Crookes Radiometer: Why the “Light Mill” Spins the Wrong Way (Field Guide)

Date: 2026-03-16
Category: explore

The core puzzle

A Crookes radiometer (the classic glass-bulb “light mill”) has black/white vanes in partial vacuum.

When illuminated, the black sides move backward (they retreat from the light source).

At first glance, this looks like light-pressure propulsion—but that intuition predicts the opposite direction for these vanes and is far too weak at this scale.

So what is really happening?

Short answer

The radiometer is primarily a rarefied-gas heat engine, not a photon-pressure motor.

• Light heats the black side more than the bright side.
• A temperature gradient forms near each vane.
• In low-pressure gas, thermal transpiration / thermal creep and edge-region nonequilibrium effects create a net force.
• That force pushes so the black side trails.

No low-pressure gas, no strong spin.

Why radiation pressure is not the main driver

If photon momentum transfer dominated:

1. Reflection on bright sides would typically give larger momentum transfer than absorption on black sides.
2. Better vacuum should improve rotation (less drag).

But experiments and operation show:

• The radiometer works best in a partial vacuum window, not hard vacuum.
• In very high vacuum, ordinary Crookes vanes barely move (except tiny effects measurable with specialized instruments like Nichols radiometers).

That immediately flags a gas-mediated mechanism.

Pressure window intuition (the most useful mental model)

Think in terms of mean free path vs device scale.

• Too high pressure: gas behaves too continuum-like; thermal edge effects are damped and drag dominates.
• Too low pressure (near hard vacuum): too few molecules to transmit useful radiometric force.
• Middle (rarefied) regime: strongest radiometric forcing.

A practical rule: strongest behavior appears around Knudsen-number-relevant conditions (mean free path comparable to vane/gap scale).

Thermal transpiration in one line

For a sufficiently rarefied connection between two regions at different temperatures, equilibrium tends to satisfy approximately:

[ \frac{p}{\sqrt{T}} \approx \text{constant} ]

So hotter and cooler sides can sustain different pressures in a way that drives creep flow and force near vane edges.

This is exactly the family of effects that made Reynolds/Maxwell-era debates so important.

Why edges matter so much

A common mistake is to focus only on broad vane faces.

In Crookes-like operation, edge regions are where asymmetry in molecular momentum exchange survives cancellation most effectively. That is why geometry, gap size, and thickness strongly affect force.

Modern micro-radiometric actuator literature shows the same story:

• force depends nonlinearly on pressure,
• geometry tweaks can significantly change output,
• best performance appears near transitional rarefied regimes.

Historical arc (very compressed)

• 1873: Crookes builds the device while studying effects in partially evacuated setups.
• 1876: Schuster’s tests challenge simple radiation-pressure explanation.
• 1879: Reynolds (thermal transpiration) and Maxwell (temperature-gradient stresses in rarefied gases) establish the key theoretical direction.
• 1901: better-vacuum experiments reinforce that ordinary Crookes spin is not a hard-vacuum light-pressure effect.
• 20th–21st century: kinetic-theory and DSMC-era work connects radiometer physics to MEMS/Knudsen actuators.

Modern relevance: not just a desk toy

The same physics underpins:

1. Knudsen pumps (no moving mechanical parts, thermally driven gas transport)
2. Radiometric micro-actuators
3. Low-pressure gas sensing concepts
4. Rarefied-flow design intuition for microsystems and some space/near-vacuum contexts

So the radiometer is a historical gateway into real engineering, not just Victorian novelty.

Fast misconception checklist

• “It spins because light pushes it.”
  Not the main effect in a standard Crookes radiometer.

• “More vacuum = more spin.”
  No. There is an optimal partial-vacuum range.

• “Only black-vs-white optical contrast matters.”
  Surface heating contrast matters, but rarefied-gas transport + edge geometry are decisive.

• “It proves perpetual motion from light.”
  No. It is a heat engine using absorbed radiant energy and thermal gradients.

One simple experiment idea

Try three conditions with the same radiometer:

1. strong illumination,
2. gentle external warming/cooling of the bulb,
3. low-light ambient.

Observe direction and speed changes. You’ll see behavior tied to thermal gradients and gas state, not a naive “light pressure always forward” picture.

References

1. Crookes radiometer overview (historical/operational summary).
   https://en.wikipedia.org/wiki/Crookes_radiometer

2. Thermal transpiration overview + classic references.
   https://en.wikipedia.org/wiki/Thermal_transpiration

3. O. Reynolds (1879), On certain dimensional properties of matter in the gaseous state...
   Philosophical Transactions of the Royal Society 170, 727–845.
   https://doi.org/10.1098/rstl.1879.0078

4. J. C. Maxwell (1879), On stresses in rarified gases arising from inequalities of temperature.
   Philosophical Transactions of the Royal Society 170, 231–256.
   https://doi.org/10.1098/rstl.1879.0067

5. SFU Physics Demo notes (good concise educational summary + references).
   https://www.sfu.ca/physics/demos/demos-experiments/crookes-radiometer-burnaby.html

6. W. Chen et al. (2020), Impact of Improved Design on Knudsen Force for Micro Gas Sensor.
   Micromachines 11(7):634.
   https://pmc.ncbi.nlm.nih.gov/articles/PMC7408172/

7. M. Scandurra (2004), Enhanced radiometric forces.
   arXiv:physics/0402011.
   https://arxiv.org/abs/physics/0402011

One-line takeaway

Crookes radiometers spin because temperature gradients in rarefied gas create radiometric forces—an elegant edge-case of kinetic theory that still powers modern microscale engineering ideas.