10 years from now, CMOS will seem as old-fashioned as vacuum tubes
By Marc Heyns
I’m very optimistic about the continuation of Moore’s Law. But in saying that, I’m speaking about Moore’s Law purely as an economic law. I believe we’ll be able to offer increasing amounts of functionality at lower and lower costs. And technological innovations as well as advances in design and application will be crucial in realizing this.
But I don’t believe a new technology will pop up and immediately replace CMOS. Compare it to the transistor and the vacuum tube. They co-existed for a long time, and each had their specific applications. The new ‘beyond CMOS’ technology will be no different. Maybe it already exists in a certain field, but we don’t ‘see’ it yet. One thing is certain; this new technology will allow us to build applications that use less energy — because energy is currently the bottleneck. We can’t even imagine what these new applications could be right now. Once they are here though, they’ll seem obvious and irreplaceable. The same thing happened with Sony’s Walkman. Back then, everyone wondered what the point was in carrying music with you wherever you go. Nowadays it’s difficult to find someone without at least a few of their favorite tunes on their smartphone or iPod.
It’s obviously impossible to predict which applications we’ll use 10 to 20 years from now, if we can’t even imagine their existence yet. But it seems likely that a lot of the ‘beyond-CMOS’ technology applications will be in health and entertainment. Affordable public healthcare is a difficult issue for any society to deal with, especially in developing countries. As such, the demand for innovation is very real. Consider biosensors. Oftentimes, their production processes are incompatible with CMOS. So why not flip the script and start from biosensors and develop electronic technology that fits?
You can already sense the arrival of the new technology through rapid succession of many important breakthroughs in physics and materials. The massive progress in manipulating electron spin and all kinds of magnetic properties of materials is an example. Or maybe the host of new materials will play an important role—like 2D materials, topological insulators, multi-layered materials with novel interactions at the interfaces, etc. One thing is certain: At some point someone will be clever enough to fuse new materials and physical concepts into a device that has something new to offer, and which will cause the formation of a large, new industry, much like how the rise of quantum mechanics and semiconductor physics led to CMOS transistors and today’s entire electronics industry.
Unfortunately, economics – rather than technology – might put the brakes on the speed of this development. Nowadays, the costs to develop and produce new technology nodes are staggering. This investment may not be feasible for a revolutionary new technology. As such, I believe the ‘beyond CMOS’ technology will have to be developed inside a niche market, where it can offer a unique functionality and become economically viable. The technology can then mature and reach out to a wider range of applications, gaining in importance alongside CMOS. Within 15 years, this new technology will be widely adopted, and CMOS will be sent to a museum. At least, that’s what my crystal ball tells me.
The person who invents the ‘beyond CMOS’ device and applications will have to be a real artist. They will need to know design, technology and applications because true innovation will only come from fusing these three, right from the start. Only a veritable Leonardo Da Vinci will be able to achieve a truly revolutionary breakthrough.
I often tell my students that I’m jealous of them, because they live in such interesting times. The challenges of my day might not have been any smaller, but the path we needed to take was more or less fixed. Moore’s Law dictated that we follow the road of downscaling transistors. If you compare a transistor from 30 years ago and a modern finFET, you’ll see essentially the same device. Nowadays, however, there is room for revolutionary innovation. Look 30 years into the future, and you’ll see a brand new world.
Prof. Marc Heyns received the M.S. degree in Applied Sciences (Electronics) in 1979 and the Ph.D. degree in 1986 from the Katholieke Universiteit Leuven, Belgium. In January 1986 he joined imec where he became Department Director and Program Director responsible for a research group working on ultraclean processing technology, advanced high-k gate stacks, metal gates, epitaxial deposition of materials, environmentally benign processing and novel high-mobility substrate materials. As program director of the “Explore” program he was responsible for performing exploratory research on nanotechnology, novel materials and devices for ultimate CMOS technology and novel memory concepts. He became an imec Fellow in 2001 and a Professor at the Katholieke Universiteit Leuven at the Department of Metallurgy and Materials Engineering in 2005. He has authored or co-authored more than 350 publications in scientific peer-reviewed journals, more than 700 contributions at scientific conferences, including more than 80 invited presentations, has edited or contributed to various books and holds more than 30 patents.