#homecmos IRC log for Monday, 2011-09-19

Action: azonenberg nibbles idly on the <100> face of a sucrose crystal01:54
pepsi`(_)o(_)02:00
azonenbergpepsi`: you didnt know about this channel?02:01
pepsi`no02:01
pepsi`what kind of dastardly chemicals do i need?02:01
azonenbergNot that bad actually02:01
azonenbergfor the metal patterning test i did, you need copper, high vacuum and high voltage, photoresist, drain cleaner, pool acid, and drugstore-grade hydrogen peroxide02:02
azonenbergWe're still working toward a full IC process, its coming along but slowly02:04
pepsi`can i do this in a spare bedroom?02:11
pepsi`where am i going to get high voltage?02:11
pepsi`and a high vacuum?02:11
pepsi`that sounds noisy02:11
azonenbergpepsi`: high meaning 120 thorugh a variac02:13
azonenbergand the metal evaporation i did in a lab on campus, i am getting a vacuum rig of my own but its not here yet02:14
swkhani wonder if the glue from copper tape messed with my cvd growth05:04
swkhanhow would you get rid of the adhesive? i used isopropanol and acetone05:04
swkhanjust standard tedpella copper tape05:05
azonenbergHmm05:17
azonenbergSingle or double sided?05:18
azonenbergAnd what sort of defects did you observe?05:20
azonenbergLooking at the MSDS and documentation the tape is using an acrylic adhesive05:27
azonenbergacetone dissolves acrylic nicely05:27
swkhanazonenberg: single sided05:30
swkhanwell i expect a thin film05:31
swkhani'm getting clusters05:31
azonenbergWhat are you depositing05:31
azonenbergand is the tape the source or the substrate?05:31
swkhanzinc oxide onto the copper tape05:32
azonenbergBy CVD05:32
swkhanyes05:32
azonenbergok05:32
azonenbergHave you deposited ZnO on Cu before?05:32
swkhannope05:32
azonenbergOk, so no baseline to compare to... hmm05:32
swkhani mean, i tried it twice on copper tape cleaned with acetone05:32
azonenbergAnd the same result?05:32
swkhanyes05:32
swkhanditto with ZnO on aluminum foil05:32
azonenbergOk, you can rule out a random fluke then05:32
swkhandefinitely05:33
azonenbergYou got the same result on foil?05:33
swkhanit's different in a way that the clusters aren't as big05:33
azonenbergBut not a film05:33
azonenbergso probably not the adhesive's fault05:33
swkhani repeatedly get weak adhesion of zinc oxide to whatever my substrate is05:33
azonenbergI suspect that you're crystallizing05:33
swkhanon silicon as well05:33
swkhancrystallizing?05:33
azonenbergor something along those lines05:33
swkhani'm trying to read every book in this field =\05:34
azonenbergand that the deposition preferentially takes place where the material is already there05:34
azonenbergyou randomly start depositing faster in one spot (uneven reactant concentrations etc)05:34
azonenbergthen that acts as a seed crystal05:34
swkhani've grown very uniform thin films of aluminum oxide before05:34
azonenbergJust a guess05:34
swkhanusing this same reactor05:34
swkhanthat's possible05:34
azonenbergbut i'd say your chemistry isnt suitable for a thin film05:34
swkhanthat's the problem with this whole field. you have to guess unless you're lucky enough to have insitu rheed or something05:34
azonenbergsomething about your reactants05:34
azonenbergit just isnt stable05:35
swkhani'm using fairly standard chemicals for zinc oxide: diethyl zinc and water05:35
azonenbergidk, CVD is one of the few deposition techniques i have not used on at least one occasion lol05:35
azonenbergi've done sputtering, evaporation, spin coating, dip coating05:35
swkhanyou're still very knowledgeable about it05:35
azonenberglol05:35
swkhanwhat's the difference between sputtering and evaporating?05:35
swkhani've sputtered platinum and gold onto glass and silicon05:36
azonenbergthey both fall under the category of physical vapor deposition, the difference is where the vapor comes from05:36
azonenbergSputtering you do in a rarefied inert gas05:36
azonenbergglow discharge through it to form a plasma05:36
swkhanno spin coating. i'm somewhat worried about film quality degradation from any sort of photoresist technique (top-bottom approach)05:36
azonenbergthe gas ions hit the target, the kinetic energy knocks metal atoms off05:37
swkhanright05:37
azonenbergand they then hit your target05:37
swkhanexactly05:37
azonenbergyour sample*05:37
swkhanyeah i read ya05:37
azonenbergthe mean free path is small enough that there is some scrambling05:37
azonenbergatoms dont travel in a straight line, they bounce a bit05:37
azonenbergso deposition tends to be pretty conformal05:37
azonenbergEvaporation you heat up the sample05:37
swkhanoh that's simpler05:38
azonenbergeither by placing it on a tungsten filament and running current through it, or by hitting it with an e-beam05:38
swkhanprobably requires a better vacuum05:38
azonenbergThis is done in deep vacuum, yes05:38
azonenberg1E-6 torr range05:38
swkhanwow05:38
azonenbergvs 1E-1 or 2 for sputtering05:38
azonenbergThe mean free path is most of the size of your chamber05:38
azonenbergatoms travel in a straight line05:38
azonenbergso you get shadowing effects05:38
swkhanbtw, i'm specifically using ald. i say cvd, but a lot of times people don't know what it is so i avoid saying ald05:38
azonenbergWhich is good if you are doing, say, lift-off05:39
azonenbergthe sidewalls dont get covered05:39
azonenbergBut bad for, say, filling vias05:39
azonenbergAlso, you can do Al2O3 deposition by sol-gel after spin coating http://www.emulsitone.com/alf.html05:40
azonenbergthey do not have such a solution for ZnO though05:40
azonenbergat least, not from that vendor05:40
