Einstein’s Gravity Survives 91 Cosmic Mergers—and Tightens the Leash on the Graviton

Somewhere far away, two compact objects spiraled together and merged. The merger shook spacetime, sending a ripple across the universe. On Earth, detectors felt that ripple—and 90 others like it. Together, these faint signals now form one of the most stringent tests yet of Albert Einstein’s general theory of relativity.

That test appears in a new preprint from the LIGO-Virgo-KAGRA collaboration. The team analyzed 91 confident gravitational-wave events: 42 new signals from the first part of the fourth observing run, called O4a, plus 49 from earlier runs. All were seen by at least two detectors and passed a strict false-alarm threshold of no more than one per thousand years. The question was simple to state, if hard to answer: do these ripples behave exactly as general relativity predicts?

No sign of new physics in the 91-event catalog

Gravitational waves are produced when massive objects accelerate. For merging black holes and neutron stars, general relativity predicts not just the overall shape of the signal but also the precise way its frequency and phase evolve as the objects spiral closer. If gravity were slightly different, that evolution would shift.

The collaboration did not simply assume the theory. It added flexible deviation parameters to the expected waveform—knobs that could absorb signs of new physics—and let the data decide whether those knobs preferred nonzero values. They focused on eight parameterized tests. These included modifications to post-Newtonian coefficients, which describe corrections to Newtonian gravity as orbital speeds rise; spin-induced quadrupole moments, which encode how a spinning object’s shape responds to its rotation; and possible changes to how gravitational waves travel.

To guard against mistakes in their waveform models, the team used two independent frameworks, TIGER and FTI, built around different ways of computing the expected signal. They also used a principal component analysis to search for combinations of deviations that single tests might miss.

The result is a null result in the best sense: no evidence for physics beyond general relativity. More than 90% of the events include the no-deviation prediction within their 90% credible intervals. The new analysis improves the bounds on deviations in the post-Newtonian coefficients by factors of 1.2 to 5.5, making many of these the tightest constraints yet from gravitational waves.

That does not mean every event is perfectly obedient. Of the 42 O4a events, four showed apparent deviations in some tests: GW230628 231200, GW231028 153006, GW231110 040320 and GW231123. But with 91 events analyzed, a few statistical outliers are expected. The paper attributes these to expected statistical fluctuations and, in specific cases, prior effects or waveform-model uncertainties. For example, GW231028’s propagation shift is attributed to a prior effect, and GW231123 was excluded from the combined propagation analyses because of waveform systematics. In most cases, the evidence still favors general relativity.

Kerr black holes still pass the no-hair test

General relativity says a black hole is remarkably simple: its mass and spin determine everything else. In particular, the way a spinning black hole’s shape responds to its rotation—its spin-induced quadrupole moment—is fixed at a specific value. Exotic alternatives, such as boson stars or gravastars, could have different values.

The team tested this by measuring a parameter that describes the combined spin-induced quadrupole of the two merging objects. The results are consistent with ordinary Kerr black holes, as general relativity predicts. The combined measurement is slightly offset from the null result because of correlations with the binary’s effective spin, but the paper concludes there is no evidence for non-black-hole objects or deviations from general relativity.

Propagation tests put a lighter leash on the graviton

Other tests probe the journey, not just the source. If the graviton—the hypothetical quantum of gravity—had a small mass, different frequency components of a gravitational wave would travel at slightly different speeds. That would distort the signal as it crossed the universe.

Combining data from many events, the collaboration updated the bound on the graviton’s mass to no more than 1.92 × 10^-23 eV/c² at 90% credibility. That is a modest improvement over the previous gravitational-wave limit of 2.23 × 10^-23 eV/c². A tighter bound comes from planetary ephemerides, at 1.01 × 10^-24 eV/c² at 99.7% confidence, but the two tests rely on different observations and are complementary.

The team also searched for a modified dispersion relation, which would make waves of different frequencies arrive at different times, and for anisotropic birefringence, in which the two polarizations of a gravitational wave travel differently. Both searches found no evidence for new physics. The birefringence analysis produced the best current constraints on 16 coefficients that describe possible symmetry breaking.

A separate test looked for line-of-sight acceleration, which could reveal that a binary is not isolated—perhaps orbiting a nearby supermassive black hole. Only a few events had signals long enough and clean enough for this test. GW170817 and GW190425 were consistent with zero acceleration. A third event, GW200115 042309, failed one of the reliability criteria, so the team did not claim a quantitative constraint.

An earlier event, GW190814, drives the combined principal-component offsets; those offsets vanish when it is left out.

Sharper ripples, sharper tests

The paper is a preprint and has not yet completed peer review. Its results are also limited by the finite sensitivity of the detectors and by the accuracy of the waveform models used to interpret the signals. Some tests apply only to subsets of the 91 events.

As the LIGO-Virgo-KAGRA observatories continue their fourth observing run and plan future observing runs, more events and better models should sharpen these tests further. For now, Einstein’s theory has passed another gauntlet. The ripples from distant black holes and neutron stars continue to arrive with the rhythm general relativity predicts. The absence of a clear deviation is not a dead end. It is a map of where new physics is not hiding—and a pointer to the faint signals and sharper analysis that might, one day, reveal a crack in the theory.