The Discovery of Alpha Centauri Bb: Three Years Later

I still have a clear memory of waiting in anticipation of the premier episode of the now classic sci-fi television show, Lost in Space, which first aired 50 years ago this season. The mission of the Jupiter 2 and the Robinson family in the then far off future date of October 16, 1997 was to start the colonization of an Earth-like planet found orbiting our closest neighboring star system, α Centauri (also known by the ancient name, Rigil Kentaurus). While hardly the first (or last) piece of science fiction to be centered on α Centauri, this show did bring this star system to my attention for the first time and helped spark my childhood interest in spaceflight in general and the nearby stars in particular.

A still showing the Jupiter 2 prior to its launch to α Centauri in the premier episode of Lost in Space which first aired on September 15, 1965. (CBS)

In reality, no such Earth-like planet was known to exist orbiting α Centauri back then and it would be another 30 years before astronomers would find the first extrasolar planet orbiting any Sun-like star – the decidedly unEarth-like hot Jupiter, 51 Pegasi b. But as planet detection techniques improved, surveys were expanded to find planets orbiting most of the nearby stars including those of the α Centauri system. As thousands of exoplanets were discovered over the coming decades, none were found orbiting our closets neighbor until October 16, 2012 when a Swiss-based team of astronomers announced the discovery of a planet orbiting α Centauri B using precision radial velocity measurements. While this planet, designated α Centauri Bb, was hardly the Earth-like world imagined in science fiction or hoped for by many, its presence demonstrated that the closest star system to us harbored at least one planet and held the promise of more to be discovered. But three years after this momentous announcement, this important discovery has yet to be independently confirmed and new data seem to have only made things more complicated.



The α Centauri system contains three stars. At its heart are a pair of Sun-like stars 4.37 light years from us that are locked in an 80-year orbit around each other. The larger component of this pair, α Centauri A, is a G2V star with 1.10 times the mass of the Sun and 1.52 times its luminosity. The smaller component, α Centauri B, is a slightly cooler K2V star with 0.93 times the mass of the Sun and 0.50 times its luminosity. The third member of this system is known as Proxima Centauri since it is slightly closer to us at a distance of 4.24 light years. This more distantly orbiting component of the system, some 15,000 AU from the pair of Sun-like stars, is a very small M5V red dwarf star with an estimated mass of only 0.12 times that of the Sun and just 0.0017 times its luminosity (see “The Search For Planets Around Proxima Centauri”).

This diagram shows the relative sizes of the Sun compared to the known members of the α Centauri system. Click on image to enlarge. (PHL)

Despite the distance between α Centauri A and B varying from 11 to 35 AU during the course of one revolution, various dynamical studies performed over the decades have predicted that regions with stable planetary orbits do exist in this system. These studies have shown that orbits out to about 3 AU, give or take, would be stable depending on their inclination to the plane of the orbit of α Centauri A and B about each other. What has not been so clear is if planets could form around this pair of stars.

A number of studies performed over the past couple of decades have been more or less evenly split on the question of whether or not planets could form around α Centauri A and B. Some studies have shown that the building blocks for planets, called planetesimals, would be able to collect themselves together into planets out to some reasonable distance. Still other studies have suggested that the presence of the two stars would have stirred up the orbits of the planetesimals too much. Instead of collecting into larger bodies, the planetesimals would tend to smash themselves apart upon contact so that planets could not form. In the end, the best way of answering this question was to look for planets.

Given the difficulty of detecting extrasolar planets even in a nearby star system like α Centauri, the first technology that offered reasonable chance of success was the precision measurement of changes in the stars’ radial velocity resulting from the small reflex motion of an orbiting planet. But after almost two decades of measurements with increasingly better instruments and refined techniques, the results of searches for planets orbiting α Centauri A and B published up to 2011 had found nothing. This null result combined with dynamical arguments only demonstrated that planets larger than Saturn or Jupiter did not orbit within about 2 AU of either α Centauri A or B. This still left a lot of possibilities including Earth-size planets orbiting comfortably inside the habitable zones of these stars but much more precise radial velocity measurements would be required to detect them (for a more detailed account of the searches performed to date, see “The Search For Planets Around Alpha Centauri – II”).

