Michael Deal mdeal at stanford.edu
Mon Jun 30 16:08:32 PDT 2003

I found one reference where they etch tungsten in a Leybold Z-401s using 
SF6 and He and get an anisotropic value of 80%.  The reference is 

At 03:46 PM 6/30/2003, Mary Tang wrote:
>Hi David --
>Just additional few questions that I think the other SpecMat members are bound
>to ask you as well:
>1.  Where is your tungsten being deposited?  (Any processing done outside of
>SNF needs to be carefully considered, because of possibly contamination
>2.  Where do your wafers go after the proposed Lampoly step?  (Any other clean
>3.  Do you have any literature references or experience with SF6 and the
>typical profiles you get with this?  (I would strongly suggest you make an
>appointment to chat with Jim McVittie, our resident etch expert...  I am
>nowhere near an expert, but has understood the prevailing wisdom is that SF6
>gives you lots of undercutting -- the anisotropic profiles attainable on the
>Lampoly are due to the Cl2 chemistry steps.  Jim can tell you what's really
>possible in our lab...)
>For reference, you should check out:
>http://snf.stanford.edu/Materials/NewMatProc.html  --- this lists all the
>questions and issues you should consider when you want to bring something new
>into the lab (some of the questions aren't applicable to your situation, but
>I'm sure you can figure out which ones are...)
>David Lieberman wrote:
> > Hi Mary, John, Mahnaz, and Jim;
> >
> > We'd like to ask permission to use the Lampoly to etch a thin layer of
> > tungsten ( ca. 100nm thick with an array of 130nm diameter holes).  We
> > require an anisotropic etch ( vertical to near vertical sidewalls) and
> > believe this is the only tool at SNF with the capability to resolve this
> > feature.
> >
> > Since the etch chemistry creates a volatile product we believe that it's a
> > safe process to run in the Lam.  The SF6 reacts with the tungsten to form
> > the volatile product WF6 which is then pumped out of the system.  With the
> > extremely small feature sizes, and the thin film, the amount of W actually
> > etched is extremely small.
> >
> > We've talked to a fab that has used the TCP etcher, normally dedicated to
> > CMOS polysilicon, for making refractory metal gates with no problems.  They
> > run a few dummies with SF6 afterward just to be extra safe.
> >
> > Looking at the history of the Lampoly it is clear the tool is underutilized
> > and if the process succeeds we would look to use the tool on a regular
> > basis.
> >
> > Thanks for your consideration,
> >
> > Dave Lieberman
> >
> > David Lieberman PhD.
> > Brion Technologies
> > 408-653-1527
>Mary X. Tang, Ph.D.
>National Nanofabrication Users' Network
>Stanford Nanofabrication Facility
>CIS Room 136, Mail Code 4070
>Stanford, CA  94305
>mtang at snf.stanford.edu

More information about the specmat mailing list