Proton exchange membrane photoelectrochemical cell for water splitting under vapor feeding

Solar hydrogen

To effectively reduce CO₂ emissions while ensuring economic growth, we must shift from current fossil fuel resources to renewable energy sources, such as solar power. Renewable energy plays a crucial role in achieving a sustainable future by providing environment-friendly alternatives to fossil fuels. Water-splitting technology for H₂ production can convert renewable energy into storable and transportable fuel to mitigate the spatial and temporal disparities in supply and demand. Although sunlight, which is abundant yet diffuse in energy density, holds vast potential for large-scale CO₂-free H₂ production, the challenge lies in achieving cost-effectiveness compared to H₂ production via steam reforming of methane. To address this issue, photocatalytic water splitting has been actively investigated as a potentially cost-effective H₂ production technology^[1]. However, the solar-to-hydrogen (STH) energy conversion efficiency is still low (approximately 1%) for state-of-the-art photocatalytic systems, and developing a visible-light-responsive photocatalytic system to enhance its efficiency remains a challenge^[2]. Photoelectrochemical (PEC) water splitting, which provides high STH energy conversion efficiency, is a promising technology for solar energy capture and utilization^[3,4]. This method has great potential for reducing the reliance on traditional energy sources and mitigating environmental impacts. In this review, we discuss a gas diffusion oxide photoelectrode system supported by surface functionalization using a perfluorosulfonic acid (PFSA) ionomer for PEC water vapor splitting under gas-phase conditions. This system is sustained by polymer electrolyte membranes without the use of liquid aqueous electrolytes.

Humidity in the air as a hydrogen source

Countries facing limitations in domestic renewables have to import chemical energy carriers derived from renewables. For example, the annual solar radiation in Japan is 1,200-1,500 kWh m^-2, but arid regions, such as Saudi Arabia and Australia, have abundant solar energy of 2,000-2,100 kWh m^-2 [Figure 1A]. Because the energy demand in these solar-abundant regions is low, there are opportunities to produce energy carriers for export.

Figure 1. (A) Global distribution of solar radiation on the surface of the Earth (©2020 The World Bank, Source: Global Solar Atlas 2.0, Solar resource data: Solargis) and energy transport to Japan. (B) Schematic of PEC water splitting using a photoanode and cathode immersed in a liquid electrolyte. (C) Schematic of a vapor-fed PEM-PEC system utilizing a porous photoanode and cathode with gas diffusion layer. (B and C) were reproduced from ref.^[13].

Hydrogen is the primary candidate for solar fuel owing to its chemical simplicity, despite its energy-efficient transportation posing a challenge. Water is a promising electron source for producing energy carriers from renewable sources. However, regions with abundant solar radiation can face challenges regarding freshwater availability, which can become a global issue in the future^[5]. To produce H₂ from seawater and wastewater, desalination and purification are required to eliminate impurities poisoning the catalyst, photoelectrode, and membrane. However, the water treatment is both energy-intensive and costly. The use of liquid water also brings about issues such as gas bubble formation, the need for pumping, freezing in low temperatures, and the leaching of toxic components.

One approach to address this challenge involves tapping into atmospheric humidity as an alternative source for hydrogen production. The relative humidity (RH) at the sea surface is consistently high (approximately 80%) throughout the year^[6]. Additionally, a recent study indicated that even in arid regions, low concentrations of atmospheric moisture can be condensed into liquids using water adsorbates, thereby facilitating water electrolysis^[5]. Several researchers have reported solid polymer electrolyte-based PEC systems, such as membrane-based artificial photosynthetic systems^[7,8]. Water vapor splitting using particulate photocatalysts has also been investigated owing to its several advantages over liquid water-based processes, such as easy scalability and suppression of the leaching of catalyst components^[9-13].

The use of water vapor has the potential advantage of reducing maintenance costs because the risk of corrosion and poisoning is reduced as water is purified through vaporization. In addition, liquid-pumping systems are not required because the natural convection of air can feed water vapor, and the risk of freezing is minimized. The vapor-fed PEC water splitting avoids the deleterious effects associated with bubble formation in aqueous media, the reduction of the contact area between the liquid and electrode, and the reflection and/or scattering of incident illumination^[14].

