Machine learning models can compute the polarizability time series needed for MD-Raman spectra at a small fraction of DFT cost, making finite-temperature Raman prediction for anharmonic materials practical.
Ab initio theory of the non-resonant Raman effect in crystals at finite temperature in comparison to experiment: The examples of GaN and BaZrS3
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
We present an ab initio theory of the non-resonant Raman scattering process in crystals at finite temperature in direct comparison with experiments. The theory incorporates the scattering geometry and polarization dependence of the Raman process and the small but finite wave vectors of the phonons for correctly describing the scattering with longitudinal optical (LO) modes in optically anisotropic solids. We implement the theory for first-order Raman scattering and showcase the approach for wurtzite Gallium Nitride and the complex chalcogenide perovskite BaZrS3 in comparison to experiment. We subsequently discuss several common estimates for second-order Raman scattering in complex materials, and highlight similarities and differences to established theoretical approaches and simulation protocols both from phonon theory and molecular dynamics.
fields
cond-mat.mtrl-sci 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Machine Learning Accelerates Raman Computations from Molecular Dynamics for Materials Science
Machine learning models can compute the polarizability time series needed for MD-Raman spectra at a small fraction of DFT cost, making finite-temperature Raman prediction for anharmonic materials practical.