Iron is the fourth most common element in the earth, where it is already oxidized into ferrous (+2) or ferric (+3) states. Among the crystal structures of iron oxide, hematite is the most thermodynamically stable form, i.e. the most commonly founded in nature. In consideration of system costs, the cost of materials for photoelectrodes needs to be treated as a serious requirement. Its electronic properties as an n-type semi-conductor make hematite a potentially promising photoanode material. The energy band gap of hematite is usually reported to be in the range of 1.9 to 2.2 eV, corresponding to 650 to 560 nm of wavelength.³⁰⁾ The absorption of yellow to UV photons that are most abundantly contained in incident photons³¹⁾ is strong for this reason. Coupled with its earth-abundance, non-toxicity, and stability in aqueous solution, its electronic properties as a semiconductor make hematite an attractive material for artificial photosynthesis. Exploration of hematite as a solar energy conversion material was first carried out by Hardee and Bard in 1976.³²⁾ They fabricated hematite thin films on Ti and Pt substrates by chemical vapor deposition. Since that time, continuing studies on hematite have been carried out with various fabrication methods, including solution-based processes.^30,33) Herein, only a few notable trials with nanostructured thin films based on solution fabrication methods will be introduced.

Researchers have tried to generate thin films of hematite with donor doping,^34-37) as hematite has a short diffusion length (L_D = 2-4 nm)^37-38) and poor majority carrier conductivity.^35,39) An increase of doping concentration in turn reduces the width of a space-charge layer between the photoelectrode and the electrolyte. When assuming classical depletion layer theory,⁴⁰⁾ only 7 nm width of depletion layer is generated in the case of Nb- doped single crystal hematite with a concentration of 5 × 10¹⁹ cm⁻³.³⁵⁾ However, although dopants were added, the relatively low adsorption coefficient for hematite^41,42) straddles carrier separation if hematite has a planar structure. Because the adsorption depth for planar hematite is in the range of 120 - 46 nm for photon wavelengths of 550 - 450 nm,⁴³⁾ fast recombination of excitons will be promoted in the case of a simple planar thin film. Consequently, overcoming these limitations has always been treated as the major barrier for hematite utilization.

In this regard, a recently reported study suggests adaptation of doping and nanostructuring simultaneously. Sivula et al.⁴⁴⁾ reported mesoporous hematite thin films on fluorine-doped tin oxide (FTO) substrate prepared by a solution-based colloidal method. For preparation of the precursor paste, a solution was synthesized by combining 400 mg of Fe₂O₃ nanopowder, with 40 μL of a 10% solution of acetylacetone (Acac) and 1-hexanol using a mortar and pestle. The prepared solution was diluted with 1% Acac in 2-propanol, and this was well-dispersed in a water-bath (10°C) by an ultrasonic tip sonicator. Finally, the dispersed solution was concentrated by evaporation and mixed with a 4 wt% solution of hydroxypropyl cellulose (HPC) in 2-propanol. The prepared precursor paste was coated on FTO substrate by doctor-blade at a thickness of 40 μm, and then annealed by a two-step process in order to remove the organics. Prior to reaching the set point (700 - 800°C), an initial heat treatment was carried out at 400°C for 10 h. Fig. 6(a), (b) shows different particle sizes of hematite with respect to the heat treatment temperature. Even though a specimen prepared at 800°C has larger particle size than at 700°C, a much higher photocurrent density was exhibited by the 800°C specimen. Analysis by XPS (X-ray photoelectron spectroscopy) and UV-Visabsorption spectroscopy showed that doping of Sn from the FTO substrate occurs, due to the high thermal energy which induces large effects on optical properties and photoactivity. Souza et al.⁴⁵⁾ also tested the effect of doping and nanostructure simultaneously by fabricating nanostructured hematite thin films using a polymeric precursor spin coating method. Both nanostructure and silicon doping concentration were controlled by the concentration of tetraethyl orthosilicate (TEOS), which has the role of silicon precursor. For the polymeric precursor, iron(III) nitrate 9-hydrate and TEOS dissolved in isopropyl alcohol were stirred in citric acid dissolved in DI water and heated in the range of 70 - 90°C. Promotion of citrate polymerization was done by stirring and heating after adding ethylene glycol to this solution. Although an optimally-controlled doping concentration of Si exhibited enhanced photocurrent density, it could be seen that the effect of Si doping shows much lower impact compared with Sn doping. For this reason, the effect of doping in hematite relies on the proper type of dopant and enough thermal energy. However, enough thermal energy promotes further growth of crystal size, which highlights the short diffusion length of hematite.