azonenbergAnyway so back to your problem - you trying to do ALD of ZnO05:40
azonenbergAnd you're doing this by alternating DEZn with H2O?05:41
swkhanyes05:46
azonenbergMy first guess is that the DEZn prefers to stick to ZnO than your substrate05:47
azonenbergfor whatever reason05:47
azonenbergso when the H2O comes in the ZnO grows by a lot and the film elsewhere by only a little05:47
azonenbergI could be wrong05:47
swkhani don't understand all the modes of growth05:47
swkhanthere's 3 epitaxial growth modes i am aware of05:48
azonenbergBut its now almost 2 AM and i have class tomorrow morning05:48
swkhanah, understandable05:48
swkhanthanks for your help =)05:48
azonenbergKeep talking and i'll read it it when i get back05:48
swkhanoh, okay05:48
swkhanthere's 3 epitaxial growth modes i am aware of: island formation, film formation, island + film formation05:48
swkhanwhat's confusing is that film formation mode has been repeatedly demonstrated by many other people for zinc oxide05:49
swkhanso trying to figure out why i get islands and not film formation is my goal05:49
azonenbergYou know, this channel is starting to turn into #microfab lol11:12
azonenbergswkhan: considering we now have folks like you who are operating out of "real" labs coming here for advice11:12
azonenbergAnd I'm not really sure as to why you'd get islands11:13
azonenbergContamination would be my first guess but if you got good Al2O3 films on the same substrate that's less likely11:13
azonenbergDo you have any photos of the failed depositions?11:19
azonenbergIt'd be interesting to see both the experimental setup and what the offending wafers look like11:19
B0101azonenberg:12:16
B0101i have to tell you something12:17
azonenberg?12:17
B0101Latech has increased its price12:17
azonenberg:(12:17
B0101wafers now cost: $18.95 USD12:18
azonenbergStill not too bad12:18
azonenbergmti is 2x that12:18
B0101and you can order 1 wafer for 29+ USD but only for the first order12:18
B0101but 2 wafers are allowed12:19
azonenbergi see12:19
azonenbergUpdate the iki then i gues12:19
azonenberggtg12:19
B0101i will, cya12:19
CIA-67homecmos r125 | wiki/Vendors.wiki | Edited wiki page Vendors through web user interface.12:26
swkhanazonenberg: aww i missed you! it may be a real lab, but it's very much in its infancy since no one including the professors or grad students know how to help15:54
swkhanazonenberg: i can show you a picture i took with an optical microscope15:55
azonenberg_workswkhan: back15:56
swkhanazonenberg: i think this is 50x http://img838.imageshack.us/img838/1576/zno.png15:56
swkhanwow perfect timing15:56
azonenberg_worklet's see...15:57
swkhani'll check again later today, but i'm pretty sure that's right. the holes you see are the underlying highly doped silicon (i HF etched the silicon dioxide)15:57
azonenberg_workScale is in what?15:57
azonenberg_workmicrons?15:57
azonenberg_worktens of um?15:57
azonenberg_work~1mm FOV, +/- 500, would make sense for 50x15:58
azonenberg_workor 100um, +/- 50, for 500x15:58
azonenberg_workIts usually nice to know what i'm looking at :p15:59
swkhani have SEM pictures as well. let me dig one up from zinc oxide on aluminum foil16:00
azonenberg_workSEM shots are always better16:00
azonenberg_workcheck out http://colossus.cs.rpi.edu/~azonenberg/images/homecmos/2011-06-07/fab001_annotated.jpg for an example of the overlay i use on my optical scope images16:00
azonenberg_worki have two cameras, one calibrated and one not16:01
azonenberg_workthis is from the calibrated one, the other one is my cheap point&shoot i use when scale isnt important16:01
swkhanhere you go: http://img546.imageshack.us/img546/5723/znoonalfoil.png16:03
azonenberg_workOoh, i know that overlay16:03
azonenberg_worki've used the same EDX system16:03
swkhanINCA?16:03
azonenberg_workYep16:03
azonenberg_workThe JSM-840 in the mat sci lab here has it16:03
azonenberg_workits what i've been using for most of my work as it's much cheaper to use than the zeiss supra 55 in the cleanroom16:04
swkhanah, i think this is a Hitachi S480016:04
swkhanso they charge you too?16:04
azonenberg_workThe cleanroom costs a ton16:04
azonenberg_workThe mat sci lab, well, i have a little arrangement with the lab manager ;)16:04
swkhani don't think we even have a clean room =\16:04
swkhani see16:05
azonenberg_workHe lets me in off the book as long as i coordinate with him for a time nobody else needs the scope16:05
azonenberg_workand dont use too much time16:05
swkhancool16:05
azonenberg_workBut he supports my work and recognizes i have no budget16:05
azonenberg_workwhich i'm thankful for16:05
swkhani have built a similar relationship with the lab manager for the sputterer. i guess i have to garner some sympathy and/or make some food for him if i want that kind of deal =)16:06
azonenberg_worklol16:06
azonenberg_worki've used rpi's cleanroom briefly16:06
azonenberg_workmostly for SEM work when i was doing consulting as an undergrad, failure analysis of some chips16:06
azonenberg_workhttp://www.rpi.edu/dept/cie/mncr/16:06
azonenberg_worktoys after toys after toys lol16:06
azonenberg_work10k ft^2 of class 100 space16:06
azonenberg_workits like half a floor of a very large lab building16:06
swkhanwow16:07
azonenberg_workWe have all kinds of fun stuff - wire and wedge bonders16:07
azonenberg_worka sem, a dual-beam SEM/FIB, an AFM16:07
azonenberg_worki dont know if there is a TEM in the cleanroom but there is one in the matsci dept16:07
swkhani just got trained on a TEM and i'm not impressed16:07
