Phase Transition of Two-Dimensional β-Ga2O3 Nanosheets from Ultrathin γ-Ga2O3 Nanosheets and Their Photocatalytic Hydrogen Evolution Activities.

Monoclinic β-Ga2O3 nanosheets hold great potential applications in electronic, optical, and photocatalytic fields. In this study, two-dimensional β-Ga2O3 nanosheets were successfully fabricated through a simple crystalline phase transition from the as-prepared ultrathin γ-Ga2O3 nanosheets. The photocatalytic hydrogen evolution reaction under UV light irradiation was achieved on the two kinds of photocatalysts. However, β-Ga2O3 with a higher crystallinity shows a lower photocatalytic activity in comparison with γ-Ga2O3. The average apparent quantum yield is calculated to be 0.29% for β-Ga2O3 nanosheets and 1.82% for γ-Ga2O3. More efficient separation and transfer rates of photogenerated carriers and larger specific areas were found in γ-Ga2O3. On the basis of the analysis of the structures of γ-Ga2O3 and β-Ga2O3, it is proposed that the disordered or defective structure contributes to the improvement of photocatalytic activity to some extent. Therefore, it is significant to develop the photocatalyst with a stable structure and a certain number of defects at the same time.

Introduction

Photocatalytic hydrogen evolution is an attractive scientific and technological goal to address the increasing global demand for clean energy and to reduce the reliance on fossil fuels and carbon dioxide emission.[1−3] Tremendous efforts have been devoted to develop a more efficient photocatalyst for hydrogen evolution during the past several decades.[4−6] As a result, various kinds of photocatalysts have been developed. Among all of these photocatalyst materials, gallium oxide (Ga2O3) has attracted intensive attention because of its high conductivity and excellent photoluminescence (PL) property resulting from the defect-rich structure.[7] Moreover, Ga2O3 with a wide band gap (∼4.7 eV) has a strong redox ability of the photogenerated holes and electrons. Thus, Ga2O3 has displayed a good photocatalytic activity for the organic pollutant decomposition.[8] However, the photocatalytic activity of Ga2O3 toward the photocatalytic H2 evolution was still very low, which was mainly due to the insufficient active sites and high recombination rate of photogenerated carriers. To enhance the photocatalytic activity, one of the popular strategies is to develop Ga2O3 photocatalysts with excellent morphology structure that can provide more efficient separation and transportation rates of the photogenerated carriers and more active sites for photocatalytic hydrogen evolution.

The ultrathin two-dimensional (2D) nanomaterials, made of several atomic layers, have received extensive interests in recent years[9−11] because of their unique and fascinating properties distinct from their bulk counterparts, including high exposed proportion of surface atoms or surface-active sites and a minimum migration distance of electrons and holes to achieve rapid carrier transport,[12] which can contribute to the high photocatalytic activity. Large numbers of nanosheet materials with single- or multilayer thickness, such as transition-metal dichalcogenides, transition-metal oxides, and layered-double hydroxides, have been prepared via the exfoliation of their bulk layered counterparts[13,14] or by the wet-chemical synthesis method.[15,16] As an important semiconductor, β-Ga2O3 has a monoclinic structure and is the most stable phase among various crystal phases of Ga2O3 and has been applied in a variety of areas, including optics,[17] electronics,[18] and particularly photocatalysis.[19] However, to the best of our knowledge, β-Ga2O3 with a nanosheet morphology has been scarcely studied because of its nonlayered structure. Our previous work has successfully developed the freestanding single-layer nanosheets of nonlayered material γ-Ga2O3 with unprecedented thickness (∼1 nm) by a facile hydrothermal method without using any kind of shape control agents.[20] However, the cubic γ-Ga2O3 is susceptible to the environment and difficult to be well preserved. Therefore, the success in γ-Ga2O3 nanosheets inspires us to prepare β-Ga2O3 nanosheets—the most stable among the five phases of Ga2O3—through a crystalline phase transition of metastable γ-Ga2O3 nanosheets.

Herein, we report a template approach for the synthesis of β-Ga2O3 nanosheets through one-step thermal treatment of ultrathin 2D γ-Ga2O3 nanosheets. The thickness of as-prepared β-Ga2O3 nanosheets is 1.35 nm, which did not change a lot in the nanosheet structure. The photocatalytic reaction indicates that the phase structure greatly affects the reaction activity. This work provides a template synthetic strategy to achieve atomically thick 2D nanosheets of the nonlayered material and will undoubtedly promote the understanding of the effects of the crystal structure on the photocatalytic reaction.

