Long-wavelength photoresponsive gallium zinc oxynitride for efficient oxygen evolution and Z-scheme water splitting reactions

Abstract

Long-wavelength photoresponsive GaN:ZnO, a solid solution of GaN and ZnO, can be obtained by reacting Ga₂O₃ and Zn₃N₂ in the presence of ZnX₂ (X = halogen) in a sealed evacuated tube. However, the activity of GaN:ZnO for overall water splitting via one-step excitation under visible light remains low due to the small particle size, which leads to uncontrollable aggregation and lack of suitable cocatalysts. In this study, well-dispersed particulate GaN:ZnO with a ZnO concentration of 66 mol%, optical absorption up to about 600 nm, and anisotropic crystalline facets was obtained using δ-Ga₂O₃ as a raw material and adding Zn to the starting material. The resulting GaN:ZnO exhibited an apparent quantum yield of 11.9% at 420 nm in the oxygen evolution reaction when loaded with an IrO_x cocatalyst. It also served as an oxygen evolution photocatalyst in a photocatalyst sheet for Z-scheme water splitting, in combination with La₅Ti₂Cu_0.9Ag_0.1O₇S₅, whose absorption edge wavelength was about 700 nm, as a hydrogen evolution photocatalyst. This work, which demonstrates the activation of long-wavelength photoresponsive GaN:ZnO, expands the possibilities of constructing photocatalytic systems that effectively utilize long-wavelength visible light for water splitting.

---

Introduction

Large-scale implementation of solar fuel production is desired in order to address increasingly serious energy and environmental problems.¹ Sunlight-driven water splitting using particulate semiconductor photocatalysts has been studied as a means of green hydrogen production on a large scale.^2–4 To produce green hydrogen in a practical manner, it is indispensable to develop photocatalytic materials that can efficiently split water into hydrogen and oxygen under long-wavelength visible light.⁵ Various semiconducting (oxy)nitrides, oxysulphides and metal-doped oxides have been extensively studied as visible-light responsive photocatalysts.^6–9 Despite having a narrower bandgap than the corresponding pure oxides, many such materials have a band structure suitable for splitting water into hydrogen and oxygen under visible-light irradiation. This is due to a negative shift of the valence band edge potential induced by nitride and sulphide ions or impurity levels associated with dopants, while the conduction band edge potential is largely unaffected.

GaN:ZnO is one of the earliest oxynitride photocatalysts shown to be capable of splitting water into hydrogen and oxygen under visible light.¹⁰ The bandgap energy of GaN:ZnO can be controlled by the concentration of ZnO. With increasing ZnO concentration up to 90 mol%, the valence band maximum shifts negatively while the conduction band minimum shifts positively.¹¹ In 2010, an apparent quantum yield (AQY) of 5.1% at 410 nm was achieved in the overall water splitting reaction using GaN:ZnO with a ZnO concentration of 13 mol% and an absorption edge wavelength of about 480 nm.¹² More recently, in 2022, GaN:ZnO with an absorption edge wavelength of around 550 nm and activity in the water splitting reaction was synthesised by applying NH₄Cl as a solid nitrogen source in a sealed evacuated tube.¹³ The activity of the resultant long-wavelength responsive GaN:ZnO for overall water splitting was rather low, with a solar-to-hydrogen energy conversion efficiency (STH efficiency) on the order of 1 × 10⁻²%. Nevertheless, GaN:ZnO could be effectively applied to visible-light-driven water splitting by constructing a Z-scheme system using two-step excitation of a hydrogen evolution photocatalyst (HEP) and an oxygen evolution photocatalyst (OEP). Namely, GaN:ZnO effectively functioned as the OEP in combination with Rh-doped SrTiO₃ as the HEP in the presence of Fe³⁺/Fe²⁺ as the redox mediator, and the resulting Z-scheme system split water with a higher STH efficiency of 3.7 × 10⁻²%. The use of long-wavelength photoresponsive GaN:ZnO allowed the wavelength range of light applicable to Z-scheme water splitting reactions to be significantly extended, because most previous work had used WO₃ and BiVO₄ as OEPs for Z-scheme water splitting, whose absorption edge wavelengths were rather short, at 470 and 525 nm, respectively.^14,15

Further improvements in efficiency can be expected by optimizing the performance of GaN:ZnO. At the same time, it is still essential to explore new long-wavelength HEP materials, because Rh-doped SrTiO₃ can utilise visible light only up to 520 nm, and its visible-light absorption induced by the Rh impurity level is rather weak.¹⁶

In our recent study, GaN:ZnO absorbing visible light up to 580 nm was synthesized by reacting crystalline Ga₂O₃ and Zn₃N₂ in the presence of ZnX₂ (X = halogen) in sealed evacuated tubes.¹⁷ The use of hydrogen-free Zn₃N₂ as a nitriding reagent allowed gram-scale synthesis of GaN:ZnO that was active for overall water splitting via one-step photoexcitation. However, the AQY remained on the order of 1 × 10⁻²% at a wavelength of 430 nm. This low activity is due to the uncontrolled particle size and morphology of GaN:ZnO, resulting in severe aggregation, which makes it difficult to refine the cocatalyst loading.