azonenberg_workSputtering rig, *two* e-beam evaporators iirc16:07
azonenberg_worki think two e-beam litho units?16:08
azonenberg_workone or two contact aligners, a stepper16:08
swkhanwow16:08
azonenberg_workPECVD/DRIE16:08
azonenberg_worknot sure if we have XeF2 capability16:08
swkhanwhat's xenon difluoride used for?16:09
azonenberg_workDry etching of silicon16:09
swkhancleaning something? etching16:09
swkhanyeah okay16:09
azonenberg_workYou sublime it into a vacuum chamber from crystals16:09
swkhanwhoa16:09
azonenberg_workvapor pressure is a few mtorr i think? i forget the exact number16:09
azonenberg_workexpose your wafer to it, then purge with argon when you're done and repeat16:10
azonenberg_workits isotropic i think16:10
azonenberg_workused for backside etching and stuff when you dont want to risk trace metal contamination16:10
azonenberg_workme, i just use KOH :P16:10
azonenberg_workscrew PN junctions, my wafer is a mechanical substrate16:10
swkhanisn't KOH anisotropic?16:10
azonenberg_workYes16:10
swkhanto silicon any way16:10
azonenberg_workVery16:10
swkhani've never seen it in person16:10
swkhani know we have potassium hydroxide around16:11
azonenberg_worksomething like 500x faster etch of <100> and <110> than <111> planes16:11
swkhanand i have wafers...16:11
swkhanwow16:11
azonenberg_worki use 30% in water16:11
swkhani don't understand why. i'll look that up later16:11
azonenberg_workput in a plastic centrifuge tube16:11
azonenberg_workwith the die16:11
swkhani'm reading this book btw: nanostructures and nanomaterials synthesis properties. it seems pretty readible so far. but it doesn't seem to give me any direct insights into what could be going on with the reactor16:11
azonenberg_workand heat that in a water bath a little below boilin16:11
azonenberg_workinteresting16:12
azonenberg_workDo you have any low-angle SEM shots at higher resolution?16:12
azonenberg_workI'm interested in what the edge profile of your clumps look like16:12
azonenberg_workI take most of my microscope pics at max resolution because disk space is cheap16:13
azonenberg_workand i might need the data later16:13
swkhanagglomeration and/or oswald ripening maybe16:13
swkhani don't have any of those yet16:13
azonenberg_workso all of my sem pics are ~2048x153616:13
swkhanwhat would the edge profile tell you?16:13
azonenberg_workWell, more importantly16:13
azonenberg_workif you do several depositions and image the same area in between (put some kind of a marker down on the wafer)16:13
azonenberg_workyou can observe how it grows16:13
azonenberg_workmostly up, mostly out, etc16:13
azonenberg_workand whether it's starting to grow in new spots16:14
azonenberg_workor only over the existing globs16:14
swkhanhmm16:14
azonenberg_workThat might provide some insight as to whats going on16:14
azonenberg_workI did something similar when trying to diagnose a contaminant in my litho16:14
azonenberg_worki imaged the same part of the die at each step of the process, after every deposition/etch16:15
azonenberg_workthen went back and looked for differences16:15
swkhani'm not sure i would know how to analyze the data16:15
swkhanwhat would i be looking for? i still don't even fully feel like i know how stuff grows16:15
azonenberg_workhttp://www.drawersteak.com/downloads/chippics/pic16f59/pic16f59_005.jpg is an example of a pic i took on the Zeiss in the cleanroom btw16:15
azonenberg_workthis is SiO2 coated Al wiring on top of an IC with some packaging particles between them16:16
azonenberg_worknote the resolution though16:16
azonenberg_workAnd i dont know how it grows either16:16
azonenberg_workThats the point16:16
azonenberg_workYou want to *watch* it grow16:17
azonenberg_workget a sequence of five or ten pics of the same clump over time16:17
azonenberg_workIn ALD you are alternating the two precursors16:17
swkhanyes16:17
azonenberg_workSo pull it out after 10, 100, or however many iterations you think is reasonable16:17
azonenberg_workimage16:17
azonenberg_workput it back in, hit it a few more times16:17
azonenberg_workimage16:17
azonenberg_workAnd use some kind of a feature on the die to make sure you are looking at the same region16:17
swkhanbut then i introduce another variable don't i? if i break vacuum and image, then i don't know if the problem is due to contamination with the outside world or due to a failure of the ALD16:18
swkhana postdoc in my lab says he used to use a scribe to etch out a letter or some kind of mark on the shiny side up and then grow16:18
azonenberg_workYes, you do - but the goal here is not necessarily to completely replicate the existing setup16:18
azonenberg_workit's to understand it16:18
azonenberg_workSo as long as you keep that in mind you're ine16:18
azonenberg_workfine*16:19
azonenberg_workFor example, you may adsorb some water onto the surface16:19
azonenberg_workYou may get some particulates16:19
azonenberg_workre water, thats used in your process anyway so you should be ok16:19
azonenberg_workparticulates, just try to keep them to a minimum16:19
azonenberg_workwhen you remove from the reactor, put in a cassette immediately and then image16:20
azonenberg_workAlso, image the wafer before you start deposition16:20
azonenberg_workto get a baseline16:20
azonenberg_workFor example, if you are scribing marks into the surface you'll get silicon chips kicked up by that16:20
azonenberg_workyou need to make sure they're removed16:20
azonenberg_worki'd suggest sonicating in ethanol16:20
azonenberg_workbefore you do the deposition16:21
