Issue link: https://iconnect007.uberflip.com/i/1116895
MAY 2019 I PCB007 MAGAZINE 21 thing, analyzes it, and can predict from the de- sign where the copper is going to be fine and where it's not on the board. He can say, "You're going to overplate it here and underplate it here." Then, he can move the circuit around or mix jobs together to get a better overall balance and predict where the copper is going to be. The same kind of pro- cess could be applied, but it's more difficult when it comes to organic materials because if you try to X-ray the glass fabric, it's really hard to see it; it's not contrasting like copper. There might be a way of analyzing it with an MRI scanner or something, but it would be tough. Happy Holden: One thing my group at Hewlett- Packard was responsible for was characteriz- ing our laminates. We would characterize all of the prepreg and cores out to 22 GHz, but we would also do it at different temperatures and humidities. Most electrical engineers didn't seem to realize that these dielectic properties change with temperature and humidity. Ther- mal simulation at one temperature doesn't tell the whole story because the electrical proper- ties change with temperature, which means we at least need a minimum and maximum tem- perature. Especially in RF and high frequency for 5G, the higher the humidity and tempera- ture, the more the properties degrade. Morgan: It's absolutely critical. There are a cou- ple of angles. One thing is moisture absorp- tion. Moisture is very lossy and is highly polar- izable, so there's a very high dissipation factor. If you get moisture in the board, you have high losses right away. Moisture inside these prod- ucts completely kills them. Many of the prod- ucts have low moisture absorption, but others absorb a lot of moisture, so you must stop that. If you allow it to absorb moisture, you have a different set of properties loss-wise and tem- perature-wise than you had when they were dry, which is a massive problem. Some prod- ucts go through phase changes, such as PTFE, between room temperature and operating tem- perature. And that changes properties as well. Thus, what you measure at 20°C and what you measure at 120°C can be very different. The other big issue is oxidation of materials. Some materials oxidize in use, so they're fine when you start. But over time, the properties of the material change because the material oxidizes and you get higher loss. One of the HDPUG tests, for example, is that we take the materials, characterize the loss on the materi- als, go through six simulated reflow cycles, and measure it again. We try to find out whether it's changed or not. Holden: Because of 5G, is there a future for a material that's made from handmade fibers? Morgan: Potentially, and it has been done be- fore, but it disappeared from the market. There's not much sign of it being resurrected, but there might be a way. Screening conduc- tors onto material rather than using photosen- sitive, screening conducts might be a way for- ward. We may end up replacing copper with conductive materials, which could be gra- phene, silver, etc. Holden: A material that we do have is the Aji- nomoto film, which other people are thinking about creating. This material doesn't have cop- per on it for semi-additive, but a film material can be packed and laminated for conventional multilayers, not just semiconductor packaging. The Ajinomoto film uses permanganate, a dif- ferent concentration, to treat the adhesion pro- file for the semi-additive process. It has only a six-minute lamination cycle, so it's a convey- orized lamination and doesn't have to sit in a press for two hours. Morgan: That could have a place now because if you're using permanganate to texture the We may end up replacing copper with conductive materials, which could be graphene, silver, etc.