#homecmos IRC log for Wednesday, 2011-10-12

learningcnot so empty here04:51
azonenbergNot at all, its picking up04:51
azonenbergslowly but surely04:51
azonenbergwe now have several people doing lab work04:51
azonenbergi'm focusing on bulk micromachining of <110> Si by KOH wet etch using Cr hardmask04:51
azonenbergas well as metal patterning04:52
azonenbergswkhan is doing research elsewhere on ALD techniques (in a real lab) but is interested in our work04:52
azonenbergB0101 is starting out with josephson junctions04:52
azonenbergashaw wants to build a laser direct-write lithography rig04:52
azonenbergthose are the main efforts i know off the top of my head04:52
azonenbergcheck out the die pics on the googlecode page04:53
learningcoh so you are most here from the same university or...?04:56
azonenbergNope04:58
azonenbergi'm at RPI04:58
azonenbergashaw is in australia04:58
azonenbergB0101 is in singapore04:58
azonenbergand swkhan is in california04:58
azonenberglol04:58
learningccool, I just saw hobbyist-friendly at the google page05:00
learningcI can make the circuits at home too05:01
azonenbergThats the idea, yes05:01
azonenbergThere will definitely be some investment in optics etc05:01
azonenbergand some basic chem knowledge will be required05:01
azonenbergbut we want to keep it simple, of the magnitude of PCB etching05:01
azonenbergbuild your own single layer PCBs at home? Fine05:01
learningchow much does it cost in tools and equipments investment though?05:01
azonenbergBuild your own 2-metal ASICs at home?05:01
azonenbergFine too05:01
azonenberglearningc: We havent gotten it working yet lol05:01
azonenbergTechniques are gradually evolving05:02
learningcwhat kind of investment in optics? what kind of microscope?05:03
azonenbergDepends on the process size you want05:03
azonenberg200um is doable dirt cheap05:04
ashawBy the way, The imaging device I am designing has a cost of about 700-100005:04
azonenberg20um costs a bit more05:04
azonenberg350nm will not be cheap but may be feasible eventually05:04
ashawI am targeting 350nm05:04
azonenbergashaw: A bit high but it will hopefulyl be one of the most expensive parts of the setup05:04
azonenbergAnd lets see how it turns out05:04
ashawthat is 12st rev05:04
azonenbergRemember you need not be able to make masks at that size05:04
azonenbergI envision building a stepper at some point05:04
azonenbergwith maybe 4x reduction05:05
azonenbergget some nice olympus microscope objectives05:05
ashawstepper?05:05
azonenbergor mitutoyo05:05
azonenbergyes, to do multiple dies on a wafer05:05
ashawoh, yeah05:05
azonenbergusing a single mask05:05
azonenbergThats much further out05:05
ashawor, as I go directly05:05
azonenberginitially i'm using my cheap microscope as a projection aligner since it gets the job done05:05
azonenbergashaw: not sure if you heard when i said a little while ago but i just got two new blank wafers05:06
ashawthe DMD device I have has a resolution of 7.6uM05:06
azonenberg2 inch <110> both sides polished05:06
azonenbergtomorrow if time permits i'll be depositing a micron of Cr on both sides of one05:07
ashawthe other expensive thing will be the CMP machine, no?05:07
azonenbergCMP? No, i think that could be made cheaply05:07
azonenbergdepending on the precision required05:07
azonenbergit could share parts with the spin coater05:07
ashawprobably not actually05:08
azonenberglike i said i'm targeting 20um05:08
azonenbergat 350nm things get a bit harder :p05:08
ashawas the wafer must be mounted upside down05:08
azonenbergCMP is on the roadmap but i dont need it yet05:08
azonenbergSo i havent put much work into it05:09
learningc2 metals asic? what type of metals involved?05:09
azonenbergi've been developing processes as they become necessary05:09
azonenberglearningc: 2 metal layers05:09
azonenbergas in, one for the gates and one interconnect layer05:09
azonenbergis a bare minimum05:09
azonenberg3 is prefrable05:09
