azonenberg changed the topic of #homecmos to: Homebrew CMOS and MEMS foundry design | Wiki: http://homecmos.drawersteak.com/wiki/Main_Page | Repository: http://code.google.com/p/homecmos/ | Logs: http://en.qi-hardware.com/homecmos-logs/
X-Scale has quit [Ping timeout: 252 seconds]
Bike has quit [Ping timeout: 272 seconds]
Bike has joined #homecmos
Textmode has joined #homecmos
chris_99 has joined #homecmos
chris_99 has quit [Remote host closed the connection]
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
osfe has quit [*.net *.split]
osfe has joined #homecmos
jamie has quit [Excess Flood]
jamie has joined #homecmos
Textmode has joined #homecmos
<whitequark> don't remember what was in the link but I tried to send it to you yesterday
<whitequark> _Sync_: I just noticed that I got a quote from a reagent shop for manganese oxide
<whitequark> and they specified it as "semiconductor quality", $5/250g, natch
<whitequark> ... where would it be used? I have no ideas at all
<_Sync_> pppppfffft
<_Sync_> eh
<_Sync_> let me mediate a bit over that
<_Sync_> +t
<_Sync_> well... I could think of maybe manganese as dopant for ii-vi and iii-v semis
<davidc__> looks like it also might be used in some CMP stuff
<_Sync_> but at that price there has to be a quantity
<davidc__> (oldschool stuff)
<whitequark> 250g is MOQ
<_Sync_> oh true
<_Sync_> I completely forgot
<_Sync_> you can use that as abrasice
<whitequark> oh
<_Sync_> ~abrasive
<_Sync_> I always just heard about zirconium oxide
<_Sync_> but manganese would be fine too
<whitequark> right, zirconia
<whitequark> never heard of manganese oxide used as abrasive
<davidc__> back when I was looking up CMP stuff, I don't recall Mn based slurries turning up
<davidc__> but there appears to be some literature on the subject
<davidc__> I'm sure pretty much everything has been tried at one point or another
<_Sync_> ^
<davidc__> I can see if it turns up in any of the books I have...
<davidc__> but eh
<_Sync_> let me get the handbook out
<_Sync_> as my prof once said "a handbook is called a handbook because you need two hands"
<_Sync_> "handbook of semiconductor manufacturing technology"
<_Sync_> which is a bit moot after all
<_Sync_> but it makes a nice impression in the bookshelf
<_Sync_> and I got it for 30 bux
<davidc__> neato.
<_Sync_> yup
<_Sync_> there it is
<_Sync_> 'In addition, manganese dioxide slurries have been tried for metal CMP, with the manganese dioxide playing the role of both abrasive and oxidizer.'
<_Sync_> it says silica, ceria and alumina for slurries
<whitequark> coool
<whitequark> oxidizer, that makes sense
<_Sync_> whitequark: the reason lab equipment is so expensive is because there usually is not a lot of it made and people know that their customers will pay the price
<_Sync_> although sometimes it is also a thing of convenience
<_Sync_> if I know I absolutely need some boron nitride in vacuum grade RIGHT NAO I can always just call up goodfellow and have it the next day
<davidc__> yes, at prices approaching extortionate rates
<whitequark> "know that their customers will pay the price" < this
<whitequark> i.e. fuck anyone who is not sitting on a fat grant
<davidc__> $171 for a 15cm x 15cm sheet of lexan.....
<whitequark> $150 for a magnetic stirrer
<_Sync_> that's pretty reasonable
<_Sync_> I mean, I paid over 500 for my ika one
<whitequark> do you live in US?
<_Sync_> no
<davidc__> I wonder if there's a website for open source / DIY lab equipment
<davidc__> I should look this up
<_Sync_> there probably is
<whitequark> davidc__: yes
<whitequark> teklalabs.org
<davidc__> because I'm pretty sure I could make a mag stirrer for way less than $200
<whitequark> it's kind of underwhelming
<_Sync_> but dunno, there is not much sense in that for me
<_Sync_> I want to get things done
<whitequark> wait, whoa, 150 designs?
<_Sync_> and not build stuff to get things done
<davidc__> _Sync_: I mean, I understand that viewpoint too, but $500 for a heat plate could be $500 to get metal bits machined for my SEM
<whitequark> ^
<_Sync_> sure
<whitequark> it's reasonable if you get paid by someone in San Francisco
<_Sync_> but I can just as easily get machining done afterwards
<davidc__> or $500 worth of optics that I can;t make myself.
