FAYETTEVILLE, Ark. -Silicon carbide devices are the hottest news in semiconductor technology and may create useful advances in automobile, plane and utility power systems. Before these devices can be widely implemented, designers need to know more about how they operate. University of Arkansas researchers Alan Mantooth and Kraig Olejniczak are developing the computer models that will make that possible.

"Compact circuit simulation models allow designers to use a technology more effectively," explained Mantooth. "Power electronics circuit designers rely on computer simulations to understand the details of their circuit operations. Some of these analyses cannot be performed by using hardware prototypes."

Mantooth, associate professor of electrical engineering, and Olejniczak, professor of electrical engineering, created a working group with a combination grant in excess of $1 million. Grant money includes $270,000 from the National Science Foundation and $450,000 from collaboration with two federal agencies, the National Institute of Standards and Technology and NASA-Glenn. The remaining $300,000 will come from industry partners Cree Research and Northrop-Grumman.

Mantooth, Olejniczak and two Ph.D. students will work with this group to study the physical properties of this emerging technology and describe them mathematically. The group will focus on power semiconductor devices that conduct from tens to hundreds of amperes of current and withstand up to 5000 volts.

Silicon carbide is ideal for power devices because it has unique thermal and mechanical properties. One of the hardest known substances, it can operate at temperatures in excess of 500°C (932°F). This makes it ideal for applications such as in car or aircraft engines or utility power systems.

In addition, silicon carbide has higher reliability, higher immunity to thermal runaway, reduced switching losses and higher current density than silicon. Because it conducts heat so well, it can result in reduced cooling costs for devices and increase efficiency.

Before circuit designers can take advantage of these properties, they must be able to test the circuits and evaluate their performance. Typical circuit analyses include nominal and worst case operating conditions, variations due to manufacturing tolerances on parts, failure modes and determination of the safe operating area of the circuit.

"Physical development of a silicon carbide circuit is expensive and time consuming," Mantooth explained. "Building and testing each physical prototype would be prohibitively expensive. Our models will expedite the implementation of silicon carbide in circuits."

Preliminary results of their work were presented at the IEEE Power Electronics Specialists’ Conference and will be published in an upcoming issue of IEEE Transactions on Power Electronics.