Photocatalytic reduction of CO₂ with water using catalysts of γ-Ga₂O₃ supported by α-Ga₂O₃: mechanism and roles of each phase

Abstract

Gallium oxide (Ga₂O₃), consisting of the mixed phases of α and β, β and γ, and α and γ, is known as a photocatalyst for the reduction of CO₂ with water, producing CO, H₂ and O₂. In previous studies, we investigated Ga₂O₃ consisting of the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃, systematically varying the contents of γ-Ga₂O₃ as catalysts for the photoreduction of CO₂ with water, and proposed a crude reaction mechanism of the photocatalytic reduction of CO₂. However, the mechanism should be refined to clarify the roles of each phase and the effects of the morphology of the mixture. To do this, we have investigated the photocatalytic activity of γ-Ga₂O₃ supported by α-Ga₂O₃ instead of their mixed phases previously examined. With increasing contents of γ-Ga₂O₃, H₂ production rates monotonically decreased, whereas CO production rates increased, reached a maximum at 60–80% of the γ-Ga₂O₃ content, and decreased significantly. These trends are consistent with those observed in the previous studies using the mixed phases. Based on the previously suggested mechanism, we have proposed the detailed mechanism as follows: (1) the surfaces of α-Ga₂O₃ and γ-Ga₂O₃ particles are hydro-oxidated to GaOOH in water, and GaOOH on the γ-Ga₂O₃ particles absorbs CO₂ as bicarbonate; (2) GaOOH on α-Ga₂O₃ is photo-decomposed to α-Ga₂O₃ producing H; (3) the produced H migrates to the γ-Ga₂O₃ particles and reduces the adsorbed bicarbonate to CO; and (4) without UV photons, the surfaces of α-Ga₂O₃ and γ-Ga₂O₃ return to their initial states of GaOOH and bicarbonate-absorbing state, respectively.

---

Introduction

Ga₂O₃ is well known as a photocatalyst for CO₂ reduction with water. Among its several different crystalline phases, the α, β, or γ phases have been used in most previous studies. In order to increase the production rate and selectivity of CO among CO₂ reduction products, metallic cocatalysts such as Ag (ref. 1–5) and others^6–8 have often been employed. In our previous work, it has been shown that mixed phases of α and β,⁹ α and γ,¹⁰ and β and γ¹¹ exhibit high photocatalytic activity without the co-catalysts.

In previous work¹⁰ reporting a detailed investigation of the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃ as the photocatalyst of the CO₂ reduction with water, we have suggested the CO₂ reduction mechanism such that water splitting dominates on α-Ga₂O₃ and its H product reduces CO₂ adsorbed on γ-Ga₂O₃ to CO. However, in the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃, it was difficult to control the morphology of the mixture and the particle sizes of both phases.

In order to verify the mechanism, we have investigated the catalytic activity of γ-Ga₂O₃ supported by α-Ga₂O₃ for the photocatalytic CO₂ reduction with water instead of the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃ previously examined. The supported catalysts made it easier to control the contents and morphology of γ-Ga₂O₃ with clear separation of γ-Ga₂O₃ and α-Ga₂O₃. Comparing the present results with the previous results obtained using the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃, we confirm the previously suggested CO₂ reduction mechanism and also provide an improved version. Furthermore, the roles of each phase in the mixed phase samples of the previous work and the present supported samples have been clarified.

Experimental

Experimental procedures

γ-Ga₂O₃ supported by α-Ga₂O₃ photocatalysts (referred to as γ-Ga₂O₃/α-Ga₂O₃ hereafter) were synthesized by an impregnation method. Their γ-Ga₂O₃ contents were determined by XAFS analysis, and their morphology (geometrical structure) was observed by transmission electron microscopy (TEM). Specific surface area was measured by the BET method. CO₂ adsorption on γ-Ga₂O₃/α-Ga₂O₃ was also examined by temperature-programmed desorption (TPD). The photocatalytic CO₂ reduction with water under UV light illumination was carried out and analyzed in terms of the γ-Ga₂O₃ content, the morphology, and the specific surface area.

Catalysts preparation

At first, Ga(NO₃)₃·8H₂O (KISHIDA Chemical Corporation, purity 99%) was dissolved in distilled water. Then, α-Ga₂O₃ powders, prepared by the calcination of GaOOH at 450 °C for 4 h as reported by Li et al.,¹² were dispersed in the solution. The dispersed solution was dried up and calcined at 450 °C for 4 h in air, resulting in γ-Ga₂O₃/α-Ga₂O₃ samples. The contents of γ-Ga₂O₃ (the γ-Ga₂O₃ contents) in γ-Ga₂O₃/α-Ga₂O₃ (nominal contents) were controlled by varying the amounts of Ga(NO₃)₃·8H₂O and α-Ga₂O₃.

