The cause is most probably the lack of flux, provided all the reflow parameters are as per specifications. It is assumed that proper reflow temperatures and TAL are achieved. One company that I know of which makes non-contact film thickness measurement system is Lumetrics. The equipment from this company can also measure multiple layers.

The easiest system to use for flux inspection is the VPI-1000 manufactured by Easy Braid Inc. It features up to 400X magnification, on screen measurement and a user configurable defect library.

It might be flux, but much more likely you are not reaching reflow temperature due to the mass involved. Run a test with a thermocouple attached to the joint and verify your process.

It is highly unlikely that a dipping process is NOT providing adequate Flux. Depending on the nature of the joint failure it may be a lack of heat transfer, poor paste, contact surface finish or flux activity. All could be playing a part.

It’s hard to say for sure without seeing the location of the non-soldered joints and analyzing some of their characteristics but Amkor has done much work on POP and has published several papers on the subject.

Check out all of their white papers.

In particular, this white paper talks about the effect of die warpage and the use of a “dipping paste” to bridge the gap and overcome coplanarity issues. This dipping paste gives greater transcription vs. flux and can expand the tolerance band of the process.

Additionally, Amkor has done work on SOP (Solder on Pad) where the top pads on the lower package are bumped and also help to eliminate open joints.

I am truly NOT the expert on this but have spoken with Moody Dreiza once or twice and he is great. Suggest a call or mail to him for more help if needed.