Photocatalysis and photoelectrochemistry

There are two primary categories of solar water-splitting devices, namely hybrid systems combining photovoltaic solar cells with a water electrolyzer and particulate photocatalyst^[1-4] PEC systems that fall between these two approaches. Particulate photocatalysts and oxide photoelectrode systems are potential cost-effective routes for scalable hydrogen production. During photocatalytic water splitting, a mixture of H₂ and O₂ gases is produced, which requires membrane separation in the later stages. In contrast, in PEC reactions, pure H₂ is obtained because a membrane separates the gases evolved at the anode and cathode [Figure 1B]. Moreover, an applied external bias or p-n junction can create a potential gradient within the light-responsive semiconductor electrode to enhance charge separation, thereby improving the internal quantum efficiency.

The Honda-Fujishima effect, which involves pairing a titanium dioxide (TiO₂) photoanode with a platinum cathode, is common in PEC systems^[15]. An n-type semiconductor can serve as a photoanode for oxygen evolution reactions (OERs), while a p-type semiconductor can serve as a photocathode for hydrogen evolution reactions (HERs). The tandem combination of an oxide photoanode with a photocathode allows for the utilization of narrow bandgap semiconductors through two-step photoexcitation. In conventional water electrolysis, the overvoltage of the OER at the anode surpasses that at the HER cathode. This emphasizes the significance of developing photoanode materials for the OER. Oxide semiconductors are often favored as O₂-evolving photoanodes because of their resistance to photocorrosion.

When a photoanode is exposed to light with energy surpassing its band gap (Eg), photoabsorption generates holes (h⁺) in the valence band, leading to oxygen production via four-electron water oxidation (2H₂O → O₂ + 4H⁺ + 4e^-). These PEC reactions typically occur on semiconductor electrodes immersed in aqueous electrolyte solutions. However, the limited availability of freshwater poses challenges in upscaling hydrogen production. Hence, researchers have investigated PEC systems in which water vapor (gaseous phase) acted as the reactant [Figure 1C]. A proton exchange membrane (PEM) was used as the solid electrolyte, with water supplied in vapor forms. In the PEM-PEC devices, it was difficult to integrate conventional photoelectrodes fabricated on two-dimensional substrates, such as sheets and plates, owing to the ion transport limitations toward the membrane. The semiconductor electrodes must exhibit porosity to facilitate the diffusion of water vapor and evolving oxygen, along with proton transport to the counter electrode through the electrode and membrane.

Historical overview of PEM-PEC devices

The PEM-PEC cell design minimized the gap between the two electrodes to reduce the electrolyte resistance and prevent the mixing of the gases evolved at each electrode. In 1996, Ichikawa reported the first PEM-PEC cell with a perforated thin film of anatase TiO₂ and an electrocatalyst separated by a Nafion membrane^[16]. The TiO₂ photoanode facilitated liquid water oxidation, generating dioxygen and protons. These protons then move to the Pt and ZnO/Cu electrocatalysts for H₂ production and CO₂ reduction, respectively. In 2009, Seger and Kamat reported a PEM-PEC cell consisting of an anatase TiO₂ photoanode and Pt cathode for H₂ production^[17]. They achieved the methanol solution oxidation using TiO₂ nanoparticles (Degussa P25) on carbon paper under UV irradiation without an external bias voltage. Later, several groups applied PEM-PEC systems for splitting water using a KOH solution or electrolyte-free pure water^[18-20].

For gas-phase reaction, Georgieva et al. first applied a PEM-PEC cell for the photooxidation of methanol vapor in air under UV and visible light illumination in 2009^[21]. They observed an increase in the photocurrent of TiO₂/tungsten trioxide (WO₃)-coated stainless steel with increasing gaseous methanol concentration when exposed to air streams saturated with water and methanol vapor. Iwu et al. reported PEC oxidation of gaseous methanol and water vapor using a TiO₂ P25 photocatalyst on carbon paper^[22]. In 2014, Rongé et al. demonstrated a PEM-PEC cell that produced hydrogen from the water vapor present in the air^[23]. The PEC oxidation of water vapor was facilitated over TiO₂ thin films coated multi-walled carbon nanotubes that were grafted onto carbon fiber felt, the surface of which was functionalized with a thin PFSA film and water-adsorbing zeolite particles. The key materials to achieve PEC water vapor splitting included photoanodes fabricated on carbon fibers and stainless steel mesh.

