It has been almost two decades since the first extrasolar planets were discovered orbiting Sun-like stars using precision radial velocity measurements. But as more accurate instruments become available and more powerful mathematical techniques are employed to analyze their data to find ever smaller planets, astronomers have been confronted by new issues that illustrate the limits of their knowledge of the stars they are studying. While the planetary interpretation of the radial velocity signals for hundreds of stars is quite secure, there are an increasing number of more marginal detections that have been questioned as the data are independently examined by others. Unfortunately, many of these questionable detections involve planets of particular interest to astronomers and general public alike – namely Earth-size planets orbiting in the habitable zone.
There are many potentially interesting extrasolar planets that stubbornly remain unconfirmed or are strongly suspected as being nothing more than stellar activity mimicking the signature of a planet. The latest “casualty” appears to be a planet orbiting the nearby red dwarf star known as Kapteyn’s Star which made headlines only a year ago with some claiming it to be the nearest and oldest potentially habitable planet known.
Background
Kapteyn’s Star (also known as Gl 191 and HD 33793) is a type M1 red dwarf located 12.75 light years away in the southern constellation of Pictor. It was named after Dutch astronomer Jacobus Cornelius Kapteyn (1851-1922) who cataloged it in 1898 and discovered that it had the highest proper motion of any star then known. It was not until the discovery Barnard’s Star (the second closest star to our Solar System currently known) in 1916 that a faster moving star was found (see “The Search for Planets Around Barnard’s Star”). Kapteyn’s Star has a radius of 0.29 times that of the Sun, an estimated mass of 0.28 times and a luminosity of about 0.012 times. With an age estimated to be in excess of ten billion years, it is over twice as old as our Sun and the closest known member of the galactic halo.
In May of 2014, an international team of astronomers announced that they had discovered two planets orbiting Kapteyn’s Star in a paper submitted for publication with Guillem Anglada-Escudé (Queen Mary University of London) as the lead author. They based their claim on a careful analysis of new and archival radial velocity measurements obtained using the HARPS, HIRES and DPS instruments. Considering the extreme age of the star and the lack of any obvious signs of stellar activity (e.g. brightness variations associated with star spots or flaring), Anglada-Escudé et al. were confident that the periodic variations they observed in the star’s radial velocity was due to the reflex motion of a pair of orbiting planets.
Details of the orbit of Kapteyn b with the system’s habitable zone shown in green. Click on image to enlarge. (PHL/UPR Aricebo)
The larger of the two planets, designated Kapteyn c, has an orbital period of 121.5 days and a minimum mass or MPsini (where i is the unknown inclination of the planet’s orbit to the plane of the sky) of about 7 times that of the Earth (or ME). The smaller planet, Kapteyn b, has an orbital period of 48.6 days and a MPsini of 4.8 ME. Given that this planet apparently orbits comfortably inside the habitable zone of Kapteyn’s Star, it received a lot of attention in the press at the time of its discovery as being the oldest and nearest potentially habitable planet yet found. I have argued that such a claim is potentially unwarranted (see “Habitable Planet Reality Check: Kapteyn b”).
First there are the usual unresolved issues of habitability associated with red dwarfs not to mention the new question raised by this discovery of the sustained habitability of ancient worlds over twice as old as the Earth. But in addition to these issues is the question of the bulk composition of large planets like Kapteyn b. Statistical analyses of Kepler data and follow up measurements of the masses of these worlds indicate that extrasolar planets transition from being predominantly rocky planets like the Earth to volatile-rich planets like Neptune at a mass of no greater than 6 ME (see “Habitable Planet Reality Check: Terrestrial Planet Size Limit”). Given that the measured minimum mass of Kapteyn b is 4.8 ME and that its actual mass is likely to be larger, it is slightly more probable that Kapteyn b is a mini-Neptune with no prospect of being habitable in the conventional sense rather than a rocky planet like the Earth. While certainly worthy of additional study, I felt that the potential habitability of this extrasolar planet was being overstated by the press and others.
