From ahazeghi at stanford.edu Wed Mar 5 16:28:14 2008 From: ahazeghi at stanford.edu (Arash Hazeghi) Date: Wed, 05 Mar 2008 16:28:14 -0800 Subject: Si etch recipe Message-ID: <20080305162814.io8r80be88sk4k8k@webmail.stanford.edu> Hi, I want to etch 1-1.2um trenches in <111> Si (UV treated 1.6um 3612 on top), does anyone know what the best recipe and time is? Thanks, Arash From kimsangb at stanford.edu Wed Mar 5 19:27:55 2008 From: kimsangb at stanford.edu (SangBum Kim) Date: Wed, 5 Mar 2008 19:27:55 -0800 Subject: Si etch recipe In-Reply-To: <20080305162814.io8r80be88sk4k8k@webmail.stanford.edu> Message-ID: <002a01c87f3a$11949500$a9b50c80@sangbumhome> The recipe that I use for <111> Si is standard poly Si etch recipe in the drytek2 book. SF6:117/F22:51/150mTorr/400W. >From my experiences few months back, 6min etch gave me 1.5um deep trenches. I used these trenches as an alignment marks for Hitachi ebeam, and it worked fine. SangBum -----Original Message----- From: Arash Hazeghi [mailto:ahazeghi at stanford.edu] Sent: Wednesday, March 05, 2008 4:28 PM To: drytek2 at snf.stanford.edu Subject: Si etch recipe Hi, I want to etch 1-1.2um trenches in <111> Si (UV treated 1.6um 3612 on top), does anyone know what the best recipe and time is? Thanks, Arash From nharjee at stanford.edu Thu Mar 13 09:04:15 2008 From: nharjee at stanford.edu (Nahid Harjee) Date: Thu, 13 Mar 2008 09:04:15 -0700 Subject: Reservation released today 9:00-11:00 am Message-ID: <2feeb6700803130904u10bd357doa52ebe43532118e8@mail.gmail.com> Despite a valiant effort, my wafers are not ready. Sorry for the late notice. nh -- Nahid Harjee Ph.D. Candidate Electrical Engineering Stanford University 408-761-8651 -------------- next part -------------- An HTML attachment was scrubbed... URL: From ahazeghi at stanford.edu Fri Mar 28 01:45:40 2008 From: ahazeghi at stanford.edu (Arash Hazeghi) Date: Fri, 28 Mar 2008 01:45:40 -0700 Subject: drytek2 inconsistent results Message-ID: Hello drytek users, I wonder if anyone else has had problems with poly etch recipe recently. Several weeks ago I used this recipe (SF6:117sccm F22:51sccm) 6:30 min to etch ~1.3um trenches in Si substrate (1.7um 3612 UV treated resist) power was 425/25W, plasma color was OK. however I later had problem picking up the marks with eBeam and RAITH. When I compared the marks with other samples that had gone through the same recipe (only power slightly different @ 400/0W) a week after it is clear that the earlier trenches were way shallower than target depth. I have attached two pictures, first is the sallow marks, second is the deep marks, resist on top is the same for both. Any ideas? Thanks, Arash ?? ------------------------------------------------------------------------ -- Arash Hazeghi PhD Candidate Stanford Center for Integrated Systems, 420 Via Palou Mall, CIS-X 300 tel: 650-725-0418 mobile: 650-353-1866 http://www.stanford.edu/~ahazeghi -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: sample1-500Aox-5PMMA-450K-200nm-badetch.jpg Type: image/jpeg Size: 963360 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: sample2-220Aox-5PMMA-950K-300nm-good etch.jpg Type: image/jpeg Size: 958197 bytes Desc: not available URL: From shott at stanford.edu Fri Mar 28 07:52:44 2008 From: shott at stanford.edu (John Shott) Date: Fri, 28 Mar 2008 07:52:44 -0700 Subject: drytek2 inconsistent results In-Reply-To: References: Message-ID: <47ED063C.4010009@stanford.edu> Arash: What is the width of these marks? My first thought is that this may be a resist/lithography problem more than an etch problem. Why? The apparent width of the lines is much smaller on your "problem" sample. This makes me wonder whether the opening was fully clear of resist or whether there was a very thin residual layer of resist that prevented the onset of etching. 10 nm of residual resist is hard to see, but would delay the start of etching until it burned through. Was the actual etch preceeded by an O2 descum step to try to remove any residual resist. While I may be coming out of left field on this one, this sort of "smells" like a resist clearing problem more than it does an etch problem. Alternative views from the drytek2 community? John From ahazeghi at stanford.edu Fri Mar 28 12:09:02 2008 From: ahazeghi at stanford.edu (Arash Hazeghi) Date: Fri, 28 Mar 2008 12:09:02 -0700 Subject: drytek2 inconsistent results In-Reply-To: <47ED063C.4010009@stanford.edu> References: <47ED063C.4010009@stanford.edu> Message-ID: <000e01c89107$2f16ef90$8d44ceb0$@edu> John, The marks are 2um wide, the fact that they look wider on the "good" sample