Habitable Planet Reality Check: Tau Ceti

I doubt that there are many space enthusiasts or even scientists out there who remember the first time they ever heard mention of the name Tau Ceti (also written τ Ceti). One of the nearest Sun-like stars known, τ Ceti has crept into popular cultural consciousness starting with science fiction stories dating back at least two thirds of a century. My first exposure to the name was probably during the early 1970s while watching syndicated reruns of the classic Star Trek episode, “Whom Gods Destroy”, or possibly any number of scifi movies that graced weekend and late-night broadcast television of the era. I would dare say most people would have similar experiences about their introduction to this star.

The interest in τ Ceti in science fiction is a reflection of the keen interest in this star among scientists. For over a century, τ Ceti has been recognized to be a nearby Sun-like star of some importance and in 1960 Frank Drake made it, along with the nearby ε Eridani, the target of the pioneering SETI experiment called Project Ozma. It was during the mid-1970s that I started reading about the scientific significance of this well known star in the growing number of popular books on the science behind the search for life beyond the Earth by various authors including Carl Sagan.

Interest in τ Ceti has only increased in recent decades as better instruments and observing techniques have become available allowing astronomers to mount serious searches for exoplanets orbiting our neighbor. On August 8, 2017, an international collaboration consisting of selected team members from several of the world’s ongoing exoplanet search programs announced the results of the latest such search. In a paper accepted for publication in The Astrophysical Journal with Fabo Feng (University of Hertfordshire, UK) as the lead author, the group described how they found evidence for four exoplanets orbiting τ Ceti including two which appear to straddle the habitable zone (HZ) of this Sun-like star – apparently the same two exoplanets originally announced by the same team in December 2012 based on an earlier analysis of a smaller set of radial velocity (RV) data. So what is the status of these new finds and can any of them really be considered potentially habitable?



With a V-magnitude of 3.50, the naked-eye star τ Ceti is located south of the celestial equator in the constellation of Cetus – the whale or, more traditionally, the “sea monster”. Also known by various other catalog designations (e.g. HD 10700 and GL 71), the common name τ Ceti was established in 1603 by the German celestial cartographer, Johann Bayer (1572-1625), in his famous star catalog called Uranometria. With a spectral type of G8.5V, τ Ceti is the closest single G-type star to us at a distance of 11.90 light years (α Centauri A is closer but it is part of a multiple star system – see the Alpha Centauri page).

An image of the nighttime sky showing the location of τ Ceti. (University of Hertfordshire)

The best current measurements of the properties of τ Ceti give it an effective surface temperature of 5,344 K, a radius 0.79 times that of the Sun, a luminosity of 0.52 times and a mass estimated to be 0.78 times. Based its 34-day period of rotation and its low surface activity, τ Ceti is calculated to be about 5.8 billion years old although other work suggests it could be much older still. All in all, τ Ceti seems to be a slightly cooler, smaller, dimmer and older version of the Sun. The biggest difference between the two stars is that τ Ceti has only about 28% of the Sun’s concentration of “metals” – what astronomers collectively label all elements heavier than helium.

Comparison of the Sun (left) and τ Ceti (right). (R.J. Hall)

Naturally, τ Ceti has been a target of a number of surveys looking for extrasolar planets. By the opening decade of the 21st century, direct imaging as well as indirect methods such as astrometry and precision radial velocity (RV) measurements looking for the reflex motion of an orbiting body had effectively eliminated the possibility of any objects about the mass of Jupiter or larger orbiting τ Ceti over a wide range of orbital distances. Smaller objects including Earth-size exoplanets orbiting in the habitable zone (HZ) were still tantalizing possibilities. A debris disk, first detected by infrared observations in 2004 (and subsequently observed with increasing resolution and sensitivity in recent years) which presumably is caused by a belt of comet or asteroid-like objects which extends from about 6 to 55 AU, strongly suggests that τ Ceti supports a family of exoplanets which has cleared the interior of this disk. But more sensitive detection methods would be needed to spot them.

