Sustainable solar seawater splitting over a BaTaO₂N photoanode enabled by chloride recirculation

Abstract

This study demonstrates that BaTaO₂N photoanodes grown on Ta substrates via bottom-up fabrication exhibit highly efficient and stable photoelectrochemical (PEC) seawater splitting under simulated sunlight. Enhanced H₂ evolution is achieved, particularly in acidic to neutral pH conditions, where the oxygen evolution reaction (OER) typically proceeds at a slower rate. This improvement is attributed to the predominant chlorine evolution reaction (CER), facilitated by a kinetically favorable two-electron transfer pathway. The high crystallinity and cuboidal morphology of BaTaO₂N, combined with uniform Co(OH)_x catalyst loading, result in over 80% photocurrent retention and a faradaic efficiency close to 100% after continuous seawater splitting for 24 h. This outstanding efficiency is attributed to the distinct band structure of BaTaO₂N, which hinders the formation of hydroxyl radicals and selectively promotes oxidation through CER and OER. Remarkably, CER outperforms OER at pH levels below 7.5, while OER becomes more dominant at higher pH levels. Moreover, the free chlorine produced during CER decomposes gradually back to O₂ and Cl⁻ through photolytic and chemical processes, which ensures continuous regeneration of Cl⁻ in the bulk electrolyte. This self-sustaining cycle facilitates prolonged photoreaction within a fixed seawater volume, which eliminates the need for continuous seawater flow configurations. By overcoming a crucial obstacle in the scalable generation of solar H₂ from seawater, these results provide a promising route for sustainable and selective PEC seawater splitting.

---

1. Introduction

Hydrogen (H₂) is an environmentally friendly and renewable energy source with a high energy density, which makes it a promising alternative to problematic fossil fuels that emit carbon dioxide. Photoelectrochemical (PEC) water splitting over semiconductors has been a promising method for producing renewable H₂ from water, as demonstrated by Honda and Fujishima, who used a TiO₂ photoelectrode in 1972.^1–3 In freshwater, the oxygen evolution reaction (OER), which proceeds through a four-electron transfer, is significantly slower than the hydrogen evolution reaction (HER), which follows a two-electron pathway. This imbalance in reaction kinetics inevitably limits the overall rate of H₂ production during the water splitting process.³ Although alkaline water electrolysis has been proposed to address the sluggish OER, the cost of producing alkaline water remains a barrier to efficient hydrogen generation.⁴ In addition, freshwater makes up less than 3% of the Earth's total water supply. The World Health Organization has projected that by 2025, nearly two-thirds of the global population will live in water-stressed regions.⁵ In light of this, seawater (which is abundantly available) offers a compelling alternative. Utilizing sunlight to split seawater, rather than freshwater, could provide a feasible and sustainable pathway for H₂ production in the future.

Seawater is rich in halide salts, with chloride (Cl⁻) being the most prevalent at around 0.55 M. The thermodynamic potential of the chlorine evolution reaction (CER) at 1.36 V_NHE is more oxidizing than that of the OER at 1.23 V_NHE, which is reflected in the following equations:⁶

2Cl⁻ ⇌ Cl₂ (g) + 2e⁻ E° = 1.36 V_NHE (1)

2H₂O ⇌ O₂ (g) + 4H⁺ + 4e⁻ E° = (1.23 − 0.0592 pH) V_NHE (2)

Although CER is thermodynamically more oxidizing than OER during seawater splitting, it can be preferentially driven due to its kinetically faster two-electron transfer compared to the four-electron transfer of OER. CER is especially likely to compete with OER in environments where OER kinetics very slow, such as acidic or neutral electrolytes.^7–10 A nanostructured WO₃ photoanode produced chlorine (Cl₂) with an average faradaic efficiency of 70% during sunlight-driven seawater splitting in aqueous 0.5 M NaCl solution at pH 2, with O₂ as a minor byproduct.⁹ Our group demonstrated enhanced seawater splitting activity using perovskite oxynitrides ABO₂N (A = La, Sr; B = Ti, Nb) in a 0.5 M NaCl electrolyte at pH 6.4 under simulated sunlight.^10,11 At neutral pH, the anodic photoactivity (including CER and OER) in artificial seawater at an applied potential of 1.23 V_RHE was more than three times higher than that of OER in aqueous buffer solution, which indicates that CER was preferentially catalyzed. Quantitative analysis showed that a large amount of O₂ was produced, which corresponds to a faradaic efficiency of approximately 50% with hypochlorite (ClO⁻) detected as a minor species—even when CER proceeded more rapidly than OER.¹⁰ These results indicate that CER dominates in seawater, although its products vary depending on pH.

In seawater splitting, the oxidation of Cl⁻ produces Cl₂, hypochlorous acid (HOCl) and/or ClO⁻, depending on the pH (Fig. S1).¹² Cl₂ evolved from CER may dissolve in water and undergo hydrolysis to form HOCl, a weak acid that partly dissociates to ClO⁻ according to the following chemical equilibria:¹³

Cl₂ + 2H₂O ⇌ HOCl + H₃O⁺ + Cl⁻ K_h = 4.2 × 10⁻⁴ M (3)

HOCl + H₂O ⇌ ClO⁻ + H₃O⁺ K_a = 3.0 × 10⁻⁸ M (4)

The concentration of Cl⁻ has a strong effect on the fraction of HOCl produced in acidic pH environments (Fig. S1). CER produces mainly hydrolyzed HOCl in acidic seawater, including 0.5 M NaCl, while it releases Cl₂ in a concentrated 3.4 M NaCl solution employed for the industrial chlor-alkali process. In addition, the increase in pH gradually alters the product ratio of CER from HOCl to dissociated ClO⁻. In addition to the dependence of pH, CER products are partly oxidized to toxic chlorate (ClO₃⁻) and/or perchlorate (ClO₄⁻) during seawater splitting due to their own oxidizing power and/or the presence of hydroxyl radicals.^12,14,15 These products can also undergo partial photolysis under ultraviolet (UV) irradiation, forming O₂, chlorite (ClO₂⁻), ClO₃⁻, and/or Cl⁻.^8,15 Because of the diversity of products and the resulting undesirable chain reactions, the faradaic efficiency of seawater splitting has remained low, and the selectivity for CER becomes unclear. This limits the development potential of solar seawater splitting for efficient H₂ production. Moreover, while the pH dependence of electrochemical seawater splitting has been investigated, the effect of pH on the activity and selectivity of PEC seawater splitting for CER and/or OER has not yet been reported.^16,17 Therefore, it is necessary to investigate the reaction behavior between CER and OER during PEC water splitting at different pH levels, and to determine the faradaic efficiency and selectivity for CER through quantitative analysis of the actual products.