Brillet et al.⁴⁶⁾ conquered this phenomenon by encapsulation with silica, whose role is to confine further growth of the hematite crystals. The experimental procedure for preparation of precursor paste was very similar to Sivula et al.,⁴⁵⁾ but titanium isopropoxide was also added for Ti ion doping when the solution was mixed with HPC. First, heat treatment was applied, to remove organic compounds. The encapsulation was then performed by chemical bath deposition. The annealed mesoporous hematite thin films were immersed at 0°C in aqueous methanol solution, hexadecyl-trimethylammonium chloride, and aqueous ammonia in sequence. The encapsulated hematite films were then annealed at 800°C in air. After heat treatment, the silica layer was etched by dilute aqueous NaOH (5 M). Brillet et al. demonstrated an adequate fabrication method to restrict further growth of particle size and an enhancement of the doping effect at high temperature.⁴⁶⁾

Deposition of an FeOOH (oxygen evolution co-catalyst) overlayer on hematite was recently tested by Kim et al. for further enhancement of the PEC performance.⁴⁷⁾ Mesoporous hematite thin films were prepared by simply immersing a FTO substrate in a solution containing 0.15 M FeCl₃·6H₂O and placing it in an oven at 100°C for 6 h. Fig. 6(f) shows the finely coated FeOOH layer. Prepared thin films were annealed via a two-step heat treatment: 550°C for 1 h followed by 20 min annealing at 800°C. The FeOOH overlayer was deposited by simply immersing hematite thin films into an aqueous solution containing 0.15 M FeCl₃·6H₂O and 1 M NaNO₃ at 100°C for 5 min. An increase in surface reaction can be captured by a decrease of onset potential and a doubling of photocurrent density.

Among oxide photoelectrode candidates, BiVO₄ has been identified as one of the most promising n-type semiconductor photoanodes for use in solar water oxidation, since it can absorb a substantial portion of the visible spectrum due to its relatively narrow band gap energy (~ 2.4 eV), and it has a favorable portion of the conduction band very near the thermodynamic hydrogen evolution potential.^3,5,8-9,48) However, inefficient charge transport and interfacial charge transfer properties are the key limiting factors for BiVO₄ photoanodes.^3-4,8,48) The presence of a porous nanostructure in the photoelectrodes provides an increased interfacial area between the electrolyte and the film. The formation of nanostructures enables improvement of the water-splitting efficiency, as the photogenerated holes will have to traverse less bulk material to be collected at the solution interface; therefore it has a lower chance of recombining before participating in the electrochemical reaction.^4,27,49) However, increasing the surface area by introducing porosity can also have adverse effects on charge separation and transport, as it can result in an increase in surface states, defect sites, and grain boundaries while decreasing crystallinity. Therefore, the morphologies and surface areas of a photoelectrode should be finely optimized to maximize the overall net positive effect.²⁷⁾

One interesting study by Xi et al.⁵⁰⁾ reports that BiVO₄ nanoplates synthesized by a hydrothermal method have monoclinic scheelite structures with well-defined facets exposed at the surface. They compared photoactivities of BiVO₄ with different morphologies such as nanorods and nanoplates. The BiVO₄ nanoplates showed higher photocurrent density than nanorods due to active surface structures of the BiVO₄ nanoplates. This study indicates that designing nanostructures with appropriate surface facets is a feasible approach for the improvement of photoactivity.

