The Search for Planets Around Barnard's Star

For over half a century, controversy has swirled around Barnard’s Star and the possibility it could harbor extrasolar planets. Back when I was growing up in the 1970s, it was widely believed that this nearby red dwarf had a pair of Jupiter-like planets – a possibility that heavily influenced many at the time and further fired my personal interest in the nearby stars as a teenager. Unfortunately, over the following decades many doubts about the existence of these planets arose as a steady stream of independent searches failed to find evidence for any planets orbiting Barnard’s Star. So where does the search for planets orbiting Barnard’s Star stand today?



Barnard’s Star (also known as BD+4° 3561 and Gliese 699) is a type M4V red dwarf star located in the constellation Ophiuchus. At a distance of 5.95 light years, it is the second closest star system after α Centauri (see “The Search For Planets Around Alpha Centauri”). Like all red dwarfs, Barnard’s Star is smaller and much dimmer than the Sun with a radius of 0.20 times that of the Sun, a mass of 0.16 times and a luminosity of only 0.0035 times. Because of its low luminosity, Barnard’s Star has a V magnitude of only 9.5 and requires a small telescope to spot despite its proximity to our solar system. Barnard’s Star has a lower concentration of “metals” (i.e. elements heavier than helium) than the Sun and is estimated to have an age in excess of ten billion years.

Barnard’s Star is named after American astronomer E.E. Barnard (1857-1923) who discovered in 1915 that it had the highest proper motion of any known star prompting the name “Barnard’s Runaway Star”. With a proper motion of 10.3 arc seconds per year, it beat the previous record holder, Kapteyn’s Star cataloged in 1898 (see “Habitable Planet Reality Check: Kapteyn b”). Because of this high proper motion , large parallax and the relative ease of observing it from major observatories around the globe due to its location near the celestial equator, Barnard’s Star has been a target of detailed investigation over the last century like no other red dwarf in the sky.

The Sproul Observatory’s 61-cm telescope used by Peter van de Kamp to measure the motions of Barnard’s Star and other nearby stars starting in 1937. (Sproul Observatory)

One of the best known observers of Barnard’s Star is undoubtedly Dutch astronomer Peter van de Kamp (1901-1995). From 1937 to 1972, van de Kamp was the director of Swarthmore College’s Sproul Observatory in Pennsylvania where he specialized in precision astrometry and became highly respected in the astronomical community for the quality of his work. A firm believer that planetary systems were common, van de Kamp pushed the limits of his equipment and measurement techniques to make not only precision measurements of the parallax and motions of nearby stars, but also to search for the telltale reflex motion of the star that would indicate the presence of unseen companions.

In 1963, van de Kamp published the results of his analysis of 2,413 photographic plates of Barnard’s Star obtained between 1916 and 1962. He found evidence for a wobble in the motion of Barnard’s Star on the order of a few milliarc seconds which he interpreted as being due to an orbiting planet with a period of 24 years and a mass 1.6 times that of Jupiter (or MJ). In 1969, van de Kamp published a new analysis which now included an extra five years of new data indicating the presence of two planets with periods of 12 and 26 years with masses of 0.8 MJ and 1.1 MJ, respectively.

Reprint of “Nearby Stars” sent to the author in 1975 by Peter van de Kamp.

Peter van de Kamp continued accumulating more data and publishing new analyses of his results throughout the 1970s and into the 1980s refining the properties of the planets he claimed he found orbiting Barnard’s Star. These finds influenced many during the 1970s interested in extrasolar planets and nearby stars especially since they orbited the second closest known star. For example, the British Interplanetary Society chose Barnard’s Star as the target for their automated interstellar spacecraft study from 1973 to 1978 known as Project Daedalus. On a personal note, these planets piqued my interest in the nearby stars as a young budding astronomer and prompted me to write to Dr. van de Kamp in 1975 for more information. He kindly sent me a reprint of his review paper, “The Nearby Stars”, which I still have in my collection of research material even after forty years. It was the first technical astronomical publication I had ever read and gave me a taste of how astronomers make their measurements and calculate properties of astronomical objects.

