azonenberg changed the topic of #homecmos to: Homebrew CMOS and MEMS foundry design | Wiki: | Repository: | Logs:
chris_99 has quit [Quit: Ex-Chat]
NullMoogleCable has joined #homecmos
Noxz has joined #homecmos
sammyb has joined #homecmos
<sammyb> :D
<Noxz> and like everyone I recognize in here is also in electronics..
<diginet> haha
<Noxz> anywho..
<Noxz> MEMS question, still trying to get a solid answer, instead of using a thick resist such as SU-8 for building molds to then electroplate tiny metal gears (or whatever) could I potentially use a thin film on silicon(which is easier to focus the photomask), control etch anisotropically to have vertical sidewalls, then electroplate/form inside of the silicon wafer, then release the part/dissolve the silicon?
<diginet> I don't think you electroplate SU8 since it's non-conductive
<diginet> (could be wrong there)
<Noxz> you don't, you use it as a mold for the sidewalls
<diginet> and yeah, what I think you're describing is the lift-off process
<Noxz> but if you put it on a silicon wafer or osmething, anythign conductive, you can then eelctroplate for it
<Noxz> it's just limited by taking time to etch vertical sidewalls through Silicon?
<diginet> but dissolving the whole wafer would be hard
<diginet> you need plasma etching to get good aspect rations
<Noxz> what about doing a steam step to add SiO2 ?
<diginet> you can use some tricks with preferential etching for certain crystallographic plans to get anisotropy, but it won't compare to anything like with plasma
<diginet> the problem with what (I think) you're describing is that it would be hard to get good aspect ratios
<Noxz> I was reading that KOH and NaOH, up to about 30% by wt with IPA to be used as a suficant etches sidewalls vertically
<diginet> the operative word here is "preferentially"
<Noxz> well, yes, silicon doesnt relaly come in big depths like you can easily put on like su-8
<diginet> if you're using SU8, the silicon is just a substrate
<diginet> for MEMS
<diginet> are you talking about soft lithography?
<diginet> oh, wait, I reread what you wrote, now I understand
<Noxz> I want to make tiny gears
<Noxz> not really MEMS, but using the same techniques
<diginet> yeah
<Noxz> as I wont have any electrical components
<diginet> just out of curiosity, what is this for?
<Noxz> I want to make a tiny mechanical watch
<Noxz> my current watch uses a silicon spring, which is sort of what got me interested
<Noxz> and the current "worlds smallest" movement is the Jaeger Le-Coultre 101
<Noxz> and I know mems can shatter those dimensions
chris_99 has joined #homecmos
<diginet> interesting
<diginet> I too am interested in horology
<diginet> that's going to be rather difficult, I'll admit
<Noxz> yes
<Noxz> but.. given I am not looking at any other uses, it's not so bad
<diginet> true
<diginet> but the problem is that it's going to be difficult to get MEMS gears to drive macroscopic watch hands I would imagine
<Noxz> I am reading up on the title's wiki now
<Noxz> some good stuff in there
<diginet> "the title's wiki" what do you mean by that?
<Noxz> [Homebrew CMOS and MEMS foundry design | Wiki:
<diginet> ohh
<diginet> are you a university student? I ask because photoresists like SU-8 are nearly impossible to acquire as an individual
<diginet> I've been trying to come up with my own formula, but progress has been slow
<diginet> azonenberg can attest to my difficulties therein
<Noxz> that's why I said without SU-8
<Noxz> and no, I am not.. but I do live in silicon valley, if that helps
<Noxz> but ho I work for does not do anything fun
<diginet> probbably not
<Noxz> just a sysadmin for a website
<diginet> (help I mean)
<diginet> unless your employer would let you ship to their office
<Noxz> grantid, my group owns over 20k baremetal servers..
<diginet> companies will not ship to individuals
<Noxz> right
<diginet> there are no openly available photoresists suitable for anything vaguely MEMS like
<diginet> even if you could acquire it, it's very expensive, $600/L or something
<diginet> granted, that would effectively last you a lifetime
<diginet> (though they do have expiration dates)
<diginet> and they won't sell in smaller quantities
<Noxz> what about dry films that I laminate on?
