Werner

Gee, I've gotta speak up regarding the grain size even if am up against
established knowledge agree with you that a fine tin grain structure leads
to a longer lifetime of a solder joint.  It is true that lap joint specimen
in a cyclic shear test do live longer if the tin grains are smaller in the
whole solder layer. However, I am sceptical about the terms used generally
about the grain size.
As you know, it is difficult to etch solder with copper around ( to be
honest I never could do it ). As an effect the distance between the lead
phases is regarded as a measure for the tin grains. This is the point I do
not agree at all. We did a lot of tests now on that ( don't wanna count the
microsections I did ) and what we found is that with thermal ageing and,
with some restrictions, cyclic load the lead phases do grow bigger. True.
The distance between the lead phases does grow bigger too. True again. But
when polished the right way you will see that the grain size of the tin
does not grow. There are more tin grain between the lead phases. On the
other hand we found new solder joints consisting of 2 or 3 giant tin
domains with the fine lead phases embedded. To verify that we did FIB
(Focussed Ion Beam) microsections across the boundaries of the domains and
we didn't find any sign of additional grain boundaries. The same thing we
found in lap shear specimen. However what we observed is the breaking up of
these large domains into bands approx. 20um wide along the interface to the
supporting structure. And we also saw that shear strain concentrates in
these bands with refined structure. When we made shear specimen with fine
tin grains the deformation was over the whole width of the solder layer.
Since we applied always the same deformation, the strain was much bigger in
the case where the refined bands formed thus causing an earlier failure.
What we saw from experiments is, that rapid cooling causes mainly large
domains. Slower cooling gives small tin grains ( large lead phases ). The
slowest cooling we did was soldering two copper plates ( 2cm x 3cm x 2mm )
together and let them cool down in an oven that was heated on 200 deg. and
turned off after the specimen was in. It took more than 10 hours for the
specimen to cool down to 50 deg.. What we found was large lead and fine tin.

For Cristophe: Regarding real solder joints I have seen all. From fine
grained ones to giant tin domains. But as Werner said ( I agree for a
change ) within a year in service it would be very difficult do distinguish
one from another. Thus I don't see a sense in giving a specification
besides the difficulty in the metallographic preparation.

The NiSn intermetallics can be made visible if you polish the microsection
with 1um paste on a soft cloth. This gives you a step from the pretty hard
intermetallic layer to the soft tin. Well visible in light microscopy maybe
supported with Differential Interference Contrast ( DIC ).

FIB and DIC I only added to bring up some new acronyms nearly nobody knows.

Have a great day

Guenter

##############################################################
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.
If you need assistance - contact Gayatri Sardeshpande at [log in to unmask] or
847-509-9700 ext.5365
##############################################################