Proxima Centauri b One Year Later: The Search for More Exoplanets Continues

It is difficult to believe that it has already been a year but on August 24, 2016, an international team of astronomers led by Guillem Anglada-Escudé (Queen Mary University of London) announced the discovery of Proxima Centauri b – an Earth-size planet orbiting inside of the habitable zone (HZ) of the closest known star to our solar system, Proxima Centauri. The discovery of such an interesting exoplanet orbiting our nearest stellar neighbor sparked a flood of interest not only among scientists but with the general public as well. Recent technological advances which might someday soon make it practical to send small probes to explore the planetary systems of nearby stars with travel times of just a few decades have also helped to enhance the interest in this new find.

Dr. Guillem Anglada-Escudé is shown during the ESO press conference in Munich, Germany during the announcement of the discovery of Proxima Centauri b on August 24, 2016. (ESO/M. Zamani)

While the existence of Proxima Centaur b had been suspected for years based on periodic variations observed in the star’s radial velocity (RV) measurements made by the European HARPS (High Accuracy Radial velocity Planet Search) team working at the European Southern Observatory (ESO), it took an intensive data collection campaign during the first quarter of 2016 called the Pale Red Dot involving observatories around the globe to gather the data required to verify its existence as well as suggest the presence of other exoplanets orbiting our neighbor. Here is the story behind this discovery and the status of the continuing search for exoplanets orbiting Proxima Centauri as part of the new Red Dots campaign.



Proxima Centauri is a V magnitude 11.1 red dwarf located, as its official IAU designation would suggest, in the southern constellation of Centaurus. This star was discovered in 1915 by Scottish astronomer Robert Innes (1861-1933) while he was the director of the Union Observatory located in Johannesburg, South Africa. Using photographs taken in April 1910 and July 1915 with a 25-cm astrographic telescope built by the pioneering English astrophotographer John Franklin-Adams (1843-1912), Innes demonstrated that the dim star shared a common proper motion with the pair of much brighter Sun-like stars in the nearby α Centauri system which has recently had its official IAU designation changed to Rigil Kentaurus. Parallax measurements by Innes and others published in 1917 convinced him that the star was actually slightly closer to us than α Centauri AB leading Innes to name it “Proxima Centauri”.

Scottish astronomer Robert Innes discovered Proxima Centauri in 1915 while working at the Union Observatory in South Africa. (South African Library)

Subsequent work by other astronomers in the decades to follow confirmed that Proxima Centauri was and still is the closest known star to our solar system at a distance of 4.24 light years – just a touch closer than α Centauri AB at a distance of 4.37 light years. Because of its similar distance, shared proper motion and close proximity in the sky, it has been assumed for the last century that Proxima Centauri is gravitationally bound to α Centauri AB about 13,000 AU distant but proving it turned out to be more difficult than expected given Proxima’s distant orbit and the comparatively large uncertainties in its position and velocity. It was not until recently that Pierre Kervella (Unidad Mixta Internacional Franco-Chilena de Astronomía/Observatoire de Paris), Frederic Thévenin (Observatoire de la Côte d’Azur) and Christophe Lovis (Observatoire Astronomique de l’Université de Genève) were able to prove definitively that all three stars were gravitationally bound. Using the latest data on the positions and motions of the stars as well as corrections for how precision radial velocity (RV) measurements are affected by various stellar phenomena, Kervella et al. were able to show that Proxima Centauri is locked in an eccentric orbit which ranges from 4,300 to 13,000 AU with a period of about 550,000 years (see “The Orbit of Proxima Centauri”).

This diagram shows the orbit of Proxima Centauri recently derived by Kervella et al.. Click on image to enlarge. (Kervella et al.)

With its distance established in 1917, it was also realized that Proxima Centauri had the lowest intrinsic brightness of any star then known. Although still dimmer and somewhat smaller stars have since been found, the best determinations of its properties pegs the luminosity of Proxima Centauri at just 0.00156 times that of the Sun with a radius of 0.14 times and a mass of 0.12 times. Despite its dimness, it still has a higher apparent brightness than most known red dwarfs (although a small telescope is required to spot it visually) and, as a result, it has been a target of study by those with an interest in this most numerous class of star.