BiVO₄ has also been prepared by a new cathodic electrode-position method, as demonstrated by McDonald, Choi et al.^51-52) A polycrystalline BiOI electrode, composed of extremely thin 2D plates, allowed for the preparation of porous BiVO₄ photoanodes using a simple chemical and thermal treatment. The resulting BiVO₄ with high porosity showed outstanding PEC efficiency. On the basis of these results, Kim and Choi⁴⁸⁾ modified a previous synthesis developed by McDonald and Choi.⁵¹⁾ It is possible to facilitate the production of nanoporous BiVO₄, as displayed in Fig. 7(a)-(f), which is composed of much smaller particles, achieving an enlarged surface area.⁴⁸⁾ The major change was the solution used as the source for conversion of BiOI to BiVO₄. They observed that the surface of BiOI film is highly hydrophobic; thus the aqueous vanadium solution used in the study by McDonald and Choi could not easily penetrate into the BiVO₄ films. As a result, the conversion of BiOI to BiVO₄ was initiated only at the top of the film surface, and the grains of BiVO₄ grew toward the bottom of the film by slow solid-state diffusion, resulting in the formation of large particles of BiVO₄ and no porosity in that direction, limiting the surface area. When the aqueous vanadium solution was replaced by a more hydrophobic DMSO solution containing vanadyl acetylacetonate as the vanadium source, the solution could easily penetrate the entire BiOI film. Therefore, upon heating, multiple nuclei could form, even within a single BiOI sheet, resulting in the formation of smaller particles, which considerably increased the porosity and the surface area. This research proposed a novel technique of electrochemical deposition and demonstrated that nanoporous morphology with high surface area effectively suppresses carrier recombination without additional doping.

Another approach to improve absorbance and charge separation efficiency of BiVO₄ is shown in Fig. 7(g)-(u). Lee et al.⁵⁾ first introduced shape-controlled gold nanoparticles (Au NPs) as plasmonic NPs on BiVO₄ thin film synthesized by pulsed electrodeposition. They reported shape-controlled Au NPs with optimum coverage that significantly promote the photoactivity of BiVO₄ in over-band gap photon energy, rather than sub-band gap photon energy. Well-defined octahedral Au NPs significantly increased photocurrent density of BiVO₄ about 3-fold, while hemispherical Au NPs reduced it, since largely-enhanced localized surface plasmon resonance electric fields (LSPR-EFs) are only observed with octahedral Au NPs. Finite-domain time-difference (FDTD) simulation also validated intensified LSPR-EFs in the entire UV-vis region from octahedral Au NPs, compared to hemispherical Au NPs, indicating that control of the shape of plasmonic Au NPs is the key to achieving high PEC efficiency.

In addition to homogeneous BiVO₄, the formation of type II heterojunctions has proven to be effective to facilitate the separation of photo-induced electron-hole pairs, enlarge the interfacial area, and maximize light absorption.³⁾ The combination of BiVO₄ and WO₃ to form a type II heterojunction not only facilitates charge separation, but also expands the light absorption capability of the composite photoelectrode.^3,11,26)

Jin et al. first designed a WO₃ thin film consisting of yolk-shell structured nanoparticles via a solution process without the use of vacuum deposition, as shown in Fig. 8(a)-(f).⁵³⁾ A thin BiVO₄ layer with a smaller band gap was coated onto the surface of and inside the WO₃ shells, providing a rationally-designed inner space between the particles and the shell for better electrolyte accessibility. The yolk-shell-shaped PEC photoanode not only induces efficient light absorption but also plays an important role in electron collection from BiVO₄ due to an enlarged contact area. The photocurrent density of the yolk-shell (Y-WO₃/BiVO₄/OER) photoanode achieved its highest value of 5.0 mA/cm² at 1.23 V vs. RHE.