Results and Discussion

The formation processes for β-Ga2O3 nanosheets are shown in Figure b. First, the ultrathin nanosheets of nonlayered material γ-Ga2O3 were synthesized by a hydrothermal method without using any shape-controlling agents, as indicated in our previous work.[20] In our previous work, we have studied the surface property and thermal stability of the γ-Ga2O3 nanosheets. With the increase of the temperature, the physically adsorbed water would be lost and the terminated OH groups reconfigured into water molecules, leading to the partial structural collapse of γ-Ga2O3. The monoclinic β-Ga2O3 was obtained through the in situ annealed reaction from the unstable cubic phase γ-Ga2O3. As is reported, the crystalline phase transition of the unstable cubic γ-Ga2O3 into monoclinic β-Ga2O3 can be easily realized at this temperature,[21] which is also demonstrated by the X-ray power diffraction (XRD) analysis shown in Figure . Similar template synthesis research has also been reported in crystalline GaN nanosheets, in which the γ-Ga2O3 nanosheets with unstable structure and cubic crystallographic symmetry, as well as the close lattice constants, enable a simple structure transition from cubic γ-Ga2O3 to GaN nanosheets.[22]Figure a,c illustrates the crystal structural models of cubic γ-Ga2O3 and monoclinic β-Ga2O3. Obviously, the nanosheet morphology of γ-Ga2O3 is well reserved, and β-Ga2O3 nanosheets are obtained, as schematically illustrated in Figure d.

Figure 1

(a,c) Crystal structural models of cubic γ-Ga2O3 and monoclinic β-Ga2O3; (b) preparation process of γ-Ga2O3 and β-Ga2O3 nanosheets; and (d) schematic diagram describing the phase transformation of β-Ga2O3 nanosheets from γ-Ga2O3 nanosheets.

Figure 3

XRD patterns of (a) γ-Ga2O3 at different calcination temperatures and (b) β-Ga2O3 in comparison with that of γ-Ga2O3.

The morphology and texture of the β-Ga2O3 nanosheets were characterized by the field emission scanning electron microscopy (FE-SEM) and transmission electron microscopy (TEM). The thickness of the samples was determined by the atomic force microscope (AFM). As shown in SEM and TEM images (Figure a,b), the as-prepared β-Ga2O3 exhibits transparent and thin morphology of 2D nanosheets. Obviously, after calcination of γ-Ga2O3 nanosheets at a high temperature (750 °C), the nanosheet morphology is well preserved, except that the edges seem to be sharper and the size of nanosheets is smaller as compared with that of γ-Ga2O3 nanosheets (Figure S1), revealing that the phase transition is not accompanied by a significant change in symmetry.[23] The HR-TEM elemental mapping images shown in Figure c,d,g,h exhibit the distribution of elements in the nanosheets. It can be clearly seen that Ga and O are uniformly distributed in β-Ga2O3. Besides, the existence of the elements was also confirmed by energy-dispersive X-ray (EDX) spectroscopy shown in Figure S2. All of the elements were clearly observed at their corresponding keV values. The HR-TEM images of the samples shown in Figure S3 illustrate that the nanosheets exhibit interplanar distances of 0.250 and 0.255 nm, which correspond to the (311) and (111) planes of γ-Ga2O3 and β-Ga2O3, respectively. In addition, the thickness of the β-Ga2O3 nanosheets evaluated by the AFM was approximately 1.35 nm, which did not change a lot in the thickness after the thermal treatment (γ-Ga2O3 nanosheets: 1.25 nm). These results reveal that it is perfectly possible to obtain the nanosheets through the phase transformation by the template synthesis. To the best of our knowledge, such ultrathin 2D nanosheets of β-Ga2O3 have rarely been reported. The optimal reaction conditions to obtain more uniform and thinner nanosheets are being explored, and we believe that the more uniform and thinner nanosheets would possibly possess more excellent photocatalytic properties.

Figure 2

SEM and TEM images (a,b), TEM mapping (c,d,g,h), and AFM (e,f) image and the corresponding height profiles of the as-prepared β-Ga2O3 nanosheets.