After examining archived photographic plates, American astronomer Harlow Shapley (1885-1972) discovered in 1951 that Proxima Centauri is a UV Ceti type variable star earning it the designation V645 Centauri. This small star experiences occasional flares that can cause the brightness of this otherwise quiescent red dwarf to increase by several percent or much more for brief periods of time. More recent photometric studies of how the brightness of Proxima Centauri varies over time indicate that it has a rather long period of rotation of about 83 days. This slow rotation and its association with the extensively studied stars α Centauri A and B suggest that Proxima Centauri has an age of around five billion years – roughly the same as the Sun and our solar system.

A view of Proxima Centauri acquired by the Hubble Space Telescope which was used to perform direct imaging and astrometric searches for companions during the 1990s. (ESA/Hubble/NASA)

Because of its small mass and relative closeness, Proxima Centauri has been the target of searches for substellar and planetary companions for decades. During the last quarter of a century, increasingly sensitive detection methods have been employed in this search capable of finding ever smaller companions and ultimately planets. However, by the opening decade of the 21st century, direct imaging and astrometric surveys looking for the subtle reflex motion of an orbiting object had failed to find anything eliminating the possible existence of anything larger than approximately Jupiter-mass (or MJ) objects with orbital periods greater than about 50 days (for a detailed summary of these earlier searches, see “The Search for Planets Around Proxima Centauri”).

Telescope array at Cerro Paranal in Chile. One its telescopes, UT2, has been used to measure the radial velocity of Proxima Centauri in search of planets using the UVES spectrometer. (J.L. Dauvergne & G. Hüdepohl/ESO)

By far the most sensitive searches to date have used precision measurements of the star’s RV to detect an orbiting unseen companion. While published results from RV surveys of Proxima Centauri go back over a decade and a half, among the most accurate RV results up until the last decade were based on the work of Michael Endl (McDonald Observatory) and Martin Kürster (Max Planck Institute) published in 2008. Endl and Kürster had observed Proxima Centauri as part of an ongoing program to survey 40 M-dwarf stars using UVES (Ultraviolet and Visual Echelle Spectrograph) on the 8.2-meter telescope designated UT2 – one of four such telescopes that make up the VLT or Very Large Telescope at ESO’s facility on Cerro Paranal in Chile.

Minimum mass upper limits as a function of period for planets in circular orbits around Proxima Centauri based on work by Endl and Kürster. The shaded area labelled “HZ” displays the approximate location of the conservatively defined habitable zone. Click on image to enlarge. (Endl and Kürster).

Based on an analysis of the 229 spectra they obtained between March 2000 and March 2007, Endl and Kürster found no statistically significant RV signature which would indicate the presence of an orbiting extrasolar planet. When considering the detection limits of their data set, it was found that there were no planets orbiting Proxima Centauri with MPsini mass values of two times that of the Earth (or 2 ME) with an orbital period of 2 days and up to 20 ME for planets with orbital periods of nearly 2,000 days. Since RV measurements alone are incapable of determining the inclination, i, of the exoplanet’s orbit with respect to the plane of the sky, only the minimum mass or MPsini value can be derived with the true mass likely being larger.


The Pale Red Dot Campaign

The UVES observation program described by Endl and Kürster was not the only campaign to make precision RV measurements of Proxima Centauri looking for orbiting exoplanets. As part of their systematic survey of nearby M-dwarf stars conducted between 2003 and 2009, the European-based HARPS team made a handful of precision RV measurements of Proxima Centauri looking for signs of variations in this and other nearby red dwarfs. Any significant changes in the RV of stars in their survey could indicate the presence of extrasolar planets that would then be followed up by a more thorough observation campaign to characterize the system.