Lee et al.³⁾ reported highly ordered one-dimensional WO₃ nanorods by changing their porosity and aspect ratio, and depositing stoichiometric BiVO₄ on the surface of the WO₃ nanorods by pulsed electrodeposition. The notable point of this research is that cross-sectional transmission electron microscopy shows dot-like BiVO₄ well coated on the entire surface of the WO₃ nanorods. BiVO₄/WO₃ type II heterojunction anodes can lead to a high photocurrent density of 4.55 mA/cm² and an incident photon-to-current conversion efficiency of 80% at 1.23 V versus a reversible hydrogen electrode without additional catalyst.

Pihosh et al.⁵⁴⁾ demonstrated this concept with a WO₃/BiVO₄+CoPi core-shell nanostructured photoanode that achieves near theoretical water splitting efficiency, as shown in Fig. 8(h)-(k). This study combines BiVO₄ with more conductive WO₃ nanorods in the form of a core-shell heterojunction, where the BiVO₄ absorber layer is thinner than the carrier diffusion length, while its optical thickness is re-established by light trapping in high aspect ratio nanostructures. Their photoanode demonstrates an ultimate water splitting photocurrent of 6.72 mA/cm² under 1 sun illumination at 1.23 V vs. RHE, which corresponds to ~ 90% of the theoretically possible value for BiVO₄.

Titanium dioxide (TiO₂) has been favored in many industrial fields for its excellent thermodynamic stability and chemical resistance.^55-59) In the case of photocatalytic applications, the discovery of photocatalytic phenomena of TiO₂ by Fujishima and Honda in 1972⁶⁰⁾ demonstrated the potential for new applicable areas, i.e. photovoltaics and photocatalysts. Since then, considerable efforts have been made to make TiO₂ a promising material for photocatalysis,^55,61-65) and today TiO₂ is considered one of the potential candidates to play an important role in solving environmental and pollution problems as a photocatalyst. Following the innovative work of Fujishima and Honda, several synthetic methods including sol-gel,^66-70) chemical vapor deposition,^71-74) physical vapor deposition,^75-77) and others^78-81) have been adapted to synthesize TiO₂ photocatalysts. But most of these cases show a lower photocurrent density than other photoanodes because of a higher resistance and larger band gap (3.0 - 3.2 eV) for TiO₂.⁸²⁾ Consequently, these disadvantages would need to be overcome in order to fabricate stable and efficient TiO₂ photoanodes. Several strategies have been proposed to maximize the potential of TiO₂ as a photoanode, and these will be further discussed here.

One approach to overcome the limit of TiO₂ photocatalysts is reinforcement of photon absorption and reaction with electrolytes by increasing surface area to volume ratio,⁸³⁾ and formation of nanostructure is always a good choice to maximize the surface area of films. Some researchers have synthesized TiO₂ nanostructure arrays by coating on different nanostructured templates such as anodic alumina membrane (AAM) and ZnO.

In 2003, Lin et al.⁸⁴⁾ fabricated TiO₂ nanowire arrays by electrophoretic deposition into the pores of an AAM. Colloidal suspensions for electrophoretic deposition of TiO₂ were synthesized by dissolving titanium tetraisopropoxide (TTIP) in ethanol and mixing with glacial acetic acid solution and the addition of nitric acid to adjust the pH to 2 - 3. An AAM with an Au substrate attached to Cu foil was used as a cathode, and platinum was used as an anode. The deposition condition for the TiO₂ coating into the pores of the AAM was a voltage of 2 - 5 V and fabricated films were annealed at 500°C for 24 h. Isolated TiO₂ nanowires could be obtained after dissolving the AAM templates in a 5 wt% NaOH solution. The length of the nanowires was dependent on the length of the AAM pores.

