Collective behavior of disordered magnetic systems

Abstract: Magnetic properties of disordered and frustrated magnets have been experimentally investigated. The magnetic systems in focus in this work are spin-glasses and frozen ferrofluids (suspensions of nano-sized magnetic particles).It has been experimentally shown that for a three dimensional Ising spin-glass an external field destroys the low-temperature phase. This result supports a real space model where the static and the dynamic magnetic properties are governed by droplet excitations. So-called memory effects have been investigated in ac-susceptibility, zero-field-cooled and field-cooled relaxation experiments. As is the case for aging phenomena, these results can also be interpreted using the same real space model.Properties of the frozen ferrofluids have been investigated by means of first and higher order harmonics of the ac-susceptibility, magnetic noise and dc-relaxation measurements. For non-interacting particles the magnetic response is compared with theoretical model predictions. By increasing the concentration of particles in the ferrofluid, the dipole-dipole interaction between particles subsequently increases. The magnetic dynamics have been investigated in a wide time window and at different temperatures for a colloid consisting of considerably interacting γ-Fe2O3 particles. Detailed information of the effects introduced by interaction have been extracted from comparisons with the corresponding dynamics of non-interacting particles. The observed aging property and the critical behavior close to a phase transition temperature suggest a low temperature phase which is spin-glass-like even for these systems. Critical exponents of a dipole-dipole interacting magnetic particle system are determined from a static scaling analysis on a concentrated sample of nearly monodispersed FeC particles.Monte Carlo simulations have been utilized to model both spin-glass systems and magnetic nano-particle systems. The advantage of using Monte Carlo simulations is that certain intrinsic properties, not easily accessible in real experiments, can be studied or varied in computer experiments.