[Fwd: Re: Lampoly recipe for silicon-rich nitride]
Mary Tang
mtang at stanford.edu
Thu Sep 25 11:43:11 PDT 2008
Hi all --
This material was approved for etch on the lampoly ("trace" Er.) OK for
amtetcher as well?
Mary
-------- Original Message --------
Subject: Re: Lampoly recipe for silicon-rich nitride
Date: Thu, 25 Sep 2008 10:06:08 -0700
From: Aaron Hryciw <ahryciw at gmail.com>
To: Mary Tang <mtang at stanford.edu>
References:
<4d36fb940809241236h456c4846r6385a3de53e53196 at mail.gmail.com>
<Pine.LNX.4.44.0809242150030.25336-100000 at cis.Stanford.EDU>
Hi Mary –
So, Jim is suggesting that AMT would be best for this (NF3 via etch).
My SRN etch-test wafers will be "clean," but since the actual device
film I'll want to etch will contain a little less than 1% erbium (which
is a non-standard metal), does this mean that AMT is completely out of
the picture? Is there some sort of decontamination procedure I could do
after my etch, which would allow me to use this tool for the Er-doped films?
Cheers!
– Aaron
On Wed, Sep 24, 2008 at 10:00 PM, Jim McVittie
<mcvittie at cis.stanford.edu <mailto:mcvittie at cis.stanford.edu>> wrote:
Aaron,
Ok, since you do not care about how much Si you etch, the Lam is not
automatically off the table. But I still would not go with the Lam since
really do not have a nitride etch process in the Lam. If I was you, I
would go with the P5000 since it is quick and uniform. The std oxide
etch
process process will work fine or you can use a nitride process. If you
already know the AMT but not the P5000, then go with the AMT. THe
via etch
shouls work fine. I would timed etch since you have non-critical
needs. If
really want to use endpoint, I can probably it up for you,
Jim
On Wed, 24 Sep 2008, Aaron Hryciw wrote:
> Hi Mary, Jim –
>
> The purpose of this etch will be to define silicon-rich nitride (SRN)
> microdisks, the pillars of which will afterwards be defined by a
TMAH etch.
> As such, my tolerance for etching into the underlying Si is not too
> critical, just so long as I punch through the SRN. To answer your
> questions:
>
> 1. What kind of film are you etching exactly? (Target composition and
> > thickness.)
> >
>
> The final process wafers will be: 1) SRN, ~49% Si, 51% N, and 2)
SRN doped
> with up to 1% Er. In both cases, the film will be ~50 nm thick.
>
> 2. What feature size and what is your masking material (thickness
and type
> > of resist, whether there's a sacrificial mask.)
> >
>
> For the first experiments, the features will be widely-spaced
circles,
> ranging from ~1 to 100 um in diameter. The intended masking
material is
> Shipley 3612, patterned using the ASML (1 um thick? 1.6 um?
with/without UV
> cure? --> will test). If necessary, however, a sacrificial mask
(Cr?) could
> certainly be used.
>
>
> > 3. What are you stopping on and what is your tolerance for
etching into
> > this?
> >
>
> I will be stopping on Si. As mentioned above, so long as I
manage to punch
> through the SRN, the tolerance for etching into the underlying Si
substrate
> is quite large; for instance, I imagine I could etch a micron
into the Si
> without any problem. Presumably, since the etch rate of Si
should be much
> larger than SRN, it should be detectable as a spike in Si-related
byproducts
> in the endpoint detection plots, correct?
>
> 4. What are your uniformity (and selectivity) requirements?
> >
>
> I will have several dies on the wafer, each with the 1 to 100 um
diameter
> (roughly equally on a logarithmic scale) disks on them, and for the
> experiments, I will be looking at single disks, so uniformity
across the
> wafer will also not be too much of an issue. Even 10% or so
should be
> fine. Since there is only a single masking/etching step, the
only material
> being etched is the SRN, so selectivity isn't really a problem,
except
> insofar as I can tell when I've reached the Si. If endpoint
detection is
> not possible, and I have to go with a timed etch (after
characterising the
> process with test wafers), so be it!
>
> 5. I'm presuming that your substrates are 100 mm silicon rounds
of standard
> > thickness?
> >
>
> Correct.
>
> Thanks again to both of you for all your help.
>
> Cheers!
