Jerome LinBy Jared Sagoff • May 10, 2016
Argonne postdoctoral researcher Jerome Lin recently began his Argonne career at the Center for Nanoscale Materials, a U.S. Department of Energy Office of Science User Facility, where he pursues his interests in semiconductor devices and new materials, nanoelectronics and nonclassical computer architectures, among others. His current research goal is to build ultra-low-power computers using new materials, devices and system architectures.
Lin comes to Argonne from Massachusetts Institute of Technology (MIT), where he received the Jin-Au Kong Award for the Best Doctoral Thesis in Electrical Engineering at MIT. He also received the 2014 Roger A. Haken Best Student Paper Award for his presentation at the 2014 International Electron Devices Meeting.
Tell me a little about your background and about the work that was honored by these awards?
I graduated from MIT about a year ago. When I was in grad school, my research was on complementary metal-oxide semiconductor (CMOS) technology, which is the basis for integrated circuits, which in turn form the foundation of the most popular electronic devices that appear in computers, mobile phones and many other devices.
The transistor technology that I have been investigating is called MOSFET — for metal-oxide-semiconductor field-effect transistors. As we reach the physical limit of our current materials, mainly silicon, the research community is looking into how to incorporate new materials.
One material I spent a long time looking at while I was at MIT is called indium gallium arsenide, or InGaAs. Our goal is to use it for future generation CMOS. While it won't be a complete replacement for silicon, it's potentially a good complement for certain parts of the chips, especially for those that require high speed. No material that we know of will be able to replace silicon completely.
What are some of the advantages of this new technology?
One reason we are exploring this new material is to extend Moore's Law, which says that chips will roughly double in computing power every 18 months. At present, it's generally recognized that silicon is reaching its physical limits, and so people are exploring different types of material that have the potential to extend Moore's Law and do what silicon can't do in terms of power and speed. InGaAs is one of the top candidates.
The advantage of this new material is that it has high electron transport properties, which translates into the potential to have high speed at low power. Over the past decade, there has been tremendous research into this worldwide. The MIT group is one of the leaders in this technology. At MIT, I worked to develop new fabrication technologies that would lead to the demonstration of one of the highest-performance InGaAs MOSFETs.
Demonstrating high-performance devices in this material system and understanding the new device physics associated with CMOS applications are important. I have made a few contributions to these areas of research, and that's why my work was recognized by a few awards.
What does Argonne offer you?
By coming to Argonne, I am looking into a new direction. I had been working on MOSFET technologies for the past eight years or so -- not just for InGaAs transistors, but also for strained-silicon and steep-slope transistor technologies.
I wanted to study something beyond CMOS. That's why I contacted [Argonne Nanoscience and Technology Division Director] Supratik Guha, to talk about new types of technology that could be used for computing.
One major area that the two of us are interested in is called "neuromorphic computing." I'm still not really able to say a lot about what this technology will be, but there's a lot of research on it going on right now. At a very basic level, it involves looking at the way that the brain handles information processing and taking those basic principles and applying them to the design of new computing architectures.
Computer scientists have been looking at these architectures, which are really brand new ways of thinking about computing. And device and system engineers are developing new methodology to make the circuit perform. I am excited about what could be done in this direction.
Neuromorphic computing is not going to replace digital computing, but it has potential advantages in fields such as machine learning. Right now, digital computers require a lot of power and computational resources when dealing with machine learning problems.
In the first year at Argonne, what do you hope to learn or achieve?
I have an electrical engineering background, so what I'm good at is making small things; I design new device concepts and turn them into reality.
Supratik is a well-respected materials scientist. The advantage of working with him is that I can learn from him new knowledge in materials science, and we will investigate the potential of using new materials to make electronic devices. One of my plans is to learn from Supratik the properties of new materials and make new devices from them. And since both Supratik and I are new to Argonne, we will work together in setting up a new lab with state-of-the-art equipment for material and device research at Argonne.
We're moving into a new area in which the materials, devices and system architectures are still under exploration. There's a lot of possibility out there, but there's also a lot of uncertainty. It's very exciting, because the more unknown there is, the more room there is for innovation. I'm hoping that by diving into a new area, and by joining Argonne, I can expand my ability to innovate.
This new field requires skills like nanofabrication and device characterization, which I have had experience in. It also requires me to work hand in hand with experts in other fields, such as computer scientists. In the year to come, there will be a lot to learn and I really look forward to it.