Suporting the Electric Vehicle Revolution: Next-Generation SiC Semiconductor Chip Development

The electric drive systems in BEVs and HEVs/PHEVs comprise batteries, motors, inverters, and power conversion equipment such as converters. These systems handle enormous amounts of power conversion within the vehicle. Power semiconductors play a crucial role here, minimizing energy loss while efficiently controlling high voltage and high current.

Energy lost during power conversion becomes heat, causing power semiconductors to heat up. Greater heat generation demands larger cooling mechanisms, creating challenges around limited cabin space and increased vehicle weight.

Enter next-generation power semiconductors using SiC and other advanced materials instead of conventional silicon (Si). Compared to traditional Si, SiC reduces power loss by approximately 70% and enables smaller, more efficient inverters and power modules. The result: extended driving range, reduced size and weight, and lower CO₂ emissions.

The power semiconductor manufacturing process begins with growing crystals—solid structures with regularly arranged atoms that determine fundamental electrical properties. These crystals are thinly sliced and polished into disc-shaped substrates called wafers. Microscopic structures are then formed on the wafer surface, designed to control current at high speed and with high efficiency. After processing, wafers are diced into small pieces to become chips—the active components responsible for power switching. Multiple chips are then combined with insulating materials, encapsulants, interconnects, and other structural components to form modularized power cards.

DENSO has been advancing third-generation SiC power semiconductor development with an eye toward practical implementation. Generation 1.5 and 2 steadily built up impressive track records in automotive applications with industry-leading performance. The third generation, however, represents a breakthrough: delivering both high performance and dramatic cost reduction. How did the development unfold?

Supporting widespread EV adoption required developing third-generation SiC power semiconductor chips compatible with both flat power cards for BEVs and stacked power cards for HEVs/PHEVs. The key requirements: larger chips to reduce the number of power cards needed while achieving overall inverter miniaturization and higher output, plus mass production capability for these advanced products.

First, stacked power cards. Achieving advanced miniaturization, cost reduction, and loss reduction for BEVs proved difficult with conventional stacked power cards, which had been shared with Si-IGBTs (insulated gate bipolar transistors)—silicon-based power semiconductors.

The issue stems from the vertical arrangement of stacked structures, which creates vertical space limitations for BEV inverters while also increasing inductance from an electrical standpoint. To fully unlock SiC's performance potential, a new structure—flat power cards—became necessary for BEVs.

Meanwhile, the HEV/PHEV market was reaching its own inflection point. Si-IGBTs, workhorses of HEV technology for years, were approaching their performance ceiling. For next-generation HEV/PHEV systems, engineers began exploring SiC power semiconductors using silicon carbide's superior heat and voltage resistance.

DENSO's SiC power semiconductors, developed through the company's own basic research, delivered world-class performance via a unique 3D structure. But this structural complexity posed a significant barrier to mass production. High performance yet difficult to manufacture at scale—resolving this contradiction was essential to meeting the demands of full-scale EV adoption.

The Power Devices Engineering Division was tasked with overcoming these hurdles. The division integrates core technologies from the R&D Division into product developments tailored to customer needs. Their role is to bridge the gap to mass production through comprehensive design, performance evaluation, and durability testing, ensuring a smooth transfer to the manufacturing division.

The approach taken by the Power Devices Engineering Division involved fusing two long-standing areas of DENSO expertise. First, the team redesigned the structure for mass production, then developed new, high-productivity processing methods while maintaining the hallmarks of high performance and reliability.

"Tadashi Misumi explains: "In vehicles operating across diverse environments, creating semiconductors that work reliably requires careful consideration—what inspections are needed? How do we guarantee quality? Our strengths lie in design informed by deep knowledge of downstream applications and reliability evaluation technologies that reflect real-world field conditions.

This expertise has been built up since the 1997 first-generation Prius IGBT development at what is now DENSO's Hirose Plant—formerly a Toyota Motor Corporation factory—working hand-in-hand with the vehicle division and refined over many years alongside DENSO's semiconductor operations."

By integrating these capabilities, the third-generation SiC power semiconductor chips achieved exceptional performance while substantially reducing costs.

Yet even after developing the third-generation chips, practical implementation demanded additional technological breakthroughs. Chief among them: larger chips.

In power semiconductors, a greater area per chip translates to higher current capacity and increased power density. Where many manufacturers rely on multiple chips, accomplishing the same work with a single chip directly enables fewer power cards and substantial inverter miniaturization with higher output.

The catch: SiC wafers contain numerous crystal defects, and enlarging chips drives down yield rates. Industry-standard maximum size at the time hovered around 5.0mm square. Producing large chips significantly exceeding this threshold at production-viable yield rates was considered technically formidable.

Ito explains their approach: "The breakthrough came from introducing our proprietary RAF method—Repeated A-Face growth, a low-defect SiC crystal manufacturing technology—into large-scale mass production for the first time. The RAF method involves slicing grown SiC crystals, extracting sections with minimal defects, and then repeating the crystal growth cycle. By minimizing defects from the seed stage during wafer production, we achieve exceptionally high-quality wafers."

