In spintronics, the electron’s spin degree of freedom, rather than its charge, is employed to process information. The efficient generation of large spin currents stands as a key requirement for future spintronics devices and applications. Spin-polarized electron currents can be used to manipulate the logic state of nanoscale magnetic memory devices known as spin valves (SVs, see figure above). These currents are orders of magnitude lower than those required for magnetic field–based control. Spin valves based on spin-transfer torque are on the verge of commercialization as the basic units of ultra-fast, low-power consumption MRAM memories, less than a decade from their discovery. In these current-based SVs, the high spin-torque efficiency is achieved by using magnetic tunnel junctions (MTJs) to separate the ferromagnetic and non-magnetic elements in the device and to enhance spin accumulation. This makes them challenging to operate reliably since not only it is complicated to manufacture large-scale memories but they also require the injection of a large enough spin current through the MTJs in order to induce magnetic switching. Continuous electrical stress at the MTJ eventually leads to device failure. To circumvent these problems, several approaches to generate pure spin currents (with no net charge current) have been proposed and are being widely investigated as the basis for the next generation of reliable and energy efficient magnetic memory devices: non-local spin injection, spin-Hall effect (SHE), and dynamical spin pumping.
In this project we investigate the advantage by dynamical spin pumping of producing pure spin currents over large (mesoscopic) areas at ferromagnetic/non-magnetic (FM/NM) interfaces and of being insensitive to a potential impedance mismatch at the FM/NM interface. Several FM/NM interfaces are being investigated, including superconducting (Nb) and actinide (U) elements.
Collaborators in this project:
Kevin Coffey (UCF, Orlando, USA)
Casey Miller (Rochester Institute of Technology, USA)