Localization-Delocalization Transitions in Bosonic Random Matrix Ensembles
Abstract
Localization to delocalization transitions in eigenfunctions are studied for finite interacting boson systems by employing one- plus two-body embedded Gaussian orthogonal ensemble of random matrices [EGOE(1+2)]. In the first analysis, considered are bosonic EGOE(1+2) for two-species boson systems with a fictitious () spin degree of freedom [called BEGOE(1+2)-]. Numerical calculations are carried out as a function of the two-body interaction strength (). It is shown that, in the region (defined by ) after the onset of Poisson to GOE transition in energy levels, the strength functions exhibit Breit-Wigner to Gaussian transition for . Further, analyzing information entropy and participation ratio, it is established that there is a region defined by where the system exhibits thermalization. The -spin dependence of the transition markers and follow from the propagator for the spectral variances. These results, well tested near the center of the spectrum and extend to the region within to from the center ( is the spectral variance), establish universality of the transitions generated by embedded ensembles. In the second analysis, entanglement entropy is studied for spin-less BEGOE(1+2) ensemble and shown that the results generated are close to the recently reported results for a Bose-Hubbard model.
Cite
@article{arxiv.1611.01970,
title = {Localization-Delocalization Transitions in Bosonic Random Matrix Ensembles},
author = {N. D. Chavda and V. K. B. Kota},
journal= {arXiv preprint arXiv:1611.01970},
year = {2017}
}
Comments
11 pages, 6 figures, Contribution to the Special Issue "Many-Body Localization" in Annalen der Physik