In general, the properties of the products of solid-state reactions are affected by the particle size and crystallinity of the reactants.¹⁸ Therefore, refinement of the raw materials is expected to lead to improved morphology control and photocatalytic activity of GaN:ZnO obtained using Zn₃N₂ as a solid nitrogen source. In the present study, well-dispersed crystalline GaN:ZnO particles with a high ZnO concentration were synthesised using δ-Ga₂O₃ as a raw material and adding Zn to the starting materials. The particles absorbed visible light up to 600 nm and exposed hexagonal crystalline facets. The resulting GaN:ZnO drove the oxygen evolution reaction with an AQY of 11.9% at 420 nm when loaded with an IrO_x cocatalyst. When used in combination with La₅Ti₂Cu_0.9Ag_0.1O₇S₅ (LTCA) as a HEP, it also enabled a Z-scheme system consisting only of non-oxide photocatalysts with an absorption edge wavelength equal to or longer than 600 nm, but still active in visible light-driven water splitting.

Results and discussion

Synthesis of GaN:ZnO

Fig. S1A in ESI† shows XRD patterns for the δ- and K-Ga₂O₃ samples. The pattern for the δ-Ga₂O₃ sample is consistent with that reported in the literature.¹⁹ The synthesised δ-Ga₂O₃ particles were aggregated to form irregularly shaped secondary particles (Fig. S1B in ESI†). TGA measurements revealed that the δ-Ga₂O₃ sample contained about 8 wt% water, as indicated by the weight loss associated with endothermic signals during the heating process (Fig. S2 in ESI†). On the other hand, the K-Ga₂O₃ was composed of a mixture of α-Ga₂O₃ and β-Ga₂O₃. The K-Ga₂O₃ particles had a high-aspect-ratio rectangular parallelepiped shape with a length of several micrometers and a thickness of less than 1 μm. The K-Ga₂O₃ contained a negligible amount of water. Fig. 1A and B respectively show XRD patterns and UV-vis DRS data for GaN:ZnO synthesised using δ- and K-Ga₂O₃ as a starting material with and without addition of Zn after washing with distilled water. The washing step was performed to investigate byproducts of the synthesis of GaN:ZnO. Single-phase GaN:ZnO was obtained from K-Ga₂O₃, while a small amount of ZnGa₂O₄ was generated as an impurity phase, as confirmed by the diffraction peak at 36.8°, when δ-Ga₂O₃ was used. This is due to oxidation of Zn₃N₂ by the moisture contained in δ-Ga₂O₃. The addition of metallic Zn to the precursor mixture prevented the formation of ZnGa₂O₄, because Zn reacted with moisture to form ZnO and H₂ in the sealed quartz tube. On the other hand, unreacted Zn remained when Zn was added to the synthesis using K-Ga₂O₃. GaN:ZnO(δ) prepared with Zn addition had similar crystallinity to GaN:ZnO(K), but exhibited diffraction peaks at lower angles closer to those for ZnO than GaN, indicating the production of GaN:ZnO with a higher ZnO concentration. In fact, the ZnO concentration estimated from the Zn/Ga ratio measured by EDS analysis was higher for GaN:ZnO(δ) prepared with Zn addition (63 mol%) than for GaN:ZnO(K) prepared with or without Zn addition (58 mol% for both). The higher ZnO concentration in the former case indicates that ZnO generated from Zn and moisture was partially incorporated in the GaN:ZnO(δ) product. In accordance with this, the GaN:ZnO(δ) sample prepared with Zn absorbed visible light up to 600 nm, which was approximately 30 nm longer than the absorption edge for GaN:ZnO(K). Note that when commercial submicron ZnO particles were added instead of Zn for the synthesis of GaN:ZnO(δ), the XRD peak for the product was not shifted relative to that for conventional GaN:ZnO(K), and instead XRD peaks attributable to ZnO were generated, indicating that externally added ZnO was not incorporated into GaN:ZnO (Fig. S3†).

Fig. 1 (A) XRD patterns (B) UV-vis DRS data for GaN:ZnO prepared using (a) δ-Ga₂O₃, (b) δ-Ga₂O₃ with addition of Zn, (c) K-Ga₂O₃ and (d) K-Ga₂O₃ with addition of Zn by heating at 1173 K for 10 h. (e) and (f) Show XRD patterns for commercial ZnO and GaN (ICDD: 00-050-792), respectively, for comparison. Peaks for Zn and ZnGa₂O₄ are indicated with asterisks (*) and circles (○), respectively.

Fig. 2 shows SEM images of GaN:ZnO(δ) and GaN:ZnO(K) synthesised with and without Zn addition. The morphology and size of the GaN:ZnO(K) particles are not very clear or uniform regardless of the addition of Zn during synthesis. In contrast, hexagonal prismatic particles reflecting the wurtzite-type crystal structure of GaN:ZnO are observed when δ-Ga₂O₃ is used because Ga₂O₃ can be more smoothly converted into GaN:ZnO. This effect is enhanced by the addition of Zn to the starting material, and allows GaN:ZnO(δ) particles to grow. This is probably related to suppression of oxidation and hydrolysis of the starting Zn₃N₂ and the intermediate Zn₂NI, as described later.

Fig. 2 SEM images of (a and b) GaN:ZnO(K) and (c and d) GaN:ZnO(δ) prepared by heating at 1173 K for 10 h (a and c) without and (b and d) with Zn addition.

