Bose-Einstein condensation in optical lattices: the superfluid-Mott-insulator transition in hexagonal lattices and the superfluid-Bose-glass universality class in 3D

In this thesis, we have studied phase transitions in ultracold atoms trapped in optical lattices. The physics of these systems is captured by Bose-Hubbard-like models, which predicts a quantum phase transition (the so called superfluid-Mott insulator, or SF-MI) when varying the potential depth of the optical lattice in a system without disorder, where atoms have short range interactions, and tunneling is allowed only between nearest neighbors. Our studies followed two directions, one is concerned with the influence of the geometry of the lattice namely, we study the changes in the phase diagram of the SF-MI phase transition when the optical lattice is hexagonal. A second direction is to include disorder in the original system. In our study of the hexagonal lattice, we obtain the phase diagram for the SF-MI transition and give an approximation for the critical point of the first Mott lobe, using a quantum Monte Carlo algorithm called Worm. We also compare our results with the ones from the squared lattice and obtained using mean-field approximation. When disorder is included in the system, a new phase emerge in the ground-state phase diagram intermediating the superfluid and Mott-insulator phases. This new phase is called Bose-glass (BG) and the quantum phase transition SF-BG was the subject of many controversies since its first studies in the late 80s. Though many progress towards its thorough understanding were made, basics characterization of critical proprieties are still under debate. Our study was motivated by the publication of recent experimental and numerical studies in three-dimensional systems [Yu et al. Nature 489, 379 (2012), Yu et al. PRB 86, 134421 (2012)] reporting strong violations of the key quantum critical relation, $\\phi= u z$, where $\\phi$ is the critical-temperature exponent, $z$ and $ u$ are the dynamic and correlation length critical exponents, respectively. We addressed this controversy numerically performing finite-size scaling analysis using the Worm algorithm, both in its quantum and classical scheme. Our results demonstrate that previous work on the superfluid-to-normal fluid transition-temperature dependence on chemical potential (or magnetic field, in spin systems), $T_c \\propto (\\mu-\\mu_c)^\\phi$, was misinterpreting transient behavior on approach to the fluctuation region with the genuine critical law. When the model parameters are modified to have a broad quantum critical region, simulations of both quantum and classical models reveal that the $\\phi= u z$ law [with $\\phi=2.7(2)$, $z=3$, and $ u = 0.88(5)$] holds true. We also estimate the order parameter exponent, finding $\\beta=1.5(2)$. (AU)