SiC MOSFETs In The Landscape Of Modern Power Devices

Over the years, low losses possible by high breakdown field madesilicon carbide (SiC) MOSFETsextremely popular amongst engineers. At present, they are mostly used in areas whereIGBTs(Insulated Gate Bipolar Transistors) have been the prevailing component of choice before. But which role do SiC MOSFETs play in today’s landscape of power devices?

With SiC MOSFETs (Metal-Oxide-Semiconductor Field-Effect Transistors), significantly lower conduction losses are possible, especially at partial loads, due to their linear output characteristic (see figure 1). This is in contrast to the IGBT situation with a knee voltage (V_{ce_sat}). Furthermore, designers could theoretically decrease the conduction losses to infinitesimally small numbers by using larger device areas. With IGBTs, this cannot be done.

Switching losses are key

If we look at switching losses, the lack of minority carriers in conduction mode eliminates tail currents, and thus very small turn-off losses are possible. Turn-on losses are also reduced compared to IGBTs, predominantly due to the smaller turn-on current peaks. Both loss types do not show an increase with temperature. However, in contrast to IGBTs, turn-on losses dominate while turn-off losses are minor, which is often the opposite of the behavior of an IGBT. Finally, engineers do not need an additional freewheeling diode, since the verticalMOSFETstructure itself contains a powerful body diode. This body diode is based on a pn diode, which in the case of a SiC device has a knee voltage of about 3 V.

You could now argue that in this case the conduction losses in diode mode are very high. However, we recommend – and that is state of the art for low-voltage silicon MOSFETs – to work in diode mode for just a short dead-time. This time should be between 200 ns and 500 ns for hard switching, and less than 50 ns for resonant topologies like ZVS (Zero Voltage Switching). The channel can then be turned on by applying a positive gate bias, which has the same advantage as in transistor mode on-state due to the lack of the knee voltage. Since the diode is a bipolar component, a small reverse recovery effect is in place as well; however, the total impact on switching losses is negligible.

Additionally, we recently introduced a650 V CoolSiC MOSFETderivative, to be deployed in a complete 650 V product portfolio. We intend this technology not only to complement IGBTs in this blocking voltage class, but also ourCoolMOStechnology. Both devices have fast switching and linear I-V characteristics in common; however, SiC MOSFETs enable body diode operation in hard switching, and at switching frequencies above 10 kHz.

Our trench-based SiC MOSFETs combine a low on-resistance with an optimized design that prevents excessive gate-oxide field stress and provides a gate-oxide reliability similar to that of an IGBT.

In the landscape of power electronics, both silicon carbide and superjunction MOSFETs have their applications

Compared to superjunction devices, they show a much lower charge in the output capacitance (Q_oss) in combination with a smoother capacitance vs. drain voltage characteristic. These features enable the use of SiC MOSFETs in high-efficiency bridge topologies like half bridge and CCM (Continuous Conduction Mode) totem pole. On the other hand, CoolMOS parts demonstrate their strength in applications where a hard commutation on a conducting body diode is either not present or can be prevented.

This sets the grounds for a successful coexistence of silicon carbide and superjunction MOSFETs in the voltage class between 600 V and 900 V. It is the designer’s decision which technology fits best to the requirements of their application.