Fig. 3 shows XRD and DRS data for GaN:ZnO(δ) prepared at different temperatures. All the synthesized samples were attributed to single-phase wurtzite-type GaN:ZnO. It can be seen that the diffraction peaks become sharper as the synthesis temperature is increased, indicating an improvement in crystallinity. Namely, the full width at half maximum (FWHM) values for the 101 diffraction peak for GaN:ZnO(δ) synthesised at 923 and 1123 K are 0.35° and 0.17°, respectively. At temperatures above 1123 K, the FWHM is largely unchanged. The onset of optical absorption observed in the DRS data was around 600 nm regardless of the synthesis temperature, and was approximately 30 nm longer than that for GaN:ZnO(K). This reflects a higher ZnO concentration in the GaN:ZnO(δ) samples due to the incorporation of ZnO generated in situ from Zn and H₂O during synthesis. In fact, elemental analysis revealed that the ZnO concentration in the GaN:ZnO(δ) samples was around 65 mol% (Table 1), while that for GaN:ZnO(K) reported in our previous work was 58 mol%.¹⁷

Fig. 3 (A) XRD and (B) DRS data for GaN:ZnO(δ) prepared with Zn addition at (b) 923, (c) 973, (d) 1023, (e) 1073, (f) 1123, (g) 1173, and (h) 1223 K for 10 h. (a) and (i) Show peaks for commercial ZnO and GaN (ICDD: 00-050-792), respectively, for comparison.

Table 1 Bulk composition of GaN:ZnO(δ) prepared at different temperatures

Synthesis temperature/K	Ga/wt%^a	Zn/wt%^a	N/wt%^b	O/wt%^b	Total/wt%	Composition
a Measured by ICP-AES. b Measured by N/O combustion analysis.
1073	28.3	49.6	5.8	13.0	96.7	Ga0.35Zn0.65N0.35O0.67
1123	28.4	50.1	5.7	12.8	97.0	Ga0.35Zn0.65N0.34O0.66
1173	28.5	50.8	5.7	12.9	97.9	Ga0.34Zn0.66N0.34O0.66
1223	28.0	49.6	5.8	13.2	96.6	Ga0.35Zn0.65N0.33O0.67

---

SEM images of GaN:ZnO(δ) synthesized at different temperatures are shown in Fig. 4. GaN:ZnO(δ) synthesised at 923 K was composed of particles smaller than 100 nm without noticeable crystal facets. Increasing the synthesis temperature led to the growth of particles with some crystal facets, and well-dispersed hexagonal prismatic particles reflecting the crystal structure and hundreds of nanometers in size were generated at 1073 K and above. Note that GaN:ZnO with such anisotropic particle shapes can also be synthesised using K-Ga₂O₃ instead of δ-Ga₂O₃ when Zn₂NI is used as the zinc and nitride ion source.¹⁷ In fact, Zn₂NI is generated in situ from Zn₃N₂ and ZnI₂ during the synthesis of GaN:ZnO. However, Zn₃N₂ is prone to oxidation and hydrolysis. In addition, Zn₂NI starts to decompose at 873 K to Zn, N₂, and ZnI₂, which are no longer reactive enough to form GaN:ZnO.²⁰ The use of δ-Ga₂O₃, which is more reactive than K-Ga₂O₃, is therefore preferable for the growth of hexagonal prismatic particles of GaN:ZnO, as the decomposition of Zn₂NX into inert byproducts prior to the formation of GaN:ZnO can be suppressed. Notably, Pt and MnO_x were reductively and oxidatively photodeposited respectively on the hexagonal and lateral planes of hexagonal prismatic particles of GaN:ZnO (Fig. S4 in ESI†). This indicates a spontaneous polarisation of GaN:ZnO like GaN and ZnO.^21–23

Fig. 4 SEM images of GaN:ZnO(δ) synthesised with addition of Zn at (a) 923, (b) 973, (c) 1023, (d) 1073, (e) 1123, (f) 1173, and (g) 1223 K.

A transmission electron microscopy (TEM) image and electron diffraction patterns for GaN:ZnO(δ) synthesised at 1173 K are shown in Fig. 5. The electron diffraction patterns obtained at different positions in the particle shown in the TEM image are similar, indicating that the GaN:ZnO(δ) particle is a single crystal. Considering the results of the elemental analysis (Table 1), the GaN:ZnO(δ) material is also stoichiometric. Nevertheless, as seen in the TEM image, the GaN:ZnO(δ) particle contains internal voids suggesting the presence of local defects and distortions.

Fig. 5 Cross-sectional TEM image and (a–e) selected area electron diffraction patterns for GaN:ZnO(δ) synthesised at 1173 K for 10 h. The positions where the diffraction patterns were acquired are labelled with letters a to e.

