≡ Menu

HARPS: Hunting for Nearby Earth-like Planets

Ever more refined radial velocity searches for exoplanets are reaching into the domain of lower and lower mass targets. It’s natural enough that we’re most interested in planets of Earth mass and even smaller, but as a new paper on the work of the European Southern Observatory’s HARPS instrument reminds us, one of the great values of this work is that we’re getting a broad view of how exoplanets form and evolve in their systems, no matter what their size. Characterizing not just planets but entire systems is becoming a profitable investigation.

But small worlds continue to fascinate us, particularly in the hopes of finding possible abodes for life. HARPS’ involvement in the hunt now includes an intense campaign to monitor ten stars that are relatively near our Sun, all of them slowly rotating and quiet solar-type stars. Mounted on ESO’s 3.6-meter instrument at La Silla Observatory in Chile, HARPS (High Accuracy Radial Velocity Planet Searcher) has produced more than 100 exoplanet candidates in its first eight years of operation, including not just Neptune-mass planets but super-Earths and intriguing systems like Gliese 581, with two possibly rocky planets near the habitable zone.

Moreover, from the system-wide point of view, the system around HD 10180 includes seven low-mass planets including the 1.5 Earth mass HD 10180 b. So when HARPS talks, we listen, and I want to quote this from the paper at the outset (internal references omitted for brevity):

… a recent investigation of the HARPS high-precision sample has shown that about 1/3 of all sample stars exhibit RV variations indicating the presence of super-Earths or ice giants… Indeed, planet formation models… show that only a small fraction (of the order of 10%) of all existing embryos will be able to grow and become giant planets. Hence, we expect that the majority of solar-type stars will be surrounded by low-mass planets.

Good news for small planets! If this is the case, we would expect that even a small sample like the current ten solar-type stars now under intense investigation by HARPS will turn up several Earth-like planets (i.e., rocky worlds in the inner system), and the new paper does not let us down. Three of the host stars involved in this program have already produced detections; these are HD 20794, HD 85512 and HD 192310. There are no giant planets here but study of the three stars has thus far yielded six low-mass planets, including three super-Earths around HD 20794 (82 Eridani), with semi-major axes of the planetary orbits measured as 0.12AU, 0.20AU and 0.35AU. The semi-major axis measures the radius of an orbit taken at the orbit’s two most distant points.

No habitable zone planets here, though, with even the furthermost planet reaching likely equilibrium temperatures of 388 K, which works out to about 115 degrees Celsius. Remember that equilibrium temperature is not the same thing as temperature at the surface. The equilibrium temperature of the Earth without an atmosphere is 255 K ( -18 degrees Celsius), but adding in the various effects of our atmosphere we come to an average of 288 K (15 degrees Celsius), so it’s clear how careful we have to be with these numbers, given how little we know about the planets in question. The surface temperature of a planet with a dense atmosphere will depend upon our atmospheric models.

That issue applies to the system around HD 85512 as well, which is described as the most stable of the stars in the HARPS sample. This star is found to have a possible super-Earth in an interesting orbit indeed, with a semi-major axis of 0.26 AU and a computed equilibrium temperature of 298 K, one that could place this potentially rocky world within the inner edge of the habitable zone. As my friend Ronald Botterweg reminds me in one of the comments to an earlier post, this equilibrium temperature is not far from that of southern France about now, but again, that has to be adjusted for atmospheric effects (for a paper analyzing different atmospheric models for this planet, see Kaltenegger et al., linked to at the end of this post).

In fact, let me go ahead and quote from the Kaltenegger paper, which calls HD 85512 b “…with Gl 581 d, the best candidate for habitability known to date.”:

We focus our analysis on HD 85512 b. We show the influence of the measurement uncertainties on its location in the Habitable Zone as well as its potential habitability. We find that HD 85512 b could be potentially habitable if the planet exhibits more than 50% cloud coverage. A planetary albedo of 0.48 +/- 0.05 for a circular orbit, and an albedo of 0.52 for e=0.11 is needed to keep the equilibrium temperature below 270K and the planet potentially habitable.

