Issue link: https://iconnect007.uberflip.com/i/1527613
OCTOBER 2024 I DESIGN007 MAGAZINE 43 istry. I'm etching down, and I'm also etching laterally. In general, on the thinner foils, you're etching half the thickness inward (reduction in width) to the thickness of the copper. If I'm etching through 18-micron foil, I'm etching 9 microns sideways, just at the top, because as you etch down through that 18 microns, you end up with a trapezoidal-shaped circuit. If I have 18-micron foil and I want to make an 18-micron circuit, I will etch the top of that trapezoid. It will be etched away. at doesn't work. Even if I'm doing a 25-micron trace and 18-micron copper, that doesn't leave me a lot at the top; it becomes almost like a pyramid at that point. at's a big problem. is is where semi-additive processing comes in. I can't put chemistry through a solid, but I can put light through photoresist. I'm making a trough that I will fill with plating, because I can expose a straight sidewall with light, which I can't do with chemistry. You know, a lot of people conflate mSAP with sub- strate or interposer, where you might be doing 6- or 8-micron features in really thin copper and everything is very small. But it changes everything when we're making a full-size PCB. ere's a technology space that's not being met today. I feel that most shops in the U.S. could do 25 microns reliably using the right materi- als. I have customers who do it. Dack: When you say it's doable, you mean with the right materials, mainly copper foil? Especially copper foil. A couple of things have converged lately that make this much more feasible today than it was even five years ago. We have better micro-thin foils on peelable carriers. e release technology has improved a lot. I know this first-hand because I have cus- tomers that want to do fluoropolymer-based dielectrics. e photoresists have improved. As I mentioned earlier, most exposure units now are capable of 25-micron resolution. Actually, a lot of them are capable of even 12-micron resolution. ose pieces of equipment have improved. Shaughnessy: It's a perfect storm, but in a positive way, not like the movie. Yes. Now we have new dielectrics where the neat resin has low CTE, low loss, and low dielectric constant. Now the laminates have become better. Of course, what's driving some of the need is the component, the BGA foot- print. All these things have come together now today, and this is a lot more feasible. Two of our Bay Area customers have done multiple part numbers where just the BGA fan-out has 25- or 30-micron features. ey're not doing hundreds of boards; they might do a lot of a dozen boards, but their customers are happy. Dack: With your customers using these processes, do they have to provide you with any special fab notes or communication, or do they just hand you the stackup and you see what you can do? I have a step-by-step procedure that I share with my customers that they'll utilize for their shop. It's really very straightforward. e big- ger issue is keeping track of the job, so that it doesn't go through the regular etch process by accident. You don't want the operator to strip " I have a step-by-step procedure that I share with my customers that they'll utilize for their shop. "