New Study: Centauri B Planets Unlikely

by Paul Gilster on November 6, 2008

Roughly twenty percent of all detected exoplanets are in binary systems, intensifying our interest in Alpha Centauri. Recent work, however, has been less than encouraging to those hoping to find one or more terrestrial worlds around these stars. Indeed, Philippen Thébault (Stockholm Observatory), Francesco Marzari (University of Padova) and Hans Scholl (Observatoire de la Côte d’Azur) have shown that in the case of Centauri A, the zone beyond 0.5 AU is hostile to the accretion processes that allow planets to form. Any terrestrial-class world that close to Centauri A would be excluded from the habitable zone, a region thought to extend from 1.0 to 1.3 AU around the star.

The same team now goes to work on Centauri B, having pointed out in the earlier paper that the mathematical modeling it used there was unique to Centauri A and could not be applied indiscriminately to other systems, not even to the second star of the Centauri binary. The authors are targeting the phase of planetary formation when kilometer-sized planetesimals accrete, and examining the effect that perturbations and gas drag have on impact velocities within a test population of such planetesimals. As with Centauri A, the results show that this early formation stage can take place only in a region within 0.5 AU.

That just might allow the needed accretion and hence planet formation to occur within the innermost part of Centauri B’s habitable zone, which is estimated to extend from 0.5 to 0.9 AU. But even here the limits are tight, and the authors believe that because planetesimals in this region would have such high relative velocities (as opposed to a system unperturbed by the close binary Centauri A), planet growth would tend to occur much more slowly.

But let me quote the paper directly on this crucial matter:

Planetesimal accretion is marginally possible in the innermost parts, ∼0.5 AU, of the estimated habitable zone. Beyond this point, high collision velocities, induced by the coupling between gas friction and secular perturbations, lead to destructive impacts. Moreover, even in the ∼ 0.5 AU region, ∆v are increased compared to an unperturbed case. Thus, ”classical”, single-star like runaway accretion seems to be ruled out.

The case is strong — in order to arrive at the subsequent planet, you first have to let accretion go to work to build it, whether in or out of the habitable zone. This work does not, then, contradict other findings that planets are feasible within the habitable zones of the Centauri stars, but does insist that they are highly unlikely to form there. So are there other ways we could wind up with planets in the habitable zone?

Two possibilities remain: Changes in planetesimal orbits as gas within the early system is dispersed, and greater separation between the two stars early in their history, which would make perturbations less pronounced. Can either leave us hope for such worlds? The first seems dubious:

…we find that later progressive gas dispersal reduces all ∆v to values that might allow accreting impacts. However, we find that the system has first to undergo a long accretion-hostile transition period during which most of the smaller planetesimals are removed by inward drift and most bigger objects are probably fragmented into small debris. Thus, the positive effect on planetesimal accretion is probably limited.

As to a wider initial separation of the Centauri stars reducing the perturbation effects, much work remains to be done. The minimum separation for accretion processes to prove favorable to the formation of a planet in the habitable zone seems to be 37 AU. Could the Centauri stars have once been separated by this amount or greater? Finding the answer to that question depends upon working out the likelihood of orbital changes in early open clusters. In short, we don’t know.

Thébault, Marzari and Scholl are doing significant work, examining as they do the crucial early stages of planet formation, and the authors point out that the even earlier phases, when the kilometer-sized planetesimals are themselves forming, have not yet been modeled. Right now we can hold out hope for the region 0.5 AU or so around Centauri B, but it is chastening to reflect that this narrow region may offer the only realistic prospect for a habitable world in this nearby system.

The paper is Thébault et al., “Planet formation in the habitable zone of alpha Centauri B,” accepted for publication in Monthly Notices of the Royal Astronomical Society and available online. Thanks to andy for the pointer.

andy November 6, 2008 at 16:56

This kind of thing is exactly why proposals for interstellar missions — especially crewed missions — should be designed to handle distances further than the 4.3 light years to the Alpha Centauri system: it may well be that it is not a suitable target. Judging by the content of our local stellar neighbourhood, such missions should be set up to be able to deal with distances in the 10-20 light year range to provide a suitable selection of target stars.

tacitus November 6, 2008 at 17:53

This kind of thing is exactly why proposals for interstellar missions — especially crewed missions — should be designed to handle distances further than the 4.3 light years to the Alpha Centauri system: it may well be that it is not a suitable target.

Well, yes and no. Unless it’s a manned mission, there are still some compelling reasons for picking the Centauri system. First it’s the closest, and scientists are an impatient bunch, even if the durations are decades in length, and doubling or tripling the length of the mission greatly increase the overall risk of failure.