As the accuracy of RV measurement techniques continued to improve during the open decade of this century, it was becoming clear that various sources of “noise” were limiting astronomers’ ability to detect planets approaching the size of the Earth orbiting Sun-like stars like τ Ceti. This noise not only included that resulting from instrumental effects but also naturally occurring “jitter” resulting from normal surface activity of the star itself. In an effort to develop better data analysis tools which deal with various sources of noise more effectively to reveal small amplitude RV variations caused by orbiting exoplanets, a team of astronomers involved with several of the world’s exoplanet survey programs collaborated using τ Ceti as a test case for a paper formally published in the March 2013 issue of Astronomy & Astrophysics with Mikko Tuomi (University of Hertfordshire, UK) as the lead author. τ Ceti was chosen for this exercise because the star’s low activity and its RV measurements obtained by various groups over many years appeared “flat”. With no obvious signs of periodic signals, τ Ceti’s data set was excellent raw material for the search for low amplitude RV variations.

The ESO 3.6m Telescope equipped with the HARPS spectrograph was used to acquire the data used to find the new planet candidates orbiting τ Ceti. (ESO/H.H.Heyer)

For this exercise, RV measurements from three different surveys were pooled together for analysis. The largest collection of data were from the European Southern Observatory’s HARPS (High Accuracy Radial velocity Planet Search) program. Out of a total of 4,864 individual RV measurements of τ Ceti acquired over 205 “epochs” (i.e. nights of observations) using the HARPS spectrograph on ESO’s 3.6-meter telescope in La Silla, Chile, the best quality 4,398 measurements from 202 epochs spanning over 2,142 days were chosen for analysis. A visual inspection of these data showed it to be flat with a standard deviation of just 1.7 meters per second. Next were a set of 978 RV measurements taken over 4,923 days using the UCL (University College London) Echelle Spectrograph on the 3.9-meter Anglo-Australian Telescope in Siding Spring, Australia. Acquired as part of AAPS (Anglo-Australian Planet Search) program, these RV measurements had a higher standard deviation of 5.0 meters per second. Finally, a set of 546 RV measurements from HIRES (High Resolution Echelle Spectrometer) on the ten-meter Keck I telescope located on the summit of Mauna Kea in Hawaii rounded out the data set. The data from the Keck Observatory covered a span of 3,446 days and had a standard deviation of 2.9 meters per second.

Here are the raw radial velocity (RV) measurements plotted as a function of Julian day from HARPS (top), AAPS (middle) and HIRES (bottom) used in the original analysis by Tuomi et al. published in 2013. Click on image to enlarge. (Tuomi et al.)

After much investigation, Tuomi et al. found that an approach which assumed a two-component Gaussian noise model plus a moving average with an exponential decay to account for correlated noise between the data points provided a significant improvement in their statistical models allowing the detection of RV signals with amplitudes less than one meter per second. Tuomi et al. found five periodic signals ranging from 14 to 643 days in their analysis of the combined data sets for τ Ceti which did not correlate with any of the activity indicators independently derived from the analysis of the HARPS spectra. The signals corresponding to the three shortest periods were also seen by analyzing just the higher quality HARPS data set alone. Tuomi et al. cautiously interpreted these five signals as being due to super-Earth mass planet candidates orbiting τ Ceti and publicly announced their results on December 19, 2012. The properties of these exoplanet candidates are listed below in Table 1 with data taken from Tuomi et al.. Calculated from these data and included in the table are the effective stellar flux or Seff values – the amount of energy each planet receives from τ Ceti compared to the amount of energy the Earth receives from the Sun.


Table 1: Properties of Exoplanets found in 2013 by Tuomi et al.
Planet MPsini


Orbital Period


Semi Major Axis




b 2.0±0.8 13.97 0.105 44
c 3.1 (+1.4/-1.1) 35.4 0.195 13
d 3.6±1.7 94.1 0.37 3.5
e 4.3 (+2.0/-2.1) 168 0.55 1.6
f 6.6±3.5 642 1.35 0.27


Since the inclination, i, of the planets’ orbits with respect to the plane of the sky can not be determined from RV measurements alone, only the minimum mass or Mpsini of these exoplanets can be determined at this time. The actual masses will likely be higher than these indicated values.