<Noxz> like I said, I will be using the silicon wafer as the mold
<Noxz> a thin film to etch the silicon
<diginet> dry films probabky wouldn't adhere well
<diginet> plus the resolution is very poor
<diginet> you'd be lucky to get 10micron or so
<Noxz> hrrmm
<diginet> (could be wrong there)
<Noxz> but yeah, given the method I propposed, I don't need a thick film for MEMS
<diginet> trust me when I say there is no easy option, I've spent a good portion of the last two years trying to find a pay to do amateur microscale patterning
<diginet> s/way/pay
<Noxz> hrmm.. well.. I have some time to pitch in ;)
<diginet> heh, well good
<diginet> definitely keep around
<Noxz> and I am fairly motivated
<diginet> we're trying to get as many people interested as possible
<Noxz> like I said, I have one specific application
<diginet> yeah
<diginet> but the difficulty in patterning is the same
<Noxz> and I know a person or two in the chip fab business in the valley around here.. not sure if that would help
<diginet> ehh, probably not
<diginet> if I had $20k to spare, I could get all I need, problem is I don't, neither do most people
<Noxz> well, one guy said he could probably get me in contact with a guy who could make masks from $50-100/hr
<Noxz> sure
<diginet> what feature size?
<Noxz> if I need one final mask in the end though..
<Noxz> dunno
<Noxz> but, MEMS was the discussion
<diginet> film masks at ~5micron feature size are about that price
<diginet> my goal at least, is to make things available to an amateur, granted it still won't be cheap, but at least possible
<Noxz> right
<diginet> but supposing you got the masks, the other stuff is non-trivial
<Noxz> an American Scientific article is what got me REALLY interested in this, a few months back
<diginet> which one?
<Noxz> lemme dig it up
<diginet> you can't really effectively electroplate bare silicon, AFAIK
<Noxz> May 2014
<diginet> my understanding is that most processes use a sputtered seed layer
<diginet> what was the subject?
<Noxz> distributed compliance
<diginet> hmm
<Noxz> the econd video on that page is of mems
<Noxz> seconds
<Noxz> urgh
<diginet> I'm a little skeptical
<diginet> MEMS is not new
<diginet> it's interesting, but that sounds a little too TED-esque for my tastes
<Noxz> sure, but he has some patents on the subject
<diginet> which further calls into question the reliability
<diginet> patents holders have a vested interest in promoting their property
<Noxz> his company:
<diginet> interesting, but there are others doing similar things
<Noxz> sure, but it got me interested
<diginet> I'm generally wary of people who think they've stumbled on the next big thing
<Noxz> nah, just a scientific american article
<Noxz> I am not sure if that's what got me interested to get a watch with a silicon balance spring or not
<Noxz> but yeah, I have a 2013 Omega Speedmaster Racing, for the sole purpose of saying/knowing that I have a silicon spring
<Noxz> and I stumbled upon an ebay auction where it was not that bad of a price
<Noxz> so, I used to be a bicycle mechanic, so I am into gears and such.. if I could have it my way, I would fly out to switzerland to go to watchmaking school
<Noxz> boring sysadmin work for now, though
<diginet> there's actually a watchmaking school here
<diginet> Ohio Institute of Technology IIRC has a watchmaking program
<Noxz> OHIO HAS ONE?!?
<Noxz> I moved from columbus
<Noxz> to san jose for this jerb
<Noxz> I still have my place in columbus
<diginet> yep!
<diginet> (I'm pretty sure)
<Noxz> I think it's in dayton, or something?
<Noxz> and also texas has one
<Noxz> no ohio listed on that page
<Noxz> but.. itnerresting
<Noxz> that's oklahoma
<Noxz> not ohio
<diginet> OH oops
<Noxz> :/
<diginet> never mind lol
<Noxz> anyways.. I can't really do fine "swiss" types of finishes on something I can't see
<Noxz> but, I will still try
<Noxz> and I will probably first start with Nickel plating, then when I have a function calibre probably go to gold/platinum.. the thing is going to be so tiny it wont cost too much
<diginet> ehhh
<diginet> the materials might not
<diginet> but making this happen will not be cheap
<Noxz> I know
<Noxz> I sort of have a budget in my head
<Noxz> of course being single only helps
<diginet> hmm?
<Noxz> no wife/gf to stop me from spending too much, or rather taking my money away
<diginet> ahh
<diginet> I'm single too
<diginet> it's still not cheap
<Noxz> yeah... I know.. hardley any hobby that is
<diginet> it's not done yet
<diginet> that's speculation
<Noxz> it looks like the stage moves, and the laser stays in the same place?
