Scanning Acoustic Microscopy - Verifying the Results

There exists a certain level of distrust of images generated with scanning acoustic microscopy due to the inherently high level of interpretation necessary to correctly understand the image. While it is true that cracks, air gaps, and delamination can all be detected due to the phase inversion of the acoustic echo generated by the SAM, it is also true that so-called “false delamination” can be generated by acoustic impedance mismatches that have nothing to do with a defect. While a trained analyst can easily distinguish between these two cases, the possibility for an improper interpretation may leave some customers desiring more concrete evidence of failure.


Though destructive, cross-sectioning is often the most direct method for verifying the results of the acoustic analysis. Using the data generated by the acoustic microscope, an analyst can pinpoint a location for cross-section, polishing away the bulk of a device until the area of interest is reached. Studying this cross-section using an electron microscope will then provide a crystal-clear image of the defect. In the above image, cross-sectioning the device after scanning acoustic microscopy confirmed delamination at the encapsulant-to-paddle interface. It is interesting to note that the SAM was able to detect this delamination, even though it is only a couple microns thick!

While scanning acoustic microscopy is a useful tool for performing non-destructive analysis on devices, there is always the question of whether the results are accurate or are misleading. An experienced analyst should always be able to discern between a real defect and a false positive due to the construction of a device; however, if any doubt remains, there is always a way to verify the result (and showcase the delamination detection power of the SAM at the same time).


Read 5935 times


Need to Determine the Root Cause of a Failure in an Electronic Component?  We get back to you with a quote in 24 hours once we have your information.

Request Failure Analysis Quote