This diagram shows the orbits of the exoplanet candidates found by Tuomi et al. orbiting τ Ceti in relation to a less conservative definition of the habitable zone (HZ). Click on image to enlarge. (PHL@ UPR Arecibo)

Two of these exoplanet candidates, τ Ceti e and f, have Seff values broadly consistent with some definitions for the habitable zone (HZ) which generated a lot of hype in the media at the time and since. However, Tuomi et al. were much more careful about what they were claiming. While they were certain that their new technique was detecting low amplitude RV variations in their data, the cause of those variations was another matter. On one hand, the RV variations could be planetary in nature and dynamical simulations suggested that it would be a packed systems that appeared to be stable in the long term. However, the RV variations detected were typically on the order of 0.6 meters per second and could easily be the result of the subtle effects of previously unrecognized surface activity on τ Ceti. This was especially true of the RV signal of τ Ceti c with a period of 35.4 days which was uncomfortably close to the 34-day period of rotation of its parent star. Even though there was no obvious signal with this period observed in the various activity indicators examined by Tuomi et al., this coincidence does raise a flag for the planetary interpretation of this signal. Choices of parameters for fitting the noise in their model could also affect the observed amplitude of the recovered signals suggesting that the modelling was incomplete.

In the end, the work by Tuomi et al. was a positive step forward in developing new mathematical techniques for analyzing precision RV measurements to detect small periodic signals. However, the claim that there were five exoplanets orbiting τ Ceti was not on the firmest of footings and requires independent confirmation. As is frequently the case in such situations, more high quality data possibly in combination with still better data analysis techniques would be needed to resolve the issue.


New Planet Search Results

After a wait of almost half a decade, the same collaboration of astronomers have published the paper of Feng et al. which follows up the earlier work of Tuomi et al.. This time around, Feng et al. started with a vastly expanded set of precision RV measurements. The bulk of these came from a collection of 9,000 publicly available RV measurements from HARPS covering a period from June 2003 to September 2013 with much of the newer data of superior quality with less noise than the data available to Tuomi et al.. After obvious outliers (i.e. data points beyond 5σ) were removed, 8,880 data points were used by Feng et al. in the new analysis – double the number in Tuomi et al. from HARPS.

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

The second set of data used in the new analysis were publicly available measurements from HIRES. Much of the data available from HIRES for this new analysis was acquired at a high-cadence during a five-day period starting on October 1, 2004 as part of an asteroseismology study of τ Ceti at the Keck Observatory. The accuracy of the short exposure spectra was affected by periodic guiding errors introducing unwanted noise in this part of the data set. For the new analysis, Feng et al. discarded the 1,597 data points acquired before May 9, 2005 leaving 752 HIRES data points of superior quality – a third again more data points than were available from HIRES to Tuomi et al..

A view of the two ten-meter telescopes of the Keck Observatory atop of Mauna Kea. Keck I is fitted with the HIRES spectrometer used to observe τ Ceti and other nearby stars. (NASA)

In addition to a greatly expanded set of precision RV data, Feng et al. also introduced an updated method of analyzing the data which takes into account how the various sources of noise may vary as a function of wavelength. In Echelle spectrographs like that used for the HARPS survey, a diffraction grating breaks up a star’s light into a repeating series or orders of overlapping spectra (unlike a simple prism which breaks up the incident light into a single spectrum). Next, a second coarser grating oriented at a right angle separates the orders in such a way that a short segment of the spectrum of one order is projected above a short segment of the next order which covers an adjacent part of the spectrum which is projected above another short segment covering the next part of the spectrum and so on. The result is a series of high resolution spectra projected one above the other which together covers a large range of wavelengths and can be recorded using sensitive two-dimensional detector arrays commonly employed in astronomical imaging. In the case of HARPS, 72 orders are recorded covering wavelengths from 378 to 691 nanometers.