is due to 300nm PMMA on top which creates this effect when spun on topographical features, I checked my samples post development and after etch and the marks looked OK. Another reason I have to believe it was the etch is that etching for >6min usually hardens resist and I have to dip it in H2O2/H2SO4 twice (each time for 20min) to get rid of the resist, for the "bad" samples all of the resist came out the first time. And I did not use any descum process. Thanks, Arash ---------------------------------------------------------------------------- - Arash Hazeghi PhD Candidate, Stanford Center for Integrated Systems 420 Via Palou Mall, CIS-X 300 Stanford, CA 94305 phone: +1-(650)-725-0418 http://www.stanford.edu/~ahazeghi/ -----Original Message----- From: John Shott [mailto:shott at stanford.edu] Sent: Friday, March 28, 2008 7:53 AM To: Arash Hazeghi Cc: drytek2 at snf.stanford.edu; Nancy Latta Subject: Re: drytek2 inconsistent results Arash: What is the width of these marks? My first thought is that this may be a resist/lithography problem more than an etch problem. Why? The apparent width of the lines is much smaller on your "problem" sample. This makes me wonder whether the opening was fully clear of resist or whether there was a very thin residual layer of resist that prevented the onset of etching. 10 nm of residual resist is hard to see, but would delay the start of etching until it burned through. Was the actual etch preceeded by an O2 descum step to try to remove any residual resist. While I may be coming out of left field on this one, this sort of "smells" like a resist clearing problem more than it does an etch problem. Alternative views from the drytek2 community? John From mcvittie at snf.stanford.edu Fri Mar 28 13:51:23 2008 From: mcvittie at snf.stanford.edu (Jim McVittie) Date: Fri, 28 Mar 2008 13:51:23 -0700 Subject: drytek2 inconsistent results In-Reply-To: <000e01c89107$2f16ef90$8d44ceb0$@edu> References: <47ED063C.4010009@stanford.edu> <000e01c89107$2f16ef90$8d44ceb0$@edu> Message-ID: <47ED5A4B.7000705@snf.stanford.edu> John and Arash, While the lack of the descum step could explain the problem, a wrong flow ratio could also explain the low etch rate. The Si etch rate is very sensitive to the SF6/F22 flow ratio. Back in 1998 when we characterized this process, we got the best trade off between etch rate and etch profile at a ratio of 3:1. At a ratio of about 1:1 the etch rate goes to zero. At ratios below 1:1, the process goes into the deposition region (negative etch rate). If no one else is seeing the problem, I would expect either there was an excess of resist scum or one of the flows was set wrong. Jim Arash Hazeghi wrote: > John, > The marks are 2um wide, the fact that they look wider on the "good" sample > is due to 300nm PMMA on top which creates this effect when spun on > topographical features, I checked my samples post development and after etch > and the marks looked OK. Another reason I have to believe it was the etch is > that etching for >6min usually hardens resist and I have to dip it in > H2O2/H2SO4 twice (each time for 20min) to get rid of the resist, for the > "bad" samples all of the resist came out the first time. And I did not use > any descum process. > > > Thanks, > Arash > > > ---------------------------------------------------------------------------- > - > Arash Hazeghi > > PhD Candidate, > Stanford Center for Integrated Systems > 420 Via Palou Mall, CIS-X 300 > Stanford, CA 94305 > phone: +1-(650)-725-0418 > > > http://www.stanford.edu/~ahazeghi/ > > > > -----Original Message----- > From: John Shott [mailto:shott at stanford.edu] > Sent: Friday, March 28, 2008 7:53 AM > To: Arash Hazeghi > Cc: drytek2 at snf.stanford.edu; Nancy Latta > Subject: Re: drytek2 inconsistent results > > Arash: > > What is the width of these marks? > > My first thought is that this may be a resist/lithography problem more > than an etch problem. Why? > The apparent width of the lines is much smaller on your "problem" > sample. This makes me wonder whether the opening was fully clear of > resist or whether there was a very thin residual layer of resist that > prevented the onset of etching. 10 nm of residual resist is hard to > see, but would delay the start of etching until it burned through. > > Was the actual etch preceeded by an O2 descum step to try to remove any > residual resist. > > While I may be coming out of left field on this one, this sort of > "smells" like a resist clearing problem more than it does an etch problem. > > Alternative views from the drytek2 community? > > John > > > >