Dependence of the giant magnetoresistance on the concentration of magnetic particles in granular composites

Abstract

We study the dependence of giant magnetoresistance (GMR) on the volume concentration of magnetic particles in a magnetic granular composite via a Monte Carlo method and by modeling the composite as a random resistor network. We assume the nanosized magnetic particles are spherical in shape and are randomly distributed in a square or cubic lattice. The uniaxial anisotropy of the particles and the classical dipolar interaction among the particles are taken into account. By considering the difference in electron scatterings for spin-up and spin-down conduction electrons at the magnetic and nonmagnetic interface, and the scatterings within the magnetic regions and the nonmagnetic host medium in the composite, the value of GMR is found to depend sensitively on the spatial distribution of the particles, the magnetic states of the particles, and the densities of the spin polarized conduction electrons. There is an optimum concentration (about 25% in two-dimensional and 30% in three-dimensional cases) of magnetic particles at which the magnetoresistance shows a maximum. This phenomenon was also observed in experiments.