An example of an Echelle spectrogram showing how the various orders are stacked one above the other to yield a single high resolution spectrum with can be recorded with a 2D imaging array. (NOAO)

The HARPS program’s standard data reduction tools such as TERRA used by Feng et al. calculates the RV for each one of the orders. In the past, these 72 RV values were simply averaged to arrive at a single high precision RV measurement. In Tuomi et al., the noise in these averaged measurements was then characterized. Feng et al. instead characterize the noise separately for each order using differential RV measurements which isolate the wavelength-dependent noise in the data. After the wavelength-dependent noise has been modelled, only then are the RV measurements weighed and averaged to arrive at a single RV measurement for that spectrum. Subsequent analysis was performed with these data at their original cadence and with the measurements averaged into one-hour bins – a tactic which helps to reduce the random noise in the data but which can mask some of the correlated noise from one measurement to another complicating the modelling of this sort of noise.

These four plots show the radial velocity (RV) variations as a function of orbital phase for the four exoplanet candidates found by Feng et al.. Click on image to enlarge. (Feng et al.)

After investigating the effects of a range of parameter choices to model the noise in such a way to optimize their results and applying the technique to various combinations of data sets, Feng et al. found strong evidence for four periodic signals in the new, expanded data set which seemed to be “Keplerian” – i.e. a periodic signal consistent with that of an orbiting exoplanet candidate. Two of those signals with periods of 163 and 636 days appear to match those of τ Ceti e and f, respectively, found a half decade earlier in Tuomi et al.. The other two signals with periods of 20 and 49 days are significantly different from anything found by Tuomi et al. and have been given the new designations of τ Ceti g and h by Feng et al. to avoid any confusion with the earlier planet candidates. The properties of these exoplanet candidates found by Feng et al. are summarized below in Table 2.


Table 2: Properties of Exoplanets found in 2017 by Feng et al.
Planet MPsini


Orbital Period


Semi Major Axis




g 1.7 (+0.3/-0.4) 20.00 0.133 29
h 1.8 (+0.7/-0.3) 49.4 0.243 8.8
e 3.9 (+0.8/-0.6) 163 0.54 1.80
f 3.9 (+1.1/-1.4) 636 1.33 0.29


Feng et al. go into some detail about the potential origins of the signals seen by Tuomi et al. with periods of 14, 35 and 94 days which corresponded to τ Ceti b, c and d. Basically, Feng et al. describe these earlier signals as being “non-Keplerian” and the result of various forms of subtle stellar activity whose signatures have been modified by uneven sampling and incomplete modelling of natural sources of noise in Tuomi et al.. In addition to the fact that the inner exoplanet candidates proposed by Tuomi et al. and now by the new analysis of Feng et al. are incompatible (i.e. they would be dynamically unstable if all of them were present), it now seems unlikely that these earlier exoplanets candidates exist and that τ Ceti g and h are better exoplanet candidates to explain the variations in RV observed in the old as well as the new and expanded data set.

Dynamical simulations of the four-planet system found by Feng at al. suggest that it is packed yet appears to be stable in the long term despite the somewhat high eccentricity values that were found by Feng et al.. However, it is pointed out that these eccentricity values are likely being overestimated due to the effects of instrument noise and may be much lower in reality helping to stabilize the system more. The long term stability of this packed system is less sensitive to the actual masses of these exoplanets which are surely larger than their minimum mass or Mpsini values. If these exoplanet candidates are in the same plane as the debris disk which has been characterized by the analysis of recent Herschel Space Observatory and ALMA data to have an inclination of around 35°, the actual masses could be about 1.74 times greater than the derived Mpsini values. Future observations and dynamical studies will help resolve this question.

