Photoelectrocatalytic hydrogen generation coupled with reforming of glucose into valuable chemicals using a nanostructured WO3 photoanode

The study addresses how coupling glucose photo-oxidation at a WO3 photoanode with proton reduction at a cathode can enhance hydrogen generation in PEC water splitting. Conventional PEC water splitting is limited by oxygen evolution kinetics and the band-edge positions of stable oxide photoanodes (Fe2O3, BiVO4, WO3). WO3, with visible-light absorption and good charge transport, is promising but still requires bias. Replacing or complementing OER with oxidation of biomass-derived substrates could increase photocurrents and enable valuable co-products instead of CO2. Prior work showed photocurrent doubling on WO3 for organics and reported glucose oxidation with limited product identification. This work aims to optimize WO3 photoanode structure and electrolyte to maximize photocurrents via photocurrent doubling, assess stability, and direct glucose reforming toward valuable products while enabling efficient hydrogen evolution.

The paper reviews efforts to improve PEC water splitting by stabilizing and enhancing photoanode performance, often via OER catalysts (e.g., Co-Pi, Ni/Fe oxyhydroxides). Stable metal oxides (Fe2O3, BiVO4, WO3) have suitable solar absorption but require bias for unassisted splitting. Biomass-derived substrates have been explored to replace OER, mainly with TiO2 photoanodes or photocatalyst suspensions, but TiO2’s wide bandgap limits visible-light use and can lead to complex byproducts (e.g., aldol condensates in alkaline ethanol oxidation). Photocurrent doubling has been demonstrated on nanostructured WO3 for C1 and larger organics, offering enhanced efficiency due to long hole diffusion length and high electron mobility. Previous WO3 thin-film studies indicated glucose oxidation with limited mineralization and unidentified intermediates. Photocatalytic systems (e.g., noble-metal-loaded TiO2, doped TiO2, CdS quantum dots) showed product formation and H2 evolution but face issues like UV-only operation, noble-metal use, or toxicity. PEC configurations, such as BiVO4 oxidation of HMF to FDCA with simultaneous H2 generation, demonstrate the advantage of product separation and controlled anode environment. These findings motivate using visible-light-absorbing, stable WO3 photoanodes to co-generate H2 while upgrading glucose to valuable chemicals via mechanisms including photocurrent doubling.

WO3 photoanode fabrication: Nanostructured mesoporous WO3 films were synthesized by a sol–gel, doctor-blade, layer-by-layer deposition of a freshly prepared tungstic acid colloid containing PEG 300 onto FTO-coated glass, followed by annealing in oxygen (550–600 °C, typically 550 °C for 30 min after each layer). Films 1.2 to ~3 μm thick were obtained by 3–8 sequential depositions; single application yields ~0.4 μm. SEM showed porous films of 20–40 nm nanoparticles and agglomerates; XRD and Raman confirmed monoclinic WO3 with preferred orientation. For alkaline-compatible operation (pH 7), a ~0.4 μm TiO2 (P25) overlayer was deposited on 1.2 μm WO3 (doctor-blade from 1 g TiO2 in 20 mL DMF with 0.4 g PVDF; dried, 100 °C solvent evaporation, then annealed at 500 °C in O2) to form WO3/TiO2 hybrid electrodes. Photoelectrochemical setup: Three-electrode Teflon cell with quartz window; WO3 photoanode (exposed area typically 0.28 cm2 in j–E/IPCE measurements; ~0.7 cm2 in long-term electrolysis), Pt counter (grid), Ag/AgCl (sat. KCl) reference; Nafion membrane separated anode and cathode compartments in product studies. Illumination: simulated AM 1.5 G, 100 mW cm−2 (Oriel 150 W simulator). j–E measurements: potential swept from ~−0.40 VRHE (with glucose) to 1.6 VRHE at 10 mV s−1 using CHI 660E. IPCE: 150 W Xe lamp with monochromator (10 nm bandwidth), calibrated silicon detector; spectra to 480 nm. Electrolytes and conditions: Screening used 0.1 M NaCl, NaClO4, NaHSO4, and CH3SO3Na at pH 2, with and without 0.1 M glucose; anodes ~1.2 μm WO3. Prolonged electrolysis (20 h at 1.23 VRHE) evaluated in 45 mL anolyte with 0.1 M glucose in 0.5 M NaCl, 0.5 M Na2SO4, or 0.5 M CH3SO3Na at pH 4, and in 0.5 M NaCl at pH 7 (for pH 7, rear-side illumination through FTO was used on WO3/TiO2). Additional large-volume stability tests used 400 mL cell (WO3 area ~1.6 cm2). Analytical methods: Glucose oxidation products identified and quantified by GC-MS (Shimadzu GC-MS-QP2010 Ultra; ZB-5MSPlus column; EI, scan 45–500 m/z; temperature program 50–280 °C; derivatization via ethanethiol mercaptalation and BSTFA/TMCS silylation). For quantitative GC-MS, glucose peak was suppressed during 25.9–26.3 min to avoid detector overload. TOC analyses (Shimadzu TOC-5050A) measured non-purgeable organic carbon to assess mineralization/volatile formation. Sample prep included neutralization of strong-acid samples before evaporation to avoid saccharide mineralization, dilution for glucose quantification, and filtration prior to TOC.