The ESO 3.6m Telescope equipped with HARPS. (ESO/H.H.Heyer)

Beginning around eight years ago, several teams employing various observing approaches are known to have started looking for lower-mass planets orbiting α Centauri A and B with instruments capable of making radial velocity measurements with uncertainties on the order of one meter per second – a factor of up to four more precise than in any previously published results for the system. The first team to announce results from their search was the European team using the HARPS (High Accuracy Radial Velocity Planetary Searcher) spectrometer on the 3.6-meter telescope at the European Southern Observatory in La Silla, Chile. According to the discovery paper by Dumusque et al., they employed a new data processing technique to extract the 0.5 meter per second signal of α Centauri Bb out of 459 radial velocity measurements they obtained between February 2008 and July 2011. These radial velocity data had a measurement uncertainty of 0.8 meters per second and contained an estimated 1.5 meters per second of natural noise or “jitter” resulting from a range of activity on the surface of α Centauri B.

The HARPS team’s analysis indicated the presence of a planet with a minimum mass or Mpsini (where i is the unknown inclination of the planet’s orbit to the plane of the sky) of just 1.1 times that of Earth, locked in a tight orbit with a period of 3.24 days and a radius of 0.04 AU. This was well below the upper limits set by earlier searches. Given sufficient observation time, the team estimated that in the future they could detect a planet with a Mpsini of about four times that of the Earth in a 200-day orbit inside the habitable zone of α Centauri B (for a more detailed account of the discovery of α Centauri Bb, see “The Search For Planets Around Alpha Centauri”)


Problems for α Centauri Bb

While the HARPS team’s result generated much excitement, it was also met with a healthy amount of skepticism in the astronomical community because of the new mathematical technique used to process the data to extract such a low-amplitude signal. One of the first critics, American astronomer Artie Hatzes (Thuringian State Observatory, Germany), performed his own analysis of the publicly available HARPS data set using two more widely employed data processing techniques to look for the radial velocity signal of α Centauri Bb. Formally published in June 2013, Dr. Hatzes’ analysis did indeed find a signal buried in the radial velocity data with a period of 3.24 days but it had a false alarm probability of a few percent – far too high to be considered a reliable detection. Furthermore, his analysis of the “random” noise in the data showed that it had periodicities in the 2.8 to 3.3 day range and amplitudes on the order of half that of the alleged planetary signal.

Given the situation of potential planetary false alarms with GJ 581, GJ 667C and Kapteyn’s Star, this finding suggested that noise in the data, especially from activity on α Centauri B, might have been mistaken for a planet (see “Habitable Planet Reality Check: GJ 581”, Habitable Planet Reality Check: GJ 667C” and “Kapteyn b: Has Another Habitable Planet ‘Disappeared’?”). Dr. Hatzes concluded that additional data were needed to better understand the nature of the noise in the radial velocity measurements and confirm the planetary nature of the radial velocity signal.

Other teams have already been taking data in order to confirm the existence of α Centauri Bb as part of their ongoing observing programs although no results have been formally published to date. A team of astronomers working with the 1.5-meter telescope at Cerro Tololo Inter-American Observatory (CTIO) in Chile are using CHIRON (CTIO Higher Resolution Spectrometer) to search for planets orbiting α Centauri A and B in part with the support of The Planetary Society. The project’s principal investigator, Debra Fischer (Yale University), was quoted in a blog on The Planetary Society’s web site posted last year by Bruce Betts that her team had insufficient data to detect α Centauri Bb when its discovery was announced in October 2012. They soon launched a renewed effort to gather much more data at a higher cadence starting in 2013 aimed specifically at detecting the purported planet’s 3.24-day signal. To date they have not announced the detection α Centauri Bb in their data but their simulations indicated that any such detection would have been marginal at best.

A diagram showing the apparent and actual trajectory of α Centauri B. The near-term minimum separation between α Centauri A and B as seen from the Earth will occur in December 2015. Click on image to enlarge. (Wikimedia Commons/SiriusB)

One of the issues complicating continuing efforts to gather more data needed to resolve the situation with α Centauri Bb is the increasing amount of stray light from α Centauri A that is degrading the quality of radial velocity measurements. As viewed from the Earth, the apparent separation of α Centauri A and B has been decreasing at an accelerating rate for about a third of a century as they move in their inclined elliptical paths around each other. The two stars are just now reaching a near-term minimum separation of only four arc seconds. The conventional wisdom has been that it will be several more years before the separation of α Centauri A and B increases enough to acquire new data of sufficient quality to confirm α Centauri Bb.

One of the other groups known to be searching the α Centauri system for planets is a team of astronomers using the HERCULES (High Efficiency and Resolution Canterbury University Large Echelle Spectrograph) spectrograph on the one-meter McLellan Telescope at the Mt. John University Observatory in New Zealand. Bergmann et al. wrote a paper published in April 2015 in the International Journal of Astrobiology where they described a new technique to reduce significantly the effects of stray light contamination in precision radial velocity measurements.