Photocatalytic CO₂ reduction with water under UV light irradiation

A sample (0.1 g) was dispersed in 100 mL of water with 0.5 M NaHCO₃ and stirred in a reaction cell made of quartz. CO₂ gas flowed into the reaction cell at 3 mL min⁻¹. UV light was provided by a Xe lamp through a UV cold mirror. The intensity of the light was 35 mW cm⁻² at 254 ± 10 nm. The produced gases (mostly H₂, CO, and O₂) were quantified by a gas chromatograph (TCD-GC, Shimadzu GC 8A). The reactions were monitored over five hours to confirm the steady-state production rates of H₂ and CO, which were determined every one hour.

Characterization

The morphology of the samples was observed by TEM, and their specific surface areas were determined by the BET method. Their crystalline structures were determined by X-ray diffraction (XRD) analysis. X-ray absorption analyses (XANES/EXAFS) were also employed and used to determine the γ-Ga₂O₃ contents of the samples.

TEM images were observed with JEM-1000K RS (JEOL Ltd.) under an acceleration voltage of 1000 kV at the High Voltage Electron Microscope Laboratory in Nagoya University. XRD patterns were recorded on Rigaku MiniFlex 600 (Cu Kα radiation, 40 kV, 15 mA) at room temperature. Ga K-edge XANES/EXAFS spectra were measured by the transmission method at room temperature at the 5S1 and 11S2 beam lines at Aichi Synchrotron Radiation Center. The ionization chambers were filled with 100% N₂ for incident X-rays (I₀) and 50% N₂ and 50% Ar for transmitted X-rays (I). Powder samples were set on a masking tape to be thick enough to measure XANES/EXAFS. Specific surface areas were determined by the BET method with N₂ adsorption at −196 °C using a Monosorb™ (Quantachrome). Samples were outgassed at 300 °C for 30 min under N₂ gas flow before measurement. FT-IR spectra were recorded with FT/IR-6100 (JASCO Co.) in the transmission mode at room temperature. Before the measurement, the sample was heated at 400 °C for 1 h. Temperature programmed desorption of CO₂ (CO₂-TPD) was carried out as follows. After drying the sample (50 mg) under He gas flow (50 mL min⁻¹) at 400 °C for 1 h, CO₂ was adsorbed under pure CO₂ gas flow at 40 °C for 1 h. Changing the CO₂ gas flow to a He gas flow (30 mL min⁻¹), CO₂ desorption profiles of the samples were measured, while increasing the temperature from 40 °C to 600 °C at a heating rate of 10 °C min⁻¹.

Results

TEM

Fig. 1 shows the TEM images and the electron diffraction patterns of α-Ga₂O₃ (a), γ-Ga₂O₃ (b), and γ-Ga₂O₃/α-Ga₂O₃ samples (c) and (d). The nominal γ-Ga₂O₃ contents of the last two were 30% and 60%, respectively. As seen in Fig. 1(a), α-Ga₂O₃ consisted of columnar-shaped particles with the length and width of about 1 μm and 0.5 μm, respectively, and each particle was fully crystallized as seen in clear diffraction spots. In contrast, as shown in Fig. 1(b), γ-Ga₂O₃ was composed of aggregates of nm-sized fine particles that were not well crystallized, giving halo rings without any clear spots of the γ-Ga₂O₃. γ-Ga₂O₃/α-Ga₂O₃ samples (Fig. (c) and (d)) clearly show that fine γ-Ga₂O₃ particles exhibiting a halo ring were deposited on the larger columnar-shaped α-Ga₂O₃ particles showing clear diffraction spots. With increasing γ-Ga₂O₃ content, the coverage of γ-Ga₂O₃ particles over the α-Ga₂O₃ particles increased, but the sizes of the γ-Ga₂O₃ particles hardly changed. In Fig. 1(d), the α-Ga₂O₃ particle was mostly covered by the γ-Ga₂O₃ particles.

Fig. 1 TEM images and electron diffraction patterns of (a) α-Ga₂O₃, (b) γ-Ga₂O₃, (c) γ-Ga₂O₃/α-Ga₂O₃ (γ = 30%), and (d) γ-Ga₂O₃/α-Ga₂O₃ (γ = 60%).

