Advances in silicon carbide and gallium nitride technologies over the past several decades have been characterized by growth, increased industry acceptance and the promise of billions of dollars in revenue. The first commercial SiC device appeared in 2001 in the form of Germany's Infineon's Schottky diode , and rapid growth ensued, with the industry expected to exceed $4 billion by 2026.
GaN first shocked the industry in 2010, when US-based Efficient Power Conversion (EPC) introduced its ultra-fast switching transistor. Market adoption is not yet on par with SiC, but power GaN revenue could reach over $1 billion by 2026.
The secret to the future success of each technology on the market lies in electric vehicles and hybrid electric vehicles. For SiC, the EV/HEV market is indeed the sweet spot; at least 60% of the over $3 billion market is expected to come from this industry.
Tesla launched the SiC power device market in 2017 when it became the first automaker to use SiC MOSFETs in a car (Model 3). This device comes from STMicroelectronics and integrates an in-house main inverter design. Other automakers quickly followed suit, including Hyundai, BYD, NIO, General Motors and others.
China’s Geely Auto has announced that it is working with Rohm Japan to develop SiC-based traction inverters for its electric vehicles. Meanwhile, automaker and semiconductor maker BYD has been developing SiC modules for its entire range of electric vehicles.
Last year, Chinese electric bus manufacturer Yutong revealed that it would use Xingneng China-made silicon carbide power modules in the bus powertrain. The modules use Wolfspeed's SiC devices.
Hyundai Motor will integrate Infineon SiC-based power modules for 800V battery platforms into electric vehicles. In Japan, Toyota is using Denso's SiC boost power modules in its Mirai fuel cell electric vehicle. General Motors has contracted Wolfspeed to supply SiC for its EV power electronics.
European carmakers have been slower to adopt SiC, but change is happening. In June, Renault and STMicroelectronics teamed up to develop SiC and GaN devices for EVs and HEVs. Importantly for Wolfspeed, Infineon, STMicroelectronics, Rohm and onsemi, automotive OEMs also prefer to purchase wafers and equipment from multiple sources to ensure reliable supply. Considering that China and other countries are pouring huge amounts of money into the SiC supply chain, sales will only continue to rise.
In the process, the thorny issue of cost is also being resolved. At the component level, silicon IGBTs are significantly cheaper than their SiC counterparts and are not going away from power applications anytime soon. But tier one manufacturers and OEMs say implementing high-power-density SiC in an inverter design, for example, can reduce system-level costs because the design will require fewer components, saving space and weight.
But how does this affect GaN? This wide-bandgap semiconductor has yet to see the success of SiC in electric vehicles. But due to its high frequency operation and high efficiency, OEMs either have a strong interest in the technology or are developing development plans.
GaN power devices can already be found in low-volume, high-end photovoltaic inverters and are increasingly used in fast chargers for a range of mobile devices, including smartphones. In fact, Ireland’s Navitas Semiconductor, the US’s Power Integrations and China’s Innoscience are all making GaN power ICs for the burgeoning fast charger market.
Given this activity, GaN power device revenue is expected to reach approximately $100 million in 2021.But that number is expected to grow to $1 billion by 2026 as GaN device suppliers seek to enter other markets to increase production. The EV/HEV market is the first to be affected.
GaN in electric vehicles is still in its early stages. Many power GaN players have developed and certified 650V GaN devices for EV/HEV on-board chargers and DC/DC conversion, and have established numerous partnerships with automotive companies.
For example, Canada-based GaN Systems supplies its devices for on-board chargers to U.S. EV startup Canoo and is also working with Canada-based EV motor driver supplier FTEX to integrate 650-V GaN power devices into electric scooter systems. Meanwhile, California-based Transphorm has partnered with automotive supplier Marelli to provide onboard charging and DC/DC conversion equipment.
STMicroelectronics is expected to provide devices that have not yet been approved for automotive certification for Renault electric vehicle applications. EPC, which currently supplies automotive-grade low-voltage GaN devices, is working with France-based Brightloop to develop affordable power converters for off-highway and commercial vehicles. Last year, Texas Instruments (TI) also qualified its 650-V GaN device for automotive applications.
As the on-board charger and DC/DC segments gain momentum, the billion-dollar question for GaN is, quite literally: Can the technology become the main inverter for electric vehicle powertrains, achieving astonishing sales comparable to SiC technology? Early industry developments show this is possible.
In February 2020, the Netherlands’ Nexperia collaborated with British consultant Ricardo to develop a GaN-based EV inverter design. Immediately following, Israel’s VisIC Technologies teamed up with German automotive supplier ZF to develop GaN semiconductors for 400-V powertrain applications.
Then, in September, GaN Systems signed a $100 million deal with BMW to provide manufacturing capabilities for GaN power devices for the German automaker’s electric vehicles, a strong demonstration of the importance OEMs place on GaN.
Another really important development is that through the merger with special purpose acquisition company Live Oak Acquisition, Navitas will become a public company with a market capitalization of $1.04 billion. The GaN power IC maker recently announced that it will supply devices for on-board chargers and DC/DC converters to Brusa HyPower, headquartered in SwitzerlandHT4. As a public company, it intends to focus on product development for EV/HEV and other markets.
