Phil Zarrow: Brook, no clean fluxes. They're designed to balance the ionic nature of the flux. In other words, residue that's left behind is meant to be nice and benign, and not reactive. However, if they're not processed properly, they can actually contribute to electrochemical migration.
Brook Sandy-Smith: When we designed no clean fluxes, we do it under the assumption that there is going to a certain thermal excursion that the assembly is going to see. That thermal excursion is critical to using up any of the active ingredients for cleaning the metal oxides off the surface, but the flux is intended to do its job and be exhausted in that amount of time.
Phil Zarrow: Okay, so the thermal excursion, we don't reach the right temperatures. Essentially, we're going to leave behind a somewhat reactive residue then, with activators still contained in it.
Brook Sandy-Smith: That's correct. Actually, the method of applying the heat is really important, too. It's different whether you're using a reflow oven that has convection, and a lot of air moving around and a controlled amount of time, in a controlled pattern of heating through the zones; versus a selective soldering operation, where you have an IR preheat, and then you're going into a selective wave; versus a rework station, where you have IR heating, but then, as soon as you turn off the IR heating it cools down really fast. All three different ways of heating things up will really affect how much of the thermal excursion the flux sees and how consistent it is from assembly to assembly.
Phil Zarrow: Okay, so we're talking about possibly not having the appropriate profile for one reason or another on a reflow operation. Selective soldering. In this particular, where we're using fixtures, palatalized fixtures, for soldering a wave soldering selectively, a previously reflowed soldered board. We're using a no clean flux. The apertures on this pallet are designed to allow solder to come through, but they're shielding off the non-solderable areas. In other words, where we previously reflowed, both from heat, and from solder, but what about flux?
Brook Sandy-Smith: What about flux? When you spray the flux liberally, it's bound to get under those areas that are supposed to be protected from heat. That's a big hurdle, because those fluxes that were intended to see the heat and be rendered benign are now not seeing the heat that you expect them to see.
Phil Zarrow: Right. Right
Brook Sandy-Smith: And those fluxes are often left on the board, because it's the last part of the process. I've seen situations where flux can go rampant underneath one of these fixtures and intrude into vias, or interact with surface finishes on other parts of the board that are not supposed to be touched with flux. There are all sorts of problems that could come from having runaway flux essentially that's unheated and unchecked.
Phil Zarrow: Right. We've run into a number of situations where, when we did our failure analysis, that it was entrapped flux, or what we sometimes call flux creep, in these areas. Exactly what you're saying. It didn't see the proper thermal excursion, and the result was a somewhat active residue there. What happens later on, it sees those wonderful elements. Moisture. We have our residue. It sees power, electricity ...
Brook Sandy-Smith: Voltage.
Phil Zarrow: And voltage, and humidity, And, it shorts.
Brook Sandy-Smith: And there you go. You've got the electrochemical migration and you have field failures because of it.
Phil Zarrow: Exactly.
Brook Sandy-Smith: Now, wouldn't it be nice, if you could have tested the assembly, after all the processes have been finishes to assess in critical areas of the board, whether the proper thermal excursion had been met, and whether any unexpected flux ended up anywhere it's not supposed to be.
Phil Zarrow: Right. I would call for a localized method of cleaning, flux extraction.
Brook Sandy-Smith: Right. Right. Similar to the C3 method, which I just worked with Forsythe recently on a paper, looking into the interaction between solder pastes. This is back to reflow ovens, and rework stations.
Phil Zarrow: Okay.
Brook Sandy-Smith: Solder paste, and how they test in the C3 test, as well SIR testing, after different thermal excursions. There were three legs to this testing: First we looked at water-soluble flux. How big of a difference in the testing is there, if you don't clean it, versus if you do.
As expected, those results were not really surprising. The extracts failed really quickly.
Phil Zarrow: Yeah. It's pretty obvious.
Brook Sandy-Smith: There's a certain amount of time for the testing, for the conductivity to increase. They failed right off the bat because they're just so active.
Phil Zarrow: No big surprise there.
Brook Sandy-Smith: No surprise there. The interesting results that we saw were in the no clean fluxes. We looked at both lead-free and tin-lead fluxes.
Phil Zarrow: Okay.
Brook Sandy-Smith: I thought it would be interesting to look at tin-lead fluxes, because the thermal excursion is lower. The melting point of the tin-lead paste is lower.
Phil Zarrow: Right.
Brook Sandy-Smith: I mean, that's one of the differences between lead-free and tin-lead. So, with the lead-free fluxes, I looked at halogen-containing, and halogen-free fluxes, and different profiles. We used a rework station with an IR heating element, for most of the samples. We also then referred it back to reflow oven profile that was similar.
With halogen-free, no-clean, lead-free flux. All of the profiles passed, which was really exciting, because it was our Indium8.9HF, which we know is very stable and reliable. This just gave us that much more confidence that, even with varying heating profiles, Indium 8.9HF will maintain a benign residue.
The halogen contained flux that we used, passed almost all of the conditions, except for the shortest profile. We really were pushing the limit of, "This is too short."
If we didn't get any failures, that we didn't test it stringently enough, so we did get a failing result on that shortest profile with a halogen-containing flux. That is not to say that across the board, halogen-containing fluxes are less reliable.
It was just in that case, there was a limit there you could be creating a good solder joint, but not have fully reacted, benign flux residues afterward. In that case, the C3 and the SIR testing always matched up. They always corroborated the same results.
Where things really got interesting, was with the tin-lead pastes. Now I chose a more reasonably formulated tin-lead version, as well as a legacy tin-lead RMA that had been used for decades to make reliable assemblies that have lived out their lifetime in the field already. I wanted to see that way that, that integrated with the new test methods as well. Actually, the legacy material was the one that really had the most variation in whether it passed C3, and SIR, between the different profiles, so even with the same condition there were sometimes when you would have one that passed, and two that failed, or the other way around. The newer material was actually very consistent, down to the shortest profile. Now, when you really got to the nitty gritty, and the cool down was really short, and the time above liquidus was shortened. That's when you really started to see failures. In both cases, they're were process conditions that we could empart of material that passed the regular condition.
Phil Zarrow: Right.
Brook Sandy-Smith: That passed as re-flowed and oven condition, but then, by changing the preheat, or by shortening the cool down cycle, then you could render the flux not benign after the process, but still have solder joints.
Phil Zarrow: Very interesting. The full discussion of your experiment and your results are contained in a paper, which can be found...