This mini-review introduces the development of porous photoanodes that utilize titanium felt as a conductive substrate, focusing primarily on our research efforts^[13,24-28]. The recent progress in the vapor-fed PEM-PEC cell is presented based on the rational design of the gas-diffusion oxide photoanodes and the triple-phase boundary concept under the gas-phase conditions. Finally, suggestions for future work are provided to improve the STH energy conversion efficiency of the PEM-PEC cell.

Porous tungsten trioxide photoanode

In polymer electrolyte fuel cells, waterproofed carbon fibers (commonly known as carbon paper) typically serve as gas-diffusion layers. Because of concerns regarding the potential oxidative degradation of carbon during the photoanode reaction, titanium fiber felt was opted for the conductive substrate for porous tungsten oxide (WO₃) electrodes. Amano et al. prepared the WO₃ layer on three-dimensional macroporous titanium fibers through a dip-coating method and subsequent calcination^[29]. The titanium felt was soaked in an aqueous precursor solution of ammonium metatungstate, including polyethylene glycol. Then, the precursor-coated felt was calcined in air at 650 °C for 1 h to obtain the porous WO₃ electrode. Titanium has superior thermal durability in oxidizing atmospheres compared to conventional transparent conductive oxide substrates^[30].

Figure 2A shows the scanning electron microscopy (SEM) images of the WO₃ electrode prepared using titanium felt as the conductive substrate (porosity: 66.6%)^[24]. The titanium fibers, with a converted diameter of 20 µm, were uniformly coated with a WO₃ layer to maintain substantial porosity ranging from several tens of microns to several microns. The WO₃ layer was composed of fine particles of diameters of ~100 nm, forming pore spaces that were several tens of nanometers in size. The bimodal porous WO₃ electrode exhibited good photoanodic properties in aqueous electrolyte solutions^[29]. The preparation of porous WO₃ photoanodes using carbon microfiber felt as a conductive substrate has also been reported^[31,32].

Figure 2. (A) Macroporous and microporous structure of gas-diffusion WO₃ electrodes. (B) Membrane electrode assembly with an electrode area of 25 cm², a stainless steel PEM-PEC cell with a 16 cm² irradiation area, and the spectrum of a blue LED light source for the vapor-fed system (wavelength: 453 nm, irradiance: 7 mW cm^-2, and temperature: 25 °C). Reproduced with permission from ref.^[24], Copyright 2019, American Chemical Society.

Triple-phase boundary for gas-fed PEC reaction

The fabricated gas-diffusion WO₃ electrode was attached to the PEM (Nafion 117 membrane) to create a membrane electrode assembly [Figure 2B]. A carbon-supported Pt electrocatalyst was used as the cathode in polymer electrolyte fuel cells. The membrane electrode assembly was then placed in a homemade electrolyzer equipped with an optical window, establishing the PEM-PEC cell. Approximately 3 vol% water vapor at ~90% RH and 25 °C was supplied to each of the two electrodes by bubbling argon at a flow rate of 20 mL min^-1 into deionized water. The irradiation area of the photoanode was 16 cm², and a blue light-emitting diode (LED) with a central wavelength of 453 nm was used as the light source [Figure 2B]. The irradiance was 7 mW cm^-2 at 453 nm. The temperature of the PEM-PEC cell slightly increased by 2 °C during light irradiation, and the experiment was conducted under conditions that did not lead to water vapor condensation.

The photocurrent response of WO₃ electrodes in a water vapor environment is shown in Figure 3^[25]. When utilizing an all-solid-state PEM-PEC cell, the photocurrent response is significantly lower than that in an aqueous electrolyte solution. To address this, a proton-conducting PFSA ionomer dispersion was applied to the WO₃ electrode and subsequently dried at 80 °C to form a thin ionomer film on the surface. While the uncoated WO₃ electrode demonstrated poor performance, the ionomer coating significantly enhanced the photocurrent response in a water vapor atmosphere [Figure 3A]. This indicates that surface functionalization is required to improve the proton conduction across the porous photoanode and membrane.

