Advancements in Wide-Bandgap Power Devices: GaN vs. SiC (Podcast)

Advancements in Wide-Bandgap Power Devices: GaN vs. SiC (Podcast)

Published: March 16, 2024

Transcript:

Welcome to PowerUP, a podcast show hosted by Maurizio Di Paolo Emilio that brings life to some of the stories on power electronics technologies and products featured on PowerElectronicsNews.com, and through other AspenCore Media publications. In this show, you’ll hear both engineers and executives discuss news, challenges, and opportunities for power electronics in markets such as automotive, industrial, and consumer. Here is your host, editor-in-chief of PowerElectronicsNews.com and EEWeb.com, Maurizio Di Paolo Emilio.

MAURIZIO DI PAOLO EMILIO: Hello everyone, and welcome to this new episode of PowerUP. Today, I would like to share the experience at APEC in Long Beach [California] last week. I am deeply grateful to everyone I had the honor of meeting during my participation at APEC. Engaging with experts in the power electronics industry was inspiring, enriching my knowledge base and motivating me to commit even more of myself to the field. Thank you very much. Special recognition goes to companies and academics in power electronics, especially GaN and silicon carbide technology experts for their remarkable achievements. Power electronics is important for tackling the climate crisis by reducing CO₂ emissions by enhanced energy efficiency. GaN and silicon carbide technologies all create promise for future advancements in power electronics, paving the way for even more efficient and sustainable solutions. In this video, we will analyze two different topics—substrates and applications for wide bandgap—and we will do this with great speakers from Texas Instruments, EPC, Power Integrations, onsemi, Infineon, and Qorvo. Let me introduce the topics and speakers. Enjoy.

Welcome to APEC 2024, showing in Long Beach. I am Maurizio, editor-in-chief of Power Electronics News. So APEC is one of the most important power electronics conferences, where designers, engineers, are meeting and talking about new challenges, new applications, new devices in terms of power electronics. So this video will delve into exciting advancements and applications about gallium nitride and silicon carbide wide-bandgap technologies. I will be conducting interviews with six speakers—so three experts for gallium nitride and three experts for silicon carbide. So the topics that we will discuss are substrates and application markets. So due to the physical constraints, carbon silicon technology is enabled to provide power products with an increased power density, downsizing, and energy conversion efficiency that the market demands in order to address environmental concerns. So significantly, efficiency increases are made possible by wide bandgap, such as gallium nitride and silicon carbide. So based on the amount of energy that is needed to move electrons from valence to conduction band, GaN and SiC are related to wide bandgap. So these energies, 1.1 electron volts [eV] for silicon, 3.2 eV for silicon carbide, and 3.4 eV for GaN. So GaN is suitable for high-frequency, high-power applications because it has high electron mobility and can switch power quickly. So silicon carbide can handle higher voltages and temperatures, so wide-bandgap technologies offer a practical implementation for electric vehicles, which is a really nice topic nowadays because higher voltages can reduce losses and also can reduce model size and weight while boosting thermal efficiency.

So let me introduce the speakers. Let me start with GaN and then with silicon carbide. In terms of GaN, I’m going to interview the following speakers: Robert Taylor, applications engineer at Texas Instruments; Michael de Rooij, vice president of applications engineering at EPC; and Balu Balakrishnan, CEO of Power Integrations. So which are the questions that I’m going to ask them? First one, when considering GaN-based power devices, could you elaborate on the significance of the choice of substrate material on device performance, reliability, and manufacturability and how researchers are addressing substrate-related challenges?

ROBERT TAYLOR: Thanks very much. That’s a great question. In terms of substrate material at Texas Instruments, we are using traditional silicon substrates for our GaN devices. So this allows us to not only ramp very quickly in terms of being able to get new devices out. It also provides us a significant cost advantage in that we’re able to use a lot of the manufacturing support staff facilities that we already have in place, versus silicon carbide, for example, the substrate material is very well-understood and very easy to work with. So we think that that provides us a great advantage in terms of being able to get out the latest and greatest types of devices for many different applications.

