Activation of Water‐Splitting Photocatalysts by Loading with Ultrafine Rh–Cr Mixed‐Oxide Cocatalyst Nanoparticles

Abstract The activity of many water‐splitting photocatalysts could be improved by the use of RhIII–CrIII mixed oxide (Rh2−xCrxO3) particles as cocatalysts. Although further improvement of water‐splitting activity could be achieved if the size of the Rh2−xCrxO3 particles was decreased further, it is difficult to load ultrafine (<2 nm) Rh2−xCrxO3 particles onto a photocatalyst by using conventional loading methods. In this study, a new loading method was successfully established and was used to load Rh2−xCrxO3 particles with a size of approximately 1.3 nm and a narrow size distribution onto a BaLa4Ti4O15 photocatalyst. The obtained photocatalyst exhibited an apparent quantum yield of 16 %, which is the highest achieved for BaLa4Ti4O15 to date. Thus, the developed loading technique of Rh2−xCrxO3 particles is extremely effective at improving the activity of the water‐splitting photocatalyst BaLa4Ti4O15. This method is expected to be extended to other advanced water‐splitting photocatalysts to achieve higher quantum yields.


S2
1(b)). 6 In this process, BaLa4Ti4O15 photocatalyst (650 mg) was added to an aqueous K2CrO4 solution (350 mL) in a quartz cell. The mixing ratio of K2CrO4 to BaLa4Ti4O15 was changed within the range of 0.05−0.20 wt% Cr ( Figure S18). After removing dissolved air by Ar bubbling, the suspension was irradiated with a highpressure Hg lamp (400 W) under an Ar flow of 30 mL/min at 25 °C for 1.5 h. The actual amounts of Cr loaded on BaLa4Ti4O15 were determined by inductively coupled plasma mass spectrometry (ICP-MS) of the aqueous solution after mixing. The solid Cr2O3/BaLa4Ti4O15 was collected by centrifugation, washed with water three times, and dried by evaporation.
The Rh−SG complex was adsorbed on Cr2O3/BaLa4Ti4O15 by mixing an aqueous solution of the Rh−SG complex with an aqueous solution of BaLa4Ti4O15 (600 mg) for 2 h at room temperature ( Figure 1c). The total volume of aqueous solution was fixed at 200 mL. The actual amount of Rh adsorbed on BaLa4Ti4O15 was determined by ICP-MS of the aqueous solution after mixing (Table S2).
The obtained Rh−SG/Cr2O3/BaLa4Ti4O15 (550 mg) was calcined under reduced pressure (>1.0×10 −1 Pa). The furnace temperature was increased at a rate of 7 °C/min and then maintained at 300 °C for 80 min to give  Rh2−xCrxO3/BaLa4Ti4O15 (impregnation method). Rh2−xCrxO3/BaLa4Ti4O15 was also prepared by impregnation method. 7 First, BaLa4Ti4O15 (600 mg) and water (4−5 mL) containing an appropriate amount of RhCl3 and Cr(NO3)3·9H2O were transferred into an evaporation dish that was subsequently placed in a steam bath. The suspension was stirred using a glass rod until the water was completely evaporated, and the resulting powder was collected and heated in air. The furnace temperature was increased at a rate of 7 °C/min and then maintained at 350 °C for 1 h.
Rh2O3/BaLa4Ti4O15. Rh2O3/BaLa4Ti4O15 was prepared by impregnation method. 7 First, BaLa4Ti4O15 (600 mg) and water (4−5 mL) containing an appropriate amount of RhCl3 were transferred into an evaporation dish that was subsequently placed in a steam bath. The suspension was stirred using a glass rod until the water was completely evaporated, and the resulting powder was collected and heated in air. The furnace temperature was increased at a rate of 7 °C/min and then maintained at 350 °C for 1 h.