Figure 3. PEC splitting of water vapor using WO₃ photoanode: (A) dependence of photocurrent response on the applied voltage and (B) time-dependent O₂ production rate over the photoanode at an applied voltage of 1.2 V vs. cathode (wavelength: 453 nm, irradiance: 7 mW cm^-2, and irradiation area: 16 cm²). Schematic illustrations of the interfacial structure of WO₃ photoanodes in a gas-phase atmosphere: (C) bare WO₃, where proton transport is the rate-determining step and (D) PEC triple-phase boundary formed by a thin PFSA ionomer film coating on the surface. Reproduced from ref.^[25], Copyright 2018 Amano et al.^[25].

Gas chromatography confirmed oxygen production at the porous WO₃ photoanode under vapor-feeding conditions [Figure 3B]. With the ionomer coating, the Faradaic efficiency (FE) of O₂ generation was 80%, indicating effective progress in water splitting. In contrast, the O₂ FE before surface coating was low, indicating that the OER did not proceed easily in the gas-phase environment. Unlike in an aqueous electrolyte solution, proton-coupled electron transfer, involving proton transport, is the rate-determining step in OERs [Figure 3C].

However, when a PFSA ionomer film is used for surface functionalization, proton-coupled electron transfer and proton transport are no longer the rate-determining step, even in gas-phase environments without liquid electrolytes. This facilitates a fast multiple-electron transfer pathway, resulting in a good photocurrent response and high O₂ FE [Figure 3D]. In vapor-fed PEC water splitting without an aqueous electrolyte, integrating processes such as photoabsorption, electron conduction, proton transport, and the diffusion of reactants and products is crucial. The design of a triple-phase boundary, including a “gas phase-electrolyte-semiconductor” boundary, plays a pivotal role in this gaseous PEC process in the absence of liquid water.

Furthermore, the PFSA ionomer and PEM serve as humidity adsorbents in the gas phase. Diffuse reflectance infrared spectroscopy revealed that the PFSA ionomer increased the amount of water adsorbed at low concentrations of water vapor and formed hydronium ions (H₃O⁺) with weakly hydrogen-bonded water molecules^[13]. The physical properties of water adsorption on semiconductor surfaces have been reviewed^[9]. The amount of water adsorbed is closely related to the RH in the system, and the adsorbed water layer affects the proton conductivity and number of contacts between water and the reaction sites. Further research is needed to separate the effects of water absorption in the ionomer and the effects of proton transport.

Titanium dioxide electrodes

The insights gained from the porous WO₃ electrode were applied to other semiconductor electrodes, including anatase and rutile TiO₂ electrodes. Anatase TiO₂ electrodes were prepared by electrochemically anodizing titanium fibers in ethylene glycol with 0.25% NH₄F and 10 vol% water at 50 V for 3 h; subsequently, they were crystallized through calcination in air at 550 °C for 1 h^[26]. The SEM image of the rectangular-shaped titanium fiber showed a layer of TiO₂ nanotube arrays with an outer diameter, a wall thickness, and a length of 110-140 nm, 15-30 nm, and ~5 µm, respectively [Figure 4A]. Stoll et al. also reported the anodization of Ti felt in ethylene glycol with 0.3% NH₄F and 2 vol% water at 30 V for 1 h, yielding TiO₂ nanotube arrays^[33]. In the nanotube structure, the photogenerated holes diffused through the wall thickness to the electrolyte interface over a short distance, while the photoexcited electrons were transported along the long axial direction of the TiO₂ nanotubes to the conductive substrate^[34].

Figure 4. (A) SEM image of the TiO₂ nanotube array formed by anodizing titanium fibers. (B) Membrane electrode assembly used in vapor-fed PEC water splitting in a humidified argon atmosphere (3 vol% H₂O), utilizing TiO₂ nanotube array photoanodes and Pt-carbon cathodes. The effect of PFSA ionomer coating number on their photocurrent response and IPCE at an applied voltage of 1.2 V vs. cathode (wavelength: 365 nm, irradiance: 56 mW cm^-2, irradiation area: 1 cm²) has also been included. Reproduced with permission from ref.^[26], Copyright 2018 Wiley.