TiO₂ nanotubes can be fabricated using the sol-gel method and AAM templates. Lee et al. in 2004⁸⁵⁾ dipped an AAM template into a TTIP solution prepared by mixing TTIP with 2-propanol and 2,4-pentanedione, and the sol-coated template was placed under vacuum until the entire volume of the solution was pulled through the AAM. The AAM was hydrolyzed by water vapor over HCl solution for 24 h, air-dried at room temperature, and then calcined at 400°C for 2 h and cooled to room temperature at a cooling rate of 2°C/h. Pure TiO₂ nanotubes were obtained after the AAM was dissolved in 6 M NaOH for several min. TiO₂ nanotubes could also be obtained by a chemical bath deposition method using AAM. Reported by Liu et al.,⁸⁶⁾ TiO₂ nanotubes could be synthesized by simply submerging AAM template in dilute TiF₄ at pH 2.1 and 60°C for 12 - 48 h.

Not only AAM, but also a ZnO nanorod array on glass substrate could be used as a TiO₂ nanotube template. Similarly, TiO₂ nanotube arrays could be obtained by etching ZnO in a ZnO nanorod template coated with TiO₂ by a dip-coating method.⁸⁷⁾ Dilute HCl was used as an etchant for the ZnO nanorods and, due to the structure of the ZnO building unit, the TiO₂ nanotubes had pores with a hexagonal shape.

In order to make TiO₂ films as photoelectrodes, it is necessary to generate a junction with a conductive substrate. However, a difference of refractive index between TiO₂ and the substrate induces a reflection of incident light which reduces light absorption.⁸⁸⁾ To overcome this phenomenon, some researchers adapted the characteristics of a rutile TiO₂ growth unit. As the crystal planes of TiO₂ have quite different surface energies, rutile TiO₂ shows preferential crystal growth when synthesized by a hydrothermal process.^56,89-91) Using this method, fabrication of a branched nanorod structure has been tried. Cho et al.⁹²⁾ fabricated branched nanorod TiO₂ films on a FTO substrate by a two-step synthesis. First, TiO₂ nanorods were grown hydrothermally on TiO₂-coated FTO substrate. Tetrabutyl titanate (TBT), 25 mL of deionized (DI) water, and 25 mL of concentrated HCl were stirred to create a hydrothermal solution. The obtained TiO₂ nanorods films were annealed at 450°C for 1 h in air, and seed nanoparticles for TiO₂ branches were deposited by a dip-coating method using the TiO₂ polymeric solution. The seeded nanorods films were immersed in an aqueous solution consisting of 10 mL of DI water, 0 - 0.5 mL of HCl, and 0.1 mL of TiCl₃ solution and kept at 80°C for 1 - 2 h. Finally, branched nanorod films were obtained by washing with ethanol and annealing at 450°C for 1 h in air. The obtained nanorods and branched nanorods showed much lower reflectance in the UV region compared with porous nanoparticles because of a change in film density from bottom to top and scattering of incident light.⁸⁸⁾

Such examples for controlling TiO₂ nanostructure have contributed to enormous advances over simple planar thin films, but still further enhancements are needed before TiO₂ photoanodes can be commercialized. Based on the belief that research on single TiO₂ structures has reached its theoretical upper limit, co-catalyst attached TiO₂ photoanodes have been examined recently. Among these, the Au nanoparticle-coated TiO₂ branched nanorod structure, reported by Su et al.,⁹³⁾ shows outstanding results. A schematic of the experimental procedure for preparing these nanorods is illustrated in Fig. 9. Similar to the above case, TiO₂ nanorods were hydrothermally synthesized on FTO with a solution composed of 30 mL of DI water, 30 mL of concentrated HCl, 1 mL of TBT and 5 mL of 5 M NaCl. Hydrothermally prepared TiO₂ nanorods were immersed in 0.2 M TiCl₄ at room temperature for 18 h and then rinsed with absolute ethanol. Specimens were annealed at 450°C for 30 min. The fabricated TiO₂ branched nanorods were immersed in 0.3 mM HAuCl₄, dispersed in DI water, and irradiated with a 300 W Xe lamp for 6 h to coat Au nanoparticles on TiO₂ surfaces. The much higher photocurrent density shown in this study demonstrates that both strategies, i.e. increasing surface area and enhancing surface reaction by a co-catalyst, can be successfully adopted at the same time.

