Improved ab initio molecular dynamics-based vibrational spectroscopy for indirect hard modeling for bulk-phase...

Area of Science:

• Computational Chemistry and Spectroscopy
• Molecular Dynamics Simulations
• Quantitative Analysis

Background:

• Ab initio molecular dynamics (AIMD) derived vibrational spectra offer a path to calibration-free quantitative spectroscopy when coupled with indirect hard modeling (IHM).
• However, AIMD spectra accuracy for bulk phases can be limited by incomplete sampling, electronic structure inaccuracies, and frequency shifts compared to experimental data.
• Strongly hydrogen-bonded systems, like aqueous acetic acid, present unique challenges due to transient associations and proton sharing.

Purpose of the Study:

• To develop and validate an improved AIMD-IHM framework for accurate, calibration-free quantitative spectroscopy.
• To address limitations in AIMD spectral reliability, including sampling, electronic structure errors, and frequency shifts.
• To demonstrate the framework's efficacy on a challenging, strongly hydrogen-bonded system: aqueous acetic acid.

Main Methods:

• Benchmarking BLYP and B3LYP/ADMM functionals for accuracy in AIMD simulations.
• Performing cluster-resolved sampling analysis using hydrogen-bond kinetics.
• Implementing a gas-phase-anchored vibrational frequency scaling strategy for bulk-phase spectra.
• Utilizing reactive-flux analysis for hydrogen-bond cluster dynamics and sampling assessment.

Main Results:

• The computationally efficient BLYP functional outperformed B3LYP/ADMM in reproducing experimental vibrational frequencies (RMSE 91 cm⁻¹ vs. 155 cm⁻¹).
• A region-specific scaling procedure significantly reduced the RMSE of BLYP-based spectra.
• Hydrogen-bond cluster analysis confirmed sufficient sampling for relevant molecular motifs.
• The scaled AIMD-derived spectra, integrated into IHM, determined experimental mixture compositions with an RMSE of 0.021, outperforming unscaled spectra (RMSE 0.034).

Conclusions:

• The improved AIMD-IHM framework enhances predictive accuracy for quantitative spectroscopy.
• The methodology effectively addresses sampling and electronic structure limitations in AIMD.
• This approach extends the applicability of calibration-free quantitative spectroscopy to strongly interacting liquid systems.