| --- Mon Feb 4 2013 | 00:00 | |
| nmz787 | azonenberg: http://ma.ecsdl.org/content/MA2012-02/59/3990.full.pdf | 10:30 |
|---|---|---|
| nmz787 | azonenberg: Low cost UV laser direct write photolithography | 10:30 |
| nmz787 | system for rapid prototyping of microsystems | 10:30 |
| azonenberg | Nice | 10:30 |
| Sync_ | that's neat | 10:32 |
| azonenberg | I want to build one | 10:32 |
| Sync_ | just lugged my mill in the cellar | 10:32 |
| azonenberg | something like that has been on my TODO for a while | 10:32 |
| azonenberg | just been too busy | 10:32 |
| Sync_ | I hopy my back does not kill me tomorrow | 10:32 |
| azonenberg | i think if/when i build one i'll add machine vision software to the camera | 10:32 |
| azonenberg | so it can automatically align the new pattern to existing alignment marks | 10:33 |
| Sync_ | should be easy enough to do with opencv | 10:33 |
| azonenberg | Yeah | 10:35 |
| azonenberg | Because if i could solve the litho problem | 10:35 |
| azonenberg | and get a vacuum chamber for metal deposition | 10:35 |
| azonenberg | then MEMS are well within reach | 10:35 |
| azonenberg | CMOS will be a lot harder due to contamination concerns | 10:35 |
| azonenberg | but a little K- ions won't hurt a comb drive :p | 10:35 |
| Sync_ | aaah ok, my back is already killing me | 10:39 |
| berndj | azonenberg, is contamination an issue regardless of feature size, or does it become less critical if you ran a 100µm process? | 19:46 |
| Sync_ | contamination is not an issue of feature size | 19:50 |
| Sync_ | particle contamination is | 19:50 |
| Sync_ | but metal ions always get you | 19:51 |
| azonenberg | berndj: Metal ion contamination will kill transistors no matter how big they are | 23:49 |
| azonenberg | But particles less than, say, half your feature size | 23:49 |
| azonenberg | aren't that big a deal | 23:50 |
| berndj | but i assume it's at least a matter of ion concentration, and not absolute number? | 23:50 |
| azonenberg | I believe so, yes | 23:50 |
| berndj | i see. so reagent purity is directly the relevant factor? | 23:51 |
| azonenberg | Yes | 23:51 |
| azonenberg | As well as cleanliness of glassware etc | 23:51 |
| azonenberg | But for MEMS you dont care about electrical performance nearly as much | 23:52 |
| azonenberg | since the Si is just a mechanical substrate | 23:52 |
| berndj | use giant glassware to take advantage of scaling factors :) | 23:52 |
| azonenberg | Which is why i think they're a better starter project | 23:52 |
| berndj | why do you suppose MEMS came only much later in the history of the tech? | 23:53 |
| azonenberg | That i'm not sure | 23:53 |
| berndj | my first instinct would be to assume that it's in some sense "harder" than ICs (i'm assuming you're using "CMOS" loosely here, and that NMOS would have the same sensitivity) | 23:54 |
| berndj | maybe it just didn't seem that interesting earlier than it became mainstream? | 23:55 |
| azonenberg | Yes, although i hear BJTs are slightly less sensitive | 23:56 |
| azonenberg | i dont understand the physics to that extent | 23:56 |
| berndj | i don't recall jeri ellsworth using crazy-level purity reagents? | 23:57 |
| berndj | i think she used conductive epoxy to make contacts and even gates. very ghetto-style! | 23:58 |
| azonenberg | Yes, and thats why i filed this under the "to learn" level | 23:59 |
| azonenberg | From what i understand metal contamination at room temp isnt as big a deal | 23:59 |
| azonenberg | the big concern is keeping things clean before you do diffusion | 23:59 |
| --- Tue Feb 5 2013 | 00:00 | |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!