XRD

Fig. 2(a) and (b) show the XRD patterns of γ-Ga₂O₃/α-Ga₂O₃, α-Ga₂O₃, and γ-Ga₂O₃ before and after use for the CO₂ reduction, respectively. Most of the sharp peaks were attributed to α-Ga₂O₃, whereas those attributed to γ-Ga₂O₃ were broad, indicating that the α-Ga₂O₃ is well crystallized, while the γ-Ga₂O₃ is poorly crystallized. This corresponds well to the diffraction patterns appearing in Fig. 1. For γ-Ga₂O₃/α-Ga₂O₃, the sharp peaks of α-Ga₂O₃ and the broad peaks of γ-Ga₂O₃ overlapped. Thus, γ-Ga₂O₃/α-Ga₂O₃—consisting of poorly crystallized fine γ-Ga₂O₃ particles supported by larger, fully crystallized α-Ga₂O₃ particles—was synthesized.

Fig. 2 XRD patterns of γ-Ga₂O₃/α-Ga₂O₃, α-Ga₂O₃, and γ-Ga₂O₃ before (a) and after (b) the reaction. The inset is an enlarged view of the lower-angle side of the α-Ga₂O₃ spectrum to show the appearance of GaOOH.

As indicated in Fig. 2(b), new peaks appearing in the XRD patterns of the samples after use were attributed to GaOOH. This suggests that the surfaces of Ga₂O₃ particles after use were covered by GaOOH. Such hydro-oxidation of the surface of Ga₂O₃ used as a photocatalyst for CO₂ reduction was reported in previous studies.^9,10

XAFS

Fig. 3(a) presents the Ga K-edge XANES spectra of α-Ga₂O₃, γ-Ga₂O₃, and γ-Ga₂O₃/α-Ga₂O₃. The differences in XANES fine structures among them were appreciable in the energy range of 10 380–10 480 eV, as shown in the enlarged inset. This allowed us to determine the γ-Ga₂O₃ content of the samples with the linear combination fitting of the XANES spectra as described below. Fig. 3(b) shows an example of the fitting. The fitted spectrum reproduces the experimental spectrum well, giving the compositions of α-Ga₂O₃ and γ-Ga₂O₃ phases of 58% and 42%, respectively.

Fig. 3 (a) Ga K-edge XANES spectra of α-Ga₂O₃, γ-Ga₂O₃, and γ-Ga₂O₃/α-Ga₂O₃ samples. (b) Least squares fitting to the observed XANES spectra with a linear combination of the spectra of α-Ga₂O₃ (58%) and γ-Ga₂O₃ (42%).

Fig. 4(a) presents Ga K-edge EXAFS spectra of α-Ga₂O₃, γ-Ga₂O₃, and γ-Ga₂O₃/α-Ga₂O₃, showing clear differences in amplitude and periodicity among the three. This made us apply linear combination fitting used in XANES analysis to determine the γ-Ga₂O₃ contents. Fig. 4(b) shows an example of the fitting with the compositions of α-Ga₂O₃ and γ-Ga₂O₃ phases of 58% and 42%, respectively.

Fig. 4 (a) Ga K-edge EXAFS spectra of α-Ga₂O₃, γ-Ga₂O₃, γ-Ga₂O₃/α-Ga₂O₃ samples. (b) Least squares fitting to the observed EXAFS spectra with a linear combination of the spectra of α-Ga₂O₃ (58%) and γ-Ga₂O₃ (42%).

In Table 1, thus determined γ-Ga₂O₃ contents by the XANES and EXAFS analyses are compared with the nominal γ-Ga₂O₃ contents calculated by the mixing ratio of reagents for the synthesis. The determined values agreed within a difference of 3% for all samples.

Table 1 Comparison of the γ-Ga₂O₃ contents determined by least squares fitting using XANES and EXAFS spectra and the nominal states determined for all γ-Ga₂O₃/α-Ga₂O₃ samples

Amount of charge γ (%)		10	20	30	40	50	53	56	60	70	80	83	86	90
Measured γ	XANES (%)	6	10	32	38	42	47	55	58	69	73	77	84	91
	EXAFS (%)	6	10	31	35	43	47	52	55	68	75	79	83	93

---

BET specific surface area

Fig. 5 shows the specific surface areas (SSAs) of all samples determined by the BET method as a function of γ-Ga₂O₃ content determined by XANES analysis. γ-Ga₂O₃ exhibited a much larger SSA than that of α-Ga₂O₃. This is quite reasonable considering the poor crystallinity of γ-Ga₂O₃. The SSA of γ-Ga₂O₃/α-Ga₂O₃ increased with an S-shaped curve showing a slower SSA increase at lower γ-Ga₂O₃ content, a roughly linear increase in the middle range, and saturation over 70% of γ-Ga₂O₃ content. This is different from the linear increase in SSA with the γ-Ga₂O₃ contents for the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃ observed in the previous work.¹⁰

Fig. 5 Changes in BET specific surface area with γ-Ga₂O₃ content (dashed lines serve as a guide to the eyes).