ESO’s 3.6m telescope equipped with the HARPS spectrograph was used to acquire the data to find exoplanets orbiting nearby stars. (ESO/H.H.Heyer

With the HARPS spectrograph attached to the ESO’s 3.6-meter telescope in La Silla, Chile, an initial batch of 27 RV measurements were made between May 2004 and February 2009. In 2012, Guillem Anglada-Escudé (then with the Carnegie Institution of Washington but now with Queen Mary University of London) and famed exoplanet hunter, R. Paul Butler (Carnegie Institution of Washington), published the results of their analysis of these data using their new HARPS-TERRA analysis software. When searching for periodicities in the RV measurements of Proxima Centauri which had an RMS variation of 2.05 meters per second, Anglada-Escudé and Butler found what they described as “a very marginal peak at 5.6 days”. In an analysis of all the M-dwarf stars in the HARPS survey published the following year with team leader Xavier Bonfils (Institut de Planétologie et d’Astrophysique de Grenoble) as the first author, it was noted that there was convincing evidence of some sort of periodic signal in the expanded data set of 32 RV measurements. While Bonfils et al. found a velocity dispersion of 2.1 meters per second in their observations with a measurement uncertainty of 1.6 meters per second, no single or multi-planet solution generated a statistically convincing match for the available data.

Plot of the radial velocity data acquired by HARPS and analyzed in 2012 by Anglada-Escudé and Butler using their old analysis method, CCF, and their new one, TERRA. Click on image to enlarge. (Anglada-Escudé and Butler)

With more data needed, two high-cadence data collections were made by HARPS in 2013. The first had total of 143 spectra obtained on three consecutive nights between May 4 and 7 with another 25 spectra acquired by May 16. A second data collection of 23 spectra from December 30, 2013 to January 10, 2014 followed. The data clearly showed a periodicity of about 11.2 days – twice the period originally found by Anglada-Escudé and Butler which, in retrospect, was probably a harmonic of this newly identified signal. Unfortunately, the false alarm probability (FAP) even with the greatly expanded data set was still too high to be considered a reliable detection. In addition, there were alternate periods of 13.6 and 18.3 days possible which were caused by a “nontrivial window function” resulting from the uneven sampling of the data over long periods. Combining these data with the earlier UVES data by Endl and Kürster produced a false alarm probability of about 1% but there were still many issues associated data sampling, stellar activity and the assumptions made about the nature of the noise in the data. What was needed was much more data over an extended period of time.

The upper panel shows the HARPS radial velocity measurements of Proxima Centauri from May 2013. The periodogram in the bottom panel shows a broad peak at 11 days. (Anglada-Escudé )

Over the next two years, Guillem Anglada-Escudé organized a dedicated, high-cadence observation campaign to look for this possible exoplanet (or exoplanets) orbiting Proxima Centauri which would eventually be known as the Pale Red Dot – a play on Carl Sagan’s reference to the habitable Earth as a pale blue dot. But more than just precision RV measurements would be required to confirm the possible signature of an orbiting planet with any confidence. At the time, there was growing concern in the exoplanet science community about the effects of natural stellar noise or “jitter” when identifying the source of low amplitude RV signals. While the readily noticeable effects of starspots rotating into and out of view over time were already appreciated, more subtle forms of chromospheric activity modulated by the rotation of the star might also mimic a planetary signature. The existence of potentially habitable planets orbiting GJ 581, GJ 667C and Kapteyn’s Star has been called into question because of such natural noise (see “Habitable Planet Reality Check: GJ 581”, Habitable Planet Reality Check: GJ 667C” and “Kapteyn b: Has Another Habitable Planet ‘Disappeared’?”). These issues are especially important with red dwarf stars like Proxima Centauri which are prone to bouts of enhanced activity and flaring.

Pale Red Dot was an international search for an Earth-like exoplanet around the closest star to us, Proxima Centauri, which employed HARP on ESO 3.6-meter telescope in La Silla, Chile (shown here) and other southern hemisphere observatories. (ESO/Pale Red Dot)

To address these concerns, Anglada-Escudé and his team enlisted the aid of additional astronomers to monitor Proxima Centauri while they were making their RV measurements using HARPS. The 40-cm ASH2 (Astrograph for the South Hemisphere II) telescope at the San Pedro de Atacama Celestial Explorations Observatory (SPACEOBS) in Chile was used for photometric observations at the narrowband wavelengths corresponding the SII and Hα which are frequently employed by astronomers to monitor stellar activity. The one-meter telescopes of (Las Cumbres Observatory Telescope network) made observations over the same time interval in the classic Johnson B and V bands. One of the BOOTES (Burst Observer and Optical Transient Exploring System) stations was originally suppose to gather photometric data as well but apparently suffered technical difficulties which prevented its participation. Analysis of these photometric data taken during the same time period as the HARPS high-cadence RV measurements would help differentiate the signal of an orbiting planet from stellar activity. With partners in place, a block of observing time was secured for HARPS during the first quarter of 2016.

