SiC and GaN Business Opportunities: Evaluation Criteria

SiC vs. GaN: Which Business Opportunity is Greater? - The Answer Changes When You Reframe the Question

The question of "SiC or GaN for next-generation power semiconductors?" has been repeatedly asked within the industry. However, this question is somewhat imprecise. While both SiC and GaN are wide-bandgap semiconductors, they differ in their voltage ranges, switching characteristics, and cost structures. The answer depends not on which is "greater," but rather on "for which application, voltage range, and timeframe are we evaluating?" This article will organize the criteria for evaluating both the SiC and GaN markets, providing a perspective applicable from technology selection to business decisions.

Differentiating by Voltage Range: SiC and GaN Are Not "Competitors"

First, let's clarify the positioning of SiC and GaN based on voltage range. GaN devices are currently seeing practical application primarily in the sub-650V domain. Their ability to achieve higher switching frequencies facilitates the creation of smaller and more efficient systems, leading to their adoption in relatively low-voltage applications such as AC adapters, server power supplies, and consumer fast chargers. Renesas Electronics' acquisition of Transphorm's GaN business is underpinned by expectations for growth in this market.

On the other hand, SiC is accelerating its adoption in high-voltage and high-power applications, primarily around 1200V, including EV inverters, solar power conditioners (PCS), and industrial UPS. onsemi offers a complete SiC portfolio encompassing SiC MOSFETs, SiC diodes, and SiC modules from 650V to 1700V, strategizing to cover a wide range of voltages from automotive to industrial applications.

The perception of "competition" arises from the overlap in the 650V range. In this band, GaN's high-frequency characteristics compete with SiC's thermal resistance and robustness. However, above 700V, practical GaN products rapidly decrease, leaving SiC in a dominant position. Conversely, in the sub-200V range, GaN-on-Si tends to have an advantage in mass production costs. Incorporating evaluation criteria specific to each voltage range is the first step toward accurately assessing business opportunities.

Considering this segmentation, the question "SiC or GaN?" is often misplaced. Narrowing down the application first will naturally determine which material is the focus of evaluation. The next question should be about the specifics of the technical selection criteria.

The Reason for Choosing SiC Isn't Just "High Efficiency" - The Design Reality of Short-Circuit Withstand Time

While "low loss and high efficiency" are often cited as reasons for adopting SiC, in the practical design of EV and industrial systems, "robustness" is an even more critical decision-making factor. Central to this is the short-circuit withstand time (SCWT, also known as Tsc).

Short-circuit withstand time refers to the duration a device can withstand a short-circuit event before failure. It can be understood as the margin of time before the protection circuit operates. Without this margin, no matter how sophisticated the protection circuit, it will not be fast enough.

The issue is that SiC has a shorter margin compared to Si. SiC devices have smaller die sizes and higher current densities, leading to faster temperature increases during a short circuit than Si devices. Under the same conditions, the time allowed for the protection circuit's response is shorter. This characteristic directly impacts the design of gate drivers and protection circuits. Microchip's 700V/1200V SiC MOSFETs specify a short-circuit withstand time of typically 3μs under certain conditions, and this figure becomes the time budget for protection circuit design.

The value of short-circuit withstand time varies with conditions such as drain voltage, gate voltage, and junction temperature. While the withstand time tends to increase with relaxed conditions, design in actual operating environments must account for worst-case scenarios. Interestingly, at high temperatures, the increase in RDSon can limit the saturation current, potentially improving short-circuit robustness. This is a behavior that runs counter to the intuition that "high temperature means degradation."

When selecting devices, short-circuit withstand time is not merely a specification but is intrinsically linked to the protection design of the entire system. This perspective is valuable for both design and procurement.

Evaluating Protection Circuits in Conjunction with Devices: Why DESAT Holds the Key

A standard protection method employed against the challenge of short-circuit withstand time is the DESAT (desaturation) function. DESAT refers to a function that monitors the drain-source voltage (VDS) during the on-state and turns off the power transistor when overcurrent is detected. It operates by determining that the VDS has risen abnormally, indicating the device has "desaturated" from the saturation region, and then shutting off the gate.

When designing DESAT protection, three key parameters are important: the DESAT threshold voltage (VDESAT), the DESAT current (IDESAT), and the short-circuit blanking time. The blanking time is a period of ignored time set immediately after switching on to prevent false detection of transient VDS increases. Shortening this time allows for faster protection but increases the risk of false tripping due to noise. Conversely, increasing it reduces false trips but consumes the margin of the short-circuit withstand time.

