Issue link: https://iconnect007.uberflip.com/i/1309864
NOVEMBER 2020 I PCB007 MAGAZINE 13 the etching side, you can't plate without hav- ing a plan on how you're going to etch it. It's very common that people blame the etcher for everything, but that's not the case. With our equipment division, we're now making our own plating equipment and etching equip- ment. We're taking advantage of a lot of the things we've learned with third-party equip- ment over at GreenSource. You'll see some very interesting things at the next productronica. Rick can talk about the etching piece a little bit and the uniformi- ty that we've developed. We're taking our plat- ing concept with the vertical line and making a second revision of it. That is being installed at one of our clients in Massachusetts. This is going to be very interesting for the market, in general. All the high aspect ratios that we've been able to achieve will be generally available from us at the next productronica show. Matties: This work is going through, and you're talking about all these different processes, ty- ing it into etching. You built a state-of-the-art factory, but how much hand-holding of a proj- ect do you have to have, or is it really done through solid process control and automation? Stepinski: There are a couple of pieces. You have to be able to measure and characterize the process. The biggest opportunity that we see is we have to move away from the elec- trical characterization of copper plating thick- ness, and we have to move away from metallur- gical cross-sections for plating thickness. Met- allurgical cross-sections don't provide enough sampling points to properly assess complicat- ed designs, so you have not enough degrees of freedom to properly assess what the thickness is everywhere. Traditionally, people have used eddy-current and CMI testers to characterize the surface. For larger areas, the problem with this is with the large number of vias that you have in complex designs. You end up measuring the filled vias and not the surface. It comes down to incor- porating other techniques to accurately char- acterize the surface with low cycle time. Be- cause you're not going to do 100 cross-sections per panel, that's totally unreasonable to do a setup. We have developed alternative metrology tools to get around cross-sections and eddy- current testing that are accurate sub-micron. This has been qualified and is being industri- alized by our automation group. Our plating lines will include measurement of surface cop- per, measurement of all diameters, how much mass of copper has been added, monitor- ing, and process control by part number with these different metrology techniques. This is a big one. People ignored plating metrology for decades, and we are going now and using new techniques that the market has not seen before but that are used very commonly in other applications—just not with PCBs. Matties: Are these inline measurements in real-time? Stepinski: Yes. When you make the recipe, you program where you want your measurements taken, and then you measure everything. Even if you do partial plating, and take it out before it's done because you don't want to scrap the first feeds, you don't want to test for every- thing and say, "I guess I scrapped it. I over- did it." Instead, you go step by step, do part of the cy- cle, measure it, characterize it, and then adapt. It's important to do this because then you can develop feedback loops and have some AI. In the next year, we are focused on putting this in place for plating and etching. A fully automat- ed etch process means it measures itself dur- ing the process and adjusts itself for the final