IGBT Modules - Technologies, Driver and Application (Second Edition) - page 55

on the effectivemagnitude of the voltage, this point may be further left (lower voltage) or
right (higher voltage). From the time t
, the current falls to the value of the tail current.
When the time t
is reached, the remaining charge carriers are removed beginning from
the collector side by recombination.
presents the individual steps involved in
the removal at times t
to t
. The time taken until all the charge carriers are finally
removed can involve values of several µs and emphasises the length in time of the tail
If the IGBT is turned off, i.e. there is either a negative voltage or the value 0V at the
gate, and the voltage at the emitter ismore positive than at the collector, the IGBT is in
reverse blocking mode and the pn-junctions J
and J
are blocking. The pn-junction J
on the other hand, is conductive. In this case the blocking behaviour is determined by J
and J
, with J
taking most of the voltage. As in forward blocking mode, in order to
achieve sufficient blocking strength, the n
-base must be wide enough and be
sufficiently low doped. When the different IGBT technologies are compared, an NPT
IGBT has approximately the same reverse blocking capability as forward blocking
capability. This is different for PT and FS IGBTs, inwhich the reverse blocking capability
is less than the forward blocking capability, i.e. the blocking behaviour of an NPT is
more or less symmetrical, while PT and FS IGBTs display asymmetrical behaviour. As
described in further detail in chapte
the reverse blocking behaviour is not specified
by themanufacturers.
As can be seen in the equivalent circuit diagram in
an IGBT has an inherent,
parasitic thyristor, which is formed by the two BJTs. To prevent the IGBT inadvertently
turning on (i.e. latching up), especially at high operating temperatures, certain
countermeasures must be taken when manufacturing and designing an IGBT, so that
the parasitic npn-transistor does not turn on. One way to do this is to short circuit the
base of the npn-transistor (p-doped zone i
with the emitter (n
-doped zone in
via the chipmetallisation, which ensures no or only a very small voltagemay
be build up between the baseand the emitter of thenpn-transistor insufficient to cause it
to turn on. This measure is further supported by, for example, partially increasing the
doping of the p-zone, which has a low ohmic resistance and therefore reduces the
inherent voltagedrop, anddesigning the emitter n
A latch-up can occur in static operation, i.e. when the IGBT is already turned on, and
dynamic operation, i.e. when the IGBT is turning off. In both cases, the current flowing
through the IGBT at the time is the relevant parameter that determines whether or not a
latch-up occurs. The higher the current, the greater the risk of a latch-up. However,
because of the measures mentioned above, almost all modern IGBTs in the specified
current range are free of latch-up effects. It should be noted that this specification is
done usually only to twice the rated current. Furthermore, even though latch-up free
operation can be assumed in most cases, this is no longer covered by the
manufacturer's specification. A specification becomes relevant again for currents that
occur during short circuit operation.
In addition, IGBTs, due to their internal design, have numerous parasitic capacitances,
some of which are independent of voltage, i.e. constant, and others which are
dependent of the collector-emitter voltage applied. Capacitances formed by the silicon
dioxide (SiO
) insulation of the gate against other layers are generally regarded as
voltage-independent (C
between the gate and the chip metallisation, C
between the
gate and the n
-base, C
between the gate and the p-channel and C
between the gate
and the n
-doped emitter area). The remaining capacitances (C
as semiconductor
capacity between the surface and the n
-base, C
as semiconductor capacity between
1...,45,46,47,48,49,50,51,52,53,54 56,57,58,59,60,61,62,63,64,65,...548
Powered by FlippingBook