Infrared spectra of neutral polycyclic aromatic hydrocarbons based on machine learning potential energy surface and dipole mapping - Archive ouverte HAL Access content directly
Journal Articles Theoretical Chemistry Accounts: Theory, Computation, and Modeling Year : 2021

Infrared spectra of neutral polycyclic aromatic hydrocarbons based on machine learning potential energy surface and dipole mapping

Abstract

The interest in polycyclic aromatic hydrocarbons (PAH) spans numerous fields, and infrared (IR) spectroscopy is usually the method of choice to disentangle their molecular structure. In order to compute vibrational frequencies, numerous theoretical studies employ either quantum calculation methods, or empirical potentials, but it remains difficult to combine the accuracy of the first approach with the computational cost of the second. In this work, we propose a new machine learning approach for the prediction of the IR properties of a class of molecules. Two artificial neural network architectures are trained to reproduce the potential energy surface and the dipole mapping of 11 PAH molecules. Altogether, these two quantities are employed to retrieve the IR properties (frequencies and intensities), using generalized second-order vibrational perturbation (GVPT2) theory, for predicting properties of 34 PAH molecules. This work lays the foundation of our approach demonstrating its overall good results and transferability capability toward unlearned molecules. Finally, we discuss the current limitations and perspectives for further improvement of the method.
Fichier principal
Vignette du fichier
2010.13686 (1.53 Mo) Télécharger le fichier
Origin : Publisher files allowed on an open archive

Dates and versions

hal-03248387 , version 1 (07-12-2023)

Identifiers

Cite

Gaétan Laurens, Malalatiana Rabary, Julien Lam, Daniel Peláez, Abdul-Rahman Allouche. Infrared spectra of neutral polycyclic aromatic hydrocarbons based on machine learning potential energy surface and dipole mapping. Theoretical Chemistry Accounts: Theory, Computation, and Modeling, 2021, 140 (6), pp.66. ⟨10.1007/s00214-021-02773-6⟩. ⟨hal-03248387⟩
71 View
12 Download

Altmetric

Share

Gmail Facebook X LinkedIn More