Technical Insight
SiC infiltrates inverter markets
As photovoltaic inverter manufacturers adopt SiC MOSFETs, when will other industries move towards wide bandgap devices, asks Compound Semiconductor
Which market segment will be the next to design SiC MOSFETs into inverters?
As the power electronics market embraces a buoyant decade, the future for wide bandgap devices looks bright. SiC diodes have stormed into the photovoltaic inverter market, and the same flavour of MOSFET is now penetrating this high-end segment. This trend looks likely to continue.
Earlier this month, France-based analysis business, Yole Developpement, predicted the entire inverter market will swell from $45 billion, in 2012, to $71 billion by 2020. As expected, the market segment of the moment for the compound semiconductor industry is PV inverters, but as Yole analyst Alexandre Avron highlights, unexpected opportunities have surfaced.
“Yes, we are seeing PV inverters first but what is a little surprising is the use of SiC [devices] in light rail traction applications,” he says. “There has been a real interest for the SiC MOSFET and several companies are already carrying out field tests.”
As Avron highlights, Alstom is testing SiC MOSFETs in its auxiliary inverters for trains. “Many more small research and development teams in major businesses, especially in Asia, are pushing for new materials,” he adds. “This is a traditional, conservative market. You just don't expect rail traction to be attracted to SiC MOSFETs.”
More predictably, Avron expects the equally conservative hybrid and electric car manufacturers to show a greater interest in SiC devices. These components can operate at much higher temperatures than silicon counterparts, removing the need for liquid cooling loops and cutting the size, weight and volume of the overall inverter system.
This is quite a bonus for the performance-driven but space-constrained car manufacturer. But still these players will first scrutinise how SiC performs in the solar industry, which as Avron puts it, is providing the “field test”.
“We've looked at these photovoltaic inverters and right now the architecture hasn't changed much,” he says. “Manufacturers have taken out the 1200V silicon IGBTs and put in either a SiC MOSFET or JFET, and changed the input boost converter a little bit. They will now see how the [device] reacts in real-life conditions at higher volumes of production.”
But change is afoot. According to Avron, new inverter architectures - that make the most of the advantages that SiC MOSFETs and JFETs can bring - are under development.
“These have, for example, higher switching frequencies, and if you compare the [designs] to today's architectures, you really see how much room for improvement there is,” he says.
Avron doesn't expect the new, improved inverter architectures will surface for two to three years yet. “We know [other industries] will go for these inverters, but first inverter manufacturers need to see exactly how SiC reacts.” he adds. “The MOSFET is really the heart of the inverter and if your heart isn't strong then you've wasted your money... but in two to three years, manufacturers will no longer be afraid to base a real product on the SiC MOSFET.”
But what about GaN - many power device manufacturers have touted progress in high power GaN diodes and transistors - yet how many products have reached the market? The long-awaited arrival of the SiC MOSFET only came after a long fanfare of announcements, and Avron is concerned GaN transistors are following the same path.
“We've seen lots of marketing about what companies are doing in-house, but still we don't have a 600V device available,” he says. “I expect there are non-disclosure agreements and the players are getting the first production batches from GaN manufacturers, but it's a little late.”
Still, progress is underway. For example, California-based Transphorm qualified its 600V HEMT last year and recently revealed a 600V GaN module for GaN-based high power PV converters. And Avron reckons industry will adopt these and future GaN devices.
“We think it will take a little bit of time... but SiC should be used in very high voltages and we could then see GaN used where SiC is now, PV inverters, electric cars and so on,” he says. “We are seeing different positioning for GaN and SiC, and more people are accepting that there is no full competition between the two [technologies].”
