Department of Physics, Engineering Physics & Astronomy

Department of Physics, Engineering Physics & Astronomy
Department of Physics, Engineering Physics & Astronomy

Metal-to-Insulator Transition in Quantum Site Percolation

Farhad Fazileh
Department of Physics, Queen's University

Wednesday, April 9, 2003
11:30 AM @ Stirling A

Abstract:

Superconductivity among spinel systems is very rare: of the 300 or so known spinels, only four of them are superconductors, only one of these four is an oxide, and that oxide, LiTi_2O_4, has the highest transition temperature (T_c ~ 13K) of any spinel. The mechanism of superconductivity in this compound has not yet been identified. The notion that Lithium Titanate's superconductivity is due to electronic correlations was first suggested by high-temperature superconductivity co-discoverer Alex Muller. As evidence supporting the hypothesis that strong electronic correlations are important in understanding the low temperature behaviour of this system, we have investigated the metal-to-insulator transitions of the cation-substituted Lithium Titanate Li_(1+x)Al_yTi_(2-x-y)O_4. In a one-electron picture, one would expect that the metal-to-insulator transition can be modelled very accurately by an electron density driven transition in a quantum site percolation model. We have studied these transitions, and find that quantitative predictions for the critical dopant concentration at which the metal-to-insulator sets in are in disagreement with those predicted experimentally; e.g., experimentally for the (x=0) LiAl_yTi_(2-y)O_4 compound, an Al concentration of y_c=0.33 produces a metal-to-insulator transition, whereas a quantum site percolation model identifies a lower bound of y_c > 0.8. One proposal that is consistent with this result is that strong correlations are ignored in a quantum site percolation (effectively Anderson) model of the transition, and thus y_c is grossly overestimated.