Spin wave computing using pre-recorded magnetization patterns

We propose a novel type of spin wave computing device, based on bilayer structure that includes “bias layer” made from hard magnetic material and “propagation with low damping, for example, yttrium garnet or permalloy. The bias layer maintains stable pre-recorded magnetization pattern, which generates field desired spatial dependence, in turn sets the equilibrium inside propagation layer. When an external source applies RF to layer, excited waves scatter magnetization's inhomogeneities resulting complex interference behavior. This scattering can be utilized perform variety mathematical operations including Vector-Matrix multiplication. dependence such patterns estimated via perturbation theory.