swkhanwould methanol work? or isopropanol? i don't have ethanol in the lab =(16:22
azonenberg_workyou see what i'm saying here? You want to understand potential failure modes and test for them16:22
azonenberg_workand methanol should work fine16:22
swkhani see16:22
azonenberg_workthe ethanol i use actually is denatured with ~5% methanol and some other stuff16:22
azonenberg_workthe "other stuff" has caused problems in some processes lol16:22
azonenberg_worki wouldn't mind 95/5 ethanol/methanol16:22
azonenberg_workstill undrinkable and untaxed, but all volatiles of similar characteristics16:22
azonenberg_workthey add all kinds of other stuff to hardware store denatured alcohol though16:23
azonenberg_workhas caused problems when i used it as a solvent for sol-gel deposition16:23
azonenberg_workActually, on that note16:23
swkhani think my current problem is the "understanding the potential failure modes" part. let's say i find even patchier growth initially16:23
azonenberg_workYour substrate is silicon?16:23
azonenberg_workor alumina16:23
swkhanthe SEM image was on aluminum foil (so i guess alumina)16:24
swkhanand the optical picture was on highly doped silicon with the native oxide removed16:24
azonenberg_workOk16:24
azonenberg_workDid you do a full RCA clean on the silicon?16:24
swkhannope16:24
azonenberg_workThat will help to eliminate some variables16:24
azonenberg_workAre you familiar with the process?16:24
swkhani am aware of it. use sulfuric acid, nitric acid, hf, and water... maybe a few more things16:24
azonenberg_workRemove particulates by sonicating in ethanol or similar (especially if you scribed it)16:24
azonenberg_worknope16:25
azonenberg_workoptional degrease in acetone if you suspect organic contamination16:25
azonenberg_workthen soak in water for a while16:25
swkhanit's organic contaminant removal then chemical oxide growth and chemical oxide removal right?16:25
azonenberg_workYes16:25
azonenberg_workSo SC1 is ammonia and water, heated16:25
swkhanah16:25
azonenberg_workand H2O216:25
azonenberg_worki use 1 part household ammonia to six parts drugstore-grade 3% H2O2 at around 70C16:25
azonenberg_worki think16:26
azonenberg_workcheck my lab notes for exact temp16:26
azonenberg_workthen you strip native oxide in HF, i use 3% Whink brand rust remover16:26
azonenberg_workthen trace metal removal in SC216:26
azonenberg_work1 part conc. HCl : 6 parts 3% H2O2 at ~70C16:26
azonenberg_work10 minutes each for SC1 and SC2, the HF is around 30 sec16:26
swkhanwill RCA clean really eliminate variables? what variables would i have right now that would mess me up?16:26
azonenberg_workThese are not the original RCA proportions, they're modified for the chemicals i have available, but funciton similarly16:27
azonenberg_workWell, you are removing trace metals and any possible organic residues16:27
azonenberg_worktrace metal is less critical for you16:27
azonenberg_workbut organic residue could provide either a crystal growth site16:27
azonenberg_workor inhibit the deposition by making the precursors not stick16:27
azonenberg_workAs a bare minimum i'd suggest a dip in acetone followed by isopropanol, then blow dry16:27
azonenberg_workALso, you said you were using aluminum foil16:28
azonenberg_workLab grade? Kitchen grade?16:29
azonenberg_workSome of them have organic release agents on them, they are not just Al + Al2O316:29
azonenberg_workI didnt see any organics in the EDS but did you actally check the spectrum for them?16:29
swkhanwhat is the organic residue?16:29
azonenberg_workAll kinds of stuff16:29
swkhanazonenberg_work: what showed up was from autoid. i don't think i explicitly checked16:29
azonenberg_workhere's a photo of some16:30
swkhanlike if you dip in acetone and then bake it off, wouldn't you just have no more (CH3)2CO?16:30
swkhanwhy would you have any residue left over?16:30
swkhani have lab grade aluminum foil16:30
swkhanwhy acetone then isopropanol by the way?16:31
azonenberg_workAcetone dissolves contaminants well16:31
azonenberg_workbut is so volatile it tends to redeposit16:31
azonenberg_workso you swish in IPA, which is less volatile, to dissolve the contaminated acetone out16:31
azonenberg_workotherwise you get streaking, i can show you pics of it16:31
azonenberg_workAnd lab grade foil probably shouldnt16:32
azonenberg_workWere you wearing gloves?16:32
azonenberg_workhttp://colossus.cs.rpi.edu/~azonenberg/images/homecmos/2011-07-11/die_f6/eds06.htm is an example of an organic particle i found on one of my dies16:32
azonenberg_worklooks like it's biological origin16:32
azonenberg_workNaCl plus some Ca and Zn, lots of carbon16:32
azonenberg_workthen Si seen through it16:33
azonenberg_workprobably dead skin cells, aka ordinary room dust16:33
swkhanit's not a clean room setup, but yes, i was wearing gloves16:33
swkhani cleaned the copper tape with IPA then acetone16:33
swkhanmaybe resulted in some problems16:33
azonenberg_worki'm not in a cleanroom either :P16:34
swkhanthe lab grade foil was not cleaned at all16:34
azonenberg_workBut i understand the implications of it and take reasonable precautions to at least compensate16:34
azonenberg_workfor example i wear a lab coat and hair net to reduce dust coming off me16:34
azonenberg_workThat may be overkill16:34
azonenberg_worki also try to avoid ever leaning directly over a sample16:34
azonenberg_worketc16:34
swkhansure sure16:34
azonenberg_workAnyway so contamination is possible but you cant be sure at this point16:35
azonenberg_workThats why i suggested imaging the wafer before deposition16:35