Unfortunately, it turns out that the planets found by van de Kamp orbiting Barnard’s Star do not exist. In 1973, George Gatewood (University of Pittsburgh) and Heinrich Eichorn (University of South Florida) published the results of their astrometric analysis of 241 photographic plates of Barnard’s Star obtained at the Allegheny Observatory and Van Vleck Observatory which found no evidence for the existence of the putative planets. Subsequent analyses of astrometric data from many other observatories including the Hubble Space Telescope published over the subsequent three decades have all failed to find any evidence for van de Kamp’s planets. Based on these null results and a detailed analysis of van de Kamp’s original plates by John L. Hershey published in 1973 while he was working at the Sproul Observatory, it is now widely believed that the wobbles observed by van de Kamp and his team were caused by subtle instrument effects resulting from upgrades and routine telescope maintenance. The search for unambiguous signs of planets orbiting Barnard’s Star continues.


Recent Search Results

With over half a century of published results of searches for planets orbiting Barnard’s Star, by far the most sensitive and comprehensive to date is a 2013 paper with Jieun Choi (University of California – Berkeley) as the lead author. Choi et al. presented an analysis of 248 radial velocity measurements of Barnard’s Star acquired between 1987 and 2012 using equipment at the Lick and Keck Observatories. While the earliest radial velocity measurements had a precision of only 20 meters per second, later ones were as good as 2 meters per second as a result of decades of improvements in equipment and refinements in measuring techniques.

Choi et al. carefully analyzed 136 Keck measurements acquired over 8 years which had a typical RMS scatter in the 2.5 to 4.2 meter per second range after correcting for the effects of secular acceleration (i.e. a geometrical effect that alters the radial velocity and proper motion of a star as a result of changing perspective as viewed from the Earth over time) which Choi et al. calculated to be 4.515±0.002 meters per second per year. The team found no evidence of radial velocity variations down to a level of 2 meters per second – the same as the estimated natural noise or “jitter” from various forms of surface activity on Barnard’s Star. Extending the analysis to the 19-year combined Lick/Keck data set, Choi et al. found no significant slope or curvature in the data with a scatter of 6.2 meters per second (the larger amount of scatter reflects the greater measurement uncertainty of the older data). Attempts to find the signature of van de Kamp’s planets orbiting Barnard’s Star (or variations of them) by Choi et al. were completely unsuccessful. As Choi et al. state in their paper, “…there can be little doubt now that van de Kamp’s two putative planets do not exist.”

This plot shows a quarter century’s worth of radial velocity measurements from the Lick and Keck Observatories compared to the signals expected from the putative planets of Barnard’s Star found by Peter van de Kamp. Click on Image to enlarge. (Choi et al.)

With no indications of long period signals in the combined Lick-Keck data set, Choi et al. performed a more detailed analysis looking for periodic signals at or below the noise level of the data. They used 15 years of Keck data to look for long period variations in the radial velocity but only the highest quality data obtained after August 2004 to search for shorter period variations. Their periodogram analysis found no peaks that can be convincingly associated with a planet orbiting Barnard’s Star with periods out to 5,000 days. Choi et al. also analyzed an independent data set of 226 published radial velocity measurements acquired using the VLT (Very Large Telescope) at the European Southern Observatory on Cerro Paranal in Chile which spanned six years with a typical uncertainty of 3.4 meters per second. Alone or in combination with the Keck data, Choi et al. again found no convincing periodic signal in the data that could be attributed to a planet orbiting Barnard’s Star.

