×

Recommended by

Indexed by

A dual-cutoff machine-learned potential for condensed organic systems obtained via uncertainty-guided active learning

Leonid Kahle1*, Benoit Minisini1, Tai Bui2, Jeremy First3, Corneliu Buda2, Thomas Goldman2, Erich Wimmer1

1 Materials Design SARL, 42 avenue Verdier, 92120 Montrouge, France

2 bp Exploration Operating Co. Ltd, Chertsey Road, Sunbury-on-Thames TW16 7LN, UK

3 bp, Center for High Performance Computing, 225 Westlake Park Blvd, Houston, TX 77079, USA

* Corresponding authors emails: leonidkahle@gmail.com
DOI10.24435/materialscloud:ed-gp [version v1]

Publication date: Aug 12, 2024

How to cite this record

Leonid Kahle, Benoit Minisini, Tai Bui, Jeremy First, Corneliu Buda, Thomas Goldman, Erich Wimmer, A dual-cutoff machine-learned potential for condensed organic systems obtained via uncertainty-guided active learning, Materials Cloud Archive 2024.121 (2024), https://doi.org/10.24435/materialscloud:ed-gp

Description

Machine-learned potentials (MLPs) trained on ab initio data combine the computational efficiency of classical interatomic potentials with the accuracy and generality of the first-principles method used in the creation of the respective training set. In this work, we implement and train a MLP to obtain an accurate description of the potential energy surface and property predictions for organic compounds, as both single molecules and in the condensed phase. We devise a dual descriptor, based on the atomic cluster expansion (ACE), that couples an information-rich short-range description with a coarser long-range description that captures weak intermolecular interactions. We employ uncertainty-guided active learning for the training set generation, creating a dataset that is comparatively small for the breadth of application and consists of alcohols, alkanes, and an adipate. Utilizing that MLP, we calculate densities of those systems of varying chain lengths as a function of temperature, obtaining a discrepancy of less than 4% compared with experiment. Vibrational frequencies calculated with the MLP have a root mean square error of less than 1 THz compared to DFT. The heat capacities of condensed systems are within 11% of experimental findings, which is strong evidence that the dual descriptor provides an accurate framework for the prediction of both short-range intramolecular and long-range intermolecular interactions.

Materials Cloud sections using this data

No Explore or Discover sections associated with this archive record.

Files

File name Size Description
data.tar.gz
MD5md5:b66aa295818467bc09875599dbe6b49b
5.2 GiB Data as tar.gz, Check README.txt for more information
README.txt
MD5md5:8d7a2b972808523498eb63f358bcbdfc
1.0 KiB README.txt which describes the data in data.tar.gz

License

Files and data are licensed under the terms of the following license: Creative Commons Attribution 4.0 International.
Metadata, except for email addresses, are licensed under the Creative Commons Attribution Share-Alike 4.0 International license.

Keywords

molecular dynamics machine learning diffusion radial distribution functions heat capacity VASP training data

Version history:

2024.121 (version v1) [This version] Aug 12, 2024 DOI10.24435/materialscloud:ed-gp