Water splitting into H₂ and O₂ over niobate and titanate photocatalysts with (111) plane-type layered perovskite structure

Abstract

Photophysical and photocatalytic properties of A₅Nb₄O₁₅ (A = Sr and Ba), Ba₃LaNb₃O₁₂, ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba), and La₄Ti₃O₁₂ with layered perovskite structures, in which a plane in parallel with (111) of a simple perovskite structure was exposed at interlayer, were investigated. These oxides were obtained by a polymerizable complex method at 973–1473 K though only A₅Nb₄O₁₅ (A = Sr and Ba) were prepared by a solid state reaction even at 1673 K. The shapes of these complex metal oxides were plate-like derived from the perovskite layered structure. These band gaps were estimated to be 3.7–4.1 eV from the onsets of diffuse reflection spectra. These oxides showed photoluminescence at 77 K. These oxides loaded with NiO cocatalysts showed activities for water splitting under UV irradiation. NiO_x/BaLa₄Ti₄O₁₅ and NiO_x/Ba₅Nb₄O₁₅ showed the highest activities among the titanates and niobates tested in the present study. NiO_x/BaLa₄Ti₄O₁₅ and NiO_x/Ba₅Nb₄O₁₅ gave 15% and 17% of quantum yields at 270 nm, respectively. Photocatalytic activities of ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba) strongly depended on the alkaline earth metal ion. Pt, Au, Ni, and PbO₂ were selectively photodeposited on basal or edge plane of the BaLa₄Ti₄O₁₅ plate-like powder while these were randomly loaded on CaLa₄Ti₄O₁₅. It was suggested that this difference in the surface property was the one of the important factors affecting photocatalytic ability for ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba).

---

Broader context

There is no doubt that solar hydrogen production from water is highly desirable. Water splitting using a particulate photocatalyst is one of the candidates for the solar hydrogen production. The photocatalyst system is not at the stage of practical use. Therefore, basic research is still important in this research field. The construction of a photocatalyst library and clarification of factors dominating photocatalytic activities are important in order to establish a guiding principle of photocatalyst design. In the present study, A₅Nb₄O₁₅ (A = Sr, Ba), Ba₃LaNb₃O₁₂, ALa₄Ti₄O₁₅ (A = Ca, Sr, Ba), and La₄Ti₃O₁₂ with a (111) plane-type layered perovskite structure in which alkaline earth metal and lanthanum cations are ordered or not at interlayer and A-site were found to be new niobate and titanate photocatalyst materials for water splitting. 15% of the quantum yield was obtained. The ordering caused anisotropic characters of photocatalytic properties that were revealed by DRS, SEM, photodeposition of metals, and photoluminescence. The relationship between photocatalytic activities, and crystal and energy structures are systematically discussed in detail. These results and discussion will be valuable in this research field of not only water splitting but also photocatalytic degradation of hazardous organic compounds.

1. Introduction

Photocatalytic water splitting using a heterogeneous semiconductor powder has been studied as an ideal H₂ production method which does not give any burdens to the environment. Many oxide photocatalysts consisting of d⁰ and d¹⁰ metal ions show photocatalytic activities for water splitting under UV irradiation.^1–35 Each material possesses unique character even if they seem similar to each other. So, finding a new material for the attractive and tough reaction, and studying its science is meaningful. Recently, metal oxynitride and nitride photocatalysts,^36,37 and Z-scheme-type photocatalysts accompanied with two-photon process^38–42 have been reported for water splitting under visible light irradiation. The final target of this research is efficient solar hydrogen production from water. The present efficiency is not satisfying for practical use. Photocatalytic water splitting is an attractive and challenging reaction. Therefore, basic research is still important in this research field. The construction of a photocatalyst library and clarification of factors affecting photocatalytic abilities are important in order to establish a guiding principle for design of a photocatalyst.⁴

NiO/NaTaO₃ photocatalyst shows high activity for water splitting under UV irradiation.⁵ Moreover, the activity of NiO/NaTaO₃ is remarkably improved by La doping (1–2%).⁶ La-doping gives nano-sized NaTaO₃ particles and characteristic surface step structure. Active sites for H₂ evolution are separated from those for O₂ evolution by the surface nano-step structure. So, the surface morphology is important. On the other hand, the metal oxides with layered structures, such as K₄Nb₆O₁₇,⁷ K₂La₂Ti₃O₁₀,⁸ A₂A′₂O₇ (A = Ca and Sr, A′ = Nb and Ta),^12,14,30 RbNdTa₂O₇,¹⁵ La₂Ti₂O₇,¹⁹ H₂La_2/3Ta₂O₇,²³ K₂Sr_1.5Ta₃O₁₀,³² H_1.8Sr_0.81Bi_0.19Ta₂O₇,³³ and LiCa₂Ta₃O₁₀³⁴ show high activities for water splitting under UV irradiation. The interlayer can function as reaction sites and separate oxidation reaction sites from reduction reaction sites in hydrated layered compounds. The anisotropy derived from the layered structure may also affect the photocatalytic ability. Therefore, it is important to examine photocatalytic properties of various compounds with surface morphology and characteristic crystal structure like a layered structure in order to obtain further information on photocatalytic water splitting.

