In spin-orbit torque experiments (i.e. switching, ST-FMR, spin pumping, etc.) for a bilayer structure of nonmagnet/ferromagnet, I assume that different choice of the ferromagnet result in the difference in the torque efficiency (or effective spin Hall angle). How do experimentalists choose a proper ferromagnetic material depending on types of the experiment? I am especially curious about differences between metallic ferromagnet thin layers such as Ni, Py, Co, Fe, CoFeB, which are most widely used.

So far, I guess these factors could be important factors in the device

- magnetic anisotropy (whether it prefers in-plane or out-of-plane magnetization)

- interface ordering, which may affect spin-flip scattering and spin memory loss

- saturation magnetization

- chemistry, which results in the difference of the hybridization with the substrate nonmagnet

I guess this question would be quite obvious to experimentalists, but as a theoretician who is ignorant of experimental details, any answer would help me a lot.