Issue link: https://iconnect007.uberflip.com/i/1284035
14 SMT007 MAGAZINE I SEPTEMBER 2020 ating an attachment between the surface of the board and the surface of the component, creating intermetallics of dissimilar materials and different surface coatings while prevent- ing oxidation. This whole process has a lot of places where things can fail. Again, going from leaded to lead-free, the actual mechani- cal properties of the lead-free solder are not as forgiving as they were with the leaded solders. You have quite a variety of different lead-free combinations in the marketplace, and it's a lit- tle bit like baking a cake—a little pinch of this and a little dab of that. All these little things go in 1–2% quantities to improve the crystalline structure of the solder so that it doesn't have quite as much stress built into it that can cause failures in the field. Johnson: For me, with a software background, I was wondering how that was modeled. It's a physical laboratory experiment. Neves: Correct. In our lab, we have a 13-zone water-cooled reflow oven through which we will run boards and coupons, and other sorts of things. We have a HATS2 tester in the facil- ity that will run reflow simulations on cou- pons, representing the boards while capturing electrical measurements during the entire pro- cess. You can look at the resistance of the vias during exposure to reflow temperatures to see whether they might be separating at the high temperature. Many times, when the materials shrinks back down to normal size again after cooling, you make what appears to be a solid electrical connection again. If you separate a via or microvia due to the stress of the materials expanding at a high tem- perature, you can actually see this disconnec- tion electrically at the high temperature. But as the PCB shrinks back down again during cooling, you get a mechanical connection that appears solid electrically, and you may never see that separation again at lower-tempera- ture testing. The only time you're going to see that problem is at the highest reflow tem- perature when the vias are physically sepa- rated due to extreme expansion. You're start- ing to see that happen more on the PCB side your coupons in the preconditioning simula- tions before your reliability testing. This makes sense for not only interconnec- tion and isolation issues, but it also makes sense if you're doing high-frequency testing because the materials will change their elec- trical properties after going through the com- ponent attachment process. Most of today's materials have a higher glass transition tem- perature (T g ) to deal with lead-free processing, but these materials are not homogeneous and contain a mix of a variety of different resins and additives. Even if you have a 180°C Tg rated material, that's not made up of a single 180°C Tg resin system. If they were to do that, it would be very brittle. The material would crack. You'd have all sorts of other problems. You have a resin Tg mixture of 130–140s, and maybe 160s, mixed into this magical group of resins to make your resin matrix for your high Tg material. High component attachment tem- peratures, especially the lead-free tempera- tures, affect the low Tg resins in that matrix quite significantly and can change the dielec- tric constant (Dk) and the dissipation fac- tor (Df) of material in the PCB. Exposure to repeated high temperatures can change the resin matrices. When the matrices in the resin system start to break down your Dk and Df, along with it, your impedance will change. For high-frequency products, it's impor- tant to expose the material to whatever high temperature you're going to do before you do your final high-frequency test because changes to the laminate resin system in the PCB from component attachment processes can affect the result. The solder joints themselves— there's another very complex process of cre- When the matrices in the resin system start to break down your Dk and Df, along with it, your impedance will change.