According to an old adage, if it seems too good to be true, it probably is. For some time now, it has been claimed that a nearby red dwarf star called GJ 667C had a compact planetary system that was packed with possibly as many as seven planets. Moreover, three of these super-Earth size planets were potentially habitable with some claiming that they were among the most Earth-like extrasolar planets known. Being a long time skeptic of overly inflated claims of the potential habitability of extrasolar planets, the situation with GJ 667C just seemed too good to be true. And after a series of papers published over the past several months and the recent submission of another critical paper on the analysis of data relating to GJ 667C, the time has come to examine the claims that it hosts potentially habitable planets.
GJ 667C is the smallest member of a triple star system located about 22 light years away in the constellation of Scorpius. At the heart of this system (also known as Gliese 667), are a pair of K-type dwarf stars, GJ 667A and B, each with a fraction of the luminosity of the Sun locked in an eccentric, 42-year orbit with a mean distance of about 12 AU. Located at a projected distance of 230 AU from GJ 667AB (its true distance is likely to be greater) is GJ 667C. This star is a type M1.5V red dwarf with an estimated mass of 0.33 times that of the Sun, a luminosity of 0.014 times that of the Sun and a surface temperature of about 3350 K. With an age estimated to be in excess of two billion years (the actual age of red dwarfs is difficult to gauge due to their very slow evolution), GJ 667 C is a fairly quiescent red dwarf and typical of the stars in the solar neighborhood.
A European-based group of astronomers, using the HARPS (High Accuracy Radial Velocity Planet Searcher) spectrograph on ESO’s 3.6-meter telescope at the La Silla Observatory in Chile, had been monitoring GJ 667C since 2004 as part of a larger observation campaign looking for variations in its radial velocity that would indicate the presence of planets. On October 19, 2009 they announced their first discovery in this system: a super-Earth designated GJ 667Cb in a close one-week orbit. The discovery of a second planet, GJ 667Cc, was announced by the HARPS team on November 21, 2011 in a preprint of a paper by Bonfils et al. which was formerly published in January 2013. GJ 667Cc has an orbital period of about 28 days and, since the inclination of its orbit to the plane of the sky is unknown, a minimum mass about four times that of the Earth. This discovery was described in greater detail in a subsequent HARPS team paper by Guillem Anglada-Escudé et al. published in 2012 along with the possible detection of a third planet – another super-Earth in a 106-day orbit designated GJ 667Cd whose presence was put on firmer ground in a 2013 paper by Delfosse et al.. In the discovery paper it was prominently noted that GJ 667Cc orbited comfortably inside the habitable zone of its sun and was considered to be potentially habitable by the authors despite the admitted uncertainty in its actual mass.
A veritable planetary jackpot for GJ 667C was announced in June 2013 in another paper by Guillem Anglada-Escudé et al.. For this work, the HARPS team collaborated with members from two other extrasolar planet search teams. They combined radial velocities from a total of 173 spectra obtained with HARPS from June 2004 to March 2012 with 23 measurements from the PFS (Planet Finder Spectrograph) on the 6.5-meter Magellan II Telescope at the Las Campanas Observatory in Chile obtained between June and October 2011 along with 17 radial velocity measurements from HIRES (High Resolution Echelle Spectrometer) at the Keck Observatory in Hawaii obtained as part of the Lick-Carnegie Exoplanet Survey. The collaboration’s analysis of this data set confirmed the presence of GJ 667Cd with a refined orbital period of 92 days along with three more super-Earths, designated e, f and g, with periods of 62, 39 and about 260 days, respectively. There were also indications of a seventh Earth-mass planet, h, with an orbital period of about 17 days but its detection was admittedly marginal at best. Calculations the team performed on the dynamical stability of this system indicated that it would remain stable for at least for one million years if the inclination of the planets’ orbits to the plane of the sky was no less than 30°, assuming the orbits were coplanar. This limits the mass of the planets to be no greater than about twice the minimum mass or mpsini inferred from the radial velocity measurements.
As important as these discoveries were, habitable planet enthusiasts were thrilled to find that two of these new planets, e and f, joined GJ 667Cc as being potentially habitable. In a press release posted on June 25, 2013, the Planetary Habitability Laboratory of the University of Puerto Rico at Arecibo ranked planets c, f and e as the second, fifth and eleventh most Earth-like planets then known, respectively.