There are many kinds of compounds with layered perovskite structure consisting of various elements, and perovskite framework and interlayer structures. Layered perovskite compounds are attractive materials as photocatalysts. Layered perovskite structure can be classified into (100) and (110) plane structures due to the plane directions of interlayer. The (110) plane structure can be classified further into (110) and (111) plane-type structures due to the plane defect structure of perovskite crystal structure. In the present paper, layered perovskite structures are classified into (100), (110), and (111) plane-type structures due to the plane defect structure (Fig. 1). Here, (100), (110), and (111) plane-types mean that the interlayer is in parallel with (100), (110), and (111) planes of a bulk type perovskite structure, respectively. These structures can also be distinguished from each other by the exposed part of MO₆ octahedra in interlayer, corner, edge, and face of the octahedra. It has been reported that K₂La₂Ti₃O₁₀,⁸ Sr₄Ti₃O₁₀,¹⁶ A₂A′Ta₂O₇ (A = H and K, A′ = La_2/3 and Sr),^18,23 La₂Ti₃O₉,¹⁹ Sr₃Ti₂O₇,²⁹ and LiCa₂Ta₃O₁₀³⁴ with the (100) plane-type layered perovskite structure and La₄CaTi₅O₁₇,¹² A₂A′₂O₇ (A = Ca and Sr, A′ = Nb and Ta),^12,14,30 La₂Ti₂O₇,¹⁹ and H_1.81Sr_0.81Bi_0.19Ta₂O₇³³ with the (110) plane-type layered perovskite structure show photocatalytic activities for water splitting under UV irradiation. In addition, the modified photocatalysts of these layered perovskite compounds have been also reported.^22,35 On the other hand, there are few reports of oxide photocatalysts with the (111) plane-type layered perovskite structure.^21,24 Moreover, a titanate photocatalyst with the (111) plane-type layered perovskite structure is not reported.

Fig. 1 Layered perovskite structures with different face defects.

A polymerizable complex method (PC method) is one of the preparation methods of functional material powders.⁴³ Various complex metal oxides can be prepared under mild condition by the PC method. We can prepare complex oxides with different crystallinity, particle size, and morphology by changing calcination temperature in a wide range. Investigation using these samples with a series of different conditions is meaningful for clarification of the factors affecting photocatalytic activities. It has been reported that the particles of K₂La₂Ti₃O₁₀⁸ and KTiNbO₅¹³ prepared by a PC method have higher crystallinity and finer crystal size than those prepared by a solid-state reaction, resulting in that these photocatalytic activities remarkably increase. We have preliminary reports that the particles of Ba₅Nb₄O₁₅ with (111) plane-type layered perovskite structure prepared by the PC method were plate-like crystals, and showed high activity for water splitting under UV irradiation.²⁶ On the other hand, A₅Ta₄O₁₅ (A = Sr and Ba) with (111) plane-type layered perovskite structure prepared by a PC method also shows high activity for water splitting under UV irradiation.^21,24 Thus, (111) plane-type layered perovskite oxides with high crystallinity and fine particle size can be prepared by a PC method.

In the present paper, we investigated photophysical properties and photocatalytic activities for water splitting of (111) plane–type A₅Nb₄O₁₅ (A = Sr and Ba), Ba₃LaNb₃O₁₂, ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba), and La₄Ti₃O₁₂ prepared by a PC method. The main characteristic point of the materials reported in the present study is the new-type layered perovskite structure in which alkaline earth metal and lanthanum cations are ordered or not at interlayer and A-site. Diffuse reflectance spectroscopy (DRS), photoluminescence, scanning electron microscope (SEM) measurements, and photodeposition of some metal and metal oxides were carried out as characterizations in order to investigate the properties based on the anisotropy of the layered structure. The relationship between photocatalytic activities, and crystal and energy structures are discussed based on the different ordering of alkaline earth metal and lanthanum cations.

2. Experimental

CaCO₃ (Kanto Chemical; 99%), SrCO₃ (Kanto Chemical; 99.9%), BaCO₃ (Kanto Chemical; 99%), La(NO₃)₃·6H₂O (Wako Pure Chemical; 99.9%), Ti(OC₄H₉)₄ (Kanto Chemical; 97%), Nb(OC₂H₅)₅ (Kojundo Chemical; 99.99%), PrCl₃·6H₂O (Rare Metallic; 99.99%), citric acid (Sigma Aldrich Japan; 99.5%), ethylene glycol (Kanto Chemical; 99.5%), and propylene glycol (Kanto Chemical; 99.0%) were employed as starting materials for PC method according to a previous report.⁴³ Metal compounds and a citric acid were dissolved in a mixed solvent of methanol and ethylene glycol or propylene glycol. The mixed solution was aged at 353 K for 2 h. At this stage, a solution containing citric complexes was formed. The solution was dehydrated to form gel by polymerization at 393 K for 12 h. A precursor was obtained by calcination of the gel containing a citric acid using a mantle heater. Various titanates and niobates were prepared by calcination of these precursors at 1073–1473 K for 10 h. Phase purity of the obtained powders was confirmed by X-ray diffraction (Rigaku; MiniFlex). Diffuse reflection spectra were measured using a UV-vis-NIR spectrometer (Jasco; UbestV-570) and were converted from reflectance to absorbance by the Kubellka-Munk method. Photoluminescence spectra of these complex metal oxides with and without Pr-doping as a luminescent probe were measured using a fluorospectrometer (HORIBA JOBIN YVON; SPEX Fluorolog-3). Temperature control of the photoluminescence measurement was carried out using a cryostat (Oxford; ITC 503S). The photocatalyst particles were observed by a scanning electron microscope (SEM, JEOL JSM-6700F). In order to investigate the surface reactivity of photocatalysts, Au, Pt, Ni, and PbO₂ particles photodeposited on photocatalysts from aqueous HAuCl₄, K₂PtCl₆, [Pt(NH₃)₄]Cl₂, Ni(NO₃)₂, and Pb(NO₃)₂ solutions were observed by SEM.