These charts compare the τ Ceti system as found by Feng et al. (top) with that of the Sun (below) with the limits of a less conservative definition of the habitable zone (HZ) indicated. Clock on image to enlarge. (Fabo Feng)

While the exoplanet candidates proposed by Feng et al. seem to be on much firmer footings than the earlier claims of Tuomi et al., it needs to be remembered that these are still just candidates requiring independent confirmation. The semiamplitudes of these RV signals are just 35 to 55 centimeters per second – a totally new realm for the RV technique which might be revealing some new phenomena which could be mimicking the signature of orbiting exoplanets. More observations, including using methods other than precision RV measurements, will be required to confirm any of these four new candidates. But assuming for the moment that these exoplanets actually exist, what is the likelihood that τ Ceti e and f, which appear to straddle many definitions of the HZ of τ Ceti, are actually habitable?


Potential Habitability

A thorough assessment of the habitability of any extrasolar planet would require a lot of detailed data on the properties of that planet, its atmosphere, its spin state, the evolution of its volatile content and so on. Unfortunately, at this very early stage, the only information typically available to scientists about extrasolar planets are basic orbit parameters, a rough measure of its size and/or mass and some important properties of its sun. Combined with theoretical extrapolations of the factors that have kept the Earth habitable over billions of years (not to mention why our neighbors are not habitable today), the best we can hope to do at this time is to compare the known properties of extrasolar planets to our current understanding of planetary habitability to determine if an extrasolar planet is “potentially habitable”. And by “habitable”, I mean in an Earth-like sense where the surface conditions allow for the existence of liquid water – one of the presumed prerequisites for the development of life as we know it. While there may be other worlds that might possess environments that could support life, these would not be Earth-like habitable worlds of the sort being considered here.

The first step in assessing the potential habitability of τ Ceti e and f is to determine what sort of worlds they are: are they rocky planets like the Earth or are they volatile-rich mini-Neptunes possessing deep hot atmospheres dominated by hydrogen overlaying layers of exotic high temperature ices with little prospects of being habitable in an Earth-like sense? Unfortunately, the only information currently available about these exoplanets that could help make such an assessment is their Mpsini or minimum mass values. Their actual masses and the radii are needed to calculate these exoplanets’ densities and get a handle on their bulk compositions. The transit method, like that employed by NASA’s highly successful Kepler mission (see the Kepler Mission page), has been effective in determining the orbit inclination, i, and the radii of exoplanets. But the odds that the orbits of τ Ceti e and f are, by random chance, aligned to produce transits visible from the Earth (i.e. their orbits have an inclination close to 90°) are about 0.7% and 0.3%, respectively. Unless τ Ceti e and f beat the odds, we will likely have to wait for the introduction of a large space-based telescope specifically designed to observe exoplanets orbiting nearby stars before we will be able to determine their orbit inclination directly and estimate their sizes.

In lieu of this vital information on the density of these exoplanets, statistical arguments can be made about the probability they have a rocky composition based on what we do know. An analysis of the mass-radius relationship for extrasolar planets smaller than Neptune performed by Rogers strongly suggests that the population of known exoplanets transitions from being predominantly rocky planets like the Earth to predominantly volatile-rich worlds like Neptune at radii no greater than 1.6 times that of the Earth or RE (see “Habitable Planet Reality Check: Terrestrial Planet Size Limit”). While rocky planets larger than this are possible, they quickly become more uncommon with increasing radius. A planet with a radius of 1.6 RE and an Earth-like composition would have a mass of about six times that of the Earth or 6 ME. More recent work by Chen and Kipping with a larger sample of exoplanets suggests that the gradual transition of the exoplanetary population from predominantly rocky planets to volatile-rich worlds starts at about 2 ME.

Considering the ~3.9 ME Mpsini values of the exoplanets of τ Ceti e and f found by Feng et al. (and keeping in mind that the actual masses are likely to be larger), they both readily exceed the 2 ME mass threshold found by Chen and Kipping indicating that they have at least some chance of being mini-Neptunes. Then again, given a randomly oriented orbit, there is something like one chance in four that the actual masses of τ Ceti e or f exceed the 6 ME threshold found by Rogers where the odds begin to heavily favor mini-Neptunes. But if the orbits of τ Ceti e and f are aligned with the observed debris disk distantly orbiting their parent star with an inclination of 35°, the actual masses of these exoplanets would be about 7 ME readily exceeding Rogers threshold.