Recently it has been shown that the use of TiO₂ can be extended to photocathodes as well as photoanodes. Andoshe et al.⁸⁸⁾ fabricated a photocathode with a heterojunction of p-Si and TiO₂ nanorod thin film. For the heterogeneous nucleation of TiO₂ nanorods during hydrothermal synthesis, a 5 nm thick seed layer of the TiO₂ film was deposited on p-Si. Subsequently, TiO₂ nanorods were hydrothermally grown on a seed layer of deposited p-Si substrate at 180 - 220°C for 0.5 - 2 h. The solution for the hydrothermal process contained 25 mL of DI water, 25 mL of concentrated HCl and 0.8 mL of TBT. As shown in Fig. 10(a), the TiO₂ nanorod/p-Si photocathode exhibits a matt black color. As illustrated in Fig. 10(b)-(h), a gradual change in nanorod density along the vertical direction causes a steady change of refractive index that restricts total reflection. TiO₂ has the additional role of passivating p-Si from corrosion. Likewise, TiO₂ exhibits a potential for heterojunction, passivation, and interface modulation materials.

Cu₂O and CuO are the most representative p-type binary oxides that have been introduced for use in solar water splitting. In addition to their favorable band gap energies that allow for the utilization of visible light, the low cost, earth abundance, and non-toxicity of Cu are additional advantages for developing Cu-based photoelectrodes.⁹⁴⁾

Cu₂O, with a direct band gap of 2.0 eV, is an attractive oxide for solar water reduction, thus it is possible to theoretically reach a photocurrent of 14.7 mA/cm² and a solar to hydrogen efficiency of 18%. The major limitation of Cu₂O is photocorrosion (e.g., Cu₂O + 2H⁺ + 2e⁻ → 2Cu + H₂O).⁹⁶⁾ Paracchino et al.⁹⁷⁾ successfully synthesized Cu₂O films by using electrochemical deposition, which was produced by cathodic reduction of Cu²⁺ ions in an aqueous alkaline (pH 12) solution. Cu₂O grains were covered with metal oxide protective layers [5 × (4 nm ZnO/0.17 nm Al₂O₃)/11 nm TiO₂] via atomic layer deposition (ALD), and finally coated with Pt nanoparticles, resulting in outstanding film quality when coated by ALD. Significantly, the protected Cu₂O photocathode exhibited substantially improved photoactivity and stability. The surface-protected Cu₂O achieved a photocurrent density of − 7.6 mA cm⁻² under AM 1.5 illumination; furthermore, the protected electrodes remained photoactive after 1 h of testing and the measured Faradaic efficiency was close to 100%.

Yang et al.⁹⁵⁾ demonstrated a Cu₂O/CuO bilayer composite synthesized by a facile method that involved electrodeposition and a subsequent thermal oxidation, as shown in Fig. 11(a)-(d). This strategy of combining electrodeposition and thermal oxidation provides an easy, low cost, and scalable strategy to prepare Cu₂O/CuO heterogeneous photocathodes for HER. The Cu₂O/CuO photocathode exhibited significantly high photoactivity and good photostability toward HER, especially at high potentials in alkaline solution. The photocurrent density for HER was 3.15 mA/cm² at a potential of 0.40 V vs. RHE, which was one of the highest reported values at the same potential on copper-oxide-based photocathodes. Good photostability was seen during the 4 h measurements.

