#homecmos IRC log for Sunday, 2011-07-24

solrizewow kristianpaul but i think you could use a converted SEM instead of that projection stuff02:18
kristianpaulwell i think thats step. but you should ask azonenberg :-)02:20
kristianpaulnext step*02:20
azonenbergsolrize: If I had a SEM lying around i'd use e-beam, sure02:21
azonenbergBut do you have one in your basement?02:21
azonenbergI know I dont02:21
solrizenoisebridge.net has a sem02:29
solrizethey're not that expensive02:29
solrizeold surplus ones02:30
azonenbergDefine "not that expensive"02:30
azonenbergLast time i checked, decently good ones sold for $20K ish02:30
azonenbergI can get access to them free or cheap at school for failure analysis, but IMO using them for processing is cheating02:30
solrizethe noisebridge one was probably donated ;)02:30
solrizehttp://benkrasnow.blogspot.com/2011/03/diy-scanning-electron-microscope.html02:31
azonenbergI read that02:31
azonenbergIts on my todo list02:31
azonenbergBut first i need a way of getting relatively hard vacuum02:31
solrizehttp://www.adafruit.com/blog/2011/07/05/diy-sem-in-operation/02:31
azonenbergAnd once i do that, my first project will be a filament evaporator for thin-film deposition02:31
solrizehmm02:31
azonenbergSince my current process requires evaporated copper02:31
azonenbergAnd i'm currently borrowing time on an evaporator in one of the labs on campus to do process development02:32
azonenbergBut i want to replace it with a homebrew unit soon (like end of year ish)02:32
solrizewow02:32
solrizewhere are you?02:32
azonenbergRPI02:32
azonenbergfirst year PhD student02:32
azonenbergthough i started the project while an undergrad02:33
solrizeneat02:33
azonenbergYesterday I did a pretty successful run using a triple-level mask process (photoresist over evaporated copper over sol-gel tantalum oxide) for wet etching patterns into Si02:38
azonenbergI had some undercut, which could be fixed by thinning the oxide layer, and some uneven exposure of the photoresist pattern (I still havent had the time to order a proper illuminator so i'm still using the mag-lite from the paper)02:39
azonenbergSo i didnt quite hit the 20 micron target, but my bigger features (maybe 30-50um across) resolved fine02:39
azonenbergAnd this is something i can work with, most of my previous runs failed horribly :P02:40
azonenbergAs a side effect i also demonstrated 20um patterning in evaporated copper, which can easily be shrunk to 10 or less based on the look of things02:41
azonenbergThat process worked beautifully02:42
azonenbergThe etch time was like five seconds :P)02:42
azonenbergMy first try at it used 60 sec and undercut severely02:42
nathan7azonenberg: \o/10:04
berndjlargely for my own reference: http://reprap.org/wiki/High_vacuum_chamber10:26
berndjazonenberg, also see http://reprap.org/wiki/MetalicaRap10:27
lekernelThe first image is of an IC that had way too much current passed through it. We expected to see just the usual destroyed substrate but instead were surprised to see this nice little fiber of silicon oxide that grew out the unfortunate device. The device became hot enough to actually oxidize the silicon and cause the fiber to grow instead of just explode. Shades of the movie "Alien".14:42
lekernelhttp://www.microwaves101.com/content/images/Mortuary/John/1%20-%20Blown%20IC.jpg14:42
nathan7cool.14:42
bart416Awesome15:16
lekernelcan we do this in air? http://en.wikipedia.org/wiki/HiPEP15:29
lekernel"a propellant throughput capability exceeding 100 kg/kW". looks rather good, no?15:31
lekernelbut "the pre-prototype HiPEP produced 670 mN of thrust at a power level of 39.3 kW ". uhm ...15:32
bart416ion thrusters in the atmosphere just seem like a bad idea15:40
bart416What you should really ask is if it works on oxygen and nitrogen15:41
bart416If the answer is yes, it should work in the earth atmosphere anyway15:41
lekernel_why are they a bad idea?15:46
lekernel_I don't see a problem with striking an ECR plasma of oxygen or nitrogen, but the increased pressure might be a problem15:48
lekernel_you can make such a plasma with an electrodeless system, so reactivity isn't even an issue (except that you might end up with toxic nitrous oxides in our case)15:49