<whitequark> I understand paying $500 for something that requires specialized equipment to create, or R&D cost
<whitequark> a hot plate has been a hot plate for at *least* fifty years
<_Sync_> yeah and they have been always expensive if you got a good one
<_Sync_> unfortunately
<whitequark> how the fuck do you make a bad hot plate
<_Sync_> for me it is just a thing that has to work
<_Sync_> make it out of steel and undersize the heater
<_Sync_> for me, lab equipment has to work
<_Sync_> it is not worth the time building it
<_Sync_> on some things it is, but mostly it is not
<davidc__> _Sync_: I find a lot of my lab equipment / tools work better because I built em myself :S
<davidc__> I mean, there are quality MFGs out there, but they charge what they're worth
<davidc__> which means that they're even further out of budget.
<_Sync_> yeah I can see the point
<davidc__> ALso, and a hot plate is sorta a dumb example of this; I like my own gear because its programmable. I just stick an ethernet interface on everything, and then I can script it from then on
<davidc__> (again, ethernet-enabled-hotplate is sorta a dumb idea, but still)
<_Sync_> sure
<_Sync_> but if I'd build everything from scratch
<_Sync_> I could not do the things I want to do
<davidc__> Yeah, and I mean, I buy plenty of stuff... just usually end up reversing/modding it
<_Sync_> shopping for machinery is even worse
<_Sync_> I have a surface grinder, so I could just make good vises myself
<_Sync_> but it is just easier to pay kurt to do it for me
<_Sync_> but yeah it is amazing what even common labware costs
<_Sync_> if you buy "retail"
<whitequark> most glassware in RU costs cents to single dollars
* whitequark is looking for some now
<whitequark> but even trivial equipment is in hundreds of $
<_Sync_> 1000ml boro lab beaker is ~$7 here in single units
<whitequark> $6.4
<whitequark> hrm
<whitequark> I guess glassware is fairly cheap everywhere
<davidc__> _Sync_: are you in north america? if so, what supplier?
<davidc__> here in .ca, all the vendor-sourced glassware is a ripoff
<davidc__> I cat get cheapstuff from ebay/amazon/etc, but who knows what its actually made of
<davidc__> "Oh, we said borosilicate glass? We just meant glass. Like window glass!"
<davidc__> "Sorry about those cuts all over your hands and that nitric acid spill....."
<_Sync_> well you in .ca don't even get proper gas so
<_Sync_> no, I'm in germany
<davidc__> _Sync_: proper gas?
<_Sync_> 91 octane
<_Sync_> is max in cali iirc
<davidc__> _Sync_: .ca is not cali :P
<davidc__> I'LL HAVE YOU KNOW :P
<_Sync_> err
<_Sync_> my bad
<_Sync_> long day
<_Sync_> eh
<davidc__> I know, just kidding, its all good
<_Sync_> but yeah, I was wondering today about how cars get along in california thus the connection popped into my mind
<_Sync_> because apparently the highest octane rating there is 91
<_Sync_> easily to get
<davidc__> maybe I'll need to get some of my .eu friends to buy and ship me some stuff... hmm
<_Sync_> I don't think that is going to be much cheaper
jamie has quit [Excess Flood]
jamie has joined #homecmos
jamie has quit [Excess Flood]
jamie has joined #homecmos
chris_99 has quit [Quit: Ex-Chat]
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."]
<SpeedEvil> Borosilicate is easy.Just take regular glass and add boric oxide.
* SpeedEvil needs some nice source of SiC elements.
<whitequark> ebay.com/usr/advancer53
<whitequark> sells off some astounding amounts of advanced NOS ceramics
<SpeedEvil> SiC heating elements
<whitequark> oh
<whitequark> well idk take that ceramic and spin thin molybdenum wire around it as well
<whitequark> do you need a source for molybdenum as well
<SpeedEvil> No, the heating element is actually SiC
<SpeedEvil> no wire
<whitequark> oh
<whitequark> cool
<whitequark> I think I want that
<whitequark> how does it work
<SpeedEvil> It goes to ~1400C IIRC in air.
