Christopher Tidy writes:

According to some articles I've found online, certain metallurgical
microscopes use a plain glass sheet which works by total internal
reflection rather than a semi-silvered mirror. If that's the case with
my microscope, it would make the glass much easier to find. I might even
be able to use a circular glass cover slip.

Or even try cutting some 0.015 Mylar sheet (overhead transparencies) with
scissors. Not really optically flat, but maybe worth experimenting with.