Topic for #homecmos is now Homebrew CMOS and MEMS foundry design | http://code.google.com/p/homecmos/ | Logs: http://en.qi-hardware.com/homecmos-logs/
wolfspra1l has quit [Ping timeout: 245 seconds]
wolfspraul has joined #homecmos
nmz787 has joined #homecmos
<nmz787> anyone here?
<azonenberg> Yep
<azonenberg> It's been slow lately but i'm still working on it
<azonenberg> was machining some microscope tube parts just a few minutes ago
<nmz787> so i think i mentioned I was interested in making microfluidics
<nmz787> and was trying to build a laser CNC to cut silicone
<nmz787> but I'm wondering how far people here have come
<nmz787> basically trying to do cost-benefit analysis before putting money down
<azonenberg> Not sure, my focus personally is mostly on lithography
<nmz787> (for my own project)
<azonenberg> i started out doing IC reverse engineering with a friend
<azonenberg> then started to realize the basics looked simple enough maybe i could do it
<nmz787> did you build a lithography setup for yourself?
<azonenberg> i've gotten patterning at 20um design rules by projection using totally homebrewed setups but i'm going to break down and buy professionally printed masks i think
<nmz787> do you have a reducer for the mask?
<azonenberg> I've used a microscope objective but my FOV is super limited as a result
<azonenberg> my new plan calls for doing contact masks 1:1 scale on 2-inch wafers
<azonenberg> laserlab.com makes 8000DPI (12.5um design rule, 3.125um lambda) masks for under $100 each
<azonenberg> on plastic film
<nmz787> aren;t masks like $500-800 even at academic prices?
<azonenberg> Thats for chrome on glass
<nmz787> electron beam ethced quartz
<azonenberg> film is a lot cheaper
<nmz787> yeah
<nmz787> whats the tradeoff?
<azonenberg> Less resolution, less rigidity. less thermal stability
<azonenberg> e-beam masks can be deep submicron features
<azonenberg> plastic limits you to a few um
<nmz787> how do you make those?
<nmz787> the film ones
<azonenberg> Photographic film, pretty much
<azonenberg> direct write with a UV laser on an x-y table
<azonenberg> I'd be buying them, $80ish for a 10x16 inch sheet is a great price
<nmz787> hmm, so you'd need a CNC
<azonenberg> considering you can cut that up and get a full mask set
<azonenberg> you can fit 10+ 2-inch wafer masks on that sheet
<azonenberg> with enough margin to hold in an aligner
<nmz787> would you build or buy an XY table?
<azonenberg> For mask making, neither - like i said i'm going to just buy professional masks
<nmz787> from where?
<azonenberg> laserlab.com
<nmz787> you wouldn't happen to be in NYC area?
<azonenberg> Upstate, actually - near albany
<nmz787> hmm, I was up in Rochester for a while, but been in NYC for ~2months
<azonenberg> Ah, nice
<nmz787> I have friends in Albany though
<azonenberg> i see
<azonenberg> i'm right across the river (RPI)
<nmz787> so are you effectively buying the film from laserlab.com?
<nmz787> or credits from them?
<nmz787> i don't see BYO-film
<azonenberg> Yeah, you buy the film
<azonenberg> from them
<azonenberg> You email the GDS and pay, a day or so later the mask shows up at your front door
<azonenberg> i havent ordered from them for this project but a group i consulted for used them for a low-res MEMS mask
<nmz787> seems pretty reasonable
<nmz787> could effectively cut out the laser CNC in my setup, if I can figure out lithography
<nmz787> i've made some transistors in a lab i audited once
<nmz787> but that used a stepper
<azonenberg> lol
<azonenberg> Yeah, i intend to start with contact litho and then make an H-line stepper in the future
<nmz787> what's H line?
<azonenberg> that will be a much bigger project and most likely wait until i have a full time job
<azonenberg> mercury vapor H line
<azonenberg> the wavelength
<azonenberg> 405nm
<nmz787> using it for mono light for interferometric positioning?
<azonenberg> alternatively I-line, 365nm
<azonenberg> No, the exposure wavelength
<nmz787> ah
<azonenberg> Either of those is a pissibility
<nmz787> where would you get the gear?
<nmz787> the stepper?
<azonenberg> Make it from scratch
<azonenberg> i mean i'd buy leadscrews and piezos etc
<azonenberg> probably use encoders for rough positioning and interferometers for fine
<nmz787> hmm, and still order the masks?