MICHAEL DE ROOIJ: EPC makes devices from 15 V through 350 V. We do that on a low-cost silicon substrate. This is a good material, because it is well-matched for the lattice structure to grow the gallium nitride on. The challenges come in when you want to go with high-voltage devices. They are thicker requirements for the heterostructure that’s needed in order to meet the blocking voltage. So for higher-voltage devices, you have four alternative options. They are gallium nitride, sapphire, silicon carbide, and QTS. The gallium nitride is expensive material, but it’s well-matched to of course the gallium nitride. It is a bit more expensive. Then you’ve got sapphire. And sapphire is less thermally conductive, so you have challenges to meet the thermal requirements, but it’s still a good match to that of the gallium nitride lattice structure. And then you’ve got the silicon carbide. It’s an expensive material. It’s a very good thermal conductor. And it’s also a good lattice match with that of the gallium nitride. And finally, you have the engineered material QTS that’s showing good promise. It also has a good thermal performance and thermal expansion coefficient to that of gallium nitride, and we’ll probably see that more in the future.

BALU BALAKRISHNAN: So the choice of materials in GaN directly relates to the performance and the cost and reliability. Our choice of materials, whether substrate or otherwise, is, we chose that so that we can get the lowest cost, number one, highest reliability, number two. Number three, it has to have very high performance in terms of not only the switching performance but also in terms of being able to go to higher voltages. The higher the voltage you can go, the higher power you can deliver, and we have announced already 1,250-V GaN, and we are planning to announce even higher voltages going forward, and that will allow us to directly compete with silicon carbide after about 10 to 20 kW.

MDPE: Second question: GaN-based power devices have gained traction in various applications, such as automotive, consumer, and renewable energy in particular. So could you elaborate on specific market segments where GaN devices are demonstrating significant advantages over traditional silicon-based solutions, and how are these advantages driving the adoption of GaN technology in those areas, and which is the technology direction of your company in various markets? So let’s listen to the speakers and their comments.

RT: In terms of GaN devices for specific applications, the biggest one that we see currently is in server power supplies in data centers. Data centers are advancing hugely because of AI and many other different types of applications. And within those data centers, the server PSU is a prime example of where GaN can have a significant advantage over silicon. It enables us to use different topologies that were no longer achievable before with silicon, which allows us to boost power density and efficiency inside of those systems. In addition, we see another huge market for GaN in the solar market. So whether that’s microinverters or other types of solar applications, that’s where we see another huge advantage. And then the final one is in automotive. There’s definitely a play within automotive for GaN as well, whether it’s with 400-V systems or 800-V systems as well. So those are the big markets that Texas Instruments is moving forward in, and we’re really excited about the technology that GaN is going to be able to provide for those.

MDR: So gallium nitride makes differences in various applications, such as DC/DC converters, automotive, motor drives, and LiDAR systems. In DC/DC, we’ve been primarily focused on 48-V to 12-V conversion, and we’re able to shrink down converters dramatically compared with silicon devices and get very high performance out of them, and making big inroads in the server market. The segue to that are the DC automotive applications with the significant increase in electrification of electric vehicles coming on the market, and they need the 48-V to 12-V conversion to bridge the gap from the 12-V–based systems to 48-V systems, and of course for higher voltages. Gallium nitride has also been demonstrated to be a very good device for LiDAR systems with very enhanced performances and very fast switching speeds that just cannot be achieved with silicon devices. Finally, gallium nitride is making improvements in motor drives, where we can switch them at higher frequency, which yields cleaner sinusoidal excitation of the motor. That significantly reduces the torque ripple motor and can improve mechanical efficiency between 11% and 14%, and of course, the slight loss in efficiency is made up by all those gains in mechanical efficiency. And at the DC end of the motor drives, you can also eliminate many large and bulky components for filtering and simply use ceramic capacitors. So gallium nitride is making a lot of inroads in various applications, and you’ll see more of that in the future.

