Many semiconductor reliability studies involve subjecting a sample to conditions above and beyond those they would normally experience – higher voltages, elevated temperature and humidity, and so on. These accelerating factors cause devices to break down much earlier than they would otherwise, eliminating the need for lengthy test programs running over the course of months or even years. Studies such as this are often run until the devices under test fail; at this point, the parts are destructively analyzed in order to determine how and why the failure occurred. Failure analysis of these units can identify the weakest point in the device’s construction; a manufacturer can take this data and determine the best approach to improve their process and create a more robust device. This sort of improvement is often an iterative process; studies will be run repeatedly, and process changes made, until a device reaches the desired level of consistent reliability. Of course, in order to make the proper changes, a manufacturer must first have an in-depth analysis of the failed devices.
A failure analysis lab has a wealth of technology and technique that can be applied to the problem of semiconductor reliability. Cross-sectioning, deprocessing, and electron microscopy are all excellent candidates for studying the construction of a device, both before and after reliability stresses; for parts that have been packaged into plastic encapsulant, acoustic microscopy can detect defects like package delamination or “popcorn” cracking resulting from the intentionally harsh environment of a reliability study. Once the results of the study have been fully understood, a manufacturer can make changes to the device, changing processing parameters or choosing different materials in order to improve the performance of the part.
Semiconductor reliability is, for obvious reasons, an absolutely vital area of study; even consumer-grade electronics are expected to last for a few years at a minimum. The analysis of devices that have been subjected to reliability stresses is a crucial part of the process of continual improvement; time and money invested in well-designed reliability studies is well-spent, indeed.
Derek Snider has been an employee at Insight Analytical Labs since 2004, where he currently works as a Failure Analyst. He is an undergraduate student at the University of Colorado, Colorado Springs, where he is pursuing a Bachelors of Science degree in Electrical Engineering.