<whitequark> fuck air, I have vacuum
<whitequark> right now I plan to heat a carbide crucible and then use the radiated IR to heat my workpiece
<SpeedEvil> I want a vacuum furnace that can do HIP
<whitequark> er
<whitequark> carbon
<whitequark> carbon will be heated inductively
<whitequark> using a watercooled copper coil
<whitequark> I even figured out how to cast thin rods, though untested yet
<SpeedEvil> To wht temperature?
<whitequark> I need 1700 for sintering alumina
<whitequark> so at least that
<SpeedEvil> F?.
<whitequark> C
<whitequark> 1700 is over working temperature for silica
<whitequark> ... well, it just melta at 1600
<whitequark> And you don't want to use it over 1300 or something
<SpeedEvil> I'd love to get to 1700C. But practically yes, that gets expensive
<whitequark> 1300 is up to what most commercial furnaces go. most even 1250. nichrome wire
<whitequark> I think it is not very expensive when doing using the technique I want to.
<whitequark> Which is, heat a graphite crucible as something that's refractory and conductive enough. Use IR to transfer heat. Use 6x molybdenum reflective shields around it to protect chamber.
<SpeedEvil> There are all sorts of fun things.
<SpeedEvil> AlN synthesis for example
<whitequark> Use formed molybdenum sheet (maybe carbon, but that's too fragile) to hang the workpiece under the coil
<whitequark> Mo is not expensive in that quantity
<whitequark> The rest is pretty much trivial to construct.
<SpeedEvil> What are you trying to heat?
<whitequark> I have currently two use cases for this furnace
<whitequark> One is sintering pure alumina, to make custom feedthroughs
<SpeedEvil> Seems machining'd be easier
<whitequark> (also because I think alumina is cool and I want to play with it)
<whitequark> mm.
<SpeedEvil> But :)
<whitequark> its hardness is 9
<whitequark> on mohs
<whitequark> it's super hard on tools.
<SpeedEvil> I thought diamond was OK
<whitequark> I'm not sure if even SiC cuts it
<whitequark> diamond is ok.
<SpeedEvil> I haven't looked at it
<whitequark> slow though.
<whitequark> also very brittle.
<whitequark> there's stuff like Macor if you want good machining
<whitequark> alumina is mainly machined abrasively, you're sort of limited in the shapes you can make.
<whitequark> and it is not a job you can just hand off to whoever.
<SpeedEvil> I guess
<SpeedEvil> HIP is cool. Take two grades of SS powder, '3d print' into a layered form. Now heat under vacuum and squeeze the voids out.
<SpeedEvil> Does really nice things for the metallurgy.
<SpeedEvil> (well, the HIP, the ther is just for prettyness)
<whitequark> ew.
<whitequark> more about alumina. I am not planning to press it into shape.
<whitequark> I will be doing direct coagulation casting. Mix it with low mw polyacrylic acid, mill, then add MgO
<whitequark> You have half an hour before it turns into a gel.
<whitequark> A sol-gel process.
<SpeedEvil> Interesting, then you just heat?
<whitequark> yeah, you can just... take it out of the cast.
<whitequark> you can even machine it!
<whitequark> it has viscosity of over 50 Pa·s
<whitequark> invented in 2008.
<SpeedEvil> So it's something like a firm jelly?
<whitequark> yes.
<SpeedEvil> Heating doesn't floppyise it?
<SpeedEvil> that's probably not a word
<whitequark> Nah.
<SpeedEvil> 'is it self-supporting'
<whitequark> yes, totally.
<SpeedEvil> neat
<whitequark> now, that's not all.
<whitequark> before I make my vacuum sintering machine (vacuum really helps for heat isolation here. otherwise argon'd do)
<whitequark> I will do some samples sintered in a commercially rented oven at 1300.
<SpeedEvil> 'with density >98%' that's damn god
<SpeedEvil> oo
<whitequark> normal alumina doesn't sinter at all at 1300
<whitequark> I'm going to blow your mind the 2nd time, then :P
<SpeedEvil> yeah - 'from solution' techniques with ~0 'particle' size can do so much better than making it from sand
<whitequark> if you add 2% of both MnO and TiO2 to alumina, then you can fire it at 1300 at nearly theoretical density as well.
<whitequark> Practically as good as at 1700.