In a water vapor environment (RH < 100%), the anatase TiO₂ nanotube arrays without ionomer functionalization initially exhibited no photocurrent response [Figure 4B]. Following the surface functionalization by the PFSA ionomer, the photocurrent monotonically increased with each subsequent coating (10 µL cm^-2 of a 5 wt% dispersion). After the third coating, the incident photon-to-current conversion efficiency (IPCE) remained nearly constant and was comparable to that of TiO₂ nanotube arrays in 0.2 mol L^-1 sodium sulfate solution (IPCE: 26.5%). This revealed that even under water vapor feeding, an ionomer modification of approximately 1.5 mg cm^-2 resulted in a performance equivalent to that in an aqueous solution^[26].

Figure 5A shows the vapor-fed PEC reaction over an anatase TiO₂ nanotube array photoanode. When an external voltage of 1.2 V was applied, the IPCE was 16% at a wavelength of 365 nm (40 mW cm^-2); the UV irradiation area was 16 cm², and the H₂ production rate was 600 µmol h^-1. The FE of H₂ generation at the cathode was close to 100%; however, the O₂ FE at the photoanode was only 86%, indicating that CO₂ was produced as a byproduct due to the oxidative decomposition of the PFSA ionomer. Energy-dispersive X-ray spectroscopy (EDS) revealed a decrease in the fluorine content on the TiO₂ surface, indicating the degradation of the ionomer during the vapor-fed PEC reaction. Similar results were obtained for the PEC reaction using a rutile TiO₂ electrode under water vapor feeding (RH < 100%)^[27]. Under a 385-nm UV illumination, the IPCE of rutile TiO₂ (8.7%) was much higher than that of the anatase TiO₂ nanotube array (4.8%) at an applied voltage of 1.2 V owing to the band gap of rutile (3.0 eV) being narrower than that of anatase (3.2 eV).

Figure 5. (A) Production rate of O₂ and CO₂ in the photoanodic compartment and H₂ in the cathodic compartment during vapor-fed PEC water splitting under humidified argon flow with a TiO₂ nanotube array photoanode (wavelength: 365 nm, irradiance: 40 mW cm^-2, and applied voltage: 1.2 V). (B) Impact of the carrier gas (humidified argon and air) in the photoanode compartment on the photocurrent response of the anatase TiO₂ nanotube array photoanode. Reproduced with permission from ref.^[26], Copyright 2018 Wiley.

The PEC water vapor splitting reaction under a humidified argon stream exhibited a high H₂ production rate (~600 µmol h^-1) at the laboratory level. In practical operations, it is necessary to utilize the humidity in the air [Figure 5B]. To confirm this hypothesis, the carrier gas in the photoanode chamber was switched from humidified Ar to humidified air. However, no difference was observed in the photocurrent response, indicating that water vapor can be oxidatively decomposed, even in the presence of O₂ in air^[26]. This finding highlights the potential of utilizing the porous OER photoanodes to harness atmospheric moisture for H₂ production.

Following the studies of our group, Zafeiropoulos et al. also reported the functionalization of porous TiO₂ and WO₃ photoanodes with proton-conducting PFSA ionomers (Aquivion and Nafion)^[35]. The optimized and functionalized photoanodes operating at 60% RH within a temperature range of 30-70 °C were able to recover up to 90% of the performance obtained at 1.23 V vs. RHE in the conventional liquid phase. The bare photoanodes, without ionomer functionalization, did not show a photocurrent response at RH < 100%, whereas Stoll et al. previously performed the PEM-PEC reaction under conditions in which water condensed (RH > 100%)^[33,36,37]. The photocatalytic water splitting reaction significantly depends on the presence of a condensed water film on the semiconductor surface for proton conductivity between the oxidation and reduction sites^[10]. Xu et al. reported a PEM-PEC system without ionomer functionalization, where liquid-like water layers formed on the surface of nanostructured TiO₂ in humid air^[38,39].