While astronomers and habitable extrasolar planet enthusiasts alike were abuzz with the prospect of so many planets in a nearby star system, there were serious doubts among many about the newest discoveries. Valeri Makarov and Ciprian Berghea (US Naval Observatory) performed their own analysis of the radial velocity data as part of a study of the system’s dynamical evolution and the spin-orbit resonance of GJ 667Cc which they submitted for publication in November 2013. Their “brute force” grid-search style of periodogram analysis definitively found planets b and c. They also detected signals with periods of about 91, 53 and 35 days which they interpreted as possibly being the result of three additional planets in eccentric orbits (seemingly corresponding to what the HARPS team had identified as planets d, e and f, respectively). However, when Makarov and Berghea included all these signals as planets in their orbit simulations, they found that the resulting planetary system was unstable and quickly broke up. Only a two-planet system consisting of planets b and c proved to be stable and they concluded that the longer period variations in the radial velocity of GJ 667C must have a nonplanetary cause.
In another paper submitted in November 2013 by Farhan Feroz and Michael Hobson (Cavendish Laboratory in Cambridge, England), an independent analysis of the radial velocity data for GJ 667C was presented that also did not support the earlier conclusions of the HARPS team. They employed a Bayesian technique to analyze the radial velocity data and found that there was a significant correlated “red noise” component in the data that the HARPS team had not properly taken into account with the “white noise” model used in their analysis. Feroz and Hobson could only confirm the presence of signals with a periodicity of 7 and 28 days corresponding to planets b and c. There was a suggestion of a signal with a period of 91 days, apparently corresponding to GJ 667Cd, which they doubted was of planetary origin.
The situation for the “packed” planetary system of GJ 667C took another serious blow with the recent submission of a paper by Paul Robertson and Suvrath Mahadevan (Pennsylvania State University). Just two months earlier, Robertson and Mahadevan along with colleagues Michael Endl (Pennsylvania State University) and Arpita Roy (McDonald Observatory, University of Texas at Austin) had published a paper with a detailed analysis of spectra of the red dwarf GJ 581 which clearly demonstrated that subtle magnetic activity on the star’s surface was affecting precision measurements of its radial velocity. Instead of there being as many six planets orbiting GJ 581 (including two that were believed by many to be potentially habitable), there were in fact only three planets present in the data after they were properly corrected for stellar magnetic activity. The other three planets, including both potentially habitable planets, were artifacts or noise and simply did not exist (see “The Disappearing Habitable Planets of GJ 581“).
As had been done with GJ 581, instead of looking at just radial velocity measurements alone, Robertson and Mahadevan examined the 171 publicly available HARPS spectra of GJ 667C from which the radial velocity measurements were originally derived. While the HARPS team had looked for the signature of star spots and other obvious signs of stellar activity that have been known for decades to generate signals that can mimic those of extrasolar planets, this new analysis was looking for much more subtle variations in magnetic surface activity by examining the Hα and Na I D lines in the spectra of GJ 667C. Robertson and Mahadevan found that there were indeed signs of magnetic activity that produced radial velocity variations that changed with the 105-day rotation period of the red dwarf. After they corrected the radial velocity data for the effects of this magnetic activity, they clearly found signals corresponding to planets b and c. However, the signal with the 92-day period that had been attributed by the HARPS team to GJ 667Cd disappeared indicating that it was not a planet after all but the result of changing surface activity modulated by the rotation of GJ 667C.
Attempts by Robertson and Mahadevan to find evidence for additional planets in the data set were not successful. Usually, after the radial velocity data are corrected for magnetic surface activity, any signals corresponding to bona fide planets become much clearer. Instead, the signals corresponding to the presumed orbital periods of planets e, f and g were marginally significant at best and their strength varied strongly depending on the assumed orbital parameters of planets b and c (especially the assumed orbital eccentricity of GJ 667Cc). Furthermore, attempts to fit the data beyond the two-planet model lead to solutions that were dynamically unstable within only a century.
Robertson and Mahadevan concluded in their paper that they can only confirm the presence of planets b and c and definitively attribute the signal that had been claimed to be GJ 667Cd to magnetic activity on the star. Further, they state in the paper “we cannot claim to have completely ruled out the planets e-g”. However, Paul Robertson is a bit more direct in an online statement in The Habitable Zone Planet Finder web site where he states “the fact that we see no sign of any of these planet candidates after the activity correction leads us to strongly doubt their existence”. As was the case with GJ 581, it seems that magnetic activity was masquerading as planets and that GJ 667C in fact has only two planets whose existence can be demonstrated by the existing set of precision radial velocity measurements. Fortunately for habitable planet enthusiasts, one of these two planets still includes the potentially habitable GJ 667Cc.
