To efficiently employ the spin transfer torque to manipulate magnetization, ideal materials have to be developed and to this end the dependence of the torques on the materials properties and device geometries needs to be understood. For this, advanced theoretical modeling is required. To rationally design multilayer materials stacks where large torques can be expected, the spin polarization will be calculated for instance for Heusler compounds. In this class of materials, 100% spin polarization can be attained by tuning the composition. An alternative are molecular magnets, where well-defined energy levels with high spin-polarization exist. Specialists apt for these calculations are theoreticians that use density functional theory, Green’s function and related approaches. To calculate the torques in these materials quantitatively, approaches such as Random Matrix Theory and Circuit Theory will be used. For thermal spin currents, the spin Seebeck coefficients will be similarly predicted and the size of the ballistic spin currents will also be estimated.

The most promising materials will by synthesized for instance using spark plasma sintering to fabricate sputter targets. These can then be used as sources for thin film deposition while also co-evaporation of materials can be used to directly synthesize thin films of the advanced materials. Growth conditions will be optimized and multilayer stacks will be grown, which form the basis of the spin torque pillar and racetrack devices. The selection of the multilayer stack materials will be done to optimize the spin-polarized transport and the resulting torques by band structure matching. And using for instance an innovative combination of perpendicularly magnetized polarizer layers and in-plane magnetized free layers, the current for switching can be reduced and the determination of the fundamental spin torque asymmetry coefficient that governs the switching is possible.