#homecmos IRC log for Saturday, 2013-02-09

--- Sat Feb 9 201300:00
nmz787anyone around?00:31
azonenbergyep01:39
azonenberghttp://mtixtl.com/SiO2-Si3N3.aspx12:57
azonenberg2-side polished 2" wafer12:57
azonenbergwith oxide + nitride12:57
azonenbergWould allow some interesting MEMS processing with no need to do film deposition (except for metal evaporation but that's a bit easier to homebrew than PECVD)12:58
azonenbergonly downside for my application is that it's (100) orientation and not (110)12:58
azonenbergbut it might be able to live with that12:58
azonenbergi think using laserlab masks and that, i could make a comb drive with my current lab setup plus the evaporator on campus12:59
azonenberga full wafer of them, no less12:59
azonenbergI'd just need some hot phosphoric acid to go through the nitride... let's see13:02
azonenbergThermal oxide in KOH is 7.7nm/min, or 23nm/min in room temp 10:1 HF13:02
azonenbergPhosphoric barely touches it13:02
azonenberg0.18nm/min13:03
azonenbergNitride in HF is slow (<1 nm/min) but phosphoric is 2.7nm/min13:03
Sync_oh dual side polished13:04
azonenbergactually that's for Si-rich, stoichiometric is 4.5nm/min and high-RI is 20nm/min13:04
azonenbergYes, and coated on both sides13:04
azonenbergSo let's assume stoichiometric for now13:04
azonenberg50nm means just over ten minutes in the wet etch13:05
azonenbergNegligible undercut for MEMS-sized features13:05
azonenbergten minutes in hot phosphoric13:05
azonenbergThen switch to HF to go through the thermal oxide, using the nitride as a mask13:05
Sync_quite expensive tho13:05
azonenbergassuming we use my very weak (~2%) HF but slightly heated it should be about the same, 23nm/min13:06
azonenbergso 300nm will be around 15 minutes13:06
azonenbergWhich will also eat about 5nm of the nitride, but not enough to matter13:06
azonenbergNow we have bare Si exposed and can do a KOH etch13:06
azonenbergAnd yes, but it means i can do MEMS work with my current lab13:06
azonenbergno need for a furnace or CVD etc13:06
azonenbergnot even spin-on glass13:07
azonenbergas long as i can live with that particular oxide-nitride stackup13:07
azonenbergnow to make sure 10 minutes in hot phosphoric won't kill my photoresist...13:07
Sync_I should ask if I can use the other cleanroom when nobody is processing there13:07
azonenbergLooks like some resists peel in hot phos13:07
Sync_I could make you some wafers for a lot less :D13:07
azonenbergetch rates of between 50 and 100nm/min13:08
azonenbergso ten minuts would be 0.5 to 1 micron13:08
azonenbergsignificant, but not enoguh to have breakthrough13:08
azonenbergthis might actually work13:08
Sync_yeah13:08
azonenbergi might want to get myself HMDS to prime it, and i still need a better spin coater13:08
azonenbergBut if i could design myself a comb drive that would work with <100> wafers using a KOH etch13:08
Sync_we had to vent the MBE yesterday :/13:09
azonenbergwhy?13:09
Sync_wiggle stick got stuck13:09
azonenbergAnd lol, it would be nice if you could make me some wafers but i think thats unlikely13:09
Sync_the threaded rod that actuates the wafer handler13:09
Sync_got unscrewed13:09
Sync_(that happens regularily)13:09
azonenberglol13:09
Sync_but is not an issue because you can screw it back in13:09
azonenbergonce you vent it13:10
Sync_but we found out if the wafer is in the clamp you cannot screw it back in13:10
Sync_no it is on the air side13:10
azonenbergoh, hmm13:10
Sync_but when the wafer is in the clamp the hole is offset at an angle13:10
Sync_so you have to vent13:10
azonenbergoh, i see13:10
Sync_I should be able to13:11
Sync_we have a company inhouse that does research on furnances13:11
Sync_so we have shitload of unused ones13:11
azonenbergof unused wafers? lol13:11
azonenbergwith known films on them?13:11
azonenbergMy current thinking is13:11
azonenbergcontact litho using LaserLab mask13:12
B0101hi, does anyone know how to polish wafers?13:12
Sync_we have some shitty wafers we don't want to use for experiments13:12
azonenbergi can get by with pretty loose alignment13:12
Sync_I can use them for free13:12
azonenbergfront to back13:12
Sync_and oxidizing is not too bad13:12
azonenbergthen front to metal will be pretty loose tolerances too13:12
Sync_B0101: on a lapping machine13:12
azonenbergWell, i need oxide + nitride13:12
