If the damage is actually being caused by the cleaving process, I'm not
sure.  However, the one time I've seen this done, the damage was caused
by exposure to the SEM imaging electron beam melting and re-flowing the
polymer.  It's not perfect, but the trick we used at the time was the
following:

* Pattern large array of test structures with a fixed spacing, and
cleave sample (PMMA on <100> Si)
* Use dummy section of sample cross section to setup SEM (focus, stig,
etc.).  Use as low an acceleration voltage as possible.
* Carefully align SEM stage (rotation, tilt) such that a single stage
axis control (X axis) moves along the cleave as precisely as possible,
without dropping out of focus.
* Slide stage along X axis until first test pattern comes into view
(and melts).  Finalize image settings.
* Blank beam, slide stage along X axis by pattern spacing.
* Unblank beam and capture an image immediately.
* Curse.  Repeat.

Our SEM had computerized stage movement, and a mode where starting an
image capture would automatically unblank the beam, but this could be
done by hand as well if your system doesn't have that.

The idea is to capture an image of the structure as early in the
melting process as possible.  It's not perfect, but was sufficient for
our purposes at the time.


Another thought: don't know what your structure is, but some shapes
might be conducive to putting down thick metal, stripping the PMMA, and
SEM imaging the inverted pattern.  (Something like a lift-off or T-gate
process.)

Good luck!

-Mike

On Jul 27, 2006, at 1:42, Jeff Kettle wrote:

> Hi all,
>
>   I have been patterning structures into PMMA using an Electron beam
> lithography system. I was wondering if anyone had any advice on
> capturing an image of the cross section for the patterned structure
> (post-development) using an SEM.
>
>   My attempts so far have failed. Cleaving sidewards across a pattern
> seems to destroy the pattern on the cleaved sample. I have also tried
> sputter coating the to ensure little damage to the resist when viewing
> under SEM. If anyone has any advice or know any "tricks" for captuign
> cross-sectional images, I would be grateful! Cheers, Jeff Kettle