Solving the Computational Puzzle: Toward a Pragmatic Pathway for Modeling Low-Energy Vibrational Modes of Pharmaceutical Crystals
Kārlis Bērziņš, Joshua J. Sutton, Sara J. Fraser‐Miller, Thomas Rades, Timothy M. Korter, Keith C. Gordon
Abstract
Five pharmaceutically relevant compounds with increasing molecular complexity, benzoic acid, diphenylacetic acid, l-tyrosine, celecoxib (form III), and carvedilol (form II), were investigated using a combination of low-frequency Raman (LFR) spectroscopy measurements and theoretical solid-state density functional theory (DFT) calculations. l-Tyrosine, specifically, was investigated in more detail to evaluate the effect and rationale for using specific computational parameters. The experimental data were used to validate the DFT simulations by probing the spectral dynamics of the low-energy vibrational modes in a broad temperature range. Principal component analysis (PCA) and individual peak analysis were also used to further elucidate temperature-induced changes. Given the nature of the conducted theoretical calculations, LFR measurements in a cold environment proved especially useful to accurately assess their quality due to complex, temperature-induced spectral dynamics for most of the investigated compounds (especially, l-tyrosine).