⚛️ general relativity

Null test of general relativity based on gravitational waves

This paper proposes a null test scheme that converts gravitational wave deviation parameters into mass-mass relations to calculate the probability of General Relativity being correct, finding that while five early binary black hole events show inconclusive deviations not driven by signal strength, the method is broadly applicable to future data.

Original authors: Dawen Chen, Zeyuan Kang, Zhoujian Cao

Published 2026-09-25

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Dawen Chen, Zeyuan Kang, Zhoujian Cao

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

For over a century, Albert Einstein's theory of general relativity has stood as the definitive description of how gravity works. It explains how massive objects like stars and planets warp the fabric of space and time, dictating the motion of everything from falling apples to orbiting planets. While the theory has passed every test thrown at it, from the bending of starlight to the precise timing of atomic clocks, many physicists suspect it is not the final word. They believe that at the most extreme scales or highest energies, the rules might change. To find these limits, scientists have turned to a new messenger: gravitational waves. These are ripples in space-time itself, created when massive objects like black holes collide and merge. Because these events happen in the most violent environments in the universe, they offer a unique laboratory to see if Einstein's equations hold up or if they begin to crack under pressure.

A team of researchers has now applied a fresh way of looking at these cosmic collisions to see if the data truly supports Einstein. They focused on five specific events where black holes merged, detected by the network of gravitational wave observatories known as LIGO, Virgo, and KAGRA. Instead of simply checking if the waves matched Einstein's predictions, the team devised a method to see if the different parts of the signal told a consistent story about the masses of the colliding black holes. In general relativity, the way the waves change as the black holes spiral together is mathematically locked to their masses. If the theory is correct, every piece of the signal should point to the exact same pair of masses. If the theory is wrong, or if the data contains hidden errors, the different pieces of the signal might point to different masses, creating a contradiction.

The researchers took the data from these five black hole mergers and broke the signal down into its various components. They treated each component as a separate clue to the masses of the black holes. If Einstein's theory is perfectly correct, all these clues should converge on a single, clear answer. However, when they mapped out the results, they found that for some events, the clues did not line up perfectly. For two of the events, the data showed a significant tension with Einstein's predictions, suggesting that the different parts of the signal were pointing to slightly different mass values. For the other three events, the data was much more consistent with the theory. The researchers calculated the probability of these mismatches, finding that for the two conflicting events, there was a greater than 90 percent credible level that the data showed tension with general relativity (equivalently, a consistency level of less than 10 percent).

Despite these intriguing discrepancies, the authors are careful not to declare that Einstein's theory has been broken. They emphasize that these tensions are not yet conclusive proof of new physics. The mismatches could be caused by subtle imperfections in the mathematical models used to describe the black hole collisions, or by noise and calibration issues in the detectors themselves. The study also revealed that the strength of the signal did not determine the size of the mismatch; even the loudest, clearest signals showed some tension, while some quieter ones fit the theory well. This suggests the issue is not simply a matter of signal quality but may stem from the complex physics of the merger or the models used to interpret it.

The work represents a new kind of check on our understanding of gravity. By transforming the complex data into a comparison of mass estimates, the team created a "null test" that asks a simple question: do all parts of the signal agree on the nature of the source? While the results show some signs of deviation, the authors conclude that current observations may contain hints of new physics, but they are equally likely to be the result of systematic errors in our models or instruments. The study serves as a powerful diagnostic tool, urging scientists to refine their models and analyze more events with greater precision before claiming that the laws of gravity have finally been rewritten.

Technical Summary: Null Test of General Relativity Based on Gravitational Waves

Problem Statement
While General Relativity (GR) has passed numerous classical and modern tests, it is widely believed not to be the final theory of gravity. Gravitational wave (GW) astronomy provides a unique tool to probe the validity boundary of GR. Existing parameterized tests (e.g., ppE, TIGER, FTI frameworks) analyze GW data by introducing deviation parameters to post-Newtonian (PN) coefficients. While these studies indicate that deviation parameters are small, a critical methodological gap remains: determining how small is small enough to definitively confirm GR. Current methods often rely on the magnitude of deviation parameters without a rigorous statistical framework to assess whether the data truly favors GR over alternative theories, particularly when systematic biases or parameter correlations are present.

Methodology
The authors propose a "null test" scheme that extends existing parameterized approaches by transforming deviation parameters into mass–mass relations. The methodology proceeds as follows:

Waveform Modeling: The study utilizes the IMRPhenomXAS waveform model (and LVK's IMRPhenomPv2 for specific data analysis) covering the full inspiral-merger-ringdown (IMR) process. The inspiral phase is parameterized by modifying the PN coefficients ( $\psi_j$ and $\psi_{jl}$ ) in the phase evolution, allowing them to vary freely alongside standard source parameters (total mass $M$ , symmetric mass ratio $\eta$ , and spins).
Null Hypothesis Construction: In GR, specific PN coefficients are strictly determined by the source parameters ( $m_1, m_2$ , spins). The authors invert this relationship: for a given measured value of a deviation parameter (or PN coefficient), they derive the corresponding mass–mass relation curve that would be required if GR were true.
Intersection Analysis:
- Region A: The posterior distribution of the source masses ( $m_1, m_2$ ) derived from the standard GW analysis.
- Region B: The "GR possible region" in the mass–mass diagram derived from the posterior distribution of a specific PN coefficient (e.g., $\phi_{2,NS}$ , $\phi_{6,NS}$ ).
- The Test: If GR is correct, Region A and Region B must intersect. The authors calculate the accumulated probability (credible level) required for these regions to first touch.
Critical Probability: The "critical probability" is defined as the minimum accumulated probability at which the mass posterior region and the GR-consistent regions (derived from one or multiple PN coefficients) share a common intersection. A high critical probability indicates that the data is consistent with GR only at a high credible level (i.e., a low probability of tension), whereas a low critical probability suggests a tension with GR at a lower credible level.
Data Application: The scheme is applied to five binary black hole (BBH) events from the LIGO-VIRGO-KAGRA (LVK) O1/O2 runs (GW150914, GW151226, GW170104, GW170608, GW170814). The analysis tests single coefficients, as well as combinations of two, three, four, and five spin-independent PN coefficients.

Key Contributions

Extension to Full IMR: Unlike previous works [32–34] that were limited to the inspiral stage, this study extends the null test idea to full inspiral-merger-ringdown signals, utilizing the entire detection data to avoid uncertainties associated with inspiral-stage cutting.
Mass-Mass Diagram Framework: The paper introduces a method to map parameterized deviation results into mass–mass diagrams, borrowing the concept from pulsar binary timing tests. This allows for a direct visual and statistical comparison of consistency across different parameters.
Multi-Parameter Consistency: The authors generalize the test to multiple coefficients simultaneously, defining a critical probability where the mass region and multiple coefficient-derived GR regions intersect.
Quantitative Threshold: The study provides a concrete metric (critical probability) to answer "how small is small enough," moving beyond simple qualitative statements about deviation magnitudes.

Results

Event-Specific Tensions: The analysis reveals varying levels of tension with GR across the five events.
- GW150914 and GW170104: Exhibit high credible levels (> $90\%$ ) of tension with GR, meaning the data is consistent with GR only at a low credible level (or conversely, the tension is significant).
- GW151226, GW170608, and GW170814: Show much smaller credible levels of tension (mostly $<60\%$ ), indicating better consistency with GR.
SNR Independence: The study finds no monotonic correlation between the network Signal-to-Noise Ratio (SNR) and the strength of the GR deviation. For instance, GW150914 (SNR=26.0) and GW170104 (SNR=13.8) both show high tension, while GW170814 (SNR=17.7) shows lower tension. This suggests the observed tensions are not simply driven by signal strength.
Multi-Parameter Behavior: In multi-parameter tests, the critical probability is often determined by the most restrictive single parameter (the one requiring the largest accumulated probability to intersect the mass region).

Significance and Claims
The paper claims that current GW observation data may contain information deviating from GR, but cautions against interpreting these deviations as evidence for a breakdown of GR. The authors explicitly state that the observed tensions are not conclusive. They attribute potential deviations to several systematic sources:

Waveform Systematics: Limitations of the IMRPhenomPv2 model in the merger-ringdown regime and differences between waveform families.
Detector Artifacts: Noise artifacts and calibration uncertainties, particularly affecting lower-SNR events.
Prior Choices: The use of wide uniform priors on PN deviation parameters can produce broad posteriors.
Parameter Correlations: Degeneracies between PN coefficients and source parameters (spins, mass ratio).

The significance of this work lies in providing a complementary diagnostic tool rather than a definitive refutation of GR. The authors conclude that their null test scheme should motivate further investigation using more accurate waveform models, independent analysis pipelines, and a larger catalog of events (extending to GWTC-1 through GWTC-5). The scheme is designed to be easily generalized to other waveform models and applied to all available GW observation data.

Technical Summary: Null Test of General Relativity Based on Gravitational Waves

More like this