Vibrational spectroscopy techniques serve as crucial tools for understanding the structure and dynamics of the condensed phase. These spectra can be accurately computed from first principles; however, for larger systems, these calculations can be prohibitively expensive. Here, we present our efforts to accelerate the prediction of condensed phase vibrational spectra with data-driven methods.<br/><br/>We present our methods and results for computing condensed phase vibrational spectra from coarse-grained representations in a mixed quantum-classical framework. We will demonstrate the efficacy of different featurizations of the coarse-grained representations, sensitivity to training set composition and size, model architecture, and the degree of transferability of the models to different parts of the phase diagram. As the framework only incorporates machine learning into one part of an overall, mixed-quantum-classical framework, we can employ this method in both single-dimensional and multi-dimensional spectra. We highlight the features that perform best for different types of spectra.

Data-driven methods for accelerating condensed phase vibrational spectroscopy modeling

Where thousands of chemistry professionals meet to share ideas and advance scientific and technical knowledge. Spring 2023 abstracts, posters, and presentations will launch at the end of March.

ACS Spring 2023

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.