Chemical Transformations Approaching Chemical Accuracy via Correlated Sampling in Auxiliary-Field Quantum Monte Carlo

摘要

The exact and phaseless variants of Auxiliary-Field Quantum Monte Carlo(AFQMC) have been shown to be capable of producing accurate ground-stateenergies for a wide variety of systems including those which exhibitsubstantial electron correlation effects. The computational cost of performingthese calculations has to date been relatively high, impeding many importantapplications of these approaches. Here we present a correlated samplingmethodology for AFQMC which relies on error cancellation to dramaticallyaccelerate the calculation of energy differences of relevance to chemicaltransformations. In particular, we show that our correlated sampling-basedAFQMC approach is capable of calculating redox properties, deprotonationfree-energies, and hydrogen abstraction energies in an efficient manner withoutsacrificing accuracy. We validate the computational protocol by calculating theionization potentials and electron affinities of the atoms contained in the G2Test Set, and then proceed to utilize a composite method, which treatsfixed-geometry processes with correlated sampling-based AFQMC and relaxationenergies via MP2, to compute the ionization potential, deprotonationfree-energy, and the O-H bond disocciation energy of methanol, all to withinchemical accuracy. We show that the efficiency of correlated sampling relativeto uncorrelated calculations increases with system and basis set size, and thatcorrelated sampling greatly reduces the required number of random walkers toachieve a target statistical error. This translates to CPU-time speed-upfactors of 55, 25, and 24 for the the ionization potential of the K atom, thedeprotonation of methanol, and hydrogen abstraction from the O-H bond ofmethanol, respectively. We conclude with a discussion of further efficiencyimprovements that may open the door to the accurate description of chemicalprocesses in complex systems.