I had an interesting idea for a perpetual motion machine and wanted to see if anyone could shoot the idea down.
http://en.wikipedia.org/wiki/Radiation_pressure
If reflected light exerts a pressure on an object, imagine two mirrors with light travelling between them (reflected light exerts double the pressure).
Imagine photons being reflected between them. Would each reflection continue to add more outward force? I have a feeling there's a quantum catch here, but I've never heard of light pressure varying depending on experiment set up.
Anyway, this acceleration could continue indefinitely, assuming the path of the photons was aligned correctly. The light will always travel faster than the mass, so no matter the distance, light will continue to cycle between them.
To make this a more realistic setup, just to show that this could provide a real gain (again assuming no quantum catches) imagine two mirrors with a rotating arm between them. If both sides of this arm had mirrors that were held parallel to the two surrounding and stationary mirrors, then you could reflect light on one side of the arm in one direction and one the other side of the mirror for the other side of the arm and generate a rotational force on the arm. Realistically you'd want to use large enough mirrors to get at least thousands of reflections, but assuming you had a source of photons that was more efficient than this (not very difficult), you should have a net gain of energy.
Ok, so does anyone know what the catch is, or is there one? I assume this effect, multiplying a reflective force for light could be verified rather easily.
Now something else to consider is that if there is a catch, it's interesting to consider what such a mechanism would imply about light.