<whitequark> Also, using 4% of either doesn't work nearly as good, which is super neat.
<whitequark> The issue with DCC is that dispersant acts as per surface area, and you need >10m²/g, in effect requiring particles of 0.3 micron or smaller
<whitequark> So, fumed alumina.
<whitequark> Linde A works, but it's sort of expensive as fuck.
<whitequark> Also surprisingly hard to obtain.
<whitequark> The 'normal' chinese alumina is 3 micron, fused alumina. Also lower limit when ball milling, too.
<SpeedEvil> I wonder what flash powder makes
<SpeedEvil> :)
<whitequark> You can't DCC that, but you can slip cast that. Slip casting means you can't take it out though, so vacuum sintering is probably out anyway
<whitequark> So I want to slip cast the 'doped' alumina, and then switch to DCC once I have the oven.
<whitequark> They did a damn good job exploring that phase space
<SpeedEvil> 'yes grad student, just mix these 10*10*10 different mixtures. It's OK, I just want them next week'
<whitequark> exactly
<whitequark> that's not all I want to do. so basically, you get alumina. cool. what next?
<SpeedEvil> https://www.youtube.com/watch?v=8YEdHjGMeho - related machine is related
<whitequark> have you fucking /seen/ the process for metallizing it? it has like half as much steps like making semiconductors
<whitequark> this is bullshit.
<whitequark> the not-bullshit process is one using active brazing solder.
<whitequark> basically, take the regular Ag/Cu solder, add 3% Ti, boom, it now wets alumina.
<SpeedEvil> hah
<whitequark> Directly.
<SpeedEvil> Now that is cool.
<whitequark> Exactly.
<SpeedEvil> Ebay can be bad/awesome.
* SpeedEvil looks at the 10* alumina crucibles 70mm*210mm in the corner.
<whitequark> No one makes or sells the active brazing solder in quantities below "call us"
<whitequark> So I will have to do it myself, probably just from raw powders.
<whitequark> The issue now is casting it.
<SpeedEvil> Alibaba annoys me. S much cheap stuff FOB china.
<SpeedEvil> At ridiculous prices.
<whitequark> You can't make that alloy not in vacuum/inert gas since Ti reacts vigorously with air at even low temps
<whitequark> But it's not like it's easy to melt and cast shit all inside your chamber
<SpeedEvil> Well, yes, but if you don't have argon, you're not trying. :)
<SpeedEvil> Will CO reduce TiO?
<SpeedEvil> I guess not
<whitequark> dunno, will need to calculate the thermodynamics to see
<whitequark> too lazy to
<whitequark> I would think that it's temp-dependent
<whitequark> CO reduction generally is
<SpeedEvil> I was wondering just add TiO ...
<SpeedEvil> but that's insane
<whitequark> Nah, that's not how it works.
<whitequark> Ti in solder reduces alumina to aluminium
<SpeedEvil> Oh - neat
<SpeedEvil> Does Al wet AlO?
<whitequark> In fact, having TiO2 in the ceramic 'might' passivate the ceramic against the solder, according to someone with PhD in that general area
<whitequark> no.
<whitequark> but now you have a grain of alumina that has a few monolayers of Al exposed on the surface.
<whitequark> maybe even one?
<whitequark> the rest of the solder can hold onto that well.
<SpeedEvil> yeah - I was wondering abou the rather lazier solution of swirling molten Al around in it
<whitequark> Anyway, so about my thoughts on making the solder rods
<whitequark> idea #1, suggested to me by a taxi driver. yes, I routinely try to talk with taxi drivers on topics such as advanced engineering ceramics
<whitequark> it works way more often than you'd think
<SpeedEvil> Make a blob of it, then wire draw?
<whitequark> exactly
<whitequark> take it to a jeweler.
<SpeedEvil> I've made fine solder with my drawplate
<whitequark> literally any commercial jeweler
<SpeedEvil> worked well
<SpeedEvil> (from coarse solder)
<whitequark> idea #2: put a sealed from one end pyrex tube into the crucible
<whitequark> put under vacuum
<whitequark> melt
<SpeedEvil> yeah
<whitequark> while molten, add argon
<SpeedEvil> then release vacuum
<whitequark> it will be sucked inside
<SpeedEvil> and smash
<whitequark> possibly combine with idea #1 to make finer wire.