Perovskite oxynitrides with the general formula AB(O,N)₃ (A = La, Ca, Sr, or Ba; B = Ti, Nb, or Ta) can absorb a wide range of visible-light wavelengths up to 750 nm. They are also capable of driving various artificial photosynthesis reactions, including (sea)water splitting, due to their favorable band structures, which enables a theoretically high solar-to-fuel conversion efficiency.^6,18 Among the perovskite oxynitrides, BaTaO₂N with a bandgap energy (E_g) of 1.9 eV (λ < 660 nm) is the most promising semiconductor for (sea)water splitting because its photoactivity under simulated sunlight has been found to be the highest (the onset potential of approximately 0.6–0.7 V_RHE and 5–6 mA cm⁻² at 1.23 V_RHE) and significantly stable for 24 h (79% retention of initial photocurrent).^19–21 Nevertheless, its high photoactivity is still limited to strong alkaline electrolytes, such as pH 13, rather than neutral pH. The splitting of neutral seawater using BaTaO₂N under sunlight should, thus, be scientifically meaningful to discuss the pH dependence of CER on its selectivity and feasibility for efficient and stable H₂ evolution. In addition, the regeneration of Cl⁻ in the bulk electrolyte via decomposition of CER products (i.e., free chlorine species indicating Cl₂, HOCl, and ClO⁻) has not been explicitly investigated during long-term PEC seawater splitting. If the Cl⁻ consumed during CER can be replenished in situ by the decomposition of free chlorine species, it would enable a closed-loop PEC cell for sustainable H₂ evolution without the need for external Cl⁻ replenishment. Such a mechanism would not only enhance the stability and scalability of solar-driven seawater splitting but also mitigate the accumulation of harmful byproducts (e.g., ClO₃⁻, ClO₄⁻). On the other hand, a gradual decrease in bulk Cl⁻ concentration over time may reduce the diffusion-limited photocurrent, thereby lowering the rate of H₂ evolution even under stable light and operating potential.²² This highlights the critical importance of continuous Cl⁻ regeneration in maintaining steady-state reactions for long-term PEC performance.

In this study, we demonstrate the sustainable long-term seawater splitting using a highly crystalline cuboidal BaTaO₂N photoanode grown on a Ta substrate (prepared via bottom-up fabrication and subsequently loaded with a Co(OH)_x electrocatalyst) in artificial seawater under simulated (AM 1.5 G) sunlight. Prior to the prolonged photoreaction, the photoactivity and selectivity for CER of the synthesized BaTaO₂N were investigated in aqueous 0.5 M NaCl electrolytes at various pH values, which were compared with those in aqueous buffered electrolytes. High faradaic efficiency and selectivity for CER were also estimated during the PEC seawater splitting of oxynitride—as determined by quantitative analyses using gas chromatography (GC) and colorimetric detection. In addition, the photo-assisted decomposition studies of CER products, i.e., free chlorine species, provide critical insights into previously ambiguous findings (particularly the unexpected excess O₂ gas detected in quantitative analyses). These observations provide valuable insights into the underlying mechanisms that govern long-term seawater splitting. Based on these results, this study aims to identify the optimal optical properties of semiconductors and the electrolyte pH values that enable enhanced solar-driven seawater splitting. Under these conditions, a Cl⁻ regeneration mechanism will be investigated. Ultimately, this work proposes a set of photoreaction strategies to sustain a continuous H₂ evolution rate and offers practical guidelines for achieving efficient and durable solar seawater splitting.

2. Experimental section

2.1. Synthesis of the BaTaO₂N/Ta photoanode

A particulate BaTaO₂N photoanode was synthesized via bottom-up fabrication, including the oxidation of a metallic substrate and subsequent nitridation, as previously reported.¹⁰ A Ta-sheet with an area of 0.9 × 1.0 cm² (0.1 mm in thickness, 99.95%, Nilaco Corp.) was ultrasonically cleaned with isopropyl alcohol, rinsed with distilled water, dried naturally, and then UV-ozone treated for 10 min. For electrochemical anodization, the two cleaned Ta sheets were mounted in parallel in glycerol (99.0%, Samchun Chemical) dissolved with 0.8 M K₂HPO₄ (99.0%, Samchun Chemical). Then, a bias of 15 V was applied between the two Ta sheets for 10 s at a constant temperature of 453 K. The anodized Ta substrate was rinsed with distilled water and dried naturally. Subsequently, the particulate BaTaO₂N was grown on the anodized Ta substrate coated with Ba precursor using two-step nitridation. Five grams of Ba(NO₃)₂ (99.0%, Kanto Chemical) were dissolved in 30 mL of ethylene glycol (99.9%, Samchun Chemical) mixed with distilled water in a 50 : 50 vol%. Thirty microliters of the Ba(NO₃)₂ solution were uniformly drop-cast onto the anodized Ta substrate, which was followed by annealing in air at 413 K for 5 min. The resulting Ta substrate was nitrided in a 100 mL min⁻¹ flow of high-purity NH₃ (6 N grade) at 1273 K for 1 h, with a ramping rate of 10 K min⁻¹. After the first nitridation, loosely bonded BaTaO₂N particles on the substrate were removed by ultrasonic treatment in a water bath. The surface-modified substrate then underwent a second nitridation under the same conditions. The final substrate, which exhibited a reddish-brown color, is hereafter referred to as the BaTaO₂N/Ta photoanode.

2.2. Photoelectrochemical seawater splitting

Prior to PEC measurements, the surface of BaTaO₂N/Ta photoanode was modified using an electrocatalyst Co(OH)_x—widely reported to lower the OER overpotential and facilitate interfacial hole transfer—via impregnation and then annealed in air at 573 K for 2 h.²³ The activity of the Co(OH)_x/BaTaO₂N/Ta photoanode for seawater splitting was measured in a conventional three-electrode system that was connected to a potentiostat (WPG100e, WonATech Co., Ltd), under chopped AM 1.5 G simulated sunlight (XES-50S2-TT, San-EI Electric). The prepared photoanode, a Pt wire, and an Ag/AgCl electrode were used as the working, counter, and reference electrodes, respectively, in a two-compartment PEC cell, which was divided by a Nafion membrane (N115, Chemours Company). The photoanode and Ag/AgCl electrode were placed in one compartment, and the Pt wire was mounted in a different compartment of the PEC cell. A 0.2 M potassium phosphate (KPi) buffer electrolyte was used in both compartments, and 0.5 M NaCl was added to the photoanode side for seawater oxidation. Different pH values in both buffer electrolytes, composed of H₃PO₄ (85.0%, Kanto Chemical), KH₂PO₄ (99.5%, Kanto Chemical), K₂HPO₄ (99.0%, Kanto Chemical), or K₃PO₄ (98.0%, Junsei Chemical), were adjusted using the Henderson–Hasselbalch equation. The electrode potential in 0.5 M NaCl aqueous electrolytes was converted to values versus a reversible hydrogen electrode (RHE) using the Nernst equation,

E_RHE = E_Ag/AgCl + 0.0592 × pH + 0.197 (5)

Linear sweep voltammograms (LSVs) were recorded by cathodically scanning the potential from 1.4 to 0.4 V_RHE at a scan rate of 10 mV s⁻¹. The Mott–Schottky (MS) analysis of BaTaO₂N/Ta photoanode was carried out in Ar-saturated 0.5 M NaCl electrolyte buffered with 0.2 M KPi at pH 5.5 in the dark, at different frequencies (0.5, 1.5, and 2.5 kHz). The quantitative amounts of gaseous H₂ and O₂ that evolved from a Pt wire and the Co(OH)_x/BaTaO₂N/Ta photoanode during seawater splitting were estimated using GC analysis (YL6500, Young In Chromass). Moreover, the quantitative amounts of free chlorine species (Cl₂(aq)/HOCl/ClO⁻), which were produced during the photoreaction, were determined using the colorimetric detection method (V-770, JASCO) utilizing a quartz cuvette with a path length of 1.0 cm, based on the oxidation of 3,3′,5,5′-tetramethyl benzidine (TMB).²⁴ The amounts of Cl⁻ and byproduct ClO₃⁻ traces were estimated using ion chromatography (Dionex ICS-5000, Thermo Fisher Scientific) in an alkaline environment.

