In the first part of “The Search for Planets Around Alpha Centauri”, general information about the α Centauri system (also known by the ancient name, Rigil Kentaurus) was presented and the claims surrounding the discovery of the planet α Centauri Bb were examined. In this second part, the searches for planets orbiting α Centauri A and B are examined. While no additional planets have been discovered so far, almost as important is what has not been found thus placing upper limits on what remains to be discovered orbiting our Sun-like neighbors.
Radial Velocity Surveys
Being bright nearby stars, α Centauri A and B are prime candidates for precision radial velocity surveys in search of extrasolar planets. Since they are relatively mature main sequence stars, they are also “quiet” in terms of various forms of surface activity that can generate unwanted noise or jitter in precision radial velocity measurements. Such noise can mask the radial velocity signature of an orbiting planet or mimic that signature. In addition, α Centauri A and B have relatively high metallicities (i.e. concentrations of elements heavier than helium which astronomers dub “metals”) which suggests that the material needed to create planets was available when these stars formed. Since dynamical studies of the α Centauri system typically show that planets with stable orbits are possible out to as far as about 3 AU from either component (give or take an AU depending on the inclination of those orbits), a search has the prospects of actually finding something assuming planets formed in this system.
Between the two stars, α Centauri B is an especially good target for searching for planets in the habitable zone using precision radial velocity measurements. It is less massive than α Centauri A resulting in a greater variation in radial velocity for a planet of a given size. In addition, its habitable zone is closer in to the star which with a smaller orbit also causes a greater variation in radial velocity. Studies have also shown that α Centauri B is exceptionally quiet so it produces less noise or jitter than its sister. As a result, planets orbiting α Centauri B are more readily detectable than those orbiting A.
While preliminary results with slightly lower precision had been presented at a conference two years earlier, the first results of a radial velocity survey of α Centauri A and B published in a peer-reviewed paper were by Endl et al. in 2001. This small international collaboration of astronomers used the Coude Echelle Spectrometer (CES) at the European Southern Observatory (ESO) in La Silla, Chile to make differential radial velocity measurements of α Centauri starting in 1992. From November 1992 to April 1998 the team acquired 205 individual radial velocity measurements of α Centauri A taken during 48 nights and 291 measurements during 43 nights of α Centauri B. With a mean measurement error of 12.3 and 9.9 meters per second for α Centauri A and B, respectively, the team found no evidence for any planets.
While no planets were found, this early search did set upper limits on what size planets could orbit α Centauri A or B and still evade detection. Since the inclination of any planet’s orbit to our line of sight, i, is not known, radial velocity measurements can only provide the minimum mass or MPsini of a planet. With this in mind, Endl et al. were able to exclude the presence of any planets with MPsini values greater than 1 and 1.5 times the mass of Jupiter, MJ, orbiting within 2 AU of α Centauri A and B, respectively. If the orbits of any planets around α Centauri A and B were coplanar with the orbit of the two stars (which provides the most stable configuration for prograde planetary orbits), the orbit inclination would be 79.23°. Assuming for the moment that is the case, these survey results exclude the possibility of “hot Jupiters” with masses greater than about 0.2 MJ within 0.06 AU of either star. By folding in the results of studies on the dynamical stability of planetary orbits as a function of inclination, Endl et al. found that planets with masses greater than 3.5 and 2.5 MJ could not exist around α Centauri A and B, respectively, no matter their inclination.
In 2011 Wittenmyer et al. published the results of an analysis of radial velocity data acquired by the Anglo-Australian Planet Search (AAPS) as part of a campaign to find low-mass extrasolar planets in short-period orbits around selected nearby stars. Included in this search were α Centauri A and B whose radial velocities were measured to an accuracy of 4.06 and 3.58 meters per second in a total of 99 and 134 measurements, respectively. The findings of their more sensitive survey indicate that there are no planets with an MPsini greater than roughly Saturn’s mass within about 2 AU of α Centauri A or B. It now seemed very unlikely that either α Centauri A or B hosts any gas giants like Saturn or Jupiter but that still left the possibility that planets larger than Neptune were in this system as well as smaller Earth-size planets.