The data for the known and suspected planets of GJ 667C, excluding GJ 667Cd whose radial velocity signal has been shown to be the result of stellar activity, are summarized in Table I. The data for planets b and c are from the latest results of the analysis by Robertson and Mahadevan. The data for planets e through h, whose existence are questionable but are included for discussion purposes, are from Anglada-Escudé et al. (2013). The insolation, Seff where Earth equals 1, for all the planets were calculated from these data making allowances for orbital eccentricity as described by Dressing et al..
Table I: Properties of the Planets of GJ 667C
|Orbit Period (days)||7.200||16.9||28.10||39.0||62.2||256|
|Orbit Radius (AU)||0.0504||0.08||0.1250||0.16||0.22||0.55|
Based on the latest models of planetary habitability by Kopparapu et al., the habitable zone for GJ 667C for a 1 Earth-mass (ME) planet runs from 0.12 to 0.24 AU corresponding to the conservative runaway greenhouse and maximum greenhouse limits, respectively. The effective insolation, Seff, ranges from 0.93 to 0.24. For a more massive 5 ME planet, the inner limit of the habitable zone is just 0.004 AU closer with an Seff of 1.00 – identical to Earth today with its hotter sun. The presence of the distant and relatively dim companions, GJ 667A and B, plays no significant role in defining the bounds of the habitable zone of GJ 667C.
Looking at the orbit as derived in the latest work by Robertson and Mahadevan, GJ 667Cc orbits comfortably inside the inner portion of the habitable zone of GJ 667C. Even with its moderately eccentric orbit which ranges from 0.091 to 0.158 AU, GJ 667Cc spends the overwhelming majority of its orbit inside the habitable zone and only briefly breaches its inner limits. Given a moderately dense atmosphere of CO2 (which GJ 667Cc is likely to have as a natural consequence of the carbonate-silicate cycle) and the presence of an ocean, the effects of the resulting factor of three variation in Seff during its short 28-day long “year” would be moderated significantly.
While a planet orbiting this close to a red dwarf would normally be expected to be a synchronous rotator with one face always pointing towards its sun, the case for GJ 667 Cc is a bit more complicated owing to its noticeably eccentric orbit. In such situations, other types of spin-orbit couplings are possible. A perfect example of this is the planet Mercury in our solar system whose rotation is in a 3:2 resonance (i.e. it rotates three times about its axis for every two orbits around the Sun) in part because of the 0.21 eccentricity of its orbit.
Valeri Makarov and Ciprian Berghea (US Naval Observatory), whose work was mentioned in the previous section, examined the question of the spin state of GJ 667Cc in their paper published in January 2014. Assuming an Earth-like composition, Makarov and Berghea found that GJ 667Cc would spin down on a time scale of only one million years. So whatever its spin state was originally and ended up being, its present state was reached long ago. Their investigation shows that there is only a 10% chance that GJ 667Cc is a synchronous rotator. It is nine times more likely that GJ 667Cc is a supersynchronous rotator. With a probability of 51%, Makarov and Berghea find that it is most likely that the spin of GJ 667Cc is locked in a 3:2 resonance with its orbit exactly like Mercury. That would give GJ 667Cc a sidereal day (i.e. the rotation period with respect to the stars) of 18.7 days. But just like Mercury, the length of the “mean solar day” (i.e. the equivalent of the average time between successive sunrises) would be two orbital revolutions long or 56.2 days long, in the case of GJ 667Cc. Since a range of models developed over the past two decades has already shown that habitable conditions can exist on synchronous rotators, this slow rotation would be unlikely to impede the habitability of GJ 667Cc. And if these models prove to be incorrect, this supersynchronous rotation state would be sufficient to prevent atmospheric freeze out.