An IR image acquired using ESA’s Herschel Space Observatory showing the debris disk around τ Ceti inclined 35° to the plane of the sky. (Lawler et al.)

But before we get too invested in any interpretation of what sort of worlds τ Ceti e and f might be, we must also remember that the current MPsini values for these exoplanet candidates have very large uncertainties reflecting not only the effects of unavoidable measurement errors but also how the various sources and types of noise in the RV data are modelled. The MPsini values for τ Ceti e and f of 4.3 and 6.6 ME, respectively, found a half decade earlier by Tuomi et al. were substantially larger than the values found for the same exoplanet candidates in the newer analysis by Feng et al. (not to mention giving the impression that it is almost certain that τ Ceti e and f are mini-Neptunes). Part of this is the result of incomplete modelling of the noise in the data which led to an overestimation of the amplitude of the RV signals for these two exoplanet candidates resulting, in turn, to larger calculated MPsini values. It might turn out that future mass determinations for τ Ceti e and f based on still better RV measurements and data analysis techniques may be even smaller than those found by Feng et al.. Until better data become available, all that can be said is that there is a substantial likelihood that τ Ceti e and f are mini-Neptunes with no chance of being habitable in the Earth-like sense but the possibility that they are instead rocky planets more similar to the Earth can not be excluded with reasonable certainty.

The next criterion that can be used to determine if a rocky exoplanet is potentially habitable is the amount of energy it receives from its sun known as the effective stellar flux or Seff. According to the work by Kopparapu et al. on the limits of the habitable zone (HZ) based on detailed climate and geophysical modeling, the inner limit of the HZ for an Earth-like rocky planet is conservatively defined by where a planet’s temperature would soar even with no CO2 present in its atmosphere as a result of a moist runaway greenhouse driven by the water vapor in the atmosphere. For an Earth-mass planet orbiting a star like τ Ceti, the Seff where this occurs is 1.05 times that of the Earth. But for a 5 ME rocky planet, the moist runaway greenhouse sets in at a higher Seff of 1.13 because the planet’s higher gravity compresses the atmospheric column somewhat lessening the greenhouse effect. At higher Seff values, the surface temperature of a planet soars hundreds of degrees above the boiling point of water while changes in the structure of the atmosphere allows for permanent water loss in a geologically brief time. With the water gone, the carbonate-silicate cycle which helps to regulate the amount of CO2 in the atmosphere of a geologically active planet breaks down allowing that gas to build up in the atmosphere. The end result is a sterilized Venus-like world with high surface temperatures caused by a CO2-dominated dry runaway greenhouse.

Looking at the Seff value of 1.80 for τ Ceti e, it is evident that it exceeds the more conservative limits for the HZ based on Kopparapu et al. and is almost as high as the 1.91 Seff value for Venus (which is most definitely not habitable). This would suggest that if τ Ceti e is a rocky planet, it is most likely to be a larger size version of Venus and not habitable. However, it is not impossible that τ Ceti e could manage to be habitable after all. Increasingly sophisticated climate models of planets with slow or even synchronous rotation (where one side of a planet always faces its sun) has shown not only that these worlds can be habitable but that they can be so at much higher Seff values than planets with faster rotation rates like the Earth. This is because various feedback mechanisms result in the formation of a permanent cloud layer over the dayside of a slowly rotating planet which helps to reflect away incident light and moderate surface temperatures. Recent models by Yang et al. suggest that a slow or synchronous rotator orbiting a star like τ Ceti could remain habitable with Seff values as high as 2.1 – 17% higher than the calculated Seff of τ Ceti e (presumably Venus’ slow rotation set in after it experienced a runaway greenhouse possibly because of interactions with its resulting dense CO2 atmosphere).

