Revised Mass and Orbit of $\varepsilon$ Eridani b: A 1 Jupiter-Mass Planet on a Near-Circular Orbit

TL;DR

This work tackles the challenge of precisely characterizing the orbit and mass of the nearby giant planet ε Eridani b by performing a thorough joint analysis of archival radial velocity and absolute astrometry from eight RV instruments and four astrometric sources. It combines careful handling of perspective accelerations, light-travel time corrections, and Gaia–Hipparcos frame correlations with a Gaussian-process model of stellar activity, implemented in a robust, multi-dataset framework. The results yield a planet mass of $m = 0.98\pm0.10\,\mathrm{M_{Jup}}$, a near-circular orbit with $e \in [0,0.10]$, a period of $P = 7.33\pm0.08$ yr, and an inclination of $i = 40^{+6}_{-5}\circ$ that is close to coplanar with the outer debris disk. These findings resolve previous tensions around eccentricity, inclination, and mass, and provide actionable predictions for imaging campaigns, supporting ε Eridani b as one of the closest Solar System analogs around a nearby star.

Abstract

The mature Jovian planet $\varepsilon$ Eridani b orbits one of the closest sun-like stars at a moderate separation of 3.5 AU, presenting one of the best opportunities to image a true analog to a solar system planet. We perform a thorough joint reanalysis and cross-validation of all available archival radial velocity and astrometry data, combining data from eight radial velocity instruments and four astrometric sources (Hipparcos, Hubble FGS, Gaia DR2, and Gaia DR3). We incorporate methodological advances that impact our findings including a principled treatment of correlation between Gaia DR2 and DR3 velocity and corrections for the changing light-travel time to this high proper motion system. We revise the planet's mass upward to $0.98 \pm 0.09 \, \mathrm{M_{jup}}$ and find that its orbit is nearly circular and close to coplanar with the outer debris disk. We further present one of the first models of an exoplanet orbit exclusively from absolute astrometry and independently confirm the planet's orbital period. We make specific predictions for the planet's location at key imaging epochs from past and future observing campaigns. We discuss and resolve tensions between previous works regarding the eccentricity, inclination, and mass. Our results further support that $\varepsilon$ Eridani b is one of the closest analogs to a Solar System planet yet detected around a nearby star.

Figures (9)

• Figure 1: Comparison of the orbit posteriors incorporating different sources of data. The blue dots mark the location of the planet on 2025-01-01. The colors of the orbit draws match those in Figure \ref{['fig:model-comparison']} for easier cross-referencing. The bottom right panel shows the outer debris disk model fit to ALMA data by epseri_alma_booth_2023. The ❉ indicates two models containing only RV data and only absolute astrometry data respectively, and whose posteriors are therefore completely independent. The models indicated by ✦ include RV data, but distinct sets of astrometric data, making the position angles they indicate on 2025-01-01 effectively independent.
• Figure 2: A comparison between posteriors from seven models orbit models, listed in the legend. We find no obvious discrepancy between any dataset. The bottom left panel is a corner plot, and the top right panel shows the predicted location on 2025-01-01, near the planets maximum separation from the star and the epoch of recent JWST observations. All contours encompass the volume contained within the $1\sigma$ 2D Gaussian equivalent. The inclination and position angle of the outer disk model from epseri_alma_booth_2023 fitted to ALMA data is over-plotted as dashed lines. The ❉ indicates two models containing only RV data and only absolute astrometry data respectively, and whose posteriors are therefore completely independent. The models indicated by ✦ include RV data, but distinct sets of astrometric data, making the position angles effectively independent. The posterior represented by black contours include all available RV and astrometric data. We draw the readers attention in particular to the $\Omega$ vs. $i$ and $\Omega$ vs. $m$ sub-panels which illuminate how each data source progressively constrains the orientation of the orbit.
• Figure 3: Maximum a-posteriori sample from the complete RV and astrometry model. Top: RV data from each instrument The scatter in the RV data are well-modeled by a Gaussian process with a quasi-periodic kernel, just as reported in previous works.
• Figure 4: Orbits sampled from the RV, Hipparcos, DR3 model (blue) and the complete RV, Hippacos, FGS, DR2, DR3 model (black) compared against all data. This shows how the addition of the DR2 and the FGS data constrains the motion of the star in the R.A. direction. The horizontal bars in the proper motion panels show the time range over which each measurement is fit by their respective catalogs. The light blue markers show binned FGS data with the size of the marker indicating the number of data points in the bin.
• Figure 5: Covariances between R.A. and Dec. velocity at each proper motion epoch, for the RV, Hip., FGS, DR2, DR3 model. The "H" epoch is the net proper motion at the Hipparcos epoch calibrated to the DR3 velocity reference by the HGCA, "G-H" is the long-term proper motion derived from the HGCA, "DR2" is the Gaia DR2 reported proper motion after correcting for frame rotation, and "DR3" is the Gaia DR3 proper motion. Both the DR2 and DR3 values have uncertainties inflated according to the multipliers given in the HGCA, DR2 hgca_brandt_2018 and eDR3 versions hgca_brandt_2021.