THEORYMERIDIAN

Charge, Fields & Light Lab · Testing Coulomb's law

Could the exponent
be 2.1?

Coulomb's law says the force falls as 1/r². Three force laws that agree at 10 cm drift apart as the distance grows. How would you tell which one nature uses, and how sure could you be?Compare the curves, take a measurement, then look inside a shell.

Hypothetical comparison
01 / Force against distance
p = 2, Coulomb's lawp = 1.9 and p = 2.1, what-ifsYour exponent, and readings made from it

Computed from the comparison lawEvery curve is F = F₀ (r/r₀)−p with the same force F₀ = 89.9 µN at r₀ = 10 cm. Only p = 2 is Coulomb's law. The other curves are what-ifs, not measurements.

Difference from inverse square at this distance

0.00%
ExponentForcevs p = 2

Start here: pick an exponent, then try each view

01View

p = 2 is Coulomb's law. Any other value is a hypothetical law, set to give the same force at 10 cm. The buttons above give finer steps such as 2.001.

04Force axis

Real experiments

Measure where the answer
should be zero.

Fitting a force curve reaches a few percent. The sharp tests look for a field inside a charged metal shell, where an exact inverse-square law predicts none. Each result is written for a law 1/r2+δ.

Limits on δ from concentric-shell experiments. Each is a bound for that apparatus and analysis, not a proof that δ is exactly zero.
ExperimentYearResult for δ
Cavendish1772–73|δ| ≤ 0.02
Maxwell1870s|δ| ≤ 5 × 10⁻⁵
Plimpton and Lawton1936|δ| < 2 × 10⁻⁹
Williams, Faller and Hill1971δ = (2.7 ± 3.1) × 10⁻¹⁶
Real experiment

What the 1971 number means

The measured value is smaller than its own uncertainty, so it is consistent with zero. It does not show that the exponent is exactly 2. It shows that any difference is below a few parts in 10¹⁶ for that apparatus and that way of writing the deviation.

E. R. Williams, J. E. Faller and H. A. Hill, Phys. Rev. Lett. 26, 721 (1971).

Why a shell

A cancellation that needs exactly 2

From any point inside a uniformly charged shell, look both ways along a narrow cone. The far patch is larger by the square of the distance ratio. If each bit of charge pushes as 1/r², the larger patch is weaker by the same factor and the two cancel. With any other exponent a small field is left over.

S. J. Plimpton and W. E. Lawton, Phys. Rev. 50, 1066 (1936); J. D. Jackson, Classical Electrodynamics, 3rd ed., section I.2.

Theory

What the theory says

Inside Maxwell's theory the exponent is exactly 2: it follows from Gauss's law and symmetry. The experiments test that theory. Today the question is often phrased as a limit on the mass of the photon, which would change the law in a different way than a shifted exponent.

A. S. Goldhaber and M. M. Nieto, Photon and Graviton Mass Limits, arXiv:0809.1003.

The reasoning, in the open

Look under the model.

The calculation

The comparison law keeps units and normalization fixed for every exponent:

F(r) = F₀ (r / r₀)−p,   F₀ = 89.9 µN at r₀ = 10 cm

F₀ is the Coulomb force between two 10 nC point charges 10 cm apart. At twice the reference distance the force is 0.2679 F₀ for p = 1.9, 0.2500 F₀ for p = 2 and 0.2333 F₀ for p = 2.1. For p = 2.1 that is 6.70 % below the inverse-square value.

"Take a measurement" places readings at distances spaced evenly on a logarithmic scale from 10 to 40 cm, multiplies each ideal value by (1 + σz) with z drawn from a standard normal distribution, and fits a straight line to ln F against ln r. The slope gives the exponent. Its standard error is σ divided by the square root of the summed squared deviations of ln r.

"Look inside a shell" adds the pushes of all parts of a uniformly charged thin shell on a probe at distance xR from the centre, assuming each part pushes as 1/sp and that pushes add. The result is shown relative to the push of a point charge with the same total charge at distance R. For p = 2 it is zero everywhere inside; the page treats anything below 10⁻¹⁰ of the reference as zero, which is far above the rounding error of the calculation.

Read the numerical model · Validation record

Sources
Limits of this model

The what-if laws are a comparison device. They are not a full theory, and superposition is assumed for them, not derived. The synthetic readings have ideal, independent, normally distributed scatter; real force measurements also suffer from charge leaking away, spheres that are not points, and nearby objects. The shell here has charge spread evenly; in the real experiments the shell is a conductor, and the measured quantity is a voltage or a charge on an inner conductor, not the field itself. The table gives published limits; this page does not reproduce those experiments.

Keyboard, motion and privacy

Use Tab to reach each control, arrow keys for sliders, and Enter or Space for buttons. Everything that can be dragged can also be set with a slider. Nothing on this page moves by itself. Results are written as text. The page runs entirely in your browser with no accounts, trackers or external code. See the privacy notice.