Strontium titanate electrode

Strontium titanate (SrTiO₃) is an effective photocatalyst used for water splitting. To prepare SrTiO₃ electrodes, the anatase TiO₂ nanotube arrays underwent further treatment by a hydrothermal reaction in 25 mM Sr(OH)₂ at 150 °C for 2 h and subsequent calcination in air at 550 °C for 1 h^[13]. Despite the collapse of the nanotube structure, particulate structures were obtained, in which the nanoparticles appeared to be interconnected to form mesoporous structures [Figure 6A]. The PEC properties of the SrTiO₃ electrode, evaluated in a 0.2 M Na₂SO₄ aqueous solution, revealed that the photocurrent onset potential shifted to the lower potential side in comparison to the anatase TiO₂ photoanode [Figure 6B]. This indicates that the PEC water oxidation reaction proceeds with the application of a smaller applied bias. Furthermore, the SrTiO₃ electrode exhibited a photocurrent response even at 0.3 V vs. RHE, whereas the TiO₂ electrode exhibited no photocurrent under the same conditions [Figure 6C].

Figure 6. Comparison of anatase TiO₂ nanotube array and SrTiO₃ photoanode. (A) SEM images of the top and side views of the oxide layers. (B) Cyclic voltammograms and (C) photocurrent response of anatase TiO₂ and SrTiO₃ photoanodes in aqueous 0.2 M Na₂SO₄ and 0.1 M NaPi solution (pH: 6.7, wavelength: 365 nm, and irradiance: 40 mW cm^-2) at varying applied potential (V vs. RHE). Reproduced from ref.^[13] with permission from the Royal Society of Chemistry.

Subsequently, PEC properties were evaluated under gas-phase conditions using a photoanode with an irradiated area of 2 cm² [Figure 7]. At an applied voltage of 0.3 V, a photocurrent response was observed for the anatase TiO₂ nanotubes and SrTiO₃ photoanodes with surface functionalization. H₂ was produced at the cathode with an FE of 100%. However, CO₂ was predominantly produced on the TiO₂ photoanode, indicating that the H₂ evolution is accompanied by the degradation of the PFSA ionomer rather than an overall water-splitting reaction^[13]. In contrast, the SrTiO₃ photoanode induced the production of H₂ and O₂ at a stoichiometric ratio of 2:1, and the produced CO₂ was undetectable (O₂ FE approached 100%). When utilizing the SrTiO₃ photoanode, water vapor completely decomposed into H₂ and O₂, and membrane separation was simultaneously achieved with the application of a small voltage. This is a major feature of the PEM-PEC reaction, which differs from photocatalytic reactions under water vapor feeding.

Figure 7. PEC properties of anatase TiO₂ and SrTiO₃ photoanodes with an irradiation area of 2 cm² in a vapor-fed water splitting system (wavelength: 365 nm and irradiance: 40 mW cm^-2). Impact of applied voltage (V vs. cathode) on (A) photocurrent response and (B) rate of H₂ formation in the cathode compartment and O₂ and CO₂ formation in the photoanode compartment. (C) Enlarged view of the product formation rate at 0.3 V vs. cathode. Reproduced from ref.^[13] with permission from the Royal Society of Chemistry.

The durability of the PEM-PEC system was evaluated over 10 h [Figure 8]. Functionalized SrTiO₃ photoanodes had a significant advantage in terms of their prolonged operational lifetime^[13]. In contrast, for anatase and rutile TiO₂ photoanodes, the surface ionomer coating responsible for proton transport was degraded under vapor-fed gas-phase conditions, which led to a gradual decrease in the photocurrent. The IPCE gradually decreased to 6.3% for the anatase TiO₂ photoanode. The O₂ evolution rate was lower than the rate of one-quarter of the electron transfer (e^-/4) owing to the formation of CO₂ from ionomer degradation. The fluorine/titanium (F/Ti) ratio, determined by SEM-EDS analysis, decreased from 1.64 to 0.41 after the vapor-fed PEC reaction, suggesting the degradation of the PFSA ionomer on the TiO₂ photoanode. The oxidative degradation is probably due to the hydroxyl radical generation on the TiO₂ surface. In contrast, the SrTiO₃ photoanode exhibited relatively stable photocurrents, maintaining an IPCE of 15.5%, even after a vapor-fed PEC reaction for 10 h. An analysis of the SEM-EDS images indicated that the F/Ti ratio remained constant after the reaction [Figure 8D]. The O₂ FE of SrTiO₃ was almost 100%, as confirmed by the O₂ formation rate, which was in agreement with the e^-/4 rate, while the CO₂ formation was negligible. The improved reaction selectivity for the OER over SrTiO₃ photoanodes increased the lifetime of the PEM-PEC system. A bias-free STH device will be developed by combination with a photocathode or by reducing the applied voltage of the photoanode in the PEM-PEC system.