2.3. Structural characterization

The crystal structure of BaTaO₂N/Ta photoanode was determined using X-ray diffraction (XRD; MiniFlex 600, Rigaku) with Cu Kα radiation at 40 kV and 15 mA. The optical property of BaTaO₂N grown on a fluorine-doped tin oxide (FTO) substrate was determined using ultraviolet-visible diffuse reflectance spectroscopy (UV-Vis DRS; V-770, JASCO). The valence band maximum (VBM) potential of BaTaO₂N grown on Ta substrate was estimated via ultraviolet photoelectron spectroscopy (UPS; AXIS Supra, Shimadzu) with He I radiation (21.2 eV) and the following equation:

E_B + E_K + Φ_analyzer = 21.2 eV (6)

Here, E_B, E_K, and Φ_analyzer represent the binding energy referenced to the Fermi level, the kinetic energy of electron, and the work function of analyzer, respectively. Furthermore, the surface morphology and elemental mapping of the BaTaO₂N/Ta and Co(OH)_x/BaTaO₂N/Ta photoanodes were examined using a scanning electron microscope (SEM; Gemini500, ZEISS) equipped with an energy-dispersive X-ray spectroscope (EDS).

3. Results and discussion

Fig. 1(a) shows the XRD pattern of oxynitride particles grown on a Ta substrate that was prepared through bottom-up fabrication. The observed peaks were primarily assigned to perovskite-type BaTaO₂N (ICSD #62619), accompanied by metallic Ta and Ta₂N peaks corresponding to the substrate and its nitrided phase, respectively. The presence of metallic Ta₂N suggests that nitrogen diffused into the conductive substrate beyond the BaTaO₂N layer during nitridation, similar to the formation of NbN and/or Nb₂N interlayers observed during the synthesis of SrNbO₂N on Nb substrates.¹⁰ A full-width-at-half-maximum (FWHM) of approximately 0.17 was observed for the (110) phase, which indicates high crystallinity of the BaTaO₂N particles. The high crystallinity was achieved through an optimized two-step nitridation (1 h + 1 h) at 1273 K, which was found to be more effective than a single-step treatment under the same total duration (Fig. S2). Its crystallinity was higher compared to that of the SrNbO₂N/Nb photoanode, which was synthesized using the bottom-up fabrication in a similar manner.¹⁰ The high crystallinity of BaTaO₂N may result from nitridation at 1273 K, which is significantly higher than the 1173 K used for SrNbO₂N. Its high crystallinity also led to strong absorbance in the visible-light region, as shown in the UV-vis DRS spectrum in Fig. 1(b). The absorption wavelength edge of the highly crystalline BaTaO₂N appeared at approximately 650 nm, consistent with previously published results.^19,20 This optical property (responsive across a wide range of visible wavelengths) is highly favorable for sunlight-driven water splitting because it enables a theoretical solar-to-hydrogen energy conversion efficiency of approximately 21%. Therefore, the actual band structure of the oxynitride synthesized via bottom-up fabrication was determined via various analyses (Fig. S3). According to the Tauc plot, the bandgap energy, E_g, of the oxynitride was determined to be 2.06 eV, corresponding to a direct allowed transition. Its flat band potential, E_fb, and VBM potential, E_VB, were estimated to be −0.28 V_NHE (−4.12 eV) and 1.6 V_NHE (−6.0 eV), respectively, based on Mott–Schottky (MS) and UPS spectral analyses. Frequency-dependent variations of the MS slope are expected for semiconductor/electrolyte interfaces, because the measured capacitance includes not only the depletion (space-charge) term but also Helmholtz/double-layer and surface-state (trap/redox) contributions with finite time constants. At lower frequencies these states exchange charge (increasing the apparent capacitance and reducing the slope), while at higher frequencies they cannot follow and the response approaches the intrinsic space-charge capacitance. Accordingly, the donor density of BaTaO₂N was extracted from the linear depletion region at an intermediate frequency where dispersion is minimal (1.5 kHz), giving N_D ≈ 4.3 × 10¹⁸ cm⁻³ (assuming ε ≈ 200 for single crystal BaTaO₂N). Using E_g, E_fb, and E_VB, the band diagram of BaTaO₂N grown on Ta was constructed, as shown in Fig. 1(c), indicating that its band edges straddle both the water redox potentials and the CER potential. Notably, the potential for water oxidation to produce hydroxyl radicals (˙OH) lies significantly more positive than the VBM potential of BaTaO₂N, as summarized in the following equation:^14,25

H₂O ⇌ ˙OH (aq) + H⁺ + e⁻ E° = 2.73 V_NHE (7)

Fig. 1 (a) XRD pattern for the BaTaO₂N/Ta photoanode synthesized through bottom-up fabrication. The empty and filled inverted triangles represent the impurity phases Ta and Ta₂N, respectively. (b) UV-vis DRS spectrum of BaTaO₂N grown on quartz glass. (c) Band structure diagram for the synthesized BaTaO₂N/Ta photoanode.

In n-type BaTaO₂N, upward band bending is formed at the interface between semiconductor and electrolyte, driving photogenerated holes toward the surface and facilitating oxidation reactions including OER and CER. Although this interfacial electric field enhances charge transfer, the E_VB of BaTaO₂N remains more negative than the potential required for ˙OH formation. Moreover, the ˙OH production at the surface of photoanodes generates chlorine radicals, such as Cl˙ and Cl₂˙⁻, during seawater splitting, which causes complex chain reactions and toxic, undesirable byproducts (ClO₃⁻, ClO₄⁻, and so on).^14,26 Therefore, these results demonstrate that BaTaO₂N crystals grown on Ta have the potential to enable sunlight-driven seawater splitting, encompassing OER and CER, while simultaneously suppressing PEC side reactions.

Fig. 2 shows SEM images and EDS elemental mapping results of BaTaO₂N particles synthesized through bottom-up fabrication. The oxynitride particles, which grew randomly on the Ta substrate, had an average size of 430 nm (Fig. S4), featuring smooth and cuboidal surfaces. Then, uneven, stick-type Co(OH)_x nanoparticles applied to enhance photoreactions were covered at the smooth surface of oxynitrides as depicted in Fig. 2(c). The cross-sectional image of the photoanode reveals that the BaTaO₂N layer on Ta was approximately 2 μm thick. An EDS elemental mapping analysis indicated that the cuboidal oxynitrides consisted of Ba, Ta, O, and N elements, which were uniformly distributed within the oxynitride layer. For the synthesized BaTaO₂N/Ta photoanode, the surface elemental composition ratios of Ba/Ta = 1.1 and O/N = 1.8—while not reflecting bulk composition—were nevertheless close to the theoretical stoichiometry (Fig. S5).

Fig. 2 Top-view SEM images of (a and b) BaTaO₂N/Ta and (c) Co(OH)_x/BaTaO₂N/Ta photoanodes synthesized via bottom-up fabrication. (d) Cross-sectional SEM image and (e) corresponding elemental EDS mapping of the prepared BaTaO₂N/Ta photoanode.

