q-AQUA: A Many-Body CCSD(T) Water Potential, Including Four-Body Interactions, Demonstrates the Quantum Nature of Water from Clusters to the Liquid Phase

Abstract

Many model potential energy surfaces (PESs) have been reported for water; however, none are strictly from “first-principles”. Here we report such a potential, based on a many-body representation at the CCSD­(T) level of theory up to the four-body interaction. The new PES is benchmarked for the isomers of the water hexamer for dissociation energies, harmonic frequencies, and unrestricted diffusion Monte Carlo (DMC) calculations of the zero-point energies of the Prism, Book, and Cage isomers. Dissociation energies of several isomers of the 20-mer agree well with recent benchmark energies. Exploratory DMC calculations on this cluster verify the robustness of the new PES for quantum simulations. The accuracy and speed of the new PES are demonstrated for standard condensed phase properties, i.e., the radial distribution function and the self-diffusion constant. Quantum effects are shown to be substantial for these observables and also needed to bring theory into excellent agreement with experiment.

A Δ-machine learning approach for force fields, illustrated by a CCSD(T) 4-body correction to the MB-pol water potential

“…And crucially, we have the CCSD(T) electronic energies needed for the correction up to 4-b, and these datasets were recently used to develop the pure many-body water potential q-AQUA. 15,16 V q-AQUA ¼…”

Many-body Δ-Machine Learning brings the accuracy of conventional force field to coupled cluster: application to the TTM2.1 water force field

“…For example, the total number of 3b interaction terms in this 256mer water system is The corresponding timing for an energy plus gradient on 8 cores for the q-AQUA potential is 0.97 sec on the same cluster (reported in Table S5 of ref. 15). This is only 2.8 times faster than the present ∆-corrected FF and most of this factor is accounted for by the timing of TTM2.1.…”

“…4. The results shown there in the Table 1 are even more accurate those from our q-AQUA potential, which is truncated at the 4-b level 15 and also more accurate than the MB-pol potential.…”

“…27 We corrected the 2-b, 3-b and 4-b interactions targeting the CCSD(T) level of accuracy. The datasets of CCSD(T) energies for 2, 3 and 4-b interactions were recently reported by us in developing a new water potential, q-AQUA, 15 that is exactly truncated at the 4-b level, which we showed is very accurate from clusters to the condensed phase. It is certainly of interest to extend to higher-body interactions, and that is achieved using the current approach, but relying on TTM2.1 for 5-and higher-body interactions.…”

“…The corrected FF has unprecedented agreement with the CCSD(T) results compared to MBpol 38 and even slightly better than q-AQUA. 15 The detailed n-body analysis presented in Table 1shows the accuracy of the ∆-corrected FF for each n and, impressively, for n > 4 TTM2.1 provides the additional small term that results excellent accuracy of the ∆-ML corrected FF. This analysis is presented graphically in Fig.…”