I have been attempting to etch through <100> Si wafers using xenon
difluoride.  I have tried to build a vapor-phase  etching system out of
standard components, including stainless steel tubing and fittings,
high-grade o-rings, and a mechanical vacuum pump.  I have succeeded to the
extent of making a couple of small holes through a Si chip, but I have yet
to develop a truly useable procedure.

I have followed the procedures listed in published articles on the subject
and their references, and in general the advice seems good.  However, my
experience has been that the procedure is not robust when etching through
hundreds of micrometers of Si.  Or perhaps another way to put it is, pretty
good lab techniques are required, and I have not achieved the proper set of
such techniques yet.

My biggest problem is that the etch stops after some tens of micrometers of
etching.  I believe that I observed visually (one time!) a reaction between
the etch product, SiF4, and pump oil vapor within my reaction vessel.  This
reaction produces a passivating layer of gunk.  The gunk can be observed on,
and scraped off of, the masking oxide near the bare Si windows that are
supposed to be etching.  The gunk is not volatile, may be water soluble, and
seems to share some characteristics with teflon and silicone--mainly the
characteristic that it coats the surface, stops the etch, and I don't want it.

Installing a liquid nitrogen cold trap seemed to suppress this problem, but
added complications especially when the trap warms suddenly and releases
lots of vapors.  In addition, the suppression may not have been complete;
the problem may still be there, just happening slower.

Any advice as to what might be stopping the etch, how to keep the etch
going, how to avoid mysterious passivation layers or reactions, what kind of
oil to used in the mechanical pump, what kind of trap to use for the pump
oil vapor, etc. would be appreciated.

Dave Read
david.read@nist.gov

Informal communication from:

     David T. Read
     Materials Reliability Division, 853
     National Institute of Standards and Technology
     325 Broadway
     Boulder, Colorado 80303
     Phone:  303 497 3853
     Fax:    303 497 5030