Raman spectroscopy is a widely used experimental technique to study the vibrational properties of solids. Atomic scale simulations can be used to predict such spectra, but reliable studies at finite temperatures are challenging, mainly due to the requirement of accurate and computationally efficient models for the dielectric susceptibility. Here, we have used molecular dynamics simulations together with a density functional theorybased model for the dielectric susceptibility to determine the Raman spectrum of barium zirconate, BaZrO3 (BZO), a well-studied oxide perovskite. At ambient conditions, where the system is cubic, we find excellent agreement with experimentally measured Raman spectra. Our study establishes that the relatively sharp spectra seen experimentally are due to second-order scattering. At higher pressures, where BZO is tetragonal, all firstorder Raman active modes are identified. Additionally, slightly below the phase transition, in the cubic phase, a broad central Raman peak appears. The origin of this type of peak is controversial and extensively debated in connection with the dynamics of the halide perovskites. Here, we show that it is also present in a hard oxide perovskite, and it originates from the highly overdamped R-tilt mode in the cubic structure.

Bismuth vanadate (BiVO _4 ) is one of the most promising photoanode materials for water splitting, yet its intrinsic oxygen evolution reaction (OER) performance is limited. Here, we employ hybrid density functional theory calculations to investigate the synergistic effects of nitrogen doping and oxygen vacancy formation on BiVO _4 for the OER. We analyze two OER mechanisms, the traditional single-site adsorption evolution mechanism (AEM) and the dual-site bridging mechanism (DSBM), to understand the enhanced catalytic activity observed experimentally. Our results show that nitrogen doping on the surface, combined with the creation of a vanadium active site through oxygen vacancy, significantly reduces the OER overpotential from 1.44 V in pristine BiVO _4 to 0.93 V (AEM) and 1.16 V (DSBM). Besides, the nitrogen dopants on the surface alter the local acid-base chemistry: proton adsorption on nitrogen becomes 0.52 eV more favorable than on oxygen, and water dissociation is 0.31 eV easier at the V site than at the Bi site. By offering low-energy proton-binding sites, nitrogen stabilizes key intermediates and favors formation of a stable O–O dimer, producing a pronounced reduction in overpotential. These findings highlight that the effective defect engineering strategies can significantly enhance the overall performance of BiVO _4 -based photoanodes in operational photoelectrochemical systems.

Transition metal oxides, such as BiVO4, have attracted significant attention for their potential in photoelectrochemical water-splitting. BiVO4, a model material in this area, is prone to charge localization in the form of small polarons. Recently, self-trapped excitons (STEs) in BiVO4 have been experimentally observed, but their precise nature remains elusive. In this study, we employ time-dependent density functional theory (TD-DFT) with a nonempirical PBE0(α) hybrid functional to investigate the localization, stability, and optical properties of STEs in BiVO4. Our results reveal two distinct localized exciton configurations with comparable energies. We show that the emission from a single STE configuration leads to multiple peaks in the emission spectrum, originating from different types of internal transitions. The positions of peaks in the calculated optical spectra are in good agreement with experimental observations.

Color centers play key roles in applications, including, e.g., solid state lighting and quantum information technology, for which the coupling between their optical and vibrational properties is crucial. Established methodologies for predicting the optical lineshapes of such emitters rely on the generating function (GF) approach and impose tight constraints on the shape of and relationship between the ground and excited state landscapes, which limits their application range. Here, we describe an approach based on direct sampling of the underlying auto-correlation functions through molecular dynamics simulations (MD-ACF) that overcomes these restrictions. The energy landscapes are represented by a machine-learned potential, which provides an accurate yet efficient description of both the ground and excited state landscapes through a single model, guaranteeing size-consistent predictions. We apply this methodology to the (VSiVC)kk(0) divacancy defect in 4H-SiC, a prototypical color center, which has been studied both experimentally and theoretically. We demonstrate that at low temperatures the present MD-ACF approach yields predictions in agreement with earlier GF calculations. Unlike the latter it is, however, also applicable at high temperatures as it is not subject to the same limitations, especially with respect to handling of anharmonicity, and can be applied to study non-crystalline materials. While we discuss remaining challenges and possible extensions, the methodology presented here already holds the potential to substantially widen the range of computational predictions of the optical properties of color centers and related defects, especially for cases with pronounced anharmonicity and/or large differences between the initial and final states.

Infrared and Raman spectroscopy are widely used for the characterization of gases, liquids, and solids, as the spectra contain a wealth of information concerning in particular the dynamics of these systems. Atomic scale simulations can be used to predict such spectra but are often severely limited due to high computational cost or the need for strong approximations that limit application range and reliability. Here, we introduce a machine learning (ML) accelerated approach that addresses these shortcomings and provides a significant performance boost in terms of data and computational efficiency compared to earlier ML schemes. To this end, we generalize the neuroevolution potential approach to enable the prediction of rank one and two tensors to obtain the tensorial neuroevolution potential (TNEP) scheme. We apply the resulting framework to construct models for the dipole moment, polarizability, and susceptibility of molecules, liquids, and solids, and show that our approach compares favorably with several ML models from the literature with respect to accuracy and computational efficiency. Finally, we demonstrate the application of the TNEP approach to the prediction of infrared and Raman spectra of liquid water, a molecule (PTAF-), and a prototypical perovskite with strong anharmonicity (BaZrO3). The TNEP approach is implemented in the free and open source software package GPUMD, which makes this methodology readily available to the scientific community.