• Nanostructured WO3 films (1.2–3 μm) exhibited strong visible response (IPCE up to ~80% at 360–380 nm for 1.2 μm; for ~3 μm films, IPCE at 440 nm nearly doubled vs 1.2 μm due to higher optical thickness). - In pH 2 electrolytes without glucose, ~1.2 μm WO3 showed photocurrent onset ~0.15 V vs Ag/AgCl (~0.47 VRHE) and plateau ~2.2 mA cm−2 at ~0.8 V vs Ag/AgCl; CH3SO3Na and NaHSO4 gave the highest plateaus. - Addition of 0.1 M glucose produced photocurrent doubling: plateaus increased to ~4.2 mA cm−2 for sulfate and chloride electrolytes, and onset shifted ~0.1–0.15 V more negative; IPCE roughly doubled vs supporting electrolyte alone. - With a ~3 μm WO3 film in 0.5 M NaCl (pH 4) + 0.1 M glucose, high plateau photocurrents near 6.5 mA cm−2 at ~1 V vs Ag/AgCl (~1.43 VRHE) were achieved under AM 1.5 G. - Long-term (20 h, 1.23 VRHE) photoelectrolysis with 0.1 M glucose showed stable operation with minor current decline (product accumulation in pores). TOC indicated similar non-purgeable organic carbon before and after electrolysis, implying minimal conversion to CO2 or volatiles. - Product distributions (Faradaic yields, FY): • pH 2 (acidic sulfate and chloride): dominant formation of gluconic acid; FYs ~16.5% (sulfate) and 14.6% (chloride). Other products (glucuronic acid, arabinose, erythrose) generally ~≤7.7% (erythrose in sulfate). Disaccharides detected but not identified. • pH 4 (0.5 M NaCl): erythrose >27% FY and gluconic acid ~20% FY; total FY of analyzed products ~52%. • pH 4 (0.5 M Na2SO4 or CH3SO3Na): increased arabinose and glucuronic acid vs NaCl, demonstrating electrolyte anion influence on selectivity. • pH 7 (0.5 M NaCl) with WO3/TiO2 under rear-side illumination: gluconic acid increased to ~33.5% FY; arabinose 4.2%, erythrose 25.5%, glucuronic acid 0.6%; total FY ~63.8% (~64%) at ~15% glucose conversion. - Mechanistic insights: Initial 2e− oxidation of glucose to gluconic acid proceeds via direct hole transfer to adsorbed glucose followed by electron injection (photocurrent doubling). Further oxidation steps likely involve electrolyte-derived radical intermediates (e.g., HSO4•, HOCl/Cl2 pathways), enabling formation of arabinose, erythrose, and glucuronic acid. - Stability: No electrode deactivation observed over extended runs; in acidic sulfate electrolytes, glucose presence mitigated WO3 deactivation otherwise associated with peroxo species formation.

The results demonstrate that coupling glucose oxidation at a WO3 photoanode significantly enhances PEC performance relative to water oxidation alone, primarily by kinetic facilitation and the photocurrent doubling mechanism. Increased photocurrents at modest bias (e.g., ~6.5 mA cm−2 at ~1.43 VRHE) indicate the feasibility of pairing the semitransparent WO3 photoanode with a single-junction PV for tandem, unassisted operation, projecting STH efficiencies >7%. Product analysis shows that electrolyte composition and pH tune the selectivity of glucose reforming: chloride media (NaCl) at pH 4–7 favor higher yields of erythrose and gluconic acid, while sulfate and methanesulfonate at pH 4 shift selectivity toward arabinose and glucuronic acid. Mechanistically, the initial formation of gluconic acid arises from direct hole oxidation with electron injection (doubling), whereas subsequent steps likely involve reactive intermediates from the supporting electrolyte (e.g., hypochlorous acid, sulfate radicals), consistent with observed differences in product distributions. The minimal mineralization observed by TOC supports selective upgrading rather than complete oxidation to CO2. The WO3/TiO2 hybrid enables operation at pH 7 while maintaining high IPCE and selectivity, expanding the operational pH window and further improving gluconic acid yield. Overall, the study validates a route to co-generate hydrogen and valuable chemicals from glucose in a PEC architecture with stable operation and tunable selectivity.

A semitransparent nanostructured WO3 photoanode efficiently photo-reforms glucose to valuable chemicals while driving hydrogen evolution in a PEC cell. Optimized film thickness (~3 μm) boosts visible absorption and, combined with photocurrent doubling, delivers high photocurrents (>6 mA cm−2 at 1.23 VRHE) under AM 1.5 G, enabling projected tandem STH >7%. Prolonged electrolysis demonstrated stable operation with minimal mineralization. Product selectivity is tunable via electrolyte anions and pH, achieving up to ~64% total Faradaic yield at ~15% glucose conversion (notably in 0.5 M NaCl, pH 7, with WO3/TiO2). Future work should focus on further optimizing selectivity, potentially via mixed-electrolyte strategies, exploring broader biomass-derived substrates, refining reactor geometry to increase conversion, and integrating with PV components for practical tandem devices.

• The stable operational pH range of monoclinic WO3 limits direct use above ~pH 4.6; operation at pH 7 required a TiO2 overlayer and rear-side illumination. - The extent of glucose conversion in 20 h runs was ~15%, constrained by electrode area-to-volume ratio and mass transport; scale-up and reactor optimization are needed. - Some products (e.g., disaccharides) were detected but not identified/quantified, leaving gaps in full carbon balance and selectivity mapping. - Mechanistic contributions of electrolyte-derived radicals vs direct hole transfer vary with conditions and are inferred rather than directly quantified. - Results were obtained under simulated sunlight; outdoor, long-term stability and performance under real solar conditions were not reported. - The study focused on glucose; generality to other biomass-derived substrates requires further validation.