Today I built a simple fixture to give me a consistent process.  It
just consisted of a piece of steel plate with a 1.5" (51mm) standing rim
into which another steel block fit.  This was to contain the foam while
it was being compressed.  The foam I used was 1" (25.4 mm) thick blue
XPS.  The top block had stops so that it would rest 1/2" from the lower
plate.  With the foam inserted and squeezed, it could not spread as it
was squeezed to half it's original thickness.   The metal parts were
heated in hot water to 200F, the foam inserted, and pressed to half it's
thickness, then cooled in hot water.   The change in properties was
impressive.   The surface will take a LOT of pressure and rebound
without permanent deformation.  It's a very different foam compared to
the original which would permanently deform without a lot of pressure.  
A sharp object will puncture the tough surface, however with a layer of
glass over it, there will not be sharp pressure points, loads will be
distributed out.  For example, I can apply a lot of thumb pressure and
it will recover, but an unsharpened pencil will penetrate fairly
easily.  The foam has far more bending resistance, the tough outer skin
preventing it from snapping until a fairly extreme deformation,   I
pulled it around a progressively tighter radius, and failure occurred at
what I would estimate to be an 9" radius.   That's pretty darn
impressive compared to original.

    When I removed the foam from the jig, it was in a pillow shape.  
1/2" at the edges, and thicker toward the center.  Size was 3" by 6".  
The sides were heated along with the top and bottom surfaces, so they
had formed the tough skin, and were restraining the foam from expansion
to it's original thickness.  I cut away the sides, and the foam
rebounded to approximately 3/4" thickness uniformly.  This was the
thickness of the test block. The foam was at about 80F when inserted,
and the only compression appears to have been the outer surface, which
is what I wanted in this case.   Clearly if I did this on full 4:8
sheets, it would be desirable to trim a small amount from the outer edge
to achieve a uniform thickness.

    The next test will involve compressing the foam more to attempt to
achieve a 50% thickness.   I'll heat the foam and the jig at the same
time to the same temperature, and use a 1/4" thickness shim, resulting
in the foam being compressed to 1/4 of it's original thickness initially.

H.W.