Raman spectroscopy is a widely used experimental technique to study the vibrational properties of solids. Atomic scale simulations can be used to predict such spectra, but trustworthy studies at finite temperatures are challenging, mainly due to the requirement of accurate and computationally efficient models for the dielectric susceptibility. Here, we have made use of molecular dynamics (MD) simulations together with a density functional theory (DFT) based model for the dielectric susceptibility to determine the Raman spectrum of barium zirconate, BaZrO_3 (BZO), a well-studied oxide perovskite. At ambient conditions, where the system is cubic, we find excellent agreement with experimentally measured Raman spectra. Our study establishes that the relatively sharp spectra seen experimentally are due to second-order scattering. At higher pressures, where BZO is tetragonal, all first-order Raman active modes are identified. Additionally, slightly below the phase transition, in the cubic phase, a broad "central Raman peak" appears. The origin of this type of peak is controversial and extensively debated in connection to the dynamics of the halide perovskites. Here, we show that it is also present in a "hard" oxide perovskite, and it originates from the highly overdamped R-tilt mode in the cubic structure.

Molecular dynamics (MD) simulations are a key tool in computational chemistry, physics, and materials science, aiding the understanding of microscopic processes but also guiding the development of novel materials.A MD simulation requires a model for the interatomic interactions.To this end, one traditionally often uses empirical interatomic potentials or force fields, which are fast but inaccurate, or ab-initio methods based on electronic structure theory such as density functional theory, which are accurate but computationally very expensive (Müser et al., 2023).Machine-learned interatomic potentials (MLIPs) have in recent years emerged as an alternative to these approaches, combining the speed of heuristic force fields with the accuracy of ab-initio techniques (Unke et al., 2021).Neuroevolution potentials (NEPs), implemented in the GPUMD package, in particular, are a highly accurate and efficient class of MLIPs (Fan et al., 2021, 2022;Fan, 2022).NEP models have already been used to study a variety of properties in a range of materials, with recent examples including radiation damage in tungsten (Liu et al., 2023), phase transitions (Fransson, Wiktor, et al., 2023) and dynamics of halide perovskites (Fransson, Rosander, et al., 2023) as well as thermal transport in two-dimensional materials (Sha et al., 2023).Here, we present calorine, a Python package that simplifies the construction, analysis and use of NEP models via GPUMD.

On the BiVO 4 photoanode surface, oxygen vacancies modified by cobalt incorporation have a remarkable impact on the water oxidation activity and charge injection efficiency.

Bismuth vanadate, BiVO4, is one of the most promising photoanode materials for the challenging oxygen evolution half-reaction in solar-driven water splitting. The material tends to be rich in oxygen vacancies, which strongly affects its photoelectrochemical properties. Experimental evidence suggests that oxygen deficiency is beneficial for the oxygen evolution reaction in the material, but the mechanism behind this enhancement is still controversial. The defects could be involved directly in the reaction if present at the surface, and the occupancy of the defect states could also play an important role. The latter is seldom considered in mechanistic studies, however. Using density functional theory, we show that the surface oxygen vacancy in bismuth vanadate is stablest when fully ionized. We investigate how this affects the oxygen evolution mechanism by mapping out the stablest reaction intermediates and compare the resulting pathway with those on the unionized oxygen-deficient surface as well as the defect-free material. The overpotentials required to drive the reaction in each case are computed to quantify whether or not vacancy formation, and subsequent ionization, improves the thermodynamics of oxygen evolution.

The influence of surface oxygen vacancies on the oxygen evolution reaction on bismuth vanadate is studied using hybrid density functional theory. Our findings reveal that the neutral charge state is thermodynamically unfavorable, leading to spontaneous ionization of VO0 into VO2+. By investigating the oxygen evolution reaction mechanism on both stoichiometric and oxygen-deficient surfaces, we find that surface oxygen vacancies reduce the reaction’s thermodynamic overpotential but only when the defects are ionized. Moreover, the reaction pathway involves the formation of surface-bound peroxide and superoxide ions as intermediates. Our work provides insight into the nature of surface oxygen vacancies and shines new light on how they may enhance the photoelectrochemical properties of semiconductors.

We study the native defects in bismuth vanadate using hybrid density functional theory. We pay special attention to where excess charges localize by considering different polaronic distortions and find that charge localization has a profound effect on the local chemical environment around certain defects. In particular, oxygen dimerization may occur in the presence of acceptor defects. On the basis of Fermi level pinning due to compensation between donors and acceptors we additionally find that intrinsic p-type conductivity is difficult to achieve in BiVO4, in good agreement with experimental observations. Our results give new insights into the defect chemistry of bismuth vanadate and act as a guide for future studies on defects in complex metal oxides.

We study oxygen vacancies in the tetragonal scheelite phase of bismuth vanadate and identify stable oxygen-deficient structures. Upon subjecting these to variable-cell optimization, we find that oxygen vacancies give rise to significant structural distortions, the degree of which exhibits a vacancy concentration dependence. Furthermore, we show that these distortions give rise to splitting of powder X-ray diffraction peaks, yielding patterns similar to that of the monoclinic scheelite phase, and that these effects are also present at finite temperatures. Our results highlight the need for characterization methods beyond X-ray diffraction for identifying the phase of synthesized bismuth vanadate samples and the importance of oxygen partial pressure control during synthesis.

We study the nature of excess electrons in CsPbBr3 and identify several single and double polaronic states. We emphasize the importance of proper inclusion of the self-interaction corrections for the stability of small electron polarons in this material. We demonstrate that spin-orbit coupling (SOC) has a significant impact on the energetics of the polaronic states. In particular, we find that SOC disfavors electron localization and leads to different polaronic geometries. Additionally, by carrying out thermodynamic integration, we show that small electron polarons are thermally stabilized in CsPbBr3. The small energy differences between the localized and delocalized electronic states could possibly reconcile the apparently conflicting properties of high charge-carrier mobilities and low recombinations rates.