BB: So the GaN devices are incredibly high-performance with the switching losses. It also can delineate very, very low R_DS(on) in a very small byte size. So anything roughly about 30 W in a power supply, we use GaN. Below 30 W, you just can’t make the GaN die small enough to be able to handle it and put in a package. So anything about 30 W, we use GaN in literally almost all of our products going forward. And that applies to products of going to adaptors, which is an early user of GaN. But going forward, we are using it in displays and zoom-in applications. But also, we will be offering products for higher power levels, like a few kilowatts for server power supplies, especially the AI servers that go up to 5 kW. And when you go to 5 kW, efficiency becomes a huge problem because you have the same space in the data center to deliver the 5 kW. So that’s a great application. The infrastructure for 5G stations also requires few kilowatts, and that’s another application. And then if you go a little bit higher-power, which we think we can go to 10 or 20 kW, then you can address things like on-board chargers in automotive applications, and also DC/DC converters in automotive. So I would say literally any power converter below 10 or 20 kW can be addressed with the technology we have already developed. Now, when you go to much higher power levels, like under 200 kW, which is where the silicon carbide shines as a wide-bandgap technology for automotive, GaN technology is not quite there yet. However, I believe GaN will get there. There are some engineering challenges to be solved, but GaN inherently is more cost-effective than silicon carbide. Silicon carbide, not only is the material expensive, but manufacturing silicon carbide is very expensive in terms of energy usage. So I don’t see silicon carbide ever being competitive directly with silicon in terms of cost-effectiveness, whereas GaN will be equal or less than silicon within the next few, maybe one or two years. And so GaN has a bright future relative to silicon carbide as the power levels go up and is able to address several-hundred kilowatts, which is a prime market for silicon carbide today.

MDPE: Thank you. So let’s switch to silicon carbide. So I’m going to interview the following speakers: Ajay Reddy Sattu, director of the product market GIS business unit at onsemi; Peter Friedrichs, vice president of silicon carbide at Infineon; and Ramanan Natarajan, product line marketing for power products at Qorvo. So which are the questions that I’m going to ask? First one: With silicon carbide substrates being used in silicon carbide–based power devices, what advancements have been made in substrate quality and manufacturing techniques to improve device performance, reduced effects, and enhanced reliability? And what more developments are needed to optimize substrate properties for future silicon carbide power devices?

AJAY REDDY SATTU: When you think about silicon carbide reliability and device performance in the context of defects, so there are generally two types of defects, right? So there are killer defects and there are also non-killer defects, right? So from that context and very generally these types of defects originate from, you have to look at both silicon carbide substrate and also the silicon carbide lapping and polishing, but also silicon carbide epi, right? So these are the three different areas where the difference can originate. Now the industry has probably resolved many of the killer defects that are generally seen in the substrates and also after epi. The non-killer defects are the ones that are generally, let’s say a possibility of escapes and so on. So at onsemi, what we have done is to develop many algorithm-based methods to screen out these devices as the material progresses through in the fab, right? So in addition to that, what we have also done is the fact that we are vertically integrated that would allow us to have shorter feedback loops from the substrate point of view and also from the epi point of view, and then the device manufacturing. Now most of the industry again today is at 6 inch, and these issues have been resolved quite nicely. But the challenge for us is, as we move from 6 inch to 8 inch, you have to consider some of the same issues, but also the issue of the 8-inch wafers, right? So when you look at it from this context, what we have been doing at onsemi is to continuously have those shorter feedback loops to improve the performance of these devices. Now in addition to this, what we also look at is develop silicon carbide technologies that are optimized by application. And what I mean by that is, sometimes I know we want to address, let’s say, our model, and we also want to address industrial applications, and also we need to have different technologies that can address these end applications.

PETER FRIEDRICHS: Yeah, thank you for the question. As a matter of fact, we have already reached meanwhile a very good silicon carbide substrate quality. The wafer diameters are compatible with silicon wafer diameters. So we are moving from about 50 mm to 200 mm a day, and we see a transition to the new wafer diameter. We also see a very good, let’s say status regarding defect entities. So we have not to take into account a step back in defect densities, as we have experienced it in the previous diameter changes. However, still, as the material quality is not where we want to have it, there is still a relatively high density of certain critical defects we need to eliminate, where we need to improve, so therefore, one of the key topics for further material improvement is lower defect entities, and the second topic regarding materials is that the mechanical properties of the modern wafers are more or less more silicon-like, meaning there are flatness, local thickness variations are topics which we need in order to use sophisticated technologies, and this should be the target for the next steps.

