Ilia Zverev

Matched pair of CoolMOST/sup TM/ transistor with SiC-Schottky diode-advantages in application

The new CoolMOS/sup TM/ C3 generation combines extremely high on-state conductivity with ultra fast switching speed at full pulse current capability. In the first generation of CoolMOS/sup TM/, due to the small chip size, one had a reduction of the saturation current at the cell level. This technique results in a reduced current capability of the device at low gate voltages. In many applications, the outstanding switching performance of the cannot be utilized due to the dynamic behaviour of CoolMOS/sup TM/ the diode. For this reason ,a whole family of SiC-diodes has been developed to get the ideal matched pair of switch and ultra fast diodes. The goal of ultra low loss applications in SMPS, power factor correction circuits and motor control units will be achieved completely.

SiC Power Devices: How to be Competitive towards Si-Based Solutions?

In this paper it will be explained with the help of examples, how SiC power devices can conquer their market. It requires a re-optimizing of power applications taking into account the sum of properties offered by this new SiC technology. This is necessary because the costs per area for SiC devices are still two orders of magnitude higher compared to competing Si devices. Introduction During the last two years a new situation arose for SiC power devices. They are now beyond the point of technology demonstration and have entered the market. Specifical ly, SiC Schottky diodes act as a precursor being introduced in Spring 2001 by Infineon [1] and also by Cree/Microsem i [2]. With this step a new focus becomes important for the SiC researchers: it’s no longer mos t i portant to show basic capabilities and new world record data of the power devices, but to consider what properties and trade off make such devices most acceptable for the system designers. One always has to be a war , that SiC devices are and will be significantly more expensive than the Si parts they shall rep lac . Of course, just looking on the physical properties it is a clear deal for SiC. It seems to be the material for power devices. But unfortunately in many cases these advantageous properties can be more than compensated by just more silicon. Here are some considerations to explain this: • The very low specific on resistance for a given breakdown volt age, which is achievable with unipolar SiC devices is explained in many papers and estimated to be u p to a factor of 700 lower than in comparable unipolar Si devices (as review see[3]). On the other hand, the today’s costs of a fully processed 6” Si wafer is more than 10 times lower than the cor r sponding costs of a 2” SiC wafer, i. e. the cost per area ratio is about 1/100. Therefore to get a figure which describes cost/performance one has to divide the above mentioned factor of 700 by 100 leaving a value of only 7. This does not take into account the still high defect density of SiC, which is limiti ng the yield especially for devices larger than 1 mm2. Further it is not considered in this simple figures, that with innovative device structures like compensation devices (e. g. CoolMOS ) even Si can do much better than expected. • SiC allows unipolar devices for much higher blocking voltages than it is reas onable with Si. So the major competition happens between u ipolar SiC and bipolar Si devices. No doubt, that such Si devices are always much cheaper for a given current rating. Furthermore bip olar devices usually show a (slightly) negative temperature coefficient in on-resistance since a lso minority carriers are involved, making them much more stable under surge current conditions with respect to un ip lar SiC devices. It is the only justification for a SiC device in this case, that switching l osses are dominant in the specific application. So as precondition for market acceptance it has to be clear, th at the system engineer gets a clear benefit by applying a comparatively high switching frequency. To finally decide, which SiC device can be competitive it is nec essary to carefully evaluate the possible cost/performance gain with the eyes of a system designer f or each device type and for specific target applications. Based on this application knowledge it is then possible to t ailor a device in a most efficient way and not to overor undervalue certain characteristics. The next barrier to be overcome for successful commercializat on of SiC devices are the design rules for power circuits which are optimized for Si devices. The introduct ion of novel device technologies requires a paradigm shift and different trade offs between various topology choices f r the same purpose. Materials Science Forum Online: 2003-09-15 ISSN: 1662-9752, Vols. 433-436, pp 805-812 doi:10.4028/www.scientific.net/MSF.433-436.805

System Design Considerations for Optimizing the Benefit by Unipolar SiC Power Devices.

Focussing on unipolar SiC power devices a variety of applications are described, where cost reduction can be a achieved on system level even for SiC device costs being several times higher than the costs of the competing Si devices. Based on the specific properties of SiC devices like Schottky diodes and JFETs it is explained with the help of these examples how this is attainable.