And this:

If clouds were increasing the albedo of HD 85512 b, its surface could remain cool enough to allow for liquid water if present. HD 85512 b is a planet on the edge of habitability.

But back to the original HARPS paper. HD 192310 has been under investigation for several seasons following the earlier discovery of a Neptune-mass planet there. HARPS confirms that earlier discovery and adds another possibly Neptune-class world, the two semi-major axes being 0.32 AU and 1.18 AU. According to the paper, we’re again bracketing the habitable zone, with equilibrium temperatures on the order of 355 K and 185 K — possibly at the very inner and outer edges of the habitable zone, respectively.

So far, then, three of the ten stars observed in this program have yielded low-mass planets. From the paper:

Although statistics is poor over only ten targets, it is interesting to note that this 30% value was already announced by Lovis et al. (2009) who based their analysis on the larger (< 200 stars) HARPS high-precision program. Theoretical works by Mordasini et al. (2009) actually forecasted that the frequency of small Neptunes and super-Earths on short and intermediated orbits would be considerably higher than that of Saturns and Jupiters. The recent amazing discoveries made by the KEPLER satellite using the transit technique further strengthen this fact. Borucki et al. (2011) report that the probability of finding low-mass planets is considerably higher than for Jupiter or Saturn-mass planets. Furthermore, when summing up the frequency of finding a planet of any mass, they end up with a probability of about 30%, again in perfect agreement with the results of Lovis et al. (2009).

All good news for finding Earth-class worlds as we push the radial velocity method into this mass range. It’s interesting, too, to look at what this paper has to say about Alpha Centauri, Centauri B being one of the ten targets on the HARPS list for the study. As the work continues, the researchers have to contend with the bright magnitude of the Centauri stars, which “may result in poorer RV precision due to incomplete light scrambling across the spectrograph’s entrance slit.” Another major issue: Alpha Centauri B is a member of a triple star system, which means the radial velocity analysis must include a complete and precise orbital model. All of this is tricky but a thorough reading of the paper yields the conviction that HARPS is up to the task.

Tau Ceti is also a member of the list — this is one of the two stars from the original Project Ozma that Frank Drake made famous back in 1960 (the other being Epsilon Eridani). Tau Ceti as yet shows no planetary signatures, and again I’m going to turn to Centauri Dreams regular Ronald Botterweg, who has been in the thick of our ongoing exoplanet discussions for many years. Ronald analyzed the metallicity of the ten stars in the HARPS sample and found that eight of them have lower metallicity than the Sun (seven, in fact, have considerably lower metallicity than Sol). Which leads Ronald to quote a recent Greg Laughlin post on systemic:

“First, among host stars with masses similar to the Sun that harbor giant planets, there’s a strong preference for metal-rich stars. This is the classic planet-stellar metallicity effect. Second, among low-mass stars, there’s a dearth of giant planet candidates. This is the known giant planet-stellar mass effect.”

Interesting stuff, and I’m pleased at the way readers here have been digging into these papers, which not only alerts me to new work but points to issues I might otherwise have missed. Solar-type stars of low metallicity are places where we find few giant planets, the latter seeming to favor high-metallicity stars of solar size and larger. Meanwhile, the relatively high metallicity content of Centauri B, which might lead us to expect a gas giant, is presumably offset by its position as a close binary. We’ll now wait with great interest to see how the HARPS work continues on the vital and fascinating question of smaller worlds in the Alpha Centauri system. With two other teams also on the case, I suspect we won’t have to wait too much longer before we learn something definitive about the planetary situation around our nearest neighbor.

The paper is Pepe et al., “The HARPS search for Earth-like planets in the habitable zone: I — Very low-mass planets around HD20794, HD85512 and HD192310,” accepted by Astronomy & Astrophysics (preprint). See also Kaltenegger et al., “A Habitable Planet around HD 85512?” submitted to Astronomy & Astrophysics (preprint).


Comments on this entry are closed.