bart416Well, air humidity might make the electric field problematic15:53
lekernel_ECR uses only microwaves15:53
lekernel_there are no multi-dozen-kV electrodes involved15:54
bart416oh, then it might just work15:54
bart416the general issue with ion thrusters is the way the charge gradient is created15:54
lekernel_yes, the idea is that ECR could make air fairly conductive, so this becomes less of a concern15:58
azonenbergWell, I think i have a pretty good idea of where to proceed from here18:55
azonenbergGonna try and set up another session on the evaporator and do some through-wafer etching18:55
azonenbergThe substrate for this run will be a quarter of a 2-inch <110> SSP wafer, coated in a thin layer of Ta2O5 after an RCA clean18:56
azonenbergAnd here i was thinking i'd never use all of the ten 2-inch cassettes i bought19:39
azonenbergOnce you start cleaving wafers and want to store the halves and quarters separately since they were all processed differently...19:39
bart416top gear is on19:39
bart416jeremy clarkson driving wrecking equipment19:40
bart416what could possibly go wrong19:40
bart416lol19:40
lekernelhttp://en.wikipedia.org/wiki/File:Electrocomposeur_1976.JPG19:52
azonenbergVery interesting19:58
Action: azonenberg wants e-beam direct write in his living room19:58
Action: azonenberg doesnt expect to get it for another 5-10 years at best19:58
azonenbergTired of guessing at layer thicknesses22:27
azonenbergSome time in a week or two i want to cross-section a few of my samples and see22:28
azonenbergMy processes are decently repeatable, I can see the thicknesses are even because its all the same color22:28
nathan7Hi azonenberg22:28
azonenbergBut without knowing the refractive index etc I have no idea what the actual value is22:28
azonenbergAnd knowing the index for bulk material doesnt usually help with thin films :P22:28
azonenberghi22:28
nathan7e-beam, the whiteboards?22:29
azonenbergI'd love to be able to do ebeam but its a ways out22:29
azonenbergWhen i build and/or buy (eventually i want both a nice commercial one plus a homebrewed one just for the fun of it) a SEM i plan to equip it to do ebeam22:29
nathan7oh, electron beam22:29
azonenbergSpecifically, electron beam direct-write lithography22:30
nathan7COOL22:30
azonenbergYou do a raster scan of the electron beam onto a polymer that functions like a photoresist (normally some kind of doped PMMA i think)22:30
azonenbergthen develop it just like PR22:30
azonenbergIts typically used for making photomasks22:30
nathan7interesting, interesting22:30
azonenbergSince you generally dont get precision alignment capability with it22:30
azonenbergBut you can make a mask and then stick that in a stepper, contact aligner, etc22:31
bart416I'm going to etch wafers using the LHC22:31
azonenberglol :p22:31
bart416I fear it might not meet the required energy to etch the wafer though22:32
azonenbergHmm, what other famous images would work nicely as test patterns besides the companion cube?22:33
azonenbergThinking of nyan cat22:33
azonenbergits already pixelated22:34
azonenbergother suggestions? Has to fit into a 500 lambda disk22:34
azonenbergwith no features smaller than 5 lambda22:34
azonenbergiow, you can make features with dimensions 5, 6, 7, ... 50022:35
bart416nyan cat would make you look like a fury though23:00
bart416maybe something symbolic23:00
azonenbergI've been doing companion cubes for too long23:01
azonenbergAnd i want to stay with the trend of memes23:01
azonenbergMake it look like a typo23:01
azonenbergI'm setting up this super-complex lab for making MEMS but somewhere along the line an extra letter got added :P23:02
azonenbergIn all seriousness the dots on his back are perfect for testing point resoluition23:02
azonenberghttp://i.imgur.com/Si3XY.png23:05
azonenbergDimensions are in microns assuming i use the 5um process, multiply by 4 for the 20um23:06
azonenbergall features are exactly 5 lambda square23:06
bart416How about an Si atom? :P23:07
azonenbergSend me mask art for one and i might do that lol23:07
nathan7\o/23:26
--- Mon Jul 25 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!