And from a scientific standpoint, the system has three different stars to study close up, plus the possibility of planets (though that could be ruled out by the time a mission is ready to launch).

I suspect that unless we find a surefire Earth analog with evidence for life within a dozen light years, the first interstellar mission will almost certainly to the Centauri system. We probably won’t even wait for that one to complete before other star systems are targeted, but it will most likely be the first.

Thomas November 6, 2008 at 18:48

I agree 100% with Andy’s post. We also need to know exactly what’s there before we go. I don’t expect any interstellar missions to be launched in this century, but if any are, I would expect by then that whatever their target, we’ll know by then if there are any planets, what kind of planets, etc.

Ron S November 6, 2008 at 21:30

When the similar discussion about ACen A came up some months ago, I wondered at the time about the long-term stability of the binary orbital parameters since any change in eccentricity could radically change the possibility of planet formation and subsequent stability. So I did a search for papers in arXiv on modeling binary systems and their planets. There were lots and I read through about a dozen of what seemed the most pertinent. Unfortunately every paper simply assumed the orbits of the stars was unchanging for all time. Worse (for me), no justifications were given.

Nevertheless maybe this is reasonable or at least highly probable. As Paul mentioned, in the critical early years there would be other neighboring stars in an open cluster (stellar systems tend to form in groups). As they disperse, each would have one chance to perturb the orbits of any binaries in the cluster. Timing and trajectory are critical. Afterward, chance encounters with other stars, especially ‘effective’ encounters should be rare.

Paul Titze November 6, 2008 at 22:14

“Simply having the capability to launch a high performance extrasolar probe—
actually a prototype starship – is not enough. For such a mission to be conducted
in the foreseeable future, there must be scientific justifications.”

http://arxiv.org/abs/0809.3535

Any interstellar mission is a major undertaking which will require lots of resources, time and money. Since there are 228 exoplanets found so far which one do we send a mission? One could argue that spending the money to send a probe or starship to any one of them will only be justified if the exoplanet shows Earth like properties.

Cheers, Paul.

daniel November 6, 2008 at 22:16

before any interstellar missions probabily they will get know by the new generation of planet hunter telescopes like TPF ,if there its any planet and after on future send interstellar probe after that if there are any interest planet, interstellar crew missions will be laurch

david lewis November 6, 2008 at 22:21

By the time we are ready for an interstellar mission for colonization purposes the presence of an habitable planet won’t matter. It would be interesting so that we could compare life there to life on earth, but it won’t affect the success or failure of the mission. All that will be needed is enough raw materials for space habitats.

Adam November 7, 2008 at 2:51

The study assumes a certain cosmogony which is yet to be proven as the only mode of planet formation. There may well be many modes of planet-making.

Ronald November 7, 2008 at 6:43

The referenced article mentions ‘a starting separation roughly 15 AU wider’ as a minimum to allow for accretion of a planet in the habitable zone (not 37).

Even so, the possibilities for that now seem to become very much like wishful thinking. The near future will tell.

What are the next realistic targets, i.e. suitable solar type stars? Tau Ceti at almost 12 ly is quite metal-poor and seems to have a failed planetary system, more like a huge and extended asteroid belt.

Epsilon Eridani at almost 11 ly has a remarkably solarsystem like system (as far as we know), but is a rather dim K2 star and quite young.

40 (= Omicron) Eridani at about 16ly: rather dim K1, not very metallic.

Sigma Draconis, almost 19 ly: variable (and not very metallic).

Eta Cassiopeia A, just over 19 ly: also rather close binary (36 AU minimum), and again not very metallic.

Delta Pavonis, almost 20 ly: very (too?) high metallicity, leaving the main sequence, i.e. becoming subgiant and variable.

My bets are that the most promising candidates for earthlike planets, either living or terraformable, within about 10 parsec (or about 33 ly), are: 82 Eridani (though metallicity estimates vary greatly), Beta Canum Venaticorum, 61 Virginis, and Alpha Mensae; maybe also Zeta Tucanae and Beta Comae Berenices.
Four of these six (that is, except the first and last one) plus 18 Scorpii at 46 ly, were also identified by Porto de Mello and others as the most promising candidates for (habitable) earthlike planets.

Ronald November 7, 2008 at 6:48

Addendum: most promising candidates within about 50 ly, that is. I would like to add the Zeta Reticuli couple to that little list, but metallicity and age estimates vary greatly.

djlactin November 7, 2008 at 8:11

w.w.w.wait! ‘habitable’ = ‘habitable to humans who live on the surface’… why should we only be interested in such bodies. curiosity has sent probes to titan and mercury. surely IF planets are discovered around ACenA or AcenB, that would be sufficient to motivate sending a probe.