Another part of the paper by Makarov and Berghea dealt with the evolution of the orbits of GJ 667Cb and c. They found that while the size of the semimajor axis of these planets stays fairly constant, the eccentricity of the orbits of these planets vary cyclically with a period of only about 0.46 years or just 170 days due to the near 4:1 mean resonance between them. In the case of GJ 667Cc, the eccentricity was found to vary between 0.05 and 0.25 (the latter being slightly lower than the 0.27 value derived by Robertson and Mahadevan but well within their estimated ±0.10 uncertainty). Given the apparent dependence on the strength of the signals for planets e, f and g that Robertson and Mahadevan found on their choice of orbital parameters of b and c, I can not help but wonder if the changes in the orbital eccentricity is partly responsible for some or all of the signals the HARP team attributes to planets e through h. This is especially true in light of their analysis using over eight years of radial velocity data unevenly spaced in time where the orbital parameters are normally assumed to remain constant and not vary on time scales of just 0.46 years. But as far as the habitability of GJ 667Cc is concerned, combining the effects of a 56-day long mean solar day and a 28-day orbit whose eccentricity varies cyclically over a time period of almost exactly six orbits would probably not compromise habitability but it would create very interesting variations in surface conditions as a function of longitude as well as latitude on time scales of a few days to a few months.
While the spin state and curious cyclically varying orbit would make for interesting diurnal and seasonal cycles on GJ 667Cc, there is a dark side to it that would hamper its habitability. Makarov and Berghea found that tidal heating resulting from the eccentricity of the orbit of GJ 667Cc would amount to 1023.7 joules per year assuming an Earth-like composition. This rather odd notation translates to about 1.6X1016 watts of heat flow or a factor of about 300 greater than Earth’s current internal heat (i.e. from the decay of radioactive elements as well as left over from its formation). Makarov and Berghea estimate that the temperature of an Earth-like mantle in GJ 667Cc would increase at a rate of 1.6 K per 100,000 years and, barring any limiting mechanism not included in their simple model, the mantle of GJ 667Cc would be rendered completely molten inside of 100 million years. Such a tidal heating rate would seriously affect the potential habitability of GJ 667Cc or even make it impossible.
Another potential obstacle for the habitability of GJ 667Cc is that it might not even be a terrestrial planet. As is the case with all planetary discoveries made using precision radial velocity measurements, only a minimum mass or Mpsini can be derived since the inclination of the planet’s orbit with respect to the plane of the sky, i, is not known from spectral measurements alone. The inclination must be determined by other means. Without a known inclination, only the probability that a planet has a mass in a certain range, such as the mass range consistent with a rocky composition, can be stated.
A recent analysis of Kepler data by Leslie Rogers (California Institute of Technology) has shown that a noticeable change in the composition of planets takes place at around 1.5 times the radius of the Earth (RE). Taking into account the uncertainties in the measured radii and especially the independently measured masses of Kepler objects with radii less than 4 RE selected for her analysis, Rogers found, at a 95% confidence level, that the majority of planets with radii greater than 1.6 RE are likely to possess a substantial volatile envelope rich in water, hydrogen and helium. In other words, such planets are more likely to be mini-Neptunes or gas dwarfs that are extremely unlikely to be habitable and would not be rocky, terrestrial-type planets. Rogers’ conclusions agree well with previous analyses of Kepler data by Marcy et al. as well as Weiss and Marcy published earlier this year. It also supports a theoretical study on the transition from terrestrial to non-rocky planets by Eric Lopez and Jonathan Fortney (University of California – Santa Cruz) that was formally published just a few days ago.
The 1.6 RE radius limit for terrestrial planets found by Rogers translates into a 6 ME upper limit for the mass of terrestrial planets assuming an Earth-like composition. Assuming an unconstrained, random orientation for the orbit of GJ 667Cc with respect to the plane of the sky and a MPsini of 4.1 ME, there is about one chance in four that GJ 667Cc exceeds this threshold. However, this 6 ME upper limit is just the maximum likely value where there is a 50-50 split between rocky and non-rocky planets. The existence of low-mass, low-density extrasolar planets like PH3 c with a mass of 4 ME discovered by Schmitt at al. and the three smallest planets known to orbit Kepler 11 discovered by Lissaeur et al. which have masses in the 2 to 3 ME range (which are all undoubtedly volatile-rich mini-Neptunes) strongly suggests that the transition from rocky to non-rocky planets as a function of mass is a gradual one. As a result, there seems to exist a non-negligible fraction of non-rocky planets below the mass threshold derived by Rogers. With this in mind there is an uncomfortably high probability that GJ 667Cc is a mini-Neptune – certainly higher than the one chance in four that its actual mass exceeds Rogers’ threshold.
