A note to those who already know this... I've just learned this and I thought it was interesting enough to post -- I'm not claiming to have discovered it!
A common statement is that "mosfets are not subject to secondary breakdown". It's one I've made multiple times and is one that appears in a lot of authoritive material.
However, after considering second breakdown in diodes, I wondered if you could indeed have second breakdown in mosfets (since, at the very least, they have a body diode).
Second breakdown in a reverse biased diode is an effect which occurs after avalanche breakdown. Essentially, after some time, the device breaks down again to a lower voltage. The cause is essentially thermal, with heating causing a defect which partially or completely shorts the device -- read the document linked to above for a more complete discussion of the discovery and explanation of the cause.
Second breakdown can occur in the forward or reverse direction. In some respects, I would consider the shorting of a diode due to massively large forward current to be caused by secondary breakdown (does anyone disagree?).
The cause, as I understand it, comes down to the fact that at higher temperatures the breakdown voltage is lowered. Because the current through the junction is heavily influenced by the breakdown voltage, reducing it slightly leads to a significantly higher current. This tends to cause the power dissipation , rather than being spread out over the entire junction, to be more and more concentrated. If not controlled, the rise in temperature increases to the point at which physical damage occurs. BJT's having junctions, are subject to this, but since mosfets do not...
If mosfets are used in linear applications and the Vgs is very slightly higher than Vgs(th), the channel is very close to being pinched off at the source end. While the resistance of the bulk material has a positive temperature coefficient, Vgs(th) has a negative temperature coefficient. This means that a slight unevenness in current flowing through the narrow region can cause localised heating which will reduce the Vgs(th) at this point. This condition is potentially unstable, leading to thermal runaway. The result is that the current flows through a narrower and narrower channel until the ohmic heating results in damage to the semiconductor. Here is one application note which covers the region of thermal instability in the mosfet SOA.
A common statement is that "mosfets are not subject to secondary breakdown". It's one I've made multiple times and is one that appears in a lot of authoritive material.
However, after considering second breakdown in diodes, I wondered if you could indeed have second breakdown in mosfets (since, at the very least, they have a body diode).
Second breakdown in a reverse biased diode is an effect which occurs after avalanche breakdown. Essentially, after some time, the device breaks down again to a lower voltage. The cause is essentially thermal, with heating causing a defect which partially or completely shorts the device -- read the document linked to above for a more complete discussion of the discovery and explanation of the cause.
Second breakdown can occur in the forward or reverse direction. In some respects, I would consider the shorting of a diode due to massively large forward current to be caused by secondary breakdown (does anyone disagree?).
The cause, as I understand it, comes down to the fact that at higher temperatures the breakdown voltage is lowered. Because the current through the junction is heavily influenced by the breakdown voltage, reducing it slightly leads to a significantly higher current. This tends to cause the power dissipation , rather than being spread out over the entire junction, to be more and more concentrated. If not controlled, the rise in temperature increases to the point at which physical damage occurs. BJT's having junctions, are subject to this, but since mosfets do not...
If mosfets are used in linear applications and the Vgs is very slightly higher than Vgs(th), the channel is very close to being pinched off at the source end. While the resistance of the bulk material has a positive temperature coefficient, Vgs(th) has a negative temperature coefficient. This means that a slight unevenness in current flowing through the narrow region can cause localised heating which will reduce the Vgs(th) at this point. This condition is potentially unstable, leading to thermal runaway. The result is that the current flows through a narrower and narrower channel until the ohmic heating results in damage to the semiconductor. Here is one application note which covers the region of thermal instability in the mosfet SOA.
