#homecmos IRC log for Wednesday, 2011-10-19

--- Wed Oct 19 201100:00
gkwhcHi, anyone here use "electric" to design ICs?19:12
azonenberggkwhc: Last time somebody asked, no19:57
azonenbergmost of us here are in the process deevelopment stage19:57
azonenbergnot ready to do full vlsi19:57
gkwhcazonenberg: oh i see. would you happen to know of a channel related to this area?19:57
azonenberggkwhc: This is pretty much the microfab/vlsi/mems channel on freenode19:58
azonenbergswkhan (not here atm) works in a real lab at nasa19:58
azonenbergmost of the rest are hobbyists19:58
gkwhcazonenberg: great!19:58
azonenbergpersonally, i'm working toward a MEMS comb drive19:58
azonenbergand then 4000 series logic19:59
azonenbergin my living room19:59
gkwhca home-fab?19:59
azonenberggkwhc: look at the channel title lol19:59
gkwhcthat's really neat19:59
azonenbergIt's still under active development and we do not yet have anything working that well20:00
gkwhci can't imagine the equipment costs, etc20:00
azonenbergI actually havent spent that much on it20:00
azonenbergi mean, my whole lab is maybe a few tens of K but most of that is the GPU cluster that isnt used for this at all20:00
azonenbergit was funded by some sponsored research a few years ago20:00
gkwhcah i see20:01
azonenbergthe nice microscope was $1k and that is the biggest single item i actually use for semiconductor stuff20:01
azonenbergprobably 5-10 total? I mean, it might be a little much for the average hobbyist who isnt being backed by sponsored work20:01
azonenbergBut for a hackerspace?20:01
azonenbergentirely feasible20:01
azonenberghttp://colossus.cs.rpi.edu/~azonenberg/images/homecmos/2011-09-20/die_i4_009.jpg20:02
gkwhcyea20:02
azonenbergThis is 20 micron half-pitch wires in copper on silicon20:02
azonenbergthe background mottling is the incompletely etched chromium adhesion layer20:02
azonenbergI did the metal deposition in an evaporator on campus but will be buying a bell jar and vacuum pump soon so i can do metal evaporation at home20:02
azonenbergits a very simple process if you can get down to decently high vacuum20:02
gkwhchm what do you use for etching? laser?20:02
azonenberghttp://colossus.cs.rpi.edu/~azonenberg/papers/litho1.pdf20:03
azonenbergThat's the exposure process20:03
gkwhcthanks!20:03
azonenbergi use the camera port method20:03
azonenbergthen the metal etch is HCl : H2O220:03
azonenbergnice simple wet etch20:03
azonenberghttp://code.google.com/p/homecmos/source/browse/trunk/lithography-tests/labnotes/azonenberg_labnotes.txt is my lab notes20:04
azonenbergwhich i actually have to update after my last (failed) experiment in Ta2O5 + Cr + Cu film patterning20:04
gkwhci am surprised that you can manually etch  20micron wires!20:05
azonenberg20 is actually doable with very good yield20:05
azonenberghttp://colossus.cs.rpi.edu/~azonenberg/images/homecmos/2011-09-20/die_i4_002.jpg is the entire test pattern20:05
azonenberga little edge roughness but 20um wires at 20um half-pitch all resolved beautifully20:05
gkwhcoh yes definitely20:06
azonenbergthese were both imaged in a JEOL JSM-840 scanning electron microscope on campus20:06
azonenbergi use the SEMs for my nice PR shots or for debugging when i need really high resolution20:06
azonenbergi can only use them when the lab manager (who i have a good working relationship with) is around and nobody else needs the tool20:06
azonenberglol\20:06
azonenbergSo for all of my in-process imaging i use the AmScope metallurgical light microscope20:07
azonenbergat home20:07
azonenbergnot nearly as sharp but the images are in color20:07
azonenbergand quick turnaround20:07
gkwhcinteresting!20:08
azonenberghttp://colossus.cs.rpi.edu/~azonenberg/images/homecmos/2011-06-07/fab002_annotated.jpg was one of my early lithography tests20:08
azonenbergthis is around 5 microns of SP24 photoresist over a tantalum oxide thin film20:08
gkwhcoh wow - 5 microns!20:09
