All-optical pump-probe experiments have been used to explore magnetization dynam-ics on the sub-nanosecond time scale. This technique is advantageous over the conventional FMR because the spin in motion could be observed directly in the time domain to enable a lucid vectorial analysis. Utilizing the ultrafast femtosecond laser, we have investigated the spin dynamics of perpendicularly-magnetized materials and the energy dissipation process-es extensively.
The damping mechanism will be categorized according to the location where the open channels of relaxation appear. The dissertation adjusts the focus on the non-local type of spin emission the magnetic heterostructures commonly undergo. We would like to propose a feasible manipulation of magnetic damping by the proper choice of neighbors.
The extrinsic damping of ferromagnetic CoFeB films sandwiched by two asymmetric adjacent layers was explored. The nearby layers have entirely-contrasting contributions to the effective Gilbert damping in terms of the non-local process, however the exchange of their relative positions to the ferromagnet does not show any difference in our system. This suggests that the influences from the top and bottom layers are indistinguishable.
Effect of the heat treatment and capping layers in CoPd multilayer system will be dis-cussed. After-annealing turns out to strengthen the magnetic property considerably. En-hanced Gilbert damping by Ta cap layer (~50%) explains that the “pumping” of spins onto the adjacent normal metal layers could diminish the magnetization motion.
CoPt multilayer with Cu spacer inserted has been investigated. We observed the in-crease of damping up to 100%, which is attributable to various probable reasons. Consider-ing that the spin favorable Cu spacer would preserve the spin motion for a long time, spin diffusion current should not add the contribution to the non-local damping. We believe that the possible mixing of Cu and Pt could expand the Pt length scale “effectively” resulting in the heavy enhancement of damping.