Normally, an Earth-size planet orbiting in or near the HZ of a Sun-like star would not become a synchronous rotator during the presumed 5.8 billion years since the formation of the τ Ceti system. However, a new paper by Rory Barnes (University of Washington) suggests that Earth-like planets which start off with a slow rotation rate may evolve into synchronous rotators more quickly. Barnes calculates that if the Earth started off with a period of rotation of about three days, its rotation would have slowed to become synchronous after 4.5 billion years assuming the Earth always had the same tidal properties we observe today and it had no large moon. Considering that τ Ceti would have been a couple of tens of percent or so dimmer after it first formed compared to today, it is possible that τ Ceti e might have started its life just inside the HZ of the infant τ Ceti system. If τ Ceti e started with a slow rotation rate and if its properties were such that tidal forces slowed its rotation over time quickly enough so that it stayed in the HZ for a slow rotator at the same time τ Ceti brightened with age, τ Ceti e might have managed to remain habitable to this day. While not an impossible scenario, on the surface this evolutionary path does not seem to be very “wide” leaving little room for deviation – the sort of deviations which would result in a Venus-like planet. There are also other definitions for the HZ with much higher Seff values but these have been criticized because they involve physically implausible conditions. While not impossible, it would seem that τ Ceti e has very poor prospects for being potentially habitable especially considering the real possibility it might be a mini-Neptune.

An artist’s depiction of the exoplanet candidates found orbiting τ Ceti. (University of Hertfordshire)

As a planet receives less energy from its sun, various processes such as the carbonate-silicate cycle allow more CO2 to build up in the atmosphere which helps to increase the greenhouse effect and maintain above-freezing surface temperatures. The outer limit of the conservative HZ, as defined by Kopparapu et al. corresponds to the maximum greenhouse limit beyond which a CO2-dominated greenhouse is incapable of maintaining a planet’s surface temperature. Instead of helping to heat the atmosphere, the addition of more CO2 beyond this point makes the atmosphere more opaque causing the surface temperatures to drop instead of increase. The latest work by Kopparapu et al. suggests an Seff value of about 0.33 for the outer limit of the HZ of a star like τ Ceti. There are some slightly more optimistic definitions of the outer edge of the HZ such as the early-Mars scenario, but these more optimistic definitions do not change the Seff for the outer limit of the HZ significantly.

Looking at the Seff value of 0.29 for τ Ceti f, it appears to orbit just outside of the more conservatively defined HZ limits of Kopparapu et al.. Given the uncertainties in the presumed properties of τ Ceti f as well as the models used to define the outer limits of the HZ, one could argue that τ Ceti f effectively orbits right at the outer edge of the HZ. However, 5.8 billion years ago when τ Ceti was dimmer than it is today, the Seff for τ Ceti f would also be lower pushing it well outside of even more optimistic definitions of the HZ for most of its existence. While τ Ceti f may still support some biocompatible environments below its surface where life might thrive, it is quite likely to be a frozen ice ball which can not support Earth-like conditions on its surface – assuming that it beats the odds and is not a mini-Neptune instead.



The new work by Feng et al., which combines an expanded set of radial velocity (RV) measurements with a new data processing technique which more effectively deals with noise in the data, have been used to identify four low-amplitude, periodic signals which might represent super-Earth-size exoplanets orbiting τ Ceti. Two of these exoplanet candidates, τ Ceti e and f, also appeared in an earlier analysis by Tuomi et al. and seem to straddle the habitable zone (HZ) prompting much media attention when their discovery was first announced in December 2012. While the exoplanet candidates proposed by Feng et al. seem to be on much firmer footings than the earlier claims of Tuomi et al., it needs to be remembered that these are still just candidates requiring independent confirmation of their planetary nature. More observations, including using methods other than precision RV measurements, will be required to confirm any of these four new candidates.