Photocatalytic water splitting reaction

Fig. 6 shows the oxygen evolution activity under visible light in an aqueous AgNO₃ solution for GaN:ZnO(δ) synthesised at 973–1223 K and loaded with IrO_x (0.5 wt% Ir). The activity is seen to increase as the synthesis temperature is raised from 973 to 1173 K due to improvements in crystallinity. However, the activity for the sample synthesised at 1223 K decreases even though there is no appreciable deterioration in crystallinity or bulk composition. This may be related to a change in the surface properties of GaN:ZnO(δ). In fact, when GaN:ZnO was synthesized at high temperature using NH₄Cl as a nitriding reagent, a decrease in the surface Zn/Ga ratio was observed, and the photocatalytic activity was eventually degraded.¹³ XPS analysis of GaN:ZnO(δ) was therefore carried out. Fig. S5 and Table S1 in ESI† show the surface composition of GaN:ZnO(δ) prepared at different temperatures. GaN:ZnO(δ) synthesised at 1223 K showed a higher Zn/Ga molar ratio on the surface than the other GaN:ZnO samples, probably due to the heating at too high a temperature in an oxygen-free atmosphere for this material. The large deviation between the bulk and surface compositions would disturb the band structure and carrier transfer, which could lower the photocatalytic activity. As seen in Fig. S6 in ESI,† the oxygen evolution activity for GaN:ZnO(δ) decreases to about 20% in the absence of an IrO_x cocatalyst, indicating that the IrO_x acts as an efficient oxygen evolution cocatalyst as reported previously.¹³ The oxygen evolution activity is almost constant for an Ir loading of 0.2 wt% or greater. The wavelength dependence of the AQY for the most active IrO_x-GaN:ZnO(δ) sample, which was synthesised at 1173 K and loaded with 0.5 wt% of Ir, for the oxygen evolution reaction is presented in Fig. 7. As can be seen, the AQY at a wavelength of 420 nm is 11.9%, which is comparable to that for GaN:ZnO synthesised using NH₄Cl as a solid nitrogen source, and outstanding among GaN:ZnO with high ZnO concentration (Table S2 in ESI†). Notably, the GaN:ZnO(δ) sample maintains a reasonably high AQY of 1% at a wavelength of 560 nm, but its activity becomes negligible under irradiation with monochromatic light at 600 nm, which represents the absorption edge. This suggests that the photocatalytic reaction proceeds through bandgap transitions of GaN:ZnO(δ) and that visible light absorption up to 600 nm can indeed contribute to the photocatalytic activity. On the other hand, GaN:ZnO(δ) prepared at 1123 K showed negligible hydrogen evolution activity from aqueous methanol solution although it generated hydrogen from aqueous ascorbic acid solution (Fig. S7 in ESI†). In previous studies, the conduction band minimum and valence band maximum for GaN:ZnO with a ZnO concentration of 50–68 mol% have been estimated to be in the range of −0.58 to −0.08 V vs. RHE and +1.80 to +2.47 V vs. RHE, respectively, with large variations probably due to differences in stoichiometry and homogeneity of the materials.^11,13,24 However, as the ZnO concentration increases, the energy offset between the conduction band minimum and the hydrogen evolution reaction potential decreases more than the energy offset between the valence band minimum and the oxygen evolution reaction potential, which may explain the low hydrogen evolution activity of GaN:ZnO(δ) in this work.

Fig. 6 Oxygen evolution activity in aqueous AgNO₃ solution (30 mM) for IrO_x-GaN:ZnO(δ) synthesised at different temperatures. The reactions were carried out under visible light irradiation (λ > 420 nm) from a 300 W Xe lamp. The amount of IrO_x loading was 0.5 wt%.

Fig. 7 AQY for GaN:ZnO(δ) synthesised at 1173 K and loaded with IrO_x (Ir 0.5 wt%) for oxygen evolution reaction as function of irradiation wavelength. The reactions were carried out in an aqueous AgNO₃ solution (30 mM) under illumination from a 300 W Xe lamp equipped with various bandpass filters.

Since IrO_x-GaN:ZnO(δ) exhibited reasonably high activity in the oxygen evolution reaction in response to visible light with relatively long wavelengths, its applicability as an OEP for Z-scheme water splitting was investigated using a LTCA/Au/IrO_x-GaN:ZnO sheet. After photodeposition of Cr₂O₃/Rh cocatalysts, the sheet evolved H₂ and O₂ stoichiometrically under visible light from water free from sacrificial reagents or additives (Fig. 8). A sheet based on GaN:ZnO(δ) without IrO_x modification exhibited lower activity, indicating the critical role of the IrO_x cocatalyst in activating GaN:ZnO(δ). The STH efficiency for the sheet was 0.05% at 5 kPa and 310 K, which was an improvement on a previously reported SrTiO₃:Rh/Fe^3+/2+/GaN:ZnO suspension system, indicating the potential of achieving higher STH efficiency by using GaN:ZnO as the OEP in a sheet system.

Fig. 8 Time course for H₂ (red) and O₂ (blue) gas evolution for photocatalyst sheets using unmodified GaN:ZnO (open circles) and IrO_x-loaded GaN:ZnO (closed squares) as the OEP in combination with LTCA and Au as the HEP and conductor layer, respectively, under visible light (λ > 420 nm).

Experimental

Preparation of GaN:ZnO

δ-Ga₂O₃ and Zn₃N₂ were synthesised as starting materials for GaN:ZnO in the laboratory. δ-Ga₂O₃ was prepared by calcining Ga(NO₃)₃·xH₂O (Kojundo Chemical Laboratory Co., Ltd, 99.9%) at 573 K for 18 h in air.¹⁹ The obtained white powder was crushed and again calcined at 573 K for 2 h. Zn₃N₂ was synthesized by heating commercial metallic Zn powder (Kojundo Chemical Laboratory Co., Ltd, particle size <75 μm, 99.9%) at 873 K for 10 h under an NH₃ flow at a rate of 100 mL min⁻¹.¹⁷ δ-Ga₂O₃, Zn₃N₂, ZnI₂ (Sigma-Aldrich, 98%), and Zn were blended at a molar ratio of 1 : 1.1 : 2 : 1 in a N₂-filled glove box (dew point < 197 K, O₂ < 1 ppm), where Zn was added to prevent unwanted oxidation of the sample during heating. Commercially available Ga₂O₃ (Kojundo Chemical Laboratory Co., Ltd, 99.9%), denoted as K-Ga₂O₃, was also used to synthesise GaN:ZnO for comparison. The precursor mixture was transferred to a quartz tube, sealed under evacuation, and heated at 923–1273 K for 10 h. After heating, the sample was allowed to cool naturally to room temperature. The obtained GaN:ZnO powder was washed using 0.5 M HNO₃ for 1 h and thoroughly rinsed with distilled water to remove excess Zn and ZnI₂. The GaN:ZnO powder was recovered by filtration and dried overnight at 313 K in an oven. The sample was subsequently calcined in air at 873 K for 1 h to reduce the density of defects in the material and improve its photocatalytic activity.²⁵ The GaN:ZnO samples synthesized from δ-Ga₂O₃ and K-Ga₂O₃ are denoted as GaN:ZnO(δ) and GaN:ZnO(K), respectively.

