Updated method for binary asteroid search in Gaia FPR

Planets are not the only celestial bodies that can have moons. We have already detected a few hundred asteroids in the Solar System that orbit the Sun while sharing the orbit with a companion. Binary asteroids, as they are called, are expected to be extremely common, ranging from an estimated ~15% of the small asteroids in the main belt to about 20-30% of the Trans-Neptunian objects (TNO) [1,2]. These objects are difficult to detect, and each technique favours different systems' configurations, creating a biased population of known binaries. 

The ~1 mas level of unprecedented precision achieved by Gaia allows us to detect signatures of an asteroid's photocenter wobble in the post-orbital fit residuals. To detect the astrometric signature due to an asteroid companion, we developed a robust and straightforward method to find and test periodicities found in the residuals of the orbital fit using only the astrometric measurements for all 30,000 asteroids in Gaia DR3 with at least one window with a minimum of 10 consecutive observations.

Our method consists of a simple model to interpret the period detected as a wobbling signal, characterised by spherical components with uniform albedo. Then, we physically characterise them by identifying celestial systems with densities and separations within acceptable parameter ranges. We perform statistical tests and physical properties filters to obtain the list of the most significant astrometric binary candidate detections. 

As a result, we obtained more than 350 astrometric binary candidates [3], where in a few cases, the candidates present characteristics that are different from most of the binaries we know, which indicates that our method might be able to find binary systems that are rarely detected with other techniques. Our method has proven to be successful with the recent confirmation of the binary nature of several objects, such as (3220) Murayama [4] with the light curve technique, and through the detection of stellar occultations ( check Raphaël Lallemand's work).

Gaia DR3 contained data from 34 months of observations for more than 150,000 asteroids, while FPR provides almost twice the observation span, with 66 months of observation for the same objects; hence, the opportunity for an optimisation in the binary search process for the Gaia FPR data and already preparing for DR4. Therefore, we revisit the method applied in the first binary search. We use a new error model by taking the combination of systematic and random components from the astrometric error Gaia FPR data as the error bars. We change the approach in the averaged residuals by calculating the weighted mean and its associated uncertainty. We also changed the distribution used in some of the steps of the procedures from uniform to Gaussian. We implemented a new statistical selection method based on the Benjamini-Hochberg procedure ("BH" procedure) that provides more confidence in the control of the rate of false detections in the selections. We performed a deeper study on the confidence intervals determination to find the optimal approach, and we updated the physical validation process by constraining the intervals of parameters we use in the selection and filtering processes.

In this work, we present the results from the binary search in FPR with the updated selection method, the comparison with the results from DR3 and the analysis of the implications of these improvements in the estimation of the binaries' physical properties.

References

[1] Pravec, P. and Harris, A. W. (2007). Icarus, 190(1):250–259.
[2] Fraser, W. C., et al. (2017). Nature Astronomy, 1(4):0088
[3] Liberato, L., et al. (2024). Astronomy & Astrophysics, 688, A50.
[4] Benishek, V., et al. (2025). Central Bureau Electronic Telegrams, No. 5507.

Acknowledgements

This work presents results based on data from the Gaia mission (ESA) processed by the Gaia Data Processing and Analysis Consortium (DPAC). DPAC is funded by national institutions, in particular those participating in the Gaia MultiLateral Agreement (MLA) (Gaia mission website and archive: https://www.cosmos.esa.int/gaia and https://archives.esac.esa.int/gaia). The project was supported by the French Agence Nationale de la Recherche, ANR, “GaiaMoons” ANR-22-CE49-0002-01, the Programme National de Planetologie, and the BQR program of Observatoire de la Côte d’Azur.