Spins for a new computing era
This event is in the past.
Join Zoom Meeting
Meeting ID: 957 8934 1303
ABSTRACT: Cognitive computing will potentially redefine everyday life, changing how individuals perform their jobs, interact with others, and make decisions. Nonvolatile memories hold the key to solve the overwhelming energy demand for such computing and ensure intelligent systems for sustainable future. Magnetic memory, with electron spin as the information token, is one of the most promising nonvolatile technologies for next generation computers. I will present a magnetic device for future neuromorphic computing. First, the experimental demonstration of spin orbit torque induced magnetic devices will be shown as the building blocks (i.e., synapses and neurons) for in-memory computing. The synaptic device shows the most important functionality – linear output resistance, and the neurons provide programmable nonlinearity, unlike any other non-volatile memories. The scaling of these devices can potentially achieve energy consumption comparable to biological synapses. Second, I will show the use of materials with antiferromagnetic interaction as domain wall conduit and/or for spin orbit torque generator can further improve the energy-efficiency of the neuromorphic building blocks. Antiferromagnetic materials also Antiferromagnetic materials are also efficient in generating spin-orbit torque due to spin Hall effect. In the final part of my talk, I will show a direct method for measuring spin Hall effect in antiferromagnetic materials using X-ray magnetic circular dichroism - photoemission electron microscopy. The devices and materials developed in this work extend in applications beyond the examples provided here, introducing versatile platforms for using electron spin in other microelectronic applications.
 S. Siddiqui et al., Nano Lett. 20, 2, 1033 (2020).
 S. Siddiqui et al., Phys. Rev. Lett., 121, 057701 (2018).