Not Just Lithium for EVs: New Semiconductors

When I toured the Ford F-150 factory in Dearborn, Detroit about six years ago, the Just-In-Time inventory system — when a company receives goods as close as possible to when they are actually needed – had been in operation at Ford for about 50 years. A missing part that could hold up the line would be rushed in by helicopter, hang the expense.

In 2022, with COVID supply chain issues and chip inventory stress causing havoc for everyone from car makers to Apple, automakers have ditched Just-in-Time to become hoarders of inventory – Just-in-Case.

Change can come fairly rapidly. So, investors should know who’s who and what’s next in the critical realm of semiconductors. ETFs in the semiconductor space saw the largest inflows of funds throughout 2021. Investors should be aware of the new type of material being used in the home, in cars, in battery energy storage systems, and in solar too.

Silicon continues to suit as the semiconductor of choice for many computing chips. But in power applications, where circuits handle power flows, things are quickly changing. Non-silicon semiconductors are growing in use, especially in power applications like electric vehicles (EVs). 

The new world of EVs has a slightly different semiconductor supply chain. It’s not just silicon chips but new and different semiconductors. Silicon carbide and gallium nitride semiconductors are taking over some of silicon’s less competent areas.

Silicon carbide

Silicon carbide (SiC) is an alternate semiconductor coming into its own period of technological sophistication after decades of limited R&D, despite an understanding of its potential. It is now in use in power electronics and devices such as power transistors, known as SiC MOSFETs.

We don’t need a detailed explanation here, but it’s worth knowing these devices make inverters better in many circumstances. Inverters turn direct current (DC) to alternating current (AC) and vice versa. Inside the phone charger we plug into the wall is an inverter, turning the 240V mains supply to DC power to charge a battery. Another example: turning DC power from solar panels to AC power for your house.

But what’s wrong with pure silicon? Again, keeping out of materials science, SiC has properties which give it a wider bandgap. SiC semiconductors tolerate higher voltages, temperatures and frequencies far better than silicon, and therefore, operate at higher efficiencies.

By allowing devices to operate at higher voltages and frequencies, SiC inverters drive other components such as magnets and coils, within the same inverter, to be much smaller. That makes devices smaller, lighter, and saves on component costs.

What’s the downside? Making the SiC crystal wafer itself is much more expensive than silicon due to the ongoing development of the material, and it’s simply harder to make without defects. SiC wafers are also much smaller than silicon wafers, and growing wafer sizes is harder. It means SiC devices are more expensive, though overall device costs are saved in other components.

Applications

Silicon carbide MOSFETs have tremendously useful applications in those inverters. Bosch Group told me in a recent interview that switching from power supplies based on silicon to new power supplies that use SiC chips adds at least 6 per cent to the battery range for EVs.

Tesla was one of the first to use SiC inverters in its vehicles in 2017, with all its models now using the material. According to a study by Exawatt, 70 per cent of passenger battery electric vehicles will use SiC MOSFETs by 2030.

SiC is also coming to inverters in solar farms, delivering better efficiencies and weight savings for certain large inverters, though upfront costs may still be higher. Wind power has similar applicable uses.

In terms of where SiC might go next, Peter Friedrichs, the vice president of SiC at Infineon, told me that battery energy storage systems are next: from homes to commercial and industrial size.

SiC can improve efficiencies in inverters or even DC-DC converters, lifting them from current efficiency ratings of 98 per cent up to 99 per cent or higher.

While that figure doesn’t sound like much on paper, using SiC offers a 50 per cent improvement in system efficiency, as Infineon pointed out. That again allows for smaller batteries in homes (or, alternately, more output at the same size), plus faster charging.

Market leaders in SiC chips include Infineon Technologies, and STMicroelectronics, which notably supplies Tesla with SiC MOSFETS.

Both will continue to seek more SiC wafer manufacturers as demand from EVs, and new applications in solar, continue to grow.

STM has bet on SiC in a significant way, building a “megafactory” in Europe with a US$3.6B capex bill attached. It estimates its annual revenue from SiC products will hit $1bn by 2024.

GaN

Gallium nitride, known as GaN, is the other wide bandgap semiconductor of interest in power electronics. It has similar properties but is useful at different configurations of power (lower than SiC) and operating frequencies (much higher than SiC).

If you’ve bought a new phone charger in the past year, and paid a little more for something better, you probably bought a “GaN charger”. This uses gallium nitride in a similar way to silicon carbide to save on size, weight, and heat. It’s why a new GaN charger is half the size of an older phone charger.

The key for GaN chips is that they’re great at high frequencies at lower power outputs, which is perfect for small inverters used in chargers for smartphones, laptops, tablets, and so on. At higher power outputs, silicon and SiC are better suited, but bets are being made for GaN chips to also be used in EV chargers, just like SiC, within a few years.

The battle between the two in some applications is interesting: GaN is a newer material with less engineering know-how, but SiC is also hard to produce.

Infineon and ST also work on GaN chips, while Navitas, founded in 2014 and which went public via a SPAC is betting almost entirely on GaN.

Silicon’s Last Gasp?

Silicon, though, won’t be thrown out overnight in the realm of power electronics.

Incumbents will keep improving, according to the great technology principle dubbed the “sailing-ship effect”. This effect, proven mathematically, is where improvements to an incumbent technology, ie. sailing ships, are intentionally sought as a new competing technology emerges ie. the steam engine and steam ships.

Silicon power devices may well follow the same route. But as a bet on the growth of renewables and EVs, silicon carbide itself stands out as an interesting way to bet on batteries, without needing to bet on lithium.