Potential Future Results
In order to push these detection limits significantly lower, a better instrument on a larger telescope was required. In October 2003, HARPS (High Accuracy Radial Velocity Planetary Searcher) was activated on ESO’s 3.6-meter telescope. Built and operated by a Geneva-based group of astronomers, HARPS is capable of making radial velocity measurements with an accuracy on the order of 1 meter per second. In theory, such an instrument would be able to detect super-Earth size planets in the habitable zones of either α Centauri A and B if sufficient data over a long enough time period were available. Using measurements obtained between February 2008 and July 2011, Dumusque et al. announced the discovery of α Centauri Bb on October 16, 2012. This planet was found to have an orbital radius of 0.04 AU and a minimum mass 1.1 times that of the Earth (ME) – well below the upper limits set by earlier published surveys (for a full discussion of this discovery, see “The Search for Planets Around Alpha Centauri”). This discovery has yet to be confirmed and doubts linger about the technique used to process the data as well as the residual effects of low levels of surface activity would have on the data even after it has been corrected.
ESO’s 3.5-meter telescope used by the HARPS team searching for extrasolar planets. (ESO)
These concerns not withstanding, Dumusque et al. believe that their purported detection of the 0.5 meter per second signal of α Centauri Bb means that they could eventually detect a planet with an MPsini of as small as 4 ME in a 200-day orbit inside the habitable zone of α Centauri B. If the orbit of such a planet were coplanar with the orbit of α Centauri AB, a planet of this mass would be near the probable upper mass limit for a rocky planet currently estimated to be about 6 ME (see “Habitable Planet Reality Check: Terrestrial Planet Size Limit”). At this time, Dumusque et al. have not stated that they have failed to detect such a planet. Instead the discovery paper states that the detection of a planet this small is possible using HARPS provided that the investment in the required observing time can be made. Even though additional analysis not included in the α Centauri Bb discovery paper will be required to make any definitive statements, it seems to be a safe bet that the existence of Saturn-size planets orbiting α Centauri B with almost any inclination has been safely excluded.
While the observing lists of various teams around the globe searching for planets are frequently fairly closely guarded secrets, the members of the HARPS team are not the only group currently known to be searching for planets orbiting α Centauri A and B. A team of astronomers based at Yale University working with the 1.5-meter telescope at Cerro Tolo Inter-American Observatory (CTIO) in Chile has been searching for planets orbiting α Centauri A and B since 2007 in part with the support of the Planetary Society. After having reduced the radial velocity uncertainty down to 5 meters per second using the refurbished Blanco Echelle spectrometer that had been retired earlier from general use at CTIO, a new purpose-built instrument called CHIRON (CTIO Higher Resolution spectrometer) was commissioned in March 2011. With a measurement precision on the order of 1 meter per second, this new system is expected to be capable of detecting planets as small as super-Earths orbiting in the habitable zone with minimum masses in the 2 to 4 ME range. The team has yet to formally publish any analyses of their results but statements from the project’s Principle Investigator, Debra Fischer (Yale University), reported on The Planetary Society’s web site indicate that they have so far failed to confirm the discovery of α Centauri Bb by Dumusque et al., although their detection would have been marginal at best with the data they have in hand.
The other group currently known to be actively observing α Centauri A and B 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. While their more modest instrument is currently capable of making radial velocity measurements with only an accuracy of about 3 meters per second, they have taken a different approach to detecting low-mass planets orbiting α Centauri. Instead of substantially improving the accuracy of a single measurement, they rely on averaging a large number of measurements taken at a high cadence during each night’s observation run. So long as the noise in their data is “white noise” (i.e. totally random noise that shows no variation with frequency), the accuracy of their final averaged measurements will be better than from an individual measurement.
