Thomas,

Yes, this ttype of phenomenon has been observed previously by anyone who
has done temperature/humidity/bias (THB) test on bare and assembled
PCBs. Look at  Figures 6.62 (a) through (d), page 223, chapter 6 in
'Failure Modes and Mechanisms in Electronic Packages', published by
Chapman & Hall ( available at http://www.wkap.nl, Kluwer Academic
Press).  Via to via, plane to via and line to line migration growth
shown are from real product failures. Pages 222- 233 discuss the
migration failur in more details.

Without looking at picture of your specific failure(s), it is difficult
to discuss your root cause. In general if you have - 1) moisture, 2)
ionic contamination and 3) voltage,  there is high probability of
migration  to occur. The mechanism goes something like this-

at anode:     meta l> metal(+) + ne(-)     and     H2O > 1/2 O2 + 2H(+)
+ 2e(-)
at cathode:    metal(+)  + ne(-) > metal     and    H2O + 2e(-) > H2 +
2OH(-)

Metal ions at cathode start accumulating and grow like a dendrite
towards anode causing an electrical short.




Thomi wrote:

>  Dear Technetters, has anyone come across a root cause for electric
> shorts between neighboring copper lines of different potential within
> a layer laminated with RCC (resin coated copper)? We have carefully
> ground away layers 1+2 from one side, and layer 4 from the other, to
> find a (ca.) 1 micrometer diameter filament connecting two neighboring
> lines in layer 3. Cross sectioning this further, we could analyse this
> filament by using SEM/EDX and found copper. The surrounding resin
> shows peaks of Bromium (probably from flame retardants) and Aluminum
> (??Al2O3 filler material??). The electric short was not detected
> during bare board testing, but occured within one week of functional
> testing of the full assembly in an environmental chamber operating at
> 120°C (I wouldn't believe it!). Is there a certain resin condition
> under which copper migration is enhanced within the RCC-resin (however
> close to the interface with the core to which the RCC is laminated)?
> There is a recommendation to ensure driving/baking out humidity from
> the multilayer, which makes me somewhat suspicious, but I don't know
> enough about the constituents (resin, hardener, flame retardant, maybe
> filler material) to try and explain what happens
> electrochemically. Best regards, Thomas Ahrens, Memellandstr. 8,
> D-24598 Boostedt
> Tel. ++(49) (0) 4393 97769   e-mail [log in to unmask]
> Bitte beachten - please note:
> - Mails groesser als 2 MB werden nicht angenommen
> - messages larger than 2 MB will not be downloaded

--
____________________________________
Regards,
Pratap Singh

Tel./Fax: 512-255-6820,
e-mail: [log in to unmask] or [log in to unmask]