NSCI Seminar: Spintronics as an Enabler for a New Paradigm in Computational Technology

Since the emergence of "Moore's Law" scaling for semiconductor transistors 50 years ago, it has been recognized that the exponential increase of transistor density would ultimately come to an end. While this end-point has been extended much farther than originally thought, the limits of scaling are indeed upon us, and improvements in computational speed and efficiency are now mainly being achieved through novel architectures and new physics, rather than through shrinking device size. Looking toward the future, new physics of device operation will play and even more important role. For this reason, spintronics emerges as a promising platform for future computational systems, since it takes advantage of very different physics of device operation that can enable both improved energy efficiency as well as novel computing paradigms. In this talk, I will describe recent advances in spintronic materials and devices, and will show how new device concepts can be realized based upon physical phenomena such as the spin Hall effect, exchange bias interactions and magnetoelectric effects. I will also compare and contrast various novel spintronic architectures both for Boolean and neuromorphic computing platforms. Finally, throughout this talk, I will provide an overview of the ongoing research in C-SPIN, a vertically integrated research center funded by SRC and DARPA, which has the mission of advancing fundamental knowledge of spintronic materials, devices and novel architectures.