azonenberg_workor whatever your substrate is16:35
swkhani'm confused though... why does acetone redeposit if it's volatile16:35
azonenberg_workIt does not redeposit16:35
azonenberg_workit evaporates16:35
azonenberg_workwhatever is dissolved in it, though...16:35
azonenberg_workhas to go somewhere16:35
swkhanoh i see16:35
azonenberg_workSo you want to dissolve the contaminant in acetone16:35
azonenberg_workthen remove the acetone rather than letting it evaporate16:36
azonenberg_workNever let a cleaning solvent evaporate off yoursample16:36
swkhanhuh, that makes sense16:36
azonenberg_workEither submerge it in something else and then let the pure solvent evaporate16:36
azonenberg_workOr blow dry16:36
azonenberg_workwhen i say blow, i dont mean forced evaporation16:36
azonenberg_worki mean use high pressure N2 etc to physically knock the droplets off16:36
azonenberg_workbefore they evaporate16:36
swkhani've seen an N2 gun. i know what you mean. i could hook that up16:37
azonenberg_workIdeally, blow dry after submerging in fresh solvent16:37
swkhani see16:37
azonenberg_workSo first you dilute the contaminated stuff, then you remove it before evaporaiton16:37
azonenberg_workThe result of letting it dry typically is a white residue consisting of salts etc16:37
azonenberg_workin a ring around the perimeter of the drop16:37
swkhanwhat if i don't clean the wafer at all?16:38
swkhanjust use it as is16:38
swkhanor... hmm i cut it so there's silica dust now16:38
azonenberg_workExactly16:38
azonenberg_workas a minimum you need to remove the particles16:38
azonenberg_workThat dust was killing my yields for several weeks16:39
azonenberg_workbefore i figured it out16:39
azonenberg_workEDS wouldnt show the contaminant because it was the same material as my substrate16:39
azonenberg_worki was bashing my head on it for a while16:39
azonenberg_worki tried RCA cleanin (which doesnt remove large dust, only trace films etc)16:39
azonenberg_worki tried wearing a surgical mask and hair net16:39
swkhanhuh, wow16:39
azonenberg_workthen finally i tried imaging a bare die before processing16:39
azonenberg_workand i saw silicon dust all over it16:39
azonenberg_workthat wasnt on an uncut wafer16:40
swkhanoh wow16:40
azonenberg_workSo this is a very real potential problem16:40
azonenberg_workif you didnt remove particles from your silicon i'd blame that right away16:40
azonenberg_workunless you were doinga full uncut wafer16:40
azonenberg_worklike i said, image the sample - however big it is - at low angle16:40
azonenberg_worklook for dust16:41
swkhanno i definitely cut the wafer16:41
azonenberg_workin secondary electron images they will show up as bright spots16:41
swkhanso sonicate it off?16:41
azonenberg_workYes thats one way16:41
azonenberg_workUse somethign with low surface tension16:41
azonenberg_workabsolutely not water16:41
swkhantake a little test tube and put some organic solvent in (acetone preferably?)16:41
swkhanlow surface tension...16:41
swkhanmeaning it tends to evaporate easily?16:41
swkhanor doesn't bead up i guess16:41
azonenberg_workno, as in it forms a film rather than beading up16:41
azonenberg_workBecause if it beads up it concentrates particles inside as it does so16:42
swkhanthat makes sense16:42
azonenberg_workif it forms a film the dust slips off16:42
azonenberg_workAlso, if you have any lint-free swabs, kimwipes, etc16:42
swkhani don't think they are called test tubes, but what are those things that have a lid on top16:42
swkhani have kim-wipes16:42
azonenberg_workyou can try soaking it in solvent and gently wiping the die with one16:42
azonenberg_workyou dont want to use much pressure16:42
azonenberg_workjust drag the corner over it16:42
azonenberg_workto avoid scratching16:42
swkhanhmm16:42
azonenberg_workThen optically image at ~400x to verify you are free of dust16:42
azonenberg_workand you mean centrifuge tubes?16:43
swkhanhow well does dust stay off?16:43
swkhanyes centrifuge tubes16:43
swkhanthat's what i was thinking of16:43
azonenberg_worki <3 those things16:43
swkhanhow long would you sonicate for? the postdoc seems to always wait 60 minutes16:43
azonenberg_workbought a case of 2000 ~1ml ones and i use them all the tie16:43
azonenberg_workmy glass test tubes havent been touched sin ethe day i opened the box lol16:43
azonenberg_workthese things are small, nonbreakable, disposable16:44
azonenberg_worknot etched by HF16:44
swkhanlol. there's a guy in ##chemistry who was desperate for some and was willing to buy from me16:44
swkhanglass not etched by HF? o_O16:44
swkhandiluted HF?16:44
azonenberg_workno, i mean the centrifuge tubes16:44
azonenberg_workAnd i actually have little experience with sonication as i dont have an ultrasound rig myself yet16:45
azonenberg_workits on the to-buy-once-this-paycheck-clears list16:45
azonenberg_workalong with a few undoped 2-inch <110> wafers16:45
azonenberg_workI can say, unless you are worried about fragile MEMS structures being damaged (certainly not the case with a blank wafer)16:45
azonenberg_workmore can never hurt16:45
swkhanwell it can in terms of time wasted16:47
swkhani wish i had a good scientific explanation for all the actions i took in lab16:47
swkhanyou seem to complement what i've been reading very well =)16:48
azonenberg_workSome of the cleaning stuff is empirical because you dont know what you are targeting16:48
azonenberg_workcontamination is often an unknown16:48
azonenberg_workSo you think "what will it take to remove XYZ?"16:48