Figure displays the XRD patterns of the as-synthesized samples. The phase transformation temperature was first investigated, as shown in Figure a. As the calcination temperature was 700 °C, the XRD indicates that β-Ga2O3 was formed and the characteristic peaks of γ-Ga2O3 were disappeared. When the temperature was further increased to 750 °C, the characteristic XRD diffraction peaks of β-Ga2O3 were observed. The strong and sharp diffraction peaks are consistent with the standard data file of β-Ga2O3. No other impurities were observed, indicating that pure-phase β-Ga2O3 was obtained. In addition, it is important to note that the diffraction peak at 2θ degree of 35.2° is stronger than that of 31.7° in the XRD pattern of β-Ga2O3, which is not consistent with the standard diffraction peaks. As we mentioned previously, the very strong diffraction peak of γ-Ga2O3 nanosheets at 36.19° demonstrates a (311) preferential orientation.[20] Therefore, it is speculated that the strong diffraction peak at 2θ degrees of 35.2° was caused by the preferential growth orientation of γ-Ga2O3 nanosheets at high temperature. The phenomenon was obviously observed in Figure b.

The Raman spectra were carried out to probe vibrational and structural properties of the as-prepared samples. Figure shows the Raman spectroscopy of the Ga2O3 samples obtained in ambient atmosphere with an excitation wavelength of 532 nm. For β-Ga2O3, peaks at 202, 350, 417, 477, 630, 651, 767 cm–1 were observed, which are characteristic peaks of monoclinic β-Ga2O3.[24] In contrast, only several weak, broad bands attributed to the bending and stretching of Ga–O bond can be observed in the Raman spectrum of γ-Ga2O3, which strongly confirms the low crystallinity of γ-Ga2O3. These results are in consistent with the XRD mentioned above.

Figure 4

Raman spectra (with an excitation wavelength of 532 nm) of the as-prepared Ga2O3 samples.

The Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda were used to determine the surface area and porous structure of the samples, respectively. As depicted in Figure , a typical IV N2 adsorption isotherm with H2- and H3-type hysteresis loops was observed for β-Ga2O3 and γ-Ga2O3, respectively, indicating the presence of mesopores in the Ga2O3 samples.[25,26] As reported in our previous work, the γ-Ga2O3 nanosheets obtained through hydrothermal reaction at low temperature result in the largest specific surface area of 227.6 m2 g–1 among ever-reported γ-Ga2O3 materials.[23] Compared with γ-Ga2O3, β-Ga2O3 obtained through the template synthesis exhibits a relatively small specific surface area (73.7 m2 g–1). Even so, the specific surface area of β-Ga2O3 nanosheets is larger than most of the β-Ga2O3 with other morphologies. Besides, the β-Ga2O3 nanosheets possess a smaller pore volume (0.020 cm3 g–1) and larger average pore diameter (5.7 nm) in comparison with γ-Ga2O3 nanosheets (volume of 0.125 cm3 g–1 and average diameter of 3.8 nm). It is expected that the differences in texture properties of β-Ga2O3 and γ-Ga2O3 would greatly influence their photocatalytic hydrogen evolution performance.

Figure 5

N2 adsorption–desorption curves and the pore size distribution plots of β-Ga2O3.

The UV–vis absorption spectra of the samples were performed in the range of 250–450 nm, which were converted to absorption spectra on the basis of the Kubelka–Munk theory,[27] as shown in Figure . The optical absorption edges of β- and γ-Ga2O3 are at 280 and 264 nm (Figure a), respectively, which can be ascribed to the excitation of electrons from valence band to conduction band. Obviously, the absorption edge of β-Ga2O3 is red shift of 16 nm in comparison with that of γ-Ga2O3, possibly resulting from the confined quantum effect of the nanosheets. The optical band gap (Eg) can be obtained using the Davis–Mott’s empirical equation: αhν = B(hν – Eg), where hν, α, B, and Eg represent the photon energy, absorption coefficient, a constant related to the material, and band gap, respectively. Moreover, n depends on the characteristics of the transition in a semiconductor (n = 1 for direct transition; n = 4 for indirect transition).[28] Therefore, the band gap energy of the Ga2O3 samples could be estimated from a plot of (αhν)2 versus the energy of light (hν). As shown in the inset of Figure a, the band gap energies of β- and γ-Ga2O3 are evaluated to be 4.43 and 4.68 eV, respectively, which agree well with the previously reported results.[29]Figure b depicts the Mott–Schottky plots of both Ga2O3 samples conducted at a frequency of 0.5 kHz. The positive slopes of the Mott–Schottky plots indicate that the two samples are n-type semiconductors.[30] The intercepts on the abscissa display the flat-band potentials of the samples. It is obvious that the flat-band potential of β-Ga2O3 is more positive than that of γ-Ga2O3 by 0.1 V. As the flat-band potential is located near the conduction band in the n-type semiconductor, it was normally considered as the conduction band of the n-type semiconductor. Combining the result from the UV–vis absorption spectra with the flat-band potentials of the samples, the approximate band-edge positions of the samples can be obtained, as shown in Figure c. The more positive conduction band minimum of β-Ga2O3 nanosheets may be detrimental to the separation and migration of photogenerated carriers, leading to a lower photocatalytic activity.[31]

