Next-Generation Computing using Spin-Based Materials

Who: Jean Anne C. Incorvia, Stanford University
Friday, October 28, 2016 - 12:00pm to 1:30pm
Hamerschlag Hall D210

We are at a time where the electronics industry is feeling pressure from two sides on the small scale we are facing the fundamental physical limits of silicon, and on the large scare we are facing new, abundant-data and distributed-data applications, such as for the internet of things. The future of computing will require both more energy efficient electronics and more big-data driven, application-specific designs.

Magnetic devices are a promising candidate for future electronics, due to their low-voltage operation nonvolatility, and low thermal budget, which can open up new energy efficient, normally-off, memory-in-computing, 3D monolithic architectures. Magnetic materials are one of the few materials systems that can me more energy efficient than silicon transistors for memory and logic, and have been shown to me more energy efficient and faster than other emerging resistive memories. Thus, they are a very important class of materials to explore for beyond-CMOS devices and systems.

I will present experimental results using three-terminal spin switches to build magnetic logic devices and circuits, and show they satisfy the requirements for beyond-CMOS devices. I will show our work on extending these spin switches to building hybrid magnetic-CMOS circuits to leverage the magnetic nonvolatility at a circuit level. I will also show our work on measuring the spin and valley Hall effect in 2D transition metal chalcogenide materials for future versions of theses switches.