|
MOSFETs have become extremely useful as high power electronic switches in a range of applications, including power supplies and motor control. While the cause of any failure may appear very similar externally, significant differences will exist internally which can help determine the root cause.
Failures in MOSFETS tend to result in short circuits between pairs of terminals – usually the source and drain. Externally the failures may appear very similar, but there are usually significant differences internally. A detailed inspection can give assistance in determining the cause of failure, and the likely magnitude of the causal parameters.
Uncovering the evidence
The most obvious variation is the extent of the damage. This depends on the current that can flow during the failure but also the duration of the event. Once a breakdown has occurred at a single point, the heat generated spreads out and creates a larger and larger area of damage as shown in Figures 1 and 2.
Figure 1. MOSFET with a small point failure
near a terminal on the chip |
Figure 2. Another MOSFET with a larger area of damage. |
In extreme cases the semiconductor chip can become cracked, as in figure 3.
Figure 3. A MOSFET with a crack through the chip between the arrows
Ultimately the damage may only stop when the current is cut off. A difficulty that can add to the problem is that fuses and contact breakers take a finite time to respond, so the time that the damage is being produced may be longer than the initiating event. This difference can obscure the original event.
|
The position of the damage gives further information on the cause such as excess temperature or current, or over-voltage spikes. On occasions the event may be so short and the damage so small that it is invisible – as is often the case with Electrostatic Discharge (ESD) damage.
Under some circumstances the position of the fault can be found by a liquid crystal technique. When the liquid crystal exceeds a critical temperature it changes phase and a black spot appears as shown in figure 4. |
Figure 4. A hot spot on a MOSFET marked by a ring and located by a liquid crystal technique.
|
| 
Figure 5. The ringed area in figure 4 after partial etching of the surface layers, revealing a small point of damage.
|
This sample was partially etched to reveal the site of the damage beneath (figure 5) thus substantiating this as the location of the failure.
|
For confirmation a sample of the deposit was sent for biological culturing and micrococci bacteria were found to be present; these act aerobically to form sulphuric acid in environments containing sulphur. The oil used in these coolers was a natural mineral oil that would allow such bacteria to thrive. A switch to synthetic oil for the coolers was therefore recommended to alleviate the problem.
So what was the cause?
It may be most tempting to blame a failure on a poor semiconductor device which can be the case – but this is often not the root cause as the external circuit design and environment need to be taken into account.
The Reliability and Failure Analysis group at Cobham Technical Services has decapsulated and examined hundreds of MOSFETs from a variety of applications and can give assistance in assessing possible causes of faults. For further information call us on +44 (0)1372 367444 or email.
|
Diagnosing failures in MOSFETs