NiO and RuO₂ cocatalysts were loaded by an impregnation method from aqueous Ni(NO₃)₂ and RuCl₃ solutions. The powder impregnated with Ni(NO₃)₂ was calcined at 543 K for 1 h in air. Pretreatment of reduction with 200 torr of H₂ at 773 K for 2 h followed by oxidation with 100 torr of O₂ at 373–573 K for 1 h was carried out for NiO-loaded photocatalysts, if necessary.⁴⁴ The powder impregnated with RuCl₃ was calcined at 673 K for 2 h in air. Water splitting reactions were carried out in a gas-closed circulation system. The photocatalyst powder (0.5 g) was dispersed in pure water (380 mL) by a magnetic stirrer in an inner irradiation reaction cell made of quartz equipped with a 400 W high-pressure mercury lamp (SEN: HL400EH-5). The amounts of evolved H₂ and O₂ were determined using on-line gas chromatography (Shimadzu; MS-5A column, TCD, Ar carrier). Apparent quantum yields at 270 nm were measured using a reaction cell with a top window made of quartz and a 300 W Xe illuminator (ILC technology; CERMAX-LX300) attached with a band-pass filter (Kenko; BPUV-270, WHM: 16 nm). The number of incident photon was determined using a photodiode (OPHIRA: PD300-UV of a head and NOVA of a power monitor). The photocatalyst powder (0.05–0.3 g) was dispersed in pure water (300 mL) for the measurement of the apparent quantum yield. The apparent quantum yields were determined by eqn (1).

3. Results and discussion

3-1. Synthesis and characterization

A single phase of A₅Nb₄O₁₅ (A = Sr and Ba) was obtained by a solid-state reaction (SSR method) at 1373 K. In construct, single phases of Ba₃LaNb₃O₁₂, ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba), and La₄Ti₃O₁₂ were not obtained even at 1673 K of the calcination temperature. It is probably caused by low reactivity of lanthanum oxide with alkali earth metal oxides. Therefore, synthesis of these compounds was attempted by a polymerizable complex method (PC method). These pure complex metal oxides were successfully prepared by calcination of the precursors at 973–1473 K. The synthetic processes by the PC method for Ba₅Nb₄O₁₅ and BaLa₄Ti₄O₁₅ that showed high photocatalytic activities as mentioned later were examined in detail.

Fig. 2 shows XRD patterns of Ba₅Nb₄O₁₅ and BaLa₄Ti₄O₁₅ prepared by a PC method. Broad diffraction patterns of both precursors were assigned to a simple perovskite structure. Almost pure phases of Ba₅Nb₄O₁₅ and BaLa₄Ti₄O₁₅ were successfully obtained by calcination at 973 and 1323 K, respectively. The diffraction peaks of (103) and (203) of BaLa₄Ti₄O₁₅ and Ba₅Nb₄O₁₅ became larger than that of (110), as the calcination temperature became high. Fig. 3 shows SEM images of Ba₅Nb₄O₁₅ prepared at different calcination temperatures. The shape of particles was plate-like derived from the layered perovskite structure. As the calcination temperature was high, the aspect ratio of particles was decreased by crystal growth to the direction in a plane parallel to the (−110) plane. This is the reason why the (110) diffraction peak became relatively small with increasing the calcination temperature. La₄Ti₃O₁₂, Ba₃LaNb₃O₁₂, Sr₅Nb₄O₁₅, and ALa₄Ti₄O₁₅ (A = Sr and Ca) were also obtained as plate-like particles.

Fig. 2 XRD patterns of (I) Ba₅Nb₄O₁₅ and (II) BaLa₄Ti₄O₁₅ prepared by a PC method. (a) and (b) precursor, powders calcined at (c) 873 K, (d) 973 K, (e) 1073 K, (f) 1273 K, (g) 1173 K, (h) 1273 K, and (i) 1323 K. Calcination time was 10 h.

Fig. 3 SEM images of Ba₅Nb₄O₁₅ powders synthesized at (a) 1073 K and (b) 1273 K.

3-2. Crystal and energy structures

Factors affecting the band gaps of layered perovskite oxides are considered as follows.

Factor (i): the bond angle between metal and oxygen ions of octahedra units (M–O–M). As the bond angle is apart from 180°, the band gap becomes wide and the excited energy state is localized.^5,14,45,46

Factor (ii): the thickness of the perovskite layer. As the thickness of the perovskite layer decreases, the two dimensionality of crystal structure becomes high, resulting in the band gap becoming wide and the excited energy state more localized.

Factor (iii): the interaction between perovskite layers. As the ionic radius of cations at the interlayer becomes large, the distance between perovskite layers becomes long, accordingly an excited energy state becomes localized by decreasing interaction between the layers.

Factor (iv): the polarization ability of cations at the interlayer towards the oxygen ions of octahedra faced at the interlayer. For example, H⁺-exchange of layered oxides sometimes causes red shift in absorption spectra due to high polarization ability of proton showing band gap narrowing.^47,48 The relationships between these factors and band gaps are summarized in Table 1.

Table 1 The relationships between band gaps and the factors affecting the band gaps

Factors	Niobate		Titanate
	Sr5Nb4O15	Ba5Nb4O15	CaLa4Ti4O15	BaLa4Ti4O15
Cation of interlayer	Sr^2+	Ba^2+	La^3+, Ca^2+	La^3+, Ba^2+
Ionic radius/Å	1.44	1.61	1.36, 1.34	1.36, 1.61
Cation of perovskite slab	Sr^2+	Ba^2+	La^3+, Ca^2+	La^3+
Ionic radius/Å	1.44	1.61	1.36, 1.34	1.36
Bond angle between metal and oxygen ions	168	180	Similar
Effect on band gap	Wider	Narrower	Similar
Thickness of perovskite layer	Four	Four	Four	Four
Effect on band gap	Similar		Similar
Interaction between perovskite layers	Large	Small	Large	Small
Effect on band gap	Narrower	Wider	Narrower	Wider
Polarization ability of cation	Large	Small	Large	Small
Effect on band gap	Narrower	Wider	Narrower	Wider
Band gaps/eV	4.04	3.91	3.79	3.85

