Another thing I wonder is this...assuming you have a 6in diameter slit-cylinder around the sample, the circumference is 18.8in or 479mm. You'd need 958 pieces of 0.5mm fiber optics arrayed around it to surround it with a "circle of light". This would probably require pieces of fiber-optics ranging from 2 to 2.5 ft long, and would be pretty expensive (not to mention a pain to assemble). I wonder if the same amount of light focused into fewer pieces of fiber-optics, arrayed in intervals around the slit and pointing to the slit's center, would approximate the effect? Presumably, light exits fiber-optics diffusely, which is why you can see the light exiting them from the sides. Hence, I'd think light exiting one piece of fiber optics should be "seen" by all parts of the sample along the plane.
Another possibility is to use a circline bulb to get light through the slit, but the a lot of light from the circline would be wasted, resulting in exorbitantly long exposures. (if going the fiber-optic route, I'd make a DIY gooseneck illuminator with a bright CFL source and a reflector bouncing most of the light into a fiber-optic bundle).