Hello,

Is anyone aware of the conditions under which buffered-HF (BOE) etching
is diffusion limited, or a reference which discusses this?

The particular situation I'm interested in is etching a 5um thick
silicon dioxide layer at the bottom of a KOH-etched pit in silicon, as
shown below. The silicon is 500um thick and base of the pit (top in the
diagram) is ~ 100um x 100um.


membrane nitride
======================================================================
dielectric oxide               |        |
--------------------------------        ------------------------------
                               /          \
                              /            \
silicon                     /              \
                            /                \
                           /                  \
--------------------------                    ------------------------
mask oxide               |                    |
==========================                    ========================
mask nitride


I have measured the etching rate of my 6:1 BOE solution to be 60nm/min
on fully exposed oxide. I suspect the etching rate in this case may be
diffusion limited due to the geometry, because, after 2h20min the oxide
still does not appear to be fully etched. My doubt is due to the fact
that it is hard to measure the the actual thickness in this case. All I
have to go on are SEM images showing undercut, which suggests at least
some etching, and TEM observations where the electron beam is scattered
so much it is impossible to focus the beam at high magnifications.

Regards,

Ken Healy

Dept of Physics and Astronomy
University of Pennsylvania