April 13, 2024

At this year’s APEC in Long Beach, California, Wolfspeed CEO Gregg Lowe’s speech was a major highlight of the conference program. Lowe, the chief of the only vertically integrated silicon carbide (SiC) company and cheerleader of this power electronics technology, didn’t disappoint.

In his plenary presentation, “The Drive for Silicon Carbide – A Look Back and the Road Ahead – APEC 2024,” he called SiC a market hitting the major inflection point. “It’s a story of four decades of American ingenuity at work, and it’s safe to say that the transition from silicon to SiC is unstoppable.”

Figure 1 Lowe: The future of this amazing technology is only beginning to dawn on the world at large, and within the next decade or so, we will look around and wonder how we lived, traveled, and worked without it. Source: APEC

Lowe told the APEC 2024 attendees that the demand for SiC is exploding, and so is the number of applications using this wide bandgap (WBG) technology. “Technology transitions like this create moments and memories that last a lifetime, and that’s where we are with SiC right now.”

Interestingly, just before Lowe’s presentation, Balu Balakrishnan, chairman and CEO of Power Integrations, raised questions about the viability of SiC technology during his presentation titled “Innovating for Sustainability and Profitability”.

Balakrishnan’s counterviews

While telling the Power Integrations’ gallium nitride (GaN) story, Balakrishnan narrated how his company started heavily investing in SiC 15 years ago and spent $65 million to develop this WBG technology. “One day, sitting in my office, while doing the math, I realized this isn’t going to work for us because of the amount of energy it takes to manufacture SiC and that the cost of SiC is so much more than silicon,” he said.

“This technology will never be as cost-effective as silicon despite its better performance because it’s such a high-temperature material, which takes a humongous amount of energy,” Balakrishnan added. “It requires expensive equipment because you manufacture SiC at very high temperatures.”

The next day, Power Integrations cancelled its SiC program and wrote off $65 million. “We decided to discontinue not because of technology, but because we believe it’s not sustainable and it’s not going to be cost-effective.” he said. “That day, we switched over to GaN and doubled down on it because it’s low-temperature, operates at temperatures similar to silicon, and mostly uses same equipment as silicon.”

Figure 2 Balakrishnan: GaN will eventually be less expensive than silicon for high-voltage switches. Source: APEC

So, why does Power Integrations still have SiC product offerings? Balakrishnan acknowledged that SiC can go to higher voltages and power levels and is a more mature technology than GaN because it started earlier.

“There are certain applications where SiC is very attractive today, but I’ll dare to say that GaN will get there sometimes in the future,” he added. “Fundamentally, there isn’t anything wrong with taking GaN to higher voltages and power levels.” He mentioned a 1,200 GaN device Power Integrations recently announced and claimed that his company plans to announce another GaN device with even a higher voltage very soon.

Balakrishnan recognized that there are problems to be solved. “But these challenges require R&D efforts rather than a technology breakthrough,” he said. “We believe that GaN will get to the point where it’ll be very competitive with SiC while being far less expensive to build.”

Lowe’s defense

In his speech, Lowe also recognized the SiC-related cost and manufacturability issues, calling them near-term turbulence. However, he was optimistic that undersupply vs demand issues encompassing crystal boules, substrate capability, wafering, and epi will be resolved by the end of this decade.

“We will continue to realise better economic value with SiC by moving from 150-mm to 200-mm wafers, which increases the area by 1.7x and decreases the cost by about 40%,” he said. His hopes for resolving cost and manufacturability issues also seemed to lie in a huge investment in SiC technology and the automotive industry as a major catalyst.

For a reality check on these counterviews about the viability of SiC, a company dealing with both SiC and GaN businesses could offer a balanced perspective. Hence, Navitas’ booth at APEC 2024, where the company’s VP of corporate marketing, Stephen Oliver, explained the evolution of SiC wafer costs.

He said a 6-inch SiC wafer from Cree cost nearly $3,000 in 2018. Fast forward to 2024, a 7-inch wafer from Wolfspeed (renamed from Cree) costs about $850. Moving forward, Oliver envisions that the cost could come down to $400 by 2028 while being built on 12-inch to 15-inch SiC wafers.

Navitas, a pioneer in the GaN space, acquired startup GeneSiC in 2022 to cater to both WBG technologies. At the show, in addition to Gen-4 GaNSense Half-Bridge ICs and GaNSafe, which incorporates circuit protection functionality, Navitas also displayed Gen-3 Fast SiC power FETs.

In the final analysis, Oliver’s viewpoint about SiC tilted toward Lowe’s pragmatism in SiC’s shift from 150-mm to 200-mm wafers. The recent technology history is a testament to how economy of scale has been able to manage cost and manufacturability issues, and that’s what the SiC camp is counting on.

A huge investment in SiC device innovation and the backing of the automotive industry should also be helpful along the way.

Related Content

  • Gallium Nitride (GaN) technology overview
  • GaN’s applications roadmap spotted at CES 2023
  • The diverging worlds of SiC and GaN semiconductors
  • SiC and resurgence of semiconductor vertical integration
  • The GaN semiconductor design view from upcoming APEC show







googletag.cmd.push(function() googletag.display(‘div-gpt-ad-inread’); );

googletag.cmd.push(function() googletag.display(‘div-gpt-ad-native’); );

The post Silicon carbide (SiC) counterviews at APEC 2024 appeared first on EDN.