Unravelling a new many-body large-hole polaron in a transition metal oxide that promotes high photocatalytic activity

The study explores how many-body interactions between charge carriers and lattice distortions (polarons) influence electronic transport and photocatalytic activity in oxides. Polarons have been implicated in phenomena such as superconductivity, metal–insulator transitions, and ferroelectricity, and are theorized to affect photocatalysis by aiding charge separation and suppressing recombination. BiVO₄ is a promising oxide photocatalyst due to its visible-light absorption and favorable conduction band edge, but undoped BiVO₄ exhibits low conductivity attributed to small-electron polaron hopping. Doping with W or Mo significantly enhances conductivity and photocatalytic performance, which cannot be explained by small-polaron hopping. Theory suggests monoclinic BiVO₄ can host a large-hole polaron that could promote photocatalysis, but this had not been experimentally observed. Efficient solar energy conversion also benefits from a small/indirect bandgap that can assist polaron-mediated charge separation via electron–phonon coupling, reducing recombination and increasing diffusion lengths. This work aims to experimentally reveal a many-body large-hole polaron in W-doped BiVO₄, elucidate its electronic/crystal origins, its interplay with an indirect bandgap, and demonstrate how these factors govern high photocatalytic activity.

Prior theoretical and experimental studies indicate: (1) Polarons play key roles in correlated oxides and may aid photocatalysis by separating photoexcited charges and preventing recombination. (2) Oxide semiconductors such as BiVO₄, TiO₂, SrTiO₃, and WO₃ show notable photocatalytic activity; however, BiVO₄’s transport is often limited by small-electron polarons. (3) Doping BiVO₄ with Mo or W enhances conductivity and photocatalytic activity, inconsistent with small-polaron hopping alone. (4) Theory predicts formation of a large-hole polaron in monoclinic BiVO₄ that could explain improved photoactivity and suggests that an indirect bandgap, assisted by lattice interactions (phonons), further aids charge separation. The energy-conversion benefit of narrowing the bandgap (e.g., 2.5 to 2.3 eV yielding up to ~40% increase in theoretical maximum photocurrent at 100% IPCE) underscores the importance of band-structure engineering in tandem with polaron physics.

Materials and film growth: W-doped BiVO₄ ceramic targets (0–5 wt% W) were synthesized via solid-state reaction from Bi₂O₃, V₂O₅, and WO₃ (calcined at 900 °C for 5 h; pressed into targets). Epitaxial BiVO₄ thin films with 0%, 0.5%, 1%, 2%, and 5% W were grown on YSZ (001) substrates by pulsed laser deposition (KrF 248 nm, 2 Hz, 80 mJ/cm²). Substrate temperature: 600 °C; O₂ partial pressure: 100 mbar. Structural characterization: Synchrotron X-ray diffraction (XDD beamline, SSLS), including 0–2θ scans and reciprocal-space mapping (HK-plane near (−204)) to assess strain, twinning, and phase (orthorhombic/monoclinic/tetragonal) evolution. Spectroscopies: High-resolution spectroscopic ellipsometry (SE) to obtain complex dielectric function ε(ω) over a broad energy range, enabling analysis of spectral weight transfer (SWT) and identification of midgap/polaronic states; X-ray absorption spectroscopy (XAS) at V L₃,₂- and O K-edges to probe unoccupied states, ligand-field splitting, hybridization (O 2p with V 3d and Bi 6sp), and hole-related prepeaks; X-ray photoemission spectroscopy for valence/chemical states (evidence of multivalent W⁵⁺/W⁶⁺). Photocatalysis: Films immersed in aqueous 0.2 M Na₂S + 0.3 M Na₂SO₃ sacrificial electrolyte in a 200 mL vessel; irradiated with visible light. H₂ evolution quantified by gas chromatography (packed MolSieve 13X column; He carrier; helium ionization detector) by sampling 100 µL headspace aliquots every 3 hours. First-principles calculations: DFT (+U) using Quantum ESPRESSO with optimized norm-conserving Vanderbilt pseudopotentials. Monoclinic BiVO₄ 24-atom cell; hole-doped by removing one electron with compensating jellium. Cutoff: 75 Ry. Structural relaxation to forces <1e−3 Ry/Bohr; SCF threshold 1e−6 Ry. k-point meshes: 6×6×6 (optimization) and 12×12×12 (DOS). Optimized lattice parameters: a = 7.151 Å, b = 11.358 Å, c = 5.056 Å, β = 135.03°. Hubbard U applied to O 2p and V 3d to capture strong correlation effects. XAS/XLD multiplet simulations performed with CTM4XAS for comparison.