Based on the theoretically possible optical properties of the BaTaO₂N/Ta photoanode, the PEC seawater splitting activity was evaluated over the oxynitride surface loaded with Co(OH)_x electrocatalyst under chopped simulated sunlight (AM 1.5 G) in aqueous 0.5 M NaCl solutions buffered with 0.2 M KPi at different pH levels ranging from 1 to 13 (Fig. S6(a) and (b)). The neutral pH was chosen based on the pK_a 7.5 of HOCl (eqn (4)). Although phosphate buffers were prepared with a total concentration of 0.2 M at each pH, the ionic strength of electrolyte slightly varied due to differences in phosphate speciation depending on pH. The seawater splitting of the oxynitride was additionally conducted at pH 5.5 using a 0.1 M KPi (Fig. S7). The photo-response was almost identical to that obtained with 0.2 M buffer, indicating that the variation in ionic strength has a negligible effect on the photoactivity. The photoactivities were compared with the splitting activity of pure water in aqueous 0.2 M KPi electrolytes, which provides a reference to evaluate the pH dependency of both the seawater splitting activity of the BaTaO₂N/Ta photoanode and the selectivity of oxidation reactions. Fig. 3(a) and (b) show bar charts illustrating the onset potentials (for 0.1 mA cm⁻²) and photocurrent densities at 1.23 V_RHE of Co(OH)_x/BaTaO₂N/Ta photoanodes at different pH levels, extracted from LSV curves acquired during the (sea)water splitting (Fig. S6). Across all pH values, the onset potentials of the BaTaO₂N photoanodes for seawater splitting were more negative compared to those for water splitting. For water splitting, the photocurrent onset exhibited a clear pH dependence; across pH 1–13 it shifted by approximately 0.16 V_RHE: as pH increased to 13, OER kinetics became faster and the required overpotential decreased, shifting the onset to more negative potentials, whereas at lower pH the slower OER kinetics shifted the onset positively. By contrast, in the NaCl-containing electrolyte, although the bulk concentration of Cl⁻ did not vary with pH (unlike OH⁻), the onset potential for seawater splitting also exhibited pH dependence, but with a smaller shift (approximately 0.1 V_RHE) than that observed for OER. Moreover, their photocurrent densities at 1.23 V_RHE were consistently high during seawater splitting in aqueous 0.5 M NaCl solutions, irrespective of the pH. It is widely acknowledged that the OER is readily catalyzed in highly alkaline electrolytes like pH 13, even in the presence of 0.5 M NaCl.^27,28 However, in this study, the presence of NaCl in the aqueous solution at pH 13 further enhanced the photoactivity, including the onset potential and photocurrent, of the oxynitride photoanode, which suggests that CER as well as OER were concurrently occurring during its seawater splitting.

Fig. 3 Bar charts for (a) onset potentials (for 0.1 mA cm⁻²) and (b) photocurrent densities at the applied potential of 1.23 V_RHE, extracted from LSV curves obtained during seawater splitting of Co(OH)_x/BaTaO₂N/Ta photoanodes in aqueous 0.5 M NaCl electrolytes buffered with 0.2 M KPi and adjusted to different pH values under AM 1.5 G simulated sunlight. Seawater splitting is compared with water splitting in aqueous 0.2 M KPi electrolytes. (c) LSV curves obtained during (sea)water splitting of the oxynitride at pH 5.5 values in darkness and under AM 1.5 G simulated sunlight. (d) A bar chart showing faradaic efficiencies estimated via quantitative analyses of products, i.e., H₂, O₂ gases, and free chlorine species (Cl₂(aq)/HOCl/ClO⁻) evolved during solar seawater splitting of the oxynitride at different pH levels.

At low pH values such as 1 and 2, the seawater splitting activity over the oxynitride was remarkably high compared to its OER activity. Its onset potential was negatively shifted by as much as 0.2 V_RHE, and the photocurrent was also more than twice as high. These results clearly demonstrate that seawater splitting over BaTaO₂N photoanodes was significantly more photoactive than pure water splitting at low pH, and that CER driven by two-electron transfer is kinetically faster than OER via the four-electron pathway despite its more oxidizing thermodynamic potential. Such fast kinetics of seawater splitting were similarly observed with increasing pH values up to the neutral pH 7.5. This is remarkable because the increase in pH accelerates the hydration of Cl₂ originating from CER to HOCl and its subsequent dissociation to ClO⁻ (Fig. S1), and the thermodynamic potentials of CER become much more positive according to the following equations:^6,12

Cl⁻ + H₂O ⇌ HOCl + H⁺ + 2e⁻ E° = (1.48 − 0.0592 pH) V_NHE (8)

Cl⁻ + 2OH⁻ ⇌ ClO⁻ + H₂O + 2e⁻ E° = (1.72 − 0.0592 pH) V_NHE (9)

In fact, even though CER requires a thermodynamic potential more positive, at least 0.25 V (eqn (8)), than 1.23 V_RHE for OER, it was kinetically favored during seawater splitting at pH 5.5. Using sulfite oxidation as a hole-scavenging benchmark, the charge injection efficiency, η_inj, was quantified from LSV curves for seawater and water splitting at pH 5.5 (Fig. S8). At 1.36 V_RHE, η_inj reached nearly 80% for seawater splitting but only 46% for water splitting (OER), indicating kinetically faster interfacial charge transfer for CER. The (sea)water splitting activity of Co(OH)_x/BaTaO₂N/Ta photoanodes was measured in the dark at the potential range higher than 1.23 V_RHE, as shown in Fig. 3(c). The seawater splitting kinetics was much faster than water splitting kinetics, consistent with the PEC result of the oxynitride. The electrochemical seawater splitting of the catalyst Co(OH)_x loaded on FTO substrate, evaluated at different pH values, also showed markedly faster kinetics, particularly at pH values below 7.5 (Fig. S6(c) and (d)). These results indicate that the fast kinetics of seawater splitting, which depends on pH, was not limited to PEC configurations. Nevertheless, the acquisition of photocurrent initiating from approximately 0.7 V_RHE clearly demonstrates the effectiveness of PEC seawater splitting, which is largely enhanced at low pH values, i.e., unfavorable environments for driving OER.