This trade-off also exists for GaN, but it is more critical for SiC. Due to SiC's smaller die size and faster temperature rise, there is less margin for setting the blanking time. The SCWT value of the device and the blanking time setting of the gate driver must be evaluated together to determine practical reliability.

Detailed information on these protection circuit design specifics can be found in gate driver IC datasheets and application notes from device manufacturers. Evaluation that includes the combination with the gate driver, not just the selection of individual devices, is crucial for practical reliability.

The Ron vs. SCWT Trade-off: How Manufacturers Are Addressing It

There is a fundamental trade-off between on-resistance (Ron) and short-circuit withstand time in SiC MOSFETs. To reduce Ron, approaches like shrinking the die size or increasing channel density are effective, but these simultaneously increase current density during short circuits, accelerating temperature rise. Lowering loss and ensuring protection margin become conflicting goals at the device design level.

Manufacturers are addressing this challenge through structural improvements. Mitsubishi Electric has significantly improved short-circuit withstand time by introducing a p-type protection layer in their trench SiC-MOSFETs. While trench structures are generally better at reducing on-resistance than planar structures, they also tend to have faster failure during short circuits due to electric field concentration. The addition of the p-type protection layer is believed to mitigate this electric field concentration.

Rohm's 4th-generation SiC MOSFETs also aim to achieve both low on-resistance (RonA) and high short-circuit withstand time through their proprietary device structure. While the specifics of these structures are detailed in each company's technical documentation, the "4th generation" designation implies performance improvements through structural revisions rather than mere miniaturization.

As an industry benchmark, Microchip's SiC MOSFETs (700V/1200V) list a short-circuit withstand time of typ. 3μs in their datasheets under specific conditions. However, this value varies significantly with measurement conditions (VDS, VGS, Tj), so unifying these conditions is a prerequisite for cross-comparison between manufacturers and grades. Evaluating parameters closer to actual usage conditions, beyond catalog values, provides more reliable decision-making data.

Criteria for Evaluating Business Opportunities: A Three-Layered View of Market, Technology, and Supply Risk

While we have examined technical selection criteria, a broader perspective is necessary for evaluating business opportunities. Organizing evaluation criteria for SiC and GaN into three layers—market size, technology maturity, and supply risk—provides a clearer overall picture for decision-making.

From a market size perspective, SiC is driven by its two major applications: automotive and industrial. Its adoption is structured to increase with the rising electrification ratio of EVs. Concrete mass production examples, such as STMicroelectronics' supply of SiC power modules to Ampere for electric powertrains starting in 2026, are accumulating and boosting market formation. While GaN's current main battleground is consumer applications like AC adapters and fast chargers, competition with SiC in data center power supplies and next-generation EV chargers is beginning to emerge.

In terms of technology maturity, SiC's evaluation methods have been largely established through its adoption in automotive applications in the 2010s. GaN, on the other hand, still faces challenges in balancing fast on/off speeds with reliability assurance, and the accumulation of long-term reliability data, particularly for industrial and automotive grades, is still in progress.

Both materials face supply risks. SiC has limited wafer suppliers, and quality stability and supplier concentration risk during the transition to 8-inch wafers are key issues. Mitsubishi Electric's joint development of 8-inch SiC substrates with Coherent can be seen as an attempt to address this supply risk. GaN, utilizing silicon fabs through GaN-on-Si technology, has a more dispersed structure in terms of wafer procurement. However, securing mass-produced wafers for high-voltage GaN-on-GaN (bulk GaN) remains difficult.

This three-layered evaluation serves as a caution against considering technology selection and procurement strategy in isolation. Systems relying on materials with unstable wafer supply, even if the individual device specifications are excellent, carry long-term business risks. Conversely, even if supply is stable, the cost of ensuring reliability may unexpectedly increase if the technology is immature.

What to Confirm Next: Implementing These Evaluation Criteria in Practice

When evaluating business opportunities for SiC and GaN, the first question to ask is "For which voltage range and application?" Once that is determined, the verification items naturally align in the order of technology maturity, selection points such as short-circuit withstand time, and supply risk.

When specifically evaluating SiC MOSFETs, confirming the SCWT value in the datasheet and tracing how it changes under actual operating conditions (VDS, Tj, VGS) is the starting point for reliability design. The combination with protection circuits should ideally be pursued in parallel with device selection for practical implementation.