A view of the one-meter McLellan Telescope on Mt. John in New Zealand that is being used with the HERCULES spectrograph to search for planets orbiting α Centauri. (MJUO)

In order to test the effectiveness of their new technique, they observed four double-line spectroscopic binaries (i.e. pairs of unresolved stars that can only be differentiated by periodic Doppler shifts in their spectral lines) whose blended images represent the worse-case scenario of “contamination”. With the new technique, they were able to recover accurate radial velocities of both components of the observed spectroscopic binaries. The New Zealand-based team plans to use their new analysis method to process the data they are continuing to gather as part of their observing campaign of α Centauri which started in 2007. Their calculations show that they should be able to detect α Centauri Bb if it exists. But as of today, no results have been published by this team either to confirm or refute the discovery of α Centauri Bb.


New Method for Confirmation

While it looks as though it might be some time before new radial velocity measurements can confirm the existence of α Centauri Bb, there is another method available to provide independent verification using current technology. Precision photometry is capable of detecting the small decrease in a star’s brightness as a result of planetary transits. This method has been successfully used by NASA’s Kepler mission to detect thousands of extrasolar planets which, by chance, have their orbits oriented to produce a transit visible from the Earth. Because of the small orbital radius of α Centauri Bb compared to the size of the star it orbits, there is a 9.5% probability that its orbit is oriented by random chance to produce observable transits of the planet across the face of α Centauri B as viewed from the Earth. Assuming an Earth-like density for α Centauri Bb, such transits would be expected to reduce the apparent brightness of α Centauri B by about 100 parts per million (ppm). While obtaining this level of photometric accuracy is not possible using any ground-based instruments, there are orbiting telescopes that are capable of making the required observations.

A view of the Hubble Space Telescope in orbit above the Earth. (STSci/NASA)

In June 2015 a paper by an international collaboration of scientists (including five of the 11 authors of the original α Centauri Bb discovery paper) with Brice-Olivier Demory (Cavendish Laboratory) as the lead author was published in The Monthly Notices of the Royal Astronomical Society which presented an analysis of photometric measurements of α Centauri B made using the Hubble Space Telescope (HST). To make their measurements of α Centauri B, Demory et al. used the Space Telescope Imaging Spectrograph (STIS) which was installed on HST during its second servicing mission in 1997 and was subsequently repaired in 2009.

STIS made almost continuous observations of α Centauri B for 26 hours from July 7 to 8, 2013 around the time a transit of α Centauri Bb was expected. Detailed analysis of the photometric measurements corrected for various instrumental effects yielded an accuracy of about 115 ppm for individual brightness measurements with greater accuracy possible by analyzing the thousands of data points collectively. After a full analysis, a very promising transit-like event about 3.8 hours long with a depth of about 90 ppm was detected in the data consistent with the transit of a planet with 0.92±0.06 times the radius of the Earth.

Plots showing the transit event observed using HST’s STIS in July 2013 (top panel) and the residuals after the data are fitted to a transit model (bottom panel). The gray dots are individual data points while the back dots are data binned into 45-minute increments. Click on image to enlarge. (Demory et al.)

With this apparently positive result, the team was able to schedule another 13.5 hours of uninterrupted observation time on HST between July 28 to 29, 2014 to reobserve α Centauri B in hopes of spotting another transit event. Employing the same data reduction and analysis procedures used for the 2013 HST data, Demory et al. observed no transit-like events in the newer data set. Given the quality of the data, a 3.8-hour long transit with a depth of about 100 ppm like that seen in 2013 should have been detected to 5σ level or better but none was present. It now seemed unlikely that a transit of α Centauri Bb had been observed in 2013 after all. But given the low probability that α Centauri Bb would produce an observable transit, this null result can not be taken as proof this planet does not exist.

A detailed analysis of the 40 hours of available HST photometric measurements from 2013 and 2014 has ruled out all explanations for the observations save for one: α Centauri B is orbited by another Earth-size planet in addition to α Centauri Bb. The best fit for the limited data hints that this new planet candidate, like α Centauri Bb, is in a tight orbit around its primary with an orbital period probably no greater than about 20 days and likely to be around 12 days (for more details on these observations and analysis, see “Has Another Planet Been Found Orbiting Alpha Centauri B?”).

