Anatomy of a Solder Joint
Life in Assembly was easier with lead and Freon. Yes, Freon is a greenhouse gas that pollutes the atmosphere. Lead poisons everything else. Life after lead and Freon promises to be better for the Earthlings not concerned with Assembly of electronics.
The Eurozone is at the root of the lead-ban in commercial products with their RoHS initiative. California, China, and a few others have rules too. The leaded solders had a more manageable process window, but that ship has sailed. Eutectic Sn63 tin/lead solder is still around, but it comes with special disposal rules as mentioned. If "eutectic" is a strange word to you, it boils down to thawing and freezing over a narrow temperature band.
Lead lowers the melting point and softens the tin. Compounds like bromide and chlorine that incorporate halogens attack the surface contaminants to improve solderability. Those are some toxic elements though. We're going halogen-free these days. It seems like everything that works well has unacceptable side effects. Kind of like food, if it tastes good, it is probably bad for you.
Pure unleaded tin has a melting point around 230 C. That's high, but not terrible for PCB assembly considering a typical lead-free solder profile peaks around 250C. Meanwhile, some metals, tin included are prone to "whiskers" where ultra-thin barbs grow out of the solder joint and can cause shorts. Toyota learned about this the hard way. That said, tin whiskers and RoHS, in general, are a bit like Climate Science; lots of deniers in spite of the proof. We're still looking for an alternative.
A variety of alloys have been created and tried in the search for an ideal solution. Tin is still widely used, but it's alloyed with silver, copper, antimony and coming soon, unobtainium. Sn96 is a popular alternative according to Kester, the General Motors of solder suppliers. 96 is the percentage of tin with silver making up the rest. Copper contributes a half percent in many formulations.
Image Source: Author
To Clean or Not to Clean
Aside from the higher melting point of modern alloys, the biggest enemy of good soldering is oxidation. Let's set aside incorrect footprints. The green patina of the Statue of Liberty is nothing but contamination of the copper skin. (I like green better than orange anyway.) She wouldn't be solderable as-is. Her makeup would have to be stripped away. Flux is the weapon of choice against this corrosive foe.
Acid flux is the most aggressive and is used on copper piping among other things. It has to be thoroughly cleaned off, and that presents problems for PCB assembly.
The miracle cure is the same stuff that violinists and baseballers use to make their bows and bats tacky. Rosin is made from the sap of pine trees and becomes acidic at high temperatures but more benign after cooling off. Resemble anyone you know?
The residue it leaves behind can be removed with detergent and water. It's often washed off for inspection purposes or aesthetics or both. You have to get to that cleaning process immediately – or not at all. No-clean solder is the not-at-all version of this process. If you've tried to wash the dried up sap off of your car, you get the point.
Rosin is ever-so-slightly conductive so it can be a concern in some applications. You might be able to make out two rosiny pins to the right of the orange relay top center of the title pic. Fuzzy but it's my picture, my toy!
To the Board Room We Go
Acronyms abound for the coatings we put on the bare copper of the PCB lands. HASL, ENIG, ENEPIG, ENIPIG, OSP, oh my! The goal is always the same. Green is for soldermask, and shiny is for soldering. Copper and silver start tarnishing almost immediately. If you try to plate gold over the copper, the atoms don't get along, so a nickel barrier was added to create ENIG.
That brought on the infamous Black Pad debacle of the '00s. The relevant spec now calls for a thinner gold layer, and that helped. The industry is still spooked. Gold's noble neighbor on the periodic table, palladium was called in to stifle the plague of black pads. There are also white and purple plagues. Gold is good but not great.
Mad scientists and metallurgists are still blowing up the labs trying to find ways to join metals in reliable, repeatable, and environmentally sound ways. Getting the right amount of solder on each connection is tricky. Solder balls and bridging are the classic signs of too much. Voids or head-on-pillow are earmarks of insufficient solder. Impurities, humidity, non-coplanarity, dross, jerky conveyor movement, asymmetric pads, component shadows or being too close to the edge can all affect the soldering mojo.
Solder defects are parsed 100 different ways. There are so many ways to do it wrong. You usually want to cater to the most expensive component on the board when it comes to IR oven profiles, paste formulations, and stencil geometry. Then, there are the dainty ceramic and wire-wound packages that come apart just because. Uniform placement such that they go through the oven broadside rather than one lead at a time helps keep your delicate parts from breaking.
Pro tip: Don't flood over the pads without using thermal relief spokes. If you must, then change the soldermask opening to the size of the land that existed before the flood.
The other hard part is coming up with dielectrics that can withstand the lead-free temperature excursions. Board warpage, delamination, measling, pad lifting and so on are all problems exacerbated by higher temperature soldering. The hidden killer: excess heat prematurely ages the whole assembly.
There's no silver bullet yet. Stopgap solutions emerge and then become untenable with the next round of miniaturization. It's a vicious cycle.
No End in Sight
Until the industry succeeds, we have to stay up with whatever is clever. The current flavor is tin-silver with an OSP finish. If there's a spec for OSP, let me know in the comments. It's the thinnest coating around but has a short shelf-life.
ENIPIG is the potential up-and-comer of the moment. The price of gold is one of the drivers. The gold over palladium is way thinner than the gold over Nickel that you'd get with ENIG. There are others, so this is far from over.
The geometries for toe, side and heel fillets or circles for BGAs are immutable requirements. The stencils are a variable. We see a lot of tinkering along with different paste compositions. Every component vendor seems to have their own recommended thermal profile. Our action item is to stay abreast of the developments and update our notes and padstacks accordingly. So make a date with the assembly house rep. Take them out and share a pitcher of truth-serum. Get them talking about the problems they face. Your assembly yield (read bottom line) depends on it.