nickel silicide process concern in SNF.
mcvittie at snf.stanford.edu
Wed Nov 3 13:20:17 PST 2004
The following materials (Mat D1) have well known deep levels in Si:
Zn, Cu, Au, Fe, and Mn.
Additional materials (Mat D2) with reporterd deep levels for Si are:
Cd, Ni, Co and S.
It is clear that we want to restrict the D1 materials. As for the D2 materials, it
is not so clear.
Our present cleaning process leaves S on the surface and does not appear to be
cause any problems. Ni and Co are currently used by industry for silicides in
current generation CMOS devices. However, it not so clear that Ni and Co would not
affect older generation type devices, like the detector devices which Sherwood
Parker work on. See note from Baylor Triplett Sept 20, 2004) below regading
possibble Ni problems. As for wet cleaning and its removal of Ni. The limited
data I have looks promising. In particular, for the etcher contamonation test we
did a few years back, we found the Lam etcher leave Ni at a level of 4.1 E11/cm2
on the wafers. After a std pre-diff clean the Ni level was measured at
0.07E11/cm2, which was the TXRF noise level for Ni. Without testing, I am not so
sure our cleaning can get Ni silicide wafer down to this level. There is data in
the literature which would be helpful. The original RCA papers did look at metal
I have one other data point regarding Ni. A few years back I looked at Ni
contamination in the Balzers and Gryphon after a sputter clean. In the Balzers we
found the Ni level to be 19E11/cm2 and in the Gryphon it was 8E11/cm2.
So what do we know.
1. Our present "Semi-clean" etching and sputter deposition tools do introduce Ni
2. Low level ( 4.1 E11/cm2) Ni surface contamination is removed in a wet clean.
3. It is not clear how much Ni contamination will affect our contamination
Michael Deal wrote:
> At 09:19 AM 11/3/2004, you wrote:
> >Great! Thanks! So, are there "official" semiclean B rules now (other
> >than the wet bench concerns?) Are they ready enough to post to
> >SpecMat? (And are there any plans for plasma etchers? I think this is
> >where Jun has concerns...)
> Here's what we decided a month or two ago:
> >>Here is the current list of materials planned for the SCT:
> >> Ni, Co, Pt, Al, Hf, Ta, Mo, W, Ti, Cr, Zr
> >>All targets except for Pt, Zr, and Cr, are already here.
> >>Al, W, and Ti are currently classified as "semi-clean." The others (Ni,
> >>Co, Pt, Hf, Ta, Mo, and Cr, and Zr) are currently being studied and used
> >>by our users and by industrial chip manufacturers as gate materials and
> >>as silicides. So far, unlike Au, Cu, alkalai metals, etc., no harmful
> >>effects have been seen on device performance with these
> >>elements. However, we should still be careful with them in case
> >>something may arise. In many cases, the users most sensitive to their
> >>effects (i.e. those building CMOS devices) are the ones who are now
> >>studying and using them. Jim and I are working with several of the
> >>students who are using these materials and we will keep a close eye on this.
> >>Here's what we propose:
> >>1. Deposition of all metals in the SCT will be limited to 100nm (may be
> >>increased to 200nm). This will limit the problems (peeling,
> >>contamination, etc.) caused by excess deposition of these materials.
> >>2. Al, Ti, and W will be classified as "semi-clean-A" and will be treated
> >>as before in terms of what equipment wafers deposited with these elements
> >>can go into subsequent to their deposition in the SCT. These include LTO
> >>(tylanbpsg), tylanfga, wbsolvent, wbgeneral, wbmetal, wbsilicide as well
> >>as many of the litho and etch equipment.
> >>3. The other elements (Ni, Co, Pt, Hf, Ta, Mo, Cr, and Zr) will be
> >>classified as "semi-clean-B." Wafers deposited with these elements in
> >>the SCT can go into all the equipment listed in the semi-clean group,
> >>EXCEPT for the etching systems. (Most can only be wet-etched anyway.)
> >>Before we do this, however, Jim will look into any issues regarding the
> >>use of wet benches and any possible cross-contamination that might occur
> >>during cleaning.
> >>Finally, many of these materials already have some "special processing
> >>rules" such as how any etching solution must be disposed of. These rules
> >>will be continued.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 422 bytes
Desc: Card for Jim McVittie
More information about the specmat