Choi et al. performed a Monte Carlo analysis to determine the upper limits of planet masses that could escape detection in their data. To do this, they injected artificial signals representing planets with various masses and orbital parameters and analyzed the data to determine if they could detect it. Since radial velocity measurements alone are unable to constrain the inclination of the planetary orbit, i, the actual mass of the planet, MP, can not be determined – only its minimum mass or MPsini value can be determined. Choi et al. found that for circular orbits with periods less than 10 days, 100 days and two years, planets orbiting Barnard’s Star with MPsini values greater than about 2, 3 and 10 times that of the Earth (or ME), respectively, can be excluded. If we assume a random orientation of the orbit, this is the equivalent of there being a 95% probability that planets with actual masses of about 7 ME, 10 ME and 32 ME do not exist with orbital periods of less than 10 days, 100 days and two years, respectively. The habitable zone of Barnard’s Star, which Choi et al. take to range from 0.05 to 0.1 AU (corresponding to periods in the 10 to 30 day range), is devoid of any planets with an MPsini values greater than 3 ME. Assuming an unconstrained inclination, this excludes planets with an actual mass greater than about 10 ME to a 95% confidence level. The limits for planets eccentric orbits are slightly higher but comparable.

The detection threshold in terms of minimum mass as a function of orbital period based on an analysis of Keck radial velocity data by Choi et al. for planets with various orbital eccentricities, e. Click on image to enlarge. (Choi et al.)

While Choi et al. have expressed a certain level of surprise by the lack of any planets detected orbiting Barnard’s Star, it is not unexpected given what we have recently learned about the planetary systems of other M-dwarf stars. A number of studies using radial velocity surveys and Kepler results in recent years have shown that gas giants are uncommon around stars with low metallicity like Barnard’s Star. As a result, the lack of any evidence of gas giants out to periods on the order of a decade or more is not surprising. As for smaller planets, a recent statistical analysis of the Kepler database for M-dwarf stars performed by Courtney Dressing and David Charbonneau (Harvard-Smithsonian Center for Astrophysics) has shown that planets with radii greater than about 2.5 times that of the Earth (corresponding to a mass of about 4.5 ME assuming a Neptune-like density) and orbital periods less than 200 days are uncommon. Planets larger than Neptune with a mass of about 17 ME are exceptionally rare in M-dwarf systems. And since the “typical” M-dwarf in the analysis by Dressing and Charbonneau is over three times more massive than Barnard’s Star (with the corresponding planets also tending to be larger), the lack of any planetary detections to date is not all that surprising (see “Occurrence of Potentially Habitable Planets Around Red Dwarfs”).



Despite the deservedly high regard the astronomical community has had for the work of Peter van de Kamp, there is now no doubt that the planets he claimed he found orbiting Barnard’s Star simply do not exist. Decades of independent astrometric measurements and, in recent decades, precision radial velocity measurements have not only failed to find these planets, they have failed to find any evidence for Jupiter-size gas giants in orbits with periods as great as a decade or more. Given that recent work strongly suggests that gas giants are uncommon around low metallicity stars like Barnard’s Star and are exceptionally rare around red dwarfs in general, the lack of gas giant orbiting Barnard’s Star is not unexpected.

The latest analysis of precision radial velocity measurements of Barnard’s Star also places stringent upper limits on the masses of any planets with orbital periods less than a few years. To a 95% level of certainty, planets with actual masses greater than about 7 ME, 10 ME and 32 ME do not exist with orbital periods of less than 10 days, 100 days and two years, respectively. While there are some who have been puzzled by this result, it is not atypical of red dwarfs given the latest statistical analysis of Kepler mission results which finds that planets of this size are not very common among the “typical” red dwarfs studied. Given that Barnard’s Star has less than a third of the mass of these “typical” red dwarfs, this finding should even be less surprising. For example, Kepler 42 is a near twin of Barnard’s Star in terms of mass and metallicity but its three known sub-Earth-size planets with orbital periods of 0.5 to 1.9 days would escape detection to date if they orbited Barnard’s Star.