This S-shaped SSA increase corresponds well to the TEM observation. At the low γ-Ga₂O₃ contents, nano-sized γ-Ga₂O₃ particles were deposited discretely without appreciable increase in SSA. In the middle range of γ-Ga₂O₃ content, both SSA and the coverage of the γ-Ga₂O₃ particles, which have a much larger SSA than that of the α-Ga₂O₃ particles, increased linearly until the surface of the α-Ga₂O₃ particles was mostly covered over 70% of the γ-Ga₂O₃ content.

Photocatalytic reduction of CO₂ with water

The reaction products were mostly H₂, CO, and O₂ for all samples. In Fig. 6, the production rates of H₂ and CO against the γ-Ga₂O₃ contents of the samples are plotted. The H₂ production rate was highest for α-Ga₂O₃ and decreased monotonously with increasing γ-Ga₂O₃ content. This is quite consistent with our previous work using the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃ and confirms that the H₂ production is dominated on α-Ga₂O₃. The CO production rates stayed small for lower γ-Ga₂O₃ contents, reached a maximum for samples containing 60–80% of γ-Ga₂O₃, and then decreased markedly for higher γ-Ga₂O₃ contents. This indicates that CO production is promoted by γ-Ga₂O₃, and that the existence of the α-Ga₂O₃ is necessary; that is, γ-Ga₂O₃ alone (without α-Ga₂O₃) showed little activity for the photocatalytic CO₂ reduction with water.

Fig. 6 Production rates of H₂ and CO plotted against the γ-Ga₂O₃ content of γ-Ga₂O₃/α-Ga₂O₃ (dashed lines serve as a guide to the eyes).

CO₂-TPD

Fig. 7 shows the CO₂-TPD profiles of α-Ga₂O₃, γ-Ga₂O₃, and γ-Ga₂O₃/α-Ga₂O₃ (γ = 77%) in the temperature range from 200 °C to 600 °C. Although a desorption peak caused by the adsorbed water appeared under 200 °C, it is not shown in the figure. The adsorbed amounts of CO₂ on γ-Ga₂O₃ and γ-Ga₂O₃/α-Ga₂O₃ were similar and significantly larger than that on α-Ga₂O₃. This indicates that CO₂ adsorption on γ-Ga₂O₃/α-Ga₂O₃ mostly originated from γ-Ga₂O₃.

Fig. 7 CO₂-TPD profiles for α-Ga₂O₃, γ-Ga₂O₃ and γ-Ga₂O₃/α-Ga₂O₃ (γ = 77%) samples.

Considering that CO₂ adsorbed on Ga₂O₃ is known to take mainly two species of carbonate and bicarbonate,¹³ the two dominant peaks appeared at around 200–300 °C and 400–500 °C could be attributed to the former and the latter, respectively. This agrees with previous reports showing that, as the precursor of CO forms, the bicarbonate desorbing at higher temperature is more favorable than the carbonate.^13–15 The bicarbonate should be formed through the interaction of CO₂ with OH species on the Ga₂O₃ surface. As depicted in Fig. 2(b), the surface of γ-Ga₂O₃ was converted to GaOOH, which very likely enhanced CO₂ adsorption as the bicarbonate (see (1)).

Discussion

Reaction mechanism

Here, we discuss the mechanism of the photocatalytic CO₂ reduction with water on γ-Ga₂O₃/α-Ga₂O₃. As seen in Fig. 6, H₂ production was dominated on α-Ga₂O₃, while CO production rates increased with the γ-Ga₂O₃ contents and reached a maximum for the samples containing 60–80% of the γ-Ga₂O₃ contents. This observation is quite consistent with the previous work using the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃.¹⁰ Considering the mechanism suggested in the previous work,¹⁰ we have claimed a slightly more detailed mechanism as follows.

I. The surfaces of the α-Ga₂O₃ and γ-Ga₂O₃ particles are hydroxylated to GaOOH in water,

Ga₂O₃ + H₂O → 2GaOOH (2)

and GaOOH on γ-Ga₂O₃ particles absorbs CO₂ a bicarbonate.