Figure 6

(a) UV–vis absorption spectra, the inset shows the plots of (αhν)2 vs the energy of light (hν); (b) Mott–Schottky plots of the samples; and (c) band structure diagram of γ-Ga2O3 and β-Ga2O3 calculated by optical absorption and typical electrochemical Mott–Schottky methods.

Figure shows the typical PL spectra of the Ga2O3 nanosheets performed at room temperature with the excitation wavelength of 250 nm. It can be found that β-Ga2O3 and γ-Ga2O3 nanosheets exhibit strong emission peaks centered at 587 and 374 nm, respectively. Compared with γ-Ga2O3 nanosheets, β-Ga2O3 nanosheets show an obvious red shift of 213 nm and a much weaker intensity. The PL of Ga2O3 has been reported by many studies.[32−34] Noteworthy, the emission wavelengths are 587 nm (2.11 eV) and 374 nm (3.31 eV) for β-Ga2O3 and γ-Ga2O3, respectively, which are smaller than the band gaps of the samples. Therefore, it can be deduced that the PL may not simply originate from the band to band transitions but the surface states of nanosheets that can influence the luminescent transition probability. It has been confirmed that the blue luminescence in β-Ga2O3 is attributed to the recombination of an electron from a donor formed by oxygen vacancies and a hole from an acceptor formed by gallium vacancies.[35] The strong emission peak of γ-Ga2O3 observed in Figure was speculated to be caused by oxygen vacancies and the much higher intensity of the emission peak suggested a higher oxygen vacancy concentration,[36] which was also demonstrated by the electron paramagnetic resonance (EPR) results shown in Figure S4. Obviously, γ-Ga2O3 exhibited an obvious signal at g = 2.001, typical character of oxygen vacancy, whereas no signal was observed in β-Ga2O3.

Figure 7

Room-temperature PL spectra of Ga2O3 nanosheets with UV fluorescent light excitation of 250 nm.

X-ray photoelectron spectroscopy (XPS) surface measurements were carried out to further investigate the surface composition and chemical states of Ga2O3 nanosheets, and the results are shown in Figure . The survey XPS spectra, as presented in Figure a, illustrate the existence of two primary elements: Ga and O. No obvious peaks for impurities were observed, except the C 1s, which can be attributed to the adventitious hydrocarbon. The C 1s peak at 284.6 eV was used as a reference to correct for specimen charging. In the high-resolution Ga 2p XPS spectrum (Figure b), the peaks of Ga 2p1/2 and Ga 2p3/2 in the β-Ga2O3 nanosheets were located at around 1146.56 and 1119.68 eV, respectively, which shift by 0.2 eV toward lower binding energy in comparison with those in the γ-Ga2O3 nanosheets (1146.77 and 1119.88 eV). These peaks are associated with the Ga–O bonding in Ga2O3.[37] The peaks of O 1s in β-Ga2O3 shown in Figure c are observed to be located at 531.20, 532.09, and 533.58 eV, exhibiting a shift toward lower binding energy compared with those peaks (531.37, 532.45, and 533.91 eV) of O 1s in γ-Ga2O3. These peaks can be attributed to the lattice oxygen Ga–O, surface-bonded OH groups attached to H2O molecules via hydrogen bonds, and molecularly absorbed H2O, respectively.[38] Obviously, the binding energies of Ga 2p in the samples are not consistent with each other, indicating that the chemical states of Ga atoms are different. The peaks in Ga 2p and O 1s XPS in γ-Ga2O3 shift to a higher binding energy, which may result from the formation of neighboring oxygen vacancies showing a high electron-attracting effect.[39] Thus, the γ-Ga2O3 nanosheets may possess higher oxygen vacancy concentrations compared with β-Ga2O3. It has been reported that compared with photogenerated holes, photogenerated electrons are preferentially transferring to oxygen vacancies.[40] Oxygen vacancies can be the centers for capturing photogenerated electrons in the photocatalytic reaction. Thus, it is speculated that the β-Ga2O3 nanosheets with lower oxygen vacancy concentrations may possess less efficient separation and transportation of photogenerated electron–hole pairs for photocatalytic hydrogen evolution.