---

Fig. 4 shows diffuse reflection spectra of A₅Nb₄O₁₅ (A = Sr and Ba), Ba₃LaNb₃O₁₂, ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba), and La₄Ti₃O₁₂. The band gaps of Sr₅Nb₄O₁₅ and Ba₅Nb₄O₁₅ with four-octahedra thickness of perovskite layer were estimated to be 4.0 eV and 3.9 eV from the onsets of absorption, respectively (Fig. 4(I)). The bond angles of Nb–O–Nb in the frameworks of perovskite structures of Ba₅Nb₄O₁₅ and Sr₅Nb₄O₁₅ are 180 and 168°, respectively.^49,50 Because the bond angle of Ba₅Nb₄O₁₅ is closer to 180° than that of Sr₅Nb₄O₁₅, the band gap of Ba₅Nb₄O₁₅ is narrower than that of Sr₅Nb₄O₁₅ according to the factor (i). The difference between bond angles of Ba₅Nb₄O₁₅ and Sr₅Nb₄O₁₅ is almost the same as those of KTaO₃ and NaTaO₃ with bulk-type perovskite structures, respectively. However, the difference in the band gap between Ba₅Nb₄O₁₅ and Sr₅Nb₄O₁₅ (0.1 eV) was smaller than that between KTaO₃ and NaTaO₃ (0.4 eV).⁵ This result implies that the factors (iii) and (iv) also affect the band gaps. The interaction between perovskite layers of Ba₅Nb₄O₁₅ would be smaller than that of Sr₅Nb₄O₁₅ because Ba²⁺ at the interlayer is larger than Sr²⁺. In addition, the energy structures of the perovskite layers are changed by different polarization abilities between Sr²⁺ and Ba²⁺ at the interlayer. The balance of these factors seems to cause the small difference in band gaps between Ba₅Nb₄O₁₅ and Sr₅Nb₄O₁₅.

Fig. 4 Diffuse reflection spectra of (I) niobates and (II) titanates; (a) Ba₃LaNb₃O₁₂, (b) Sr₅Nb₄O₁₅, (c) Ba₅Nb₄O₁₅, (d) CaLa₄Ti₄O₁₅, (e) SrLa₄Ti₄O₁₅, (f) BaLa₄Ti₄O₁₅, and (g) La₄Ti₃O₁₂.

The band gaps of CaLa₄Ti₄O₁₅, SrLa₄Ti₄O₁₅, and BaLa₄Ti₄O₁₅ were 3.78 eV, 3.82 eV, and 3.85 eV, respectively. This order in the band gap was reverse to that for ATiO₃ (A = Ca, Sr, and Ba) with bulk-type perovskite structure. The positions of La³⁺, Ca²⁺, and Ba²⁺ in crystal structures of ALa₄Ti₄O₁₅ (A = Ca and Ba) give some information to consider this difference in the order of band gaps. The ratio of Ba²⁺ to La³⁺ at the interlayer is 1 : 1 in BaLa₄Ti₄O₁₅ while only La³⁺ exists in the perovskite layer as shown in Fig. 5.⁵¹ On the other hand, Ca²⁺ and La³⁺ randomly exist at the interlayer and perovskite layer in CaLa₄Ti₄O₁₅. The distortion of the perovskite framework structure of BaLa₄Ti₄O₁₅ would be close to that of CaLa₄Ti₄O₁₅ because the ionic radius of La³⁺ (1.36 Å) is nearly equal to that of Ca²⁺ (1.34 Å). Therefore, the factor (i) would be negligible. The band gap of BaLa₄Ti₄O₁₅ is wider than that of CaLa₄Ti₄O₁₅ due to an increase in the distance between layers (factor (iii)) and the different polarization ability of Ca²⁺ and Ba²⁺ (factor (iv)) because the ionic radius of Ba²⁺ is larger than that of Ca²⁺. The band gap of Ba₃LaNb₃O₁₂ with layered perovskite structure with three-octahedra thickness was estimated to be 4.1 eV. The perovskite layer of Ba₃LaNb₃O₁₂ is thinner than that of Ba₅Nb₄O₁₅. Therefore, the band gap of Ba₃LaNb₃O₁₂ was wider than that of Ba₅Nb₄O₁₅ according to the factor (ii). The band gap of La₄Ti₃O₁₂ with layered perovskite structure with three-octahedra thickness was also wider than that of ALa₄Ti₄O₁₅ with layered perovskite structure with four-octahedra thickness.

Fig. 5 Crystal structures of ALa₄Ti₄O₁₅ (A = Ca and Ba) with layered perovskite structure.⁵⁰

Ba₅Nb₄O₁₅ shows a broad yellow emission with maximum at 580 nm.⁵² The photoluminescent properties of A₅Nb₄O₁₅ (A = Sr and Ba), Ba₃LaNb₃O₁₂, ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba), and La₄Ti₃O₁₂ were measured as shown in Fig. 6. These materials did not show any emission at room temperature, but did so at 77 K. The onsets of excitation spectra of these materials agreed with those of absorption spectra, indicating that these emissions were due to the band gap excitation.

Fig. 6 Excitation and emission spectra of (I) niobates and (II) titanates; (a) Sr₅Nb₄O₁₅, (b) Ba₅Nb₄O₁₅, (c) CaLa₄Ti₄O₁₅, (d) SrLa₄Ti₄O₁₅, and (e) BaLa₄Ti₄O₁₅ at 77 K.

Information on the migration of excited energy in the layered oxide photocatalysts can be obtained by monitoring the emission of some doped lanthanide cations as guests.^53–55 Photoluminescent properties of ALa₄Ti₄O₁₅ (A = Ca and Ba) doped with 1% of Pr³⁺ were examined as shown in Fig. 7. Pr³⁺ would be doped in the site of La³⁺ at an A-site of the perovskite layer and interlayer, judging from the ionic radius and electric charge of Pr³⁺. Luminescence of the doped Pr³⁺ was observed for ALa₄Ti₄O₁₅:Pr(1%) (A = Ca and Ba) by the host excitation at room temperature. This result indicates that the photo-generated carrier or excited energy in the host material can migrate to the Pr³⁺ guest. The ratio of an excitation band of a host around 280 nm to that of a Pr³⁺ guest at 449.7 nm for BaLa₄Ti₄O₁₅ was larger than that for CaLa₄Ti₄O₁₅. It suggests that excited energy in the perovskite framework migrates to the Pr³⁺ guest more easily for BaLa₄Ti₄O₁₅ than CaLa₄Ti₄O₁₅. It may also be due to the amount of Pr³⁺ in the perovskite framework of BaLa₄Ti₄O₁₅ being greater than that in CaLa₄Ti₄O₁₅. Reported niobate and titanate photocatalysts for water splitting show photoluminescence at 77 K.^20,53–55 Moreover, the photo-generated carrier or excited energy can migrate in the perovskite layer. Therefore, A₅Nb₄O₁₅ (A = Sr and Ba), Ba₃LaNb₃O₁₂, ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba), and La₄Ti₃O₁₂ were expected to work as photocatalysts for water splitting.

