(719) 570-9549 [email protected]

 When examining a contemporary integrated circuit, an electronic failure analyst must face a myriad array of challenges; metal interconnects can be too dense for traditional isolation techniques to be of any value, critical dimensions may be too small to be thoroughly examined in any but the most cutting edge of microscopes, and layers are often spaced so finely as to make planar deprocessing a nerve-wracking, pulse-pounding undertaking, in which one slip of the finger can result in irreparable damage to a device undergoing analysis. As if these hurdles weren’t enough to contend with, analysts must also grapple with a rapidly expanding segment of the microelectronics market: semiconductor devices that incorporate moving parts as a part of their operation. These devices, referred to as MEMS (Micro-Electro-Mechanical Systems), offers a unique challenge from the standpoint of semiconductor failure analysis, largely due to their markedly different construction.

MEMS devices are most often used to provide an interface between the physical world and the electronic world – put differently, their largest niche is in the field of sensors. These sensors can be designed to react to many different varieties of stimulus – vibrations caused by a source of sound, for example, or changes in orientation or acceleration. This stimulus is translated into an electrical change, like an increased capacitance or a modulation of a resonant frequency, based on the way the physical stimulus acts upon the moving parts of the device. These moving parts may be interdigitated capacitive fingers of silicon, gossamer-thin silicon membranes, or any number of other delicate configurations. The inherent fragility of these devices naturally poses a challenge for any semiconductor failure analysis procedures; alternate approaches must be developed for decapsulation and inspection of any MEMS.

One of the greatest challenges in performing any sort of failure analysis on a semiconductor MEMS device is the problem of inspecting the device without inadvertently damaging it; since the device is not only sensitive to any sort of electrical stresses, but also to physical inputs, special care must be taken not to introduce any sort of undue mechanical stress to the device. Even common procedures like blowing debris off the surface of a chip with compressed air can be too harsh for some MEMS, causing the delicate sensor elements to snap and float away in the breeze. For this reason, it is often necessary to perform the vast majority of the work using secondary inspection tools like infrared inspection systems, which allow an analyst to look through the silicon lids common to so many MEMS devices, looking for problems like foreign particulate or stiction that may be interfering with the proper operation of the device.

With the wide variety of MEMS-based silicon sensors on the market, designed to detect an equally vast array of physical stimuli, it is difficult to write a straightforward procedure for performing any semiconductor failure analysis. The most vital elements in such an undertaking are a certain degree of creativity, a willingness to improvise, and a good base of experience to draw upon; fortunately, there are failure analysis labs with a wealth of all three traits.