Find in Library

Spin pumping experiments using ferromagnetic metals have reported highly efficient spin–charge conversion in Bi and at the Bi/Ag interface, possibly due to the inverse Rashba–Edelstein effect. However, longitudinal spin Seebeck effect experiments using the yttrium iron garnet ferrimagnetic insulator in Bi films and Bi/Ag bilayers do not show evidence of appreciable spin-to-charge conversion except the large Nernst signal inherent to Bi. These contrasting conclusions highlight the differences between magnetic metals and magnetic insulators as spin current injectors. Only the detected voltages that adhere to the inverse spin Hall effect of jC = (2e/ℏ)θSH jS × σ are due to spin currents.