Request to use lampoly or p5000 to etch semi-clean wafers

Jim McVittie mcvittie at cis.Stanford.EDU
Fri Oct 9 10:07:05 PDT 2009


Back at the beginning of time, we wanted to keep all metals of the Lam and
push requests like this to the P5000. The idea was to reserve the Lam for
frontend processes where being metal free is most important. However, yrs
back when we did TXRF measurements on the Lam, we found that it was no 
better then the P5000 Si chamber and the key was the post etch cleaning. 
So I do not think it make much difference from a metal contamination 
viewpoint whether this work is done in the Lam or P5000. Since this is not 
a high resolution etch, I would recommand he use the P-5000. By the way, 
the Lam intrically has Al comtamination because of its clamp.


On Fri, 9 Oct 2009, Mary Tang wrote:

> Hi all --
> We really need to get back to this question of how we are going to 
> accommodate requests for metal-containing materials in clean dry etch 
> tools.  This request strikes me as being an incremental step in that 
> direction....
> As for as lampoly is concerned (and I suspect would also apply to 
> P5000etch), this seems reasonable.  The EBR margin should be sufficient 
> (a wafer without EBR risks sticking to the lampoly clamp ring.)  And the 
> clamp is made from alumina, so I don't think there's an appreciable 
> contamination risk from whatever trace Al may remain on exposed wafer 
> surfaces.
> I vote that we OK this request -- with the proviso that we update the 
> documentation to describe in detail the rationale and conditions under 
> which semi-clean wafers can be processed in lampoly (and p5000etch.)
> So...  do we have a quorum?
> Mary
> Nahid Harjee wrote:
> > SpecMat,
> >
> > I am designing a process to fabricate single crystal silicon 
> > cantilevers that are 3 um thick. The narrowest cantilevers are 14 um 
> > wide and have 4 "legs" at the clamped end that are 2 um wide. I am 
> > trying to select a tool to etch the 3 um of silicon to define the 
> > cantilever that will produce straight sidewalls. In the past, I have 
> > used stsetch to define wider cantilevers. However, I am concerned that 
> > the scalloping resulting from this tool will make it difficult to etch 
> > the 2 um legs of the new design. Thus, I am writing to propose that I 
> > perform this etch with HBr/Cl2 chemistry in lampoly or p5000. What 
> > makes this process non-standard is that my wafers will be semi-clean 
> > at this point. The step prior to defining the cantilevers is 
> > deposition of 1500 A of Al in gryphon which is then patterned with a 
> > wet etch in Al-11. During the Si etch, the Al will never be exposed to 
> > the plasma (it will be covered by 3 um of resist). However, there is a 
> > chance that there will be trace Al on the exposed Si from the Al-11 
> > bath. At the most recent process clinic, Keith Best raised the point 
> > that there may be exposed Al in the EBR region of my wafers. In order 
> > to minimize this possibility, I can use 5 mm EBR for the Al litho and 
> > 2 mm EBR for the cantilever litho, ensuring there's a buffer of 3 mm 
> > of resist with no Al below it.
> >
> > I look foward to hearing your decision on the proposed process or if 
> > there is a tool that is better suited for this process, I welcome any 
> > suggestions.
> >
> > Thanks,
> >
> > nh
> >
> > -- 
> > Nahid Harjee
> > Ph.D. Candidate
> > Electrical Engineering
> > Stanford University
> > 408-761-8651

James (Jim) P. McVittie, Ph.D.	        Sr. Research Scientist 
Paul G. Allen Building                  Electrical Engineering
Stanford Nanofabrication Facility       jmcvittie at
Stanford University             	Office: (650) 725-3640
Rm. 336X, 330 Serra Mall		Lab: (650) 721-6834
Stanford, CA 94305-4075			Fax: (650) 723-4659 

More information about the specmat mailing list