Failure Analysis of Semiconductor Devices – LEDs, Photocells, and Beyond
The most complicated semiconductor devices, such as microprocessors with millions of transistors, carefully constructed printed circuit boards, or the incredibly accurate silicon sensors known as MEMS, are frequently the focus of discussions of failure analysis services (micro-electro-mechanical systems). Yet in practice, the great majority of electronic parts are far simpler. These range from straightforward active parts like light-emitting diodes (LEDs) and power transistors to passive discrete like resistors and capacitors. For instance, almost every television remote control system makes use of a mix of LEDs and photodetectors to enable channel surfing. These kinds of components, despite their relative widespread use, are just as prone to failure as any other, and failure analysis can be just as helpful in helping to improve them.
While there is just one failure mode that can be observed in semiconductor LEDs—insufficient or nonexistent light output—failure analysis can be rather complicated. Photoemission microscopy is ideal for detecting flaws in LEDs for obvious reasons. What the naked eye sees as a lower-than-specified light output is frequently actually uneven illumination, as shown by the photoemission system (in other words, only part of the LED is lighting up, not the whole device).
A cross-section of the failed device may frequently reveal a junction that has been spiked with metal, generating a localised short, an ESD flash point creating a leaky junction, or even a processing error, which can all be the cause of this uneven lighting.
Read more: Emission Microscopy – Advanced Applications of Photoemission Systems
Similar to this, semiconductor photodetectors’ failure analysis frequently relies on utilising their special qualities. For instance, an oscilloscope may be used to monitor the output of the device while a pulsed light source is shone onto the surface of a photovoltaic die (often an array of diodes that creates a voltage when exposed to light). An analyst may be able to identify the specific sort of failure affecting the array by characterizing the photovoltaic output. For instance, an output that is stuck low may indicate something quite different than an output that reaches its nominal voltage but rapidly decays back to zero. To properly continue with the failure analysis at this point, an analyst must integrate experimental data with expertise and knowledge of semiconductor physics.
Although disassembling cutting-edge technology is undoubtedly the most exciting aspect of semiconductor failure research, the examination of more widely used devices, such as LEDs, is equally crucial. Tracking down an LED or photocell failure can be at least as difficult as locating a CPU failure.