A view of the one-meter McLellan Telescope on Mt. John in New Zealand. (MJUO)
As of January 2014, the New Zealand-based team have acquired over 26,000 spectra of α Centauri A and 19,000 spectra of α Centauri B since they started their observing campaign in 2007. Based on simulations performed by Endl et al. made available earlier this year in a preprint, HERCULES should be capable of confirming the presence of α Centauri Bb. In addition, their approach with the existing instrument should also be able to detect super-Earth size planets in the habitable zone. The one draw back to this team’s approach is the assumption that the uncertainty in their measurements is due primarily to random measurement errors. Any systematic errors in their measurements would not be eliminated by their approach. Only time will tell if this team will be successful in their search for planets orbiting either component of α Centauri.
While these searches for planets orbiting α Centauri and other nearby stars continues, efforts are being made to improve the precision of radial velocity measurements beyond the current 1 meter per second limit in hopes of detecting still smaller planets. Currently, ESPRESSO (Echelle SPectrograph for Rocky Exoplanet and Stable Spectroscopic Observations) is under construction. When used with ESO’s 8.2-meter Very Large Telescope (VLT), ESPRESSO should be capable of measuring radial velocity to an accuracy of 10 centimeters per second. There is also the Harvard-Smithsonian Astrophysical Observatory’s G-CLEF (GMT – CfA Large Earth Finder) to be used with the Giant Magellan Telescope (GMT) being built at the Las Campana Observatory in Chile. With an effective collecting area of 22-meter telescope, G-CLEF also has the goal of measuring radial velocity to an accuracy of 10 centimeters per second.
As impressive as these instruments may be, we may be approaching the practical limits of the radial velocity technique for detecting extrasolar planets. Twice in recent months, with the nearby red dwarfs GJ 581 and GJ 667C, small variations in radial velocity that had been interpreted as being the result of super-Earth size planets have instead been found to be from the previously unrecognized effects of subtle surface activity modulated by the rotation of these stars (see “The Disappearing Habitable Planet of GJ 581” and “Habitable Planet Reality Check: GJ 667C”). Since such “jitter” is not random (i.e. it is not white noise with a flat spectrum), simple averaging of multiple measurements will not average out the noise in the data. In the end, the practical limits of the radial velocity detection technique will depend on how accurately surface activity can be characterized, modeled and its effects removed from the raw radial velocity measurements. Assuming ESPRESSO and G-CLEF can meet their measurement goals, it might be a while before the needed data processing techniques are developed so that Earth-size planets can be detected orbiting α Centauri… if ever.
Direct Imaging
While astronomers may be approaching the practical limits of the radial velocity technique for finding small extrasolar planets orbiting α Centauri or other stars, there is progress being made in another area: direct imaging. As is the case with other planet search techniques, the proximity of α Centauri makes it an ideal candidate for direct imaging of extrasolar planets.
One of the most sensitive direct-imaging searches for small companions in the α Centauri system to date was performed by French astronomers Pierre Kervella (Observatoire de Paris-Meudon) and Frederic Thevenin (Observatoire de la Côte d’Azur). They used the SUSI2 (Super Seeing Imager 2) camera on ESO’s 3.6-meter NTT (New Technology Telescope) to obtain sequences of images in 2004 and 2006 in the V, R, I and Z bands (corresponding to wavelengths of 551, 658, 806 and 900 nm, respectively). They were looking for dim objects a few tens of arc seconds or more away from and out of the glare of α Centauri A and B. Such a distant object would not orbit either star alone in what is called an “S-type” orbit. Instead it would be orbiting both stars around the barycenter of the α Centauri AB system in what is called a “P-type” orbit. Kervella and Thevenin found no faint comoving companions of α Centauri AB in their search eliminating the possibility of objects with a mass greater than 15 MJ orbiting α Centauri AB at distances of 100 to 300 AU as well as objects larger than 30 MJ between 50 and 100 AU. Earlier dynamical studies have shown that objects much closer to α Centauri AB would not have a stable orbit. This search thus precludes the existence of small to moderate-mass brown dwarfs in P-type orbits around α Centauri AB.