Telescopes like this one-meter telescope were used during the Pale Red Dot campaign to acquire multi-band photometery of Proxima Centauri to monitor this star’s activity as HARPS took its radial velocity measurements. (LCOGT)

As part of the Pale Red Dot campaign, one spectrum of Proxima Centauri was obtained near the end of almost every night between January 19 and March 31, 2016. Out of 60 scheduled nights, 54 nights yielded a total of 56 spectra (with two nights generating two spectra each). These new measurements from 54 “epochs” were combined with 90 earlier observations from HARPS and 72 measurements obtained between 2000 and 2008 by UVES. The combined data set clearly showed a signal with a period of 11.2 days and a false alarm probability (FAP) of 7X10-7. The somewhat marginal signal observed in the pre-Pale Red Dot data had the same phase as that which was present in the new data with a much superior false alarm probability confirming that the periodic signals observed in both sets of data were in fact the same. Detailed analysis of the near-simultaneous photometric data showed a periodicity of about 80 days which is close to the previously determined 83-day rotation period of Proxima Centauri. No periodicities were noted in the photometric data or other activity indicators near 11.2 days. With stellar activity eliminated as the likely cause of the observed RV variations, the only other explanation was the presence of a planet with an orbital period of 11.2 days.

This plot shows how the radial velocity of Proxima Centauri and how it varied during the Pale Red Dot campaign of the first quarter of 2016. Click on image to enlarge. (ESO/G. Anglada-Escudé)

Anglada-Escudé et al. published their results in the August 25, 2016 issue of the peer-reviewed science journal, Nature. Fitting a Keplerian orbit to their RV data, Anglada-Escudé et al. found a planet with an orbital period of 11.186 +0.001/-0.002 days and a Doppler semiamplitude of 1.38±0.21 meters per second. Combined with the presumed properties of its host star, the new exoplanet, dubbed Proxima Centauri b, has a minimum mass or MPsini of 1.27 +0.19/-0.17 ME and mean orbital radius of 0.0485 +0.0041/-0.0051 AU – a mere 7 million kilometers from its host star and squarely inside most definitions of the habitable zone (HZ) for a star as dim as Proxima Centauri (for a full accounting of the early discussions about the potential habitability of Proxima Cenaturi b, see “Habitable Planet Reality Check: Proxima Centauri b”).

This diagram compares the orbit of Proxima Centauri b with the same region of our Solar System. Also indicated is the position of the habitable zone (HZ) around Proxima Centauri. Click on image to enlarge. (ESO/M. Kornmesser/G. Coleman)

While the world hailed the discovery of Proxima Centauri b, it was noted in the original discovery paper that the residuals in the RV data suggested that maybe there was a second, possibly super-Earth-size planet with an orbital period in the 60 to 500 days range and a Doppler semiamplitude of less than three meters per second. Subsequent analysis suggested a range of 40 to 400-days along with hints of periodic signals with periods less than six days. Based on a recent statistical analysis of Kepler finds by Ballard and Johnson, about half of red dwarf planetary systems contain five or more planets in coplanar orbits so it would be expected that Proxima Centauri could have more than just one planet. Once again, more data were needed to characterize these signals and determine whether they are the result of stellar activity or the presence of additional planets orbiting Proxima Centauri.


The Red Dots Campaign

After the success of the Pale Red Dot campaign, Guillem Anglada-Escudé and his team were able to secure yet another block of valuable observation time on telescopes around the globe for a second, now expanded campaign which started in June 2017 called Red Dots. In addition to observing Proxima Centauri to search for more exoplanets, two more nearby red dwarf stars have been added to the new campaign: Barnard’s Star and Ross 154 (see “Red Dots: The Search for Nearby Earth-Size Exoplanets”). This new project to search for nearby exoplanets nicely complements the partnership ESO announced on January 9, 2017 with the Breakthrough Initiative which aims to demonstrate proof of concept for a new technology that will enable ultra-light unmanned spaceflight at 20% of the speed of light. Such nanocraft could be sent to exoplanets found in ESO searches of nearby star systems such as α Centauri.