The IrO_x cocatalysts (0.5 wt%) were loaded onto the surface of GaN:ZnO as an oxygen evolution cocatalyst by a solvothermal reaction using a microwave reactor (Monowave 200, Anton-Paar). GaN:ZnO was dispersed in ethylene glycol containing Na₃IrCl₆·nH₂O (Kanto Chemical, 99.5%), followed by heating at 423 K for 10 min.^7,13 The sample was washed with distilled water and recovered by filtration. The IrO_x-loaded GaN:ZnO powder was collected by filtration and dried at 313 K.

Rh was loaded onto IrO_x/GaN:ZnO by the photodeposition for the measurement of hydrogen evolution activity. The IrO_x/GaN:ZnO was suspended in a aqueous 10 mM ascorbic acid solution. To the suspension, Na₃RhCl₆·nH₂O (Mitsuwa Chemicals, 18.3 wt% Rh) was added so that the fraction of Rh became 0.5 wt% with respect to the photocatalyst sample. The suspension was evacuated to completely remove dissolved air and then exposed to visible light (λ ≥ 420 nm) from a Xe lamp equipped with a dichroic mirror and a cut-off filter.

To investigate the possibility of the facet-oriented electron/hole migration, Pt and MnO_x were reductively and oxidatively photodeposited on GaN:ZnO, respectively. Firstly, Pt (0.5 wt%) was loaded by reductive photodeposition. GaN:ZnO was dispersed in 10 vol% methanol solution containing H₂PtCl₆·6H₂O (Kanto Chemical Co., Inc., >98.5%), followed by irradiation with visible light (λ ≥ 420 nm). Subsequently, MnO_x (Mn 0.5 wt%) was loaded by oxidative photodeposition. Pt-loaded GaN:ZnO was dispersed in 10 mM NaIO₃ solution containing Mn(NO₃)₂·6H₂O (Fujifilm Wako Pure Chemical Co., 98.0%), followed by irradiation with visible light (λ ≥ 420 nm). The reactant solution was evacuated to remove air, and 10 kPa Ar was added prior to photodeposition.

Preparation of LTCA

Mg- and Al-codoped LTCA was synthesized by a conventional solid-state reaction in a sealed quartz tube similarly to a previous report.²⁶ TiO₂ (rutile, Kanto Chemical Co. Inc., 99.99%) was first impregnated with Mg(NO₃)₂·6H₂O (Fujifilm Wako Pure Chemical Co., 99.5%) and Al(NO₃)₃·9H₂O (Fujifilm Wako Pure Chemical Co., 99.5%) as Mg and Al dopant sources, respectively, and calcined at 1073 K for 1 h in air. The Mg (Mg⁻¹ + Al + Ti) and Al/(Mg + Al + Ti) molar ratios were both 1%. The resulting material is denoted as Mg- and Al-codoped TiO₂. Before mixing with other chemicals, La₂O₃ (Kanto Chemical Co., Inc., 99.99%) was calcined in air at 1273 K for 10 h to remove moisture and carbonates. La₂O₃, La₂S₃ (Kojundo Chemical Laboratory Co., Ltd, 99.9%), Mg- and Al-codoped TiO₂, Cu₂S (Kojundo Chemical Laboratory Co., Ltd, 99%), Ag₂S (Kojundo Chemical Laboratory Co., Ltd, 99%) and sulfur (Kojundo Chemical Laboratory Co., Ltd, 99.99%) were mixed at a molar ratio of 2 : 3 : 4 : 0.9 : 0.1 : 0.5 in a N₂-filled glovebox. The precursor mixture was sealed in an evacuated quartz tube and heated from room temperature to 473 K in 9 min, from 473 to 673 K in 100 min, and from 673 to 1323 K in 53 h, and then maintained at 1323 K for 96 h. After the annealing process, the resulting solid chunk was ground into powder. The obtained Mg- and Al-codoped LTCA is denoted as LTCA for simplicity.

Fabrication of photocatalyst sheet

Photocatalyst sheets employing LTCA and GaN:ZnO as the HEP and OEP, respectively, were fabricated by the particle transfer method.^27,28 Briefly, LTCA and IrO_x-loaded GaN:ZnO (denoted as IrO_x-GaN:ZnO) (10 mg each) were suspended in 0.5 mL of 2-propanol. After sonication for a few minutes, the solution was dropcast onto a 3 × 3 cm glass plate. After the plate was dried at room temperature, an Au layer was deposited by vacuum evaporation (VFR-200 M/ERH, ULVAC) to a thickness of 0.3 μm. The plate was then heated at 523 K for 10 min in air to reduce the contact resistance between the particles and the Au layer. The primary glass plate was then attached to a second glass plate using adhesive carbon tape to peel off the powder/Au assembly, and then sonicated for a few seconds to remove the loosely bound particles to obtain a working photocatalyst sheet. To conduct Z-scheme water splitting reactions, the resulting sheet, denoted as LTCA/Au/IrO_x-GaN:ZnO, was successively loaded with Rh (0.1 μmol) and Cr₂O₃ (0.2 μmol as Cr) by photodeposition under visible light irradiation (λ ≥ 420 nm) for 1 h each.