<Noxz> based on the description
<diginet> yes, that's the idea
<Noxz> okay
<Noxz> just verifying
<diginet> moving the laser beam, e.g. with galvo mirrors would cause it to lose focus
<Noxz> that is similar reason why i am interested in a thin resist
<Noxz> just wait till you can make your own MEMS steppers
<Noxz> and replace those ball screws
<Noxz> btw, is the wiki yours?
<diginet> that's going to be a long, long, long-time coming
<diginet> no, it's azonenberg's
<Noxz> ahh, okay
<Noxz> I am well aware of him from ##electronics
<Noxz> I've spoken to him about su-8 before
<Noxz> spoken==chatted
<diginet> yeah
<Noxz> irc'd
<Noxz> I've also been in that channel for several years at this point
<Noxz> originally for designing a motor for an electric car
<diginet> anyway, if you're going to try and do MEMS I don't think wet etching will provide a high enough apsect ratio
<Noxz> .. which.. to add to my projects, I purchased an old electric car a few months ago
<diginet> and if you're going for plasma etching, you have bigger problems than resist thickness
<Noxz> did you check out the one paper I linked?
<diginet> the SciAm one?
<Noxz> vertical sidewalls acheived with KOH and NaOH
<diginet> oh, it wouldn't load for
<Noxz> with IPA as a surficant
<diginet> me
<Noxz> it just downloaded again for me.. I can slap it somewhere
<diginet> I'll take your word for it though
<diginet> no it's fine
<Noxz> okay
<Noxz> basically they use isopropyl as a surficant
<Noxz> and somewhat high concentration
<Noxz> and low temperature
<Noxz> which, yes, slows down etching
<diginet> anyway, IMO the most plausible way you could get this to happen is to shoot for something like 10micro resolution, and use a film mask
<Noxz> but vertical sidewalls are acheived
<diginet> *micron
<diginet> that might be doable
<Noxz> what do you mean by film mask?
<Noxz> instead of laser?
<diginet> yeah
<Noxz> ah, right
<diginet> a film mask is probably what your friend was referring to
<diginet> film masks are basically photographic film
<Noxz> what about a stepper to reduce the mask.. or a 1:1 mask?
<diginet> contact lithography is going to be much easier
<diginet> and should be suitable for MEMS
<diginet> (resolution wise)
<diginet> a company uses a photoplotter to scan silver halide film
<Noxz> and it helps that the mask is close with a thick resist
<diginet> it's in direct contact
<diginet> this helps resolution
<Noxz> sure
<Noxz> but I woudl't think that UV diffuses much?
<diginet> building a stepper is difficult
<diginet> it does
<Noxz> okay
<diginet> contact litho with a film mask is by far the most plausible means of doing what you want
<diginet> IMO anyhow
<Noxz> that was sort of the route I was looking at
Noxz_ has joined #homecmos
<Noxz_> VPN kicked me off, I didn't want to enter in the code again
<Noxz_> done with work for now
<diginet> $135 for a 7um mask
<Noxz_> ahh, nice
<diginet> that should be fine for MEMS
<Noxz_> what do you think a workable area should be, anyways?
<diginet> you'll need to, at the very least build some sort of spin coater, and figure out a way to get photoresist
<Noxz_> 1cm^2 ?
<diginet> no idea
<Noxz_> sure.. or test some laminated ones on the market
<Noxz_> I like the idea of the simple-ness involved in them
<Noxz_> versus spin coating, pre-baking, exposing, post baking, etc etc etc
<diginet> there's no way, unless I'm totally mistaken, laminated masks will be good enough
Noxz has quit [Disconnected by services]
Noxz_ is now known as Noxz
<diginet> not to mention I imagine adhesion to Si would be a huge issue
<Noxz> the reason is adhesion?
<diginet> among other things
<diginet> environmental instability is another
<Noxz> filling the gaps.. having air bubles
<diginet> and really, just the chemistry
<Noxz> oh
<diginet> thermal expansion, etc
<Noxz> I was trying to look at DJ DevCorp SUEX thick film resist, specifically for MEMS, but there isn't too much on it
<Noxz> no price sheet anywhere
<Noxz> 5 or so papers reference using it
<diginet> rule of thumb: if you have to ask for a quite, it's probably very expensive and not sold to individuals
<Noxz> quote* ;)
<diginet> Also, I recommend against using new/unusual stuff unless you have a very good reason
<diginet> right hahah
<diginet> mostly because we're already basically on our own, not many people are attempting this
<Noxz> sure
<diginet> at least with the more established resists there's some people who might be able to help
<Noxz> everyone interested in it is going to school for it, right?
