Failure analysis of consumer electronics can pose a wide variety of challenges, due to the multitude of different failure mechanisms that might befall a device. Environmental factors, mistreatment, and even the way that the device is packaged can contribute to the untimely demise of a device. While the vast majority of integrated circuits are packaged using a plastic or epoxy based mold compound, some high-reliability devices – especially those used in aerospace applications – are encased in hermetically sealed tombs of ceramic and metal. Performing electronic failure analysis of these hermetic packages poses a new set of challenges, as there are certain failure mechanisms and tests that are applicable only to this type of packaging.
Generally speaking, hermetic packages are sealed under a dry, neutral environment, to prevent the ingress of contaminants that might reduce the device’s reliable lifespan. Ideally, this seal is a completely impermeable metallic weld, preventing even the smallest molecule of unwanted material from entering the die cavity and wreaking havoc with the integrated circuit within. However, it should be no surprise that these seals do not always approach ideality, and may allow gases to seep into the cavity over time. If water vapor or other harmful contaminants can penetrate the cavity, a device’s lifespan may be drastically reduced; as a result, it is very beneficial to have a way to test hermetic packages for potential leaks when performing electronics failure analysis of these devices.
Gross and fine leak testing can be performed in a multitude of different ways – using radioactive tracers, relatively inert gases like helium, fluorocarbons, or any number of different materials – but the basic method is the same. A device is placed in a chamber that has been pressurized with the tracer of choice and allowed to “soak”, to give the tracer time to wend its way into the device cavity. The device is then removed from the chamber and exposed to a detection mechanism, keyed to look for the signature of the particular tracer used in the test; a radioactive tracer might be detected with a Geiger counter, for example. Gross leaks are detected by immersing a device in the fluid after the pressure soak; as the tracer escapes the device cavity, a stream of bubbles issues forth from the site of the breach.
Even if the device has been thoroughly welded shut, there may still be trouble inside the device cavity. Small metallic particles, fragments of substrate material inadvertently chipped off during the packaging process, or other materials may be sealed inside the hermetic cavity along with the integrated circuit. These particles are, in many cases, only one bump or jostle away from creating catastrophic failure, shorting pins and bouncing off the glassy surface of the semiconductor die. These particles can, in many cases, be so microscopically small or constructed of a material such that they are difficult to detect with x-ray imaging; in order to verify that a particle is present, then to extract it for analysis, requires specialized tools.
The most common way of identifying the presence of such particles for electronics failure analysis is a Particle Induced Noise Detection (PIND) system. The system consists of a platform capable of dealing short, sharp, shocks or sustained vibrations to a sample (in order to jar the particle loose from any crevasse it may have wedged itself in) and a sensitive transducer (similar to a microphone) that picks up even the softest of sounds. The suspect part is placed onto the PIND tester and is subjected to a series of shocks and vibrations, shaking the particle and bouncing it off the walls of its metallic prison. With each impact, the sound is detected by the transducer and amplified many times over; the transducer output is displayed as an oscilloscope trace and is also played on a speaker for the analyst’s listening pleasure (since particles are often possessed of an impeccable sense of rhythm). Of course, simply verifying that there is a particle inside the cavity is not enough; an analyst must also be able to extract and identify the particle, in order to determine where in the process it was introduced. To this end, a hole is punctured in the device lid, and covered with tape. The device is returned to the PIND tester and vibrated until the particle can no longer be detected, at which point it has most likely bounced onto the tape. The tape is then removed and the particle is inspected.
While leak testing and PIND are two methods of finding failures in a hermetic package, they are far from the only analytical methods available. An analyst’s greatest asset is versatility; while the tests detailed here are targeted specifically at finding defects in hermetically sealed packages, there are many other techniques in the analytical toolbox that can be adapted to find these types of failures with equal success.