[Fwd: Lampoly with Aluminum]

Jim McVittie mcvittie at cis.Stanford.EDU
Thu Oct 2 22:19:05 PDT 2008


My concern is the case where a user lets the etch go through the resist
and exposes Al. The Al would be attacked by the Br and Cl and end up on
the walls where it is would increase the recombination rates and affect
subsequent Si etch rates and uniformity. THe resist selectivity is only
3:1 at best in the Lam.


On Thu, 2 Oct 2008, Mary Tang wrote:

> Hi all --
> Think of this as another tangible test case.  What do you think?
> Mary
> -------- Original Message --------
> Subject: 	Lampoly with Aluminum
> Date: 	Thu, 2 Oct 2008 17:14:02 -0700
> From: 	Joey Doll <jcdoll at stanford.edu>
> To: 	Mary Tang <mtang at stanford.edu>
> CC: 	Bryan Petzold <petzold at stanford.edu>
> Hi Mary -
> Here's a paragraph re: metal in lampoly:
> In order to release microcantilevers fabricated on SOI wafers, a 
> frontside silicon etch is required. Ideally this etch would be uniform 
> within a few % across the wafer, stop on the buried oxide and have 
> vertical sidewalls without scalloping (as in DRIE) for low noise. In 
> order to minimize the topography on the wafer for lithography steps, it 
> is best to release the cantilever as one of the last steps, after the 
> deposition and patterning of aluminum bondpads. We propose allowing the 
> introduction of wafers with standard metals (Al, Ti, W) into lampoly if 
> the metal is protected by photoresist and not exposed to the plasma, 
> much like the policy on STSEtch. As in the case of STS, the etch chamber 
> is aluminum already, so keeping aluminum at least out of the chamber is 
> somewhat of a moot point, and the wafer is cooled on the backside, 
> keeping temperatures down and the metals at the surface of the wafer so 
> that they can be removed by a subsequent diffusion clean. However, the 
> metal on the wafer would not be exposed to the plasma under normal 
> operation, and again much like STS and drytek2, it should be possible to 
> keep the tool in the clean/semiclean category so that clean wafers 
> coming into it could still go back into the tubes afterwards while 
> semiclean wafers coming in remain semiclean going out.
> Let us know what you think. We have wafers with some aluminum on them 
> right now that we would be interested in etching as soon as possible, 
> but titanium is something we'll be using in about two weeks.
> Thanks,
> Joey

Jim McVittie, Ph.D.    			Senior Research Scientist 
Allen Center for Integrated Systems     Electrical Engineering
Stanford University             	jmcvittie at stanford.edu
Rm. 336, 330 Serra Mall			Fax: (650) 723-4659
Stanford, CA 94305-4075			Tel: (650) 725-3640

More information about the specmat mailing list