II. GaOOH on α-Ga₂O₃ is photo-decomposed to α-Ga₂O₃, producing H, as reported by Aoki et al.⁹

2GaOOH + hν → α-Ga₂O₃ + 2H + O (3)

III. The produced H migrates to the γ-Ga₂O₃ particles and reduces the absorbed bicarbonate, resulting in CO. (It is not clear whether this process is photo-assisted or not).

IV. Without photons, the surfaces of α-Ga₂O₃ and γ-Ga₂O₃ return to their initial states of GaOOH and the bicarbonate-absorbing state, respectively.

This mechanism is schematically illustrated in Fig. 8.

Fig. 8 Schematic drawing of the mechanism of photocatalytic CO₂ reduction with water on γ-Ga₂O₃/α-Ga₂O₃.

Effect of sample morphology on the reaction rate

Although the present results are quite similar to those observed in the previous work using the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃, the present production rates of H₂ and CO were clearly smaller. As seen in Fig. 5, the SSA shows an S-shaped increase with γ-Ga₂O₃ content, while a linear increase was observed in previous work. This seems reasonable considering the morphology difference, whereby the γ-Ga₂O₃ particles cover the α-Ga₂O₃ particles in the present work, while α-Ga₂O₃ and γ-Ga₂O₃ particles were mixed in the previous work. The CO production rates of both studies were hardly correlated with the SSA. This also supports the reaction mechanism described above.

In the present work, the γ-Ga₂O₃ content giving the maximum production rate of CO was around 70%, which was clearly higher than 40% from the previous work, while 0.8 mmol h⁻¹ g⁻¹ of the former was clearly less than 3 mmol h⁻¹ g⁻¹ of the latter. This difference could be attributed to the difference in morphology between the supported γ-Ga₂O₃/α-Ga₂O₃ and the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃. Since in the former, the γ-Ga₂O₃ particles cover the α-Ga₂O₃ columnar particles, the surface area of the α particles was comparatively less than that of the latter, resulting in less H production and, consequently, less CO production. In addition, in lower γ-Ga₂O₃ content samples of the former, the γ-Ga₂O₃ particles covered the α-Ga₂O₃ columnar particles rather discretely. Hence, H produced on the α-Ga₂O₃ particles should migrate a longer distance compared with the mixed-phase samples, resulting in less CO production.

There is another geometrical factor. In the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃ samples, both phases are directly exposed to the UV light, whereas in the supported γ-Ga₂O₃/α-Ga₂O₃ photocatalyst, the α-Ga₂O₃ particles are partially or fully covered by the γ-Ga₂O₃ particles. In addition, owing to the smaller band gap of the γ-Ga₂O₃, the γ-Ga₂O₃ particles covering the α-Ga₂O₃ particles would shield the UV light from the α-Ga₂O₃ particle beneath,^16–18 and thus reduce the H formation on the α-Ga₂O₃ particles and consequently lead to lower CO production. These morphological effects further support the proposed reaction mechanism and emphasize the importance of phase arrangement and morphology in determining photocatalytic performance.

Conclusions

In this work, γ-Ga₂O₃ supported by α-Ga₂O₃ (γ-Ga₂O₃/α-Ga₂O₃) photocatalysts with different γ-Ga₂O₃ contents were synthesized by impregnation followed by calcination. TEM and XRD analysis showed that nano-sized γ-Ga₂O₃ particles were deposited on the surface of columnar-shaped α-Ga₂O₃ particles. The γ-Ga₂O₃ content was successfully determined by Ga K-edge XAFS analysis.

Photocatalytic CO₂ reduction with water was carried out to investigate the change in H₂ and CO production rates with γ-Ga₂O₃ content. The H₂ production rate decreased with the γ-Ga₂O₃ content, whereas the CO production rate reached a maximum at 60–80% of γ-Ga₂O₃ content. These results indicate that H₂ production is dominated on α-Ga₂O₃, while CO production is promoted on γ-Ga₂O₃, which absorbs a much larger amount of CO₂ in the form of bicarbonate compared to α-Ga₂O₃.

