Electric cars are all vying for SiC...but why?

It was not long ago that Renault Group and STMicroelectronics announced they had reached a cooperative agreement on the supply of SiC (silicon carbide) and GaN (gallium nitride) products for new-energy vehicles from 2026 to 2030. It is no coincidence, then, that Vitesco Technologies (Germany) also received an order for 800V SiC inverters from the Hyundai Motor Company for hundreds of millions of Euros earlier this year; JAC Motors and Bosch also signed a strategic agreement for cooperation in the field of SiC inverters.

In the SiC field, the recent actions of other leading electric vehicle companies and new vehicle manufacturers have also been attracting attention. One example is Tesla, who released a new model Model S Plaid using an SiC inverter; While launching the "BYD Han," its first model with SiC technology, BYD Company also announced that by 2023 it will achieve full replacement of silicon (Si)-based IGBTs with SiC automotive power semiconductor devices; NIO has also expressed that it will adopt an electric drive system based on SiC technology on its new ET7 models for 2022.

Various signs indicate that in the next few years, SiC power devices will usher in a small market surge in the field of electric vehicles. Therefore, everyone is planning ahead and “making pairs" in advance to bolster their supply chains.

Why use SiC?

In fact, considering the cost of the electric drive inverter in an electric vehicle, if SiC power devices are used to replace mature Si-based IGBTs, the cost of a single vehicle will increase by US$200 to 300. That being the case, why are so many people willing to spend more money to adopt this more "expensive" plan? The answer starts with the characteristics of the SiC device itself.

In the field of power electronics, the power device responsible for switch control is the key to performance. For a long time, Si materials have occupied a dominant position in the field, but along with increasing power density, higher levels of switching speed (frequency), and increasingly demanding power requirements in applications, the performance of Si devices is being "squeezed" closer to its theoretical limits, so people began to look into materials, searching for new semiconductor materials that can replace Si. As a result, SiC and GaN, two materials used in wide-bandgap (WBG) semiconductors (also known as third-generation semiconductors) have gradually entered people’s field of vision. Of these materials, SiC has many unparalleled advantages in semiconductor devices of power from 650V to 3.3kV.

Figure 1: Performance and application range of devices of various semiconductor materials (Image source: Infineon)

As shown in Figure 2, the wide-bandgap (WBG) of SiC is 3 times that of Si, its dielectric breakdown field strength (critical field strength) is nearly 10 times that of Si, its thermal conductivity is 3 times that of Si, and its saturated electron mobility is 2 times that of Si.... The application of these characteristics in power devices means:

• High WBG: The greater the WBG, the greater the critical breakdown voltage, the more suitable for high-voltage and high-power applications.
• High saturated electron mobility: The higher the value, the faster the switching speed of the device, which makes the drive power required for high-frequency operation under high voltage smaller with lower energy loss. In addition, peripheral devices of a smaller size are allowed to be used in high frequency circuits, which also contributes to system miniaturization.
• High thermal conductivity: Additional cooling systems can be avoided, which is conducive to the optimization of costs and dimensions.
• Small drain-source on resistance per unit area: Can effectively reduce loss.

Figure 2: Comparison of key characteristics of various semiconductor materials (Image source: Infineon)

Specific to automotive applications, some analyses show that replacing Si-based devices with SiC devices in electric drive inverters can reduce the energy efficiency loss of the device-level driver by 80%. According to estimation by Cree, the use of SiC power devices in electric vehicle inverters can reduce vehicle power consumption by 5%-10%. Taking into account overall considerations, although the cost of inverter modules will increase, battery costs and heat dissipation costs, as well as the cost of space usage, will be significantly reduced—the cost for an entire vehicle can be reduced by US$2,000. In addition to inverters, SiC power devices can also be used in many aspects, such as on-board chargers (OBC) and power conversion systems (DC/DC) of electric vehicles. It really is no wonder that everyone is rushing to adopt SiC.

Tesla's success story

It is true that everyone has long known about the performance advantages of SiC devices, but if these discussions were to remain at the theoretical level with no real-world success cases, then people would inevitably be hesitant when making technical decisions. Therefore, as to the reason why many car companies have been decisive in "embracing" SiC today, in addition to technological progress in SiC in recent years, Tesla's demonstrative role cannot be underestimated.