- Structural evolution with W doping: Reciprocal-space maps reveal phase changes with doping. Pure BiVO₄ shows an orthorhombic, twinned structure with a = 5.101 ± 0.003 Å, b = 5.192 ± 0.001 Å, c = 11.697 ± 0.001 Å; average in-plane diagonal ≈ 7.275 Å (matching YSZ ≈ 7.278 Å). With light W doping (0.5–1%), twinning weakens and the film becomes monoclinic; above 2% W, films are fully strained tetragonal with no twinning. - Photocatalytic performance: Monoclinic 1% W-doped BiVO₄ exhibits the highest H₂ evolution under visible light, achieving approximately 3–4× higher activity than other phases/dopings measured over 15 h. - Discovery of a many-body large-hole polaron: XAS at the V L₃-edge shows a strong, polarization-dependent prepeak at ~514.8 eV whose intensity is maximized at 1% W; this feature corresponds to a hole-related midgap state indicative of a large-hole polaron. Multiplet simulations cannot account for the polarization-dependent prepeak, supporting its hole-polaron nature. The main V L₃-edge peaks at ~515.8, ~516.8, and ~517.9 eV reflect ligand-field splitting of unoccupied V 3d states. - O K-edge XAS shows hybridization and anisotropy: Low-energy (<531 eV) features correspond to V 3d–O 2p hybridized states (triplet split), while high-energy (≥531 eV) features arise from O 2p–Bi 6sp hybridization. The 1% W-doped sample shows the largest Bi 6p/s energy splitting, indicating strong anisotropy due to BiO₈ dodecahedral distortion. The hole-related hybrid states occupy the lowest absorption onset, and the large-hole polaron intensity peaks at 1% W. - Spectroscopic ellipsometry and spectral weight transfer (SWT): ε₂(ω) is partitioned into regions: I (Drude), II (indirect transitions and large-hole polaron), III (direct transitions), IV (higher-energy interband). A new midgap absorption appears at ~2.63 eV (region II). From 0% to 1% W, SWT shifts from region III to II and IV; at 5% W, spectral weight in region II transfers back to regions III and I. This anomalous SWT over a broad energy range evidences strong electronic correlations and screening effects. - Correlation-driven band-structure changes: On-site Coulomb interactions in O 2p (Upp) and V 3d (Udd), enhanced by hole creation (via W doping and vanadium vacancies), along with distortions of BiO₆/BiO₈ units, lift Bi 6s toward the valence band maximum, producing an indirect bandgap and enabling formation of a localized large-hole polaron midgap state composed of O p hybridized with V d and Bi sp. - Optimal doping and transport: The large-hole polaron state has weak binding energy, preserving band-like transport for electrons/holes and suppressing recombination. The 1% W-doped monoclinic phase shows the most pronounced polaron features, strongest indirect bandgap characteristics, reduced O 2p–Bi 6s/6p hybridization, lifted Bi 6p/6s at the valence top, improved conductivity, and the highest photocatalytic activity.

The work addresses how many-body interactions—specifically the formation of a large-hole polaron coupled to lattice distortions—govern charge transport and photocatalytic activity in BiVO₄. W doping introduces multivalent W (W⁵⁺/W⁶⁺), stabilizes V vacancies, creates holes, and modifies the crystal field via distortion, enabling a large-hole polaron that manifests as a midgap state. SE reveals broad spectral weight transfers, a hallmark of electronic correlations and screening, showing that states near the Fermi level are strongly renormalized by high-energy bands through on-site Coulomb interactions (Upp, Udd). XAS at V L- and O K-edges directly ties the polaron to O 2p–V 3d–Bi 6sp hybridization and to anisotropic BiO polyhedral distortions that lift Bi 6s, creating an indirect bandgap. The indirect gap, assisted by the polaron, facilitates charge separation by involving phonons in absorption/recombination processes, thereby reducing electron–hole recombination and increasing diffusion lengths. The combination of weakly bound, large-radius hole polarons (preserving band-like transport) and an indirect bandgap maximizes photocatalytic efficiency, consistent with the observed 3–4× enhancement in H₂ evolution for the 1% W-doped monoclinic phase. Excessive W doping (e.g., 5%) alters structure and hybridization, redistributing spectral weight away from the polaronic region and diminishing photocatalytic gains, highlighting the delicate balance among charge, orbital, and lattice degrees of freedom.

This study experimentally reveals a many-body large-hole polaron in W-doped BiVO₄ and clarifies its interplay with an indirect bandgap as key to achieving high photocatalytic activity. Using SE, XAS, and XRD supported by DFT+U, the authors show that W-induced holes, vanadium vacancies, and BiO polyhedral distortions drive strong correlations (Upp, Udd), lift Bi 6s toward the valence-band maximum, and produce an indirect bandgap with a localized midgap polaron state composed of O p–V d–Bi sp hybridization. The effect is maximized at 1% W in monoclinic BiVO₄, which demonstrates 3–4× higher H₂ evolution, improved conductivity, and suppressed recombination due to weakly bound large-hole polarons retaining band-like transport. Future work could optimize polaron–bandgap interplay via controlled strain, defect engineering, and dopant concentration; extend time-resolved spectroscopies to directly track polaron dynamics; and explore analogous mechanisms in other correlated oxide photocatalysts.

Limitations are not explicitly detailed in the provided text. The study focuses on thin-film samples on YSZ substrates and specific W doping levels (0–5%), and relies on DFT+U and spectroscopic proxies for polaron identification; broader compositional ranges, substrate effects, and direct time-resolved measurements of carrier dynamics are not discussed.