RhNP/BaLa4Ti4O15
. RhNP was loaded on BaLa4Ti4O15 by photodeposition method. 8 First, BaLa4Ti4O15 photocatalyst (600 mg) was added to an aqueous RhCl3 solution (350 mL) in a quartz cell. After removing dissolved air by evacuation through Ar bubbling, the suspension was irradiated with a high-pressure Hg lamp S3 (400 W) under Ar flow of 30 mL/min at 25 ℃ for 1.5 h. The solid RhNP/BaLa4Ti4O15 was collected by centrifugation and then washed with water 3 times and subsequently dried by evaporation.
After removing dissolved air by evacuation through Ar bubbling, the suspension was irradiated with a highpressure Hg lamp (400 W) under Ar flow of 30 mL/min at 25 ℃ for 1.5 h. The solid RhNP@Cr2O3/BaLa4Ti4O15 was collected by centrifugation and then washed with water 3 times and subsequently dried by evaporation.

Characterization
The ultraviolet-visible absorption spectra were collected by a spectrometer (JASCO, V-670 or Shimadzu, UV 3600). The wavelength-dependent optical data (I(w)) were converted to energy-dependent data (I(E)) by the following equation, which conserved the integrated spectral areas: The Fourier transform infrared spectrometer (FT-IR) attenuated total reflection (ATR) spectra were recorded in the region between 500 and 5000 cm −1 using a spectrometer (JASCO FT/IR-4600-ATR-PRO ONE) equipped with a DLATGS detector as the average of 50 scans at 4 cm −1 resolution.
Rh K-and Cr K-edge X-ray absorption fine structure (XAFS) measurements were performed at beamlines BL01B1 and BL37XU at the SPring-8 facility of the Japan Synchrotron Radiation Research Institute (proposal number 2018A0910, 2018A0919, and 2018B1422). The incident X-ray beam was monochromatized by Si (111), for Cr K edge, and Si(311), for Rh K edge, double-crystal monochromator, respectively. XAFS spectra of Rh0.5Cr1.5O3, CrO3, Cr2O3 and Cr foil (Cr K edge), and Rh0.5Cr1.5O3, Rh2O3 and Rh foil (Cr K edge) as references were recorded in transmission mode using ionization chambers. Rh K-edge and Cr K-edge XAFS spectra for photocatalyst samples were measured in fluorescence mode using a 19-element Ge solid-state detector at room temperature. The X-ray energy was calibrated for the Rh K-and Cr K-edges using Rh foil and Cr2O3, respectively. The X-ray absorption near-edge structure (XANES) and extended X-ray absorption The X-ray photoelectron spectroscopy (XPS) data were collected by using an electron spectrometer (JEOL, JPS-9010MC) at a base pressure of ∼2 × 10 −8 Torr. X-rays from the Mg-Kα line (1253.6 eV) were used for excitation. The spectra were calibrated with the peak energies of Au 4f 7/2 (83.7 eV).

S4
ICP-MS was performed with an Agilent 7500c spectrometer (Agilent Technologies, Tokyo, Japan). Bi was used as the internal standard.
The high-angle annular dark field scanning transmission electron microscope (HAADF-STEM) images were obtained by ultra-high-resolution transmission electron microscope (The FEI Titan Themis 80−200) operating at 200 kV, with a beam convergence semi angle of 25 mrad and HAADF collection angle from 56−200 mrad. Elemental maps were acquired using a super X detector and low-background sample holder.
The transmission electron microscope (TEM) images were recorded with a JEM-2100 electron microscope (JEOL) operating at 200 kV, typically using magnification of 400,000−600,000.
The X-ray diffraction (XRD) patterns of the samples were measured with a Rint2500 diffractometer (Rigaku) using Cu-Kα source operated at 50 kV and 100 mA. A reflection-free silicon plate was used as a substrate.

Measurement of photocatalytic activity
Water-splitting. The photocatalytic water-splitting reaction was performed at room temperature using an experimental apparatus built in-house consisting of a high-pressure Hg lamp (400 W) and quartz cell. 6 The reaction was performed with an Ar gas flow rate of 30 mL/min. Before the measurements, the reaction solution containing the prepared photocatalyst (500 mg) in mili-Q water (350 mL) was purged with Ar gas for 1 h to ensure complete removal of air from the reaction vessel. The apparent quantum yield at 270 nm was estimated using a top-irradiation cell made of quartz and 300-W Xe lamp (Asahi Spectra; MAX-302) equipped with band-pass filter. The number of incident photons was determined using a photodiode (Ophir: PD300-UV head and NOVA power monitor). The photocatalyst powder (100 mg) was dispersed in pure water (300 mL) to measure the apparent quantum yield.
O2-photoreduction reaction. The decrease of the amount of evolved H2 was examined to investigate the likelihood of the O2-photoreduction reaction. In particular, the photocatalytic water-splitting reaction was investigated under a gas flow consisting of a 7:3 mixture of Ar to air instead of pure Ar. Under these experimental conditions, sufficient O2 is included in the reaction system for the O2-photoreduction reaction to occur.
Back reaction. The back reaction between H2 and O2 to produce H2O in the gas phase in the dark was examined using a gas-tight circulation system with a dead volume of 400-500 mL. In this experiment, 100 mg of photocatalyst was used.