azonenberg_worktry it16:48
azonenberg_workif it works, odds are XYZ is your contaminant16:48
azonenberg_workand you now know how to remove it16:48
azonenberg_workAssuming your problem is contamination16:48
azonenberg_workwhich i consider likely but far from definite16:48
swkhanso okay let's consider this situation... i go down to the optical microscope + raman and see no dust. take it to the SEM and see no noticeable silica dust, proceed to ALD without any cleaning or maybe a water soak?16:49
azonenberg_workSounds reasonable if you are worried about dust16:50
azonenberg_workBut i'd suggest an acetone + IPA dip to rule out oils etc16:50
azonenberg_worksay, you brushed a finger of you glove against it that had touched your bare hand while putting it on16:50
azonenberg_workthat might leave fingerprint oilis16:50
azonenberg_workoils*16:50
azonenberg_workThen go on to the ALD and image after say 20% of your normal deposition time16:51
azonenberg_workput back in, run another 20%, image16:51
azonenberg_workobserve some things16:51
swkhani'm fairly certain i didn't do that, but if i just sprayed the acetone onto the silicon sample and then used a kim wipe to wipe away the contaminants, what would you say?16:51
azonenberg_work* How soon does it clump? Are you getting a film that degenerates into clumping?16:51
swkhansure16:51
azonenberg_work* How do the clumps grow? DO they form early on and just get bigger, or do new ones appear? When they grow, do they grow out or up or both?16:52
azonenberg_workand i'd suggest dipping it in a beaker and then blowing dry16:52
azonenberg_workthats what i usually do16:52
swkhani'm trying to arm my self with as many tools as possible ahead of time16:52
azonenberg_workradioshack duster spray, i kid you not16:52
swkhanregarding dipping, doesn't that result in more contaminated acetone?16:52
azonenberg_workis my poor man's solution for the lack of an N2 gun16:52
azonenberg_workandif you swish it around16:52
azonenberg_workyou dilute the contaminant a lot16:52
azonenberg_workthen you blow it off16:53
swkhanso let's say i get clumps that form early on and just get bigger. that's ostwald ripening right?16:54
azonenberg_worki dont know the technical term off the top of my head16:55
swkhanit's where little things form bigger things because it is thermodynamically more stable to be that way16:55
azonenberg_workYeah16:55
azonenberg_workThats my guess as to what might be going on if you are not having contamination issues16:55
swkhanso what does that imply? i get frustrated by knowing this theoretical stuff and not knowing how to apply it16:56
azonenberg_workwell, brief readin suggests ostwald ripenin is mostly a liquid phase process16:56
swkhani'm going to sputter zinc on quartz today. maybe using purer copper foil too16:56
swkhani should get the heck into lab. it's 10 am!16:56
azonenberg_workits 13:00 here lol16:57
azonenberg_worki'm midway between my two classes of the day16:57
azonenberg_workand also, here's a question16:57
azonenberg_workYou desire a ZnO thin film on something16:57
azonenberg_workMust it be done by CVD?16:57
azonenberg_workOff the top of my head i'm wondering if reactive sputtering might work16:57
azonenberg_worki know its done for nitrides and i think oxynitrides16:58
azonenberg_workBecause that should be quite uniform16:58
azonenberg_worknot saying to abandon all your work, but at least consider alternatives if you cant get this process to work16:59
swkhanazonenberg_work: that's something i was planning on trying as well. sputtering zinc oxide as a seed on top of various things... copper, aluminum foil, sputtered zinc even17:05
swkhanazonenberg_work: my professor wants to be sure that the ALD i put together isn't a fluke with aluminum oxide17:06
swkhanzinc oxide is a weird material though17:06
swkhani guess i'm insecure about it not being single crystal17:06
azonenberg_workI dont know the microstructure17:06
azonenberg_workto me it's sunscreen and diaper rash cream :p17:06
azonenberg_workBut as a metal oxide i'd assume its amorphous17:07
azonenberg_workALD is normally used as an epitaxy technique so i'm a little confused as to how it works with amorphous materials17:07
swkhanbut if it's epitaxial growth, and i have silicon (100) underneath, shouldn't i get ZnO (100)?17:11
swkhanbut... since zinc oxide does not have a cubic crystal structure, it's a bit weirder17:11
azonenberg_workExactly, thats why its wierd lol17:12
azonenberg_workNot only is it not the same structure17:12
swkhanhmm i wonder about alumina then17:12
azonenberg_workthe lattice spacing is different17:12
azonenberg_workThe alumina is obviously not monocrystalline17:12
azonenberg_workMy guess is that you are growing at grain boundaries or something like that17:12
azonenberg_workor lattice defects17:12
azonenberg_workwhere you have a couple of atoms aligned at the right spacing17:12
swkhanoh what the heck it's trigonal Al2O3?17:13
azonenberg_workthen you nucleate there and it starts growing epitaxially on top of that little speck of ZnO17:13
swkhanthat seems intuitive to me as well17:13
azonenberg_workI'm now starting to suspect that might be your problem17:13
swkhanthe problem i'm actually trying to address is why i get poor adhesion17:14
azonenberg_workAdhesion, oh17:14
azonenberg_workThats a horse of a different color :p17:14
azonenberg_workFails the scotch-tape test i assume?17:14
azonenberg_workWell, to begin - if you are only growing on lattice defects you are getting strong bonds at a handful of sites17:15
