What failure mechanisms can be falsely induced by reliability testing?

EESemi.com FAQ

- questions most frequently asked by visitors of www.EESemi.com

Failures Falsely Induced by Rel Testing

Can a reliability test induce false failures?

Test methods to assess the reliability of devices are designed to make sample units fail by subjecting the parts to conditions that accelerate specific failure mechanisms. What the reliability engineer must understand though is that reliability testing, if not performed properly, can induce false failures not representative of the true reliability of a population of devices. Needless to say, a reliability engineer must be aware of what these false failures are and how they can be prevented.

Burn-in, with the ability to subject parts to high temperature in the presence of electrical excitation, has long been used to accelerate oxide failures and transistor degradations. Burn-in, however, requires extensive handling of the samples, since these need to be loaded onto and unloaded from burn-in boards. Burn-in can therefore induce false failures related to loss of electrical contact due to lead deformation. Burn-in can likewise result in EOS issues, i.e., when the burn-in board or a power supply used is questionable.

A few of the industry-standard reliability tests such as HAST and THB subject parts to moisture, and therefore require deionized (DI) water as utility input. Unfortunately, the inadvertent use of contaminated water as input can result in metal corrosion in the samples, whether at package- or die-level. Thus, corrosion issues found after reliability testing that affect both the qualification and control samples must prompt the reliability engineer to check the DI water used, just to make sure that the corrosion failures were not simply due to the input water used.

ESD is another failure mechanism that can be falsely induced by the reliability testing itself. The preparation and handling of samples, if not done in an ESD-preventive manner, can cause the samples to fail by, well, ESD. Thus, ESD controls that are implemented in production lines must also apply to the reliability lab.

Excessive intermetallic growths and excessive oxidation of metal surfaces, on the other hand, can result from excessive environmental chamber temperatures. An equipment or set-up problem that results in a higher temperature inside the chamber than what's required will falsely accelerate the mentioned failure mechanisms. Discoloration of the samples are obvious signs of exposure to excessive temperatures, but this may not always be present. Failures due to excessive intermetallic formation must therefore trigger the reliability engineer to check if the oven used is providing the correct thermal stress to the samples.

These are just some of the common failure mechanisms falsely induced by the reliability testing itself. Every reliability engineer must therefore consider the possibility of failures that are falsely induced by the rel test itself whenever conducting investigations on qualification or reliability monitor failures.