The products obtained during the seawater splitting activity of Co(OH)_x/BaTaO₂N/Ta at the applied potential of 1.23 V_RHE for 1 h were quantitatively analyzed using GC and colorimetric detection. The anodic current density remained relatively stable, without sudden degradation during the photoreaction, regardless of measured pH conditions (Fig. S9). Moreover, the quantitative amounts of evolved products resulting from seawater splitting at pH 5.5 were repeatedly estimated every 15 min and compared with those from pure water splitting (Fig. S10). The gaseous H₂ was detected in the Pt wire compartment, whereas gaseous O₂ and liquid HOCl were observed in the compartment of the Co(OH)_x/BaTaO₂N/Ta photoanode. The amounts of evolved gases increased linearly during stable seawater splitting for 1 h, being two times larger than those during pure water splitting—consistent with the photocurrent result shown in Fig. 3(b). The faradaic efficiency of seawater oxidation (CER and/or OER) was determined by the total amounts of O₂ and half of HOCl evolved from four- and two-electron transfer, respectively (Fig. S10(c)). The faradaic efficiency of seawater splitting was almost unity, demonstrating that the anodic photocurrent acquired over the BaTaO₂N photoanode indeed resulted from sunlight-driven seawater splitting. It is worth noting that the amount of HOCl product was negligibly detected in the initial 15 min and increased over time, while a significant amount of O₂ was estimated and decreased over time. The amount of HOCl product, exclusively corresponding to approximately 14% of faradaic efficiency, was estimated in 1 h photoreaction. These results contradicted the faster kinetics of CER in seawater splitting compared to OER. In addition, the faradaic efficiency of seawater oxidation (estimated by total detection of O₂ and HOCl) was slightly lower (approximately 2%) than that of seawater reduction producing H₂. Tiny ClO₃⁻ traces were also detected from the electrolyte obtained after seawater splitting for 1 h using ion chromatography (IC) in an alkaline environment (Fig. S10(d)). Although it is unclear whether the HOCl originating from CER was partly oxidized to ClO₃⁻ during the photoreaction or the alkaline IC measurement, the presence of ClO₃⁻ traces would slightly lower the faradaic efficiency of seawater oxidation. Therefore, the seawater oxidation, which involves only CER and OER and exhibits a high faradaic efficiency (approximately 98%) in this study, was attributed to the optical band structure of BaTaO₂N, which is incapable of generating ˙OH photoelectrochemically—thus preventing undesirable chain side reactions.

Fig. 3(d) shows the faradaic efficiencies of seawater splitting operated for 1 h at different pH levels (1, 2, 7.5, 10, and 13 as well as pH 5.5), which were determined by quantitative analysis of the evolved products in the same manner as above (Fig. S11). In this study, the identification of Cl₂(aq)/HOCl/ClO⁻ products originating from CER was determined by calculating the composition fraction of free chlorine species, as determined by pH dependence (Fig. S1). The seawater splitting of Co(OH)_x/BaTaO₂N/Ta photoanode resulted in faradaic efficiencies exceeding 95% at all pH values, generating free chlorine species through CER and O₂. These results clearly prove that the increase in anodic photocurrent, driven by the rapid kinetics of CER, was fully utilized to boost the H₂ production rate, especially at low pH levels. At pH 1 and 2, the dissolved amount of Cl₂(aq) was relatively high (corresponding to a faradaic efficiency of approximately 25%) compared to the levels of HOCl/ClO⁻ detected under conditions with a pH above 5.5. This difference may be attributed to the higher seawater splitting activity in comparison to OER under lower pH conditions. Interestingly, a measurable amount of ClO⁻ was observed at a high pH of 13, where OER kinetics is considered dominant, despite a slight increase in photocurrent, as shown in Fig. 3(b). Conversely, it is noteworthy that the detected amount of O₂ remained considerably high regardless of pH, exceeding a faradaic efficiency of 83% even at pH 5.5, where rapid CER kinetics were observed. Nevertheless, considering the high faradaic efficiency, the fact that the high photocurrent observed during seawater splitting in neutral and acidic electrolytes indeed enhanced the rate of H₂ production is highly significant.

The photo-assisted decomposition of free chlorine species was investigated at different pH values to determine why a high proportion of O₂ gas was observed in quantitative analyses of sunlight-driven seawater splitting, as mentioned above. The constant concentrations of free chlorine species in aqueous 2.7 μM NaOCl solutions, buffered with 0.2 M KPi at different pH values, were irradiated under AM 1.5 G simulated sunlight, and the remaining amount of free chlorine functioned with time was quantitatively estimated using colorimetric detection (Fig. S12). Fig. 4(a) shows decomposition curves of the free chlorine species at different pH values (2, 5.5 and 7.5). The free chlorine species are well known to undergo decomposition through several mechanisms, such as photolysis, chemical disproportionation, and thermolysis.^14,29 In this study, considering the substantial amount of O₂ detected after seawater splitting, it is most plausible that the detected O₂ originated from the decomposition of HOCl (which is the predominant free chlorine species at pH below 7.5). The decomposition of HOCl is typically attributed to either a spontaneous chemical reaction—i.e., disproportionation (eqn (10))—or radical-mediated photolysis (eqn (11)), as illustrated by the following reaction pathways:^29–31

2HOCl → 2Cl⁻ + O₂ + 2H⁺ (10)

2HOCl → Cl₂O + H₂O (10-1)

Cl₂O + H₂O → 2Cl⁻ + O₂ + 2H⁺ (10-2)

2HOCl + hv → 2Cl⁻ + O₂ + 2H⁺ (11)

HOCl + hv → ˙OH + ˙Cl (11-1)

4˙OH → 2H₂O₂ → O₂ + 2H₂O (11-2)

˙Cl + H₂O → Cl⁻ + ˙OH + H⁺ (11-3)

Fig. 4 (a) Decomposition curves of free chlorine species in aqueous 2.7 μM NaOCl electrolytes buffered with 0.2 M KPi adjusted to different pH levels of 2, 5.5, and 7.5 under AM 1.5 G simulated sunlight. (b) Decomposition curves of free chlorine species at pH 5.5 under AM 1.5 G simulated sunlight equipped with and without a 400 nm cut-off filter (λ > 400 nm). (c) O₂ quantification curve simultaneously estimated via GC analysis during the decomposition of free chlorine species at pH 5.5 under AM 1.5 G simulated sunlight.

In addition, the free chlorine species itself was found to act as an oxidant and further reacts to produce ClO₃⁻ as shown in the following reaction:¹⁴

2HOCl + ClO⁻ → ClO₃⁻ + 2Cl⁻ + 2H⁺ (12)

The presence of ClO₃⁻ byproduct was qualitatively determined after seawater splitting at pH 5.5 for 1 h using ion chromatography, as discussed in Fig. S10. Nevertheless, it was evident that the free chlorine was primarily decomposed to O₂ because only a negligible amount of ClO₃⁻ was detected in this study. As shown in Fig. 4(a), the decomposition rate of free chlorine species increased with decreasing pH, which indicates strong pH dependence. The fast decomposition of major species HOCl at lower pH may be attributed to its increased oxidizing power and the acid-favored formation of reactive intermediate Cl₂O (eqn (10-2)), which facilitates the subsequent O₂ evolution via chemical disproportionation.^32,33 In addition, the photolysis of HOCl was promoted under acidic conditions due to its higher molar absorptivity and more efficient radical generation compared to its conjugate base, ClO⁻, thereby accelerating O₂ evolution via a radical-mediated pathway.³⁴ Since the initial decomposition rate showed a linear relationship in the plot of log[free chlorine] versus time, the reaction can be described as pseudo-first-order. In general, the HOCl photolysis is known to follow pseudo-first-order kinetics with respect to its concentration.^34,35 Therefore, the initial stage of O₂ evolution is likely driven primarily by the photodecomposition of HOCl.