azonenberg_workbut otherwise not very strong17:15
azonenberg_workExplore growth on some other materials17:15
azonenberg_workOff the top of my head, a lot of things do not stick that well to Si17:15
azonenberg_workHave you read "Etch rates for micromachining processing, part 2"?17:15
azonenberg_workThere's a section in there talking about film adhesion of the materials they tested17:16
swkhanazonenberg_work: it fails the pick-it-up-with-a-pair-of-tweezers-without-scrapping-it test =(17:16
azonenberg_workevaporated Cr is a very popular adhesion layer17:16
azonenberg_workTi works too17:16
azonenberg_workand do you mean without it falling off completely?17:16
swkhanhave you figured out why?17:16
swkhanand no i haven't17:16
azonenberg_workOr you scratch it17:16
azonenberg_workMetal tweezers will scratch all kinds of stuff17:16
swkhanazonenberg_work: i can take a napkin and wipe it off17:16
azonenberg_work...17:16
azonenberg_workok, thats even worse than my evaporated copper without Cr lol17:17
swkhanlol17:17
azonenberg_workWithout Cr, tape takes it rigth off17:17
azonenberg_workand it floats off during etch17:17
azonenberg_workwith Cr it sticks well, tape doesnt even touch it17:17
azonenberg_workJust as a totally random idea17:17
swkhanyeah, i still haven't figured out why Cr and Ti are so frequently used as adhesion layers17:17
azonenberg_workEvaporate or sputter Cr onto something17:17
azonenberg_workthen try growth on that17:17
swkhanwhat about platinum? we don't have chromium to my knowledge17:18
azonenberg_workyou'll be gettin a nice polycrystalline layer of a single material17:18
azonenberg_workHmm, i think Pt needs an adhesion layer :p17:18
azonenberg_workCr, Ti, and some other odd materials17:18
azonenberg_workRead the paper i mentioned17:18
swkhanokay17:18
azonenberg_workthey alk about adhesion of a lot of transition metals17:18
azonenberg_workWhat are you studying again btw?17:19
azonenberg_workee? chem? phys?17:19
swkhanEE17:20
swkhanthey just state it makes a good adhesion layer17:20
azonenberg_workYeah, nobody knows exact details - the very strong Cr-O bond is something to do with it i think17:21
azonenberg_workand i just find it kinda funny that i'm helping *you* with this stuff as someone who hasnt taken any EE classes except one on networking :p17:21
swkhani know right. i feel so incompetent =(17:23
swkhani spent the greater part of this year building the ALD17:23
swkhanlots of hardware and software17:23
swkhanno chemistry or physics really17:23
azonenberg_workyeah... lab stuff is something you cant really do in a lecture17:24
azonenberg_workyou have to just go do it17:24
azonenberg_workIn any case, another idea is just to go and find papers talking about ZnO17:25
swkhani've been doing that all week17:25
swkhani forgot to mention earlier17:25
swkhanone paper said that the deposition rate is roughly 2x what i thought it was17:25
swkhanso i may have been depositing 100 nm films rather than 50 nm films17:26
azonenberg_workok17:26
swkhanso i'm going to make a list of stuff i want to do today17:26
azonenberg_workHave you checked if there is any ZnO thin film in the open areas?17:26
azonenberg_workiow, you are getting clumps17:26
azonenberg_workBut is it *all* in the clumps?17:26
azonenberg_workdo you have clumps and blank?17:26
azonenberg_workor is it clumps over fil,17:26
azonenberg_workfilm*17:26
swkhanit could be clumps with a thin film under17:27
swkhanthe copper was more clearly segregated17:27
swkhanhttp://img88.imageshack.us/img88/6938/znooncutape.png17:30
azonenberg_workI cant see too much17:31
azonenberg_workYour contrast is too high and the images are too small17:31
swkhansorry17:31
azonenberg_workJust saying, spend some more time on the SEM and play lol17:32
swkhani had another person begging me to get off on this one17:32
azonenberg_workI'm not the greatest myself17:32
swkhanmy al foil one was way better17:32
azonenberg_worki still have trouble stigmating17:32
swkhanit was small because that was what the edx gave and i didn't take the sem image off17:32
swkhanthe hitachi makes it a bit easier to see if you're getting rid of astigmatism17:32
azonenberg_workhow so17:33
azonenberg_workOh, so something i didnt notice before in your copper tape image?17:40
azonenberg_workYou're not depositing shells over a big clump17:40
azonenberg_workYou're depositing tiny (~500nm) sopheres17:40
azonenberg_workWhich is interesting17:41
azonenberg_worknot sure what it means yet17:41
swkhanme neither17:42
swkhanso things to do today: 1. try to get a "time lapse" view of what is going on 2. properly clean stuff (in lieu of RCA, sonicate with acetone + silicon sample in centrifuge tube sealed and placed in warm water) 3. sputter zinc on quartz and deposit ZnO 4. deposit on very high purity copper17:44
swkhani think i'm going to have to hunt for H2SO417:45
swkhanbut it seems like everyone uses an RCA clean process17:45
azonenberg_workYeah, its the industry standard for silicon cleaning17:55
azonenberg_workbeen used since the 1950s17:55
mrdataHNO3?17:56
azonenberg_workmrdata: hno3 is an oxidizing acid17:57
azonenberg_workyou run the risk of growing too much sio217:57
mrdataoh17:58
mrdatafinishing companies use HNO3 when cleaning aluminum for plating17:58
mrdataHF also, ithink17:58
swkhani think i'm also going to change my deposition temperature17:59
swkhanalso my # of cycles17:59
swkhani got a loaded day ahead of me17:59
swkhanit's 11 am... i'm going to be in lab until 8 pm for sure18:00
azonenberg_worklol have fun18:01
azonenberg_worki'm in an incredibly boring class18:01
