≡ Menu

Alpha Centauri B: A Close Look at the Habitable Zone

The dreams of Alpha Centauri I used to have as a boy all focused on visual effects. After all, the distance between Centauri A and B ranges from 11.4 to 36.0 AU. What would it be like to have a second star in our Solar System, one that occasionally closed to a little more than Saturn’s distance from the Sun? What would a day be like with two stars, and even more, what would night be like with a star that close lighting up the landscape? I also wondered about how much effect a second star would have on the planets in our system, curious as I was about gravitational effects and even the possible repercussions for weather and seasonal change.

Image: The Alpha Centauri star system and other objects near it in the sky. Image copyright Akira Fujii / David Malin Images.

You can imagine, then, that Duncan Forgan’s new paper hit close to home. Forgan (University of Edinburgh) has taken discussions of habitability around Centauri B to a new level by analyzing the effect of Centauri A on habitability using latitudinal energy balance models that allow him to study how small changes in the properties of a planet can affect the overall climate there. Such models have been useful in studying things like climate variability due to orbital eccentricity and other factors, and Forgan puts them to work to chart the effect of a binary companion.

Alpha Centauri in the Last Fifteen Years

Before I get into the results of the habitability study, though, I want to go through some of the more recent work on Alpha Centauri, all summarized carefully in the Forgan paper. Indeed, I point you to this paper with great assurance that if you are interested in the Centauri stars, you’ll find a useful bibliography and summary here that will quickly get you up to speed (though the bibliography would be better if it listed paper titles along with the rest of the citations). Let’s run through some of the more salient work — in most cases I’ll skip the authors and citations in this discussion, knowing that Forgan’s work containing all of these is freely available at the arXiv site.

Centauri A and B, being high metallicity stars, are presumably prime candidates for circumstellar disks with a high solid material component, making the building blocks of planets readily available, and deepening the spectral lines for improved precision in radial velocity studies. Another useful factor for observations is that the binary is inclined by only 11 degrees with respect to our line of sight, an important fact because it means that any planets we discover through RV methods will yield a mass that is fairly accurate, assuming that the planets around these stars have formed in the same orbital plane. Without such knowledge, the mass figures from RV studies vary widely depending on assumptions about the target system’s inclination.

Studies on planet formation have shown that both Centauri A and B should be capable of forming terrestrial planets even when the perturbations caused by the binary companion are taken into account. Early studies on this question have found that the planetesimal disks seem to be stable out to about 3 AU of the parent stars, assuming a reasonable inclination of the disk relative to the binary plane, meaning something less than 60 degrees. More recent work by Thébault and colleagues has shown that the later stages of accretion may not be efficient because the binary companion can inhibit the growth of larger objects outside 0.75 AU (Cen A) and 0.5 AU (Cen B).

What does this mean? Most likely that the formation of gas giants is unlikely here (a finding that squares with previous radial velocity surveys), while if we can get past the problem of forming larger planetesimals referred to above, Earth-mass planets should be able to form in the habitable zone of Centauri B, assuming an eccentricity of no more than 0.3. A 2009 study I’m not familiar with by Michtchenko & Porto de Mello makes the case that any terrestrial planets that do form in Centauri B’s habitable zone should be dynamically stable despite perturbations from Centauri A under certain conditions of eccentricity and orbital inclination, but planets with inclinations to the orbital plane larger than about 35 percent should experience strong instability.

So where is the habitable zone around Centauri B? Kasting and team used a model that assumed Earth-mass planets with similar atmospheric composition and found a habitable zone ranging from 0.5 to 0.9 AU, although this 1993 study did not include the perturbing influence of Centauri A. But Forgan notes this with regard to the light reaching Centauri B planets:

If main sequence relations for the luminosity of each object are assumed, the insolation experienced by planets in the habitable zone of α Cen B due to α Cen A would be no more than a few percent of the total insolation of the α Cen AB system at the binary’s periastron, and around one tenth of a percent at apastron. This insolation can be diminished further by eclipses of α Cen A by α Cen B, the duration of which is estimated to be of order a few Earth days.

