Technical Insight
SiC MOSFETs tipped for the top
Which SiC transistor will industry choose? Compound Semiconductor talks to Jeff Casady, ex-chief technology officer of SemiSouth and now at Cree, to find out more.
In November this year, US-based Cree unveiled a 1200V SiC power module based on MOSFETs and Schottky diodes. Operating at up to 100kHz frequencies and targeting high power converters, industrial motor drives, solar inverters and uninterruptible power supplies, the device is significant.
It not only joins the industry's increasing army of SiC devices targeting tomorrow's high voltage, energy efficient power converters, but crucially, it uses MOSFETs.
Jeff Casady, one-time chief technology officer at recently shuttered SiC JFET developer, SemiSouth, and now product portfolio planning manager at Cree, says we can expect more.
“As far as Cree and the rest of the market goes, we are now seeing substantial performance, availability and reliability improvements in MOSFETs, almost on a weekly basis,” he says. “Costs are falling fast and there are going to be several more announcements coming early next year from Cree. I imagine our [MOSFET] competitors, primarily Rohm, will be doing the same thing.”
For more than a decade, myriad organisations have honed SiC semiconductors to tolerate higher voltages and temperatures than silicon, while ensuring the devices switch more rapidly with lower losses. Cree, Rohm, ST Microelectronics and many more have focused on the tried-and-tested MOSFET, Infineon is forging ahead with its JFET, Fairchild recently released a 1200V SiC BJT, and United SiC is working on IGBTs.
But with industry consolidation setting in - as evidenced by SemiSouth's recent demise - will this diverse range of transistors remain? As Casady now says: “When you talk to customers, they always wanted the MOSFET from probably the early days.”
According to Casady, this transistor was once a technical challenge for SiC, but not any more. “We're seeing a lot of aggressive design-in activity and we think the MOSFET market in SiCs is going to be very large,” he adds.
So why does he think the MOSFET could emerge as the hot favourite? Put simply, the device is a known entity for designers used to driving silicon HBTs and MOSFETs.
“These people want a device to replace silicon, and the SiC MOSFET is something they really understand,” explains Casady. “They understand how to drive it and it's also got a body diode as most silicon sets do. We know that competitors looking at bipolar [transistors] and JFETs have a much more difficult conversation with their customers as they are not used to those parts.”
But it's not all about silicon-based semiconductors. Manufacturers of GaN ICs, including International Rectifier, Efficient Power Conversion, Transphorm and Fujitsu Laboratories, have already made in-roads into higher voltage markets, with 600V devices readily available.
Many players assert that 1200V GaN ICs will be good-to-go within a few years, and crucially, the cost of these devices will drop more rapidly than SiC versions. Could the GaN IC outpace both silicon and SiC in high power, high voltage applications, including the all-important hybrid electric vehicle market? Casady thinks not.
Focusing on electric vehicles he says: “I don't think GaN will do well here as the reliability requirements are so stringent.”
As he asserts, SiC devices have been in the market for at least a decade - compared to the few years notched up by GaN - and are already used in high reliability applications from solar and wind turbine inverters to industrial power supplies.
“Manufacturers from the automotive market needs years of field data from these high reliability markets before they choose a new technology,” he says. “[SiC] is already going into markets that require twenty years of reliability [data]. It will be really challenging for GaN to come in and do that.”
But perhaps most pertinent is that Casady's new found employer has a choice. And it's not GaN.
As he highlights, between LEDs and RF power devices, Cree makes most of its money on GaN devices. “People at Cree are really objective about this... and we could choose to use GaN or SiC in any of these applications,” he adds. “But for hybrid electric vehicles, and all these 600V and higher power electronics applications, from my vantage point, it seems SiC is the clear choice.”
.jpg)
Image caption:
SiC transistors allow higher efficiency, compact and lighter weight systems, cutting total system costs in applications such as solar power, compared to conventional silicon-based technologies. Credit: USAF
It not only joins the industry's increasing army of SiC devices targeting tomorrow's high voltage, energy efficient power converters, but crucially, it uses MOSFETs.
Jeff Casady, one-time chief technology officer at recently shuttered SiC JFET developer, SemiSouth, and now product portfolio planning manager at Cree, says we can expect more.
“As far as Cree and the rest of the market goes, we are now seeing substantial performance, availability and reliability improvements in MOSFETs, almost on a weekly basis,” he says. “Costs are falling fast and there are going to be several more announcements coming early next year from Cree. I imagine our [MOSFET] competitors, primarily Rohm, will be doing the same thing.”
For more than a decade, myriad organisations have honed SiC semiconductors to tolerate higher voltages and temperatures than silicon, while ensuring the devices switch more rapidly with lower losses. Cree, Rohm, ST Microelectronics and many more have focused on the tried-and-tested MOSFET, Infineon is forging ahead with its JFET, Fairchild recently released a 1200V SiC BJT, and United SiC is working on IGBTs.
But with industry consolidation setting in - as evidenced by SemiSouth's recent demise - will this diverse range of transistors remain? As Casady now says: “When you talk to customers, they always wanted the MOSFET from probably the early days.”
According to Casady, this transistor was once a technical challenge for SiC, but not any more. “We're seeing a lot of aggressive design-in activity and we think the MOSFET market in SiCs is going to be very large,” he adds.
So why does he think the MOSFET could emerge as the hot favourite? Put simply, the device is a known entity for designers used to driving silicon HBTs and MOSFETs.
“These people want a device to replace silicon, and the SiC MOSFET is something they really understand,” explains Casady. “They understand how to drive it and it's also got a body diode as most silicon sets do. We know that competitors looking at bipolar [transistors] and JFETs have a much more difficult conversation with their customers as they are not used to those parts.”
But it's not all about silicon-based semiconductors. Manufacturers of GaN ICs, including International Rectifier, Efficient Power Conversion, Transphorm and Fujitsu Laboratories, have already made in-roads into higher voltage markets, with 600V devices readily available.
Many players assert that 1200V GaN ICs will be good-to-go within a few years, and crucially, the cost of these devices will drop more rapidly than SiC versions. Could the GaN IC outpace both silicon and SiC in high power, high voltage applications, including the all-important hybrid electric vehicle market? Casady thinks not.
Focusing on electric vehicles he says: “I don't think GaN will do well here as the reliability requirements are so stringent.”
As he asserts, SiC devices have been in the market for at least a decade - compared to the few years notched up by GaN - and are already used in high reliability applications from solar and wind turbine inverters to industrial power supplies.
“Manufacturers from the automotive market needs years of field data from these high reliability markets before they choose a new technology,” he says. “[SiC] is already going into markets that require twenty years of reliability [data]. It will be really challenging for GaN to come in and do that.”
But perhaps most pertinent is that Casady's new found employer has a choice. And it's not GaN.
As he highlights, between LEDs and RF power devices, Cree makes most of its money on GaN devices. “People at Cree are really objective about this... and we could choose to use GaN or SiC in any of these applications,” he adds. “But for hybrid electric vehicles, and all these 600V and higher power electronics applications, from my vantage point, it seems SiC is the clear choice.”
Image caption:
SiC transistors allow higher efficiency, compact and lighter weight systems, cutting total system costs in applications such as solar power, compared to conventional silicon-based technologies. Credit: USAF
