I have had two experiences with shorting under the capacitor, and the
adhesive was a contributing factor(at a previous company).

1.   A dual dot pattern was used to secure the component to the PCB.  We
switched from an adhesive developed for low volume dispensing to an
adhesive designed for high speed dispensing.  Due to the different
chemistries and viscosities, when the new adhesive was used and the
component placed on the adhesive, the adhesive kept its two distinct dot
shapes, it did not always get "mashed" together.   ( With the previous
adhesive, when the component was placed, the adhesive "flowed" together to
make one big mass.)  This left a small channel between the two dots, under
the component.  During the wave solder process, flux would be sucked under
the component, between the dots (due to capillary action).  The solder
would then also wick under the component, following the flux.  Usually a
small amount of solder was left under the component, and this shortend the
distance required to create a short under the component.    For the short
to occur, a number of variables have to fall into place.  This problem
occured on a random basis, but was usually limited to the same components.

For the short to occur, the circuit has to have a voltage applied.  The
solder under the component decreased the distance between the ends of the
component or pads for the short to grow, the remaining flux under the
component,, moisture from the board, or outside humidity, will provide the
medium for a small dendritic growth to complete the short.  If the
component is shifted on the pads, it decreases the distance required for
the growth also.  Like I said, this was a random occurance, it took us a
while to solve the issue.    We solved the problem  by changing to a single
dot pattern (but see the next experience below).  Another solution is to
put a dummy trace between the two pads, under the component.  The dummy
trace and the dual dot pattern acts like a dam.  Several SMT books discuss
this solution.

2.   About a year after solving the above problem, switching to a single
dot, the problem occured once again.  The root cause was voids in the cured
adhesive dot and the size of the dot.  If the dot was slightly oversized,
and was touching one of the pads, during the wave solder process, solder
would wick into the voids  from the molton pad and travel under the
component due to the capillary effect of the voids.  As stated above, this
shortened the distance required to create a short under a component.  When
the dot size was such that it touched both pads, we had actual shorts.
Once again, this occured on a random basis.

The voids in the adhesive occured due to improper storage; it froze, laid
on its side. etc.  We also discovered a certain amount of moisture inherent
in the PCB, especially those from overseas (solved by requiring sealed
desicatted shipping packages and proper inhouse storage).  We also switched
to a improved adhesive from the same company that was designed to be less
hydroscopic (obviously others had seen the same problem and they had
developed a new formulation).  The cure profile was also a contributer.
Instead of curing quickly at a high temperature(150?C, we cured at a lower
temperture 130?C for a slightly longer time.  At the high temperture, it
was determined that the entrapped moisture in the PCB was flashing to
steam, and sometimes the vapor path from under the solder mask was through
the adhesive dot.  This created voids in the glue as the adhesive cured.
The lower temperature allowed the moisture to escape while the adhesive was
still soft, so it would eventually solidify into a solid dot.  However, the
key improvement was reducing the size of the dot so it never touched the
pad.  You can have voids in the adhesive, but if it never touches the pad,
the solder will not wick into the dot.

General Observation:
The capacitor short could be caused by a number of reasons.  In this case,
several variables worked together to create the problem on a widely random
basis.  The voids in the adhesive, the dot size under the component, the
shifted component and the environment.  The solder wicking under the
component shortened the distance required for a dendrite to grow.  In this
case, the humid environment, the contaminants on the board or under the
component, the shortened distance and a constant voltage provided the means
for a dendritic growth under the component.  For your information, you can
take a capacitor, apply 12 volts across it and place DI water to the top,
and eventually, you will create a dendritic growth (try it, it is fun to
watch)

Hope this helps, you can contact me for more information if required.  John
Maxwell (SMT Consultant) helped us solve this issue several years ago.  He
monitors this forum also.

Ed Holton
Hella Electronics
[log in to unmask]

##############################################################
TechNet Mail List provided as a free service by IPC using LISTSERV 1.8c
##############################################################
To subscribe/unsubscribe, send a message to [log in to unmask] with following text in the body:
To subscribe:   SUBSCRIBE TECHNET <your full name>
To unsubscribe:   SIGNOFF TECHNET
##############################################################
Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional information.
For the technical support contact Dmitriy Sklyar at [log in to unmask] or 847-509-9700 ext.311
##############################################################