Tuning the Model for Centauri B

Kasting was using a global radiative balance model (GRBM), but he and other researchers later deployed latitudinal energy balance models (LEBMs) of the kind Forgan uses in his new study, the latter being more complex and incorporating assumptions about latitude and season and other properties that would be temperature dependent. Forgan adjusts the model to include the effects of the binary (neglecting the distant M-dwarf Proxima Centauri). From the paper:

A planet in global radiative balance is not in general in local radiative balance, and by extension habitability is not a discrete concept (i.e. either habitable or uninhabitable), but a continuous one, where a certain fraction of the planet’s surface will be habitable at any given time. In the LEBM, the evolution of the planet’s temperature T (λ) is described by a diffusion equation made nonlinear by the addition of the heating and cooling terms, as well as an albedo which makes a rapid transition from low to high as temperature decreases past the freezing point of water. As a result, small changes in the properties of a planet can strongly affect the resultant climate.

The latitudinal energy balance model, then, seems the best approach for asking how the perturbations caused by Centauri A might affect planets in the habitable zone of Centauri B.

So what does Forgan find? It turns out that calculating the habitable zone of Centauri B’s inner and outer boundaries can be roughly correct if we leave Centauri A out of the picture — the dimensions of the habitable zone remain more or less the same. But adding Centauri A does create oscillations in the planet’s climate that happen when Centauri A is at its closest to Centauri B. The temperature variations caused by Centauri A are no more than several K, and could alter the fraction of habitable surface on planets at the habitable zone boundaries by about 3 percent, a figure made flexible depending on the size of oceans or planetary obliquity.

The paper goes on to note the possible effect on life (science fiction writers take note):

It is reasonable to speculate that if life were to exist on planets around α Cen B, that they may develop two circadian rhythms (cf Breus et al. 1995) corresponding to both the length of day around the primary, and the period of the secondary’s orbit (approx 70 years). Altering the available habitat by a few percent may also influence migration patterns and population evolution.

Small changes over time, though, can lead to big results, as Forgan goes on to remind us:

While we have demonstrated that the temperature fluctuations for planets around α Cen B due to α Cen A are relatively small, the consequences of a periodic temperature forcing of a few K to long term climate evolution cannot be fully understood from this work. To fully appreciate the impact on (for example) ocean circulation and carbonate-silicate cycles requires further investigation with more advanced climate models.

Simulations and Their Limitations

The paper analyzes the results of the simulation for different classes of planets, from fully habitable worlds to uninhabitable hot planets, uninhabitable snowball planets, and two other classes — eccentric transient planets and binary transient planets — that are both partially habitable, with the habitability oscillating according to either the planet’s orbit around Centauri B (eccentric transients) or the period of Centauri A (binary transients). Forgan is careful to comment on the limitations of the LEBM model, which is not sensitive to long-term climate processes, and he notes that adding clouds and a carbonate-silicate cycle into the mix would potentially extend the outer edge of the habitable zone. Another limitation: The model is not sensitive to planets with extremely slow rotations.

Nevertheless, previous work with such modeling has shown its effectiveness, and the picture of the potential Centauri B system that emerges is one in which habitable worlds could well flourish. On the latter score, one other note:

The inner edge of the habitable zone is less well-defined than the outer edge – atmospheric changes could allow liquid water above 373 K, and the runaway greenhouse effect may become important at temperatures nearer 350 K (Spiegel et al. 2008 and references within). In any case, the outer edge is likely to be more interesting from an astrobiological standpoint, as current and future instrumentation will be more capable of prob[ing] spectral features of planets at larger semi-major axes (see e.g. Kaltenegger & Selsis 2010).

What’s needed now, of course, are radial velocity results from the ongoing studies of Alpha Centauri, which will begin to tell us whether or not rocky terrestrial worlds actually exist there. This is tricky work — radial velocity methods are much happier with huge gas giants in close orbits than with small rocky planets, which demand a much longer analysis. The paper is Forgan, “Oscillations in the Habitable Zone around Alpha Centauri B,” accepted for publication in Monthly Notices of the Royal Astronomical Society (preprint) and highly recommended to anyone with an interest in planets around our nearest stellar system.


Comments on this entry are closed.

