Artificial intelligence-enhanced quantum chemical method with broad applicability

Peikun Zheng,Pavlo O Dral,Olexandr Isayev,Roman Zubatyuk,Wei Wu

doi:10.1038/s41467-021-27340-2

Peikun Zheng, Pavlo O Dral + Show 3 more

Open Access

https://doi.org/10.1038/s41467-021-27340-2

Copy DOI

Abstract

High-level quantum mechanical (QM) calculations are indispensable for accurate explanation of natural phenomena on the atomistic level. Their staggering computational cost, however, poses great limitations, which luckily can be lifted to a great extent by exploiting advances in artificial intelligence (AI). Here we introduce the general-purpose, highly transferable artificial intelligence–quantum mechanical method 1 (AIQM1). It approaches the accuracy of the gold-standard coupled cluster QM method with high computational speed of the approximate low-level semiempirical QM methods for the neutral, closed-shell species in the ground state. AIQM1 can provide accurate ground-state energies for diverse organic compounds as well as geometries for even challenging systems such as large conjugated compounds (fullerene C60) close to experiment. This opens an opportunity to investigate chemical compounds with previously unattainable speed and accuracy as we demonstrate by determining geometries of polyyne molecules—the task difficult for both experiment and theory. Noteworthy, our method’s accuracy is also good for ions and excited-state properties, although the neural network part of AIQM1 was never fitted to these properties.

Highlights

High-level quantum mechanical (QM) calculations are indispensable for accurate explanation of natural phenomena on the atomistic level
The artificial intelligence–quantum mechanical method 1 (AIQM1) method consists of three main parts (Fig. 2): (1) semiempirical QM (SQM) Hamiltonian, (2) neural network (NN) correction to the potential, and (3) dispersion corrections
We have chosen the orthogonalization- and dispersion-corrected method 2 (ODM2) Hamiltonian[26], which provides the most consistent and accurate predictions across different properties among other SQM methods, those based on neglect of diatomic differential overlap (NDDO) approximation

Summary

Introduction

High-level quantum mechanical (QM) calculations are indispensable for accurate explanation of natural phenomena on the atomistic level. We introduce the general-purpose, highly transferable artificial intelligence–quantum mechanical method 1 (AIQM1) It approaches the accuracy of the gold-standard coupled cluster QM method with high computational speed of the approximate low-level semiempirical QM methods for the neutral, closed-shell species in the ground state. AIQM1 can provide accurate ground-state energies for diverse organic compounds as well as geometries for even challenging systems such as large conjugated compounds (fullerene C60) close to experiment This opens an opportunity to investigate chemical compounds with previously unattainable speed and accuracy as we demonstrate by determining geometries of polyyne molecules—the task difficult for both experiment and theory. ANI potentials are much less transferable than general-purpose QM methods, because they are limited to closed-shell, neutral organic compounds, and the use of the local approximation imposes further limitations on their transferability, e.g., to large, highly conjugated systems (Fig. 1b, c). This approach has already given rise to an increasing number of hybrid AI/QM methods[7,8,20,21,22], most of them are either proofof-principle or based on relatively slow DFT or trained on data of a b

Methods

Results

Conclusion