azonenbergand probably filament evaporated aluminum05:09
ashawmy goal is CMP with electrolytic copper05:10
ashawpersonally05:10
learningcusing this technique to build transistors?05:10
azonenbergashaw: the thing about copper is that you need a barrier layer and everything05:10
azonenbergand adhesion layers05:10
azonenbergsputtered Ta will work05:10
ashawno, that is front end of line05:10
azonenbergi was thinking damascene actually05:10
azonenbergfor when i got to copper05:11
azonenbergsol-gel SiO205:11
azonenberglitho with negative of metal pattern05:11
azonenbergsputter Ta then Cu without breaking vacuum05:11
azonenberginto the trenches05:11
learningcwhat are the semiconductors used?05:11
azonenbergthen CMP until its flat05:11
azonenberglearningc: silicon05:11
learningcdoped with?05:12
azonenbergThe ones i have are mostly P type B doped05:12
azonenbergbut i've been doing mostly MEMS05:12
azonenbergso i dont really care05:12
ashawprobably p doped05:12
azonenbergMy handful of CMOS masks are for an N-well process on P type wafers05:13
azonenbergbut there's plenty of room for change between now and then05:13
azonenbergin any case its 0100 here and i need sleep lol05:13
azonenbergKeep the discussion going, i'll be back in the morning05:13
ashawlearningc05:14
ashawWhat are you interested in about this?05:14
learningcI'm exploring the ideas only05:14
ashawmy LDI head will work with PCBs too :)05:15
learningcand see how far we can make our own devices05:15
ashawsame machine05:15
learningcLDI?05:15
ashawLaser direct imagnig05:15
learningcoh, I have no access to a laser though05:15
ashawBluray laser05:16
learningcof those05:16
ashawdiode based05:16
learningcwhat's the power of these?05:16
ashaw300mW05:16
azonenbergYou dont need much05:16
azonenbergthats actually overkill05:16
ashawyeah05:16
azonenbergyou need aroudn 200 mJ/cm^2 for my photoresist05:16
ashawhuge05:16
learningcyou can go far with 300mW05:17
azonenbergso 20 mW would need 10 sec for a 1cm^2 die if you could get it all onto the die05:17
azonenberg300mW would need under a second and would actually probably have to be PWMed05:17
ashawyou would probably need 1mW05:17
learningcbut I suppose the optics is critical?05:17
azonenberglearningc: yes05:17
ashawand driving the DMD device05:18
learningcor do you use the same optics in the bluray instead?05:18
ashawno different05:18
ashawvery05:18
ashawmore like a digital projector05:18
ashawbut using a bluray laser as a light source05:18
ashawand not expanding it05:19
learningcdid you do any prototype of the laser imaging device?05:20
ashawI am designing it now.05:20
learningcwhat about others?05:21
ashawand will build it soon05:21
ashawnot a new idea05:21
ashawused for a long time in PCB industry05:21
learningcbut it's a challenge if you DIY05:21
ashawI am scaling it down, in cost, and in feature size05:21
ashawyeah05:21
learningcis it like a cnc machine?05:22
ashawthe main movement stage is, yes05:22
learningcand what kind of resolution are you targetting?05:23
learningcor precision05:23
ashawI want to image at 500nm05:24
ashaw10 allow a 1um process05:24
azonenbergashaw: Hmm16:49
azonenbergYou want to put the laser head on an x-y table?16:50
azonenbergI thought you were going to do single 1024x768 or whatever masks16:50
azonenbergYou will have alignment issues since your stages will probably not be that accurate16:50
azonenbergI mean, it'd be fine at 5-10um scales probably16:50
azonenbergBut going smaller will not be easy16:50
ashawI intend to have a movable head, but before magnification22:11
ashawazonenberg22:11
azonenbergh/o busy22:12
learningcazonenberg: which microchip wifi module do you recommend since I'm believing you have a bit of experience with them?22:31
learningcazonenberg: never mind, only 2 versions one with antenna build-in and the other external22:46
learningcI think I'll go for the external one22:47
--- Thu Oct 13 201100:00

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