MLTB: Enhancing Transferability and Extensibility of Density Functional Tight-Binding Theory with Many-body Interaction Corrections

Overview

Journal J Chem Theory Comput

Publisher American Chemical Society

Date 2025 Jan 29

PMID 39876631

Authors

Daniel J Burrill

Chang Liu

Michael G Taylor

Marc J Cawkwell

Danny Perez

Enrique R Batista

Nicholas Lubbers

Ping Yang

Affiliations

Soon will be listed here.

Abstract

We present a hybrid semiempirical density functional tight-binding (DFTB) model with a machine learning neural network potential as a correction to the repulsive term. This hybrid model, termed machine learning tight-binding (MLTB), employs the standard self-consistent charge (SCC) DFTB formalism as a baseline, enhanced by the HIP-NN potential as an effective many-body correction for short-range pairwise repulsive interactions. The MLTB model demonstrates significantly improved transferability and extensibility compared to the SCC-DFTB and HIP-NN models. This work provides a practical computational framework for developing reliable SCC-DFTB models with additional many-body corrections that more closely approach the DFT level of accuracy. We illustrate this method with the development of an accurate model for the thorium-oxygen system, applied to the study of its nanocluster structures (ThO).

References

van Lenthe E, Baerends E . Optimized Slater-type basis sets for the elements 1-118. J Comput Chem. 2003; 24(9):1142-56. DOI: 10.1002/jcc.10255. View

Unke O, Chmiela S, Sauceda H, Gastegger M, Poltavsky I, Schutt K . Machine Learning Force Fields. Chem Rev. 2021; 121(16):10142-10186. PMC: 8391964. DOI: 10.1021/acs.chemrev.0c01111. View

Aebersold L, Wilson A . Considering Density Functional Approaches for Actinide Species: The An66 Molecule Set. J Phys Chem A. 2021; 125(32):7029-7037. DOI: 10.1021/acs.jpca.1c06155. View

Panosetti C, Engelmann A, Nemec L, Reuter K, Margraf J . Learning to Use the Force: Fitting Repulsive Potentials in Density-Functional Tight-Binding with Gaussian Process Regression. J Chem Theory Comput. 2020; 16(4):2181-2191. DOI: 10.1021/acs.jctc.9b00975. View

Krishnapriyan A, Yang P, Niklasson A, Cawkwell M . Numerical Optimization of Density Functional Tight Binding Models: Application to Molecules Containing Carbon, Hydrogen, Nitrogen, and Oxygen. J Chem Theory Comput. 2017; 13(12):6191-6200. DOI: 10.1021/acs.jctc.7b00762. View

Kulichenko M, Smith J, Nebgen B, Li Y, Fedik N, Boldyrev A . The Rise of Neural Networks for Materials and Chemical Dynamics. J Phys Chem Lett. 2021; 12(26):6227-6243. DOI: 10.1021/acs.jpclett.1c01357. View

Elstner M, Seifert G . Density functional tight binding. Philos Trans A Math Phys Eng Sci. 2014; 372(2011):20120483. DOI: 10.1098/rsta.2012.0483. View

Spiegelman F, Tarrat N, Cuny J, Dontot L, Posenitskiy E, Marti C . Density-functional tight-binding: basic concepts and applications to molecules and clusters. Adv Phys X. 2020; 5(1):1710252. PMC: 7116320. DOI: 10.1080/23746149.2019.1710252. View

Schutt K, Arbabzadah F, Chmiela S, Muller K, Tkatchenko A . Quantum-chemical insights from deep tensor neural networks. Nat Commun. 2017; 8:13890. PMC: 5228054. DOI: 10.1038/ncomms13890. View

10.

Stocks R, Barnard A . Enhancing classical gold nanoparticle simulations with electronic corrections and machine learning. J Phys Condens Matter. 2021; 33(32). DOI: 10.1088/1361-648X/ac0751. View

11.

Batista E, Martin R, Hay P, Peralta J, Scuseria G . Density functional investigations of the properties and thermochemistry of UF6 and UF5 using valence-electron and all-electron approaches. J Chem Phys. 2004; 121(5):2144-50. DOI: 10.1063/1.1768518. View

12.

Panosetti C, Annies S, Grosu C, Seidlmayer S, Scheurer C . DFTB Modeling of Lithium-Intercalated Graphite with Machine-Learned Repulsive Potential. J Phys Chem A. 2021; 125(2):691-699. DOI: 10.1021/acs.jpca.0c09388. View

13.

Perdew , Burke , Ernzerhof . Generalized Gradient Approximation Made Simple. Phys Rev Lett. 1996; 77(18):3865-3868. DOI: 10.1103/PhysRevLett.77.3865. View

14.

McSloy A, Fan G, Sun W, Holzer C, Friede M, Ehlert S . TBMaLT, a flexible toolkit for combining tight-binding and machine learning. J Chem Phys. 2023; 158(3):034801. DOI: 10.1063/5.0132892. View

15.

Behler J, Lorenz S, Reuter K . Representing molecule-surface interactions with symmetry-adapted neural networks. J Chem Phys. 2007; 127(1):014705. DOI: 10.1063/1.2746232. View

16.

Kranz J, Kubillus M, Ramakrishnan R, Anatole von Lilienfeld O, Elstner M . Generalized Density-Functional Tight-Binding Repulsive Potentials from Unsupervised Machine Learning. J Chem Theory Comput. 2018; 14(5):2341-2352. DOI: 10.1021/acs.jctc.7b00933. View

17.

Kelley M, Deblonde G, Su J, Booth C, Abergel R, Batista E . Bond Covalency and Oxidation State of Actinide Ions Complexed with Therapeutic Chelating Agent 3,4,3-LI(1,2-HOPO). Inorg Chem. 2018; 57(9):5352-5363. DOI: 10.1021/acs.inorgchem.8b00345. View

18.

Porezag , Frauenheim , KOHLER , Seifert , Kaschner . Construction of tight-binding-like potentials on the basis of density-functional theory: Application to carbon. Phys Rev B Condens Matter. 1995; 51(19):12947-12957. DOI: 10.1103/physrevb.51.12947. View

19.

Lubbers N, Smith J, Barros K . Hierarchical modeling of molecular energies using a deep neural network. J Chem Phys. 2018; 148(24):241715. DOI: 10.1063/1.5011181. View

20.

Ammothum Kandy A, Wadbro E, Aradi B, Broqvist P, Kullgren J . Curvature Constrained Splines for DFTB Repulsive Potential Parametrization. J Chem Theory Comput. 2021; 17(3):1771-1781. PMC: 8023658. DOI: 10.1021/acs.jctc.0c01156. View