Red Dots is the successor of the 2016 Pale Red Dot campaign which is continuing the search for exoplanets orbiting Proxima Centauri as well as the nearby red dwarfs, Barnard’s Star and Ross 154. (ESO/Red Dots)

At the heart of the new Red Dots campaign is a hundred-day observing run using the HARPS spectrograph in La Silla starting on June 21, 2017. The goal is to obtain spectra of the three target stars on a total of 90 nights, weather permitting, which can then be used to derive precision RV measurements. This new data set will be ideal for resolving the orbital period of the second suspected exoplanet orbiting Proxima Centauri as well as search Barnard’s Star and Ross 154 for possible exoplanets. These new data will be analyzed together with archived RV measurements from HARPS and UVES.

The HARPS spectrograph shown during laboratory tests with its vacuum tank open so that some of its high-precision components can be seen. (ESO)

As was done during the 2016 Pale Red Dot campaign, nearly simultaneous multiband photometry has been acquired of the three target stars starting on June 15 by a now expanded global network of modest-size observatories. As part of a public outreach effort, the Red Dots campaign has also enlisted the help of the American Association of Variable Star Observers (AAVSO) so that amateur astronomers can participate in the photometric monitoring of these target stars (for those interested in participating, see this AAVSO Alert Notice). These additional photometric observations will help fill the inevitable gaps in coverage from the professional observatories involved in Red Dots especially for the young and highly active Ross 154 with its short three-day period of rotation.

For the new Red Dots campaign, multiband photometric data for Proxima Centauri will be acquired by two observatories which had originally participated in the Pale Red Dot Campaign. The 40-cm ASH2 telescope at SPACEOBS in Chile is making measurements in the V and R Johnson bands. In addition,’s one-meter and 40-cm telescopes located in southern hemisphere sites like Australia (Siding Spring Observatory), South Africa (Sutherland) and Chile (Cerro Tololo Observatory) are taking data in the standard B and V Johnson bands as well as the rˈ and iˈ Sloan bands employed by the Sloan Digital Sky Survey (SDSS). Photometric observations in support of Red Dots are expected to continue through October 5. Combined with various activity indicators derived from the HARPS spectra used to determine the RV, these new photometric data will help scientists differentiate between stellar activity and orbiting exoplanets as they attempt to identify the causes of any observed periodic variations in RV.

View of SPACEOBS in Chile which includes the 40-cm ASH II telescope which is acquiring photometric observations of Proxima Centauri for the Red Dots campaign. (Red Dots)

Unlike the Pale Red Dot campaign whose data were not publicly revealed until the final analysis was published following peer review, the Red Dots campaign is being run as an open notebook science experiment – the practice of making the entire primary record of a research project publicly available online as it is recorded. The public will have the opportunity to read updates of what is being observed and discussion of the results on the Red Dots website. According to the ESO announcement, however, “any observations presented during this time will of course be preliminary only and they must not be used or cited in refereed literature. The team will not produce conclusive statements, nor claim any finding until a suitable paper is written, peer-reviewed and accepted for publication.”

Based on the public releases of data and associated commentary up to about the halfway point of the current Red Dots campaign, progress is definitely being made collecting new RV measurements of Proxima Centauri despite a slow start. Bad weather at La Silla during June meant that data were acquired only during 10 out of the first block of 22 epochs. Subsequent improvements in the weather meant that by August 10, the number of observations had grown to 33 out of a possible 50 epochs. Hopefully the weather will continue to cooperate through the end of September so that about another 40 RV measurements can be made before the end of the current campaign.

Here are the RV data of Proxima Centauri acquired for the Red Dots campaign up to August 10, 2017. Also plotted are the RV signals for Proxima Centauri b with a period of 11 days and a long-term signal with a period of 185 days which might correspond to a second exoplanet in a more distant orbit. Click on image to enlarge. (Red Dots)

While the 11-day signal corresponding to Proxima Centauri b is difficult to detect in the new data alone (which was not unexpected given the comparatively small number of data points), its signal is becoming even more pronounced when the new data are combined with the earlier UVES and HARPS data sets. Hopefully this will allow the properties of Proxima Centauri b to be refined with noticeably lowered uncertainties for key parameters like its MPsini value and orbital eccentricity.