  • henk August 23, 2011, 9:02

    It is nice to see that they are looking at the nearby stars, of course it is still far away. That is one of the things that i do not like about Kepler that we find out that there is a habitable planets a 1000 lightyears away. We can never visit those in the foreseeable future. It is better to have a large telescope like the Terrestrial Planet Finder that can find small planets at nearby stars, but that telescope is canceled.

    I know it is a small chance, but if alpha centauri does not have any planets. How long must they study that star to come with that conclusion?

  • Dunkleosteus August 23, 2011, 10:22

    You may be interested in this paper, which describes the discovery of 18 Jovians around retired A stars by the Keck team. Although there may not be many giant planets around low-mass stars, it is certainly not true for stars more massive than the Sun.

  • andy August 23, 2011, 13:21

    Regarding the case of HD 10180, see this paper where they don’t find evidence to support the existence of candidate b (helpfully they then proceed to rename planets c-h as b-g).

  • Erik Anderson August 23, 2011, 13:30

    Ronald may want to check out my recently completed data compilation of Hipparcos stars, posted at CDS just last week. There are columns for metallicity and exoplanet-flags in main.dat.

    See: http://cdsarc.u-strasbg.fr/viz-bin/Cat?cat=XHIP

    For instance, the average [Fe/H] content of exoplanet host stars in Hipparcos with metallicity data available is 0.052 dex (i.e., +13% of solar value).

    – Erik

  • Martin J Sallberg August 23, 2011, 13:47

    About metallicity, that lower metallicity stars rarely have giant planets may simply be because their planets form too small to evolve into giants, but may still have habitable environments. I do not know what “considerably lower” means, but I have read that the average star has 75% of Sols metallicity. And yes, atmospheric character may trick temperature calculations both ways.

    One thing the article ignored is that all high-pressure life (including animal life) are facultative super-ultrathermophiles due to the similarity between molecules being crushed by pressure and molecules slamming into each other due to heat, accounting for the many deepsea animals surviving in hydrothermal vents, which opens the possibility for alien life (including complex) in pressurized hot water environments.

  • Daniel August 23, 2011, 18:21

    Great news! Wise find the first 6 Brown Dwarfs Y-class…


  • Daniel August 23, 2011, 18:36

    another Better link about the 6 Y-class Brown Dwarf , and 100 new Brown Dwarfs one at 9 light years…


  • Bounty August 23, 2011, 19:51

    “keep the equilibrium temperature below 270K and the planet potentially habitable” This confuses me.

    I have a lot going through my head on this, so I’ll try to make it quick.
    What is a good estimate for the equilibrium temperature range, that would allow for some existing liquid water on a super earth? (With or without factoring in the atmosphere, but be specific.)

  • Daniel August 23, 2011, 21:32

    The scientific paper It’s online already at arxiv:

    The First Hundred Brown Dwarfs Discovered by the Wide-field Infrared Survey Explorer (WISE)


    The Discovery of Y Dwarfs Using Data from the Wide-field Infrared Survey Explorer (WISE)


  • David August 24, 2011, 0:13


    The following gives a brief explanation:

    Note that this gives the effective temperature assuming that the planet in question radiates like a black body with no atmosphere and no internal source of energy. Not mentioned in the wikipedia article is that strictly speaking, the temperature given by the derived equation only applies at the sub-solar point which is the spot on the planet where the star (or Sun if we are talking about a planet in our solar system) is directly overhead. For more information, see any
    introductory astronomy text aimed at science students.

  • bigdan201 August 24, 2011, 1:24

    Surface gravity is a big concern of mine when it comes to planets. If it’s too low, there are long term health problems. If it’s too high, it will crush you. I’d like to understand what our gravity tolerance really is – how much lower or higher than 1G can the gravity be while allowing us to survive, adapt, and live comfortably? It seems that there isn’t enough research on this.

    Another question is figuring out the surface gravity of exoplanets in the first place. The mass can give a general idea, but we need its radius as well. However, I haven’t seen many answers on this front either.

