A Perturbation Expansion Method to Study Highly Correlated Spins

This paper proposes a new numerical algorithm to study dynamical spin dependent properties of local highly correlated structures. The method consists in diagonalizing a finite cluster containing the many-body terms of the Hamiltonian and embedding it into the rest of the system, the Embedding Cluster Approximation (ECA), combined with Wilson’s ideas of logarithmic discretization of the representation of the Hamiltonian, the Logarithm Discretization Embedded Cluster Approximation (LDECA). The physics associated to a dot and a side-coupled double dot connected to leads are discussed in detail.     

Influence of long-range interactions on superconductivity and phase separation

The t-J model is studied including a long-range 1/r repulsive interaction. It is observed that charge-density-wave states become stable as the strength of the 1/r term, V_coul, is increased. Due to this effect, the domain of stability of the superconducting phase that appears near phase separation at V_coul=0 is not enlarged by a 1/r interaction as naively expected. Nevertheless, superconductivity exists in a finite region of parameter space, even if phase separation is suppressed.