The situation with the potential habitability of GJ 667Ce and f, assuming for the moment that they even exist, is better than the situation with GJ 667Cc. With the orbits as derived by Anglada-Escudé et al. (2013), these planets fall comfortably inside the habitable zone of GJ 667C. With their presumably more circular orbits, they would be more likely to become synchronous rotators but, as models have shown, this is probably not a barrier to habitability. Since the tidal heating rate decreases rapidly with increasing orbital radius as well as with decreasing planet radius and orbital eccentricity, this pair of planets would experience tidal heating rates more than three orders of magnitude less than GJ 667Cc amounting to just a fraction of Earth’s internal heat flow. So overheating will not be an issue. If we are to believe the orbit stability analysis results in Anglada-Escudé et al. (2013) which limits the orbit inclination to a minimum value of 30°, the maximum masses for these planets would be less than the approximately 6 ME mass limit for terrestrial planets found by Rogers meaning that it is highly probable that these would be terrestrial planets.
Unfortunately, the planets GJ 667Ce and f as described by the HARPS team and their collaborators in Anglada-Escudé et al. (2013) probably do not exist. The independent analyses of radial velocity data by Makarov and Berghea and by Feroz and Hobson as well as the newest analysis of spectra by Robertson and Mahadevan all point to the radial velocity variations that had been interpreted as being the result of these planets are instead caused by time varying magnetic activity on GJ 667C modulated by its rotation, noise in the data and, I would speculate, the fast evolution of the orbits of GJ 667Cb and c as found by Makarov and Berghea (effects that are not normally included in the analysis of radial velocity data). While this system was found to be dynamically stable in Anglada-Escudé et al. (2013), the stability of such a packed system depends critically on the assumed orbit parameters. Makarov and Berghea as well as Robertson and Mahadevan have independently found that the planetary systems they derive from the data that include more than just GJ 667Cb and c are typically very unstable. Taken together, it seems highly unlikely that GJ 667Ce and f even exist never mind are potentially habitable.
Once again we find ourselves in a situation where the potential habitability of an extrasolar planet has been overstated. While GJ 667Cc does orbit comfortably inside the habitable zone of its sun, it fairly high probability of being a mini-Neptune, gas dwarf or larger volatile-rich planet than it is to be a terrestrial planet owing to its currently uncertain mass. And if GJ 667Cc is a terrestrial planet, the near 4:1 resonance with its massive neighbor, GJ 667Cb, increases the eccentricity of its orbit to the point where the planet would experience excessive tidal heating that would render it sterile. But if GJ 667Cc is a terrestrial planet, if there is some process that limits its tidal heating and if it avoids a list of other potential issues with habitability orbiting a red dwarf, GJ 667Cc would experience an interesting mix of climate variations due to its 28-day long orbital period, likely 56-day long solar day in combination with an orbit whose eccentricity varies markedly over a 170-day cycle.
As for GJ 667Ce and f, their prospects for potential habitability seem better than for GJ 667Cc except for the fact that it now appears likely they do not even exist. The variations in the radial velocity measurements that had been interpreted as being the result of these two planets would appear to be the result of time varying magnetic activity on the surface of GJ 667C and other sources of noise. I would speculate further that the relatively rapid, cyclic changes in the orbits of GJ 667Cb and c, whose existence has been independently verified, also plays a role in generating the spurious signals that had been interpreted as being the result of planets e through h in this system. In addition, studies on the stability of this planetary system show that the inclusion of these two planets most likely renders it unstable. Given the stability issues and possible nonplanetary interpretations of the radial velocity data, it seems unlikely that GJ 667Ce and f exist. Obviously additional data for this star and more refined analysis techniques will be required to definitively explain the observed variations in the radial velocity of GJ 667C.
Follow Drew Ex Machina on Facebook.