Estimation of the constitute metal atoms included in Rh2−xCrxO3 particles
Assuming that Rh2−xCrxO3 particles are supported hemispherically on the photocatalyst as M2O3 (M = Rh or Cr), the number of total metal atoms in the particle can be obtained from equations 1 and 2.

S5
Here, n, ρ, NA, MW, and D indicate the number of total metal atoms, density of M2O3, Avogadro constant, molecular weight of M2O3, and particle diameter, respectively. The density of M2O3 and the molecular weight of M2O3 were changed in the range of Rh:Cr = 1:4−4:1 and thereby the number of Rh atoms in each Rh2−xCrxO3 particle was estimated ( Figure S15).  Figure 3(d).

Additional Figures
Scheme S1. Potential future energy conversion system consisting of photocatalysts and fuel cell to produce electrical energy without consuming fossil fuels and producing carbon dioxide.    (Figure 1(c)). The surface without the adsorption of Rh-SG complex is selected in this figure, in contrast to Figure 4B(a) and 4C(a). As can be seen, the Cr2O3 layer is mainly loaded at the edge of BaLa4Ti4O15 by the photodeposition method, because the reduction of the metal ions occurs easier at the edge of BaLa4Ti4O15 than at the flat surface of BaLa4Ti4O15. 2 This means that the Rh2−xCrxO3 particles could also be formed preferentially at the edge of BaLa4Ti4O15 ( Figure S25(a)), which is the different from the case of the loading of Rh2−xCrxO3 particles by impregnation methods (Figure S25(b)). 7 Figure S6. FT-IR spectrum of Rh−SG complex together with that of GSH. In figure, "ν", "νas", "νss", and "σ" indicate the stretching vibration, asymmetric stretching vibration, symmetric stretching vibration, and in plane bending vibration, respectively. 11−13 Figure S5. HR-TEM image of Cr2O3/BaLa4Ti4O15 with 0.1 wt% Cr observed at high magnification for the edge of BaLa4Ti4O15. Figure S7. Assignments of ESI mass spectrum of Rh−SG complex. As shown in Figure 3(d), the main peak group (i and ii) is composed of Rh2(SG)2 containing Rh(II) (i) or Rh(III) (ii). The sharp peaks (# and ##) are due to the fragments and adducts of Rh2(SG)2, respectively. The asterisk peaks (*) are due to the species which do not include Rh, (GSH)2. The assignments of the peaks *, #, and ## are summarized in Table S1.  Figure 6C(b)), the Rh−SG complex is estimated to be conglomerated on the surface (Figure 4B(a) and 4C(a)(c)), forming Rh−SG aggregates, [Rh2(SG)2]n ( Figure S15).  . Relation between calcination temperature and the water-splitting activity obtained for Rh2−xCrxO3/BaLa4Ti4O15 photocatalysts including 0.09 wt% Rh and 0.10 wt% Cr ( Figure S10). These results indicate that the apparent aggregation of the particles ( Figure S10(c)) leads to the decrease of watersplitting activity ((c)). The relation between the particles size and the water-splitting activity is consistent with that obtained from the comparison between Rh2−xCrxO3(1.3 nm)/BaLa4Ti4O15 and Rh2−xCrxO3(3.0 nm)/BaLa4Ti4O15 (Figure 9(a) vs. 9(d)).