mrdataoh? what class18:01
azonenberg_workBut the time passes faster when i'm working on FPGA stuff during the lecture P18:01
azonenberg_workdatabase systems18:01
Action: mrdata is never bored; there's always something to do while listening to dull lectures18:02
azonenberg_workI said the class was boring18:02
azonenberg_worknot that i was bored18:02
mrdataare they teaching about normalization?18:02
azonenberg_worksubtle difference18:02
mrdatatuning?18:02
azonenberg_workand theyre just getting into how to choose keys18:02
azonenberg_workits all noob-level stuff18:03
mrdatamarvelous18:03
mrdatai could teach that18:03
mrdatado they use any UML?18:03
azonenberg_workNot yet, thank god18:04
azonenberg_worki've been exposed to it in software design & doc already18:04
azonenberg_workand want nothing more to do with it :p18:04
mrdatalol. uml isnt terrible at showing data models18:08
mrdatai make heavy of of those in db design18:09
mrdatas/of/use/18:09
berndjin real life nobody uses UML; everyone just draws boxes and random lines on whiteboards18:16
azonenberg_workberndj: yeah18:17
swkhanwell i'm still looking stuff up. seems like temperature is a big deal for the quality of the ZnO crystallinity18:23
swkhanThe degree of preferred orientation (I002/I101) for the ZnO thin film frown at 170 °C is much higher than that grown at 130 °C. -> what does I002/I101 mean? that the 002 plane and the 101 plane are equally preferred?18:23
azonenberg_workhmm18:24
azonenberg_worki think that it prefers 002 to 10118:25
swkhanhuh18:28
swkhanwell this is really insightful18:28
swkhan"It may be said that the crystallinity of the ZnO film grown by ALD with a process time per cycle of 10 s at 170 °C is better than that grown by ALD with a process time per cycle of 4.35 s at the same technique. In general amorphous films are obtained if the film growth rate is high or if the substrate temperature is low in an ALD process. This may be because there is no time enough for reactants to move to equilibrium atomic sites under such conditi18:28
swkhanthat last line!18:29
swkhantotally makes sense18:29
azonenberg_workNice18:30
azonenberg_workSo the cycle times are very important18:35
swkhansorry, bit off topic but i've had this song stuck in my head all day: http://www.youtube.com/watch?v=6WlX7LlpjBI18:35
azonenberg_workno audio on this box lol18:35
azonenberg_workmaybe later18:35
swkhani think i got everything i want18:52
swkhanthanks a lot azonenberg_work =)18:52
swkhantake care18:52
Action: azonenberg_work glances up from far-too-long critial path to see another ER diagram on the blackboard19:36
azonenberg_workgaah19:36
mrdataER diagrams are the bread n butte of DB design19:39
mrdata*butter19:40
mrdatayou must get the relationships the right way around19:40
mrdataand the right cardinality19:40
azonenberg_workYeah19:41
azonenberg_workBut she's beating a dead horse so many times19:41
azonenberg_workit seems like thats been half of the lecture19:42
azonenberg_worktalking about what they are, rather than how to use them19:42
azonenberg_workand lol, given your name19:42
azonenberg_workof course you'd say that :p19:42
mrdataso, maybe she's into sado-necro-bestiality?19:44
azonenberg_worklol19:45
azonenberg_workidk... class is done in 5 mins and thats all i care about now19:45
azonenberg_worki have a mountain of exams to grade :P19:46
azonenberg_workand then i have to pop all of my recent dies under the microscope tonight19:46
azonenberg_workto decide which ones i am going to SEM tomorrow19:46
mrdatanow, it's interesting when you get to the parts of design where privacy constraints trump convenient access19:46
mrdataoh, good.19:46
Action: mrdata wants to see more slides19:47
mrdatai was browsing logs for this channel19:47
mrdatagood stuff19:47
azonenberg_workThe last couple of tests have been 20um half-pitch lines in copper on silicon19:47
azonenberg_workas well as attempts at KOH etching through the copper as a mask19:48
cheaterazonenberg_work: what is a critical path?19:48
azonenberg_workcheater: The portion of a digital system that sets your maximum clock frequency19:48
azonenberg_workin other words, the longest path (in time, not space) between two flipflops19:48
cheateris that just the longest wire so to speak?19:48
cheateryeah19:48
azonenberg_workLongest string of gates plus interconnect wire19:49
azonenberg_workit may not be the most gates or the longest wire19:49
cheateryeah19:49
cheaterby the way19:49
mrdatayou want the critical path to be as short as possible19:49
cheaterERDs are useful, but it's important to understand normal forms too19:50
azonenberg_workcheater: yeah, i know19:50
azonenberg_workup to 3nf just seems like common sense design19:50
azonenberg_workeven 4nf19:50
cheaterif you want a good ERD tool use mysql workbench19:50
azonenberg_workI've used it before19:50
cheateri use it even for things that aren't mysql, because it's so good19:50
azonenberg_worki'm not new to database stuff by any stretch of the term19:50
cheateri know people from the team19:50
cheateroh ok then19:50
azonenberg_worki'm taking the class because its a phd qualifier requirement :p19:50
cheatersorry about lecturing you19:51
azonenberg_worklol no worries19:51
mrdataphd qualifier, really?19:51
mrdatawhat phd area?19:51
azonenberg_workMost of the classes i'm in this semester are absurdly easy for me19:51
azonenberg_workclass is out, will be bac online in half an hour or so19:51
--- Tue Sep 20 201100:00

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