Resumen de Monte Carlo study of quantum phase transitions at zero temperature
Oleg Osychenko
The Thesis is devoted to simulations of quantum phase transitions by means of Quantum Monte Carlo techniques. Quantum phase transition is a transition between phases at zero or low enough temperature, where quantum effects play an important role. The recent advances in the field of ultracold atom manipulation and optical lattices allowed to produce the systems with unique properties. This opened a perspective to observe quantum phase transitions in many-body systems with non-trivial interparticle interactions in a wide range of the system's characteristic physical parameters and geometries. First, we develop the explicit expressions for the Ewald sums in systems with an interaction potential of a generic 1/r^k type, and in 3D, 2D and 1D geometry. These generalizations can be useful in simulating systems with important interaction potentials as the dipole-dipole, van der Waals interaction, etc. In this Thesis we give the functional forms for the terms of the Ewald sums, ready for implementation in a code. The derivation and the functional form of the results differ in the cases of short-ranged, long-ranged and "marginal" forces, and for a jellium model. It is argued that in the case of some short-range potentials the Ewald method can be advantageous with respect to a direct summation due to a faster convergence rate. We also give a discussion of the convergence properties of a quasi-neutral Coulomb system. We have obtained the zero-temperature phase diagram of bosons interacting through Yukawa forces. We have used a diffusion Monte Carlo simulation starting from a good approximation to the optimal variational ground-state wave function obtained by solving the corresponding Euler-Lagrange hypernetted chain equations. The phase diagram shows that any fermionic mixture of pure elements will always be seen in gaseous form, as the mass ratios required for crystallization of weakly bound fermionic molecules are far beyond the ones that can be achieved in nature. We investigate an alternative mechanism based on the confinement of one of the species to a deep optical lattice which increases its effective mass. The resulting mass ratio of the mixture created in this way can then be tuned at will and could be used to check experimentally the predicted phase diagram both in the gas and crystal (superlattice) phases. We performed a QMC study of the system, comrised of Rydberg atoms. The applied QMC techniques allowed to parametrize a model with isotropic van der Waals interactions into a universal phase diagram. We have characterized the phase diagram of Rydberg atoms by considering a model of bosons with repulsive van der Waals 1/r^6 interaction, and determined solidification and Bose-Einstein condensation conditions. Relaxation mechanisms other than thermal motion should be considered if one considers Rydberg systems on timescales of several tenths of microseconds. We have also studied the excitation spectrum within the approximation of a classical harmonic crystal. We also discuss that interactions between Rydberg excitations open a possibility of new supersolid scenarios. In the last Chapter of the Thesis I present a study of the system of para-hydrogen atoms at low temperatures below the point of crystallization by means of QMC methods. The zero-temperature simulation was performed in order to investigate the properties of a metastable liquid phase and to find the fraction of the Bose-Einstein condensate in the relevant range of densities. The methods of choice for the zero-temperature simulations of the para-H2 system were VMC and DMC techniques. The results of the zero-temperature simulations suggest that the metastable liquid para-hydrogen is a strongly correlated liquid, which again serves as an evidence of high instability of this hypothetical system. The calculation of the Bose-Einstein condensate shows that the condensate fraction is substantially lower than in the liquid helium He4.
