrTiO 3 [27]. Furthermore, the increased amount of oxygen vacancies can improve the efficient charge transport in Q-SrTiO 3 , followed by the Fermi level shift toward to the CB of the Q-SrTiO 3 , facilitating the separation of photogenerated electron-hole pairs and resulting in enhanced photocatalytic activity [19,27]. In the VB XPS profile, the VB maxima were estimated by  The O 1s high-resolution X-ray photoelectron spectra of SrTiO3 and Q-SrTiO3 showed two typical components of SrTiO3 (Figure 4c). The two peaks located at 529.2 and 531.5 eV were assigned to bulk oxygen and surface oxygen in the samples, respectively. Based on previous research, it was assumed that the peak intensity of surface oxygen was related to the concentration of oxygen vacancies on the surface of SrTiO3 and Q-SrTiO3 [19,36]. The peak intensity at 531.5 eV become stronger for Q-SrTiO3, indicating that the concentration of oxygen vacancies on the surface of Q-SrTiO3 increased after the ethanol-quenching process. The introduction of more oxygen vacancies on the surface of Q-SrTiO3 lattice resulted in impure/defect states in the band gap, enhancing the visible and near infrared-light absorption of Q-SrTiO3 [27]. Furthermore, the increased amount of oxygen vacancies can improve the efficient charge transport in Q-SrTiO3, followed by the Fermi level shift toward to the CB of the Q-SrTiO3, facilitating the separation of photogenerated electron-hole pairs and resulting in enhanced photocatalytic activity [19,27]. In the VB XPS profile, the VB maxima were estimated by linear extrapolation of the peaks to the baselines (Figure 4d). Both, SrTiO3 and Q-SrTiO3 displayed identical VB band positions at 2.3 eV below the Fermi energy, indicating no shift in the VB edge.

Photocatalytic Activity
The photodegradation of RhB in aqueous solution under visible-light irradiation was used to evaluate the photocatalytic activity of the Q-SrTiO3, with SrTiO3 as the control (Figure 5a). After adsorption-desorption equilibrium between the photocatalyst and the RhB solution was achieved in the absence of light, even a slight adsorption of RhB over the samples resulted in a slight decrease in the concentration of RhB. During the photodegradation process, SrTiO3 showed no appreciable reduction in the RhB concentration in aqueous solution; however, Q-SrTiO3 showed higher photocatalytic activity than SrTiO3 in reducing the concentration of RhB in aqueous solution. After 370 min of visible-light irradiation in the presence of Q-SrTiO3, RhB was decomposed by about 20%; in contrast, SrTiO3 caused only 3% decomposition of the dye. The reaction constant (Kapp) was calculated from the slope of the linear regression obtained from the plot of −ln (C/C0) vs. time (Figure

Photocatalytic Activity
The photodegradation of RhB in aqueous solution under visible-light irradiation was used to evaluate the photocatalytic activity of the Q-SrTiO 3 , with SrTiO 3 as the control (Figure 5a). After adsorption-desorption equilibrium between the photocatalyst and the RhB solution was achieved in the absence of light, even a slight adsorption of RhB over the samples resulted in a slight decrease in the concentration of RhB. During the photodegradation process, SrTiO 3 showed no appreciable reduction in the RhB concentration in aqueous solution; however, Q-SrTiO 3 showed higher photocatalytic activity than SrTiO 3 in reducing the concentration of RhB in aqueous solution. After 370 min of visible-light irradiation in the presence of Q-SrTiO 3 , RhB was decomposed by about 20%; in contrast, SrTiO 3 caused only 3% decomposition of the dye. The reaction constant (K app ) was calculated from the slope of the linear regression obtained from the plot of −ln (C/C 0 ) vs. time (Figure 5b). These results suggested that Q-SrTiO 3 shows better activity than SrTiO 3 for the photodegradation of RhB under visible light. 5b). These results suggested that Q-SrTiO3 shows better activity than SrTiO3 for the photodegradation of RhB under visible light. The UV-light photocatalytic activities of Q-SrTiO3 were investigated by monitoring the decomposition of RhB in an aqueous solution, with SrTiO3 as the control (Figure 5c,d). After 270 min of UV-light irradiation, the RhB dye was almost completely decomposed (~90%) by Q-SrTiO3, (Figure The UV-light photocatalytic activities of Q-SrTiO 3 were investigated by monitoring the decomposition of RhB in an aqueous solution, with SrTiO 3 as the control (Figure 5c,d). After 270 min of UV-light irradiation, the RhB dye was almost completely decomposed (~90%) by Q-SrTiO 3 , (Figure 5c). The reaction constant (K app ) was shown in Figure 5d, this results also suggested that Q-SrTiO 3 shows better photocatalytic activity than SrTiO 3 for the photodegradation of RhB. The photocatalytic activity of Q-SrTiO 3 for the photolysis of water to produce H 2 in 100 mL 10% aqueous methanol solution was also studied under solar irradiation, using SrTiO 3 as the control. Figure 5c,d present the time course of H 2 generation for SrTiO 3 and Q-SrTiO 3 under solar light irradiation. Q-SrTiO 3 steadily produced H 2 gas at the rate of 42.12 µmol g −1 h −1 , which was almost 6.2 times higher than that observe with SrTiO 3 (6.83 µmol g −1 h −1 ). All these results demonstrated that Q-SrTiO 3 possesses higher photocatalytic activity than SrTiO 3 .

Conclusions
In this paper, a facile and general method has been introduced to modify the surface environment of SrTiO 3 through an ethanol-quenching process. Q-SrTiO 3 showed higher photocatalytic activity than did SrTiO 3 for the degradation of RhB and the photolysis of water to produce H 2 under the irradiation by visible, UV or solar light. Results of spectroscopic characterization revealed that after rapid ethanol quenching, a high concentration of oxygen vacancies was introduced on the surface of the Q-SrTiO 3 lattice. Consequently, in order to maintain the regional charge balance in Q-SrTiO 3 , the redundant Sr 2+ is likely to substitute for Ti 4+ . Moreover, oxygen vacancies play an important role in enhancing the photocatalytic performance of Q-SrTiO 3 by not only inducing localized states into the band gap of Q-SrTiO 3 , but also acting as photoinduced charge traps. Consequently, the light absorption ability is increased and the recombination rate of photogenerated electron-hole pairs is decreased, thus enhancing the photocatalytic activity of Q-SrTiO 3 .