Intermetallics are a necessary evil in the metal-to-metal bonding world, which definitely includes soldering. There are two basic ways that metal will "chemically" bond to another metal: 1) solid solution 2) intermetallic. We will focus just on intermetallics for the moment as that is the most pertinent to the soldering world.
Many people confuse or interchange "wetting" for intermetallic formation (bonding). Wetting is just wetting. Just because a solder "wets" to a surface does not mean that an intermetallic "bond" has been formed. For example, and I have done this myself, 55.5Bi 44.5Pb can be melted onto a piece of copper. The molten BiPb will flow and "wet" to the surface of the copper. However, upon solidification (cooling) of the alloy, the BiPb can be peeled off. Why?... because no intermetallic was formed between the BiPb and the copper surface.
In order for an intermetallic to form, some amount of the surface metallization must dissolve into the molten solder. For this reason, Sn (tin) has long been a critical component of solder alloys. Molten Sn (tin) is an excellent solvent of many other metals. And, conveniently for us, those "many other metals" include elements like copper, gold, silver and, to a lesser degree, nickel. The rates at which these other metals dissolve into molten tin (solder) will differ. Gold dissolves readily into solder; whereas nickel does so slowly. So, because the rate of dissolution is different for each metal, the rate of intermetallic formation is also different. I have dealt with companies that have a long history of soldering to copper, and, for whatever reason, they are forced to switch to an ENIG (Electroless Nickel / Immersion Gold ) surface. (It is important to note that the gold layer is very thin and only applied to protect the nickel from oxidation. This gold layer readily dissolves completely into the molten solder and the "bond" is actually made to the nickel surface). When they make the change they sometimes encounter a number of issues such as incomplete wetting, poor bond strength, etc. and do not know why. They are not aware that the same reflow profile (time and temperature) that yielded a good (intermetallic) bond to copper is not sufficient to get the same intermetallic bond to nickel. Once they adjust their profile (more time and/or higher temperature) to allow for sufficient intermetallic formation , they are able to achieve acceptable solder joints. Keep in mind that dissolution, the phenomenon of a solid dissolving into a liquid, is effected by both time and temperature. Generally speaking, more time and more temperature allows for more dissolution and, hence, more intermetallic formation.
As mentioned in my opening line, intermetallics are a necessary evil. Why "evil"? Because they tend to be the most brittle part of the solder joint. Some intermetallics are more brittle than others. (This should be taken into consideration when choosing a solder alloy for a particular metallization). For example, intermetallics that form between Sn and Au are often extremely brittle. Being brittle, they can be subject to fracture, etc. This is a case where more is not always better. Yes, you need an intermetallic to get a "bond". Too thin of an intermetallic layer can be bad; but too thick of an intermetallic layer can be just as bad, if not worse. Believe it or not, the solder may not adhere well to its own intermetallic layer. Intermetallics are generally crystalline and chemically-stable structures....they do not really react with anything else once they have formed. If you have ever looked at a fractured solder joint, you may have noticed that the fracture likely took place right at the interface between the intermetallic layer and the bulk solder.
One other possible outcome of an excessively thick intermetallic layer is "voiding" at the interface. Why? Well, we first need to look at the reaction products. There are two basic types of reaction products that form the intermetallic layer between Sn and Cu. They are Cu3Sn and Cu6Sn5. In the f
irst case there are 3 Cu atoms to every Sn atom and in the second case 6 Cu atoms to every 5 Sn atoms. In both cases the Cu is being consumed faster than the Sn atoms. Because of this disparity in the reaction, in an exaggerated scenario, little holes or vacancies ("voids") can form in the copper surface.
Intermetallic formation is not only limited to the solder process. Metal atoms can diffuse even in the solid state. And that movement can cause the metal atoms to interact, react, and form intermetallics or cause the existing intermetallic layer to thicken. "Ageing" experiments are often performed to measure how much the intermetallic layer will change and what effect it will have on the mechanical nature of the joint.
It is well beyond the scope or purpose of this blog post to provide an exhaustive discussion of intermetallics. Whole books could be written on the topic. So, I am far from doing justice to the topic of intermetallics. I can only hope to shed a little light on the subject.
Comments or questions are very welcome.