Abstract:
The class of quantum impurity models, which involves a subsystem with a limited number of degrees of freedom (an impurity) coupled to a much larger system (a bath) with a quasi-continuum of degrees of freedom, has been of longstanding interest. Especially, the continuous development of dynamical mean-field theory (DMFT), in which the correlated lattice problem is mapped self-consistently to an effective impurity model, necessitates an efficient and accurate solution of the impurity ground state and one-body Green’s functions. In this talk, I will discuss our efforts in tackling such challenges using tensor network states: i) Optimizing the ground state entanglement through single-particle basis rotation and subsequently introducing a tailored tree tensor network that optimizes this entanglement structure for general impurity problems; ii) Utilizing a hybrid purification and minimal entangled typical thermal states (METTS) sampling framework for finite temperature calculations; iii) Accurately calculating low-energy (long-time) dynamics by conducting time evolution in the complex plane, complemented by methods to reconstruct the associated real-time correlation functions.