Accuracy and limitations of the bond polarizability model in modeling of Raman scattering from molecular dynamics simulations.

This article systematically evaluates the bond polarizability model (BPM) for calculating electronic polarizability and Raman spectra from molecular dynamics (MD) simulations across a range of molecular and solid-state systems. BPM, which assumes that total polarizability is the sum of independent bond contributions depending only on bond lengths, accurately reproduces overall Raman spectral features and peak positions for simple and highly symmetric molecules such as CO₂ and CH₄, but shows quantitative inaccuracies for molecules with significant bond-angle variations (e.g., H₂O) and complex solids like BaTiO₃ (BTO) and CsPbBr₃ (CPB) perovskites. The study finds that BPM's neglect of bond-angle effects and bond interactions limits its accuracy, especially for asymmetric molecules and solids with strong distortions, although second-order expansions improve results somewhat. Consequently, while BPM serves as a computationally efficient baseline for qualitative Raman spectral modeling, the authors conclude that more sophisticated atomistic models incorporating bond interactions and angular dependencies are necessary for quantitatively reliable predictions in complex systems.