APS Logo

Benchmark all-electron ab initio quantum Monte Carlo calculations with Antisymmetrized Geminal Power (AGP) and Pfaffian Ansatz for G2-set molecules using TurboRVB package

ORAL

Abstract

In this work, we report accurate binding energy calculations for 55 molecules in the G2 set, using lattice regularized diffusion Monte Carlo (LRDMC) within TurboRVB quantum Monte Carlo package. We employ both the more traditional Ansatz namely the Jastrow Slater Determinant (referred here as JDFT) as well as the more flexible Ansatz namely the Jastrow antisymmetrized geminal power with singlet correlation (JAGPs) and the Pfaffian (Pf) including both singlet and triplet correlations. Being the most general electron pairing function, we expect the Pfaffian Ansatz to be the most efficient in determining the correlation energy. The many-body wave functions are first optimized at the variational Monte Carlo (VMC) level (Jastrow factor + nodal surface optimization), followed by the LRDMC projection of the Ansatz. Remarkably, for many molecules the LRDMC binding energies obtained using the JAGPs Ansatz reach the chemical accuracy (~1 kcal/mol) and for most of the other cases the binding energies are accurate within ~5 kcal/mol. We further expect to improve upon the JAGPs binding energies by using the Pfaffian Ansatz. This work shows the effectiveness of these more flexible Ansatz in TurboRVB code for binding energy calculations and electronic simulations in general.

Publication: Planned paper: Benchmark all-electron ab initio quantum Monte Carlo calculations with Antisymmetrized Geminal Power (AGP) and Pfaffian Ansatz for G2-set molecules using TurboRVB package

Presenters

  • Abhishek Raghav

    School of Information Science, JAIST

Authors

  • Abhishek Raghav

    School of Information Science, JAIST

  • Ryo Maezono

    School of Information Science, JAIST, School of Information Science, JAIST, Nomi, Ishikawa, Japan, School of Information Science, JAIST, Nomi, Ishikawa, Japan.

  • Kenta Hongo

    Research Center for Advanced Computing Infrastructure, JAIST, Research Center for Advanced Computing Infrastructure, JAIST, Nomi, Ishikawa, Japan, Research Center for Advanced Computing Infrastructure, JAIST, Nomi, Ishikawa, Japan.

  • Sandro Sorella

    International School for Advanced Studies, Condensed Matter Theory group, SISSA

  • Kousuke Nakano

    1. International School for Advanced Studies 2. School of Information Science, JAIST, School of Information Science, JAIST, School of Information Science, JAIST, Nomi, Ishikawa, Japan., Condensed matter theory, SISSA