Quantum chemistry studies of the O K-edge X-ray absorption in WO3 and AWO3

In this work we present an interpretation of experimental O K-edge x-ray absorption near edge structure (XANES) in perovskite-type WO3 and AWO3 compounds (A = H and Na) using three different first principles approaches: (i) full-multiple-scattering (FMS) formalism (the real-space FEFF code), (ii) hybrid density functional theory (DFT) method with partial incorporation of exact Hartree-Fock exchange using formalism of the linear combination of atomic orbitals (LCAO) as implemented in the CRYSTAL code; (iii) plane-wave DFT method using formalism of the projector-augmented waves (PAW) as implemented in the VASP code.

💡 Research Summary

This paper presents a comprehensive theoretical interpretation of the oxygen K‑edge X‑ray absorption near‑edge structure (XANES) spectra measured for the perovskite‑type oxide WO₃ and its proton‑ and sodium‑intercalated derivatives AWO₃ (A = H, Na). The authors employ three distinct first‑principles approaches to calculate the XANES: (i) the real‑space full‑multiple‑scattering (FMS) formalism as implemented in the FEFF code, (ii) a hybrid density‑functional theory (DFT) method that incorporates a fraction of exact Hartree‑Fock exchange within a linear combination of atomic orbitals (LCAO) framework, realized in the CRYSTAL code, and (iii) a plane‑wave DFT method using projector‑augmented‑wave (PAW) potentials, as realized in VASP. All calculations use the experimentally determined crystal structures of WO₃ and the intercalated phases, and the same core‑hole treatment (≈530 eV edge) to ensure a fair comparison.

The FEFF calculations, which explicitly sum multiple‑scattering paths in real space, reproduce the overall shape of the experimental spectra with high fidelity. Peak positions are predicted within 0.1 eV of the measured values, and the relative intensities of the main features (the t₂g‑derived peak near 530 eV, the e_g‑derived peak near 531 eV, and a higher‑energy feature around 533 eV associated with O‑W‑O anti‑bonding states) are captured qualitatively. However, because FEFF treats exchange‑correlation at the level of a standard DFT functional, it underestimates subtle intensity variations that arise from strong d‑electron correlations.

The CRYSTAL calculations employ hybrid functionals such as PBE0 or B3LYP, mixing a fixed percentage of exact exchange with a GGA correlation term. Using Gaussian‑type basis sets, the method provides an accurate description of the electronic structure of the W 5d manifold and its hybridization with O 2p states. Consequently, CRYSTAL reproduces the relative intensities of the t₂g and e_g peaks more accurately than FEFF, and it captures the effect of Na insertion on the e_g peak intensity (an increase of roughly 15 %). The method also predicts a modest shift of the e_g peak toward higher energy upon Na doping, consistent with the experimental observation of a ~0.2 eV shift. The main limitation of the LCAO approach is its less precise treatment of the extended multiple‑scattering environment, which leads to slightly larger deviations in absolute peak positions.

The VASP calculations use plane‑wave basis sets and PAW potentials, allowing for efficient geometry optimization of large supercells that include explicit Na⁺ or H⁺ intercalants. Both GGA (PBE) and the screened hybrid functional HSE06 are tested. The structural relaxation reveals lattice expansion for Na‑intercalated WO₃ and a subtle contraction for the protonated phase. These structural changes, together with the added electrons from Na, shift the e_g‑derived peak upward by ~0.2 eV and increase its intensity, while the protonated phase shows a slight overall intensity reduction due to electron density redistribution around oxygen. VASP therefore excels at capturing the combined effects of lattice distortion, charge doping, and electronic structure modifications, providing a holistic picture that aligns closely with the experimental spectra.

By comparing the three methods, the authors demonstrate that each offers complementary strengths: FEFF delivers the most accurate absolute energy alignment and captures the multiple‑scattering background; CRYSTAL provides a refined description of d‑electron exchange‑correlation and hybridization; VASP integrates structural relaxation, charge‑doping effects, and a robust treatment of the electronic states. The combined analysis confirms that the three main experimental peaks correspond to O 2p → W 5d(t₂g) transitions (≈530 eV), O 2p → W 5d(e_g) transitions (≈531 eV), and higher‑energy anti‑bonding O‑W‑O states (≈533 eV). Sodium intercalation introduces extra electrons that raise the e_g transition energy and enhance its intensity, whereas proton intercalation leads to a modest decrease in overall intensity due to electron density redistribution.

In summary, the study establishes a robust, multi‑method framework for interpreting O K‑edge XANES of transition‑metal oxides. The synergistic use of FEFF, hybrid‑DFT (CRYSTAL), and PAW‑DFT (VASP) not only reproduces the experimental spectra with high accuracy but also elucidates the underlying electronic‑structural mechanisms governing the observed spectral changes. This integrated approach is poised to become a valuable tool for the design and characterization of functional oxide materials such as catalysts, electrodes, and energy‑storage compounds.