Fig. 7 Diffuse reflection spectra and photoluminescence spectra of (I) BaLa₄Ti₄O₁₅:Pr(0.1%) and (II) CaLa₄Ti₄O₁₅:Pr(0.1%) at room temperature. (a) and (b): the diffuse reflection spectra, (c) and (d): the excitation spectra were obtained by monitoring luminescence of Pr³⁺ at 606 nm, (e) and (f): the emission spectra were obtained by host excitation at 280 nm.

3-3. Photocatalytic water splitting

Table 2 shows photocatalytic activities of A₅Nb₄O₁₅ (A = Sr and Ba), Ba₃LaNb₃O₁₂, ALa₄Ti₄O₁₅(A = Ca, Sr and Ba), and La₄Ti₃O₁₂. The amounts of NiO-loaded and pretreatment conditions were optimized for the photocatalytic water splitting over these photocatalysts. The activities of these photocatalysts without cocatalysts were low. However, when NiO cocatalysts were loaded, the activities were improved more than two orders of magnitude.

Table 2 Water splitting over various metal oxides with layered perovskite structures^a

Photocatalyst	Preparation condition	Surface area	Band gap	NiO loaded	Pretreatment^b	Activity/µmol h^−1
		/m^2 g^−1	/eV	(wt%)		H2	O2
a Catalyst: 0.5 g, pure water: 380 mL, inner irradiation cell made of quartz, 400 W high-pressure mercury lamp. b R and O indicate temperatures in K for H2reduction for 2 h and O2oxidation for 1 h, respectively.
La4Ti3O12	1473 K, 10 h	3.4	3.95	None	No	9	0
La4Ti3O12	1473 K, 10 h	3.4	3.95	0.6	No	357	179
La4Ti3O12	1473 K, 10 h	3.4	3.95	0.6	R673–O373	330	150
CaLa4Ti4O15	1373 K, 10 h	6.2	3.79	None	No	3	0
CaLa4Ti4O15	1373 K, 10 h	6.2	3.79	0.2	No	11	4
CaLa4Ti4O15	1373 K, 10 h	6.2	3.79	0.2	R673–O373	593	276
SrLa4Ti4O15	1373 K, 10 h	6.2	3.82	None	No	2	0
SrLa4Ti4O15	1373 K, 10 h	6.2	3.82	0.7	No	23	11
SrLa4Ti4O15	1373 K, 10 h	6.2	3.82	0.7	R673–O373	1171	546
BaLa4Ti4O15	1323 K, 10 h	6.1	3.85	None	No	5	2
BaLa4Ti4O15	1323 K, 10 h	6.1	3.85	0.5	No	118	62
BaLa4Ti4O15	1323 K, 10 h	6.1	3.85	0.5	R673–O373	2300	1154
Ba3LaNb3O12	1073 K, 10 h	5.8	4.07	None	No	2	0
Ba3LaNb3O12	1073 K, 10 h	5.8	4.07	0.5	R773–O473	1185	588
Sr5Nb4O15	1073 K, 10 h	7.9	4.04	None	No	3	0
Sr5Nb4O15	1073 K, 10 h	7.9	4.04	0.7	R773–O473	2200	1100
Ba5Nb4O15	1073 K, 10 h	7.2	3.91	None	No	3	0
Ba5Nb4O15	1073 K, 10 h	7.2	3.91	0.5	R773–O473	4021	1972
Ba5Nb4O15	1273 K, 5 h	1.9	3.87	0.1	R773–O473	939	453

---

Fig. 8 shows the band structures of these titanate photocatalysts and a NiO cocatalyst.⁵⁶ The band positions were estimated by the band gaps and the valance band maxima that were supposed to be +3.0 eV.⁵⁷ Photogenerated electrons in a conduction band of a host material must transfer to a NiO cocatalyst that is a hydrogen evolution site. Therefore, it is required that the conduction band level of a photocatalyst is higher than that of a NiO cocatalyst. If the band level of photocatalyst does not satisfy the requirement, the activation pretreatment of H₂reduction and subsequent O₂oxidation is usually needed.⁴⁴ The activation pretreatment for the NiO cocatalyst was effective for all photocatalysts as well as previously reported titanate and niobate photocatalysts.^{7,8,12,14,16,20,29,44,58} On the other hand, La₄Ti₃O₁₂ showed high activity without the activation pretreatment because of the high conduction band level. NiO_x(0.5 wt%)/BaLa₄Ti₄O₁₅ and NiO_x(0.5 wt%)/Ba₅Nb₄O₁₅ showed the highest activities for water splitting among titanates and niobates, respectively. The photocatalytic activity of Ba₅Nb₄O₁₅ calcined at 1073 K was higher than that at 1273 K. The SEM images indicated that Ba₅Nb₄O₁₅ calcined at 1073 K had higher aspect ratio and smaller particle size than that at 1273 K as shown in Fig. 3. In the case of layered perovskite photocatalyst of which the interlayer space is not hydrated, the photogenerated e⁻ and h⁺ have to reach the surface to react with water molecules. Therefore, the photocatalytic activity is improved by the preparation giving high crystallinity and fine particle. Charge separation of photo-generated e⁻ and h⁺ might be enhanced by high anisotropy of the crystal structure.¹⁴ These factors mean that the photocatalytic activity of Ba₅Nb₄O₁₅ that had been calcined at 1073 K was higher than that treated at 1273 K.