This technology reduced SiC crystal defects to one-third of competitor levels. The result: large chips measuring 9.0mm square or greater—far exceeding the 5.0mm industry standard—at high yield rates. By pairing SiC's superior heat and voltage resistance with large-area semiconductor chips, DENSO engineered a design that cuts the number of power cards while delivering higher efficiency than conventional approaches.

"This project demanded producing high-quality chips at an unprecedented scale," notes Ito. "Without the extraordinary dedication and collaboration from our manufacturing division and production sites worldwide, mass production on this scale would have been impossible."
ーIto

Yet the RAF method alone couldn't fully realize inverter miniaturization.

Wafer supply for inverters came from two sources: those produced using RAF technology and those from mass production partners. Quality varied subtly between them. Crystal defect quantity, type, and distribution—these differences impact post-manufacturing inverter electrical characteristics and yield rates. Designs assuming stable, uniform wafer quality risked unexpected problems during mass production.

Another challenge emerged: automotive semiconductors must withstand high currents and high-speed switching, yet the team encountered unexpected device behaviors virtually unknown even in academic research.

DENSO's vertically integrated monozukuri approach proved crucial to overcoming these hurdles. The Power Devices Engineering Division understood phenomena at both the material property and crystal structure levels of SiC, implementing countermeasures by fine-tuning characteristics to suppress anomalous behaviors. The inverter division, intimately familiar with real-world vehicle operation, proposed control-based solutions to manage these behaviors at the inverter level. The research division joined forces to investigate and explain the unexpected phenomena.

"When device challenges arise, we don't limit ourselves to device-only solutions," says Misumi. "We can develop countermeasures spanning devices, inverters, and modules. That's the core advantage of DENSO's vertically integrated monozukuri approach."
ーMisumi

Today, third-generation SiC power semiconductors form the technological backbone of two major vehicle electrification projects at DENSO, with implementation advancing in forms optimized for BEVs and HEVs/PHEVs, respectively.

The first is the eAxle-integrated inverter project for BEVs. This project is a three-company collaboration, with DENSO primarily responsible for inverter development.

eAxles (electric axles) installed in EVs integrate three critical components into a single drive unit: motor, inverter, and transaxle (combining transmission and differential).

DENSO's SiC power semiconductors serve as the heart of inverters embedded in these eAxles, delivering tangible results across both BEV and HEV/PHEV applications.

For BEVs, the technology powers multiple dedicated models. For HEVs/PHEVs, it enables simultaneous inverter miniaturization and dramatic power conversion efficiency gains—enhancing fuel economy and power performance while elevating the overall appeal of hybrid vehicles.

What's next for these next-generation SiC power semiconductors? Ito outlines the vision ahead:

"Through the third generation, we've mass-produced using 6-inch wafers. However, transitioning to larger-diameter wafers is essential to enhance cost competitiveness. While global competition is intense, we aim to develop compelling devices that ensure DENSO SiC is the choice for customers worldwide. Beyond that, we envision applications extending far beyond the automotive sector. Given their high efficiency and reliability, I am confident our SiC power semiconductors hold tremendous value for the green energy sector and social infrastructure. To achieve this, we will continue to develop even more competitive devices by capturing needs from various regions and collaborating with research institutions around the globe."
ーIto

Misumi, who brings over 20 years of power semiconductor development experience to the team, reflects on the journey:

"When I started in power semiconductor development, I never imagined HEVs would achieve such widespread adoption. The knowledge and expertise we gained solving Si-IGBT development challenges, combined with the technology accumulated through years of SiC-focused basic research, represent the core competitive advantages of DENSO's semiconductors. I'm confident SiC power semiconductors still have significant room for advancement. We're pursuing higher performance and lower costs to serve customers worldwide and contribute meaningfully to carbon neutrality."
ーMisumi

Yamamoto has been involved in SiC R&D since his second year at the company—back when wafers were just 2 inches in diameter and devices were manually fabricated in beakers.

"SiC power semiconductors were adopted for BEV motor-driving inverters in the second generation, and with the third generation, we reached the cost-reduction phase for mass-market models," Yamamoto notes. "SiC technology still has enormous growth potential. My goal is to advance performance and reduce costs so that DENSO-made SiC power semiconductors power EVs around the world."
ーYamamoto

Development continues with 8-inch wafers while R&D explores next-generation power semiconductors using new materials and architectures beyond SiC.

“The market environment is challenging, but our long-term vision remains crystal clear,” says Ito. “We will develop high-quality, cost-effective power semiconductors that respond to customer needs and deliver them globally. We aim to realize a mobility society that prioritizes environmental sustainability and safety on a global scale, contributing to climate change mitigation. Our immediate priorities are advancing 8-inch wafer production and next-generation device development to drive further cost reductions and productivity improvements. These innovations are expected to expand applications beyond mobility into renewable energy, industrial equipment, and many other sectors.”
ーIto