Gaia Exoplanet Orbits, Demographics, and Evolution Survey (GEODES): Characteristics of Three Long-Period Companions Accelerating their Host Stars

Technical Summary: Gaia Exoplanet Orbits, Demographics, and Evolution Survey (GEODES). I. Characteristics of Three Long-Period Companions Accelerating their Host Stars

Problem Statement
The upcoming fourth data release (DR4) from the European Space Agency's Gaia mission is projected to yield thousands of giant planet candidates, potentially revolutionizing the statistical understanding of cold Jupiters, their eccentricities, and occurrence rates. However, a significant fraction of these candidates are expected to be false positives, primarily arising from near-equal-flux binary stars. Furthermore, even genuine detections will require extensive follow-up to precisely determine orbital parameters and masses, as Gaia astrometric solutions alone may not constrain orbits to better than 20% for the majority of systems. To realize the demographic potential of Gaia DR4, a rigorous pre-release vetting process is required to identify promising targets, exclude stellar binaries, and characterize substellar companions using complementary data.

Methodology
The authors introduce the Gaia Exoplanet Orbits, Demographics, and Evolution Survey (GEODES), an observational campaign designed to identify and characterize the most promising potential hosts of planetary and substellar companions prior to the release of Gaia DR4. The survey targets stars exhibiting astrometric variations suggestive of low-mass companions, selected from three primary sources: Hipparcos-Gaia accelerations, the G23H catalog, and Gaia Renormalized Unit Weight Error (RUWE) and Astrometric Excess Noise (AEN) metrics.

For this initial paper, the authors present a joint analysis of three specific systems (HIP 18512, HIP 45839, and HIP 81991) that exhibit significant radial velocity (RV) trends and astrometric accelerations. The methodology combines:

1. Archival and New Observations: The study utilizes a multi-epoch dataset including Hipparcos, Hipparcos-Gaia, Gaia DR2, and Gaia DR3 absolute astrometry. It incorporates precision RVs from Keck/HIRES, La Silla/HARPS, and the Automated Planet Finder (APF) Levy spectrograph.
2. High-Resolution Imaging: Adaptive optics (AO) imaging was conducted using ShaneAO/ShARCS at Lick Observatory and archival Keck/NIRC2 data to detect or rule out stellar and brown dwarf companions.
3. Orbital Modeling: The authors employ two distinct orbital fitting codes to ensure consistency and robustness:
   • ethraid: A brute-force importance sampling approach exploring a broad range of candidate properties (mass, semi-major axis, eccentricity) to generate mass-separation joint posterior distributions.
   • octofitter: A full orbit fitting code using a non-reversible parallel tempered sampler (Pigeons.jl) that jointly fits absolute astrometry (including HGCA proper motion anomalies, Gaia DR2-DR3 proper motions, and Hipparcos intermediate astrometric data) and RV data.

Key Results
The paper presents the characterization of three systems, each representing a distinct outcome of the survey strategy:

• HIP 18512: A nearby (15.3 pc) K4V dwarf. The study recovers a known stellar companion (M2.5 V) at a projected separation of $10.87 \pm 0.07$ arcseconds ($\sim$166 AU). The combined RV and astrometric accelerations are fully consistent with the known binary companion, which has a mass of $0.552^{+0.064}_{-0.059} M_\odot$ and an orbital separation of $145^{+32}_{-14}$ AU. This system serves as a validation of the survey's ability to identify stellar binaries masquerading as planetary signals.
• HIP 45839: A K4V star at 24.2 pc exhibiting strong RV acceleration ($13.6$ m/s/yr) and astrometric acceleration. AO imaging detected a faint source ($K_s = 12.5$ mag) at $5.3$ arcseconds. However, photometric and astrometric analysis rules out this source as the cause of the acceleration (it is likely a background object or a white dwarf inconsistent with the system's age). The joint RV and astrometric analysis constrains the true companion to a semi-major axis of $a = 17.9^{+4.8}_{-2.7}$ AU and a mass of $M = 45.2^{+10.5}_{-12.7} M_{Jup}$. The $octofitter$ analysis indicates a $\sim$100% probability that this companion is a brown dwarf.
• HIP 81991: A metal-poor G5V star at 43.8 pc. The data indicate a companion with a separation of $a = 6.4^{+0.6}_{-0.3}$ AU and a mass of $M = 9.5^{+5.4}_{-2.2} M_{Jup}$. Unlike the other two systems, the companion here is more likely to be a giant planet ($65\%$ probability) than a brown dwarf ($35\%$ probability), based on the $octofitter$ results which incorporate short- and medium-baseline astrometric data to rule out high-mass, high-separation models.

Significance and Claims
The authors state that this work represents the first results of the GEODES survey, demonstrating the program's capacity to identify promising targets for post-DR4 validation. The paper claims that by combining astrometry, RVs, and direct imaging, the survey can effectively distinguish between stellar binaries, brown dwarfs, and giant planets. Specifically, the study highlights that:

1. A significant fraction of Gaia candidates will be stellar binaries, necessitating the vetting procedures employed here.
2. Joint analysis of RV and astrometric data is critical for constraining the masses and orbits of long-period companions that cannot be fully characterized by astrometry alone.
3. The survey strategy successfully recovers known companions (HIP 18512) and identifies new substellar candidates (HIP 45839 and HIP 81991) with distinct mass regimes, providing a template for the larger demographic studies anticipated with Gaia DR4.

The paper concludes that thorough pre-release vetting and follow-up are essential to realize the full potential of Gaia DR4 for exoplanet demographics, particularly for cold Jupiters and brown dwarfs.