Fig. 8 Band structures of various titanate photocatalysts.

Table 3 shows the photocatalytic activities of BaLa₄Ti₄O₁₅ loaded with various metals and oxides. NiO_x/BaLa₄Ti₄O₁₅ obtained by the impregnation method followed by the activation pretreatment showed the highest activity for water splitting. The activities of Ni/BaLa₄Ti₄O₁₅ obtained by photo-deposition method and NiO_x/BaLa₄Ti₄O₁₅ prepared by photo-deposition method followed by oxidation pretreatment were low compared with that of NiO_x/BaLa₄Ti₄O₁₅ prepared by the impregnation method. Metallic Au and RuO₂ cocatalysts were also effective for BaLa₄Ti₄O₁₅.

Table 3 Water splitting over BaLa₄Ti₄O₁₅ photocatalyst loaded with various metals and oxides^a

Cocatalyst (wt%)	Loading method	Pretreatment^b	Activity/µmol h^−1
			H2	O2
a Catalyst: 0.5 g, pure water: 380 mL, inner irradiation cell made of quartz, 400 W high-pressure mercury lamp. b R and O indicate temperatures in K for H2reduction for 2 h and O2oxidation for 1 h, respectively.
None	—	—	5	2
Ni (0.5)	Photo-deposition	None	10	0.5
NiO (0.5)	Impregnation	O543	118	62
NiOx (0.5)	Impregnation	R773–O473	1333	673
NiOx (0.5)	Photo-deposition	R773–O473	432	203
RuO2 (1.0)	Impregnation	O673	575	265
Au (1.0)	Photo-deposition	None	36	10
IrO2 (1.0)	Photo-deposition	None	3	0
Pt (1.0)	Photo-deposition	None	35	1
PbO2 (1.0)	Photo-deposition	None	5	0

---

The time course for water splitting on optimized NiO_x(0.5 wt%)/BaLa₄Ti₄O₁₅ photocatalyst is shown in Fig. 9. At the initial stage, NiO_x(0.5 wt%)/BaLa₄Ti₄O₁₅ produced H₂ and O₂ at the rates of 2.3 and 1.14 mmol h⁻¹ using a 400 W high-pressure mercury lamp, respectively. The turnover number of the amount of reacted electrons/holes to the molar quantity of BaLa₄Ti₄O₁₅ was 30 at 4 h of the reaction time. It clearly indicated that the reaction proceeded photocatalytically. The apparent quantum yield was 15% at 270 nm.

Fig. 9 Photocatalytic water splitting over optimized NiO_x (0.5 wt%)/BaLa₄Ti₄O₁₅. (a): H₂ and (b): O₂. Catalyst: 0.5 g, pure water: 380 mL, inner irradiation cell made of quartz, 400 W high-pressure mercury lamp.

Table 4 summarizes metal oxide photocatalysts with layered perovskite structures for water splitting. Although the detailed experimental condition was not the same for each other, the (111) plane-type layered perovskite oxide photocatalysts found in the present paper showed relatively high activities compared with layered titanate and niobate photocatalysts previously reported.

Table 4 Comparison in photocatalytic activity for water splitting among various metal oxides with layered perovskite type structures^a

Photocatalyst	Interlayer structure type	Cocatalyst	Band gap	Activity/µmol h^−1		Apparent quantum yield	Ref.
			/eV	H2	O2	(%)
a Light source: 400–450 W high-pressure mercury lamp, reaction cell: inner irradiation cell made of quartz, reactant: pure water. NiOx indicates a pretreated NiO cocatalyst.
K2La2Ti3O10	(100)	NiOx	3.4	2186	1131	—	8
Sr3Ti2O7	(100)	NiOx	3.2	144	72	—	29
Sr4Ti3O10	(100)	NiOx	3.3	170		4.5	16
La2Ti3O9	(100)	NiOx		386		—	20
La2Ti2O7	(110)	NiOx	3.8	441	220	12	20
La2Ti2O7:Ba	(110)	NiOx	3.8	5000		50	20
La4CaTi5O17	(110)	NiOx	3.8	499		20	12
BaLa4Ti4O15	(111)	NiOx	3.85	2300	1154	15	This work
KCa2Nb3O10	(100)	RuO2		96	47	—	22
Ca2Nb2O7	(110)	NiOx	4.3	101		7	13
Sr2Nb2O7	(110)	NiOx	4.0	110	36	—	14
Ba5Nb4O15	(111)	NiOx	3.91	4021	1982	17	This work

---

3-4. Difference in photophysical and photocatalytic properties between CaLa₄Ti₄O₁₅ and BaLa₄Ti₄O₁₅

The order of photocatalytic activities of ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba) was Ba > Sr > Ca. This order seems to depend on the ionic radii of cations at interlayer. Photocatalytic properties would depend on the anisotropy of layered perovskite structure. The difference in the framework distortion of perovskite structure between ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba) photocatalysts is not significant for the activity as discussed in section 3-2. The shapes of these particles are similar to each other. Therefore, the difference in photocatalytic activities between ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba) is due to the conduction band level and the anisotropy that are caused by A-site cations as shown in Fig. 5. It has been reported that the reactivity of e⁻ and h⁺ on a TiO₂ surface can be examined by observing on which plane Pt and PbO₂ are photodeposited reductively and oxidatively, respectively.⁵⁸ Therefore, the features of photodeposited metals and PbO₂ were examined to see the effect of the anisotropy as shown in Fig. 10. Metallic Au was selectively deposited from AuCl₄⁻ not on a basal plane but on an edge plane. In contrast, PbO₂ was oxidatively photodeposited from Pb²⁺ on the basal plane. Metallic Ni was reductively photodeposited from Ni²⁺ only on the basal plane with the characteristic whisker-like shape. Metallic Pt was non-selectively photodeposited from [PtCl₆]²⁻ on the whole surface. When metallic Pt was photodeposited from [Pt(NH₃)₄]²⁺, the particle size on the basal plane was smaller than that on the edge plane. In contrast, the plane-selective photodeposition was not observed for CaLa₄Ti₄O₁₅. Moreover, A₅Nb₄O₁₅ (A = Sr and Ba) in which A-site cations is the same as interlayer cations did not show the selective deposition neither. This unique selective photodeposition is due to the difference in adsorption property for ions between edge and basal planes, and the charge separation promoted by the structural anisotropy due to the different distribution of alkali earth metal cations. This character resulted in that BaLa₄Ti₄O₁₅ showed higher activity than CaLa₄Ti₄O₁₅.

