Advisor: AD Rollett, MSE
|
|
 |
We propose to perform serial sectioning, with
electron back scatter diffraction scanning (EBSD)
on a film of tin with the ultimate objective of
predicting where hillocks and whiskers will form. The
local objective is to identify grains whose boundaries
have normals all pointed down into the film. The
(pure Sn) films of interest have strongly columnar
microstructures so at the surface there will be
the occasional small island grain that will have
these characteristics. The underlying assumed
mechanism driving hillock formation is diffusion
creep in a film under biaxial compressive stress,
such that any grain that can add material on its
boundaries and be pushed out/up of the film can
lead to hillock formation. The diagram above
is borrowed from a recent presentation by Bill
Boettinger (NIST). The objective for the
EBSD scanning would be to further identify (island)
grains whose boundaries are all high angle boundaries
relative to the average misorientation in the system. Here
the idea is that high angle boundaries are expected
to be high diffusivity paths and/or boundaries
that can easily migrate into the surrounding grains
(which is associated with hillock formation). Rollett & de
Graef at CMU-MSE have recently installed a Robomet
which is a robotic metallography system (supplied
by UES). We will use the Robomet to perform
the serial sectioning and use image registration
tools developed at CMU to align the sections and
extract boundary normals.
|
