Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: Quantum chemical accuracy from density functional approximations via machine learning
Abstract Kohn-Sham density functional theory (DFT) is a standard tool in most branches of chemistry, but accuracies for many molecules are limited to 2-3 kcal ⋅ mol −1 with presently-available functionals. Ab initio methods, such as coupled-cluster, routinely produce much higher accuracy, but computational costs limit their application to small molecules. In this paper, we leverage machine learning to calculate coupled-cluster energies from DFT densities, reaching quantum chemical accuracy (errors below 1 kcal ⋅ mol −1 ) on test data. Moreover, density-based Δ -learning (learning only the correction to a standard DFT calculation, termed Δ -DFT ) significantly reduces the amount of training data required, particularly when molecular symmetries are included. The robustness of Δ -DFT  is highlighted by correcting “on the fly” DFT-based molecular dynamics (MD) simulations of resorcinol (C 6 H 4 (OH) 2 ) to obtain MD trajectories with coupled-cluster accuracy. We conclude, therefore, that Δ -DFT  facilitates running gas-phase MD simulations with quantum chemical accuracy, even for strained geometries and conformer changes where standard DFT fails.  more » « less
Award ID(s):
1856165
PAR ID:
10222055
Author(s) / Creator(s):
; ; ; ;
Date Published:
Journal Name:
Nature Communications
Volume:
11
Issue:
1
ISSN:
2041-1723
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; (, Chemical Science)
    With dual goals of efficient and accurate modeling of solvation thermodynamics in molten salt liquids, we employ ab initio molecular dynamics (AIMD) simulations, deep neural network interatomic potentials (NNIP), and quasichemical theory (QCT) to calculate the excess chemical potentials for the solute ions Na + and Cl − in the molten NaCl liquid. NNIP-based molecular dynamics simulations accelerate the calculations by 3 orders of magnitude and reduce the uncertainty to 1 kcal mol −1 . Using the Density Functional Theory (DFT) level of theory, the predicted excess chemical potential for the solute ion pair is −178.5 ± 1.1 kcal mol −1 . A quantum correction of 13.7 ± 1.9 kcal mol −1 is estimated via higher-level quantum chemistry calculations, leading to a final predicted ion pair excess chemical potential of −164.8 ± 2.2 kcal mol −1 . The result is in good agreement with a value of −163.5 kcal mol −1 obtained from thermo-chemical tables. This study validates the application of QCT and NNIP simulations to the molten salt liquids, allowing for significant insights into the solvation thermodynamics crucial for numerous molten salt applications. 
    more » « less
  2. ; ; ; (, Physical Chemistry Chemical Physics)
    Rational design of molecular chelating agents requires a detailed understanding of physicochemical ligand–metal interactions in solvent phase. Computational quantum chemistry methods should be able to provide this, but computational reports have shown poor accuracy when determining absolute binding constants for many chelating molecules. To understand why, we compare and benchmark static- and dynamics-based computational procedures for a range of monovalent and divalent cations binding to a conventional cryptand molecule: 2.2.2-cryptand ([2.2.2]). The benchmarking comparison shows that dynamics simulations using standard OPLS-AA classical potentials can reasonably predict binding constants for monovalent cations, but these procedures fail for divalent cations. We also consider computationally efficient static procedure using Kohn–Sham density functional theory (DFT) and cluster-continuum modeling that accounts for local microsolvation and pH effects. This approach accurately predicts binding energies for monovalent and divalent cations with an average error of 3.2 kcal mol −1 compared to experiment. This static procedure thus should be useful for future molecular screening efforts, and high absolute errors in the literature may be due to inadequate modeling of local solvent and pH effects. 
    more » « less
  3. ; ; ; (, Chemical Physics Reviews)
    Density functional theory (DFT) has been applied to modeling molecular interactions in water for over three decades. The ubiquity of water in chemical and biological processes demands a unified understanding of its physics, from the single molecule to the thermodynamic limit and everything in between. Recent advances in the development of data-driven and machine-learning potentials have accelerated simulation of water and aqueous systems with DFT accuracy. However, anomalous properties of water in the condensed phase, where a rigorous treatment of both local and non-local many-body (MB) interactions is in order, are often unsatisfactory or partially missing in DFT models of water. In this review, we discuss the modeling of water and aqueous systems based on DFT and provide a comprehensive description of a general theoretical/computational framework for the development of data-driven many-body potentials from DFT reference data. This framework, coined MB-DFT, readily enables efficient many-body molecular dynamics (MD) simulations of small molecules, in both gas and condensed phases, while preserving the accuracy of the underlying DFT model. Theoretical considerations are emphasized, including the role that the delocalization error plays in MB-DFT potentials of water and the possibility to elevate DFT and MB-DFT to near-chemical-accuracy through a density-corrected formalism. The development of the MB-DFT framework is described in detail, along with its application in MB-MD simulations and recent extension to the modeling of reactive processes in solution within a quantum mechanics/MB molecular mechanics (QM/MB-MM) scheme, using water as a prototypical solvent. Finally, we identify open challenges and discuss future directions for MB-DFT and QM/MB-MM simulations in condensed phases. 
    more » « less
  4. ; ; ; ; (, Nature Communications)
    Abstract Interatomic potentials derived with Machine Learning algorithms such as Deep-Neural Networks (DNNs), achieve the accuracy of high-fidelity quantum mechanical (QM) methods in areas traditionally dominated by empirical force fields and allow performing massive simulations. Most DNN potentials were parametrized for neutral molecules or closed-shell ions due to architectural limitations. In this work, we propose an improved machine learning framework for simulating open-shell anions and cations. We introduce the AIMNet-NSE (Neural Spin Equilibration) architecture, which can predict molecular energies for an arbitrary combination of molecular charge and spin multiplicity with errors of about 2–3 kcal/mol and spin-charges with error errors ~0.01e for small and medium-sized organic molecules, compared to the reference QM simulations. The AIMNet-NSE model allows to fully bypass QM calculations and derive the ionization potential, electron affinity, and conceptual Density Functional Theory quantities like electronegativity, hardness, and condensed Fukui functions. We show that these descriptors, along with learned atomic representations, could be used to model chemical reactivity through an example of regioselectivity in electrophilic aromatic substitution reactions. 
    more » « less
  5. ; ; ; ; (, The Journal of Chemical Physics)
    Dimer interaction energies have been well studied in computational chemistry, but they can offer an incomplete understanding of molecular binding depending on the system. In the current study, we present a dataset of focal-point coupled-cluster interaction and deformation energies (summing to binding energies, De) of 28 organic molecular dimers. We use these highly accurate energies to evaluate ten density functional approximations for their accuracy. The best performing method (with a double-ζ basis set), B97M-D3BJ, is then used to calculate the binding energies of 104 organic dimers, and we analyze the influence of the nature and strength of interaction on deformation energies. Deformation energies can be as large as 50% of the dimer interaction energy, especially when hydrogen bonding is present. In most cases, two or more hydrogen bonds present in a dimer correspond to an interaction energy of −10 to −25 kcal mol−1, allowing a deformation energy above 1 kcal mol−1 (and up to 9.5 kcal mol−1). A lack of hydrogen bonding usually restricts the deformation energy to below 1 kcal mol−1 due to the weaker interaction energy. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Text Encoded Conversion
  • XML
  • Save / Share this Record
  • Send to Email