  • Ronald February 21, 2012, 12:27

    With regard to age, metallicity etc.of Alph Cent A and B:

    Mamajek & Hillenbrand put the age anywhere between about 4.5 and 6.5 gy.
    Metallicity is about 60% higher than solar.

    Since the mass of Alph Cen A is 10% greater than our sun’s, the higher luminosity (about 1.56 times solar) can partly be explained by this (luminosity correllates with the 3rd power of mass). However, even so its luminosity is on the high side. Maybe then this is also a result of its slightly advanced age in comparison with our sun, enhanced by the fact that larger mass stars age more quickly. Not a very pleasant prospect for (higher) life on a planet orbiting it. This relative disadvantage of A in comparison with B and our sun is emphasized by the fact that the present Habitable Zone of A extends from about 1.2 AU outward, where terrestrial planet formation may have been more problematic and more easily disrupted by companion B.

    Alph Cen B, being smaller in mass (about about 0.9 times solar) and luminosity (about 0.48 times solar), its stable main sequence lifespan will be much greater than A’s and its present age, be it 4.5 or 6.5 gy, no problem. Furthermore, the HZ extends from about 0.65 AU outward, where terrestrial planet formation was less easily disrupted by companion A.

    Because of the high metallicity it seems likely that giant planets could and would have formed. However, any planet in the outer system (beyond about 3 AU) is likely to have been severely disrupted (swallowed or ejected), if the protoplanetary material (planetesimals etc.) was not already absorbed by the parent stars.
    And previous studies have practically ruled out giant planets in the inner systems.
    The question remains, however, in view of mentioned high metallicity, whether any remaining planets in stable orbits in the inner systems of A and B will appear to be very large planets of the ‘subgiant’ class (Neptune class, ice giant seems an odd name so close to the star) and/or super-earth class.

  • jkittle February 21, 2012, 12:42

    Bacteria constantly exchange genetic information and in fact it is very difficult to define bacterial “species” havet are not carrying at least genes that have been “horizontally transferred” from even very distantly related bacteria. Thus you can trace the ancestor/ phylogenetic relationships of individual genes easier than the relationship between bacterial. There are cares where two seemingly different bacteria share 20, 30 or even 50% of their genes( with very few mutations) in some parts where as other parts of the genome are very dissimilar. It is really genome “mix and matc” in the prokaryotic world. The main difference I can see is that most bacteria are haploid most of the time ( but some yeast are haploid too…) and that the genomes of most prokaryotes are MUCH smaller, ( though there are some yeasts that have very compact genomes.) There are even yeasts that lack mitochondria. So the spectrum of living organism genomes is very broad and not a sharply defined into categories as the textbooks would oversimplify.

  • Ronald February 21, 2012, 12:43

    My previous comment also brings me to a former question, again, whether super-earths of roughly 5 – 10 Me are the (remaining) rocky cores of ice (sub) giants and gas giants, of which the gaseous/liquid envelopes have been eroded away by stellar irradiation or were not even able to form in the first place, because of proximity to the mother star.
    Maybe, in the near future we will learn a simple and logical rule that a (potential) gas/ice giant within a certain minimum distance from its star (the snow line?) is or becomes a rocky super-earth.

    Does anyone have to say anything interesting about this?

  • FrankH February 21, 2012, 16:01

    Alpha Centauri A is 10% more massive than the Sun, which accounts for the greater luminosity. It’s still well within the main sequence.

    That’s a good article., and brings up some interesting points in the discussion section. The ESO article:


    is where the 4.8BY age comes from, but they don’t say where the figure comes from.

    The diameters are based on this article:


    The Yildiz article is pretty solid, so the age has to lie somewhere in the 4.8 – 5.6BY range.

  • Eniac February 21, 2012, 16:02


    I stand by my comments about the emergence of the Eukaryote being a singular rare event. One of the reasons why it was such a rare event is because the composite cell had genes that were not necessary for its survival until the Oxygen environment was created, which took millions of years. Bacteria are under strong selective pressure to shed superfluous genes in order to optimize proliferation rate. This is why bacteria tend to have minimal genomes. The initial hybrid bacteria (which lead to the Eukaryote cell) was not optimized in this manner. Thus, its survival was a matter of extreme happenstance.

