Free-energy perturbation in the exchange-correlation space accelerated by machine learning: Application to silica polymorphs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Forslund, Axel, Jung, Jong Hyun, Ikeda, Yuji, Grabowski, Blazej
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916716540329984
author Forslund, Axel
Jung, Jong Hyun
Ikeda, Yuji
Grabowski, Blazej
author_facet Forslund, Axel
Jung, Jong Hyun
Ikeda, Yuji
Grabowski, Blazej
contents We propose a free-energy-perturbation approach accelerated by machine-learning potentials to efficiently compute transition temperatures and entropies for all rungs of Jacob's ladder. We apply the approach to the dynamically stabilized phases of SiO$_2$, which are characterized by challengingly small transition entropies. All investigated functionals from rungs 1-4 fail to predict an accurate transition temperature by 25-200%. Only by ascending to the fifth rung, within the random phase approximation, an accurate prediction is possible, giving a relative error of 5%. We provide a clear-cut procedure and relevant data to the community for, e.g., developing and evaluating new functionals.
format Preprint
id arxiv_https___arxiv_org_abs_2505_00789
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Free-energy perturbation in the exchange-correlation space accelerated by machine learning: Application to silica polymorphs
Forslund, Axel
Jung, Jong Hyun
Ikeda, Yuji
Grabowski, Blazej
Materials Science
We propose a free-energy-perturbation approach accelerated by machine-learning potentials to efficiently compute transition temperatures and entropies for all rungs of Jacob's ladder. We apply the approach to the dynamically stabilized phases of SiO$_2$, which are characterized by challengingly small transition entropies. All investigated functionals from rungs 1-4 fail to predict an accurate transition temperature by 25-200%. Only by ascending to the fifth rung, within the random phase approximation, an accurate prediction is possible, giving a relative error of 5%. We provide a clear-cut procedure and relevant data to the community for, e.g., developing and evaluating new functionals.
title Free-energy perturbation in the exchange-correlation space accelerated by machine learning: Application to silica polymorphs
topic Materials Science
url https://arxiv.org/abs/2505.00789