Algorithm boosts SiC transistor reliability

January 11, 2019 //By Christoph Hammerschmidt
Algorithm boosts SiC transistor reliability
Researchers at the Friedrich Alexander University Erlangen-Nuremberg (FAU) have developed a simple and precise method for detecting defects in silicon carbide (SiC) transistors for higher power, more reliable and more efficient devices.

SiC has high-voltage resistant, high-temperature resistant, chemically robust and suitable for higher switching frequencies, which can ensure a further increase in efficiency. In power electronics, SiC components are being used as an alternative to silicon in an increasing number of applications.

Field-effect transistors (MOSFETs) obtain their functionality from the boundary layer between SiC and a very thin layer of silicon oxide applied to it. It is precisely this layer that confronts researchers with major challenges: Here, unwanted defects occur during production that trap electrical charge carriers and thus reduce the current in the component. The investigation of these defects is therefore extremely important in order to exploit the potential of the material.

Conventional methods for investigating the properties of MOSFETs, mostly from the silicon world, do not take these defects into account at all. Other, more complex measurement methods are either not practicable on a large scale or cannot be applied to finished components at all. For this reason, researchers at the FAU's Chair of Applied Physics have been looking for new ways to better investigate these defects. They noticed that the interfacial defects always follow the same pattern. "We represented this pattern with a mathematical formula," explains doctoral student Martin Hauck. "In this way we can include the interfacial defects so cleverly in the calculation that not only the results of the usual parameters such as electron mobility or input voltage can be precisely determined. In addition, the concentration and distribution of the defects is determined almost incidentally, so to speak".

Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.