500e
If as BM is saying these could be due to micro bubbles, the way I read the literature, it does not go down to the level that is starting these problems.
The size of these bubbles is small, and individually they are much smaller than the tolerable defect size determined during certification testing. However, collectively the sum total of the voids can well and truly exceed the TDS. In the case of extensive micro-voiding the adhesive surrounding any disbond may have a shear strength of only 50% of the prisitine adhesive bond. Further, the reduced strength may result in fatigue of the small segments of adhesive between the bubbles, leading to disbonding. A further complication is that when an OEM undertakes certification tests they use a teflon insert in an otherwise pristine bond, where the inserted "dsibond" is surrounded by good adhesive. The same approach is used for damage tolerance analysis using computer modelling, where the properties of the adjacent adhesive are unaltered for the purposes of modelling. Most of the literature considers only on this type of large voids (macro-voids).
Hence the certification basis for the original damage tolerance analysis and the computer modelling are compromised if the adhesive has extensive micro-voids.
I am preparing a paper for a conference in Brisbane in July on this and other risks to safety as a result of the application of damage tolerance analysis to adhesive bonds. I'll post it on my web site after the conference.
Regards
blakmax