On a positive note, the presence of a planet with a minimum mass or MPsini greater than 3 ME has been excluded in the habitable zone of Barnard’s Star. Given that recent analyses of Kepler data have shown that planets with masses greater than about 6 ME are more likely to be volatile-rich mini-Neptunes with little possibility of supporting life as we know it (see “Habitable Planet Reality Check: Terrestrial Planet Size Limit”), the search results to date exclude the presence of mini-Neptunes or larger worlds to something like an 87% confidence level. It is more likely that if a planet orbits inside the habitable zone of Barnard’s Star, it is a terrestrial planet with a chance of being potentially habitable.

The future holds much promise for finding planets orbiting our neighbor. Continued improvements in the accuracy of radial velocity measurements could reduce the current upper limits for MPsini by a factor of several over the coming years. Even greater sensitivity improvements for potential planets in longer period orbits should be possible in a few years when astrometric measurements from ESA’s Gaia mission become available. We might only be a few years away from finally discovering bona fide planets orbiting Barnard’s Star after over a half a century of false starts.


Follow Drew Ex Machina on Facebook.


Related Reading

“The Search for Planets Around Proxima Centauri”, Drew Ex Machina, February 23, 2015 [Post]

“The Search For Planets Around Alpha Centauri”, Drew Ex Machina, August 11, 2014 [Post]

“The Search For Planets Around Alpha Centauri – II ”, Drew Ex Machina, September 25, 2014 [Post]

“Habitable Planet Reality Check: Kapteyn b”, Drew Ex Machina, June 6, 2014 [Post]

“Occurrence of Potentially Habitable Planets Around Red Dwarfs”, Drew Ex Machina, January 12, 2015 [Post]

“Habitable Planet Reality Check: Terrestrial Planet Size Limit”, Drew Ex Machina, July 24, 2014 [Post]


General References

A Bond and A.R. Martin, “Project Daedalus: the mission profile”, Journal of the British Interplanetary Society, Vol. 29, pp. 101-112, 1976

Jieun Choi et al., “Precise Doppler Monitoring of Barnard’s Star”, The Astrophysical Journal, Vol. 764, No. 2, Article id. 131, February 2013

Courtney D. Dressing and David Charbonneau, “The Occurrence of Potentially Habitable Planets Orbiting M Dwarfs Estimated from the Full Kepler Dataset and an Empirical Measurement of the Detection Sensitivity”, arVix 1501.01623 (submitted to The Astrophysical Journal), January 7, 2015 [Preprint]

George Gatewood and Heinrich Eichorn. “An unsuccessful search for a planetary companion of Barnard’s Star (BD + 4° 3561)”, Astronomical Journal, Vol.78, No. 8, pp. 769-776, October 1973

John L. Hershey, “Astrometric analysis of the field of AC +65° 6955 from plates taken with the Sproul 24-inch refractor”, Astronomical Journal, Vol. 78, No. 5, pp. 421-425, June 1973

Peter van de Kamp, “Astrometric Study of Barnard’s Star”, Astronomical Journal, Vol. 68, p. 295, No. 5, June 1963

Peter van de Kamp, “Parallax, Proper Motion, Acceleration, and Orbital Motion of Barnard’s Star”, Astronomical Journal, Vol. 74, No. 2, pp. 238-240, March 1969

Peter van de Kamp, “Alternate Dynamical Analysis of Barnard’s Star”, Astronomical Journal, Vol. 74, No. 6, pp. 757-759, August 1969

Peter van de Kamp, “The Nearby Stars”, Annual Review of Astronomy and Astrophysics, Vol. 9, pp.103-126, 1971

Peter van de Kamp, “Astrometric study of Barnard’s star from plates taken with the Sproul 61-cm refractor”, Astronomical Journal, Vol. 80, No. 8, pp. 658-, August 1975

Peter van de Kamp, “Dark companions of stars – Astrometric commentary on the lower end of the Main Sequence”, Space Science Reviews, Vol. 43, pp. 211-327, April 1986