Unipolar and mixed power semiconductors (part 1)

In unipolar semiconductor devices, current transport is provided solely by the majority carriers of the large, lightly-doped "base" region that gives devices their voltage withstand capability. This base region is therefore not modulated in conductivity, so the performance "compromise" offered between blocked voltage and current flow is a priori less favorable, for the same crystal surface, than for bipolar devices.

On the other hand, in the absence of minority carrier storage phenomena, unipolar components are intrinsically faster. Two related advantages should be highlighted:

• very high input resistance of controlled devices (transistors), thanks to the possibility of controlling the flow of majority carriers by field effect (junction field effect, Metal-Oxide-Semiconductor field effect);

• high lateral thermal stability under direct polarization, due to the negative temperature coefficient of carrier mobility; this stability enables the production of high-current components with a large active surface area, by parallel integration of elementary cells.

The possibilities of unipolar and bipolar effects appear complementary. Combining them in hybrid component structures offers additional degrees of freedom when it comes to arbitrating performance trade-offs between voltage withstand, current capacity, switching time and losses. The most industrially important example is the IGBT (Insulated-Gate-Bipolar -Transistor), but many other "mixed" devices of varying potential, which have appeared over the last two decades, are based on this approach.

The division into two articles of the analyses relating to unipolar and mixed components is motivated solely by space constraints. They are in fact a whole.