Fwd: Re: request to use PDMS in P5000/gryphon

Ed Myers edmyers at stanford.edu
Thu Jan 5 10:16:36 PST 2006


A number of good points have been made during the discussions.  Alissa is 
correct in her recognition of increased requests for PDMS.  It would 
benefit our community if SpecMat could categorized PDMS in to one of our 
contamination classifications.  The question being if it needs to be gold 
contaminated or we can place it in either Semiclean or SemicleanB, since I 
don't think Clean is appropriate.  We will need to understand the 
differences between the many different variations of PDMS.  I'll work with 
Mary and see what we can find with regards to trace elements and the 
chemical and thermal stability of the PDMS.

The more pressing mater is to respond to Alissa's request.  From trace 
element contamination, Mary's memory seems to indicate Alissa's requests 
are reasonable.  I am very concerned about whether the process is 
manufacturable.  Various concerns have been expressed over a number of the 
process steps.  These concerns include the ability to expose 0.7um features 
due to surface undulations in the PDMS, adhesion of the Al to the PDMS and 
the potential for the Al to peel at numerous locations including the 
Gryphon deposition chamber, P5000 and resist strip and the influence of the 
PDMS on the etch process in the P5000.

As a proposal why don't we let Alissa get started on her PDMS casting and 
lithographic process development.  I would recommend starting with Al 
deposition from the Innotec.  The Innotec should be the most compatible 
deposition tool, since the wafer heating will be the lowest.  If she is 
able to get satisfactory patterns, we will be at a decision node where we 
either let the Innotec film in to the P5000 or we allow the PDMS in to the 

Let's get another iteration of comments and try to resolve the request this 


>Hello SpecMat'ers --
>I don't know how hot the gryphon can get, but do think that  outgassing is 
>likely lot less of a problem for PDMS than for photoresist or polyimide 
>tape, if temperatures remain fairly low -- and if the PDMS has been 
>sufficiently cured.  Again, I'm not sure of the purity, but I think the 
>data exists somewhere...
>Alissa M. Fitzgerald wrote:
>>Hi Mary, Jim,
>>Thanks for the info.  Based on your information, and Jim's most recent
>>email, does this mean it is approved for use in the P5000?  (Also, as a side
>>note, based on the number of inquiries I get regarding PDMS, SpecMat may
>>want to consider and publicize a general policy with regards to this
>>material.  I think it's an important material that is gaining popularity in
>>MEMS, esp. with regards to medical and biotech applications.)
>>The process is aggressive and experimental.  Honestly, I am not sure this it
>>is going to work, but the customer is interested in trying it out. We may
>>need to start with different PDMS thickness, bigger CD's, etc.
>>We will need permission to put PDMS in the gryphon, too.  We need to use
>>>-----Original Message-----
>>>From: Mary Tang [mailto:mtang at stanford.edu] Sent: Wednesday, January 04, 
>>>2006 7:47 AM
>>>To: Jim McVittie
>>>Cc: Alissa M. Fitzgerald; SpecMat at snf.stanford.edu
>>>Subject: Re: request to use PDMS in P5000
>>>Hi Alissa, Jim --
>>>I think I have this information somewhere, from a previous request.  As 
>>>I vageuly recall, PDMS (Dow Corning Sylgard 182 or 184) is based on a 
>>>Pt-catalyzed reaction, although very little Pt is actually present.
>>>Other metals and impurities, such as sulfur, will prevent 
>>>polymerization.  So, other than the Pt, PDMS is actually pretty clean -- 
>>>although perhaps not by electronics-grade standards, it's cleaner than 
>>>your ordinary plastics.  I'll see if I can find the info.  I think that 
>>>Claudia Richter provided it, so I'll also check with her.
>>>Just on a side note, I'm personally less concerned about the potential 
>>>contamination than the process flow itself (Alissa, perhaps you've got 
>>>experience or references on this already.) 500 microns of PDMS is pretty 
>>>thick...  It's got a high thermal expansion coefficient, so I'm not 
>>>entirely sure that you could put 0.5 microns of Al on it without having 
>>>it peel off due to stress differences, even with an adhesion layer 
>>>(although having thin lines might help).   I think Claudia or Neville 
>>>Mehenti may have experience in depositing metals on PDMS in our lab 
>>>(although I'm pretty sure they would have used metalica or
>>>innotec.)  By the way, does your request entail using gryphon for Al 
>>>Also, PDMS is a darn good insulator -- I think the Al etch rates and 
>>>profiles may be very different than they would be on silicon due to 
>>>differences in plasma behavior (at least, I understand that P5000 
>>>etching of films on quartz is very different from etching on silicon.) I 
>>>would suggest that if you have problems, a thinner PDMS layer (tens of 
>>>microns -- you may have to dilute and spin coat) might help.
>>>Constrained PDMS (by adhesion at the Si/PDMS interface) won't expand as 
>>>much and electronic effects on plasma *might* be reduced.
>>>Again, I'll if I still have the purity info, and if not, I'll drop a 
>>>note to Claudia.  I think that Dow provided this info to Claudia (or 
>>>whomever it was who gave it to me) so you might try asking them.
>>>Jim McVittie wrote:
>>>>Hi Alissa,
>>>>During the overetch, the Al etch chamber will be contaminated by the 
>>>>decomposition products of the PDMS. So the important
>>>question what is
>>>>in PDMS and is it a problem to other users of the chamber.
>>>My concern
>>>>is what metals at in PDMS and at what level. Can you find a purity 
>>>>analysis for PDMS?
>>>>   Thanks, Jim
>>>>"Alissa M. Fitzgerald" wrote:
>>>>>  Part 1.1    Type: Plain Text (text/plain)
>>>>>          Encoding: 7bit
>>>Mary X. Tang, Ph.D.
>>>Stanford Nanofabrication Facility
>>>CIS Room 136, Mail Code 4070
>>>Stanford, CA  94305
>>>mtang at stanford.edu
>Mary X. Tang, Ph.D.
>Stanford Nanofabrication Facility
>CIS Room 136, Mail Code 4070
>Stanford, CA  94305
>mtang at stanford.edu

More information about the specmat mailing list