>
> – Aaron
>
>
> On Wed, Sep 24, 2008 at 10:55 AM, Mary Tang <mtang at stanford.edu
<mailto:mtang at stanford.edu>> wrote:
>
> > Hi Aaron --
> >
> > Jim is absolutely right -- my apologies, I think I confused
yours with
> > another project and didn't realize you were stopping on
silicon. I don't
> > know about this machine specifically, but silicon etches about
an order of
> > magnitude faster with SF6 than silicon nitride (stoichiometric).
> >
> > As I neglected to perform due diligence before by asking you a
bunch of
> > questions, I'm going to ask right now (and no doubt, Jim will
have more.)
> >
> > 1. What kind of film are you etching exactly? (Target
composition and
> > thickness.)
> > 2. What feature size and what is your masking material
(thickness and type
> > of resist, whether there's a sacrificial mask.)
> > 3. What are you stopping on and what is your tolerance for
etching into
> > this?
> > 4. What are your uniformity (and selectivity) requirements?
> > 5. I'm presuming that your substrates are 100 mm silicon rounds
of standard
> > thickness?
> >
> > Again, sorry about that --
> >
> > Mary
> >
> >
> > Jim McVittie wrote:
> >
> >> Hi,
> >>
> >> I just sent you a copy of a note I sent Mary. We never got
great Nitride
> >> results in our Lam. The other tools gave better results so we
never push
> >> the Lam to get a good niride process. I would NOT use the Lam
to stop on
> >> Si. It is a Si etcher and not optimized to stop on Si. You
need to use an
> >> oxide etcher which has the polymer deposition which will slow
down the
> >> etch rate when you hit Si. SF6 is not a good chice since it
loves to etch
> >> etch Si. I can help you if you want to use the AMT or P5000.
> >>
> >> Jim
> >>
> >>
> >> On Wed, 24 Sep 2008, Aaron Hryciw wrote:
> >>
> >>
> >>
> >>> Hi Jim –
> >>>
> >>> I need an anisotropic etch recipe for silicon-rich nitride (SRN);
> >>> ultimately, I'll need to etch through a 350-nm-thick SRN
film, stopping
> >>> on
> >>> Si. Mary suggested using the lampoly tool with a SF6-based
chemistry,
> >>> and
> >>> said that you would likely be able to offer some advice as
how best to
> >>> proceed. She said that recipe 99 (a "clean" recipe with SF6
and Cl with
> >>> no
> >>> bias, for general chamber cleaning) might be a good place to
start, and
> >>> she
> >>> also sent me a link to the standard nitride recipe (SF6 and
He) for
> >>> etching
> >>> on the Lam1 tool in Berkeley's Microlab (although this is quite a
> >>> different
> >>> tool compared to SNF's).
> >>>
> >>> I'm going to ask Maurice for some dummy low-stress nitride
(which are
> >>> apparently quite Si-rich) to use as etch-test wafers, and I
was hoping
> >>> you
> >>> could give me a few lampoly "rules of thumb" (pressures, RF
power, bias,
> >>> etc.) for anisotropic SF6-based etching which I could use as
a starting
> >>> point. Many thanks!
> >>>
> >>> Cheers!
> >>>
> >>> – Aaron
> >>>
> >>>
> >>>
> >>>
> >>>
> >>
> >>
> >>
> >
> >
> > --
> > Mary X. Tang, Ph.D.
> > Stanford Nanofabrication Facility
> > CIS Room 136, Mail Code 4070
> > Stanford, CA 94305
> > (650)723-9980
> > mtang at stanford.edu <mailto:mtang at stanford.edu>
> > http://snf.stanford.edu
> >
> >
>
>
>
--
--------------------------------------------------------------
Jim McVittie, Ph.D. Senior Research Scientist
Allen Center for Integrated Systems Electrical Engineering
Stanford University jmcvittie at stanford.edu
<mailto:jmcvittie at stanford.edu>
Rm. 336, 330 Serra Mall Fax: (650) 723-4659
Stanford, CA 94305-4075 Tel: (650) 725-3640
--
Aaron Hryciw
Postdoctoral Scholar
Geballe Laboratory for Advanced Materials
Stanford University
476 Lomita Mall (04-490)
McCullough Building, Rm. 325
Stanford, CA 94305-4045
Tel.: (650) 723-5840
Fax.: (650) 736-1984
--
Mary X. Tang, Ph.D.
Stanford Nanofabrication Facility
CIS Room 136, Mail Code 4070
Stanford, CA 94305
(650)723-9980
mtang at stanford.edu
http://snf.stanford.edu
More information about the specmat
mailing list