Figure S12. Comparison of Rh K-edge FT-EXAFS spectra between Rh2−xCrxOy/BaLa4Ti4O15 (Figure 1(d)) and Rh2−xCrxO3/BaLa4Ti4O15 (Figure 1(e)) together with those of Rh foil, Rh2O3, and Rh0.5Cr1.5O3. Coordination numbers of Rh−O bonds estimated for Rh2−xCrxOy/BaLa4Ti4O15, Rh2−xCrxO3/BaLa4Ti4O15, Rh2O3, and Rh0.5Cr1.5O3 are listed in Table S3. S11 Figure S13. Comparison of Cr K-edge XANES spectra between Rh2−xCrxOy/BaLa4Ti4O15 (Figure 1(d)), Rh2−xCrxO3/BaLa4Ti4O15 (Figure 1(e)), CrO3, and Cr2O3. In the Cr K-edge XANES spectra of Rh2−xCrxOy/BaLa4Ti4O15, a weak peak exists at ~5992 eV similar to that of CrO3. This peak is attributed to tetrahedral chromium oxide species, which is characteristic of chromium oxides with a high degree of oxidation, indicating that some of the chromium oxide was further oxidized during calcination. This peak disappeared after the light irradiation, indicating that chromium oxides with a high degree of oxidation was reduced by the light irradiation to Cr(III). Figure S14. Comparison of XRD patterns for (a) Cr2O3/BaLa4Ti4O15 (Figure 1(b)), (b) Rh-SG/Cr2O3/BaLa4Ti4O15 (Figure 1(c)), (c) Rh2−xCrxOy/BaLa4Ti4O15 (Figure 1(d)), and (d) Rh2−xCrxO3/BaLa4Ti4O15 (Figure 1(e)) together with that of (e) BaLa4Ti4O15 for comparison. The PDF number of BaLa4Ti4O15 is 01-073-7800. Figure S15. Tentative counting of the number of constitute Rh atoms in one Rh2−xCrxO3 particle of Rh2−xCrxO3/BaLa4Ti4O15 (Figure 1(e)). It is estimated that Rh2−xCrxO3 particle of 1.3 nm includes around 24 metal atoms on the basis of the densities of bulk Rh2O3 and Cr2O3 and the atomic weights of Rh, Cr, and O (see section 1.6 in the supporting information). If each Rh2−xCrxO3 particle would contain the same number of Rh and Cr, the number of Rh atoms in each Rh2−xCrxO3 particle of 1.3 nm is estimated to be around twelve. This means around six Rh2(SG)2 gathered in one Rh−SG aggregate of Rh-SG/Cr2O3/BaLa4Ti4O15 (Figure 1(c)). However, at present, the exact chemical composition of Rh2−xCrxO3 particle was not elucidated.     Figure S20 and S21. Figure S20. Measurement of the back reaction between H2 and O2 to produce H2O ( Figure S19(a)) in the gas phase in the dark over Rh2−xCrxO3(1.3 nm)/BaLa4Ti4O15 including 0.09 wt% Rh and 0.10 wt% Cr. This experiment was conducted in the gas phase in the dark using a gas-tight circulation system with a dead volume of 400-500 mL. Both H2 and O2 were hardly reduced for 5 h, indicating that the back reaction ( Figure S19(a)) is relatively well suppressed on Rh2−xCrxO3(1.3 nm)/BaLa4Ti4O15 including 0.09 wt% Rh and 0.10 wt% Cr. S14 Figure S21.    In (a), Rh is not necessarily used for the formation of Rh2−xCrxO3 but also used for the formation of Rh2O3 at the flat surface of BaLa4Ti4O15. Since the Rh2O3 particles without including Cr cannot suppress the O2photoreduction reaction, those particles have low activity for the water-splitting reaction ( Figure S24(a)) and thereby the use of Rh for the formation of those particles leads to the decrease of the activity. In (a), there are also the excess loading of Cr2O3 layer, which might cause the decrease of absorption of light and thereby the decrease of the activity. On the basis of these results, it can be considered that Rh2−xCrxO3(1.3 nm)/BaLa4Ti4O15 showed higher water-splitting activity than Rh2−xCrxO3(3.0 nm)/BaLa4Ti4O15, because positive effects caused by miniaturization of Rh2−xCrxO3 cocatalysts ( Figure 6C(b)) and selective formation of Rh2−xCrxO3 cocatalysts at the edge of BaLa4Ti4O15 (Figure 2, S3, S4, and S25(a)) overcomes the negative effects caused by the formation of Rh2O3 particles without including Cr and the excess loading of Cr2O3 layer on the BaLa4Ti4O15 (Figure S25(a)).