#homecmos IRC log for Monday, 2013-02-18

Sync_what's so bad about it, once it's gone it's gone00:00
Sync_use it while we can00:00
Sync_huh 10" conflat blanks are ... spendy00:01
diginetbecause we lose a tonne of things like MRIs, particle accelerators, cryogenics in general, etc00:42
Sync_I guess there will be an effort to make better superconducting magnets that work at LN2 temperatures00:45
diginetit's not the critical temperature that is the issue, the superconductors which need LHe have a much higher critical field, and also have much better mechanical properties01:20
Sync_yes, but that is just an engineering problem ;)01:21
diginetthe latter yes, but not the former01:22
diginetalthough, one potential option could be MgB2, which only needs LH201:22
Sync_I guess if we really run out sensors will have improved01:24
Sync_we're currently also making a mass spec for breathing air01:24
Sync_to detect trace acetone01:24
diginetazonenberg: are you around?01:25
Sync_which appears in a few ppm just before you have a sugar shock or epileptic seizure01:26
diginetI was wondering, what is the DPI on photomasks generally? Because I wondering, would FLCDs work as a sort of digital photomask? They have ridiculously high pixel density01:26
Sync_but the vacuum system is the main problem01:26
Sync_good question diginet01:32
diginetif an FLCD could work as a photomask, that would be my number one choice, if not, then e-beam01:32
Sync_ebeam would be mine because that is pretty easy to make01:33
diginetsure, but vacuum01:35
diginetthen again, the mask is just PMMA in a solvent, which is absurdly cheap/simple/safe01:35
Sync_vacuum is easy01:35
diginetis it though? you need a pretty hard vacuum for e beam01:39
Sync_yeah but getting in the 10^-7mbar range is not too bad01:40
Sync_which is acceptable for ebeam01:40
diginetwell, you can cannibalize a rotary vane from an old fridge for the first stage, and an ion beam pump might be DIYable01:40
Sync_not really01:41
Sync_that will not be sufficient01:41
diginetI don't know enough about them, but I wonder if a turbomolecular pump could be made from a turbocharger for a car01:41
Sync_just get a two stage rotary and a turbo01:41
diginetturbos at least $5k01:41
diginetI don't call that DIYable, at all01:41
diginetand they're insanely fragile and tempermental01:41
Sync_then why do I have 7 rotaries and 2 turbos?01:41
Sync_somewhere there should also be a tsp and an ion pump01:42
diginetyou do?01:42
diginetwhere did you get them?01:42
Sync_mostly ebay01:42
Sync_or fron work01:42
Sync_~from even01:42
diginethow much?01:43
Sync_my first rotary turbo set was 450¬01:43
diginetstill pricey though01:43
Sync_thinking about getting a cryo pump01:43
diginetcryo pumps could be DIYed I think01:43
diginetit's just a stirling engine in reverse01:43
Sync_but it is quite a challenge to get it working at 12K01:45
Sync_it is easier just to buy one01:45
Sync_nobody said it would be cheap but it is not really expensive01:47
diginetdo rotaries have oil?01:47
Sync_they have to01:47
diginetare there any oil-less first stage pumps?01:48
Sync_yes membrane pumps01:48
Sync_why do you want to go oilless?01:48
Sync_the problem with membrane forepumps is low throughput and the turbo needs to be a modern hybrid one01:49
diginetbecause oil creates a lot of problems01:49
diginetwhat about sorption pumps? they seem rather simple01:50
Sync_yes but you need to have access to ln201:51
Sync_oil isn't too much of a problem01:51
diginetLN2 isn't either01:51
Sync_yeah but you need a supply of it01:51
diginetLN2 is cheaper than water01:51
Sync_oil just sits in the pump01:51
diginetI'm building an air liquefaction machine01:51
diginetthe problem is when it gets into the chamber, which is especially a problem for semiconductors01:52
Sync_yes for ultra modern devices and mass production it is an issue01:52
Sync_but not for a hobbyist or research01:52
Sync_two of our evap chambers are oil pumped and it is all good01:53
Sync_I'd have a scroll pump here just in case tho01:53
diginetyeah, those are cool01:54
Sync_again, you complain about vacuum being expensive and then complain about the cheapest way to get there by using oil sealed rotaries...01:54