    I do not follow this argument at all. The reason it was such a rare event is because the composite cell had too many genes and would have lost them before they would have been useful, except by chance it didn’t? What if a composite were instead formed while it would have been useful? This is how things usually evolve, and your argument would fall flat in this case. Perhaps I misunderstand.

    BTW, diving at 5 bar pressure (40 meters down) can only be done for a limited time (1-2 hours) before you start to have Nitrogen narcosis. Also, the partial pressure of Oxygen is way too high for optimal health. This is why Helium-Hydrogen gas mixtures are used by people who do this diving commercially or who have lived in the undersea habitats like SeaLab and the like.

    This is all true, but not really relevant. We have already accepted we’ll need artificial air, because there is no oxygen at all. So, we can use the appropriate oxygen partial pressure, leave out the nitrogen, and do whatever else we please to make it suitable for the pressure, or not?

  • Rob Henry February 21, 2012, 16:40

    Eniac, brings up two points regarding sexual reproduction.

    1) “The key difference, I think, is that a given diploid organism has an entire population in its ancestry”
    There are several ways that I can treat this statement, but I will examine just two.

    a) The diploid condition allows robust health due to buffering. The problem with that is that without buffering bad genes are removed faster, and that more than makes up for the effect.

    b) The choice of gene combinations available allow ordinary individuals to produce extraordinary offspring, and sub par ones – then the extraordinary ones pass on more genes. Unfortunately this is (nearly) perfectly cancelled by natural selection on asexual populations being more closely matched to the genes that are passed on. Actually this is just a restatement of Muller’s ratchet.

    2) “The many disadvantages you mention just underscore the point, since practically all higher organisms take them in stride. They MUST be balanced by substantial advantages, and while there is no real consensus what those are, they nevertheless exist.”
    And that is a truly great point, and I believe the reason that so few are prepared to take on faulty pro sexual reproduction oversimplifications. So far their have been only two prominent replies to this question that have come to prominence, and note that each might just be peculiar to Earth or terrestrial life.

    a) That the diploid state allows rapid switching between sets of genes that are optimised to two radically different states that the environment oscillates between on time scales much greater than individual lifetimes. This has been thought to be disproved for our case ever since it was shown that rapidly varying environments are more likely to contain asexually reproducing variants of more flexible groups.

    b) That the parasite load was so high for a progenitor group that shared a common parasite that one subgroup would have a high advantage if they adopted sexual reproduction. This *Red Queen” has been rapidly adopted after the fall of the above theory and clung to despite problems.

  • Rob Henry February 21, 2012, 17:04

    I have just realised that I might have missed a subtly that has falls outside being good for the whole population, but may still be good for the entire course of evolution. With sexual reproduction populations of the most robust complex organisms can remain outside large populations and thus resist purifying selection. This must help speciation a lot, even if I have earlier implied the balance of evidence indicates that it is no help to adaptation on the timescales needed for terrestrial life.

  • Eniac February 22, 2012, 2:39


    The diploid condition allows robust health due to buffering. The problem with that is that without buffering bad genes are removed faster, and that more than makes up for the effect.

    The first part, in my opinion, is not the key advantage here, although it may well be essential to supporting larger genome sizes. This second part would probably be true if it was always immediately clear which genes are bad and which are good. There may be a lot of good genes that only work in particular environments, or in particular combinations with other genes. A “gene pool” from which gametes are drawn in random combinations keeps genes around much longer and thus gives them a much greater chance to prove themselves. Truly bad genes will still be eliminated in due time, they will simply be crowded out by the ones that prove themselves good. I do not agree with you that bad genes staying around longer “more than makes up” for the positive effect.

    You seem very well read on this, and mine may be an antiquated, disproved notion. If so, don’t hesitate to point this out and comment on how it was disproved.

  • kzb February 22, 2012, 14:26

    Ronald: interesting comments about the age and luminosity of “A”. It’s increased luminosity relative to the sun is only partly explained by its 10% greater mass. Also its age is greater than solar. So it’s luminosity could already be on a rapidly increasing trend, and a previously habitable planet could now be on the way to being cooked !

