Diagnosing phase transitions through time scale entanglement (2507.11276v1)

Abstract: We propose a new method based on the compression of arbitrary multi-point correlators into quantics tensor trains and studying its so called bond dimension as an indicator of time scale entanglement in the system. This quantics tensor train diagnostics approach allows us to i) identify quantum phase transitions (ground state crossings) and thermal crossovers, ii) distinguish between these two by allowing flexibility through iii) using arbitrary available correlators and iv) the possibility to deal with low-accuracy input data through varying the quantics tensor train cutoffs. We verify the validity of the approach by studying the phase diagram of the Hubbard dimer, the four-site Hubbard ring with and without next-nearest neighbor hopping and the single-impurity Anderson model. We find time scale entanglement to be a system-inherent property becoming maximal at phase transitions and crossover, which translates to peaks in the quantics tensor train bond dimension of correlator functions. Additionally, the connection of the method to other physical quantities, such as various entanglement measures and susceptibilities, is shown.