A false color image made from the VRI-filtered SUSI2 images of the region around α Centauri AB acquired in 2004. The glow in the center of the image is the point spread function of the bright stars with the scattered dots being dim background stars. The scale bar in the lower left is one arc minute. (Kervella & Therenin)
In order to image planets closer to α Centauri A and B, more advanced imaging methods are required. Males et al. recently submitted for publication a study on the ability to use adaptive optics (AO) on Earth-based telescopes to image planets in the habitable zones of nearby stars. And to illustrate the potential power of modern AO techniques, they used actual images of α Centauri A obtained in April 2014. For their exercise, they used the 6.5-meter Magellan AO telescope equipped with the VisAO camera operating at a wavelength range of 0.4 to 1.0 μm and the Clio2 infrared camera operating at 1 to 5 μm. Males et al. used the images they obtained of α Centauri A to perform simulations that showed that they could detect planets with contrast ratios (i.e. the ratio between the brightness of the planet and the star it orbits) in the 10-6 to 10-7 range inside the habitable zone of α Centauri A which ranges from about 0.55 to 1.55 arc seconds from the star as seen from Earth. They believe that with improvements to their AO system and imaging techniques, they should be able to detect planets with contrast ratios of about 10-8.
Unfortunately, an Earth-like planet in the habitable zone of a Sun-like star such as α Centauri A would have a contrast ratio on the order of about 2X10-10 which is far beyond the capability of any ground based system to detect. In fact, even a Jupiter-like planet would not be detectable orbiting near the habitable zone of α Centauri A with a system limited to a contrast ratio in the 10-6 to 10-7 range. However, with the hoped-for detection threshold of 10-8, even a Neptune-size planet would be expected to be detectable near the habitable zone of α Centauri A. The detection of smaller planets will require the development of a space-based system such as the proposed 30-centimeter class ACESat (Alpha Centauri Exoplanet Satellite) telescope.
Conclusions
Aside from the unconfirmed discovery of α Centauri Bb, no other planets have been found orbiting either α Centauri A or B so far. Radial velocity survey results published to date seem to exclude the possibility of Saturn to Jupiter-size gas giants orbiting either α Centauri A or B inside about 2 AU (corresponding roughly to the outer limit of a stable orbit around either star). Direct imaging results also seem to exclude objects larger than small to moderate-size brown dwarfs orbiting around α Centauri AB from 300 AU to about 50 AU (corresponding roughly to the inner limit of a stable P-type orbit around α Centauri AB). Despite the lack of additional detections, the published results to date do not exclude a wide range of possible planets including Earth-size planets orbiting inside the habitable zones of α Centauri A and B.
Numerous radial velocity surveys are currently being conducted that should be capable of detecting super-Earth size planets with minimum masses in the 2 to 4 ME range in the habitable zone which, depending on the orbit inclination to our line of sight, would approach the largest possible size of terrestrial planets. While future improvements in the accuracy of radial velocity measurements using instruments currently under construction could theoretically detect smaller planets still, issues associated with the natural noise or jitter in the radial velocity measurements caused by surface activity on α Centauri A and B as well as the ability of scientists to model and remove their effects may make such detections difficult or impossible.
Direct imaging of planets in or near the habitable zones of α Centauri A and B are more promising in the long run. While current systems seem to be incapable of detecting planets of any size by reflected light alone in or near the habitable zone at this time, only modest improvements to AO systems on large ground-based telescopes would allow for the detection of Neptune-size planets. Unfortunately, the direct imaging of potentially habitable Earth-size planets orbiting even nearby stars like α Centauri A and B will have to wait for the introduction of a new generation of space-based instruments designed specifically for the task.
The next post in this series, “The Search for Planets Around Proxima Centauri“, examines the searches for planets orbiting the small and distant neighbor of α Centauri A and B, Proxima Centauri.
