Density matrix embedding: a simple alternative to dynamical mean-field theory.

Area of Science:

• Quantum chemistry
• Condensed matter physics
• Computational materials science

Background:

• Accurate computation of ground-state properties for infinite quantum systems is computationally challenging.
• Existing methods like dynamical mean-field theory (DMFT) have limitations in formulation and computational cost.

Purpose of the Study:

• Introduce Density Matrix Embedding Theory (DMET) as a novel quantum embedding approach.
• Develop a computationally efficient method for frequency-independent properties of infinite systems.

Main Methods:

• DMET formulates the problem using the frequency-independent local density matrix.
• It employs a finite, minimal bath with one bath site per impurity site, avoiding bath discretization errors.
• The method maps the bulk system to a simpler impurity model, exact in noninteracting and atomic limits.

Main Results:

• DMET successfully reproduces total energies and correlation functions for 1D and 2D Hubbard models.
• Accurate prediction of metal-insulator transitions was achieved.
• The method demonstrated high accuracy at a significantly reduced computational cost compared to benchmarks.

Conclusions:

• DMET provides a computationally simple and efficient quantum embedding method.
• Its frequency independence and minimal bath design make it suitable for large-scale electronic structure calculations.
• DMET shows promise for accurate prediction of material properties.