The decomposition rate of HOCl decreased gradually over time, following an initially rapid degradation phase, regardless of the pH. The photolysis of HOCl is known to be wavelength-dependent, with rapid O₂ generation driven by ˙OH formation upon absorption of ultraviolet (UV) light (eqn (11-1)), particularly at shorter wavelengths (around 250 nm).^33–35 The simulated sunlight (AM 1.5 G) used in this study provided only UV light with wavelengths above approximately 350 nm, which inherently limited the photodecomposition rate of HOCl. Fig. 4(b) shows the decomposition behavior of free chlorine under visible light irradiation (λ > 400 nm) at pH 5.5 using a 400 nm cut-off filter. The decomposition continued even under visible light exposure, and an analysis of the initial decomposition rate revealed a linear relationship in the plot of 1/[free chlorine] versus time, which suggests a second-order reaction. This finding is consistent with the second-order kinetics of spontaneous chemical decomposition of HOCl with respect to its concentration.^30,31 In this study, both photolysis and spontaneous chemical degradation of free chlorine species co-occurred under simulated sunlight irradiation. Fig. 4(c) shows the quantification of O₂ gas generated during the sunlight-assisted decomposition of free chlorine at pH 5.5 through GC analysis. Nearly complete decomposition was observed at the initial stage within 1 h, followed by a period of stable O₂ evolution, and a gradual decrease during the final phase of decomposition. The decrease in decomposition rate over time, as indicated by the slope of the decomposition curve, likely reflects the concentration dependence of free chlorine decomposition—through chemical pathways following second-order kinetics and/or photolysis proceeding via a pseudo-first-order pathway—for O₂ evolution. In other words, the reduced concentration of free chlorine led to a slower decomposition rate. However, the prolonged exposure of the dilute free chlorine under sunlight appeared to result in other, undetected decomposition pathways rather than the continued O₂ evolution. It presents the involvement of additional pathways in the decomposition, possibly including oxidation reactions that yield byproducts such as ClO₃⁻, as qualitatively detected by IC in this study. Despite these parallel decomposition routes, the quantitative result clearly shows that O₂ gas was the dominant product through the decomposition of free chlorine species under simulated sunlight. In other words, the decomposition data provide compelling evidence why the predominant portion of products was identified as gaseous O₂ during seawater splitting of the BaTaO₂N photoanode under neutral and acidic pH environments—even though CER was kinetically more favorable than OER in seawater.

The reactant Cl⁻ for CER is regenerated through the subsequent decomposition of free chlorine species as a CER product into O₂ gas (eqn (10) and (11)). Therefore, long-term solar seawater splitting was investigated to explore the feasibility of continuous H₂ production under a constant Cl⁻ concentration. Fig. 5(a) displays the photocurrent curve of a Co(OH)_x/BaTaO₂N/Ta photoanode functioned with time acquired during seawater splitting for 24 h at the applied potential of 1.23 V_RHE in an aqueous 0.5 M NaCl solution buffered with 0.2 M KPi under AM 1.5 G simulated sunlight. The seawater splitting of BaTaO₂N was operated at pH 5.5, where the highest O₂ evolution, corresponding to a faradaic efficiency of more than 83%, was observed as discussed in Fig. 3(d). Simultaneously, the concentration of reactant Cl⁻ in the aqueous electrolyte was quantified using both IC in an alkaline environment and the Mohr titration method based on silver nitrate. After 24 h of operation, approximately 81% of the initial photocurrent was retained, which demonstrates the high photocurrent stability of the oxynitride photoanode. In contrast, the previously reported SrNbO₂N photoanode, prepared via a bottom-up fabrication, exhibited limited long-term activity for PEC seawater splitting.¹⁰ Whereas prior BaTaO₂N photoanodes have shown high photoactivity and, in some cases, appreciable long-term stability under strongly alkaline electrolytes such as pH 13, sustained long-term operation under neutral pH environment has not been reported previously (Table S1). Thus, in this work, the markedly enhanced stability of the BaTaO₂N photoanode at pH 5.5 is particularly noteworthy. The improvement can be attributed to the more uniform deposition of Co(OH)_x on the smooth, cuboidal BaTaO₂N particles (Fig. 2), contrasting with the porous morphology of the SrNbO₂N particles. The enhanced crystallinity of BaTaO₂N crystals, as discussed in Fig. 1, may also facilitate efficient charge separation during the photoreaction. Therefore, the uniform loading of the electrocatalyst on highly crystalline BaTaO₂N particles likely contributed to the sustained photocurrent and enhanced long-term durability during the seawater splitting. Furthermore, the amounts of H₂, O₂, and HOCl produced during the photoreaction were quantitatively analyzed in the same manner as above, as shown in Fig. 5(b). The faradaic efficiencies of oxidation (O₂ and HOCl) and reduction (H₂) reactions consistently remained above 98% throughout the 24 h operation, respectively. On average, the crystalline, cuboidal Co(OH)_x/BaTaO₂N/Ta photoanode produced H₂ at a rate of 31 μmol h⁻¹ cm⁻². These results clearly indicate that the photocurrent generated during prolonged seawater splitting predominantly contributes to H₂ evolution, which demonstrates that sustainable H₂ production through solar seawater splitting is feasible even under competitive CER and OER pathways.

Fig. 5 Solar seawater splitting performance of the Co(OH)_x/BaTaO₂N/Ta photoanode. (a) Chronoamperometry curve recorded during 24 hour seawater splitting at an applied potential of 1.23 V_RHE in Ar-saturated aqueous 0.5 M NaCl electrolyte, buffered with 0.2 M KPi and adjusted to pH 5.5, under AM 1.5 G simulated sunlight. The Cl⁻ concentration was simultaneously monitored using the Mohr titration method. (b) Time profiles of O₂ and free chlorine (HOCl), and H₂ generation over the Co(OH)_x/BaTaO₂N/Ta photoanode, and a Pt wire during solar seawater splitting. Dashed lines represent the expected quantities of H₂, HOCl (e⁻/2), and O₂ (e⁻/4) assuming 100% faradaic efficiency. (c) Bar chart showing faradaic efficiencies estimated from the quantitative analyses of the products shown in (b).

In this study, the selective contribution of CER to the measured photocurrent could not be directly determined due to the rapid initial decomposition of free chlorine species. However, a bar chart showing faradaic efficiencies over time, displayed in Fig. 5(c), reveals that HOCl was detected with a faradaic efficiency of approximately 3% in the initial 15 min of seawater splitting, which gradually increased to 23% over 24 h. This result indicates that CER continued to proceed during prolonged seawater splitting and that the accumulation of its products became more significant over time. In contrast, O₂ gas accounted for more than 95% of the faradaic efficiency in the early stage but decreased to approximately 75% by 24 h. As shown in Fig. 5(a), the Cl⁻ concentration remained nearly constant at 0.5 M during the initial phase but exhibited decreasing behavior over time. Nonetheless, the final Cl⁻ concentration remained significantly higher than what would be expected if CER had proceeded with complete conversion of Cl⁻ to free chlorine, which was not decomposed to O₂ gas. These findings collectively suggest that the majority of O₂ generated during the initial stage resulted from the rapid decomposition of free chlorine, as indicated by the high rate of Cl⁻ regeneration. However, the observed decrease in O₂ production over time, along with the concurrent increase in free chlorine accumulation, implies that the formation of free chlorine by CER became more pronounced relative to the decomposition rate of its products as the reaction proceeds. As a result, the slow decomposition rate of free chlorine increasingly hindered the complete regeneration of Cl⁻ as the photoreaction proceeded.