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 Proxima Centauri”, Drew Ex Machina, February 23, 2015 [Post]
“The Discovery of Alpha Centauri Bb: Three Years Later”, Drew Ex Machina, October 16, 2015 [Post]
“Publication Watch: The Stability of Planets in the Alpha Centauri System”, SETIQuest, Vol. 3, No. 3, p. 16, Third Quarter 1997 [Article]
“Publication Watch: Habitable Planet Formation in Binary Star Systems”, SETIQuest, Vol. 4, No. 1, p. 21, Second Quarter 1998 [Article]
“Habitable Planet Reality Check: Terrestrial Planet Size Limit”, Drew Ex Machina, July 24, 2014 [Post]
General References
Bruce Betts, “Update on the search for planets in the Alpha Centauri system”, The Planetary Society Blogs, April 4, 2014 [Link]
Xavier Dumusque et al., “An Earth-mass planet orbiting α Centauri B”, Nature, Vol. 491, pp. 207-211, November 8, 2012
M. Endl et al., “The planet search program at the ESO Coude Echelle spectrometer II. The α Centauri system: Limits for planetary companions”, Astronomy and Astrophysics, Vol. 374, pp. 675-681, August 2001
Michael Endl et al., “The Mt. John University Observatory Search For Earth-mass Planets In The Habitable Zone Of Alpha Centauri”, arXiv: 1403.4809 (accepted for publication in the International Journal of Astrobiology), March 19, 2014 [Preprint]
P. Kervella and F. Thevenin, “Deep imaging survey of the environment of α Centauri. II. CCD imaging with the NTT-SUSI2 camera”, Astronomy and Astrophysics, Vol. 464, No. 1, pp.373-375, March II 2007
Jared R. Males et al., “Direct imaging of exoplanets in the habitable zone with adaptive optics”, arXiv: 1407.5099 (to appear in Proc. SPIE 9148), July 18, 2014 [Preprint]
Andrei Tokovinin et al., “CHIRON – A Fiber Fed Spectrometer for Precise Radial Velocities”, Publications of the Astronomical Society of the Pacific, Vol. 125, No. 933, pp.1336-1347, November 2013
Robert A. Wittenmyer et al., “The Frequency of Low-Mass Exoplanets. III. Toward η⊕ at Short Periods”, The Astrophysical Journal, Vol. 738, ID 81, September 1, 2011
I guess astrometry is also going to be extremely challenging (have any planet discoveries made by astrometry withstood follow-up?), though at least it has the advantage of being able to determine the inclination. Apparently recent results (arXiv:1409.3593) indicate that Alpha Centauri B has a substantial obliquity with respect to the binary orbit, which would make me rather wary of assuming that the planetary orbits would be in the binary plane.
To the best of my knowledge, there has never been an extrasolar planet detected by astrometry. While there were alleged discoveries of extrasolar planets by astrometry back in the 1960s and 70s by Peter van de Kamp, they were all subsequently refuted (a subject for future post). All I ever recall seeing in recent decades has been upper limits set on planet masses detected by RV measurements set by astrometry primarily using Hipparcos data (although in the past year or so I recall seeing a paper questioning the ability of Hipparcos data to be reliably used in such a way). The just-started Gaia mission will likely be the first to detect and possibly discover extrasolar planets by astrometery although, unfortunately, Alpha Centauri A and B are too bright to be observed by this instrument.
As for the inclination of the rotational axis of Alpha Centauri B, I am not surprised that is does not seem to be aligned with its orbit around Alpha Centauri A. Recent work has suggested that such an alignment is not expected for wide binaries like Alpha Centauri. As for the inclination of the orbit of Alpha Centauri Bb (assuming it exists), recent observation suggest that a migrating planet like this (especially one in a binary system) will not necessarily line up with the equator of its primary. Given all the unknowns and open questions, only future observations will tell what exactly is going on in this system and if confirmed planets exist.
As regards astrometry, there were some recent claims for VB 10 that got fairly swiftly demolished. The only astrometric candidates I’m aware of that haven’t been are HD 176051 (Muterspaugh et al. 2010) and a claim for Luhman 16 (Boffin et al. 2014), though as far as I can tell neither has been followed up to confirm the detections — the Boffin et al. paper has been cited for its parallax determination rather than the claim of a third object in the system.