  • Rob Henry February 22, 2012, 16:27

    Eniac, I wasn’t saying that sex was disadvantageous in Earth’s case, only that the reason that it is advantageous to higher life is mysterious, and may not apply on other biospheres. Keeping old genes about is definitely useful if the environment varies sufficiently on a scale of a thousands of generations. Also it is useful if it keeps around deactivated formally dominant genes that can be reactivated via point mutation to readaptation of a population if the environment ever reverts to a millions of years old states. Or
    perhaps all populations of complex animals on Earth tend to periodically go through bottlenecks of small numbers that would accentuate Muller’s ratchet, and we just haven‘t yet experienced what causes these contractions.

    Your ideas aren’t so much antiquated as they are difficult to fit for our case. And that brings me back to the Red Queen, which is a harder nut to crack. Our fossil record implies a certain *punctuated equilibrium* pattern of evolution, that can be most easily explained by small isolated populations forming around larger populations, and then one of these replacing the entire larger population. The only problem is that their lower genetic diversity would preclude such a replacement by these precepts.

    I don’t claim that I know what‘s happening, but I do ask, that if we ever find an ETI, we shouldn’t be to surprised if their biological progenitors were asexual.

  • andy February 22, 2012, 18:01

    There’s also a metallicity effect in the stellar evolution: higher metallicity stars are cooler, less luminous and have longer main sequence lifetimes at a given mass.

    Running the Yonsei-Yale stellar evolutionary tracks interpolator for a 1.1 solar mass star with solar and 1.51 times solar metallicity, the solar metallicity track has a main sequence lifetime of 6.1 Gyr versus 6.9 Gyr for the metal-rich track. Current luminosity occurs at 5 Gyr or so along the metal-rich model, so it’s got some time yet to go before things really speed up in terms of luminosity evolution.

    Increase in luminosity for Alpha Centauri A over the main-sequence lifetime appears to be roughly 50%, versus roughly 30% for the Sun. Alpha Centauri B works out as roughly 30% luminosity increase due to the slower evolution at lower masses.

  • Rob Henry February 22, 2012, 18:34

    Again I feel the want to add a speculation of my own as an afterthought. The fact that the sex determining mechanisms are unrelated in such closely related species as turtles, birds and mammals implies to me that sexual reproduction is an ongoing advantage on Earth. Let me explain.

    The simplest way to determine gender for a sexually reproducing species that adopts a dioecious state is to do so via a single gene switch. If so this gene must be transmitted asexually. The fact that this setup has proved so unstable over such a deep ancestry must surely indicate that on Earth, sexual reproduction has a very large advantage.

  • Eniac February 22, 2012, 19:31

    Rob, you have a good point. Let us make for a moment the assumption that the sole reason for the diploid genome is the increased possibility of proof-reading and error correction it provides, without which larger genomes could not be supported. This is reasonable, since replication error rate and maximum supportable genome size are inversely related.

    In this case, we could imagine a different biochemistry the DNA analog of which is much more stable than ours, and which naturally supports a much lower error rate without diploid proofreading. Then, it would be reasonable to conclude that the evolution of diploidy might never happen, and there would not be sexual reproduction.

    However, there can be diploidy without sexual reproduction, which would seem to imply that there is more to it than just proofreading and error correction. In my view, redundancy and proofreading are both nice advantages of a diploid genome, but the shuffled gene pool is the real deal, the game changer that eclipses both of the others.

    You also alluded to speciation as a consequence of sexual reproduction and the cross-breeding barrier it entails. It appears to me that might be a fourth item to consider, which could also be a big one in terms of impacting evolution itself.

  • kzb February 23, 2012, 8:20

    Thanks for looking at the stellar evolution of “A” Andy. I was getting a bit worried that any colony established in the AC system would have a substantially lower future life expectancy than Earth. The model supports the lower age estimates for the system.
    However, it’s worth checking what the 50% increase in luminosity over its main sequence life means for habitable planet evolution ?

  • Ronald February 23, 2012, 11:55

    kzb (February 22, 2012 at 14:26), with regard to ‘A’:
    “(…) So it’s luminosity could already be on a rapidly increasing trend, and a previously habitable planet could now be on the way to being cooked !

