IBM Research – Almaden, San Jose, CA 95120
Since magnetic domain wall racetrack memory was proposed 15 years ago, remarkable progress have been made (1) to pack more domain walls in a given footprint (density), (2) to inject them reliably at lower power (write), (3) to shift them back and forth in lockstep way (access), (4) to move them more efficiently (access), and (5) to detect them more effectively (readout). Especially a series of recent breakthroughs such as discovery of spin-orbit torques and exchange coupling torques in chiral domain walls have made the racetrack memory more promising for commercial products in the near future. In addition, the application to three-terminal one-bit devices has emerged to provide a new opportunity for racetrack memory thus attracting enormous attention in magnetic communities and industries since the technology is much simpler, more viable and useful for e.g. development of ultrafast high endurance memory and neuromorphic devices compared to the original proposal.
In this talk, I will review all of these updating the audience and provide its outlook/perspective in the end.
Dr. See-Hun Yang is a Research Staff Member working for IBM Research – Almaden, San Jose, CA in the US. He is leading projects on devices research and development based on current driven manipulation of magnetic nanostructures such as ultrafast one-bit racetrack memory and magnetic neuromorphic devices for deep machine learning and neural networks. He has published more than 3 patents and 120 papers, and given more than 30 invited/plenary talks. Before joining IBM, he worked for Advanced Light Source, Lawrence Berkeley National Laboratory as a postdoctoral research fellow. He received Ph.D., M.S., and B.S. degrees from department of physics at Seoul National University in South Korea. He received IBM Research Division Award for the Discovery of Giant Tunneling Magnetoresistance in MgO-based Magnetic Tunnel Junctions.
San Jose, California