We perform quantum Monte Carlo simulations to study the 2D hard-core Bose–Hubbard model in a random potential. Our motivation is to investigate the effects of randomness on the Kosterlitz–Thouless (KT) transition and its quantum critical phenomenon. The chemical potential is assumed to be random, by site, with a Gaussian distribution. The KT transition is confirmed by a finite-size analysis of the superfluid density and the power-law decay of the correlation function. By changing the variance of the Gaussian distribution, we find that the transition temperature decreases as the variance increases. We obtain the phase diagram showing the superfluid and the disordered phases, and estimate the quantum critical point (QCP) and the critical exponent. Our results on the ground state show the existence of the Bose glass phase. Finally, we discuss what the variance of the QCP indicates from the viewpoint of quantum percolation .
We study the universal critical behavior of two-dimensional (2D) lattice bosonic gases at the Berezinskii-Kosterlitz-Thouless (BKT) transition, which separates the low-temperature superfluid phase from the high-temperature normal phase. For this purpose, we perform quantum Monte Carlo simulations of the hard-core Bose-Hubbard (BH) model at zero chemical potential. We determine the critical temperature by using a matching method that relates finite-size data for the BH model with corresponding data computed in the classical
XY model. In this approach, the neglected scaling corrections decay as inverse powers of the lattice size
L, and not as powers of
1/lnL, as in more standard approaches, making the estimate of the critical temperature much more reliable. Then, we consider the BH model in the presence of a trapping harmonic potential, and we verify the universality of the trap-size dependence at the BKT critical point. This issue is relevant for experiments with quasi-2D trapped cold atoms.