Dynamical mean field theory with the density matrix renormalization group.

摘要: A new numerical method for the solution of dynamical mean field theory's self-consistent equations is introduced. The uses density matrix renormalization group technique to solve associated impurity problem. algorithm makes no a priori approximations and only limited by number sites that can be considered. We obtain accurate estimates critical values metal-insulator transitions provide evidence substructure in Hubbard bands correlated metal. With this algorithm, more complex models having larger degrees freedom considered finite-size effects minimized. (10). While very stable allowed detailed investigations fundamental prob- lems such as transition model (3,11), its main drawback access real frequency quantities spectral func- tions requires recourse less controlled techniques analytic continuation Green functions. second based on Wilson's renormal- ization (12,13). This formulated both at T � 0 finite (small) providing results small frequencies; thus it adequate investigation metallic phases with heavy effective-mass quasiparticles. cost pay description high energy features involve approx- imations cannot so accurately obtained (14). goal present work introduce DMFT use another powerful methodology many-body Hamilton- ians: (DMRG) (15). method, like NRG, has appealing feature making possibility systematic improvement quality solutions. However, unlike not low-frequency asymptotic (14) may equally reliable solutions gapless gapfull phases. More significantly, provides distributions intensities are relevance analysis x-ray photoemission optical conductivity experiments. shall illustrate formulation solu- tion the, now classic, Mott model. show criti- cal interactions metal-to-insulator insulator-to-metal obtained, and, interestingly, we find phase.