The effort to confirm the existence of a second exoplanet in a longer period orbit (which would be designated “Proxima Centauri c” if it is confirmed) has been making noticeable progress as well. After the effects of Proxima Centauri b are removed, analysis of the RV data definitely indicate the presence of another periodic signal with a low false alarm probability (FAP). But at this stage there is no clearly unique solution with periods of 91, 111, 193 or 210 days being possibilities – all significantly longer than the 11-day orbital period of Proxima Centauri b suggesting a large orbital radius for this new exoplanet (if indeed it is an exoplanet) well beyond the “snow line” where water would freeze solid as rock. It is hoped that the new data yet to be gathered during the current Red Dots campaign will allow for a robust solution with one of these periods. While the RV data from last year’s Pale Red Dot and this year’s Red Dots campaigns clearly show some sort of long term variation, other data will be required to determine if this is the result of an orbiting exoplanet or due to some sort of subtle form of stellar activity modulated by the rotation of Proxima Centauri and sampled irregularly in time.

This periodogram shows the four most likely periods of the RV variations remaining after the effects of Proxima Centauri b are removed. Click on image to enlarge. (Red Dots)

After the effects of Proxima Centauri b and the longer term variations in the RV are taken into account, there is no indication in the existing data for the presence of any other RV variations with periods greater than two days including the previously suspected signals with periods of less than six days. This does not mean that there are no other exoplanets orbiting Proxima Centauri. It simply means that any additional exoplanets that might be present are probably beyond the detection capability of the current instruments. In the mean time, we still have just over a month left in the current Red Dots campaign. With luck, the discovery of another exoplanet orbiting Proxima Centauri might be expected to be officially announced with the release of a peer-reviewed paper of the formal analysis results of Red Dots probably shortly after the beginning of 2018. Stay tuned!


Follow Drew Ex Machina on Facebook.


Related Reading

For more details on Proxima Centauri, check out this site’s Proxima Centauri page.


General References

Guillem Anglada-Escudé and R. Paul Butler, “The HARPS-TERRA Project. I: Description of the Algorithms, Performance and New Measurements on a Few Remarkable Stars Observed by HARPS”, The Astrophysical Journal Supplement, Vol. 200, No. 15, June 2012

Guillem Anglada-Escudé et al., “A terrestrial planet candidate in a temperate orbit around Proxima Centauri”, Nature, Vol. 536, No. 7617, pp. 437-440, 25 August, 2016

Guillem Anglada-Escudé, “Proxima HARPS data release #1”, Red Dots website, July 14, 2017 [Post]

Guillem Anglada-Escudé, “Proxima HARPS data release #2 and more”, Red Dots website, August 11, 2017 [Post]

Sarah Ballard and John Asher Johnson, “The Kepler Dichotomy among the M Dwarfs: Half of Systems Contain Five or More Coplanar Planets”, The Astrophysical Journal, Vol. 816, No. 2, Article id. 66, January 8, 2016

X. Bonfils et al., “The HARPS search for southern extra-solar planets XXXI. The M-dwarf sample”, Astronomy & Astrophysics, Vol. 549, ID A8, January 2013

M. Endl and M. Kürster, “Toward detection of terrestrial planets in the habitable zone of our closest neighbor: Proxima Centauri”, Astronomy and Astrophysics, Vol. 488, No. 3, pp.1149-1153, September 2008

P. Kervella, F. Thévenin and C. Lovis, “Proxima’s orbit around Alpha Centauri”, Astronomy & Astrophysics, Vol. 598, ID L7, February 2017

Cristina Rodriguez, “Photometry log #1: firsts observations”, Red Dots website, July 31, 2017 [Post]

“VLT to Search for Planets in Alpha Centauri System – ESO Signs Agreement with Breakthrough Initiatives”, Organization Release ESO 1702, January 9, 2017 [Release]

“Photometry requested for Red Dots campaign”, AAVSO Alert Notice 583, June 16, 2017 [Notice]

“Red Dots: The Live Search for Terrestrial Planets around Proxima Centauri Continues”, ESO Announcement 17036, June 19, 2017 [Announcement]

Red Dots web site [Web Site]