From eap at gloworm.Stanford.EDU Sat Oct 3 19:54:40 2009 From: eap at gloworm.Stanford.EDU (Eric Perozziello) Date: Sat, 3 Oct 2009 19:54:40 -0700 (PDT) Subject: Why vinyl gloves are bad for your wafers In-Reply-To: <4AC7F39E.2040400@snf.stanford.edu> Message-ID: These wands didn't used to have anything other than a metal tube pressed into the quartz tip. They used to rotate around (annoyingly) when they would get loose. I only noticed the "glue" when this rainbow wafer showed up late last year. It was only noticable because the oxide grown was <300A, and thus, the thin film of silicones or whatever produced the rainbow. You couldn't see it on thicker oxides. There should be some coral documentation on this. It was initially reported as an "oxide uniformity problem" as I recall. (I think it might be 11/19/2008 on tylan2, which fits the description perfectly: rainbow on left side of wafer, flat up). This message's "clear" indicates that it was a recurring issue. The "glue" is still on all of the wands (tylan1/2, tylan3, and thermco1/2) as of last night. The application looks as though it's done by hand, not by machine. This was observed on wafers coming out of the furnace, before being picked up "hot". Something transferred to the wafer cold, and then redistributed during the process, from the wand contact point. On Sat, 3 Oct 2009, maurice wrote: > The "sealant" may have been the glue used by the manufacturer to hold > the wand tip to the metal shaft. If wafers are unloaded from the hot > furnaces without letting them cool it melts the glue and can get on the > tip. I have not seen it on the tip but I have seen it the shaft. > > You probably have seen wands that spin on the metal shaft at the oxide > furnaces. If you look close you will see the glue is on the shaft. The > LPCVDs open at a lower temperature so we don't see it as often there. > > We replace the wands as soon as it is reported and we stressed the > importance of letting wafers cool during our trainings. But I would > guess we change 3 wands a week (some of that is from wands breaking or > getting dropped). > > On the photo you attached, is that the front side or the back side of > the wafer? > > I can't imagine any of our staff trying to fix a vacuum wand. It is > easier to get one from the stockroom than putting silicone in that > little hole in the wand. > > I would expect that if anyone spots a damaged or questionable wand, they > would report it on coral and if needed, use clean tweezers to > load/unload their wafers. > > -maurice (and Alexander) > > > > > Eric Perozziello wrote: > > The same thing should go for use of silicones on "clean" vacuum wands. > > > > The attached photo shows a user's wafer that came out of a furnace that > > was loaded with a "clean" vacuum wand. When the user complained, > > the staff response was "Try cleaning your wafers better". But on further > > testing and investigation, it was from sealant (silicone ?) applied to > > the "clean" vacuum wand, presumably to "fix" it instead of replacing it! > > > > > > On Fri, 2 Oct 2009, Mary Tang wrote: > > > > > >> Dear labmembers: > >> > >> > >> > >> > >> > >> It has come to our attention that some labmembers are touching wafers with gloved hands > >> instead of using tweezers or vacuum wands at "clean" stations, such as furnaces and wet > >> benches. Although there may be special circumstances where manual handling is warranted, > >> this violates all known standards of good clean room practices. This is no-gloves-on-wafers > >> policy as was stated firmly in this labmembers' note from 2001: > >> > >> http://snf.stanford.edu/cgi-bin/ezmlm-cgi?mss:147:ebffjnejjanhflmfgmjp > >> > >> > >> > >> > >> > >> In this lab, vinyl gloves from the bag are "clean" for using wafer handling tools. The gloves > >> we use are certified Class 1 so don't shed particles and are metal ion-free. So, these are good > >> for handling cassettes, handles, wands, tweezers and other wafer handling tools for which > >> we want to avoid cross-contact with other surfaces. Certainly, as anyone can see from the > >> acid-etched buttons at any wet station, we are not as diligent as we should be with our glove > >> hygiene. So, we encourage frequent glove changes in order to ensure that contaminants > >> don't travel from one surface to another. > >> > >> > >> > >> > >> > >> However, vinyl is polyvinyl chloride, or PVC, which is a hard, brittle material. To make it > >> flexible, plasticizers are added -- up to 35-45% by weight in gloves. Plasticizers are low molecular > >> weight (<400), oily materials. DEHP or DOP is the most common in cleanroom vinyl gloves. > >> It is widely recognized in the cleanroom industry that vinyl gloves will leave a trace plasticizer > >> residue on contact with surfaces (see page 270 of Cleanroom Technology: Fundamentals of > >> Design, Testing, and Operation, available on Google books at > >> http://books.google.com/books?id=-ufEtmr1sBgC&printsec=frontcover&source=gbs_navlinks_s#v=onepage&q=&f=false ) > >> > >> > >> > >> > >> > >> Here is a simple demonstration. A vinyl glove was pressed firmly down on 1/2 of a clean > >> silicon wafer and then removed. Then, contact angle measurements were made. On the > >> bare silicon half, contact angles were too low to measure (<5 degrees), indicating very > >> good wetting with the native surface oxide. On the other half, the average contact angle > >> was 19 degrees (stdev=2 degrees), indicating a significant change in the surface energy. > >> The wafer was then run three times through lampoly, using recipe #1 (60" main etch, silicon > >> to resist selectivity ~ 3.6.) As can be seen in this photo, glove residue masks the etch. > >> (The small, aligned dots show where the contact angle was measured.) > >> > >> > >> > >> > >> > >> Granted, this was a simple and crude demonstration. For a more rigorous approach, see > >> the attached paper describing effects of DOP plasticizer on electronic device performance. > >> > >> > >> > >> > >> > >> Please understand that there is generally a good technical rationale behind our lab policies > >> -- similar no-gloves-touching-wafers policies for CMOS-clean stations exist at the Berkeley, > >> MIT, and Cornell cleanrooms. We trust that these should be convincing reasons as to > >> why we should not use gloved hands to touch wafers. > >> > >> > >> > >> > >> > >> > >> > >> Your SNF staff > >> > >> > > > > > From shott at stanford.edu Sun Oct 4 09:27:34 2009 From: shott at stanford.edu (John Shott) Date: Sun, 04 Oct 2009 09:27:34 -0700 Subject: Why vinyl gloves are bad for your wafers In-Reply-To: References: Message-ID: <4AC8CCF6.6080901@stanford.edu> Eric: Yesterday afternoon I also looked at all of the wands and am aware of the white adhesive. I have also, however, opened a brand new tip ... I had to cut a seal on the plastic box to open it ... and can confirm that this is the way that the vacuum wands are coming from H-Square. This is definitely not a "home improvement" on our part. I have also already asked Mary to contact H-Square to find out what that stuff is. My guess, however, is that it is a high-temperature ceramic adhesive that is similar to the stuff shown here: > http://www.cotronics.com/vo/cotr/ca_fastset.htm Certainly this stuff looks a lot like the off-white stuff that H-Square is using and it is expressly advertised as being good for sealing metals to glass and capable of withstanding very high temperatures. My guess is that it IS a hand-operation used to assemble these things at H-Square .... someone probably paints a ring of this stuff near the end of the steel tube (so they don't plug the vacuum) and then stuffs the tube into the end effector. It also appears as if at some point they somehow shape any of the adhesive that has squished out so that it does not protrude beyond the broad, flat surfaces of the end effector. We will investigate this stuff further to be able to more definitively say what it is ... but I'm guessing that it is not the problem. > Something transferred to the wafer cold, and then > redistributed during the process, from the wand contact point. Of course, even if this adhesive gets a clean bill of health, there is nothing that would prevent other sources of contamination from getting on the tips. Heck, I suppose that grabbing the tip of a vacuum wand with a "clean" vinyl glove could even transfer some plasticiser to a wafer that would likely result in non-uniformity .... or certainly there could be other ways of transferring contaminant to the end effector of the wand and then onto the wafers. In any event, we'll try to learn more about the details of the off-white adhesive that H-Square is using and report back. Thanks, John From eap at gloworm.Stanford.EDU Sun Oct 4 14:08:09 2009 From: eap at gloworm.Stanford.EDU (Eric Perozziello) Date: Sun, 4 Oct 2009 14:08:09 -0700 (PDT) Subject: Why vinyl gloves are bad for your wafers In-Reply-To: <4AC8CCF6.6080901@stanford.edu> Message-ID: On Sun, 4 Oct 2009, John Shott wrote: > > Something transferred to the wafer cold, and then > > redistributed during the process, from the wand contact point. > > Of course, even if this adhesive gets a clean bill of health, there is > nothing that would prevent other sources of contamination from getting > on the tips. Heck, I suppose that grabbing the tip of a vacuum wand > with a "clean" vinyl glove could even transfer some plasticiser to a > wafer that would likely result in non-uniformity .... or certainly there > could be other ways of transferring contaminant to the end effector of > the wand and then onto the wafers. Like paint chips from where the wand is normally stored.... -------------- next part -------------- A non-text attachment was scrubbed... Name: IMG00128.JPG Type: application/octet-stream Size: 55742 bytes Desc: URL: