Budget Amount *help |
¥1,100,000 (Direct Cost: ¥1,100,000)
Fiscal Year 2003: ¥400,000 (Direct Cost: ¥400,000)
Fiscal Year 2002: ¥700,000 (Direct Cost: ¥700,000)
|
Research Abstract |
[1]We have re-examined the microscopic formulation of the transport through small interacting systems connected to noninteracting leads based on the Kubo formalism, and have shown quite generally that the many body transmission coefficient Τ(ε), by which the dc conductance is expressed in a Landauer type form g=(2e^2/h) ∫ dε (-∂f/∂ε) Τ(ε) where f(ε) is the Fermi function, can be related to a three-point retarded function defined in the real time. [2]We have applied the linear-response formulation to the quantum-dot superlattice, which is modeled by a two-dimensional Hubbard cluster connected to two noninteracting leads. The results, which are obtained by using the second-order self-energy in the perturbation expansion with respect to the Coulomb interaction U, demonstrate how the structure of resonance peaks is affected by the interaction. Especially, the width and the position of the resonance are sensitive to the correlation effects. [3]We have also studied nonequilibrium properties of the Kondo effects in quantum dots based on the Anderson model under a finite bias voltage V. We have shown that for small but finite V the low-energy behaviors of the excitation spectrum and nonlinear response of the current can be described by the local Fermi-liquid theory. Furthermore, we have deduced the exact asymptotic behavior of the Green's function in the limit of the large bias voltage V. [4]We have also launched nonperturbative calculations of the transport properties using the numerical renormalization group method. With this approach, we have studied the Kondo effect of a quantum dot in a Josephson junction, and have clarified precise features of the quantum phase transition between the non-magnetic singlet and magnetic doublet ground states. We are now applying the method to a Mott-Hubbard insulator of a nanometer scale.
|