Students Invited to Help Shape the Future of Semiconductors

July 12, 2024

By Mike Peña

As AI weaves itself into more areas of life, scientists and engineers at UC Santa Cruz have come together to develop a better component that serves as the "neuron" in brain-inspired computing. The imperative to make such processing much more energy efficient, and faster, means that those who can design and build these next-generation devices will be in high demand on the job market.

That's why the project team invites college students across the Bay Area, including at UC Santa Cruz, to attend their July 26-27 workshop on campus. World-class scientists will discuss the work that needs to be done to drive innovation, and along with that, the opportunities that this project will create for students to learn the knowledge and skills needed to advance brain-inspired computing.

The team is focused on a fundamental electronic component called a memristor, a building block for circuits that retains memory of energy flow without power. They aim to show that a memristor device created at UC Santa Cruz that consists of atomically thin, two-dimensional semiconducting material combined with transition metal oxide thin film can revolutionize the field of neuromorphic computing, and computing more broadly.

While these devices have been studied beyond UC Santa Cruz, the memristors fabricated in the lab of assistant physics professor Aiming Yan are some of the first created by this team. The 2-D material makes these designs more reliable than traditional memristors, which are made of thicker material that makes them more prone to physical and performance degradation over time.

"This workshop is just as much about workforce development as it is about the future of semiconductors," said Yan, the project's principal investigator. "Being so close to Silicon Valley, we want to help students across the region realize that this is a promising area to pursue a career in."

This project also demonstrates UC Santa Cruz's effort to engage in semiconducting technologies that are prominent in Silicon Valley and in high demand worldwide, Yan added. By equipping students from throughout the region with the knowledge and skills to advance brain-inspired computing, Yan said this effort will showcase UC Santa Cruz as a source for the next wave of high-tech workers for industry leaders like Google, IBM and Intel.

Speaking at the workshop will be project team members and UC Santa Cruz professors Jason Eshraghian, Nobby Kobayashi, Ricardo Sanfelice, and Jairo Velasco, Jr. In addition to the lectures, the event will feature student-research presentations, panel discussions among faculty members and students about transferring from a community college to a four-year university, the graduate school admission process and grad-student life, as well as information on the semiconductor industry.

In August 2023, the National Science Foundation awarded UC Santa Cruz a nearly $300,000 grant to fund the project for two years. In their proposal, the team described the key role that memristors play in brain-inspired computing, as well as the potential for neuromorphic computing to transform how information is processed and stored.

The silicon technology in today’s computers segregates memory and data processing. So when information is processed, it must be pulled from memory and transferred to the central processing unit (CPU). The information must then be put back in storage, and it is that retrieval and return of information that increases both processing time and power consumption. Meanwhile, a memristor can both store and process information—akin to an actual neuron—significantly speeding up computing while lowering energy needs.

As we increasingly rely on digital devices to perform evermore complex tasks, a solution for faster and more energy efficient processing is critical. To illustrate the exponential energy requirements for neuromorphic computing, Yan said the human brain only needs about as much as a 60-watt light bulb to do its job.

“Memristor-based computers can be potentially 10,000 times lower in energy consumption compared to conventional computers using complementary metal-oxide-semiconductors,” said Yan, who focuses on atomic and nanoscale structural and property tailoring, and the study of functional materials. “However, the most significant improvement for our proposed device geometry is to combine the 2-D materials and transition metal oxides to make the memristors even more energy efficient and robust.”

This project seeks to advance memristor design by incorporating 2-D semiconductors, such as transition metal dichalcogenides—each layer just three atoms thick. In addition to leaving little room for degradation, a semiconductor that thin would allow many more transistors to be packed onto a computer chip.

This interdisciplinary project draws on faculty with deep expertise in physics, materials science, and electrical engineering, so students who participate are sure to have a holistic and high-quality experience, Yan said. She added that one of the project’s main goals is to engage students from underrepresented backgrounds.