April 5, 2023
Engineering large spin-orbit torque in silicides for low-power spintronics applications
Magnetic memory has been a strong candidate for next-generation low power memory solutions. In fact, magnetic memory has been commercialized in both standalone and embedded memory markets for the past few years. However, there are still several challenges in the current spin-transfer torque technology in magnetic memory applications. As a result, spin-orbit torque emerges as a candidate to replace spin-transfer torque due to advantages such as separate read-write current path, faster switching speed and larger material design space. To this regard, efforts in studying spin-orbit torque have been geared toward addressing several issues including material compatibility with silicon platform, trade-off between conductivity and spin-torque efficiency, and engineering field-free switching designs. In this talk, I will discuss how we propose to use silicides, a decades-old CMOS technology, for efficient spin current generation to manipulate magnetizations. In both amorphous iron silicide and cobalt silicide, we discovered a strong spin current generation efficiency that rivals the largest efficiency values discovered in topological insulators. Moreover, we discovered a clear Fe (Co) concentration dependence of spin current generation efficiency in both systems that can potentially be signatures of Fermi level tuning in amorphous materials for manipulating spin current generation. Finally, we show that the scaling trend between conductivity and spin current generation efficiency in both silicides are opposite to the conventional scaling trend. In the conventional scaling trend, the efficiency increases as the conductivity decreases. This is a fundamental bottleneck to adopt spin-orbit torque technology for magnetic memory applications. In silicides, the trend is opposite with an increasing efficiency as the conductivity decreases. This novel trend can lift the conventional trade-off and further enable the adoption of spin-orbit torque in magnetic memory applications. This work presents an efficient way of generating spin current in a CMOS compatible material system – silicides – for energy efficient computing.
Jason Cheng-Hsiang Hsu is currently a PhD candidate in EECS at Berkeley working with Prof. Sayeef Salahuddin on studying energy efficient magnetic switching in novel materials and device structures. His research interest sits at the intersection of materials and device applications for energy efficient and novel computing. He obtained his bachelor's in electrical engineering with Honors in Research from the University of Wisconsin-Madison in 2017 where he was a Hilldale Undergraduate Fellow. He is the recipient of the Best Student Presentation Award in the 2022 Magnetism and Magnetic Materials conference, 2022 Demetri Angelakos Memorial Achievement Award and the 2022 APS Stanford Ovshinsky Student Travel Award. He has held an internship position at Western Digital working on plasmonic devices for heat-assisted magnetic recording technology.