Figure 8

High-resolution XPS spectra: (a) survey XPS spectra scanned from 0 to 1200 eV and (b,c) high-resolution spectra of Ga 2p and O 1s.

The photoelectrochemical measurements including transient photocurrent response and electrochemical impedance spectra (EIS) were employed to investigate the migration ability of the photogenerated carriers. Figure a shows the I–t curves of the samples under the irradiation of 300 W Xe lamp. Compared with γ-Ga2O3, β-Ga2O3 exhibits a decrease in the transient photocurrent, indicating a lower photogenerated separation and transportation rates of carriers in the β-Ga2O3 nanosheets. It is further demonstrated by the EIS Nyquist plots performed in the dark, as shown in Figure b. The radius of β-Ga2O3 is significantly increased compared with that of γ-Ga2O3. The result is well consistent with the analysis in PL and XPS.

Figure 9

(a) Transient photocurrent response of the samples under the irradiation of 300 W Xe lamp and (b) Nyquist plots of electrochemical impedance spectroscopy of the samples in the dark.

It has been reported that the structure and surface properties have great effects on the photocatalytic activity of the samples.[41] Because the two kinds of Ga2O3 possess different structures and similar morphology, it would be interesting to evaluate their photocatalytic activity. The photocatalytic performance of the samples is evaluated by the hydrogen evolution from the water–methanol mixed solution. Figure a shows the photocatalytic activity of the samples under the irradiation of UV light for H2 evolution in 6 vol % methanol aqueous solution. The β-Ga2O3 nanosheets display a H2 evolution rate of 4.5 mmol·g–1·h–1, which is two times lower than that of γ-Ga2O3 (10.0 mmol·g–1·h–1). The produced H2 increases steadily in proportion to the irradiation time. After reaction for 450 min, the amount of H2 reaches to 11.3 mmol·g–1 for β-Ga2O3 and 25.0 mmol·g–1 for γ-Ga2O3. Average apparent quantum yield (AQY) was also calculated to further evaluate the photocatalytic activity of the Ga2O3 samples. The number of incident photons was calculated in the range of 250 nm (initial wavelength of the UV light source) to the optical absorption edge of the samples (280 nm for β-Ga2O3 and 264 nm for γ-Ga2O3). The AQY of β-Ga2O3 nanosheets was 0.29% in the range of 250–280 nm, whereas the value of γ-Ga2O3 nanosheets was 1.82% in the range of 250–264 nm. Obviously, the AQY of β-Ga2O3 nanosheets for H2 evolution is much smaller than that of γ-Ga2O3 nanosheets. It should be noted that the BET surface area of γ-Ga2O3 nanosheets (227.6 m2 g–1) is larger than that of β-Ga2O3 nanosheets (73.7 m2 g–1). However, the enhancement of AQY exceeds the effect of the difference in the specific surface area for β-Ga2O3 and γ-Ga2O3 nanosheets. Therefore, the specific surface area is not the key factor to determine the difference in the photocatalytic hydrogen evolution activity between β-Ga2O3 and γ-Ga2O3 nanosheets. The more active sites resulting from the unsaturated gallium atoms and higher separation and transfer rates of photogenerated carriers of γ-Ga2O3 nanosheets, which were confirmed by XPS and transient photocurrent response, were speculated to contribute to the higher photocatalytic hydrogen evolution activity. The stability of the samples during the photocatalytic process is of vital importance. As depicted in Figure b,c, the photocatalytic hydrogen evolution activity remains unchanged after three cycle runs.

Figure 10

(a) Photocatalytic H2 evolution over samples under UV light irradiation. (b,c) Stability measurements of γ-Ga2O3 and β-Ga2O3 nanosheets over 450 min.

