Most of you have probably seen videos of countless Mentos/Diet Coke experiments all over the internet. If you are one of the few who have not, this is a good site to teach you the basics: Steve Spangler Science. Essentially what you get is an explosion of foam due to the rapid outgassing of the soda. I am still not sure why diet soda works better than regular soda. If you know why, please comment or email me at tjensen@indium.com.
What do Mentos and Diet Coke have to do with voiding you ask? Truthfully, not much but it is just one example of things that shouldn't be mixed (unless you are intentionally looking for a mess). Mixed alloys is another example. This is particularly true when you mix Sn/Pb and Pb-free for BGA assemblies. Today, the most common mixing is using Sn/Pb solder paste but using a Pb-Free bumped BGA. There are concerns about overall reliability of these mixed alloys, but the most common problem people have encountered is high amounts of voiding. This comes as a direct result of people trying to improve the reliability by raising the Sn/Pb reflow profile to around 225-230 C and allowing complete mixing of the two alloys. This definitely gives a more homogeneous microstructure, but most Sn/Pb pastes weren't designed for that high of a reflow temperature. High peak temperatures result in more flux outgassing and, therefore, more voiding.
To avoid the excessive outgassing, you could eliminate mixed alloys. Everyone would like to do this, but it is often not possible. Therefore, you best option is to select a solder paste that is more thermally stable and has a high oxidation barrier. This will reduce the outgassing at the elevated peak temperatures and allow you to focus on diet coke outgassing rather than that of the solder paste.