An artist’s depiction of NASA’s TESS satellite which will systematically search the bright stars of the sky for planetary transits starting in 2017. (NASA)

Unfortunately with such a poorly constrained orbit, three weeks of nearly continuous photometric monitoring of α Centauri B will be required to confirm this hypothesis. HST is too busy to accommodate a dedicated search of this length and no other space telescope currently available is capable of making the needed observations. In addition, since the radial velocity signature for this planet would be expected to be maybe half that of α Centauri Bb, this method has little likelihood of providing independent confirmation of this sighting any time soon. Once again, we will have to wait for a few more years for new telescopes to become available such as NASA’s TESS (Transiting Exoplanet Survey Satellite) mission or ESA’s CHEOPS (Characterizing Exoplanets Satellite) which are both scheduled for launches in 2017 and may be capable of making the required observations of such a bright target.



Three years after the discovery of α Centauri Bb, we still find ourselves waiting for independent confirmation of this important find. And in the process of trying to observe a transit of this world using precision photometry, the waters have only been muddied further with the apparent detection of what might be a second extrasolar planet. The lack of independent confirmation of α Centauri Bb or, indeed, the absence of the publication of any follow up observations by even the HARPS team should not be taken as evidence that this extrasolar planet does not exist. Instead, it is merely a reflection of just how difficult such a detection is using precision radial velocity measurements and how much more difficult it has become due to the current small apparent separation of α Centauri A and B.

For the time being, we might have to resign ourselves to the fact that it could be a few more years before we finally see any new results of radial velocity and photometric measurements to confirm or refute the discovery of α Centauri Bb as well as the sighting of a potential second planet by HST. And if these new measurements continue to produce ambiguous results, it will likely be a couple of more decades before our technology advances enough to directly image this pair of extrasolar planets. By that point, however, we would have already been able to search for Earth-sized and larger planets in much wider orbits including out into the habitable zone. No matter how this situation unfolds, we are at the threshold of learning much more about the planetary systems of our nearest neighbors.



Just a few days after this article was originally posted, a new paper by Vinesh Rajpaul, Suzanne Aigrain and Stephen J. Roberts of the University of Oxford was accepted for publication in The Monthly Notices of the Royal Astronomical Society which cast grave doubts about the existence of α Centauri Bb. They present a convincing case that the radial velocity signature that had been interpreted as being the result of α Centauri Bb is instead an artifact of noise in the data and how it was sampled irregularly over time. Even HARPS team member and staunch defender of the claim that α Centauri Bb exists, Xavier Dumusque, has now publicly conceded that α Centauri Bb probably does not exist. The search for planets in the α Centauri system (and an explanation for the transit-like event observed by Demory et al.) continues.


Follow Drew Ex Machina on Facebook.


Related Reading

“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]

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

“Has Another Planet Been Found Orbiting Alpha Centauri B?”, Drew Ex Machina, March 28, 2015 [Post]

“Habitable Planet Reality Check: GJ 581”, Drew Ex Machina, March 9, 2015 [Post]

“Habitable Planet Reality Check: GJ 667C”, Drew Ex Machina, September 7, 2014 [Post]

“Kapteyn b: Has Another Habitable Planet ‘Disappeared’?”, Drew Ex Machina, May 14, 2015 [Post]


General References

Christoph Bergmann et al., “Searching for Earth-mass planets around α Centauri: precise radial velocities from contaminated spectra”, International Journal of Astrobiology, Vol. 14, No. 2, pp. 173-176, April 2015

Bruce Betts, “Update on the search for planets in the Alpha Centauri system”, The Planetary Society Blogs, April 4, 2014 [Link]

Brice-Olivier Demory et al., “Hubble Space Telescope search for the transit of the Earth-mass exoplanet Alpha Centauri Bb”, Monthly Notices of the Royal Astronomical Society, Vol. 450, No. 2, pp. 2043-2051, June 2015

Xavier Dumusque et al., “An Earth-mass planet orbiting α Centauri B”, Nature, Vol. 491, pp. 207-211, November 8, 2012

Artie P. Hatzes, “Radial Velocity Detection of Earth-Mass Planets in the Presence of Activity Noise: The Case of α Centauri Bb”, The Astrophysical Journal, Vol. 770, No. 2, Article ID 133, June 2013

V. Rajpaul et al., “Ghost in the time series: no planet for Alpha Cen B”, arXiv 1510.05598 (accepted for publication in The Monthly Notices of the Royal Astronomical Society), submitted October 19, 2015 [Preprint]