IGBTs can be modeled as a MOSFET controlled bipolar
junction transistor (BJT). The MOSFET provides a highimpedance
gate control, and the BJT portion provides low
ON-state resistance for the ability to conduct high current
density levels. In an NPT IGBT, a p-n-p BJT is formed from
a heavily doped p-type (p+) layer added to the collector side
of the basic MOSFET structure. The BJT portion enables conductivity
modulation in the lightly doped n-type (n−) epitaxial,
or drift region, layer with injection of holes from the (p+)
collector. The drift region supports the bulk of the depletion
region formed by the OFF-state collector voltage, but also
contributes a significant portion to the ON-state loss from the
finite drift region resistance [3]. As the IGBT voltage hold off
capacity increases, the thickness of the drift region generally
increases. Therefore, the drift region requires deeper ambipolar
diffusion lengths at high-level injection levels for a sufficient
carrier concentration across the entire drift region [2]. High
minority carrier concentration is typically present in the higher
voltage-rated IGBT drift regions during the ON-state to mitigate
the drift region resistance and forward saturation voltage. As a
result, injection and recombination of large amounts of charge
will result in slower turn-on and turn-off transitions. This is
especially evident in the decaying current tail during last portion
of the turn-off transition [4]..