Characterisation of materials

The GaN:ZnO samples were analysed after washing with 0.5 M HNO₃ and calcining in air at 873 K, unless otherwise noted. X-ray diffraction (XRD) patterns were acquired using a Rigaku MiniFlex 300 with Cu Kα radiation operating at 30 kV and 10 mA. UV-vis diffuse reflectance spectroscopy (DRS) data were obtained using a spectrometer (V-670, JASCO) equipped with an integrating sphere. Scanning electron microscopy (SEM) images were captured using a Hitachi SU8000 system. TEM images and selected area electron diffraction (SAED) patterns were recorded using a JEOL JEM-2800. The cross-sectional sample for TEM observation was made by Ar ion milling using a JEOL EM-09100IS ion Slicer. Thermogravimetric analysis (TGA) measurements were performed using a Rigaku Thermo Plus TG 8120. The measurements were conducted by raising the temperature from room temperature to 1173 K at 10 K min⁻¹ in ambient air. Energy-dispersive X-ray spectroscopy (EDS) was performed using an analytical system integrated with a desktop SEM instrument (Phenom Pharos, ThermoFisher Scientific). XPS analysis was carried out using a PHI Quantera II spectrometer with an Al Kα X-ray source. The binding energy was calibrated using the C 1s peak (285.0 eV) The bulk composition of the samples was analysed by inductively coupled plasma-atomic emission spectroscopy (ICP-AES, ICPS-8100, Shimadzu) and oxygen/nitrogen analysis (EMGA-620W, Horiba).

Photocatalytic reactions

Water splitting reactions were carried out in a Pyrex top-irradiation-type reaction vessel connected to a closed gas circulation system. The oxygen evolution reaction was performed using 0.1 g of IrO_x-loaded GaN:ZnO powder dispersed in an aqueous 30 mM AgNO₃ solution, while the hydrogen evolution reaction was performed using 0.1 g of Rh- and IrO_x-coloaded GaN:ZnO dispersed in an aqueous ascorbic acid (10 mM) or methanol (10 vol%) solution. The reactant solution was evacuated to remove air and irradiated with visible light (λ ≥ 420 nm) from a 300 W Xe lamp equipped with a dichroic mirror and a cut-off filter after introduction of Ar (10 kPa) into the reaction system. The reactor temperature was maintained at 283 K by circulating cooling water in the reactor jacket. The amount of evolved gases was analysed using a gas chromatograph (GC-8A, Shimadzu) equipped with Molecular Sieve 5A columns and a thermal conductivity detector.

The overall water splitting reaction was carried out using the same reaction system. Distilled water (40 mL) was added to the reactor, and a Cr₂O₃/Rh-loaded photocatalyst sheet was immersed. After degassing the reaction system thoroughly, the background pressure was adjusted to 5 kPa by introducing Ar. The photocatalyst sheet was illuminated with visible light (λ ≥ 420 nm) from a 300 W Xe lamp or simulated sunlight generated by a solar simulator (HAL-320, Asahi Spectra Co., Ltd).

Apparent quantum yield measurement

The AQY for the oxygen evolution reaction was calculated as
where A, R and I denote the number of electrons involved in the reaction, the number of gas molecules generated, and the number of incident photons in a given time span, respectively. In the oxygen evolution reaction from water, the value of A is 4 because it is a four-electron reaction. The oxygen evolution reaction was carried out in a 10 mM AgNO₃ aqueous solution and under irradiation from a Xe lamp equipped with a series of bandpass filters to produce monochromatic light. The number of incident photons was determined using a LS-100 spectroradiometer (EKO Instruments Co., Ltd). In the Z-scheme water splitting reaction using LTCA/Au/IrO_x-GaN:ZnO sheet, A = 4 and R is the rate of H₂ evolution. The coefficient of 4 reflects a two-step photoexcitation process to drive the hydrogen evolution reaction, which is a two-electron reaction. Z-scheme water splitting was carried out under the same conditions described in the previous section except for the use of a bandpass filter (λ = 420 nm) instead of a cut-off filter.

Solar-to-hydrogen energy conversion efficiency measurement

The STH efficiency for the photocatalyst sheet during Z-scheme water splitting was measured using the same experimental apparatus but with illumination from a solar simulator (HAL-320, Asahi Spectra Co., Ltd). The STH efficiency was calculated as
where R(H₂), ΔG, P, and S are the rate of hydrogen evolution from the Z-scheme water splitting system, the change in Gibbs free energy that accompanies water splitting, the energy intensity of the solar light irradiation, and the irradiation area (∼6.25 cm²), respectively. A value of 237 kJ mol⁻¹ was used for ΔG_f, corresponding to the standard condition at 298 K. The energy intensity of the simulated sunlight (P) was adjusted to match that of Air Mass 1.5 Global (AM1.5 G) irradiance in the wavelength range of 350–800 nm and was assumed to be 100 mW cm⁻².

