Excited state optimization for strongly correlated quantum defects using ensemble variational Monte Carlo
arXiv:2607.27377
The paper applies ensemble variational Monte Carlo to optimize excited‑state wavefunctions of strongly correlated point defects such as NV and SiV centers in diamond and Fe/Cr impurities in AlN, showing that hybrid functional orbitals and full parameter optimization improve predicted excitation energies.
Abstract
Using the recently introduced ensemble variational Monte Carlo (VMC), we study optimized wave functions for strongly correlated point defects, including nitrogen-vacancy and silicon-vacancy centers in diamond and substitutional iron and chromium impurities in aluminum nitride. We study the effects of fully optimized determinant expansion parameters, orbitals, and Jastrow correlation factors on these systems. We find that orbitals from the hybrid functional PBE0 are much better (have lower objective functional) than semilocal PBE, which results in changes in the excitation energies up to 0.5 eV. Further improvements can be made by directly optimizing the objective functional, resulting in changes in excitation energies up to 0.2 eV. The most important parameter varies from defect to defect, reinforcing the necessity of optimizing all parameters to obtain accurate excited states in strongly correlated defects.