Efficient Formulation of Ab Initio Quantum Embedding in Periodic Systems: Dynamical Mean-Field Theory

Area of Science:

• Condensed Matter Physics
• Computational Materials Science
• Quantum Chemistry

Background:

• Accurate simulations of electron correlation in solids are crucial for understanding material properties.
• Dynamical Mean-Field Theory (DMFT) is a powerful method for capturing local electron correlations.
• Efficient and quantitative ab initio implementations of DMFT are needed for realistic material systems.

Purpose of the Study:

• To develop and present an efficient ab initio dynamical mean-field theory (DMFT) implementation for quantitative simulations in solids.
• To enable the study of complex materials with realistic quantum chemical basis sets and large local environments.
• To validate the accuracy of the developed ab initio DMFT scheme against established methods and experimental data.

Main Methods:

• Implementation of ab initio DMFT using Hamiltonians for unit cells or supercells.
• Utilization of Hartree-Fock for double counting correction and Intrinsic and Projected Atomic Orbitals (IAO + PAO) for local embedding.
• Employment of efficient integral transformation and coupled-cluster Green's function impurity solvers.
• Application to hexagonal boron nitride, crystalline silicon, and antiferromagnetic nickel oxide.

Main Results:

• The ab initio DMFT scheme successfully handles embedded impurity problems with hundreds of orbitals.
• Calculations were performed for hexagonal boron nitride, crystalline silicon, and nickel oxide, involving up to 104 impurity orbitals.
• The method produced spectral functions that show excellent agreement with benchmark coupled-cluster computations and experimental spectra.

Conclusions:

• The presented ab initio DMFT implementation offers an efficient and accurate approach for quantitative simulations of solids.
• The scheme effectively captures electron correlation effects in diverse material systems.
• This work paves the way for more reliable theoretical predictions of material properties.