Although I think most astronomers have assumed Proxima Centauri was bound to the central binary at Alpha Centauri, the case wasn’t definitively made until fairly recently. Here we turn to Pierre Kervella (Observatoire de Paris), Frédéric Thévenin (Côte d’Azur Observatory) and Christophe Lovis (Observatoire Astronomique de l’Université de Genève). We last saw Dr. Kervella with reference to a paper on aerographite as a sail material, but his work has appeared frequently in these pages, analyzing mission trajectories and studying the Alpha Centauri system. Here he and his colleagues use HARPS spectrographic data to demonstrate that we have at Centauri a single gravitationally bound triple system. This is important stuff; let me quote the paper on this work to explain why (italics mine):
Although statistical considerations are usually invoked to justify that Proxima is probably in a bound state, solid proof from dynamical arguments using astrometric and radial velocity (RV) measurements have never been obtained at a sufficient statistical significance level. As discussed by Worth & Sigurdsson (2016), if Proxima is indeed bound, its presence may have impacted planet formation around the main binary system.
This is a six-year old paper, but I want to return to it now because a new paper from the same team will tighten up its conclusions and slightly alter some of them. We’ve gone from resolving whether Proxima is bound to the A/B binary to pondering the issues involved in the dynamical history of this complex system. That in turn can inform the ongoing search for planets around Centauri A and B at least in terms of explaining what we might find there and how these two systems evolved. The original paper on this work lays out the challenges involved in tracing the orbit of the red dwarf. For HARPS is exquisitely sensitive to the Doppler shifts of starlight, and these data, obtained between 2004 and 2016, contain potential booby traps for analysis.
Image: Pierre Kervella, of the Observatoire de Paris/PSL.
Convective blueshift is one of these. We’re looking at the star’s spectral lines as we calculate its motion, and some of these are displaced toward the blue end of the spectrum because of the structure of its surface convection patterns. The lifting and sinking of hot internal gases has to be factored into the analysis and its effect nulled out. The spectral lines are displaced toward the blue, in effect a negative radial velocity shift, although the effect is stronger for hotter stars. In the case of Proxima, Kervella’s team finds a relatively small convective blueshift, though still one to be accounted for.
A similar though more significant issue is gravitational redshift, which occurs as photons climb out of the star’s gravity well. Here the effect is “an important source of uncertainty on the RV of Proxima” whose value can be established and corrected. How the astronomers went about making these corrections is laid out in a discussion of radial velocities that aspiring exoplanet hunters will want to read.
Image: Orbital plot of Proxima showing its position with respect to Alpha Centauri over the coming millenia (graduations in thousands of years). The large number of background stars is due to the fact that Proxima is located very close to the plane of the Milky Way. Credit: P. Kervella/ESO/Digitized Sky Survey 2/Davide De Martin/Mahdi Zamani.
Out of all this we learn that Proxima’s elliptical orbit around Centauri A and B’s barycenter extends from 800 billion kilometers when closest (periastron) to 1.9 trillion kilometers at apastron, its farthest distance, with an orbital period of approximately 550,000 years. The orbital phase is currently closest to apastron.
The Astronomy & Astrophysics site (this is the journal in which the paper above appeared) is currently down, so I’m quoting from the version of the paper on arXiv, which after noting that the escape velocity of Alpha Centauri at Proxima’s distance (545 +/- 11 m/s) is about twice as large as Proxima’s measured velocity, goes on to speculate in an intriguing way:
Proxima could have played a role in the formation and evolution of its planet (Anglada-Escudé et al. 2016). Conversely, it may also have influenced circumbinary planet formation around αCen (Worth & Sigurdsson 2016). A speculative scenario is that Proxima b formed as a distant circumbinary planet of the αCen pair, and was subsequently captured by Proxima. Proxima b could then be an ocean planet resulting from the meltdown of an icy body (Brugger et al. 2016). This would also mean that Proxima b may not have been located in the habitable zone (Ribas et al. 2016) for as long as the age of the αCen system (5 to 7 Ga; Miglio & Montalbán 2005; Eggenberger et al. 2004; Kervella et al. 2003; Thévenin et al. 2002).