Sync_it is much less a problem than one would think01:55
Sync_you can even inject nitrogen into the forepump line to keep it at an elevated pressure01:55
diginetsorry, I didn't mean to complain, :P01:56
Sync_what approach are you using for your liquifier?01:56
diginethampson-linde cycle, with pre-cooling01:56
diginetit's not the most efficient, but it's not so bad01:57
Sync_ah lame01:57
diginetwhy do you say that?01:57
Sync_because that's the easy way01:58
diginetwhat would you propose using then? reverse-brayton?01:59
Sync_I bought a high pressure compressor for that but then changed my mind, an autocascading precooler is just easier01:59
Sync_but that project is on ice since I just can open the ln2 tap at work02:00
diginethmm, is a cascade cooler more efficient?02:01
Sync_that really depends02:02
diginetbut you say cascade is easier?02:03
Sync_we have a small ln2 generator in one lab because getting a ln2 line installed is too expensive and it uses a cascade02:03
Sync_well, you only have to get the gas mix right02:03
Sync_which is not too bad02:03
diginetI guess that makes sense considering there is no compressor (the compressor is definitely expensive)02:03
Sync_yes no dicking around with high pressure air too02:04
diginetone thing I've considered is starting it off with a membrane to at least eliminate some of the O2, I'd rather not have to deal with LOX02:04
Sync_lox is fun02:05
Sync_it is also blue :D02:05
diginetsure, but it's explosive and dangerous02:05
diginetbut it is pretty, no denying that02:05
Sync_it is only supporting combustion very well so just keep it away from combustibles02:06
Sync_it is also quite magnetic02:06
Sync_which is also fun to demonstrate02:06
diginetyes! I've seen that before, well, diamagnetic technically02:07
Sync_you can have very good fun burning things quite fast with lox02:07
diginetthis is insane, but what I really want is some DIY Ar02:08
diginetthat stuff is pricey (even though it shouldn't be)02:08
Sync_20l bottle refill is around 40¬02:08
Sync_which is quite good02:10
diginetthat's almost $70, not cheap IMO02:10
Sync_you're getting a gas, known to be oil free, 4 9s or better in quality02:11
Sync_I mean really, that's cheaper than you could do it02:11
diginetI don't need 4N though02:12
Sync_what do you want to use it for?02:12
diginetproblem is purge and fill means I need several times the volume of the chamber for one session02:14
Sync_use a load lock02:14
Sync_a 50l bottle of n2 would probably be the easiest02:15
Sync_yeah 50l n2 is +-50¬ per fill02:19
Sync_for a continious application that'd be quite a lot but then n2 concentrators are quite spendy02:20
azonenbergdiginet: For SEM coating you can sputter/evaporate almost anything10:52
azonenbergsputtered Au can be done with a less high vacuum and cheaper equipment10:52
azonenbergas with sputtered Pt10:52
azonenbergwith deeper vacuum you can evaporate C but that doesnt work well on samples with complex topography10:52
nmz787diginet: what's an FLCD?11:30
B0101azonenberg: do you test your wafer conductivity when you get them?12:17
azonenbergB0101: I have not done any semiconductor work per se12:17
azonenbergso i've never needed to12:17
azonenbergall of my fab to date has been purely mechanical12:17
azonenberglithography testing, MEMS, etc12:18
azonenbergThat way i dont care about metal contamination ;)12:41
B0101I may sound crazy, but do you think you can grow oxide layers using a normal oven?12:49
B0101I want to pump steam into a normal oven, but I think it will be a slower process than if I were to use a kiln at temperatures above 1000 deg C, am I right?12:51
azonenbergIt's exponential12:53
azonenbergYou will get oxide growth even at room temperature12:53
azonenbergBut it wont be usefully fast below 600C or so12:53
azonenbergI would suggest you look into spin-on glass if you want to avoid thermal oxidation12:53
diginetFLCD is ferrofluid liquid crystal display13:53
diginetso I've been working on an intersting-ish side project, basically, DIY e-paper14:03
azonenberghow far have you gotten?14:03