azonenbergKOH eats Si for breakfast and likes SiO2 for dessert13:12
azonenbergbut nitride is too crunchy13:13
Sync_I'll ask around13:13
Sync_but I think it'd be possible13:13
azonenbergAnyway, no immediate need13:13
azonenbergi want to try with the MTI wafer since thats something i can readily source13:13
azonenbergand that's contingent on me getting a better spin coater13:13
Sync_and haha, the ebeam evap control got a bit out of hand on the other MBE13:14
Sync_and melted a bit of the crucible13:14
azonenbergI know a guy wh odid that here13:14
Sync_it was quite annoying to pound that Si block out of there13:14
azonenberggot in some trouble for it :p13:14
B0101hmm, I am not sure about getting wafers off ebay13:14
azonenbergSync_: http://www.tedpella.com/chemical_html/chem3.htm#anchor31221613:14
azonenbergthats not a bad price13:14
Sync_yeah that is acceptable13:14
azonenbergthen i need to find more photoresist13:15
Sync_well yeah azonenberg but the machine did it on itself13:15
azonenbergmine is probably degraded by now13:15
azonenbergbeen sitting aroudn a long time13:15
azonenbergAny idea where to buy resist?13:15
Sync_now we have added an end stop so that cannot happen anymore13:15
Sync_good question13:15
azonenbergi had been using PCB-grade13:15
azonenbergbut thats meant to stick to metals13:15
azonenbergnot oxides13:15
Sync_spi supplies?13:16
azonenbergtedpella doesnt have it13:16
azonenberglet me try spi13:16
azonenberghttp://www.2spi.com/catalog/chem/chem2a2.shtml13:16
azonenbergok, thats a better price on HMDS13:16
azonenberg$78 for 250ml from ted pella or $15.62 for 100ml from SPI13:17
Sync_yeah13:17
Sync_spi usually has prices that are insane13:17
Sync_but that's ok13:17
azonenbergNo photoresist though13:17
Sync_but their website is very VERY web -1.013:17
azonenberghttp://www.krackeler.com/products/sigma/ALDRICH/654892?highlight=photoresist13:19
azonenbergNegative resist13:19
azonenberghmm13:19
azonenbergwith proprietary thinner/developer/stripper13:19
Sync_yeah13:19
Sync_notwant13:19
azonenberghttp://www.lithotalk.com/13:20
azonenbergooh, good forum13:20
Sync_ah nice13:20
Sync_oh yeah13:21
Sync_we get our supplies from microchemicals13:21
azonenbergYes, but do they sell direct?13:21
Sync_idk13:21
Sync_oh btw, someone managed to smash a view port on a vacuum chamber while moving something13:22
azonenberg:o13:22
azonenberghow bad was it?13:22
azonenbergi assume the wafer(s) got shredded13:22
Sync_left quite large scratches in the chamber13:22
Sync_there was nothing in it13:22
Sync_it was just pumped to prevent contamination13:22
azonenbergoh13:22
azonenbergi thought you meant they bumped a sample into it13:22
azonenbergand scratched the inwodw or something13:22
azonenbergthe window*13:22
Sync_no13:23
Sync_they moved a shelf and hit the viewport+13:23
Sync_B0101: buying wafers off ebay is not a good idea13:23
Sync_even buying from reputable companies that distribute wafers from bankrupt fabs sell crap13:23
Sync_we have a LOAD of wafers that are completely useless13:23
Sync_some have really bad metal ion contamination, others are not doped and again others have all the wrong conductivity13:24
Sync_but they were cheep13:25
Sync_:D13:25
azonenberglol13:25
azonenbergnice for mechanical / litho testing though13:25
azonenbergdeposition process development13:26
azonenbergetc13:26
Sync_we do rate calibration on them13:26
azonenbergYe13:26
azonenbergYep*13:26
azonenbergGood for something still then :p13:26
azonenbergSo i think i might try to stick with my PCB resist a bit longer13:26
azonenbergmaybe buy another bottle from the same vendor13:26
Sync_yep13:26
azonenbergthe solvent chemistry looks very similar to normal electronic resist13:27
azonenbergits probably not trace-metal grade13:27
azonenbergbut who cares about metal ions for MEMS :D13:27
Sync_yeah13:27
azonenbergSo i think the next step would be to make a contact mask aligner13:27
azonenbergsuitable for a single 2" wafer13:27
azonenbergand a spin coater13:27
Sync_oh, I bought a mill btw13:28
azonenbergooh13:28
azonenbergi think when i have osme time i'll develop a full process for this13:30
Sync_meh I should start working on the spin coater idea I had13:31
azonenbergi was going to make one but got busy with school... silly thesis :p13:33
azonenbergall of this fab work is totally unrelated to my "real" research in OS/CPU design and security lol13:34