Figure 8. Durability of anatase TiO₂ and SrTiO₃ photoanodes in vapor-fed PEC water splitting. (A) Photocurrent response (the numbers in the figure are the IPCE just before the light was turned off), (B) O₂ production rate, and (C) CO₂ production rate (wavelength: 365 nm, irradiance: 40 mW cm^-2, irradiation area: 16 cm², and applied voltage: 1.2 V vs. cathode). (D) SEM-EDS images of fluorine in the ionomer-coated SrTiO₃ photoanode before and after the PEC reaction. Reproduced from ref.^[13] with permission from the Royal Society of Chemistry.

Bismuth vanadate electrode

Bismuth vanadate (BiVO₄) with an E_g of 2.4 eV is recognized as a visible light-responsive photoanode. To enhance the performance of vapor-fed PEC water splitting under visible light irradiation, our group prepared porous BiVO₄ electrodes using Ti felt as a conductive substrate by a dip-coating of the precursor solution of Bi(NO₃)₃ and VO(acac)₂ and subsequent calcination at 500 °C^[28]. The photocurrent response of each porous photoanode was investigated in a phosphate buffer solution using a light source (blue LED) with a central wavelength of 454 nm [Figure 9A]. The performance of a single BiVO₄ was inferior to that of a single WO₃ (E_g = 2.7 eV), which was also responsive to blue light. When a BiVO₄ layer was coated onto the WO₃ photoanode, an increase in the photocurrent was observed. Upon doping Mo species into the BiVO₄ layer, the IPCE significantly increased to 10.7%, owing to the increased electron density resulting from the substitution of Mo⁶⁺ for V⁵⁺. Furthermore, the WO₃/Mo-BiVO₄ heterojunction photoanode exhibited a photocurrent response at a wavelength of approximately 500 nm, which is longer than that of the WO₃ photoanode [Figure 9B].

Figure 9. PEC properties of WO₃/Mo-doped BiVO₄ photoanodes under visible light irradiation. (A) Photocurrent response under 454 nm irradiation and (B) IPCE action spectra in aqueous 0.1 M NaPi solution at 1.2 V vs. RHE. (C) Photocurrent in the vapor-fed PEC water splitting using WO₃/Mo-BiVO₄ photoanode at an applied voltage of 1.2 V (light source: blue LED, wavelength: 454 nm, and irradiance: 6.5 mW cm^-2). Schematics of (D) the vapor-fed PEC system under visible light irradiation and (E) PEC water splitting with applied voltage, depicted as ΔE. Figures (A-D) were reproduced with permission from ref.^[28], Copyright 2020, American Chemical Society.

During the vapor-fed process in a PEM-PEC cell with the porous WO₃/Mo-BiVO₄ heterojunction photoanode, H₂ was produced with an IPCE of 10% at 454 nm and 1.2 V under blue-light irradiation [Figure 9C and D]. This was achieved with an FE of approximately 100 and 94% for H₂ and O₂ generation, respectively. However, the H₂ production rate in PEM-PEC cells gradually decreased during light irradiation. One cause of this degradation is the photocorrosive dissolution of BiVO₄ under acidic conditions^[40]. When using strongly acidic PEM or PFSA ionomers, it is crucial to consider the chemical stability of the semiconductor electrode materials, including the cocatalysts. Furthermore, a porous W-doped BiVO₄ photoanode demonstrated solar H₂ production from ambient air with humidity (RH: 100%) in the PEM-PEC reactor, retaining a current density of 1.55 mA·cm^-2 at 1.23 V vs. RHE under AM1.5 simulated solar light^[41]. For STH energy conversion, the applied voltage must be less than 1.23 V [Figure 9E]. Moreover, much research effort is needed for narrow-bandgap photoelectrodes with visible light absorption edges longer than 500 nm to increase STH efficiency.