What is Semi-Clean

Jim McVittie mcvittie at snf.stanford.edu
Wed Aug 18 11:24:00 PDT 2004


Here my initial thoughts. Mike Deal and I need to get together and consolidate
our thoughts on the tool.

Semiclean is a term that Margaret Prisbe came up some years back. For the SCT
system, it means that we will only allow metals, wafers and procedures
compatible with CMOS devices.

In terms of metals, the list use to be pretty short, however with the push
toward hi-k and metal gates the list has gotten longer and more uncertain. The
uncertainty means that we are going be trying out new metals to see how well
they work from a CMOS device standpoint and how well we can keep them from
getting into or cross-contaminating metal depositions where we do not want
them. An important issue with the SCT is that it is a tool where we want to get
reproducible device properties. The initial concern is work function control,
which means interface control. This requirement means that, like the Balzer, we
probably will not allow organics. Thick depositions may also be a problem. The
SCT is the replacement for the Balzer, which was a device research tool. The
main difference the SCT and Balzers is that the target cost is much lower and
we are not restricted by the needs of the Physics group. The restrictions will
now be set by the CMOS device needs.

In terms of wafers, no gold (or sodium like metals) contaminated wafers will be
allowed. Because the importance of interface control, organics on wafers will
be restricted. Other problem metals will be prohibited or  have restrictions in
terms of sputter etching and heating.

In terms of procedures, I am concerned with any procedure that would affect
contamination of deposited films or interfaces. Areas of concern are deposition
thickness, sputter etching and substrate temperature. Thick depositions cause
flaking which increases surface area and cross-contamination. Sputter etching
can introduce contamination material from wafers. Finally, heating of
substrates can cause contamination problem. At this point I am not sure exactly
how these concerns should translate into restrictions.

Regarding nitride and oxide deposition, there is no question that we will allow
reactive metal nitride deposition in the tool. As for oxide deposition, I think
we need some experience with the tool and a device need before addition this


Ed Myers wrote:

> SpecMat Committee,
> We are getting close having a couple of students spearhead the use of the
> SCT sputter.  This is great, since they can take the lead at understanding
> the influence of RF and heat during deposition.  The only conflict I see is
> the materials they want to use is outside of the current definition of
> Semi-Clean metals.
> Metals that are currently being requested include: Ni, Mo, Hf and
> reactively sputtered TiN.  These materials and metals such as Co have been
> successfully introduced into high volume semiconductor manufacturing
> facilities.  There is also an increase in the use conductive metal oxides,
> such as Ru and Ir for contacting high dielectric constant materials.
> If we allow these materials in a Semi-Clean sputter, we will have opened a
> cascading effect for the balance of the Semi-Clean tool set.  I think we
> should take this opportunity to revisit the what metals should comprise the
> Semi-Clean list.
> Ed Myers
> >X-Sieve: CMU Sieve 2.2
> >Delivered-To: emyers at snf.stanford.edu
> >Date: Mon, 16 Aug 2004 08:49:57 -0700 (PDT)
> >From: Abhijit Pethe <pethe at stanford.edu>
> >To: emyers at snf.stanford.edu
> >Subject: Request for training and use on the SCT
> >
> >
> >Hi Ed,
> >
> >My name is Abhijit and I am working with Prof. Saraswat on the properties
> >on the Metal-Germnanium Ohmic Contacts. The cleanliness of the apparatus
> >and the substrate is extermely crucial for studying these
> >interfaces. We would hence want to use an equipment belonging to the
> >semi-clean category as opposed to the gold-contaminated Innotec.
> >
> >We plan to study the following metals
> >1. Ni
> >2. Hf/W (Hf capped with W)
> >3. Ti/TiN
> >
> >It would be great if you could train me on the system once it is available
> >for use. Please let me know if there is any other information I need to
> >supply.
> >
> >Thanks
> >Abhijit
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mcvittie.vcf
Type: text/x-vcard
Size: 422 bytes
Desc: Card for Jim McVittie
URL: <http://snf.stanford.edu/pipermail/specmat/attachments/20040818/0b9933fe/attachment.vcf>

More information about the specmat mailing list