Ron Hager November 7, 2008 at 8:59

Below is a link to a famous rant by Jim Mora when a sports reporter asked him a stupid question. That is my reaction when someone begins talking about a manned mission to the stars as if it were a reality, just substitute the words “Manned Mission” every time he says the word “Playoffs”. Hell, I just want to get to Mars and back in the next hundred years without losing anyone.
http://www.youtube.com/watch?v=Qwq7BYOnDrM

ljk November 7, 2008 at 10:57

What are the closest objects that radiate in the infrared or other
non-optical areas only?

I bet there are some brown dwarfs (and black holes?) closer to
Sol than the Alpha Centauri system. And we do know of some
BDs that have their own planets.

And maybe some of those BDs aren’t brown dwarfs.

ljk November 7, 2008 at 11:06

As for interstellar missions, we may also have to get out of the
paradigm of the big one-shot starship ala Daedalus.

As we continue to build smaller and smaller computers with more
power, storage, and processing capacity, we may be better served
by sending out swarms of little ships such as have been described
elsewhere on CD to explore a multitude of star systems.

The little probes can be programmed to focus on specific areas
and if some get lost during the mission, it won’t end the whole
enterprise.

I am sure there are plenty of hurdles to get over yet with sending
probes to other star systems, but one area that might solve a lot
of issues is to get past the notion of one big star probe or an even
bigger vessel full of hundreds of humans.

John Hunt November 7, 2008 at 13:19

If we develop the infrastructure to launch one interstellar mission, it will cost much less to launch a second and a third and a fourth… If we therefore have the ability to launch multiple interstellar missions then I would have to think that we would choose to send one of those to the A.C. system simply because we can get there the earliest.

David Lewis makes an excellent point:

All that will be needed is enough raw materials for space habitats.

That criteria is really very low. The question would then not be, is there an Earth-like planet (i.e. warm oceans and continents) in the A.C. system but rather can you get asteroid and cometary material together and construct a space habitat and start spinning it?

Going another step, why should anyone care about the habitable zone? Mars is outside the habitable zone but we’re certainly going to be inhabiting it. If we really had to, we could probably establish a manned base on Mercury in a crater near its ice caps. This means that certain planets with water ice now becomes a “habitable” planet including those well within the “habitable zone”.

andy November 7, 2008 at 15:57

Epsilon Eridani has several advantages over Alpha Centauri as a destination: for a start there is a known planetary system (the dust discs imply there is matter floating around in nice manageable chunks rather than being forced to dive down deep gravity wells and deal with 6000-degree hydrogen plasma), and secondly it is closer to the ecliptic, which lowers the energy requirements. The fact it is a single star presumably makes figuring out a good trajectory into the system somewhat easier as well.

philw1776 November 7, 2008 at 16:21

Adam Says: “The study assumes a certain cosmogony which is yet to be proven as the only mode of planet formation. There may well be many modes of planet-making.”

Excellent observation. While I respect their work and have no evidence that they’re wrong, it’s hubris of the highest order to claim we REALLY understand planetary formation today. I do not mean to discourage such attempts at simulation however, just do not accept them as certain fact.

andy November 7, 2008 at 17:56

Sure it is just a simulation and we don’t have the observations to support or refute it. Then again it is also misleading to say that trying to make predictions is “hubris”: while no-one is saying we completely understand planet formation, it is also incorrect to say we have no idea what is going on and the models are valueless and divorced from reality. From my reading of the literature, there does seem to be significant progress in reproducing the characteristics of the known exoplanet population with the computer models.

(As an aside, it’s interesting that the fact these are all just models was not raised when the models predicted habitable worlds could exist in hot Jupiter systems, which I guess is a conclusion people like, but did get raised when the models predict things we don’t like, e.g. poor prospects for habitable planets around the Alpha Centauri stars, or that solar system-type architectures may be rare…)

What this should highlight is that pinning interstellar hopes on Alpha Centauri having a useful set of planets and doing engineering that will only get us across 4.3 light years is not the safest of assumptions. I doubt there’d be much public support for an extremely expensive mission to a system devoid of planets. Sure the scientists would like it, but I doubt the public would be particularly impressed with something that would look qualitatively similar to our own Sun.

John Hunt November 7, 2008 at 19:18

If we really had to, we could probably establish a manned base on Mercury in a crater near its ice caps. This means that certain planets with water ice now becomes a “habitable” planet including those well within the “habitable zone”.