The changes in the production rates of CO and H₂ with γ-Ga₂O₃ content are consistent with previous work using the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃ as photocatalysts. Based on the previously suggested mechanism, a slightly more detailed mechanism is given as follows: (1) the surface of α-Ga₂O₃ and γ-Ga₂O₃ particles are hydro-oxidated to GaOOH in water, and GaOOH on the γ-Ga₂O₃ particles absorbs CO₂ as bicarbonate; (2) GaOOH on α-Ga₂O₃ is photo-decomposed to α-Ga₂O₃, producing H; (3) the produced H migrates to the γ-Ga₂O₃ particles and reduces the adsorbed bicarbonate to CO. Without UV photons, the surfaces of α-Ga₂O₃ and γ-Ga₂O₃ return to their initial states of GaOOH and bicarbonate-absorbing state, respectively; and (4) without UV photons, the surfaces of α-Ga₂O₃ and γ-Ga₂O₃ return to GaOOH and bicarbonate-absorbing states, respectively. Still, the detailed pathways of CO production are unclear and need to be further researched.

Compared with the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃ photocatalyst reported in the previous study, both H₂ and CO production rates of γ-Ga₂O₃/α-Ga₂O₃ in the present study were lower. This difference is successfully attributed to the difference in the morphology, i.e., g-Ga₂O₃ particles covered the α-Ga₂O₃ particle in the latter, while the mixed phases of α-Ga₂O₃ and γ-Ga₂O₃ particles in the former. These results demonstrate that not only phase composition but also the spatial arrangement of α and γ-Ga₂O₃ plays a crucial role in controlling the CO₂ reduction activity of Ga₂O₃-based photocatalysts.

Conflicts of interest

The authors declare no conflicts of interest.

Data availability

Raw data were generated at Nagoya University. Derived data supporting the findings of this study are available from Tomoko Yoshida on request.

Acknowledgements

This work was supported by JST, CREST Grant Number JP24031877 and JSPS KAKENHI Grant Number JP20KK0116, Japan.

Notes and references

1. M. Yamamoto, T. Yoshida, N. Yamamoto, T. Nomoto, Y. Yamamoto, S. Yagi and H. Yoshida, J. Mater. Chem. A, 2015, 3, 16810.
2. N. Yamamoto, T. Yoshida, S. Yagi, Z. Like, T. Mizutani, S. Ogawa, H. Namiki and H. Yoshida, e-J. Surf. Sci. Nanotechnol., 2014, 12, 263.
3. Y. Kawaguchi, M. Akatsuka, M. Yamamoto, K. Yoshioka, A. Ozawa, Y. Kato and T. Yoshida, J. Photochem. Photobiol., A, 2018, 358, 459.
4. K. Yoshioka, M. Yamamoto, T. Tanabe and T. Yoshida, e-J. Surf. Sci. Nanotechnol., 2020, 18, 168.
5. M. Yamamoto, S. Yagi and T. Yoshida, Catal. Today, 2018, 303, 334.
6. Y. Pan, Z. Sun, H. Cong, Y. Men, S. Xin, J. Song and S. Yu, Nano Res., 2016, 9, 1689.
7. S. Kikkawa, K. Teramura, H. Asakura, S. Hosokawa and T. Tanaka, J. Phys. Chem. C, 2018, 122, 21132.
8. H. Yoon, J. Yang, S. Park, C. Rhee and Y. Sohn, Appl. Surf. Sci., 2021, 536, 147753.
9. T. Aoki, M. Yamamoto, T. Tanabe and T. Yoshida, New J. Chem., 2022, 46, 3207.
10. N. Ota, Y. Takashiro, M. Yamamoto, T. Tanabe and T. Yoshida, J. Mater. Chem. A, 2025, 13, 6663.
11. M. Akatsuka, Y. Kawaguchi, R. Itoh, A. Ozawa, M. Yamamoto, T. Tanabe and T. Yoshida, Appl. Catal., B, 2020, 262, 118247.
12. L. Li, W. Wei and M. Behrens, Solid State Sci., 2012, 14, 971.
13. B. Zhao, Y. Pan and C. Liu, Catal. Today, 2012, 194, 60.
14. K. Teramura, K. Hori, Y. Terao, Z. Huang, S. Iguchi, Z. Wang, H. Asakura, S. Hosokawa and T. Tanaka, J. Phys. Chem. C, 2017, 121, 8711.
15. Y. Wang, Y. Sun, X. Liu and F. Dong, PNAS Nexus, 2024, 3, 339.
16. J. Lyons, ECS J. Solid State Sci. Technol., 2019, 8, 3226.
17. T. Oshima, T. Nakazono, A. Mukai and A. Ohtomo, J. Cryst. Growth, 2012, 359, 60.
18. D. Shinohara and S. Fujita, Jpn. J. Appl. Phys., 2008, 47, 7311.

---