The idea of Proxima b as a captured planet has not to my knowledge appeared anywhere else in the literature. I was fascinated, enough so that I dashed off a quick email to Dr. Kervella asking about this as well as the current status of the orbital calculations. And indeed, his response indicates new work in progress:
… we identified a mistake in our 2017 determination of the orbital parameters of Proxima. In the papier, they are expressed in the Galactic coordinate system, and the orbital inclination is thus not directly comparable to that of the Alpha Cen AB orbit. We are preparing a new publication with revised orbits and parameters for all three stars. The main difference is that now the orbital plane of Proxima is better aligned with that of AB. The gravitationally bound nature of Proxima with Alpha Cen AB is also strengthened, as we include new astrometry and radial velocities.
I’ll cover the new paper as soon as it appears. Dr. Kervella also observes that confirming a scenario of Proxima b as a captured planet would be difficult (Proxima b has ‘forgotten’ the history of its orbital evolution, as he puts it), meaning that working with astrometric data alone will not be sufficient. But the arrival of telescopes like the Extremely Large Telescope, now under construction in Chile’s Atacama Desert with first light planned for 2028, should signal a treasure trove of new information. A spectrum obtained by ELT could show us whether Proxima b is indeed an ocean planet.
The paper on Proxima Centauri’s orbit is Kervella, Thévenin & Lovis, “Proxima’s orbit around α Centauri,” Astronomy & Astrophysics Vol. 598 (February 2017), L7 (abstract/preprint).
Hope that we can convey to the authors an appreciation for an interesting and, in some ways, an unusual study. Since Alpha and Proxima are so much in the foreground, the properties of the analysis do not look like the usual textbook example.
For one thing, thing the paper notes an orbital rate on order of a half a kilometer per second. And the Proxima orbital path projected to the terrestrial observer is largely overhead; hence, a lot of the info obtained is astrometric.
But since the campaign for radial velocity appears to have resorted to compensating for surface convection and relativistic red shift effects ( red dwarfs, being concentrated or dense, though not as much as white dwarfs, there will be a strong surface gravity), the how and why of the radial velocity extraction becomes an interesting report in itself. It in effect orients Proxima’s orbit in space.
Just in passing, the eccentricity of A and B is about the same as Proxima’s. That could be an effect brought on by the 3 body dynamics settling into a somewhat stable orientation. Maybe early on, say in formation, but Alpha Cen A and have a period of only 80 years. Could it be that in an earlier age, this system was more closely bound? When you consider the other alternatives (as is now or Proxima captured from space), I would hazard to say, Yes.
Another matter: We presume we have in our solar system a well populated Oort Cloud out to tens of thousands of AUs. Though in principle, one expects Oort Clouds elsewhere, and we have studies of circumstellar disks and proto planets in early days, a star here and there which have dust and and planets… But I haven’t seen anything about analogs to an Oort Cloud or Kuiper Belt with regard to our nearest neighbor(s). All three bodies A, B and C should have a distinctly different interaction with such structures, but that does not necessarily mean that all that material would be swept away. If so, we would not have any positive IDs on planets – or suspects.
Reflecting on the nature of the 3 body system and the history of what could have been a similar Oort Cloud to that surrounding Sol, here is one possibility:
If the binary Alpha Cen A and B had a cumulative mass loss over aeons, then Proxima could have slowly spiraled out. Why the two binary stars would have ejected mass from their surfaces, I have no idea as of now, but if the circumstellar disks surrounding them were large and interacted – that could have been an initial separating mechanism.
In any case, we would have discovered Jovian mass planets surrounding A and B by now and planets at Jovian distances from the two would be unstable. Proxima having a planet is one reason not to give up hope for Alpha Centauri exoplanets of a terrestrial nature. If one or more exists with reasonable offset from its primary, its nature will likely be distinctively different from Earth, but it will certainly be an interesting case. Maybe even interesting enough to visit.
Here was my synopsis of Kervella, Thévenin & Lovis preprint from seven years ago (has it been that long already?): https://www.drewexmachina.com/2016/11/29/the-orbit-of-proxima-centauri/
I had the same reaction. It really does seem like yesterday!