Conclusions

Porous but well-dispersed single crystal GaN:ZnO particles absorbing visible light up to 600 nm and exposing hexagonal crystal facets were synthesised in a sealed evacuated tube by using poorly crystalline δ-Ga₂O₃ as a raw material and adding metallic Zn to the starting materials. The Zn reacted with moisture generated from δ-Ga₂O₃ to form ZnO in situ, preventing oxidation of the nitriding reagents during the synthesis of GaN:ZnO. The in situ generated ZnO was incorporated into the GaN:ZnO material, increasing the ZnO content to about 65 mol%, which allowed extension of the visible light absorption edge to 600 nm. Moreover, the use of δ-Ga₂O₃ favored the growth of hexagonal prismatic GaN:ZnO particles by suppressing decomposition of the reactants and the associated production of inert byproducts.

The obtained GaN:ZnO evolved oxygen from an aqueous AgNO₃ solution with an AQY of 11.9% at 420 nm when loaded with an IrO_x cocatalyst. Moreover, IrO_x-GaN:ZnO acted as an OEP in a photocatalyst sheet for Z-scheme water splitting in combination with doped LTCA as a HEP and an Au thin film as a conductor layer. The resulting photocatalyst sheet exhibited a STH efficiency that was superior to that for a previously-reported Z-scheme suspension system using IrO_x-GaN:ZnO. There is still much room for clarification and optimization of the properties of GaN:ZnO and photocatalyst sheets containing it. This is expected to lead to improvements in the activity of photocatalytic sheets based on narrow bandgap nonoxide materials such as GaN:ZnO and LTCA.

Data availability

The data supporting this article have been included as part of the ESI.† Data are also available upon request from the authors.

Author contributions

T. H. and K. D. designed and supervised the research. N. I., H. S. and S. N. carried out the experiments. N. I., H. S., S. N., M. N., T. T., T. H. and K. D. discussed the results. N. I., S. N., T. H. and K. D. wrote the manuscript with contributions from the other authors.

Conflicts of interest

The authors declare no competing financial interests.

Acknowledgements

This research was supported by JST, PRESTO, Japan (grant no. JPMJPR20T9), the Artificial Photosynthesis Project (ARPChem) of the New Energy and Industrial Technology Development Organization (NEDO), Advanced Research Infrastructure for Materials and Nano-technology in Japan (ARIM)” of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan (Grant Number JPMXP1223UT0004), research equipment shared in MEXT Project for promoting public utilization of advanced research infrastructure (program for supporting construction of core facilities, grant number JPMXS0441000024), and Mitsubishi Chemical Corporation. The authors thank Ms. Keiko Kato of The University of Tokyo and Ms. Michiko Obata of Shinshu university for their assistance with ICP-AES and XPS measurements.