Sync_http://stores.ebay.com/InterLogicTrading?_trksid=p2047675.l2563 uh huh, pakistan he says13:34
B0101Darn: I don't get why certain students actually tried consuming CuSO4, darn, report coming up...13:38
azonenbergconsuming?13:38
azonenberglike, ingesting?13:38
azonenbergthat can't be healthy13:38
B0101ingesting? yes, though according to the other lab tech, the student only consumed a little bit13:39
azonenbergo_O13:41
azonenbergthese are... college students?13:41
azonenbergthey should know better...13:41
Sync_no13:41
Sync_they don'T13:41
Sync_also, why not taste it?13:41
azonenbergDarwin award?13:41
Sync_#yolo *hides*13:41
Sync_but being a college student doesn't mean that you know anything tbh13:42
azonenberglol13:43
Sync_hm CuSO4 was used as an emetic agent13:43
Sync_can't be that bad.13:43
B0101I don't think CuSO4 taste sweet, does it? o_O13:44
Sync_no13:44
Sync_it tastes bitter13:44
azonenbergmakes sense given the sulfate13:45
azonenbergstill, you have to wonder what said person was thinking13:45
soul-di read copper is verry toxic when i was researching it a bit13:45
Sync_it works because it irritates the stomach walls like most cytotoxins13:45
azonenbergsoul-d: copper ions are not exactly healthy13:45
B0101according to the original incident report, the student claims that CuSO4 tasted sweet13:45
azonenbergB0101: ...13:46
Sync_then it was not coppersulfate13:46
Sync_or he has a very strange taste13:46
Sync_heavy metal ions in your body are generally not too good13:46
azonenbergTe lab accidents i normally hear about are things like "poured chemical into wrong beaker, blew up"13:46
azonenbergthe*13:46
azonenbergnot "decided to randomly drink some chemical" lol13:47
Sync_I'm annoyed to some degree by a lot of that safety craze13:49
Sync_especially when it gets in the way of working safely13:49
B0101here, I am shocked: the students don't even get latex gloves when working with chemicals, especially corrosive ones13:51
B0101Imagine handling HCl, or even HNO3 w/o gloves13:52
B0101I have written letters to the safety department, but no responses even after the 3rd letter13:53
Sync_huh13:53
Sync_hcl is not bad13:53
Sync_I have some scars on my hands where NaOH tried to soap up my hands13:54
B0101ah, NaOH... some students even teacher's think that basic chemicals can be treated like innocous chemicals13:55
Sync_it could be worse13:58
Sync_if the solution is not too awfully concentrated you just wash it off your hands13:58
Sync_I can see where the "hurr wear gloves for everything" comes from, but there usually is no real need for them14:05
B0101wow... just to do hydrogen test: 'Placed zinc metal into test-tube, poured conc. HCl into test-tube, student stoppered off test-tube. Started heating up test-tube from bottom,causing test-tube to explode"14:12
B01012 incidents in 1 day, same lab session...14:12
Sync_that is quite an achievent to stopper it off so hard that it explodes14:13
B0101I don't know how a student managed to do that, as all the time we tried that, the stopper pops off the test-tube14:17
Sync_it is possible, probably overheated the tube14:18
B0101no matter what the cause, the lab techs are going to be in trouble once science dept asks14:19
B0101speaking about safety: I believe that labs should be safe, except when people do stupid things14:25
Sync_well, most safety aspects are about training, there are not that many constructive safety measures you can take in an R&D lab14:29
Sync_you can provide proper workspaces, appropiate for the job, like fume hoods and such but apart from that14:29
B0101but then here, the war on terror is getting so absurd that people have difficulty setting home labs here14:32
B0101in fact, if they suspect you being a terrorist, they arrest you and under my country's law, the can detain anyone without trial for 2 years and that may be extended depending on the government14:33
B0101just as long as 1 officer suspects, even without evidence, that still applies14:34
B0101so actually, the so-called safety rules here just hinder the school lab from operating with audits and checks from the government14:38
B0101they think as if some student or lab tech is going to make explosives or something14:38
Sync_but that's what they do :D14:39
Sync_we made all kinds of explosives in chemistry class in school14:39
