Using multideterminant methods and quantum Monte Carlo to characterize a strongly correlated material*

Understanding and accounting for strong correlations in ab initio simulations is essential for predicting properties of novel materials.

Density functional theory has been the workhorse of simulation for several decades and has demonstrated incredible ability at predicting a wide range of phenomena. However, the method is not as reliable for strongly correlated materials or van der Waals dominated materials, requiring some approximation/correction which can hinder its predictive capability.

In recent years, many-body methods such as diffusion Monte Carlo (DMC) have shown great success at reproducing ground state and excited state properties of many molecules and solids containing 3d and 4d metals.

While the accuracy of DMC depends on the nodes of a trial wavefunction, the error due to an inexact nodal surface can be systematically improved by using larger multi-determinant expansions in trial wavefunctions.

In this work, we present results of DMC calculations on a strongly correlated transition metal material including the use of selected configuration interaction (sCI) to construct better trial wavefunctions.