Conclusions

In summary, the 2D β-Ga2O3 nanosheets were successfully fabricated from ultrathin 2D γ-Ga2O3 nanosheets through the one-step template synthesis. The process of structure transition from cubic γ-Ga2O3 to monoclinic β-Ga2O3 did not bring significant changes of the nanosheet morphology. Compared with γ-Ga2O3, β-Ga2O3 nanosheets exhibit a decreased photocatalytic hydrogen evolution activity. The PL spectra, XPS surface measurements, and N2 adsorption–desorption curves indicate that the β-Ga2O3 nanosheets with less disordered structure possess inefficient charge separation and transfer rates and lower specific surface area than γ-Ga2O3, which decreased the photocatalytic activity. Through the analysis of γ-Ga2O3 and β-Ga2O3 nanosheets, we proposed that our research may open promising prospects for the development of sheet-like nanostructures through template approach and bring more in-depth understanding of the effects of crystal structure and morphology on the photocatalytic reaction.

Experimental Section

Preparation of the Catalysts

Analytical grade commercial gallium nitrate and ammonia solution were purchased from Sigma-Aldrich and Sinopharm Chemical Reagent Co. Ltd, respectively, and were used without further purification. The γ-Ga2O3 nanosheets were synthesized by the ammonolysis of the gallium nitrate. In detail, 1 mmol gallium nitrate (0.2557 g) was simultaneously dropwise added into 20 mL solution (containing 10 mL ammonia solution and 10 mL H2O) under magnetic stirring at room temperature. After 30 min, the suspension solution was transferred into a Teflon-lined stainless-steel autoclave and maintaining the autoclave at 120 °C for 10 h in an electric oven. After cooling to room temperature naturally, the obtained sample was centrifuged and washed several times with deionized water and then collected and dried at 80 °C for 12 h, giving the white power (γ-Ga2O3). The γ-Ga2O3 nanosheets were converted to β-Ga2O3 by calcination at 700 °C for 3 h (with a heating rate at 5 °C min–1) under air. After the calcination process, the nanosheets were changed into smaller and sharper ones.

Characterization

The crystalline phases and structure of the obtained Ga2O3 samples were characterized by XRD on a Bruker D8 Advance X-ray diffractometer equipped with a Cu Kα X-ray source (λ = 1.5406 Å). The Raman spectra were collected using an inVia-Reflex micro-Raman spectroscopy system (Renishaw Co.) with an excitation wavelength of 532 nm at room temperature. The size and morphology of the samples were observed by FE-SEM (JEOL JSM-6701F and Nova NanoSEM 230) and TEM (Tecnai model G2 F20 S-TWIN transmission electron microscope), respectively. The thickness of the samples was evaluated by a tapping-mode AFM (NanoScope MultiMode IIIa, Veeco Instruments) with a Si-tip cantilever. The N2 absorption–desorption isotherms were measured on a Micromeritics apparatus model ASAP 2020. The UV–vis diffusion reflectance spectroscopy was recorded at room temperature by a Cary 500 Scan spectrophotometer with BaSO4 as a reflectance standard, and the room-temperature PL spectra were performed on a FL/FS 920 TCSPC fluorescence spectrophotometer (Edinburgh) with a pulsed xenon flash lamp excited at 250 nm. Electron spin resonance was recorded over a Bruker ESP 300E EPR spectrometer. XPS measurements were conducted on a PHI Quantum 2000 XPS system with a monochromatic Al Kα source and a charge neutralizer. Photoelectrochemical measurements were performed in a conventional three-electrode electrochemical cell with a working electrode, a Pt counter electrode, and a saturated Ag/AgCl electrode as the reference electrode. During the experiment, the three electrodes were immersed in a sodium sulfate electrolyte solution (0.2 M).

The photocatalytic hydrogen evolution reaction was carried out in a photoreaction vessel made of quartz connected to a closed glass gas circulation system. The photocatalytic reaction was performed in a dispersion of 20 mg of photocatalyst in 155 mL of CH3OH solution (containing 145 mL of distilled water and 10 mL of CH3OH). The system was evacuated using a mechanical pump several times to completely remove the air. A 125 W high-pressure mercury lamp was used as the light source for photocatalytic hydrogen generation reaction. The reaction temperature of the photocatalytic system was controlled at 5 °C by a continuous flow of cooling water. The evolved gases were analyzed by an online gas chromatograph (argon as the carrier gas), equipped with a thermal conductive detector. The average AQY of the samples was measured under the same photocatalytic reaction conditions. The incident light intensity of UV light source was measured by SpectriLight ILT950, and the total number of incident photons was measured using a calibrated silicon photodiode. The irradiance spectrum of the high-voltage mercury lamp with 125 W has been added in the Supporting Information (Figure S5). The AQY of H2 evolution over the Ga2O3 nanosheets was calculated according to the following equations: