Astrophysicist Kyu-Hyun Chae of Sejong University has developed a new technique to test gravity at very low accelerations using the full 3D velocities of wide binary stars. Unlike previous methods limited to 2D sky-projected data, Chae's approach leverages Bayesian inference and Markov Chain Monte Carlo simulations to derive the probability distribution of a gravity parameter directly from relat by Riko Seibo
Tokyo, Japan (SPX) May 28, 2025
Astrophysicist Kyu-Hyun Chae of Sejong University has developed a new technique to test gravity at very low accelerations using the full 3D velocities of wide binary stars. Unlike previous methods limited to 2D sky-projected data, Chae's approach leverages Bayesian inference and Markov Chain Monte Carlo simulations to derive the probability distribution of a gravity parameter directly from relative stellar velocities.
"The new method overcomes all these limitations. It is a sort of revolutionary and ultimate method for wide binaries whose motions can only be 'snapshot-observed'," said Chae, stressing the method's need for accurate radial velocity data.
Xavier Hernandez, a pioneer in using wide binaries to test gravity, called the method "a fully rigorous Bayesian approach which will surely become the standard in the field." The transition to 3D velocity data, he noted, significantly boosts accuracy.
Chae applied the method to about 300 binaries from ESA's Gaia DR3 with relatively accurate radial velocities. The results align with prior findings from both Chae and Hernandez, reinforcing that at accelerations below about 1 nanometer per second squared, gravitational strength appears 40 to 50 percent greater than predicted by Newtonian physics. The deviation, statistically significant at 4.2 sigma, mirrors predictions made by modified Newtonian dynamics (MOND).
Pavel Kroupa of the University of Bonn commended the study, describing it as "an impressive study of gravitation using very wide binaries as probes taken to a new level of accuracy and clarity."
MOND creator Mordehai Milgrom said Chae's findings underscore a fundamental departure from Newtonian expectations in low-acceleration systems-one that cannot be attributed to dark matter. "The magnitude of the anomaly they find is also consistent with the generic predictions of existing MOND theories," Milgrom said.
Chae's team is now incorporating precision radial velocity data from GEMINI North, Las Cumbres Observatory, and other sources. In parallel, Hernandez's group is conducting speckle photometry to exclude systems with hidden stars that could skew results. These efforts aim to sharpen the Bayesian analysis further and validate the modified gravity signal.
"With new data on radial velocities, and results from speckle photometric observations, the Bayesian inference is expected to measure gravity sufficiently precisely, not only to distinguish between Newton and MOND well above 5sigma, but also to narrow theoretical possibilities of gravitational dynamics," said Chae.
Research Report:Low-Acceleration Gravitational Anomaly from Bayesian 3D Modeling of Wide Binary Orbits: Methodology and Results with Gaia Data Release 3
Related Links
Sejong University
The Physics of Time and Space