Elevating Density Functional Theory to Chemical Accuracy for Water Simulations through a Density-Corrected Many-Body Formalism

Abstract

Kohn-Sham density functional theory (DFT) has been extensively used to model the properties of water. Albeit maintaining a good balance between accuracy and efficiency, no density functional has so far achieved the degree of accuracy necessary to correctly predict the properties of water across the entire phase diagram. The recent development of the strongly constrained and appropriately normed (SCAN) functional has renewed the interest in ab initio simulations of liquid water, yielding promising results that are, however, still unable to reproduce all the experimental data. Here, we present density-corrected SCAN (DC-SCAN) calculations for water which, minimizing density-driven errors, elevate the accuracy of the SCAN functional to that of coupled cluster theory, the "gold standard" for chemical accuracy. Building upon the accuracy and efficiency of DC-SCAN within a many-body formalism, we introduce a data-driven many-body potential energy function, the MB-SCAN(DC) PEF, that is able to quantitatively reproduce coupled cluster reference values for interaction, binding, and individual many-body energies of water clusters. Importantly, the properties of liquid water calculated from molecular dynamics simulations carried out with the MB-SCAN(DC) PEF are found to be in excellent agreement with the experimental data, which thus demonstrates that MB-SCAN(DC) is effectively the first DFT-based model that correctly describes water from the gas to the condensed phase. Since the many-body formalism adopted by the present MB-SCAN(DC) PEF for water is general, we believe it can open the door to the routine development of data-driven many-body PEFs for predictive simulations of generic (small) molecules in the gas, liquid, and solid phases.Its anomalous behavior 1 and importance to life 2 make water one of the most studied chemical compounds. Among its many unique properties is the high value of the heat capacity which allows water to resist sudden temperature changes, thus permitting living organisms to survive without experiencing significant temperature fluctuations. 3 In addition, the dynamic nature of the water hydrogen-bond network plays a central role in several fundamental processes, including transport and diffusion in bulk solutions and at interfaces, and hydration of hydrophobic and hydrophilic solutes. 4 For example, protein folding is thought to be driven by the hydrophobic effect. 5 Finally, countless chemical reactions involving charged species take place efficiently in liquid water due to

Liquid water simulation using hydrogen bond corrected SCAN and neural network potentials.

“…One option to correct these issues is to go to higher levels of theory such as the random phase approximation, density corrected or hybrid SCAN or MP2. 8,[10][11][12][13] But these can be prohibitively expensive even to obtain the training data to run simulations and inaccuracies can remain.…”