B0101heck, If I try to do that here, I'm doomed14:41
Sync_way more fun than regular high school chemistry14:42
B0101heh, too bad I have no chance to experience that here14:42
Sync_you just have to have cool teacher14:42
B0101here, law doesn't allow it14:43
Sync_here too14:44
Sync_but who cares14:44
berndji thought the royal navy gave their sailors CuSO4 to prevent them from catching gayitis16:26
berndjour high school science teacher regularly drpped slices of sodium into a tube of bromine16:30
berndjmade a nice bang16:31
berndji can't find a reference for CuSO4 in the royal navy; quite possibly a crappy HS textbook / bogus storytelling from the teacher16:34
B0101hmm, I am not sure how they deposit copper onto boards to make copper boards for PCB making, anyone has any idea?22:38
Sync_electroplating22:40
Sync_a thin foil is glued on in the first place and then the rest is deposited22:41
B0101hmm, I not sure if we can create different layers of CuO and Cu2O to make devices22:43
Sync_wut22:43
B0101I was thinking that since both CuO and Cu2O are semiconductors, if we can create different layers of the 2 on one board, we may be able to make simple things like transistors22:45
Sync_probably not22:45
B0101ok then, and can we melt bits and pieces of silicon to make ingots?22:48
Sync_probably but what would a Si ingot do?22:48
B0101well, I am looking for ways to obtain Si and other semiconductor materials as the supplier I usually go to has increased the price22:52
Sync_manufacturing your own wafers is quite impossible or rather very unlikely due to the cleanliness issues22:53
B0101I am not sure about ebaythough i did find something: http://www.ebay.com.sg/itm/25-Silicon-Wafer-Wafers-8-200mm-Polished-and-Films-High-Resistivity-P-Type-/200831220349?pt=LH_DefaultDomain_0&hash=item2ec2793a7d22:56
Sync_what do you want to do with them?22:57
B0101I want to try to make a simple NMOS transistor22:58
Sync_then just get wafers from ebay22:59
Sync_but do not expect anything from the transistor22:59
Sync_the problem more or less with wafers from unknown sources is that you really do not know what specs they have or how good they are qualitywise23:05
Sync_which are things you want to eliminate in semiconductor manufacturing23:06
B0101darn, now I have to risk buying potentially impure wafers or buy a pack of 25 for $500023:12
Sync_for a very very basic transistor grade of the wafer is not the prime concern23:13
berndjwould a FET of either polarity be easier to make than a BJT?23:15
Sync_yes23:15
Sync_two contacts, oxide, gate23:15
Sync_much easier23:15
berndjB0101, i think i read in a book somewhere that they do some electroless nickel (or even paladium? but maybe that's just for plated vias) on a board and then plate the rest of the thickness with copper23:16
Sync_yep23:16
berndjSync_: metal gate, i'm guessing?23:16
Sync_most fabs use a palladium process23:16
Sync_it doesn't really matter23:16
Sync_use anything conductive23:16
berndjit matters if you want to DIY :)23:17
Sync_but yes a metal gate is probably what you want to use23:17
Sync_no23:17
Sync_:D23:17
Sync_I guess you could use conductive epoxy23:17
berndjpaper + fibreglass epoxy + copper foil :)23:17
berndjthen you can use cheap HSS drills too23:17
berndjyeah, jeri ellsworth used conductive epoxy for her FETs23:17
Sync_I'd just evap some copper23:18
berndj(btw azonenberg you need to get her into this channel!)23:18
berndjevap? sputter? i thought the pcb material might not be so high vacuum compatible23:18
Sync_wat, I was talking about the fets23:18
Sync_pcb material is just so cheap to buy23:19
berndjoooh sorry, got confused there23:19
Sync_there is no need to diy23:19
berndjfor fun!23:19
Action: berndj points Sync_ at the first 4 letters of the channel name23:19
Sync_yes, but there really is no point besides fun23:20
Sync_I do plated vias at home but that's as far as I'd go23:20
B0101wow, even optics shops carry silicon wafers??? O_o23:20
Sync_Si is IR transparent23:20
berndji might still make DIY PCB just because it makes drilling easier, because then i can use paper, for example. HSS-friendly. HSS drills seem to come smaller than TC drills (while still fitting my budget)23:20
berndjyou plate vias with that palladium process?23:21
Sync_yes23:21
berndjcool!23:21
Sync_you can just get phenolic pcbs23:21
