LDECA and DMRG: Feasible Combination to study quantum transport.?

CLAUDIO GAZZA

Niteroi

Workshop; CIAM WORKSHOP 2013; 2013

COLABORACIÓN INTERAMERICANA EN MATERIALES CIAM. NSF(USA), CNPq(Brasil) and CONICET(Argentina)

With the advent of new techniques for fabrication of mesoscopic systems, and the development of new materials (especially carbon based, like carbon nanotubes and graphene), the variety of ´man made´ nanostructures presenting many-body interactions has advanced tremendously, challenging the suitability of current numerical techniques able to solve multi-impurity Anderson models and extract their charge and spin transport properties. Here we present a new numerical technique that is a natural development and adaptation, along 15 years, of different ideas: cluster embedding, logarithmic discretization, and a recently presented way to calculate very efficiently green functions within DMRG. This method, dubbed LDECA+DMRG [1-3], is here applied to a double quantum dot system presenting a two-stage Kondo effect. Given the large difference in energy scales presented in the sequence of Kondo effects (as temperature is powered through many orders of magnitude) and the large range of gate voltages taken by the crossover region between a simple molecular state and the TSK regime, this problem presents an ideal benchmark for testing the new technique. In comparing our results with published NRG results, the authors show the advantages and limitations of the new technique and illustrate a few problems where its application would be advantageous. [1] LDECA stands for logarithmic-discretization embedded-cluster approximation, E. V. Anda et al., Phys. Rev. B 78, 085308 (2008). [2] S. R. White, Phys. Rev. Lett. 69, 2863 (1992). [3] Adaptive Lanczos-vector method for dynamic properties within the DMRG. P. E. Dargel, et al., Phys. Rev. B 83, 161104(R) (2011)