MDPE: The second question is: With the increasing demand for high-power and high-temperature applications, silicon carbide–based power devices have emerged as promising alternatives to silicon devices. Can you discuss specific market sectors or applications where silicon carbide devices are proving to be particularly beneficial and where your company is investing? So let’s listen to the speakers and their comments.

ARS: So in terms of different applications for silicon carbide, you know, again for automotive, it’s well-understood that the key application areas are for traction inverters, on-board chargers, and also for high-voltage DC/DC converters. And on the other side for industrial applications, the key end areas are DC fast chargers, UPS energy storage, and also solar. But when you look at it in the context of each of these applications, how onsemi has been optimizing silicon carbide technology, let me give you a few examples. So for on-board chargers and for high-voltage DC-to-DC converters in automotive, high power density is a must. So obviously, customers would require a technology that is capable of high-frequency switching with lower losses, right? So we have developed a switching flavor in our Gen 3, which is M3S, and on the traction inverter application, we have developed a technology that is more traction-capable, or N3T, or a variation of it. Now similarly, on the industrial side, since most of the applications will require high-frequency switching applications anyway, so we have M3S technology that can improve switching performance in UPS, energy storage ESS, and also in solar. Now for the motor drive market in industrial, we again can use the traction flavor of the silicon carbide technology that will improve end-system performance.

PF: Yeah, fortunately, silicon carbide is not any longer focused to selected market segments. So meanwhile, we are supplying different existing markets for power semiconductors and many emerging markets for power semiconductors. In particular, of course, topics like EV charging or solar power conversion, storage applications, all the applications where power density and high efficiency are extremely important, are the ones which have already a very large penetration with silicon carbide. So of course, we as Infineon focus on that. But we intend to serve all the markets interested in silicon carbide. Therefore, recent developments for instance as a segment of traction propulsion, in the segment of standard motor drives, servo motor drives, UPS systems, power supplies, these are all applications which we see as the most important ones for silicon carbide, and of course, usings also will add on. So let’s say circuit breakers, a very interesting application for SiC, and not to forget also applications like aircraft, electrified aircraft, electrified ships, etc.

RAMANAN NATARAJAN: Well, silicon carbide devices have really re-energized the power-conversion world, and they are definitely bringing benefits over silicon devices. They switch faster. They’re able to provide lower on-resistance in the same package and essentially help customers make things more efficient. And different applications are taking advantage of this in different ways. At Qorvo, we are primarily focused on a broad set of industrial applications, such as battery chargers, battery test equipment, test and measurement equipment, renewable energy applications like solar inverters, energy storage systems, and of course electric vehicles, charging stations, on-board chargers, and DC-to-DC converters. And these applications are adopting silicon carbide for a variety of reasons. Data center power suppliers want to reduce operational expenses of the data center. Electric-vehicle manufacturers want to charge their cars faster and make their vehicles lighter. Renewable energy applications want to be able to run at higher temperatures and have a longer lifetime and higher efficiency with their products. So really, it’s a really thrilling time to be in power electronics and power semiconductors, and wide-bandgap devices like silicon carbide are really making the way for the world here.

MDPE: So the wide-bandgap industry is investing substantial resources in developing efficient technologies to drive an energy revolution in a more sustainable and environmental direction. So the market is going to grow in the next few years. I look forward to new devices and applications. So thank you to the speakers for your time. Thanks a lot for this opportunity, and thank you for listening.

That brings us to the end of this episode. Stay tuned with more news and technical aspects about power electronics. If you are listening to this on the podcast page at EETimes.com or PowerElectronicsNews.com, links to articles on topics we have discussed are shown on this page. PowerUP is brought to you by AspenCore Media. The host is Maurizio Di Paolo Emilio, and the producer is James Ead.

The post Advancements in Wide-Bandgap Power Devices: GaN vs. SiC (Podcast) appeared first on Power Electronics News.

[ad_2]

Source link