New Interpretations of Old Data
It has been known for decades that certain types of stellar activity can mimic the signature of an orbiting planet. While astronomers searching for extrasolar planets make every effort to eliminate these sources of error in their data analysis, techniques like precision radial velocity measurements are pushing the limits of not only the instruments and data analysis methods, but also our understanding of the stars being studied. This issue is becoming increasingly important as the precision of radial velocity measurements now regularly exceeds the natural noise or “jitter” caused by normal surface activity in even well behaved, quiescent stars.
While red dwarfs are known to display obvious signs of surface activity like star spots, they can also exhibit more subtle forms of activity that can mimic a planetary signature. (David A. Aguilar – CfA)
While there had been a number of papers published questioning the presence of some smaller planets orbiting particular red dwarfs, the first clear signs of trouble came in July 2014 when a team of astronomers lead by Paul Robertson (Pennsylvania State University) published their analysis of data for the nearby red dwarf known as GJ 581. Instead of looking at just the radial velocity measurements as other investigators had done, Robertson and his team examined the original spectra looking for more subtle signs of stellar activity. They found that instead of being orbited by as many as six planets including a pair that some claimed were potentially habitable, the data indicated that GJ 581 has only three planets with low levels of surface activity modulated by the 130-day rotation period of the star mimicking the radial velocity signature of the other planets (see “The Disappearing Habitable Planets of GJ 581”).
HARPS team members Guillem Anglada-Escudé and Mikko Tuomi recently published a response to the claims of Robertson et al. and continue to assert that one of the more interesting of their finds, GJ 581d, actually exists. While they have rightfully taken issue with some of the analysis techniques employed by Robertson et al. in their original paper, the fact remains that the 66-day orbital period of GJ 581d is one-half of the star’s rotation period making a planetary interpretation of the radial velocity signature questionable by the standards typically followed by those in this field including the HARPS team (see “Habitable Planet Reality Check: GJ 581d”).
In August 2014, Paul Robertson and team member Suvrath Mahadevan (Pennsylvania State University) submitted a similar analysis for another nearby red dwarf star, GJ 667C. Once again, the HARPS team working with other astronomers claimed to have found as many as seven planets orbiting GJ 667C based on their analysis of radial velocity measurements. As with GJ 581, independent analysis of the data performed by others raised some doubts about these interpretations. When Robertson and Mahadevan examined the publicly available HARPS spectra of GJ 667C from which the radial velocity measurements were originally derived, they once again claimed to have found clear evidence for magnetic surface activity in the red dwarf modulated by its 105-day rotation period. After they corrected the radial velocity measurements for these effects, the signature of the potentially habitable GJ 667Cd with a 92-day orbital period completely disappeared. In fact, Robertson and Mahadevan could only find clear evidence for just two planets orbiting GJ 667C (see “Habitable Planet Reality Check: GJ 667C”).
Robertson and Mahadevan were rejoined by Arpita Roy (who had participated in the analysis of data for GJ 581) to review the data for Kapteyn’s Star. Since normal stellar activity tracers like Ca II H&K lines and V-band photometry have been shown not work well with M dwarfs, Robertson and his team once again examined the Hα and Na I D lines in 95 publicly available HARPS spectra for Kapteyn’s Star along with the Hα lines from 20 HIRES spectra in the Keck Observatory archive. Their analysis of these tracers of stellar activity in the combined data sets showed a clear 143-day periodicity. They interpret this as being the result of subtle surface activity on Kapteyn’s Star modulated by its rotation with a period of 143-days – a reasonable value given this star’s age and other properties.
Seeking to avoid the issues they encountered in their treatment of the noise in the data for GJ 581 caused by surface activity in that star, Robertson et al. this time concentrated on looking for correlations between their surface activity tracers and the radial velocity measurements. They noted that stellar activity was strongly correlated with the radial velocity signal with a period of about 48 days – the same as the purported orbital period of Kapteyn b and exactly one-third of the period of rotation derived by Robertson et al.. In similar cases in the past with other stars, this would normally be sufficient grounds to question the planetary interpretation of the radial velocity signature.
