SEMICONDUCTOR TECHNOLOGY Could silicon soon be a thing of the past?
Recent advances in semiconductor research have led to the emergence of compound semiconductors. They’re lighter, faster, smaller, and more efficient. But will they lead to the replacement of pure silicon in power electronics applications?
Power electronics has been ruled by a concept known as Moore’s Law for almost six decades: the observation of Intel co-founder Gordon Moore made that the number of transistors in a silicon-based integrated circuit had doubled roughly every year and would continue to do so indefinitely every two years. Although advancements in silicon-based power technologies has enabled us to develop electronic devices that Sir Ambrose Fleming himself would have thought impossible, the power electronics field is rapidly changing as engineers switch to devices based not on silicon chips — which are quickly reaching their theoretical limitations under Moore’s Law — but on new materials that can handle electricity more efficiently.
The emergence of post-silicon devices
So-called post-silicon devices based on materials like silicon carbide and gallium nitride are already in circulation, and they’re becoming more relevant at a time when silicon supply chains are severely restricted. These materials emerged in 2017 when Tesla faced a critical moment in its history. The EV maker had already released two successful car models but facing the need to become a major automaker and serve the mass market, it needed to make a cheaper, mass-market vehicle: the Tesla Model 3. When the Model 3 was released, it held a significant technical advantage over the competition: it used a relatively unknown material, silicon carbide (SiC). The Model 3 was not only revolutionary due to its status as one of the first luxury EVs accessible to the mass market, but also its ground-breaking power electronics which proved that electrified vehicles could work at scale.
"The next generation in power semiconductors will be driven by Silicon Carbide technology"
Wide bandgap semiconductors
SiC and other wide bandgap semiconductors, also known as “compound semiconductors”, like gallium nitride (GaN) are much more useful than standard silicon in power electronics applications since they can move power more efficiently, because more energy is required to switch the material between its two states — “ON” (i.e., conductive) and “OFF” (i.e., insulating).
Indeed, silicon carbide is one of the more advanced wide bandgap materials, having been in development as a transistor material for several decades now. In that time, researchers have also begun using “younger” wide bandgap materials like GaN, ultimately leading to three scientists being awarded the 2014 Nobel Prize in Physics for their use of GaN to create the world’s first bright blue LEDs that are now ubiquitous in devices like screen and lightbulbs.
Researchers have more recently began using GaN to improve power electronics. In just a few years, it has reached fruition in applications such as smaller, lighter, faster, and more efficient adapters for charging phones and computers. According to Jim Witham, the CEO of GaN Systems that supplies Apple with its GaN laptop chargers, “a typical charger that you buy for your computer is 90 % efficient… gallium nitride is 98 % efficient. You can cut power losses by four times.”
Does this mean the end for silicon?
”The advent of compound semiconductors is a game-changer that has the potential to be as transformational as the internet has been for communications,” said Stephen Doran, CEO of the UK’s Compound Semiconductor Applications Catapult in a July 2018 interview with TechRadar. Doran points to the fact that next-gen semiconductors like SiC and GaN have the potential to be as much as 100 times faster than standard silicon, and as a result could power the continued explosion of connected devices.
Current trends show that there’s truth to this assertion. The power gallium nitride (GaN) device market is currently growing at a compound annual growth rate (CAGR) of 59 % from US$126 million in 2021 to US$2 billion in 2027. This is according to recent estimates by market research and strategy consulting company Yole Développement in its annual report, ‘Power GAN 2022’, illustrating that more and more manufacturers are favoring next-gen materials over silicon for the benefits they deliver and the latter’s theoretical limitations.
As for what this means for silicon, nobody knows. There’s no single answer as to when so-called Moore’s Law will “end” or, in other words, when the demands of modern power electronics exceed what can be achieved with silicon transistors alone. Compound semiconductors alongside other innovations such as quantum computing, atomic computing, and cold computing are all likely to play a major role in the future of electronics — whether these will completely eradicate silicon, however, remains to be seen. Nobody knows what tomorrow’s electronics will look like.