Machine learning force fields based on local parametrization of dispersion interactions

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
Degree programme
Physical Review B, Volume 104, issue 5
We present a comprehensive methodology to enable the addition of van der Waals (vdW) corrections to machine learning (ML) atomistic force fields. Using a Gaussian approximation potential (GAP) [Bartók et al., Phys. Rev. Lett. 104, 136403 (2010)10.1103/PhysRevLett.104.136403] as a baseline, we accurately machine learn a local model of atomic polarizabilities based on Hirshfeld volume partitioning of the charge density [Tkatchenko and Scheffler, Phys. Rev. Lett. 102, 073005 (2009)10.1103/PhysRevLett.102.073005]. These environment-dependent polarizabilities are then used to parametrize a screened London-dispersion approximation to the vdW interactions. Our ML vdW model only needs to learn the charge density partitioning implicitly by learning the reference Hirshfeld volumes from density functional theory (DFT). In practice, we can predict accurate Hirshfeld volumes from the knowledge of the local atomic environment (atomic positions) alone, making the model highly computationally efficient. For additional efficiency, our ML model of atomic polarizabilities reuses the same many-body atomic descriptors used for the underlying GAP learning of bonded interatomic interactions. We also show how the method enables straightforward computation of gradients of the observables, even when these remain challenging for the reference method (e.g., calculating gradients of the Hirshfeld volumes in DFT). Finally, we demonstrate the approach by studying the phase diagram of C60, where vdW effects are important. The need for a highly accurate vdW-inclusive reactive force field is highlighted by modeling the decomposition of the C60 molecules taking place at high pressures and temperatures.
Funding Information: The authors acknowledge funding from the Academy of Finland, under Projects No. 310574, No. 330488, and No. 329483 (M.A.C.), 321713 (M.A.C., P.H.-L., and H.M.), No. 308647 (X.C.), No. 314298 (X.C. and H.M.), and the QTF Center of Excellence program Grant No. 312298 (T.A.-N.). M.A.C., P. H.-L., and H.M. also acknowledge a seed funding grant from the Aalto University Materials Platform. Computing time from CSC–IT Center for Science for the MaCaNa project and from Aalto University's Science IT project is gratefully acknowledged. Publisher Copyright: © 2021 American Physical Society.
Other note
Muhli , H , Chen , X , Bartók , A P , Hernández-León , P , Csányi , G , Ala-Nissila , T & Caro , M A 2021 , ' Machine learning force fields based on local parametrization of dispersion interactions : Application to the phase diagram of C60 ' , Physical Review B , vol. 104 , no. 5 , 054106 .