Interesting concept, here is a article that goes into a little more detail on binary or double star systems and their ability to eject planets.
Tilted Tatooines can serve as a source of strange, solo planets.
https://astrobites.org/2023/11/01/tilted-planets-ejected-form-free-floating-rogues/
Just wondering if New horizon could use parallax to get a more exact position of the stars with its camera. Just wait for a flare and then triangulate with ground based telescopes to get the data.
Do we know yet whether the three body system is a stable one without doubt?
Does it matter practically to us?
If we ever interacted with the system in any way it would be on time scales irrelevant to the orbital period of Proxima.
Even if it is stable now, will it remain stable for a full orbit of Proxima? At half a million years, the orbital period is so long that it seems possible, if not likely, that another body would enter the equation and disrupt or alter equilibrium prior to a complete orbit of Proxima.
All naïve thoughts from a non qualified enthusiast, so thanks for the understanding and responses in advance.
Proxima Centauri is a fifth a light year from Alpha Centauri which is one trillion miles or 13,000 AU. It has planets because there is no problem with tidal forces like with Alpha Centauri.
In his book One Hundred Billion Suns, Rudolf Kippenhahn writes that double star systems which are at least a Sun Saturn distance apart don’t have any planets because the two stars grab all the angular momentum and there is none left for planets to form an accretion disk. All that forms is a ring of gas with stars at opposite ends. The center of the ring has no gas and therefore no accretion disk. This idea or theory was tested with a computer simulation in 1983 the year of that book. Although the computing power was much less and today, the idea is still valid and a testable hypothesis. There may be some accretion for each star, but that would have to be close to each star. Consequently, tidal forces might prevent any planets from forming in such star systems. Ibid. We still have to prove this hypothesis and make a thorough search. There is a range limit for such systems because if the stars are far enough apart, then there will be no tidal problems. Double star systems where the stars are really close to each other done have that problem because they act as one central mass instead of competing masses for gas, dust, etc. so they can form an accretion disk around them. Planets have already been discovered around such star systems.
I’m afraid that One Hundred Billion Suns is far outdated. I have a copy here and remember reading it when it came out. The existence of stable orbits in systems like Alpha Centauri AB has been fully modeled and accepted since the late 1990s (Wiegert and Holman did the key work), so I would not be surprised to find planets around both Centauri A and B, limited to orbital distances not a great deal further than that of Mars. But that leaves plenty of room in the habitable zone of each star.
Paul Gilster it’s not the stable orbits that are invalid but the impossibility of planets forming in that type of star system in the first place. Even if they did, the stable orbits will be near the stars and the number of planets would definitely be limited. Ideas based on First principles are never outdated. Many of the ideas he uses in that book are still valid today which at first surprised me when I re read it forty years later with a much more developed intuition in physics and science than when I was a young man. It’s a great book by a Brilliant astrophysicist. I’m glad I read it. The computers are outdated, but not that book. A theory is potentially true until is it proved invalid and we have yet to prove that planets exist around Alpha Centauri. I hope they do as I like the Tatooine situation.
It would be interesting if the the stars rotate in the same direction which would cause the planet making discs to interfere with each others formation reducing the possibility of planets due to collisions.
Looking at the track of Proxima about the two more massive elements of the Centauri system, I thought I would check to see what the inclination of AB was to the observer. I did notice it in Wikipedia entry, so I went back to Burnham’s Celestial handbook where I thought I saw it last. And it was a surprise, since as observed the ellipse of 0.52 eccentricity is very narrow, “tilted about 11 degrees to the edge on position”. Thus Alpha and Proxima, though they share much the same eccentricity do not share the same plane. But if one or the other has an associated Oort Cloud, they would be in a cross fire with each other at intersections. At the distance that Alpha and Proxima are from each other, there might still be a precession mechanism. But more likely, if they were closer together early on, they might not have been in the same plane then – or else Proxima was knocked out of it. Early F and A stars illuminate vast accretion disks in comparison to the current day sun, say with the Kuiper Belt. So there are still some worthwhile areas of investigation about resources for planet forming. After all, A and B mass would add up to an F…