Assuming for the moment that τ Ceti e and f actually exist and have the properties found by Feng et al., it seems unlikely that either of these exoplanet candidates have particularly good prospects for being potentially habitable in an Earth-like sense contrary to some of the media hype over the last half decade. While the actual masses of these exoplanets are poorly constrained given the large measurement uncertainties and unknown orbital inclination, there is the distinct possibility that either or even both of them are not rocky planets like the Earth. Instead they could be volatile-rich mini-Neptunes with deep hot atmospheres dominated by hydrogen on top of layers of exotic ices which only form under high pressures and temperatures – definitely not the kind of place to find Earth-like habitable conditions amenable to life as we know it.

But even if τ Ceti e and f are rocky planets like the Earth, their prospects for being habitable are still not very good. The Venus-like effective stellar flux or Seff of τ Ceti e exceeds that for conservative definitions of the limits of the HZ. While one could invent various evolutionary scenarios which result in a habitable τ Ceti e today, it seems more likely that it is simply a larger version of Venus. The situation with τ Ceti f is just the opposite: its Seff is so low that it orbits just beyond the outer edge of most definitions of the HZ and would have been even farther outside of the HZ in the past when τ Ceti was dimmer. While there could be subsurface environments which could support life on τ Ceti f, on the surface this exoplanet would be a frozen ball of ice incapable of supporting Earth-like conditions. Combined with the likelihood that they could be mini-Neptunes, neither of these exoplanets seem to have particularly good chances of being potentially habitable. And given the packed nature of the planetary system as described by Feng et al., it would seem unlikely that τ Ceti could have any potentially habitable worlds orbiting stably inside of the HZ which have so far escaped detection.

While this assessment is not very rosy, it must be remembered that τ Ceti e and f have yet to be confirmed. And even if they are confirmed, their properties may end up being different enough from those found in the new analysis by Feng et al. to change the assessment of their potential habitability. But no matter what, the continued detailed study of any exoplanets found orbiting τ Ceti is sure to add to our understanding of planetary systems of Sun-like stars. In addition, the new data processing techniques presented by Feng et al. are a vital step in detecting Earth-size exoplanets orbiting other Sun-like stars as astronomers continue to push the practical limits of precision RV measurements.


Follow Drew Ex Machina on Facebook.


Related Reading

“Prevalence of Earth-Size Planets Around Sun-Like Stars”, Drew Ex Machina, November 3, 2015 [Post]

“Earth Twins on the Horizon?”, Drew Ex Machina, January 9, 2015 [Post]

“Habitable Planet Reality Check: Kepler’s New Planet Candidates”, Drew Ex Machina, June 22, 2017 [Post]


General References

Rory Barnes, “Tidal Locking of Habitable Exoplanets”, arXiv 1708.02981 (Accepted for publication in Celestial Mechanics and Dynamical Astronomy), August 9, 2017 [Preprint]

Jingjing Chen and David Kipping, “Probabilistic Forecasting of the Masses and Radii of Other Worlds”, The Astrophysical Journal, Vol. 834, No. 1, Article id. 17, January 2017

Fabo Feng at el., “Color difference makes the difference: Four Planet candidates around τ Ceti”, arXiv 1708.02051 (Accepted for publication in The Astrophysical Journal), August 7, 2017 [Preprint]

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

S.M. Lawler et al., “The debris disc of solar analogue τ Ceti: Herschel observations and dynamical simulations of the proposed multiplanet system”, Monthly Notices of the Royal Astronomical Society, Vol. 444, No. 3, pp. 2665-2675, November 2014

Leslie A. Rogers, “Most 1.6 Earth-Radius Planets are not Rocky”, The Astrophysical Journal, Vol. 801, No. 1, Article id. 41, March 2015

M. Tuomi et al., “Signals embedded in the radial velocity noise: Periodic variations in the τ Ceti”, Astronomy & Astrophysics, Vol. 551, Article ID A79, March 2013

Jun Yang et al., “Strong Dependence of the Inner Edge of the Habitable Zone on Planetary Rotation Rate”, The Astrophysical Journal Letters, Vol. 787, No. 1, Article id. L2, May 2014