Qi et al.⁹⁴⁾ introduced NiFe-layered double hydroxide (NiFe-LDHs) overlayers as co-catalyst on Cu₂O electrodes via electrodeposition and studied their PEC behavior, as shown in Fig. 11(f)-(j). The electrodeposition enabled the synthesis of uniformly-anchored NiFe-LDH layered nanoplates onto the Cu₂O surface. The resulting Cu₂O/NiFe-LDH exhibited a remarkable 7-fold enhancement of photocurrent density under an applied voltage as low as − 0.2 V vs. Ag/AgCl. Also, long-term photostability tests revealed that Cu₂O/NiFe-LDH photocathodes showed no photocurrent loss after 40 h of operation under light at − 0.2 V vs Ag/AgCl low bias condition, which makes Cu₂O/NiFe-LDH photocathode a good candidate for low bias PEC water splitting.

CuO is another well-known photocathode with an indirect band gap of 1.2 - 1.8 eV. Due to its smaller band gap, it is possible to achieve a higher photocurrent than with Cu₂O, but CuO has received less attention than Cu₂O for use as a photocathode in solar water splitting. This lack of attention is probably because of uncertainty as to whether the photo-excited electrons in CuO could reduce water to H₂ (the conduction band minimum of CuO is reported to be between 0 and − 0.2 V vs RHE, which is more positive than that of Cu₂O).⁹⁸⁾ Also, photocorrosion of CuO is another limiting factor for use in PEC water splitting. Recent studies showed suppressed photocorrosion of CuO.^98-99)

Cardiel et al.100) developed novel electrochemical synthesis methods to produce copper hydroxyl double salt (Cu-HDS) films with four different intercalated anions (NO³⁻, SO₄²⁻, Cl⁻, and dodecyl sulfate) as pure crystalline films, deposited as (Cu₂NO₃(OH)₃, Cu₄SO₄(OH)₆, Cu₂Cl(OH)₃, and Cu₂DS(OH)₃), as shown in Fig. 12. These methods are based on p-benzoquinone reduction, which increases the local pH at the working electrode and triggers the precipitation of Cu²⁺ and appropriate anions as Cu-HDS films on the working electrode. The resulting Cu-HDS films could be converted to crystalline Cu(OH)₂ and CuO films by immersing them in basic solutions. The CuO films prepared from Cu-HDS films have unique low-dimensional nanostructures, creating high surface areas that cannot be obtained by direct deposition of CuO, which has a 3D atomic-level crystal structure, since Cu-HDS films were composed of 2D crystals as a result of the atomic-level layered structure of HDS. The CuO films recorded great onset potential (> 1.0 V vs RHE) very close to their flat band potential (~ 1.2 V vs RHE), which shows great promise for use as photocathodes in PEC cells for solar water splitting.

To date, metal oxides as photocathodes have been focused primarily on Cu-based oxides such as Cu₂O, CuO, CuFeO₂ and CuBi₂O₄.^28,101-102) However, the presence of copper in most p-type oxides is known to be a cause for photocorrosion, therefore these materials need a protection layer. Recent studies have searched for a breakthrough to solve the limitations of Cu-based metal oxides as photocathodes.

Perovskite-type lanthanum iron oxide (LaFeO₃) is a p-type oxide that has several attractive characteristics such as an ideal band gap of 2.1 eV and band edger positions for use in PEC water splitting. Wheeler et al.103) reported nanoporous LaFeO₃ prepared by electrochemical co-deposition of La(OH)₃ and Fe(OH)₂ via nitrate reduction, as shown in Fig. 13. The significant finding of this study is that, after 1 h of PEC measurement for water reduction at 0.5 V vs. RHE, there is a negligible portion of surface electrons being used for water reduction, and no sign of photocorrosion unlike Cu-based metal oxides. This confirmed that LaFeO₃ is a rare p-type oxide that does not suffer from photocorrosion. Therefore, the enhancement of the PEC performance of LaFeO₃ can be expected with systematic doping studies and by combining it with a proper hydrogen catalyst.