In addition to these deals, partnerships and mergers, early work on GaN modules shows that compound semiconductors are following in the footsteps of SiC and industry players are preparing for wider industry consolidation. GaN Systems, for example, is offering power evaluation module kits to design engineers, while Transphorm has been working with Fujitsu General Electronics to develop GaN modules for industrial and automotive applications.
So what’s next for SiC and GaN? With power SiC device makers anticipating the electric vehicle market to reach billions, will GaN achieve the same success? Widespread adoption of GaN in powertrain inverters by OEMs will fundamentally impact market forecasts. But for now, we can only wait and see.
Unleashing the Economic-Driven Potential of Compound Semiconductors
Advances in silicon-based semiconductors and computing technology have transformed our world in a short period of time. Nonetheless, we have now reached the limits of where silicon alone can take us, and innovative approaches are needed to achieve the necessary advancements in our digital lives.
Our collective experience with COVID-19 leaves no doubt that connectivity is no longer optional. Fully aware of our individual role in this emerging, interconnected society requires that we participate in increasingly powerful and intertwined networks through our smart devices.
New technologies from the Internet of Things, Industry 4.0 and 5G/6G to electric/autonomous vehicles and augmented reality place unprecedented demands on existing semiconductors. Rising performance requirements have left traditional silicon semiconductors unable to meet the challenges.
So, what is the solution? There is a way forward at the level of semiconductor materials itself: we have to look beyond silicon. And there is a complementary set of semiconductor materials that could provide opportunities to go where silicon cannot.
Compound semiconductors refer to compounds formed from two or more elements, including materials such as gallium nitride, gallium arsenide, and indium phosphide. These semiconductors already power many of the technologies our connected world relies on. At the same time, their adoption is accelerating rapidly (growing twice as fast as silicon), and they are becoming mainstream in power electronics, sensing, connectivity and advanced display applications.
But the growth story doesn’t stop there. To enable the next wave of innovation, engineers are combining compound semiconductors with silicon to deliver a new generation of products that will be highly disruptive in delivering performance attributes at a scale that will enable mass adoption of compound semiconductors.
The growth of compound semiconductors is achieved through epitaxial growth, epitaxial growth is the basic production process used to build atomically engineered structures from which chip companies build electronic and optoelectronic devices. By selecting the right compounds and precisely controlling how the materials are deposited on the substrate, epiwafer manufacturers have a unique materials "tool set" that allows them to customize their wafers to enable a variety of device types, from power amplifiers to sensor lasers. The level of customization enabled by epitaxy is unparalleled; therefore, it is a key component of the compound semiconductor ecosystem.
The success of the silicon industry and its manufacturing plants is built on increasingly advanced photolithography, mastering successive process nodes to produce ever smaller transistors technologies - 5nm, 2nm, etc. Epitaxy is to compound semiconductors what photolithography is to silicon. Therefore, epitaxy will be at the core of our efforts to move beyond traditional feature scaling based on Moore's Law .
But how can you turn the potential of compound semiconductors into a reality—and educate an industry that has relied on silicon-based solutions for decades? In short, we need open foundries to start producing state-of-the-art compound semiconductors.
For example, compare Intel with TSMC in the silicon field. How many designs does Intel process per month? One or two? Now consider TSMC, which is constantly adapting its processes to accommodate various designs, forcing it to continually learn, adapt, and improve. These “best practices” benefit everyone in the chain, from the manufacturing plant all the way to the end user.
The same situation also occurs in the field of compound semiconductors. Some vertically integrated chip manufacturers believe that producing epitaxy in-house is the best option. However, they will never be able to learn as quickly or as quickly as an open foundry that is constantly working with multiple customers on multiple designs. This process of accumulating experience and knowledge gives the open access foundry model an important advantage in the innovation race.
In addition, the environmental issues of chip manufacturing cannot be ignored. The semiconductor industry must be "carbon responsible" and minimize the life cycle impact of its materials and processes. Resources can no longer be considered unlimited. Energy consumption has consequences. The semiconductor industry must find a way to build a digital, connected world that accommodates the global drive toward net-zero emissions. Advanced compound semiconductor materials are driving industry innovation and, due to their well-documented performance and efficiency attributes compared to silicon, provide a key building block in achieving net zero.
I recently spoke with several key figures in the silicon industry, and they highlighted these trends. They increasingly recognize the importance of compound semiconductors and acknowledge the strategic importance of epitaxy. In particular, there is recognition of the support that compound semiconductors provide to the silicon industry.
As a result, advances in silicon and compound semiconductors are no longer independent pursuits; instead, the two industries are collaborating more broadly to deliver disruptive technologies at scale. Extension plays a vital role in this work.
Semiconductors are vital to the world economy. With the growing focus on energy efficiency and ubiquitous connectivity, compound semiconductors are expected to provide next-generation technology solutions that silicon cannot.
