Crookes radiometer spin direction

Settled in 1879 by Osborne Reynolds (thermal transpiration, 1879); refined quantitatively by Albert Einstein (1924).

William Crookes invented his radiometer ('light mill') in 1873 while investigating chemical weighing errors caused by radiant heat, and was initially convinced the spinning vanes demonstrated direct mechanical pressure from light. However the device requires only a partial vacuum and spins the wrong way for that explanation, since the dark sides move away from the light rather than being pushed toward it as true radiation pressure would predict. Osborne Reynolds's independent 1879 work on thermal transpiration - gas flow driven by a temperature gradient along a surface - supplied the real mechanism: the vanes' warmer edges impart a net tangential push to gas molecules that drives rotation, with Albert Einstein providing a more complete quantitative theory of the radiometer force in a 1924 paper.

The device that looked like a beautiful desktop proof of light's physical push turned out instead to be a demonstration of a much subtler effect - the same warm-edge gas physics used today in tiny vacuum-pump Knudsen pumps.

Explanations that circulated

• Crookes himself initially believed the vanes were turned by direct radiation pressure from light striking the dark (absorbing) sides harder than the light (reflecting) sides; actual radiation pressure at the vacuum levels used is far too weak to explain the observed torque, and the true mechanism is thermal, not photonic.
• James Clerk Maxwell initially argued that thermal effects should cancel exactly at equilibrium and could not produce net rotation; this was corrected once thermal transpiration/edge effects (gas slipping preferentially at the warmer vane edges) were properly accounted for by Reynolds and later refined quantitatively by Einstein in 1924.

References

• On certain dimensional properties of matter in the gaseous state, Osborne Reynolds, Philosophical Transactions of the Royal Society of London, 1879