<diginet> yes, but it's not just that
<diginet> azonenberg is a PhD student for example
<Noxz> my previous job I had a "Student" position, even though I was not, at Chemical Abstract Services.. they wre THE chemistry registry database, that "cas: 123-456-789" number you seee often after chemical names
<diginet> university research tends to assume university environments
<diginet> i.e. access to fancy labs, chemicals, etc
<Noxz> but not always true
<diginet> so there's a lot less to draw on that's applicable to people like you and me
<Noxz> sure
<Noxz> it like needing a 7.5mm screw and all thati's offered is a 7mm and 8mm.. so you have to redesign based on wht is available
<diginet> something like that, yes
<diginet> I've been working on a water-developable resist formulation, which uses PVA and a chemically amplified design. The idea is to avoid dangerous stuff, but it's still not easy
<diginet> I'm not a chemist, so that doesn't help either
<diginet> but probably your best bet is to see if any of your contacts can get you some resist
<Noxz> ah, sure
<diginet> unfortunately that's not really a long-term solution
<diginet> some of those chemicals are *nasty*
<diginet> I highly advise against working with them without a fumehood, for example
<Noxz> well, my initial idea seems decent enough, except for the adheshion issue
<Noxz> right, like KOH
<diginet> KOH is not as bad
<diginet> I'm more referring to things like HF
<Noxz> ohh, that too, that's on my lsit to disolve SiO2 if needed to have a release layer
<diginet> I don't think that's going to work very well
<Noxz> eh, I'm not there yet
<Noxz> ;)
<diginet> true, true
chris_99 has quit [Quit: Ex-Chat]
<Noxz> I'm giving my buddy a half hour to call me (going to buy a backgammon set.. maybe, plus dinner) before I get food myself
<diginet> ahh
<Noxz> a friend thought it would be funny to dump a bunch of ahlloween candy on my couch before boardgames last night.. and now I have a headache
<Noxz> agh, gotta VPN back in.. I'll be sure to add this channel onto my desktop for connecting
Noxz has quit [Quit: Page closed]
sammyb has quit [Remote host closed the connection]
Textmode has joined #homecmos
Textmode has quit [Ping timeout: 255 seconds]
Textmode has joined #homecmos
Noxz has joined #homecmos
Noxz has quit [Ping timeout: 246 seconds]
sammyb has joined #homecmos
qi-bot has quit [*.net *.split]
Textmode has quit [*.net *.split]
kristianpaul has quit [*.net *.split]
kanzure has quit [*.net *.split]
kanzure has joined #homecmos
Textmode has joined #homecmos
kristianpaul has joined #homecmos
qi-bot has joined #homecmos
Textmode has joined #homecmos
Textmode has quit [Quit: "It was one dev, naked in a room with a carton of cigarettes, a thermos full of coffee and bourbon, and all his summoned angels."]
qi-bot has quit [*.net *.split]
Textmode has quit [*.net *.split]
kristianpaul has quit [*.net *.split]
qi-bot has joined #homecmos
Textmode has joined #homecmos
kristianpaul has joined #homecmos
osfe has quit [Remote host closed the connection]
osfe has joined #homecmos
Speedevil is now known as Guest37431
BitEvil_ has joined #homecmos
Guest37431 has quit [Ping timeout: 272 seconds]
BitEvil_ is now known as SpeedEvil
mrdata has quit [*.net *.split]
berndj has quit [*.net *.split]
XgF has quit [*.net *.split]
mrdata has joined #homecmos
berndj has joined #homecmos
XgF has joined #homecmos
mrdata has quit [Ping timeout: 258 seconds]
mrdata has joined #homecmos
Textmode has quit [Quit: "It was one dev, naked in a room with a carton of cigarettes, a thermos full of coffee and bourbon, and all his summoned angels."]
chris_99 has joined #homecmos
SpeedEvil has quit [Quit: No Ping reply in 180 seconds.]
SpeedEvil has joined #homecmos
SpeedEvil has quit [Changing host]
SpeedEvil has joined #homecmos
chris_99 has quit [Remote host closed the connection]
chris_99 has joined #homecmos
superkuh has quit [Ping timeout: 265 seconds]
superkuh has joined #homecmos