Thermal stress cracking can be completely avoided by CO2 laser cutting thin alumina sheets underwater. I also show how to formulate and apply silver paste, then sinter in a kiln to produce double-sided ceramic printed circuit boards with conductive vias.
60W CW CO2 laser at 80% power. 10mm/sec. Standard lens focal length (50mm). 2mm water above ceramic. 180 passes to cut through 0.75mm thick alumina.
Silver paste: 97% silver powder, 3% glass powder by mass. Particle size 1 micron or less. Add poly vinyl alcohol mold release until desired consistency reached.
Paste applied with 4 mil thick vinyl stencil. Dried in air 10 minutes, then rapidly brought up to 900*C, held for 10 minutes, then rapidly brought back down to room temperature. Total cycle about 45 minutes.
I measured electrical conductivity of the finished traces from my process with vinyl stencils: 4 milliohms per square at 10 micron final thickness. This is pretty close to the Dupont published spec ( less than 2 milliohm/sq at 16 micron thick)
Underwater CO2 laser cutting reference: https://sci-hub.se/10.1016/J.JEURCERAMSOC.2011.06.015
60W laser cutter: https://www.ebay.com/itm/294386493292
Ceramic sheets at McMaster: https://www.mcmaster.com/alumina/nonporous-alumina-ceramic-sheets-and-bars/
Ceramic PCB prototypes: https://ceramic-pcb.com/product/alumina-pcb-al2o3-pcb-prototypes-online/
Quickfire kiln: https://kruegerpottery.com/products/par_quikfire6
Raspberry Pi picoReflow oven controller: https://apollo.open-resource.org/mission:resources:picoreflow
Silver powder: https://www.ebay.com/itm/122525930519
Glass glaze: https://www.amazon.com/dp/B0044SCR6O
Dupont silver paste: https://www.dupont.com/content/dam/dupont/amer/us/en/transportation-industrial/public/documents/en/LF131.pdf
Binders for ceramic powder: https://digitalfire.com/article/binders+for+ceramic+bodies
PVA mold release: https://www.tapplastics.com/product/fiberglass/mold_releases/tap_pva_mold_release_liquid/67
Applied Science on Patreon: https://www.patreon.com/AppliedScience
It might be reasonable to add a bit of gallium o the paste. Gallium wets glass and would form an alloy with silver.
ReplyDeleteMost likely would also wet the glass particles from the glaze.
On the other hand, if wetting is sufficient, it might turn out that glass in the paste isn't needed any longer. Plus, the resulting alloy would have way lower melting temperature.
It's all a matter of experimentation. Some interesting reference to gallium alloys in dentistry: https://nvlpubs.nist.gov/nistpubs/Legacy/RPT/nbsreport4313.pdf
While looking at the Coefficient of Thermal Expansion (CTE) for alumina, higher temperatures have greater CTE, that seems to explain why heating it didn't work. It might be worth a try to cut the alumina sheet on dry ice. The CTE at dry ice temperatures is about half of what it is at room temperature. A possible draw back is that the thermal conductivity increases as you drop the temperature. I suspect that could be dealt with by running at a faster cutting speed.
ReplyDeletehttp://www.matweb.com/search/DataSheet.aspx?MatGUID=c8c56ad547ae4cfabad15977bfb537f1
Hey Ben! Would love to get in contact with you about an exciting opportunity. Please email me at mary@madeinnetwork.com for more details! Thanks so much.
ReplyDeleteHello. Sorry for wrong place but i could not find way to contact You otherwise.
ReplyDeleteCould You try making video about influence of alpha/beta/gamma radiation on at least few common chips?
Ideally i would like to see 22nm chips, but also common uC's like 328.
Greetings
Hey Ben, after reading these comments I feel a bit bad as it seems, that many people try to contact you through them and I hate to join them as it might be annoying.
ReplyDeleteI just need some technical advice and hoped you could point me in the right direction. I'm really into electronics repair and fault analysis and I now have multiple products (a keyboard and a smartwatch) where corrosion has damaged flex PCBs. The metal has been eroded from the foil carrier and I've been banging my head to try and find way to reconnect the traces. Soldering is no option unfortunately, because the foil seems too heat sensitive. Do you have any technology you could point me towards that's precise, low temp (ish) and stable with which I could reconnect these broken traces ? (I've tried conductive Ink pens, but they were VERY imprecise and had an extremely high resistance )
I'd be very grateful if you could find the time for a reply. Thanks
Peter
Peter, repairing flex circuits can be quite a challenge. For fine-pitch work, I'd recommend using a single strand of wire taken from a stranded cable. The goal is to replace the missing/corroded metal, so you'll need to use new fresh copper, and that comes from the strand. Connecting the strand to the remaining bits of circuit is probably best done with a tiny dab of solder paste and a very quick tap from a soldering iron. It takes a fair bit of practice, and even then, your particular flex may not have enough bulk left to survive the repair. You could also try 2-part silver conductive epoxy, but it is less strong and less conductive than solder. For much larger pitch work, like the flex inside a full size keyboard, 2-part silver conductive epoxy may be a good choice. You could also try conductive adhesive tape, or "Z-tape" called ACF in the industry. A lot of the skill for making these repairs comes from experience and practice with manual dexterity. It's rewarding when you manage to successfully repair a tiny fragile thing!
Delete