azonenbergthe Ta2O5 is transparent (yellow collor from diffraction), the reddish spots are pinholes in the film20:09
azonenbergthats thickness20:09
azonenbergthe feature size is around 2520:09
gkwhci'll need to read more about the subject area20:09
azonenberghttp://colossus.cs.rpi.edu/~azonenberg/images/homecmos/2011-06-07/fab004_annotated.jpg20:10
azonenbergthats a nicer shot with the amscope20:10
azonenbergi actually have two cameras, one of them is calibrated but a pain to set up (i need to haul my laptop out to the bench to connect to it)20:11
azonenbergthe other is a point and shoot that i hold up to the eyepiece, good for quick in-process imaging if i dont care about exact dimensions20:11
azonenberghere's an image from that cam http://colossus.cs.rpi.edu/~azonenberg/images/homecmos/2011-08-06/S7301603.JPG20:11
gkwhcthe resolution of that is impressive!20:12
gkwhcwhat camera do you use?20:12
azonenbergThe calibrated camera is an amscope MD190020:13
azonenbergcheap optics, a nikon/zeiss/mitutoyo would be way better20:13
azonenbergbut for $1200 camera + scope it was a great price20:13
azonenberga mitutoyo FS-60 (my dream scope) would run over ten grand20:13
gkwhcthats really expensive, but what you have currently looks great already!20:15
azonenbergYeah20:15
azonenbergThe other camera is a $50 samsung point and shoot20:15
azonenbergthats the one i took the nyan cat shot with20:15
azonenbergPuns about "nyanotechnology" were too good to pass up lol20:15
gkwhchaha true!20:15
azonenbergso when i got bored of my other mask art i made this one20:16
azonenbergAnd he is truly a nanoscale cat20:16
azonenbergaround 600 microns x 300, which is big20:16
azonenbergbut only 200nm deep20:16
azonenbergfrom top of copper to the surface of the silicon20:16
azonenberg(approximately, that was the intended thickness when i ran the deposition but I havent actually done profilometry on it)20:16
azonenbergIn the next day or two i am going to be doing Cr evaporation on a wafer20:18
gkwhcexciting!20:18
azonenberglift-off patterning of 800-1000nm Cr over a photoresist pattern20:18
azonenbergto be followed by a KOH wet etch20:18
azonenbergif successful i can do a through wafer pattern20:18
gkwhcabout how long would the whole process take?20:18
azonenbergwell, i am going to be doing several dies by lift-off as well as a half-wafer to test etching of the Cr20:21
azonenbergThe evaporator on campus is in the mat sci department, it takes an hour or so for the vacuum pump to warm up20:22
azonenbergthen maybe 20 minutes to pump down the chamber, two or three for the actual deposition20:22
azonenbergspin coating photoresist on a batch of dies is maybe 5 mins20:22
azonenberglithography is 5-10 minutes per die20:23
azonenbergmetal etch is fast20:23
azonenbergthen the through-wafer etch is a good couple of hours20:23
azonenbergbut i can remove the die after 10 mins or so to see if its working20:23
gkwhci see20:30
gkwhcthats pretty fast20:30
azonenbergThe evaporation is something i do a big batch (usually half or a whole 2-inch) every few weeks since i just need films to pattern and they last me a while20:30
azonenbergi'm not doing any multilevel metalization yet20:31
kristianpaulI'll buy a scope as soon i see this work can be asilly reproducesable and sourcable :)20:31
azonenbergkristianpaul: We're working on that20:31
kristianpaullast pic looks nice, where came froma ctually?20:31
azonenbergi'm still hoping to have a comb drive by end of year, i have most of the essentials donated20:31
azonenbergnyan cat?20:31
kristianpaulahhh !20:31
azonenbergdonated? what am i saying20:31
azonenbergdesigned20:31
kristianpaulfrom fab004  i tought was another thing :)20:32
azonenbergand the last pic is one of my earliest tests when i was experimenting with litho techniques20:32
azonenberglol20:32
kristianpaulindeed nice camera resolution20:33
kristianpaulalso bundle price i guess :)20:33
azonenbergyep20:33
--- Thu Oct 20 201100:00

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