Notes and references

1. M. R. Shaner, H. A. Atwater, N. S. Lewis and E. W. McFarland, A comparative technoeconomic analysis of renewable hydrogen production using solar energy, Energy Environ. Sci., 2016, 9, 2354–2371.
2. S. Nandy, S. A. Savant and S. Haussener, Prospects and challenges in designing photocatalytic particle suspension reactors for solar fuel processing, Chem. Sci., 2021, 12, 9866–9884.
3. Q. Wang, C. Pornrungroj, S. Linley and E. Reisner, Strategies to improve light utilization in solar fuel synthesis, Nat. Energy, 2022, 7, 13–24.
4. H. Nishiyama, T. Yamada, M. Nakabayashi, Y. Maehara, M. Yamaguchi, Y. Kuromiya, Y. Nagatsuma, H. Tokudome, S. Akiyama, T. Watanabe, R. Narushima, S. Okunaka, N. Shibata, T. Takata, T. Hisatomi and K. Domen, Photocatalytic solar hydrogen production from water on a 100-m² scale, Nature, 2021, 598, 304–307.
5. T. Hisatomi and K. Domen, Reaction systems for solar hydrogen production via water splitting with particulate semiconductor photocatalysts, Nat. Catal., 2019, 2, 387–399.
6. Z. Wang, Y. Inoue, T. Hisatomi, R. Ishikawa, Q. Wang, T. Takata, S. Chen, N. Shibata, Y. Ikuhara and K. Domen, Overall water splitting by Ta₃N₅ nanorod single crystals grown on the edges of KTaO₃ particles, Nat. Catal., 2018, 1, 756–763.
7. K. Chen, J. Xiao, J. J. M. Vequizo, T. Hisatomi, Y. Ma, M. Nakabayashi, T. Takata, A. Yamakata, N. Shibata and K. Domen, Overall Water Splitting by a SrTaO₂N-Based Photocatalyst Decorated with an Ir-Promoted Ru-Based Cocatalyst, J. Am. Chem. Soc., 2023, 145, 3839–3843.
8. Q. Wang, M. Nakabayashi, T. Hisatomi, S. Sun, S. Akiyama, Z. Wang, Z. Pan, X. Xiao, T. Watanabe, T. Yamada, N. Shibata, T. Takata and K. Domen, Oxysulfide photocatalyst for visible-light-driven overall water splitting, Nat. Mater., 2019, 18, 827–832.
9. R. Asai, H. Nemoto, Q. Jia, K. Saito, A. Iwase and A. Kudo, A visible light responsive rhodium and antimony-codoped SrTiO₃ powdered photocatalyst loaded with an IrO₂ cocatalyst for solar water splitting, Chem. Commun., 2014, 50, 2543–2546.
10. K. Maeda, T. Takata, M. Hara, N. Saito, Y. Inoue, H. Kobayashi and K. Domen, GaN:ZnO Solid Solution as a Photocatalyst for Visible-Light-Driven Overall Water Splitting, J. Am. Chem. Soc., 2005, 127, 8286–8287.
11. Y. Li, L. Zhu, Y. Yang, H. Song, Z. Lou, Y. Guo and Z. Ye, A Full Compositional Range for a (Ga_1-xZn_x)(N_1-xO_x) Nanostructure: High Efficiency for Overall Water Splitting and Optical Properties, Small, 2015, 11, 871–876.
12. K. Maeda and K. Domen, Photocatalytic Water Splitting: Recent Progress and Future Challenges, J. Phys. Chem. Lett., 2010, 1, 2655–2661.
13. K. Liu, B. Zhang, J. Zhang, W. Lin, J. Wang, Y. Xu, Y. Xiang, T. Hisatomi, K. Domen and G. Ma, Synthesis of Narrow-Band-Gap GaN:ZnO Solid Solution for Photocatalytic Overall Water Splitting, ACS Catal., 2022, 12, 14637–14646.
14. R. Abe, Development of a New System for Photocatalytic Water Splitting into H₂ and O₂ under Visible Light Irradiation, Bull. Chem. Soc. Jpn., 2011, 84, 1000–1030.
15. H. Kato, Y. Sasaki, N. Shirakura and A. Kudo, Synthesis of highly active rhodium-doped SrTiO₃ powders in Z-scheme systems for visible-light-driven photocatalytic overall water splitting, J. Mater. Chem. A, 2013, 1, 12327–12333.
16. R. Konta, T. Ishii, H. Kato and A. Kudo, Photocatalytic Activities of Noble Metal Ion Doped SrTiO₃ under Visible Light Irradiation, J. Phys. Chem. B, 2004, 108, 8992–8995.
17. N. Iwasa, Z. Teng, G. Ma, T. Hisatomi and K. Domen, Synthesis of narrow bandgap gallium zinc nitride oxide solid solutions for photocatalytic water splitting under visible light, Chem. Mater., 2024, 36, 2917–2924.
18. K. Maeda and K. Domen, Solid Solution of GaN and ZnO as a Stable Photocatalyst for Overall Water Splitting under Visible Light, Chem. Mater., 2010, 22, 612–623.
19. A. Sharma, M. Varshney, H. Saraswat, S. Chaudhary, J. Parkash, H. J. Shin, K. H. Chae and S. O. Won, Nano-structured phases of gallium oxide (GaOOH, α-Ga₂O₃, β-Ga₂O₃, γ-Ga₂O₃, δ-Ga₂O₃, and ε-Ga₂O₃): fabrication, structural, and electronic structure investigations, Int. Nano Lett., 2020, 10, 71–79.
20. X. Liu, C. Wessel, F. Pan and R. Dronskowski, Synthesis and single-crystal structure determination of the zinc nitride halides Zn₂NX (X=Cl, Br, I), J. Solid State Chem., 2013, 203, 31–36.
21. Z. Li, L. Zhang, Y. Liu, C. Shao, Y. Gao, F. Fan, J. Wang, J. Li, J. Yan, R. Li and C. Li, Surface-Polarity-Induced Spatial Charge Separation Boosts Photocatalytic Overall Water Splitting on GaN Nanorod Arrays, Angew. Chem., Int. Ed., 2020, 59, 935–942.
22. M. Huang, J. Lian, R. Si, L. Wang, X. Pan and P. Liu, Spatial Separation of Electrons and Holes among ZnO Polar {0001} and {101̅0} Facets for Enhanced Photocatalytic Performance, ACS Omega, 2022, 7, 26844–26852.
23. Y. Chen, H. Zhao, B. Liu and H. Yang, Charge separation between wurtzite ZnO polar {001} surfaces and their enhanced photocatalytic activity, Appl. Catal., B, 2015, 163, 189–197.
24. M. W. Liao, H. T. Jeng and T. P. Perng, Formation Mechanism and Bandgap Reduction of GaN–ZnO Solid-Solution Thin Films Fabricated by Nanolamination of Atomic Layer Deposition, Adv. Mater., 2023, 35, 2207849.
25. K. Maeda, K. Teramura and K. Domen, Effect of post-calcination on photocatalytic activity of (Ga_1−xZn_x)(N_1−xO_x) solid solution for overall water splitting under visible light, J. Catal., 2008, 254, 198–204.
26. S. Nandy, T. Hisatomi, M. Nakabayashi, H. Li, X. Wang, N. Shibata, T. Takata and K. Domen, Oxide layer coating enabling oxysulfide-based photocatalyst sheet to drive Z-scheme water splitting at atmospheric pressure, Joule, 2023, 7, 1641–1651.
27. Q. Wang, Y. Li, T. Hisatomi, M. Nakabayashi, N. Shibata, J. Kubota and K. Domen, Z-scheme water splitting using particulate semiconductors immobilized onto metal layers for efficient electron relay, J. Catal., 2015, 328, 308–315.
28. T. Minegishi, N. Nishimura, J. Kubota and K. Domen, Photoelectrochemical properties of LaTiO₂N electrodes prepared by particle transfer for sunlight-driven water splitting, Chem. Sci., 2013, 4, 1120–1124.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ta03576c

‡ These authors have contributed equally.

---