Abstract
The metal-insulator transition in arc physics is studied theoretically using a disordered one-dimensional model which combines the arc welding features of the Longini model and the Coulomb correlation effects of the Hubbard model. The model is binary in nature; one specie of atoms will simulate the bulk metal while the other represents the plasma, tantamount to a vacuum. The number of atoms of one species follows a linearly-decreasing function within fixed intervals in the model. The numerical parameters used are chosen with the physical characteristics of copper in mind. The band structure of the half-filled non-degenerate d band is calculated using self-consistent Hartree-Fock approximation for both the non-spin- and spin-polarized cases. The density of states is analyzed on the basis of the Mott-Hubbard bands. These bands are used to demonstrate the presence of metal-insulator transition.Results indicate that the metal-insulator transition does indeed occur in the model. The location, in space, for the onset of the metal-insulator transition is determined from conductance calculations. The conductance is calculated using both the WKB approximation in the spirit of Longini's calculation and the Landauer conductance formula. Results show that the former is quite inadequate for the determination of the onset of metal-insulator transition. By using the results of the Landauer conductance, it is found that the arc model is successful in extending the region of metallic conductivity far into the vacuum, as indicated by the "delayed" point for the onset of the metal-insulator transition.