<nmz787> oh, is that how they'd do it
<azonenberg> That's a long-term project
<azonenberg> first step is to do contact litho
<nmz787> see i got confused after the interferometer and thinking of how much vibration the whole system would have
<nmz787> and then needing an optical vibe dampened table
<azonenberg> i would definitely use a vibration dampening table if i went that far
<azonenberg> at 1um features its not that bad
<nmz787> but I'm only aiming for <10 micron sized features now
<azonenberg> Even that is a bit tricky
<nmz787> hmm
<nmz787> cool
<azonenberg> smallest i've done reliably with my homebrew setup was 20, i hit 5 experimentally
<nmz787> so maybe I can try this laserlab.com and contact lith
<azonenberg> but most of that was my cheap objective i was using
<nmz787> that was you making your own masks though, right?
<azonenberg> i was using inkjet or laser pinted 200um masks
<azonenberg> then reducing with a horribly hacked setup
<azonenberg> contact litho with professional masks should give good results
<nmz787> sure, so contact should be fine with the professional film masks
<nmz787> cool
<azonenberg> thats the plan
<nmz787> so is it actually in contact, or separated slightly?
<nmz787> i imagine the photoresin is liquid or tacky
<azonenberg> No
<azonenberg> you bake off the solvent first
<azonenberg> its quite solid
<azonenberg> Normally you bring the wafer into contact, back off a micron or so, then align
<azonenberg> then bring into full contact for exposure
<nmz787> are there any chemistries that are suited to home/home-business?
<azonenberg> for photoresist?
<nmz787> i'd just be doing soft-lithography
<azonenberg> i dont know as much about fluidic stuff
<azonenberg> DNQ-novolac can be obtained for PCBs from injectorall.com
<azonenberg> the solvent is a little nasty so you probably want exhaust in the area you're actually coating
<nmz787> where you expose the resist, develop and clean off, then pour silicon on top, cure, peel off
<azonenberg> once dried its pretty harmless
<nmz787> silicone*
<azonenberg> the developer is dilute NaOH / KOH / TMAH or, hypothetically but not tested, dilute ammonia
<nmz787> hmm
<azonenberg> and i dont know anything about microfluidic patterning of silicone or polymers in general
<azonenberg> i've always coated it over a metal layer, then etched
<nmz787> hmm, well thanks for the site links
<nmz787> this page says 'emulsion side'
<azonenberg> Yeah
<nmz787> does that mean it may have a 3D surface patterned on it?
<azonenberg> The film is a solid plastic base with a photosensitive later on top
<azonenberg> layer*
<azonenberg> you typically want that layer against your wafer to avoid diffraction around it
<azonenberg> rather than having the plastic in between
<azonenberg> In most cases you want a PD mask
<azonenberg> VPD is the ordering code at 8000 DPI
<nmz787> i'm wondering if I can just pour silicone on top of that
<azonenberg> on top of the resist?
<azonenberg> i dont know anything about patterning silicone but you could potentially do a lift-off type process
<azonenberg> coat a glass substrate with photoresist
<azonenberg> expose all of the areas you do not want channels
<azonenberg> develop, now your channels are high spots where there's resist and background is low spots
<azonenberg> then coat silicone over the glass
<azonenberg> after it cures, peel off the glass plate
<azonenberg> and remove the resist with a solvent
<nmz787> ya
<nmz787> but i was saying pouring it on the film from laserlab
<nmz787> if its already 3D
<azonenberg> Oh
<azonenberg> no, odds are its not
<azonenberg> its just reacting internally
<azonenberg> also, you dont want to screw up your mask in the event of a process defect :P
<nmz787> yeah
<nmz787> what if the resist peels off the glass when you start peeling off the silicone?
<azonenberg> Doesnt matter
<azonenberg> you just wash it off with a solvent
<azonenberg> i dont think acetone eats PDMS
<nmz787> i'm not sure
nmz787 has quit [Quit: Leaving.]
nmz787 has joined #homecmos
Guest19125 has quit [Read error: Operation timed out]
Charlie has joined #homecmos
Charlie is now known as Guest56798
fpgaminer has quit [Ping timeout: 256 seconds]
fpgaminer has joined #homecmos
fenn_ is now known as fenn
lekernel_ has joined #homecmos
lekernel has quit [Ping timeout: 276 seconds]
nmz787 has quit [Quit: Leaving.]
lekernel_ is now known as lekernel
nmz787 has joined #homecmos
nmz787 has quit [Quit: Leaving.]
hideo has quit [Remote host closed the connection]
hideo_ has joined #homecmos
nmz787 has joined #homecmos
nmz787 has quit [Quit: Leaving.]
kristianpaul has quit [Ping timeout: 272 seconds]
kristianpaul has joined #homecmos
kristianpaul has joined #homecmos
kristianpaul has quit [Changing host]
wolfspraul has quit [Ping timeout: 246 seconds]
wolfspraul has joined #homecmos
superkuh has quit [Remote host closed the connection]