    Indeed! Although andy rightly mentions, that a high metallicity star has a longer main-sequence lifespan, and A is probably around 5gy, it’s higher mass more than offsets this. And the fact that A is presently so bright (about 1.56 * solar luminosity) does not bode well for a planet in the (former) HZ.
    Our own sun, at about 4.6 gy of age, gets about 10% brighter per gy and the earth will probably become to hot for (higher) life in about 0.5 – 1 gy.
    It is likely that A with its higher mass is already much more advanced in this process.
    After all, the higher the stellar mass, the faster its evolution and the shorter the ‘habitable lifespan’, meaning the duration of a planet in its HZ. In other words, although brighter stars have a wider HZ, the Continuously HZ (CHZ, defined as the HZ during at least 3 gy) becomes narrower beyond a certain mass and hence the window of opportunity for (higher) life.

  • Ronald February 23, 2012, 11:57

    I meant: and hence the window of opportunity for (higher) life *also becomes shorter*.

  • Ronald February 23, 2012, 12:38

    What makes Alph Cent A and B even more fascinating is the fact that it is a *solar* binary (ignoring distanc component C), i.e. both components are solar type stars.
    That is very rare: the only other examples of such solar binaries (or multiples) that I could find in our galactic neighborhood (within about 70 ly) are Zeta 1 and 2 Reticuli (resp. G2 and G1) at almost 40 ly, and 16 Cygni A and B (resp. G1.5 and G2.5) at about 71 ly. Both are (very) wide binaries/triples (16 Cygni A is also an intermediate binary with an M dwarf, C).
    But 16 Cygni A and B are much older than our sun, estimated at least 8 gy and probably around 10 gy. Although their mass is around solar and their metallicity slightly higher (1.1-1.2 * solar), their luminosity is resp. 1.6 and 1.3 times solar.
    Interestingly, 16 Cygni B has a super-Jupiter in the HZ (just, toward the outer edge), habitable moons and so…

  • kzb February 23, 2012, 14:21

    Ronald and Andy:
    Is it conceivable that the apparent high metallicity of the AC system (and maybe others Ronald mentioned) is actually due to swallowed hot Jupiters adding their metals to the outer layers of the star?

    I also wonder if planets could’ve been scattered into very wide orbits encompassing both A and B. I’m not sure if this is dynamically feasible or not. If it is feasible, I wonder if the orbits are wide enough to resolve any large planets separately to the stars?

  • andy February 23, 2012, 14:53

    Latest age estimate for 16 Cygni from asteroseismology using Kepler photometry gives around 6.8 Gyr (reference).

    What are you counting as solar-type stars? Depending on your definition there may well be a few other stars that meet the requirements. Xi Ursae Majoris is a multiple consisting of two G-type stars, each of which has a lower-mass companion. Xi UMa Ab is an M-dwarf, while the nature of Xi UMa Bb has not yet been conclusively determined: it may be a K-dwarf, and there may be additional stars in the Xi UMa B subsystem though this is far from certain.

  • Ronald February 26, 2012, 19:24

    @andy: “Latest age estimate for 16 Cygni (…) gives around 6.8 Gyr”.
    Ok, even so, it is on the old side for a G2 star, which could (partly) explain its high luminosity.
    The giant planet near 16 Cygni B is at least 1.7 Mj, possibly much more and its orbit is highly elliptical (eccentricity almost 0.7), which would disrupt the orbit of any planet outside 0.3 AU. So, with regard to habitable planets there, it is either a large habitable moon or nothing.

    Xi Ursae Majoris (also known as Alula Australis): you are right that this (A and B) is also a solar couple (G0). However, Aa is a close binary with Ab,an M dwarf, with a separation varying from 0.8 to 2.6 AU. Ba is also a very close binary with a very weak M dwarf or brown dwarf.
    The two pairs Aab and Bab have a very elliptic separation varying from 12.5 to 40 AU, which is aggravated (especially with a view to circumbinary planetary orbits) by the fact that their orbits are not nearly coplanar.
    In other words, I would not readily expect stable planetary systems there.