Dear Colleagues ..

I am facing an issue with the BGA part where in three locations we found separation between the BGA solder balls and pads on the PCB (ENIG finish).
This condition was discovered during the cross-section that we have requested from our local lab.
I can share few pictures I have taken from our report (can't attach the whole document as this report is confidential to our business).

Basically this BGA is the plastic ACTEL package 896 pins, SAC305 finish, soldered with leaded process using 62Sn/36Pb/2Ag solder paste. Temperature was slightly elevated to accommodate SAC305 solder finish from the BGA package.

The separation condition we have observed was only found in location A1, A16 & T16 which are the corners of the BGA package. As well all 3 separations occurred after the reflow process was completed because there is an evidence of completed inter-metallic fusion.  

In location A16 & T16 you can see the copper is protruding of the pad but I have been told this is not an issue and cause of the separation.
I was considering package or PCB warpage as the cause of this separation or something related to cooling the board after reflow.
I asked our lab to measure the height of the solder balls across row A & T and they all deem to be ok.
I have asked to check for the phosphorus content as the pad level to see if this is related to black-pad symptom and the phosphorus was measured as 10.3%.

So far the board assembly house is adamant their reflow profile is correct.
PCB manufacturer is confident with their PCB and pointing to the assembly house.
There is a thought about poor sample preparation where the separation issue would be caused by the lab (during cross-section).

We are attempting to perform another cross-section on different board but here I am not really sure if this is the right way to go (extra cost).
Because I have never seen such a defect, I am not sure what should be the next step to take.
Any advice guys ?

Andy