An artist’s impression of the size and appearance of Kapteyn b compared to the Earth – assuming that Kapteyn b is actually a terrestrial planet and not a mini-Neptune. (PHL/UPR Arecibo)
The conclusion of Robertson, Mahadevan and Roy in their new work is that the stellar activity and the particular way it was sampled in the HARPS and HIRES data sets created a false planet-like signal with a period of 48 days. This analysis suggests that Kapteyn b does not exist and is nothing more than an artifact. While there does not seem to be any obvious correlation between surface activity and the radial velocity signature for Kapteyn c, its 121-day orbital period is suspiciously close to the 143-day rotational period of the star that they derive. The team recommends continued observation of Kapteyn’s Star to confirm the planetary interpretation of this radial velocity signal.
Conclusions
Once again we find ourselves in a situation where there are conflicting interpretations of the subtle radial velocity variations of a red dwarf star. While members of the HARPS team will undoubtedly respond with a critique of the analysis by Robertson, Mahadevan and Roy, the fact that the orbital period of Kapteyn b is exactly one-third of the derived rotational period of Kapteyn’s Star casts grave doubts on the planetary interpretation of the radial velocity data. Unless there is convincing evidence to the contrary presented by the HARPS team or others, it appears that Kapteyn b simply does not exist.
As the controversy over the purported small planets of GJ 581, GJ 667C and Kapteyn’s Star continues to play out, it is clear that new precautions in the acquisition of radial velocity data and its analysis need to be taken to avoid false positives when searching for planets orbiting low-mass M-dwarf stars. And as this controversy as well the questions about the existence of α Centauri Bb (see “The Search for Planets Around Alpha Centauri”) and planets orbiting other more Sun-like stars illustrate, we still have much to learn about the effects of stellar activity on radial velocity even as scientists create instruments with ever improving accuracy.
Follow Drew Ex Machina on Facebook.
Related Reading
“Habitable Planet Reality Check: Kapteyn b”, Drew Ex Machina, June 6, 2014 [Post]
“The Disappearing Habitable Planets of GJ 581”, Drew Ex Machina, July 7, 2014 [Post]
“Habitable Planet Reality Check: GJ 667C”, Drew Ex Machina, September 7, 2014 [Post]
“Habitable Planet Reality Check: GJ 581d”, Drew Ex Machina, March 9, 2015 [Post]
General References
Guillem Anglada-Escudé and Mikko Tuomi, “Comment on ‘Stellar activity masquerading as planets in the habitable zone of the M dwarf Gliese 581’”, Science, Vol. 347, No. 6226, p. 1080, March 6, 2015
Guillem Anglada-Escudé et al., “Two planets around Kapteyn’s star: a cold and a temperate super-Earth orbiting the nearest halo red-dwarf”, Monthly Notices of the Royal Astronomical Society Letters, Vol. 484, No. 1, p. L89-L93, September 2014 [Preprint]
Paul Robertson, Suvrath Mahadevan, Michael Endl and Arpita Roy, “Stellar activity masquerading as planets in the habitable zone of the M dwarf Gliese 581”, Science, Vol. 345, No. 6195, pp. 440-444, July 25, 2014
Paul Robertson and Suvrath Mahadevan, “Disentangling Planets and Stellar Activity for Gliese 667C”, The Astrophysical Journal Letters, Vol. 793, Article ID. L24, October 1, 2014 [Preprint]
Paul Robertson, Suvrath Mahadevan, Michael Endl and Arpita Roy, “Response to Comment on ‘Stellar activity masquerading as planets in the habitable zone of the M dwarf Gliese 581’”, Science, Vol. 347, No. 6226, p. 1080, March 6, 2015
Paul Robertson, Suvrath Mahadevan and Aripita Roy, “Stellar Activity Mimic a Habitable Zone Planet Around Kapteyn’s Star”, arXiv 1505.20778 (submitted for publication in The Astrophysical Journal Letters), submitted May 11, 2015 [Preprint]