diginetpretty far actually14:03
diginetit is surprisingly simple14:03
diginetit uses electrochromism instead of electrophoresis14:03
diginetprussian blue pigment sandwich between ZnO:Al coated glass (patterned of course)14:04
azonenbergI guess the low resolution but cheap option14:04
diginetor PEDOT14:04
azonenbergwould be thermochromism :P14:04
azonenbergnice little grid of heating elements14:04
azonenbergbacking a thermochromic dye on a plate14:05
azonenbergrefresh rates on the order of 0.5 Hz14:05
azonenbergmassive power consumption14:05
azonenbergresolution of 0.5cm or so14:05
diginetwell prussian blue is insanely cheap, and PEDOT and/or ZnO:Al are as well14:05
azonenbergSounds interesting14:05
azonenbergWhy ZnO/Al instead of ITO?14:05
diginetbecause ITO is stupidly pricy14:05
diginetand I prefer to use sustainable materials14:05
azonenbergHmm, how about like sputtered Au or something14:05
azonenberga few nm of it cant be that expensive14:06
diginetwouldn't work, too thin14:06
diginetthus too high resistivity14:06
azonenbergAs in, to be thick enough to have good resistivity it'd be opaque?14:06
diginetAZO is already in use though, it's much more common in the solar cell industry14:07
azonenbergI see14:07
diginetthe only problem is that it needs a moisture barrier because it is more susceptible to atmospheric water vapour than ITO14:07
azonenbergYeah, i was just reading that14:07
diginetseriously though, ITO is a thorn in the side of the display industry14:08
azonenbergHow about just spin-on glass?14:08
diginetyeah, or acrylic or whatever14:08
azonenbergWell i was thinking14:08
azonenbergZnO/Al film14:08
azonenbergpattern it14:08
azonenbergcoat the whole surface with a transparent insulator of some sort, say SOG or reactive sputtered SiO214:08
azonenbergThen etch windows in the glass where you want "bond pads"14:09
azonenbergand put down, say, a thick aluminum or copper layer14:09
diginetI think what is often used is PEDOT, which is of course also conductive, and protects it (where you need contact with the electrolyte)14:09
diginetPEDOT:PSS is interesting, very easy to make, the problem is it only has about 1/10 the conductivity as ITO14:10
diginetthat being said, electrochromics are bistable, so maybe it doesn't matter14:10
diginetyou know what's cool though? prussian blue is the same pigment used in blueprints/cyanotypes, so this would be literally, "e-blueprints"14:11
diginete-ink is amusing, because it's /way/ over-engineered14:12
diginetthere have been much simpler options around for a long time14:12
diginetthe best part is that the bistability means no need for active matrix14:14
diginetwhat I don't understand is, considering how simple this stuff is, why it isn't in use14:17
azonenbergwell the wiki article suggests that AZO is difficult to etch14:18
azonenbergYou might want to consider a lift-off process14:18
azonenbergNot that those have great resolution either14:18
azonenbergHow is it normally deposited - sputtering?14:19
diginetno, not AZO, i mean electrochromics14:19
azonenbergAnd is that reactive sputtering of an Al/Zn target in Ar+O2?14:19
digineteven using ITO14:19
diginetI mean, electrochromics are absurdly simple14:19
diginetazonenberg: my preference is sol-gel, which has been demonstrated14:20
diginetalthough I'm still considering PEDOT14:20
azonenbergblack to transmissive14:20
diginetuses a PEDOT relative as well14:21
azonenbergHow about switching speed14:22
diginetcan be down to 200 milliseconds14:22
azonenberg"Recent advancements in modified porous nano-crystalline films have  enabled the creation of electrochromic display. The single substrate  display structure consists of several stacked porous layers printed on  top of each other on a substrate modified with a transparent conductor  (such as ITO or PEDOT:PSS)."14:22
azonenbergLooks like it's an ongoing research problem14:22
Sync_hmm yeah14:37
B0101Anyone here knows about optics?22:47
Sync_just ask22:50
B0101I want to know how to set up a proper system in which take a photomask and shrink the pattern down for exposure22:53
--- Tue Feb 19 201300:00

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