National Physical Laboratory has completed a three-year project investigating the impact of thermal cycling regimes on the shear strength of lead-free solder joints.

The project undertook a comparison of accelerated test regimes for accessing the reliability of solder joints, in particular those made using lead-free solders. Samples were subjected to six different cycling regimes to investigate the effect of thermal excursions, ramp rates and temperature dwells. The most damage to joints per cycle was found to be caused by thermal cycling between 55-125°C, with a 10°C/min ramp rate and 5 minute dwells.

Similar degrees of damage in the lead-free solder joints were experienced from thermal shock regimes with ramp rates in excess of 50°C/min. However, these regimes, although faster to undertake, appeared to cause different crack propagation paths than observed with the thermal cycling regimes, although importantly still remaining within the solder.

Since this is a small difference, thermal shock testing may still be used to differentiate between, or enable ranking of, the effects of changes to materials or processes on the reliability of the solder joints. Hence, if a wide range of conditions are to be tested, a first sift can be completed using thermal shock, with the final work using more typical thermal cycling conditions. The difference between the SAC (95.5Sn3.8Ag0.7Cu) and SnAg (96.5Sn3.5Ag) solder alloy results across all types of cycles showed very little difference in rates of joint degradation.

Dr. Chris Hunt, one of the investigators, said, "One of the major findings to come out of this work is that the thermal shock cycles do produce a failure that is different to ordinary thermal cycles, but the failure is still in the solder, hence it gives us enough confidence to use it as a process to sort out reliability on a range of assemblies. It requires us to use the more usual thermal cycles which are slower ramp and dwell for final qualification, but the thermal shock is definitely a useful tool in getting through lots of different assemblies and evaluating & ranking reliability. The thermal shock cycle will typically be at least a third of the time, so you will get there more than three times quicker. So, if you are cycling, typically up to 2000 cycles, significant time can be saved—down from two months to two weeks."

For more information contact Dr. Chris Hunt: chris.hunt@npl.co.uk.

www.npl.co.uk

Copyright 2004, UP Media Group. All rights reserved.