Abstract
Recently, the quantum topological energy partitioning method called interacting quantum atoms (IQA) has been extended to MPn (n = 2, 3, 4) wave functions. This enables the extraction of chemical insight related to dynamic electron correlation. The large computational expense of the IQA‐MPn approach is compensated by the advantages that IQA offers compared to older nontopological energy decomposition schemes. This expense is problematic in the construction of a machine learning training set to create kriging models for topological atoms. However, the algorithm presented here markedly accelerates the calculation of atomically partitioned electron correlation energies. Then again, the algorithm cannot calculate pairwise interatomic energies because it applies analytical integrals over whole space (rather than over atomic volumes). However, these pairwise energies are not needed in the quantum topological force field FFLUX, which only uses the energy of an atom interacting with all remaining atoms of the system that it is part of. Thus, it is now feasible to generate accurate and sizeable training sets at MPn level of theory. © 2019 The Authors. Journal of Computational Chemistry published by Wiley Periodicals, Inc.
Highlights
Where V is the correlation energy between any two atoms A and B, d is the correlated part of the 2PDM, the G functions are Gaussians arising from products of Gaussian primitives originally centred at nuclei, K refers to the product pre-factor, and there are Nbasis Gaussian primitives
We first consider the size of grid necessary to achieve a given recovery error in the correlation energy of the whole system
The fact that the recovery error obtained through the 6D algorithm heavily depends on the system size creates an unsustainable situation to transition FFLUX from small molecules to large biological systems. This is not the case for the 3D electrostatic potential (ESP) algorithm because the recovery error is largely independent of the system size
Summary
Atomistic force fields continue to be developed[1–10] given their growing use and impact in biological and materials science. Extensive comparisons[11–14] between, traditional force fields can offer guidance in this development, as well as comparisons[15,16] between machine-learnt potentials. Within this grand scheme of force field development, the modeling of dispersion[17] energy has received special attention over the last decade and more. It is possible to circumvent perturbation theory and focus directly on the electron correlation, from a single supermolecular system. This is the route we follow here. We calculate dynamical correlation energy extracted from the two-particle density-matrix (2PDM) of this single supermolecular (quantum) system. Substituting eq (1) into this sum leads to the derivation shown in eq (2), X
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
