Fwd: XR1541 bonding

Ed Myers edmyers at stanford.edu
Mon Aug 15 14:59:19 PDT 2011


I am concerned with the statement in the Dow document which says 
Resists are comprised of hydrogen silsesquioxane(HSQ) resin.  This 
makes it sound like an organic film, which may not be stable in the 
diffusion clean.  I do not have any experience with the material, so 
I can not make a sound recommendation.


At 02:34 PM 8/15/2011, Mary Tang wrote:
>Hi all --
>Here's a request from J Provine, relating to a process that Jae 
>Hyung Lee has apparently been running in the lab.
>The questions are:
>1.  Is electronics grade HSQ, cured at 400 C, sufficiently clean to 
>be processed in wbdiff?
>2.  Is it sufficiently clean to be processed in tylans 1/2?
>We've approved electronics grade SOG for processing in back-end/low 
>temperature tools through SpecMat.  I don't believe we've officially 
>approved it for clean/front-end tools.  Should we?
>My humble opinion:  Wbdiff is OK -- it's as clean as the chemicals 
>used and chemicals will get changed @ 4 hours.  Tylans 1/2 are OK if 
>the process has been vetted in tylan3/4.  My concern is wafers 
>debonding and flaking stuff everywhere.  However, I suspect that Jae 
>has already been doing this and tested the process, if he's got a 
>special slow recipe.
>-------- Original Message --------
>Subject: XR1541 bonding
>Date: Mon, 15 Aug 2011 14:16:57 -0700
>From: J Provine <mailto:jprovine at stanford.edu><jprovine at stanford.edu>
>To: Ed Myers <mailto:edmyers at stanford.edu><edmyers at stanford.edu>, 
>Mary Tang <mailto:mtang at stanford.edu><mtang at stanford.edu>, maurice 
>stevens <mailto:maurice at stanford.edu><maurice at stanford.edu>
>hi mary, maurice, and ed,
>i've had a chance to talk with mary and ed about using the HSQ based 
>ebeam resist XR1541 (a type of spin on glass) to do a moderately 
>high temperature bonding of two oxidized wafers.  i'll talk with 
>maurice when he returns on wednesday.  this email is to get everyone 
>the same page and document the proposal for what i would do.
>i would start with two clean oxidized SOI, that have been through 
>clean CMP and then decontamination.
>1. spin XR1541 (about 150nm) at headway.
>2. bake on hot plate at 90C (2 min) -> 150C (2 min) -> 400C (2 
>hours) with weight applied by the litho weights covered in clean new 
>foil and the hotplate covered in new foil.
>3. clean bonded wafers in wbdiff
>4. anneal at 900C for 2-3 hours in tylan1/2 using the recipe slow900a
>for step 3 i could go to another, less clean bench if needed, such 
>as wbnonmetal.
>for step 4 i could go to tylan3/4 if needed for my process, but i 
>would like to stay clean and only go contaminated if necessary.
>is this ok?  what additional information can i provide to help in 
>the decision.
>attached is a document about XR1541 and its cleanliness.

More information about the specmat mailing list