An intriguing aspect is that the faradaic efficiency of seawater oxidation remained nearly constant at approximately 98% for 24 h. It indicates that the unique band structure of BaTaO₂N, which is insufficient to generate ˙OH, predominantly produces free chlorine and O₂ gas driven by CER and OER. The decomposition of free chlorine primarily led to the evolution of O₂, with other side reactions occurring at negligible levels. As discussed in Fig. 4, the gradual decrease in the concentration of free chlorine over time reduced its decomposition rate to O₂ while facilitating alternative (yet unidentified) decomposition pathways. In contrast, during the prolonged seawater splitting process of 24 h, the continuous generation and accumulation of free chlorine instead resulted in a concentration-dependent acceleration of its decomposition to O₂, which suppressed competing decomposition pathways. This finding indicates that under the steady-state photoreaction, the sustained formation of free chlorine by CER promoted preferential O₂ production over other undesired byproducts.

In addition, although a high faradaic efficiency of nearly 100% for H₂ evolution was maintained for 24 h, a gradual decline in photocurrent resulted in a decrease in the H₂ evolution rate. In general, the depletion of reactant Cl⁻ in the bulk electrolyte could decrease the diffusion-limited photocurrent.²² In this study, despite the initial recirculation of Cl⁻ concentration during the early stage of the photoreaction, the photocurrent gradually decreased. Furthermore, although the recirculation of Cl⁻ became slower as the reaction proceeded, the decreasing rate of the photocurrent remained essentially unchanged. These observations suggest that while Cl⁻ recovery contributed to photocurrent behavior, it was not the sole determining factor in its long-term stability. After 24 h of photoreaction, a relatively high surface coverage of Co(OH)_x remained on the BaTaO₂N surface (Fig. S13). However, the SEM analysis revealed that the Co(OH)_x particles became smaller and more porous compared to their initial state prior to the reaction, as shown in Fig. 2. The decreasing photocurrent may be mainly attributed to the corrosion of the Co(OH)_x electrocatalyst during the photoreaction. Therefore, in order to sustain a stable H₂ evolution rate, it is essential to not only identify a corrosion-resistant electrocatalyst for CER but also to maintain a consistent concentration of Cl⁻ in artificial seawater. To address this problem, a flowing system of seawater could be introduced to replenish Cl⁻; however, this approach may introduce spatial limitations and incur additional costs for seawater splitting. Alternatively, this study demonstrates that the reactant Cl⁻ was gradually replenished in the bulk electrolyte during solar seawater splitting, although the decomposition of free chlorine into O₂ occurred slowly. In conclusion, as shown in Fig. 6, if the generation rate of free chlorine during CER can be matched with its decomposition rate into O₂, a steady Cl⁻ concentration can be maintained, which enables a stable long-term H₂ production rate under constant volume of seawater (similar to conventional water splitting systems).

Fig. 6 Schematic illustration of Cl⁻ regeneration and closed-loop operation during solar seawater splitting using a BaTaO₂N photoanode under sunlight.

4. Conclusions

This study provides clear experimental evidence that BaTaO₂N photoanodes, grown on Ta substrates via bottom-up fabrication, enable highly efficient and stable PEC seawater splitting under simulated sunlight. Most notably, these oxynitrides demonstrated significantly enhanced H₂ evolution activity in seawater at acidic to neutral pH conditions, where the sluggish OER would otherwise limit performance. This enhancement was attributed to the preferential CER, which is driven by a kinetically favorable two-electron transfer mechanism. As a result, the anodic photocurrent, amplified by the fast CER kinetics despite its thermodynamic disadvantage, was effectively harnessed for H₂ production, achieving a faradaic efficiency approaching 100%. This exceptional efficiency stems from the favorable band structure of BaTaO₂N, which suppresses hydroxyl radical formation and selectively enables oxidation via CER and OER. The selectivity of CER over OER was especially pronounced below pH 7.5, where free chlorine evolution increased, while OER became dominant at higher pH. These pH-dependent trends confirm that CER kinetics play a central role in enhancing H₂ evolution, particularly under low-pH conditions. In addition, the high crystallinity and well-controlled morphology of BaTaO₂N, combined with uniform Co(OH)_x catalyst loading, supported long-term PEC operation, retaining over 80% of the initial photocurrent after continuous seawater splitting for 24 h.

Crucially, this work also reveals that free chlorine species generated via CER gradually decompose back to O₂ and Cl⁻ through photolytic and chemical pathways, which enables in situ regeneration of Cl⁻ in the bulk electrolyte. Although this decomposition is not immediate, the accumulation of free chlorine over time is self-limiting due to its eventual breakdown. This closed-loop mechanism enables the sustained long-term storage of seawater in a fixed electrolyte volume, eliminating the need for flowing seawater configurations. Finally, this study addresses a significant bottleneck in solar H₂ production from saline water: the challenge of selectivity and stability in non-alkaline media. The combination of CER-dominant oxidation pathways under mildly acidic to neutral pH and the self-regenerating behavior of Cl⁻ presents a practical and scalable route to efficient, durable, and sustainable solar-driven H₂ generation directly from seawater.

Conflicts of interest

The authors declare no conflicts of interest relevant to this study.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information (SI) is available: analysis results for the band structure of BaTaO₂N, PEC and EC activities, quantitative analysis results at different pH. See DOI: https://doi.org/10.1039/d5qi01561h.

Acknowledgements

This research was financially supported by Basic Science Research Program (No. RS-2023-00243439) and Global-Learning & Academic Research Institution for Master's PhD students, and Postdocs (LAMP) Program (No. RS-2024-00442775) through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (MOE).