“Habitable Planet Reality Check: Terrestrial Planet Size Limit”, Drew Ex Machina, July 24, 2014 [Post]
“The Disappearing Habitable Planets of GJ 581”, Drew Ex Machina, July 7, 2014 [Post]
“The Extremes of Habitability”, SETIQuest, Volume 4, Number 2, pp. 1-8, Second Quarter 1998 [Article]
“The Transition from Super-Earth to Mini-Neptune”, Drew Ex Machina, March 29, 2014 [Post]
“Habitable Planet Reality Check: Kepler 186f”, Drew Ex Machina, April 20, 2014 [Post]
“Habitable Planet Reality Check: Kapteyn b”, Drew Ex Machina, June 6, 2014 [Post]
“GJ 832c: Habitable Super-Earth or Super Venus?”, Drew Ex Machina, June 27, 2014 [Post]
“Abundance of Earth Analogs”, Drew Ex Machina. June 25, 2014 [Post]
“Habitable Planet Reality Check: 55 Cancri f”, Drew Ex Machina, May 7, 2014 [Post]
“The Search for Planets Around Alpha Centauri”, Drew Ex Machina, August 11, 2014 [Post]
Guillem Anglada-Escudé et al., “A Planetary System Around the Nearby M Dwarf GJ 667C with at Least One Super-Earth in Its Habitable Zone”, The Astrophysical Journal Letters, Vol. 751, ID L16, May 20, 2012
Guillem Anglada-Escudé et al., “A dynamically-packed planetary system around GJ 667C with three super-Earths in its habitable zone”, Astronomy and Astrophysics, August 2013
X. Bonfils et al., “The HARPS search for southern extra-solar planets XXXI. The M-dwarf sample”, Astronomy & Astrophysics, Vol. 549, ID A109, January 2013
X. Delfosse et al., “The HARPS search for southern extra-solar planets XXXIII. Super-Earths around the M-dwarf neighbors Gl 433 and Gl 667C”, Astronomy & Astrophysics, Vol. 553, ID A8, May 2013
Courtney D. Dressing et al., “Habitable Climates: The Influence of Eccentricity”, The Astrophysical Journal, Vol. 721, No. 2, pp. 1295-1307, October 1, 2010
F. Feroz and M.P. Hobson, “Bayesian analysis of radial velocity data of GJ667C with correlated noise: evidence for only two planets”, Monthly Notices of the Royal Astronomical Society, Volume 437, Issue 4, p.3540-3549, February 2014
R. K. Kopparapu et al., “Habitable zones around main-sequence stars: new estimates”, The Astrophysical Journal, Vol. 765, No. 2, Article ID. 131, March 10, 2013
Ravi Kumar Kopparapu et al., “Habitable zones around main-sequence stars: dependence on planetary mass”, The Astrophysical Journal Letters, Vol. 787, No. 2, Article ID. L29, June 1, 2014
Jack J. Lissaeur et al., “A closely packed system of low-mass, low-density planets transiting Kepler–11“, Nature, Vol. 470, Issue 7332, pp. 53-58, February 2011
Eric D. Lopez and Jonathan J. Fortney, “Understanding the Mass-Radius Relation for Sub-Neptunes: Radius as a Proxy for Composition”, The Astrophysical Journal, Vol. 792, No. 1, September 1, 2014
Valeri V. Makarov and Ciprian Berghea, “Dynamical evolution and spin-orbit resonances of potentially habitable exoplanet. The Case of GJ 667C”, The Astrophysical Journal, Vol. 780, No. 2, article id. 124, January 2014
Geoffrey W. Marcy et al., “Masses, Radii, and Orbits of Small Kepler Planets: The Transition from Gaseous to Rocky Planets”, The Astrophysical Journal Supplement, Vol. 210, No. 2, Article id. 20, February 2014
Abel Mendez Torres, “A Nearby Star with Three Potentially Habitable Worlds”, Planetary Habitability Laboratory Press Release, June 25, 2013 [Press Release]
Paul Robertson, Suvrath Mahadevan, Michael Endl and Arpita Roy, “Stellar activity masquerading as planets in the habitable zone of the M dwarf Gliese 581”, Science, Vol. 345, No. 6195, pp. 440-444, July 25, 2014
Paul Robertson and Suvrath Mahadevan, “Disentangling Planets and Stellar Activity for Gliese 667C”, The Astrophysical Journal Letters, Vol. 793, Article ID. L24, October 1, 2014 [Preprint]
Paul M. Robertson, “More on stellar activity: an investigation of Gliese 667C”, The Habitable Zone Planet Finder, September 1, 2014 [Link]
Leslie A. Rogers, “Most 1.6 Earth-Radius Planets are not Rocky”, Submitted to The Astrophysical Journal, July 16, 2014 [Preprint]
Schmitt et al., “Planet Hunters VII: Discovery of a New Low-Mass, Low Density Planet (PH3 c) Orbiting Kepler-289 with Mass Measurements of Two Additional Planets (PH3 b and d)”, The Astrophysical Journal, Vol. 795, No. 2, ID 167, November 10, 2014
Lauren M. Weiss and Geoffrey W. Marcy, “The Mass-Radius Relation for 65 Exoplanets Smaller than 4 Earth Radii”, The Astrophysical Journal Letters, Vol. 783, No. 1, Article id. L6, March 2014