A Bottom-Up Approach towards Intelligent Magnetic Memory:

from Materials Design and Devices Property to Application Performance

Xiang (Shaun) Li, [email protected]

Postdoctoral Scholar, Electrical Engineering and Materials Science Engineering

Stanford University

May 14, 2020

The key to the recent success of machine learning lies in the power of computing hardware, a huge amount of data to learn from, and bio-inspired algorithms. However, the existing electronics hardware is failing to process and store the ever-growing learning data and algorithm weights in a fast, cheap, and energy-efficient way. The major bottleneck lies in the low speed and high energy cost incurred when data is transferred between memory storage and logic processing units. One promising approach to solving this bottleneck is a high-density on-chip cache memory. Magnetic memory technologies are one of the leading candidates. [1] However, research and development of new materials for high-performance magnetic memory across academia and industry remain a trial and error process, lacking frameworks that link material-level chemical design, device-level electrical properties, and application-level energy and delay metrics.

In this talk, I first establish a predictive framework to guide the complex materials design space explorations for magnetic memory with desired device properties. [2] This framework is realized by leveraging state-of-the-art nanoscale chemical and structural characterization techniques. Using this framework, I demonstrate one of the highest performing voltage-controlled magnetic memory devices. In the second part of this talk, I introduce a cell-level physics-based model that illustrates materials requirements for reaching application-level energy and delay performance. [3] This model suggests that two materials parameters are critical and enables a benchmarking of write energy and speed based on various materials reported in the literature. These two bottom-up frameworks show the power of bridging materials, devices, and applications to guide memory materials, devices, and circuits research in the future.

Reference:

[1] X. Li, et al., “Voltage-controlled magnetoelectric memory and logic devices,” MRS Bulletin, vol. 43, pp. 970-977, 2018.

[2] X. Li, et al., “Predictive Materials Design of Magnetic Random-Access Memory Based on Nanoscale Atomic Structure and Element Distribution,” Nano Letters, vol. 19, pp. 8621-8629, 2019.

[3] X. Li, et al., “Materials Requirements of High-Speed and Low-Power Spin-Orbit-Torque Magnetic Random-Access Memory,” IEEE Journal of the Electron Devices Society, (Early Access) 2020. DOI: 10.1109/JEDS.2020.2984610

Biography:

Xiang (Shaun) Li received his Ph.D. degree in Electrical and Computer Engineering from UCLA in 2018, and his B.S. in Physics degree from Peking University, China in 2013. He is now a postdoctoral scholar at Stanford University jointly with Electrical Engineering, and Materials Science and Engineering departments. He is interested in novel materials and devices for non-volatile memory and neuromorphic applications. He has authored magnetic devices chapter of the IEEE Industry Roadmap for Devices and Systems (IRDS) and published 30+ papers with 1000+ citations in Nature Communications, Nano Letters, IEEE, APL, MRS Bulletin, PRL, and PRB.

He also has extensive experience in entrepreneurship and technology commercialization. He participated in the Stanford Ignite Certificate Program in Innovation and Entrepreneurship at Stanford University Graduate School of Business in 2020. He acted as the Chief Technology Officer at Inston Inc., a startup spin-off from UCLA from 2018 to 2019. He was also a Technology Fellow at UCLA Technology Development Group from 2016 to 2017.

Time: May 14, 2020 10:00 AM Hong Kong
Registration URL

Click here is YouTube is not available.