References

1. A. Fujishima and K. Honda, Electrochemical Photolysis of Water at a Semiconductor Electrode, Nature, 1972, 238, 37–38.
2. T. Hisatomi, J. Kubota and K. Domen, Recent advances in semiconductors for photocatalytic and photoelectrochemical water splitting, Chem. Soc. Rev., 2014, 43, 7520–7535.
3. C. Jiang, S. J. A. Moniz, A. Wang, T. Zhang and J. Tang, Photoelectrochemical devices for solar water splitting - materials and challenges, Chem. Soc. Rev., 2017, 46, 4645–4660.
4. K. Zeng and D. Zhang, Recent progress in alkaline water electrolysis for hydrogen production and applications, Prog. Energy Combust. Sci., 2010, 36, 307–326.
5. S. Jjoshi, F. Ilievski and J. Pujara, Knowledge-Powered Recommendation for an Improved Diet Water Footprint, ArXiv, 2024, 2403, 17426.
6. J. Seo and K. Domen, Perovskite and layered perovskite oxynitrides for efficient sunlight-driven artificial synthesis, Mater. Chem. Front., 2024, 8, 1451–1479.
7. W. Luo, Z. Yang, Z. Li, J. Zhang, J. Liu, Z. Zhao, Z. Wang, S. Yan, T. Yu and Z. Zou, Solar hydrogen generation from seawater with a modified BiVO4 photoanode, Energy Environ. Sci., 2011, 4, 4046.
8. S. Iguchi, Y. Miseki and K. Sayama, Efficient hypochlorous acid (HClO) production via photoelectrochemical solar energy conversion using a BiVO4-based photoanode, Sustainable Energy Fuels, 2018, 2, 155–162.
9. M. Jadwiszczak, K. Jakubow-Piotrowska, P. Kedzierzawski, K. Bienkowski and J. Augustynski, Highly Efficient Sunlight–Driven Seawater Splitting in a Photoelectrochemical Cell with Chlorine Evolved at Nanostructured WO3 Photoanode and Hydrogen Stored as Hydride within Metallic Cathode, Adv. Energy Mater., 2019, 10, 1903213.
10. V.-H. Trinh and J. Seo, Porous Cuboidal SrNbO2N Crystals Grown on a Nb Substrate as an Active Photoanode for Neutral Seawater Splitting under Sunlight, ACS Sustainable Chem. Eng., 2023, 11, 1655–1665.
11. T. T. T. Tran and J. Seo, Enhanced Neutral Seawater Splitting on Less-Defective, Two-Dimensional LaTiO2N Photoanodes Prepared from Layered Perovskite BaLa4Ti4O15, J. Mater. Chem. A, 2024, 12, 17128–17141.
12. R. K. B. Karlsson and A. Cornell, Selectivity between Oxygen and Chlorine Evolution in the Chlor-Alkali and Chlorate Processes, Chem. Rev., 2016, 116, 2982–3028.
13. N. N. Greenwood and A. Earnshaw, Chemistry of the Elements, Butterworth-Heinemann, 2nd edn, 1997.
14. M. S. Koo, X. Chen, K. Cho, T. An and W. Choi, In Situ Photoelectrochemical Chloride Activation Using a WO3 Electrode for Oxidative Treatment with Simultaneous H2 Evolution under Visible Light, Environ. Sci. Technol., 2019, 53, 9926–9936.
15. B. Rao, N. Estrada, S. McGee, J. Mangold, B. Gu and W. A. Jackson, Perchlorate production by photodecomposition of aqueous chlorine solutions, Environ. Sci. Technol., 2012, 46, 11635–11643.
16. S. Shin, T. U. Wi, T. H. Kong, C. Park, H. Lee, J. Jeong, E. Lee, S. Yoon, T. H. Kim, H. W. Lee, Y. Kwon and H. K. Song, Selectively Enhanced Electrocatalytic Oxygen Evolution within Nanoscopic Channels Fitting a Specific Reaction Intermediate for Seawater Splitting, Small, 2023, 19, e2206918.
17. X. Wu, S. Zhou, Z. Wang, J. Liu, W. Pei, P. Yang, J. Zhao and J. Qiu, Engineering Multifunctional Collaborative Catalytic Interface Enabling Efficient Hydrogen Evolution in All pH Range and Seawater, Adv. Energy Mater., 2019, 9, 1901333.
18. J. Seo, H. Nishiyama, T. Yamada and K. Domen, Visible-Light-Responsive Photoanodes for Highly Active, Stable Water Oxidation, Angew. Chem., Int. Ed., 2018, 57, 8396–8415.
19. K. Ueda, T. Minegishi, J. Clune, M. Nakabayashi, T. Hisatomi, H. Nishiyama, M. Katayama, N. Shibata, J. Kubota, T. Yamada and K. Domen, Photoelectrochemical oxidation of water using BaTaO2N photoanodes prepared by particle transfer method, J. Am. Chem. Soc., 2015, 137, 2227–2230.
20. J. Seo, M. Nakabayashi, T. Hisatomi, N. Shibata, T. Minegishi and K. Domen, Solar-Driven Water Splitting over a BaTaO2N Photoanode Enhanced by Annealing in Argon, ACS Appl. Energy Mater., 2019, 2, 5777–5784.
21. X. Luo, Y. Xiao, B. Zhang, C. Feng, Z. Fan and Y. Li, Direct synthesis of BaTaO2N nanoparticle film on a conductive substrate for photoelectrochemical water splitting, J. Catal., 2022, 411, 109–115.
22. J. G. Vos, Z. Liu, F. D. Speck, N. Perini, W. Fu, S. Cherevko and M. T. M. Koper, Selectivity Trends Between Oxygen Evolution and Chlorine Evolution on Iridium-Based Double Perovskites in Acidic Media, ACS Catal., 2019, 9, 8561–8574.
23. J. Seo, T. Hisatomi, M. Nakabayashi, N. Shibata, T. Minegishi, M. Katayama and K. Domen, Efficient Solar–Driven Water Oxidation over Perovskite–Type BaNbO2N Photoanodes Absorbing Visible Light up to 740 nm, Adv. Energy Mater., 2018, 8, 1800094.
24. Y. Guo, Q. Ma, F. Cao, Q. Zhao and X. Ji, Colorimetric detection of hypochlorite in tap water based on the oxidation of 3, 3′, 5, 5′-tetramethyl benzidine, Anal. Methods, 2015, 7, 4055–4058.
25. S. Siahrostami, G. L. Li, V. Viswanathan and J. K. Norskov, One- or Two-Electron Water Oxidation, Hydroxyl Radical, or H₂O₂ Evolution, J. Phys. Chem. Lett., 2017, 8, 1157–1160.
26. J. Kiwi, A. Lopez and V. Nadtochenko, Mechanism and Kinetics of the OH-Radical Iintervention during Fenton Oxidation in the Presence of a Significant Amount of Radical Scavenger (Cl-), Environ. Sci. Technol., 2000, 34, 2162–2168.
27. M. Benedet, G. A. Rizzi, O. I. Lebedev, V. Roddatis, C. Sada, J.-L. Wree, A. Devi, C. Maccato, A. Gasparotto and D. Barreca, Advances in photo-assisted seawater splitting promoted by green iron oxide-carbon nitride photoelectrocatalysts, J. Mater. Chem. A, 2023, 11, 21595–21609.
28. L. Yu, Q. Zhu, S. Song, B. McElhenny, D. Wang, C. Wu, Z. Qin, J. Bao, Y. Yu, S. Chen and Z. Ren, Non-noble metal-nitride based electrocatalysts for high-performance alkaline seawater electrolysis, Nat. Commun., 2019, 10, 5106.
29. C. K. Remucal and D. Manley, Emerging investigators series: the efficacy of chlorine photolysis as an advanced oxidation process for drinking water treatment, Environ. Sci.: Water Res. Technol., 2016, 2, 565–579.
30. L. C. Adam, I. Fabian, K. Suzuki and G. Gordon, Hypochlorous acid decomposition in the pH 5–8 region, Inorg. Chem., 2002, 31, 3534–3541.
31. D. M. Bulman, S. P. Mezyk and C. K. Remucal, The Impact of pH and Irradiation Wavelength on the Production of Reactive Oxidants during Chlorine Photolysis, Environ. Sci. Technol., 2019, 53, 4450–4459.
32. D. Angyal, I. Fabian and M. Szabo, Kinetic Role of Reactive Intermediates in Controlling the Formation of Chlorine Dioxide in the Hypochlorous Acid-Chlorite Ion Reaction, Inorg. Chem., 2023, 62, 5426–5434.
33. Y. Shiraishi, Y. Shimabukuro, K. Shima, S. Ichikawa, S. Tanaka and T. Hirai, Sunlight-Driven Generation of Hypochlorous Acid on Plasmonic Au/AgCl Catalysts in Aerated Chloride Solution, JACS Au, 2023, 3, 1403–1412.
34. Y. Feng, D. W. Smith and J. R. Bolton, Photolysis of aqueous free chlorine species (HOCl and OCl-) with 254 nm ultraviolet light, J. Environ. Eng. Sci., 2007, 6, 277–284.
35. H. Zhou and D. Xiao, Role of NOM in the Photolysis of Chlorine and the Formation of Reactive Species in the Solar/Chlorine System, ACS Omega, 2022, 7, 7769–7776.

---