Posted by Rivs: Sat Jul 07, 2007 8:26 am    Post subject: PQFN 3 x 3, 3.3 x 3.3 & 5 x 6 Package Crack After TC

Dear All,

Just want to seek your expertise help, We are encountering package cracking on our PQFN packages after temperature cycling of 100, 250, 500 to 1000 cycles. TC condition is -65 to 150 30 minutes per cycle.

We have not encountered this during our initial build (2006) and for verification we'ved evaluated both 2006 & 2007 build wherein package crack is only seen on the 2007 build parts and none on the 2006 build.

No materials nor process procedure changed since 2006.

Hope someone here can give us inputs on how we can identify the rootcause of this.

Posted by Paula: Mon Jul 09, 2007 10:16 am    Post subject: pqfn

Do you perform ATE test after TCT? Maybe it is ATE-induced.

Posted by Rivs: Mon Jul 09, 2007 10:23 am    Post subject: Re: pqfn

Hi Paula,

Thanks for the response. No we did not perform ATE nor manual testing. The main purpose of the evaluation (TC loading) is to verify the solder cracking (both on die and clip attach).

Posted by FARel Engr: Tue Jul 10, 2007 4:17 pm    Post subject:

Hi Rivs, request your FA guys to trace the origin of the crack and its general direction of propagation. This should give you an idea of the nature of the stress experienced by your devices. Hypothesize on what could be inducing these stresses, and conduct simulations to confirm or eliminate each hypothesis in terms of being the most likely root cause. By the way, do they exhibit any internal delam prior to TC?

Good luck to the team!

Posted by Rivs: Tue Jul 10, 2007 4:39 pm    Post subject:

FARel Engr thanks for the respone, actually I'm handling both FA & Reliability, package crack is originating from the leadframes half etch area, no delamination seen prior TC on the packages.

Posted by mel_c: Fri Jul 13, 2007 10:09 pm    Post subject:

Do extensive visual inspection using an optical microscope or SEM and focus on the crack pattern and propagation. After that, imagine the forces acting on the package that could result to this crack. You can also do cross-section to verify the origin of the crack. Was the crack induced thermally or mechanically? A CSAM data before and after TC can be of help also. If ever you CSAM can do QBAM or virtual cross-sectioning, much better. I hope this helps.
_________________
mel_c

Posted by Pauljnoonan: Fri Jul 20, 2007 11:43 pm    Post subject:

Also do you have any time zero units? I would be suspicious of the assembly process. It is likely that these crack could have been caused during assembly (perhaps during lead forming/lead shearing process). Also check the tool history on these machines if possible. It's normal for this type of package to have issues during assembly if the tools are worn. If this is the case you should see issues on time zero units. See if evidence of the beginning of this crack exists on time zero units. Mel c made some good suggestions above. if you could perform CSAM or virtual cross section on time zero units.