Actually I meant to say “inside of the habitable zone” meaning like Mercury. My point is that even if Alpha Centauri A or B only have small planets up to 0.5 AU, so long as we can calculate that they should have a polar ice caps then they become candidates for human colonization and hence a good candidate for the first interstellar mission (science probe or “manned”).

Planets with warm oceans would, of course, be great candidates to look for exobiology. But, by the time our probe arrived at such distances we’ll probably be able to analyze for atmospheric signatures of life out to much greater distances. Exobiology would already likely be detected although we would naturally want to get a close-up look.

Brian Schmidt November 7, 2008 at 21:01

While this study rules out planetisimal migration forming planets, what about planetary migration? A near miss between two planets, both orbiting <.5 AU, could send one of them further out.

Randy McDonald November 7, 2008 at 23:49

“My point is that even if Alpha Centauri A or B only have small planets up to 0.5 AU, so long as we can calculate that they should have a polar ice caps.”

You mean ice caps like the ones theorized for Mercury and the Moon? Um.

If close-orbiting terrestrial worlds like Mercury and the Moon exist, I can imagine them being objects of study, but objects of habitation?

“82 Eridani (though metallicity estimates vary greatly), Beta Canum Venaticorum, 61 Virginis, and Alpha Mensae; maybe also Zeta Tucanae and Beta Comae Berenices.
Four of these six (that is, except the first and last one) plus 18 Scorpii at 46 ly, were also identified by Porto de Mello and others as the most promising candidates for (habitable) earthlike planets.”

82 Eridani and Zeta Tucanae are heading towards the sub-giant stage as well. I’m not sure, though, that a star being a subgiant necessarily means it couldn’t have a garden world. What if a world was a glacier planet for most of the system’s existence until its sun started to heat up? Super-Mars could do well in this scenario.

Beta Canum Venaticorum and Beta Comae Berenices are both Sun-like stars. They’re actually quite close to each other in the bargain.

andy November 8, 2008 at 12:10

Another thing to point out here is that 0.5 AU for the inner edge of the habitable zone around Alpha Centauri B is something of a stretch. A planet located at this distance would receive about the same amount of radiation from the star than Venus (a planet which is noted for its pleasant surface conditions and verdant biosphere) does from the Sun. To receive Earthlike amounts of radiation, the planet would have to be located at roughly 0.7 AU.

Ronald November 8, 2008 at 20:38

@Randy McDonald:
“82 Eridani and Zeta Tucanae are heading towards the sub-giant stage as well. I’m not sure, though, that a star being a subgiant necessarily means it couldn’t have a garden world. What if a world was a glacier planet for most of the system’s existence until its sun started to heat up? Super-Mars could do well in this scenario.”

I suppose that depends on the rate of change (increasing insolation and heating) during this sub-giant process .

“Beta Canum Venaticorum and Beta Comae Berenices are both Sun-like stars. They’re actually quite close to each other in the bargain.”

But Beta Comae Berenices is quite a bit brighter: luminosity 1.42 – 1.49 times solar, against about 1.20 – 1.22 for Beta Canum Venaticorum.

spaceman November 9, 2008 at 2:50

UNLIKELY TO BE THE LAST WORD, LET’S NOT THROW IN THE TOWEL YET

I still think these astronomers need to be careful about the types of environments in which they say planets cannot be found. If we have learned anything in the young field that is extrasolar planet scientist it is this: planets are not only abundant, but they form in a greater diversity of environments than anyone would have thought possible prior to the 1990s. The Kepler mission website puts it quite well:

“The formation of stars and planets is complex, making it almost impossible to predict the diversity of planetary systems from first principles (Boss, 1995, Lissauer 1995).”

And yet, studies like the one mentioned in this paper do precisely that—they are based on first principles and they try to predict the limits of planetary system diversity. If you asked an astronomer 30 years ago if he or she thought that gas gaints would be found in torch orbits, or if a planet could exist in the Gamma-Cephei system, or if planets could form around pulsars or brown dwarfs they would probably laughed at you with incredulity for even asking such a question. Again, if there is one thing that we have learned in this golden age of exoplanet science it is this: the diversity of planetary systems is greater than anyone would have predicted using studies like yours. So, I don’t think it will be the final word on the issue of terrestrial planets being able to form at ~1 A.U. around either Alpha Centauri A or B. counter-examples exist not for your study specifically, but for other previous pessimistic assessments of the prospects of planet formation around close binary systems.

Ronald November 9, 2008 at 18:29

Further to the sub-giant issue raised yesterday (how suitable are sunlike stars moving off the main-sequence to subgiant stage for biological life?), when I mentioned in response to Randy McDonald: “I suppose that depends on the rate of change (increasing insolation and heating) during this sub-giant process “.