In the use of SiC power devices in electric vehicles, Tesla is likely the first car company to "take the plunge." In 2018, Tesla adopted the 650V SiC MOSFET introduced by STMicroelectronics in the inverter on its Model 3. It is said that compared to the earlier Model X models that use Si-based IGBTs, this move provides a boost in the efficiency of the inverter equal to 5%–8%, which is indispensable for improving the range of the vehicle. Then, on the Model Y launched in 2020, Tesla also adopted SiC MOSFETs in the rear-wheel drive of its power module. In addition to the Model S Plaid, Tesla currently has 3 models that use SiC technology. Of the three, with superior high-voltage, high-temperature, and high-frequency performance provided by SiC MOSFETs in the electric drive inverters, the Model S Plaid only needs 2.1 seconds to accelerate to 100 kph, earning it the reputation as the world's fastest accelerating mass-produced car. Such a "moniker" has undoubtedly become the best endorsement of SiC.

With the maturity of SiC products and technologies, expansion of its application in the field of electric vehicles is not only reflected in the expansion of the penetration range, but also in the deepening of application methods. In early electric drive inverters for new energy vehicles, a mixed architecture of Si-based IGBTs and SiC-SBDs was generally adopted, but that architecture is now evolving into using pure SiC inverters. In 2017, Rohm's pure SiC power module helped the VENTURI team to create a new inverter whose size has been reduced by 43% and weight by 6kg. Such a successful case makes the future of pure SiC inverters very promising.

Will new trends be lacking?

According to HIS Markit's forecast, the market size of SiC power devices is expected to exceed US$10 billion by 2027, with a compound annual growth from 2018 to 2027 of nearly 40%! And the new energy vehicle market is the most important driving force.

However, the increase in demand will also bring about some concern, that is, "whether the outbreak of demand will cause a shortage of supply," especially considering how in these past few years the automotive electronics field has suffered from "core shortage," the psychological shadow of which still lingering. Understandably, concerns have progressed.

From the current point of view, the main factors restricting the rapid expansion of SiC device production capacity include:

• SiC still has a hard time competing with Si in the preparation of basic materials such as substrate wafers and epitaxial wafers. For example, substrate wafers are mostly 4 and 6 inches (while the mainstream process of Si devices is 8 and 12 inches); Gas phase epitaxy rate is low, and liquid phase epitaxy production is low...Before these technical problems find a breakthrough solution, production capacity will certainly be limited.
• From the SiC device manufacturing process, forming a good ohmic contact in the fabrication of electrodes is still a difficult point.
• On the layout of the SiC industrial chain, the key process technologies of the past are in the hands of a few companies. The entire market is small, and it is far from forming a large-scale standardized division of labor like the Si-based process.

The above-mentioned bottlenecks will throttle the rapid ramp-up of production capacity and the reduction of costs. Looking at SiC substrate wafers as an example, it may be seen that the current cost of SiC is 4 to 5 times that of Si, and it is expected that the price will gradually drop to about twice that of Si over the next 3-5 years. In this process, short-term production capacity and short supply may be unavoidable.

Fortunately, good expectations for the market’s development have inspired confidence. We can see that the industry’s investment in increasing production capacity is also increasing. For example, STMicroelectronics’ acquisition of Norstel and Infineon’s acquisition of Siltectra, an emerging company in the field of SiC wafer dicing, etc.

Moving forward, what is a reasonable prediction for the development of SiC devices? The previous point of view of a person in the industry was more objective. We may as well quote it here—IGBT has been developed for a total of 30 years, from 1990 to the present, and has gone through 7 generations of technology. The final cost has been reduced to one-fifth of the what it used to be. The development of SiC from an emerging technology to a general-purpose technology will similarly be a very long process, and SIC technology will also require time for technical polishing.

Therefore, for SiC, on the one hand, we must actively follow this big trend; on the other hand, "patience" is also a virtue in this respect. Once one can perceive the rhythm, the entire process of technological upgrading will be more seamless and smooth.