Occupancy Fluctuation Effect on Metal-Insulator Transition in 3D Hubbard Model within Dynamical Mean-Field Theory Framework

Dynamical Mean-Field Theory (DMFT) is among the most widely used techniques in computational condensed-matter physics, especially in addressing electron-electron interactions. In this regard, Quantum Monte Carlo may be considered as the most accurate DMFT impurity solver, but it usually requires a high numerical cost. On the contrary, lower numerical-cost technique, such as mean-field theory, often gives inaccurate results due to complete suppression of quantum fluctuations. In this study, we propose a new potentially low numerical-cost impurity solver taking a full account of fluctuation effects. Applying this new algorithm on Hubbard model, we formulate local self-energy that depends on the occupancy fluctuations. These fluctuations act as the semi-classical degree of freedom that needs to be integrated to obtain the averaged fully interacting Green functions. We test our algorithm by addressing temperature dependence of the density of states that reveals metal-insulator transition.