Fig. 10 SEM images of various metals and PbO₂ loaded on BaLa₄Ti₄O₁₅ photocatalysts by a photo-deposition method; (a) Au, (b) Ni, (c) PbO₂, (d) Pt from [PtCl₆]²⁻, (e) Pt from [Pt(NH₃)₄]²⁺, and (f) magnification of (e).

The reduction with H₂ at 773 K for 2 h followed by oxidation with O₂ at 473 K for 1 h was carried out for the Ni/BaLa₄Ti₄O₁₅ prepared by the photodeposition. The pretreatment gave NiO_x loaded only on the basal plane of BaLa₄Ti₄O₁₅. The shape of this NiO_x was similar to the NiO_x loaded by an impregnation method. The activity of this NiO_x/BaLa₄Ti₄O₁₅ photocatalyst was lower than that of NiO_x/BaLa₄Ti₄O₁₅ prepared by the impregnation method. It suggests that the NiO cocatalyst loaded on the edge plane was more effective than that on the basal plane.

4. Conclusions

A₅Nb₄O₁₅ (A = Sr and Ba), Ba₃LaNb₃O₁₂, ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba), and La₄Ti₃O₁₂ with a (111) plane-type layered perovskite structure were found to be new niobate and titanate photocatalyst materials for water splitting. A PC method was advantageous for the preparation of these photocatalysts at wide-range temperature. These photocatalysts showed photoluminescence at 77 K. Moreover, energy migration in the layered perovskite structure was confirmed by probing luminescence of doped Pr³⁺. This energy structure revealed by photoluminescent properties is preferable for photocatalysts. BaLa₄Ti₄O₁₅ showed the highest activity for water splitting among ALa₄Ti₄O₁₅ (A = Ca, Sr, and Ba) photocatalysts. Ba²⁺ and La³⁺ exist at interlayers of BaLa₄Ti₄O₁₅ while only La³⁺ exists at an A-site of the perovskite layers. In contrast, Ca²⁺ and La³⁺ randomly exist at the interlayers and the perovskite layers of CaLa₄Ti₄O₁₅. The anisotropic surface reactivity was confirmed by selective photodeposition of Pt, Ni, Au, and PbO₂ for BaLa₄Ti₄O₁₅ but no for CaLa₄Ti₄O₁₅. These results indicated that the structural anisotropy due to the ordered positions of Ba²⁺ and La³⁺ in BaLa₄Ti₄O₁₅ was one of the major factors for the high activity. The application of plate-like BaLa₄Ti₄O₁₅ particles prepared by a PC method to other photocatalytic reactions besides water splitting is expected.

Acknowledgements

This work was supported by Core Research for Evolutional Science and Technology (CREST) of Japan Science and Technology Agency (JST) and a Grant-in-Aid (No. 14 050 090) for the Priority Area Research (No. 417) from the Ministry of Education, Culture, Sports, Science and Technology, Japan.