berndjdidn't think you could still get those (but i'm a n00b)23:21
Sync_and carbide tooling is not too expensive23:21
berndjwell, i can buy 0.3mm drills for pocket change at a local shop; i wouldn't even know where to get the same size in carbide23:22
berndjnever mind what it'll cost me23:22
Sync_a very high quality one will run you 14¬23:24
Sync_but it'll last a long time23:24
Sync_you could also resharpen the drills23:24
berndj14 euro??? where!?23:24
Sync_hoffmann23:25
berndjlast quote i had from a local supplier was R700 (just under $100 i guessconvert) for a 0.5mm carbide drill23:25
Sync_R?23:25
berndjsouth african rand23:25
Sync_58¬23:25
Sync_that is quite a ripoff23:26
berndjit's possible it's optimized for different uses, namely metal fabrication, not making holes in PCBs23:26
Sync_a .3mm HSS drill should run you around 6¬ in a good quality23:27
Sync_the only thing that can be changed is the cutting geometry23:27
berndjyeah, something close to that, not 5x more23:27
Sync_and that does not matter too much when drilling pcbs23:27
berndjand the length of the spirally bit, forget its name23:27
berndjthe flutes23:28
Sync_yeah the flute angle is something to be considered but for hobby production it does not matter at all23:28
berndjguy i spoke to said they can go up to 70:1 l:D ratio (which is wholly unneeded for PCB)23:28
Sync_yeah it is possible but only with specialized geometry23:28
berndji mean the *length* of the flutes23:28
berndjwhich at 0.3mm would be a 21mm hole23:29
Sync_yeah with very small bits it tends to be a problem but with regular pcbs every drill should cut it23:29
berndjdo the PCB shops drill one board at a time or do they drill a whole stack at a time?23:30
Sync_haha23:32
Sync_they do one at a time23:32
Sync_drilling a stack will break your drill23:32
berndjand they run their drills at an unholy rpm, right?23:34
Sync_yeah23:34
Sync_sometimes over 100000rpm23:34
Sync_but that's nothing too special23:34
berndjno? lol, so what is?23:34
Sync_well you can buy spindles made for that on every corner23:35
berndji still need to build a PCB drill. biggest obstacle is figuring out how to chuck it23:36
berndjprobably a collet is easiest23:36
Sync_just get a dremel or proxxon tool23:36
berndjlol, you must live in a first-world country23:36
berndji thought those tools have rather suboptimal runout?23:36
Sync_they work fine for diy pcbs23:37
Sync_I just call kavo or gmn and they are happy to sell me a spindle that can do it23:38
berndjthe joys of being in a first-world country :-/23:39
Sync_(I actually do not want to pay their prices but it'd be possible)23:39
Sync_I'd think they'd be happy to ship to south africa23:40
berndji'm sure. i just don't feel like running the credit card gauntlet. (i don't have one (yet?) and don't really want one - rarely miss having one)23:41
Sync_but really a dremel tool or such will do it for you23:41
berndji guess that's my fallback. those are easy to get here23:42
Sync_we use a mb140 and the proper drill for it at our hackerspace and it is very good23:43
berndjin what city are you btw?23:44
Sync_hannover23:45
berndji'm also german23:49
Sync_I suspected that23:49
Sync_:D23:49
berndjmy family's from ex-germany though (bautzen and further east even)23:49
Sync_cool23:50
Sync_that's close to my home town23:50
berndjjust don't say uhyst?23:50
Sync_not that far east but colditz23:51
berndji recognize that name23:51
Sync_close to grimma23:52
Sync_how did you end up in SA?23:52
berndjmy mother's family had a schloss in pannewitz 100+ years ago, but i guess we gambled the estate away or something23:52
berndjwe ended up in the SA because of the 7 day war! suez canal was closed so the tug my dad worked on had to go via cape town23:53
berndjfell in love with the place and decided to move here23:53
berndjanyway, i'm jealous of you first world people with your proper hackerspaces :)23:55
Sync_move back to the civilized world  :P23:55
berndjwe have one here, but it only just got started and it's mostly just software people. i guess we don't have the density of engineers that you have in europe / north america23:55
Sync_most of the people in hackerspaces do not have an academic engineering background23:56
berndji mean "engineer" more broadly - maker types23:56
--- Sun Feb 10 201300:00

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