Parametrization of Nonbonded Force Field Terms for Metal–Organic Frameworks Using Machine Learning Approach

Abstract

The enormous structural and chemical diversity of metal−organic frameworks (MOFs) forces researchers to actively use simulation techniques as often as experiments. MOFs are widely known for their outstanding adsorption properties, so a precise description of the host−guest interactions is essential for high-throughput screening aimed at ranking the most promising candidates. However, highly accurate ab initio calculations cannot be routinely applied to model thousands of structures due to the demanding computational costs. Furthermore, methods based on force field (FF) parametrization suffer from low transferability. To resolve this accuracy−efficiency dilemma, we applied a machine learning (ML) approach: extreme gradient boosting. The trained models reproduced the atom-in-material quantities, including partial charges, polarizabilities, dispersion coefficients, quantum Drude oscillator, and electron cloud parameters, with accuracy similar to the reference data set. The aforementioned FF precursors make it possible to thoroughly describe noncovalent interactions typical for MOF−adsorbate systems: electrostatic, dispersion, polarization, and short-range repulsion. The presented approach can also readily facilitate hybrid atomistic simulation/ML workflows.

Open-Source Machine Learning in Computational Chemistry

“… 144 The force field precursors for metal–organic frameworks (FFP4MOF) tool was developed for use in materials research and is able to predict nonbonded parameters for metal-containing systems. 145 Molecule-specific (i.e., ammonium perchlorate, pentafluoroethane, difluoromethane) examples of optimizing Lennard-Jones parameters through multiobjective surrogate-assisted Gaussian process regression and support vector machine workflows can be found in ref ( 146 ) and its two cited GitHub repositories. Similarly, PREMSO uses a presampling-enhanced, surrogate-assisted global evolutionary optimization strategy that allows the use of features at different scales (e.g., single-molecule and bulk-phase observables).…”

PACMAN: A Robust Partial Atomic Charge Predicter for Nanoporous Materials Based on Crystal Graph Convolution Networks

“…Korolev et al have used the gradient boosting decision tree (GBDT) method to train a model to predict DDEC charges based on intrinsic elemental properties and structural descriptors of the site. Additionally, the development of the connectivity-based atom contribution (CBAC) approach can be considered a significant advancement in the structural descriptors used in their study. , The predicted partial charges show an average absolute deviation of 0.01e based on the CoRE MOF 2014 DDEC database with 2932 MOFs. The performances of the developed ML models have been benchmarked in the literature against each other.…”

An automated protocol to construct flexibility parameters for classical forcefields: applications to metal–organic frameworks

“…The present article focuses exclusively on parameterizing the intracluster bonded interactions (i.e., parameterizing the exibility terms) up to second-order derivatives in the energy. The intracluster nonbonded interactions and intercluster nonbonded interactions have been partly studied in several previous publications (co)authored by one of us [51][52][53][54][55][56][57][58][59][60][61][62][63][64] and will be further studied in some of our upcoming publications.…”

“…, parameterizing the flexibility terms) up to second-order derivatives in the energy. The intracluster nonbonded interactions and intercluster nonbonded interactions have been partly studied in several previous publications (co)authored by one of us 51–64 and will be further studied in some of our upcoming publications.…”