I realize now that there are actually two issues here at the same time:

- one is how suitable an earthlike planet could be for colonization *now*. Indeed, it is quite well possible, that planets that were once too cold, are becoming balmy as their mother star gets brighter and hotter and hence the habitable zone moves outward to encompass them, as some time in the distant future will happen to Mars (even without human terraforming intervention). This stage of ‘secondary’ habitability could still last for long periods of time (millions of years?), interesting enough to justify attempts to interstellar travel and settlement. Randy mentioned 82 Eridani and Zeta Tucanae as examples, Delta Pavonis is another well-known one.

- the other issue, however, is whether the passing and dynamic sub-giant stage will be long ánd stable enough to wake up a dead-frozen world and allow life to arise and develop. For this, even some millions of years would be a very brief period, hardly allowing for even the most primordial lifeforms. Main question regarding this is therefore how long such an earthlike planet would remain in the continuously habitable zone of its star.

andy November 9, 2008 at 19:47

Ronald, regarding the habitability of subgiant and giant stars, you may find this paper interesting: Lopez, Schneider and Danchi (2005) “Can Life Develop in the Expanded Habitable Zones around Red Giant Stars?”

Ronald November 10, 2008 at 6:35

Thanks, andy, interesting indeed!

The authors conclude, that there might be time enough for the development of at least primitive life on a planet passing through the expanding HZ of a subgiant;

“For a 1 Msol star at the first stages of its post–main-sequence evolution, the temporal transit of the habitable zone is estimated to be several times 10^9 yr at 2 AU and around 10^8 yr at 9 AU. Under these circumstances life could develop at distances in the range 2–9 AU in the environment of subgiant or giant stars, and in the far distant future in the environment of our own solar system.”

In an included table of class IV stars within 30 pc, Delta Pavonis is specifically mentioned, however, Zeta Tucanae and 82 Eridani are not (maybe very early stages, still considered class V).

Encouraging.

Tim G November 11, 2008 at 7:58

Tidal strength (the rate at which gravity changes) is proportional to the stars mass and inversely proportional to the cube of the distance from the star. Therefore tidal strength would be more than seven times as strong at 0.5 AU around B than at 1 AU around our sun. Tides cause Mercury and Venus to have rotations that are in resonance with their year, resulting in extremely long days. I would suspect a planet at 0.5 AU from B to have such a resonance if not be tidally locked.

John Hunt November 12, 2008 at 0:26

Randy & Ronald mention some specific stars which are promising candidates for ‘habitable’ earthlike planets. When one is looking for exobiology, of course, you are excited to find a planet in the ‘habitable zone”. But for an interstellar mission we are likely going to send our first craft to whatever system is the closest which has any planets. Here’s why.

Imagine that we can get a craft up to 0.1c. Here’s how long it would take to reach some of those promising candidates:
  200 yrs – 82 Eridani
  270 yrs – Beta Canum Venaticorum
  280 yrs – 61 Virginis
  330 yrs – Alpha Mensae
  280 yrs – Zeta Tucanae
  300 yrs – Beta Comae Berenices
  460 yrs – 18 Scorpii

But what if one of the cleared zones around Epsilon Eridani had a planet but which was calculated to have frozen instead of liquid water? Or perhaps there were multiple planets including a Jupiter equivalent which likely had many moons?

Here’s the choice:
  A) 105 years to Epsilon Eridani or
  B) 200-460 years to an Earth-like planet.

If science return is more than 200 years away then it is probably best to wait 50 or so years until we have a craft which can travel at 0.2c. But if we target a system only 105 years away then there’s really no benefit to waiting 50 or so years until we build a craft that can travel at twice the speed.

Of course we’ll eventually send probes and even humans out to 20+ light-years. But for the first craft we’ll likely choose one of the solar systems closest to us.

Ronald November 12, 2008 at 5:14

Tim: are you sure it is cube of distance, not square (as in Newton’s law of gravity)?
And is the slow rotation of Venus really a result of (near-)tidal locking, I mean at that distance of 0.7 AU?

andy November 12, 2008 at 9:30

Ronald: yes, tidal forces follow an inverse cube law: they are caused by different gravitational forces felt at different points on an object, so they depend on the differential of the gravitational force. Since the result of differentiating 1/r^2 with respect to r is proportional to 1/r^3, the tidal force follows an inverse cube law.

andy November 18, 2008 at 19:34

Greg Laughlin’s posted his view about these results over at systemic.

Comments on this entry are closed.

{ 2 trackbacks }