The Radial Velocity TATOOINE Search for Circumbinary Planets: Planet Detection Limits for a Sample of Double-lined Binary Stars - Initial Results from Keck I/Hires, Shane/CAT/Hamspec and TNG/Sarg Observations

We present preliminary results of the first and on-going radial velocity survey for circumbinary planets. With a novel radial velocity technique employing an iodine absorption cell we achieve an unprecedented RV precision of up to 2 m/s for double-lined binary stars. The high resolution spectra collected with the Keck I/Hires, TNG/Sarg and Shane/CAT/Hamspec telescopes/spectrographs over the years 2003-2008 allow us to derive RVs and compute planet detection limits for ten double-lined binary stars. For this initial sample of targets, we can rule out planets on dynamically stable orbits with masses as small as ~0.3 to 3 MJup for the orbital periods of up to ~5.3 years. Even though the presented sample of stars is too small to make any strong conclusions, it is clear that the search for circumbinary planets is now technique-wise possible and eventually will provide new constraints for the planet formation theories.

💡 Research Summary

The paper reports the first results of the “TATOOINE” radial‑velocity (RV) survey aimed at detecting circumbinary planets around double‑lined spectroscopic binary (SB2) stars. Traditional RV techniques have struggled with SB2 systems because the spectra of the two stellar components overlap, making it to disentangle their individual velocities difficult. The authors overcome this limitation by employing an iodine (I₂) absorption cell in conjunction with a novel data‑reduction pipeline that simultaneously models the spectra of both stars and the iodine reference lines. This approach yields an unprecedented RV precision of ≤2 m/s for each component, a factor of two to five better than previous attempts on similar systems.

Observations were carried out between 2003 and 2008 using three high‑resolution echelle spectrographs: Keck I/HIRES (10 m), the Telescopio Nazionale Galileo (TNG) SARG (3.58 m), and the Shane 3 m telescope with the CAT/Hamspec spectrograph. The combined data set consists of roughly 200 spectra covering ten SB2 targets. All targets are relatively bright, nearby binaries with orbital periods ranging from a few days to several tens of days, and they exhibit clearly separated spectral lines that enable the double‑line analysis.

The reduction pipeline first derives precise orbital solutions for the binary stars themselves, fitting for period, eccentricity, semi‑amplitude, and other Keplerian elements for both components simultaneously. Once the binary motion is removed, the residual RV time series are examined for periodic signals indicative of a third body orbiting the barycenter of the pair. To quantify detection limits, the authors perform extensive Monte‑Carlo simulations that inject synthetic planetary signals into the residuals, assuming dynamically stable circumbinary orbits as defined by the Holman & Wiegert (1999) stability criterion. By varying planetary mass, orbital period, and phase, they determine the minimum mass that would have been detectable at a given period with the existing data.

The analysis shows that, for the ten systems studied, planets with masses between roughly 0.3 and 3 MJup can be ruled out on stable orbits with periods up to about 5.3 years (≈2000 days). The sensitivity is strongest for short‑period planets (≤1 year), where the data can exclude even sub‑Jupiter masses; for longer periods the detection threshold rises toward the upper end of the quoted range. No statistically significant planetary signals were found in any of the residual RV series.

Although the sample size is modest, the work demonstrates that high‑precision RV measurements of SB2 stars are now technically feasible. The authors argue that expanding the survey to a larger, more diverse set of binaries and extending the time baseline beyond a decade will improve sensitivity to lower‑mass, longer‑period circumbinary planets. Moreover, coupling RV results with direct imaging or transit searches, as well as with observations of circumbinary protoplanetary disks, could provide a comprehensive picture of planet formation in binary environments. The paper concludes that the methodology opens a new observational window that will eventually place robust empirical constraints on competing planet‑formation theories—particularly those addressing how solid cores and gas envelopes can assemble and survive in the dynamically complex gravitational potential of a close binary star system.