2013/02/08 by Kumar, Sanjeev, Chakraborty, Prabuddha B.
#Disordered Systems and Neural Networks (cond-mat.dis-nn) #FOS: Physical sciences #Strongly Correlated Electrons (cond-mat.str-el) #Superconductivity (cond-mat.supr-con)
paper · doi:10.48550/arxiv.1302.1967
We study the effect of disorder symmetry in a disordered s-wave superconductor. We begin with an attractive Hubbard model on a two-dimensional square lattice in the presence of diagonal- and off-diagonal disorder. The model is studied numerically using Bogoliubov-deGennes mean-field approach. We find that the off-diagonal disorder is much more efficient in destroying superconducting order. A detailed analysis brings out very distinct qualitative pictures in the two cases: diagonal disorder leads to formation of small scale isolated superconducting islands separated by large region of normal metal, whereas off-diagonal disorder leads to a connected network of superconducting region. Naively, it seems that the connected network of superconducting regions should have a large average value for superconducting order parameter, but this does not hold true for large disorder, as we demonstrate explicitly. Our qualititive picture is supported by real-space data on local order parameters.