The Nordtvedt Effect and Lunar Laser Ranging: Testing How Gravity Falls

What the Nordtvedt effect is: the strong equivalence principle on trial

Galileo's famous claim — that all bodies fall alike regardless of composition — is the weak equivalence principle. Einstein's general relativity demands something stronger: that gravitational binding energy itself also falls at the universal rate. This is the strong equivalence principle (SEP). A rock and a planet may be made of the same atoms, but the planet is held together by an enormous cloud of negative gravitational self-energy, and general relativity insists that this energy weighs exactly as much inertially as it does gravitationally.

In 1968 the physicist Kenneth Nordtvedt showed that many alternative theories of gravity — notably scalar–tensor theories like Brans–Dicke — quietly break this rule. In those theories, a body's gravitational self-energy contributes unequally to its inertial mass (its resistance to acceleration) and its passive gravitational mass (how hard gravity pulls it). The result: two bodies with different fractional self-energy fall toward a third mass at slightly different rates. That differential acceleration is the Nordtvedt effect, and its size is set by a single dimensionless number, the Nordtvedt parameter η.

The mechanism: why Earth and Moon would fall differently toward the Sun

Write a body's passive gravitational mass as m_g = m_i × (1 + η × Ω/mc²), where m_i is the inertial mass and Ω/mc² is the fraction of the mass-energy stored as gravitational self-binding. In general relativity η = 0, so m_g = m_i exactly. If η ≠ 0, the ratio m_g/m_i depends on how tightly self-gravitating the body is.

The Earth and Moon fall together around the Sun, but they are not equally self-bound:

• Earth: Ω/mc² ≈ −4.6 × 10⁻¹⁰
• Moon: Ω/mc² ≈ −2 × 10⁻¹¹, about twenty times smaller

With η ≠ 0, the more strongly bound Earth is pulled toward the Sun with a slightly different acceleration than the Moon. The difference, Δa ≈ η × (Ω_E − Ω_M)/mc² × g_Sun, is tiny — but it is directional. It always points along the Earth–Sun line, so as the Moon orbits, the perturbation oscillates with the synodic month. The lunar orbit becomes polarized: pulled slightly toward the Sun at new Moon and away at full Moon, a periodic shift in the Earth–Moon range.

The numbers: a 13-meter yardstick and the η = 4β − γ − 3 relation

In the parametrized post-Newtonian (PPN) framework that catalogs weak-field gravity, the Nordtvedt parameter is a specific combination of the PPN parameters: η = 4β − γ − 3 (with additional preferred-frame and preferred-location terms in the fully general form). Here γ measures how much spacetime curvature a unit mass produces and β measures the nonlinearity of gravity. General relativity fixes β = γ = 1, giving η = 0 exactly.

The observable signature in lunar ranging is a range oscillation:

• δr ≈ A × η × cos(D), where D is the Moon's angular elongation from the Sun and the amplitude A ≈ 13 meters.
• The oscillation period is the synodic month, 29.53 days.

Worked scale: if η were as large as 10⁻³, the effect would displace the orbit by about 13 mm — comfortably above today's few-millimeter ranging precision. The current null result, |η| ≲ few × 10⁻⁴, therefore corresponds to keeping any such polarization below the millimeter level over decades of tracking. Combined with the Cassini bound γ − 1 ≈ (2.1 ± 2.3) × 10⁻⁵, LLR tightly pins β − 1 as well.

How it is observed: firing lasers at Apollo mirrors

The test is possible because five corner-cube retroreflector arrays sit on the Moon, each reflecting light straight back toward its source:

• Apollo 11 (1969), Apollo 14 and Apollo 15 — the last being the largest array — placed by U.S. astronauts.
• Lunokhod 1 and Lunokhod 2, carried by Soviet robotic rovers.

Observatories such as the McDonald station in Texas, the Côte d'Azur (Grasse) station in France, and the APOLLO apparatus at Apache Point, New Mexico fire short laser pulses at these targets. Of the roughly 10²⁰ photons launched, only a handful return per pulse, but timing the ~2.5-second round trip yields the Earth–Moon distance. Over the past half-century the precision has improved from a few hundred millimeters to a few millimeters. Analysts then fit a comprehensive orbital model — tides, relativity, planetary perturbations — leaving η as a free parameter. Any residual cos(D) term at the synodic period would be the Nordtvedt signal. So far the fitted amplitude is consistent with zero.

How it compares: cousins of the Nordtvedt test

The Nordtvedt effect is a strong-equivalence test, which distinguishes it from its relatives:

• Weak equivalence principle tests (Eötvös torsion balances, the MICROSCOPE satellite reaching ~10⁻¹⁵) compare bodies whose self-gravity is negligible — they probe composition, not binding energy. Only bodies as massive as planets, stars, or neutron stars can test the SEP.
• Binary pulsar tests: a millisecond pulsar orbiting a white dwarf, and especially the pulsar triple system PSR J0337+1715, subject a neutron star (self-energy fraction ~10⁻¹, ten billion times Earth's) to an external field. These extend the SEP test into the strong-field regime.
• The Cassini γ measurement uses radio time delay near the Sun to isolate γ, complementing LLR's constraint on the β–γ combination.

A crucial subtlety: a nonzero η can masquerade in the data. Because the Nordtvedt effect and a possible time-variation of Newton's constant G both leave slow imprints on the lunar orbit, LLR analyses fit them jointly, and the same data also bound Ġ/G to about 10⁻¹³ per year.

Significance and open questions

The Nordtvedt effect turned a diplomatic-era stunt — mirrors on the Moon — into a precision gravity laboratory. Its historical importance is hard to overstate: the very first analyses in 1976 by Williams, Shapiro and collaborators already gave η = 0.00 ± 0.03, and later work by J. G. Williams, Slava Turyshev, and Dale Boggs drove the bound down to the 10⁻⁴ level, confirming the strong equivalence principle and squeezing scalar–tensor alternatives.

Open frontiers remain:

• Next-generation ranging. New active retroreflector concepts and transponders aim for sub-millimeter, potentially tightening η by another order of magnitude.
• Disentangling systematics. Lunar interior, tidal dissipation, and thermal effects on the arrays are now the limiting uncertainties, not the lasers.
• Cosmological motivation. If dark energy is a light scalar field, it could induce a small but nonzero η; screening mechanisms in modified-gravity models predict SEP violations that LLR and pulsars are actively hunting.

To date every measurement is consistent with Einstein: gravitational energy falls exactly like everything else.

Equivalence principles and how the Nordtvedt effect relates to what lunar laser ranging measures

Principle / quantity	What it states or measures	Status / value
Weak equivalence principle (WEP)	Test bodies (negligible self-gravity) fall identically	Confirmed to ~10⁻¹⁵ (MICROSCOPE satellite, 2022)
Strong equivalence principle (SEP)	Even gravitational self-energy falls at the same rate	Tested by the Nordtvedt effect via LLR
Nordtvedt parameter η	Amplitude of SEP violation; η = 4β − γ − 3	|η| ≲ few × 10⁻⁴ (LLR); 0 in GR
Earth gravitational self-energy	Fractional binding energy Ω/mc²	≈ −4.6 × 10⁻¹⁰ of Earth's rest mass
Moon gravitational self-energy	Fractional binding energy Ω/mc²	≈ −2 × 10⁻¹¹ (≈ 20× smaller than Earth)
Ranging precision	Earth–Moon distance uncertainty	Improved from ~decimeter (1969) to few mm (today)