References

1. K. Domen, J. N. Kondo, M. Hara and T. Takata, Bull. Chem. Soc. Jpn., 2000, 73, 1307 and references therein.
2. H. Kato and A. Kudo, Catal. Today, 2003, 78, 561 and references therein.
3. J. Sato, H. Kobayashi, K. Ikarashi, N. Saito, H. Nishiyama and Y. Inoue, J. Phys. Chem. B, 2004, 108, 4369 and references therein.
4. A. Kudo and Y. Miseki, Chem. Soc. Rev., 2009, 38, 253 and references therein.
5. H. Kato and A. Kudo, J. Phys. Chem. B, 2001, 105, 4285.
6. H. Kato, K. Asakura and A. Kudo, J. Am. Chem. Soc., 2003, 125, 3082.
7. S. Ikeda, A. Tanaka, K. Shinohara, M. Hara, J. N. Kondo, K. Maruya and K. Domen, Microporous Mesoporous Mater., 1997, 9, 253.
8. S. Ikeda, M. Hara, J. N. Kondo, K. Domen, H. Takahashi, T. Okubo and M. Kakihana, Chem. Mater., 1998, 10, 72.
9. H. Kato and A. Kudo, Catal. Lett., 1999, 58, 153.
10. A. Kudo, S. Nakagawa and H. Kato, Chem. Lett., 1999, 28, 1197.
11. C. Mitsui, H. Nishiguchi, K. Fukamachi, T. Ishihara and Y. Takita, Chem. Lett., 1999, 28, 1327.
12. H. G. Kim, D. W. Hwang, J. Kim, Y. G. Kim and J. S. Lee, Chem. Commun., 1999, 1077.
13. H. Takahashi, M. Kakihana, Y. Yamashita, K. Yoshida, S. Ikeda, M. Hara and K. Domen, J. Alloys Compd., 1999, 285, 77.
14. A. Kudo, H. Kato and S. Nakagawa, J. Phys. Chem. B, 2000, 104, 571.
15. M. Machida, J. Yabunaka and T. Kijima, Chem. Mater., 2000, 12, 812.
16. Y. G. Ko and W. Y. Lee, Catal. Lett., 2002, 83, 157.
17. N. Sato, H. Kadowaki, H. Kobayashi, K. Ikarashi, H. Nishiyama and Y. Inoue, Chem. Lett., 2004, 33, 1452.
18. K. Shimizu, Y. Tsuji, T. Hatamachi, K. Toda, T. Kodama, M. Sato and Y. Kitayama, Phys. Chem. Chem. Phys., 2004, 6, 1064.
19. J. Kim, D. W. Hwang, H. G. Kim, S. W. Bae, J. S. Lee, W. Li and S. H. Oh, Top. Catal., 2005, 35, 295.
20. Y. Miseki, H. Kato and A. Kudo, Chem. Lett., 2005, 34, 54.
21. K. Yoshioka, V. Petrykin, M. Kakihana, H. Kato and A. Kudo, J. Catal., 2005, 232, 102.
22. Y. Ebina, N. Sakai and T. Sasaki, J. Phys. Chem. B, 2005, 109, 17212.
23. K. Shimizu, S. Itoh, T. Hatamachi, T. Kodama, M. Sato and K. Toda, Chem. Mater., 2005, 17, 5161.
24. H. Otsuka, K. Kim, A. Kouzu, I. Takimoto, H. Fujimori, Y. Sakata, H. Imamura, T. Matsumoto and K. Toda, Chem. Lett., 2005, 34, 822.
25. M. Machida, T. Mitsuyama and K. Ikeue, J. Phys. Chem. B, 2005, 109, 7801.
26. Y. Miseki, H. Kato and A. Kudo, Chem. Lett., 2006, 35, 1052.
27. T. Kurihara, H. Okutomi, Y. Miseki, H. Kato and A. Kudo, Chem. Lett., 2006, 35, 274.
28. R. Abe, M. Higashi, K. Sayama, Y. Abe and H. Sugihara, J. Phys. Chem. B, 2006, 110, 2219.
29. H. Jeong, T. Kim, D. Kim and K. Kim, Int. J. Hydrogen Energy, 2006, 31, 1142.
30. S. Ikeda, M. Fubuki, Y. K. Takahara and M. Matsumura, Appl. Catal., A, 2006, 300, 186.
31. H. Kadowaki, N. Saito, H. Nishiyama and Y. Inoue, Chem. Lett., 2007, 36, 440.
32. W. Yao and J. Ye, Chem. Phys. Lett., 2007, 435, 96.
33. Y. Li, G. Chen, C. Zhou and Z. Li, Catal. Lett., 2008, 123, 80.
34. T. Mitsuyama, A. Tsutsui, T. Hara, K. Ikeue and M. Machida, Bull. Chem. Soc. Jpn., 2008, 81, 401.
35. O. C. Compton, C. H. Mullet, S. Chiang and F. E. Osterloh, J. Phys. Chem. C, 2008, 112, 6202.
36. Y. Lee, H. Terashita, Y. Shimodaira, K. Teramura, M. Hara, H. Kobayashi, K. Domen and M. Yashima, J. Phys. Chem. C, 2007, 111, 1042.
37. K. Maeda, K. Teramura, D. Lu, T. Takata, N. Saito, Y. Inoue and K. Domen, Nature, 2006, 440, 295.
38. K. Sayama, K. Mukasa, R. Abe, Y. Abe and H. Arakawa, Chem. Commun., 2001, 2416.
39. R. Abe, K. Sayama, K. Domen and H. Arakawa, Chem. Phys. Lett., 2001, 344, 339.
40. H. Kato, M. Hori, R. Konta, Y. Shimodaira and A. Kudo, Chem. Lett., 2004, 33, 1348.
41. R. Abe, T. Takata, H. Sugihara and K. Domen, Chem. Commun., 2005, 3829.
42. M. Higashi, R. Abe, K. Teramura, T. Takata, B. Ohtani and K. Domen, Chem. Phys. Lett., 2008, 452, 120.
43. M. Kakihana and K. Domen, MRS Bull., 2000, 25, 27.
44. K. Domen, A. Kudo, T. Onishi, N. Kosugi and H. Kuroda, J. Phys. Chem., 1986, 90, 292.
45. M. Wiegel, M. H. J. Emond, E. R. Stobbe and G. J. Blasse, J. Phys. Solids, 1994, 55, 773.
46. Z. Zou, J. Ye and H. Arakawa, Chem. Phys. Lett., 2000, 332, 271.
47. A. Kudo and T. Sakata, J. Phys. Chem., 1996, 100, 17323.
48. A. Kudo and E. Kaneko, Microporous Mesoporous Mater., 1998, 21, 615.
49. M. Weiden, A. Grauel, J. Norwig and S. Horn, J. Alloys Compd., 1995, 218, 13.
50. T. A. Vanderah, T. R. Collins, W. Won-Ng, R. S. Roth and L. Farber, J. Alloys Compd., 2002, 346, 116.
51. Y. Tohdo, K. Kakimoto, H. Ohsato, H. Yamada and T. Okawa, J. Eur. Ceram. Soc., 2006, 26, 2039.
52. A. M. Srivastava, J. F. Ackerman and W. W. Beers, J. Solid State Chem., 1997, 134, 187.
53. A. Kudo and T. Sakata, J. Phys. Chem. B, 1995, 99, 15963.
54. A. Kudo, Chem. Mater., 1997, 9, 664.
55. A. Kudo and E. Kaneko, Microporous Mesoporous Mater., 1998, 21, 615.
56. M. P. Dare-Edwards, J. B. Goodenough, A. Hamnett and N. D. Nicholson, J. Chem. Soc., Faraday Trans. 2, 1981, 77, 643.
57. D. E. Scaife, Sol. Energy, 1980, 25, 41.
58. T. Ohno, K. Sarukawa and M. Matsumura, New J. Chem., 2002, 26, 1167.

---

Footnote

† Current address: Materials and Structures Laboratory, Tokyo Institute of Technology, 4259 Nagatsuda, Midori-ku, Tokohama, 226-8503, Japan.

---