  • spaceman August 24, 2011, 2:55

    This is more evidence to suggest that our galaxy has scads of planetary systems. One thing that is important to remember about the newly reported HARPS results is that the radial velocity technique is highly biased in favor of finding planets that are closer in and more massive. Thus, it is amazing that even with the aforementioned bias, 3 out of the 10 monitored stars have been found to harbor Super-earths and/or ice giants. I would be willing to bet that with greater precision and more years of observing, the majority (>5) of these 10 stars will show evidence of planets (smaller planets and planets with longer orbital periods).

    As for this work’s implications for finding planets around Alpha Centauri B, I think that the jury is still out. The authors note the main sources of difficulty facing radial velocity planet hunters who observe this system. Maybe I am being naively optimistic, but I think they will find terrestrial mass planet(s) around Alpha Centauri B within a few years time. We shall see! Exciting times.

  • Ronald August 24, 2011, 7:29

    Fascinating post! (but of course I am observationally biased ;-) ). This stuff is what makes me tick, including the people here.

    Martin J Sallberg:
    – “About metallicity, that lower metallicity stars rarely have giant planets may simply be because their planets form too small to evolve into giants, but may still have habitable environments”.
    Sure, there is no contradiction there. I was actually thinking along the same lines: the rocky cores remain to small (< 5 – 10 Me) to gather a significant gas envelope to develop into a true gas giant. Interestingly, the fact that there seems to be a rather sharp boundary between the two categories suggests that there are threshold values for stellar parameters (metalicity, mass) below which gas giants cannot form. And hence that there a two (or more) really distinct planetary categories, in mass, nature and origin, which are not just a continuing range. I mean the terrestrial planets (including super-earths) plus maybe the ice subgiants (Neptune class), and the gas giants. Maybe the ice subgiants are a separate category as well. Interesting question then remains (at least to me): are the super-earths oversized terrestrials (i.e. simply in the same continuum) or are they left-over cores of ice subgiants (Neptunes)?

    "I do not know what “considerably lower” means": 6 out of 10 have metallicities < 50% solar, 1 has about 65%. That is significantly below average, as you suggest yourself. 1 has about 90%, but is a K3V, hardly solar type.
    And Alph Cen B is a close binary, so that one hardly counts (with its high metallicity I mean).
    And, finally, Delta Pavonis has always been a stellar oddball with its very high metallicity, relatively high mass for its spectral class (G5-8) and already seeming to move off the main sequence, or at least advancing in it. Even the aliens agree on this ;-)

  • Ronald August 24, 2011, 8:18


    I briefly looked at the publication you mentioned, interesting stuff as well. What is clear from the summary tables 19 and 21 (at the end) is:

    – Indeed, all 18 are very bright (low Mv, high L) stars; all but 3 are high mass (1.3 – 1.95 Msol) as you would expect. 2 (HD82886 and HD96063) have just above solar mass and are rather old for an A star. 1 (HD152581) is really exceptional, only 0.93 Msol and very old. Is this really an old A star?

    – The 3 before-mentioed oddballs (are they really A stars?, their B-V color index is lower) have significantly lower metallicities (-0.3 to -0.46), all other 15 have metallicities from slightly lower to (mostly) higher than solar.

    – All planets are true gas giants of Jupiter class, or (3) even considerably larger. Remarkably, none is a real hot gas giant, the closest being at about 0.8 AU, all others at > 1 AU (1.1 to 4.3 AU). Maybe the aggressive stellar wind and the like of these bright stars disables formation of close gas giants?

    All in all this pub strongly supports the mass-and-metallicity/gas giant planet relationship. I wonder whether such stars also host one or more (closer-in) terrestrial planets or whether these have been sucked up/perturbed by the gas giants.

  • Ronald August 24, 2011, 8:24

    Erik Anderson August 23, 2011 at 13:30:
    “Ronald may want to check out my recently completed data compilation of Hipparcos stars, posted at CDS just last week.”

    I’d love to!!!

    Just give me some time (few days). Like many, I suffer from the injustice of needing (and having) a regular job and have just spent my lunch time. In the coming days (or rather evenings) I will try to neglect my family, social life and night’s rest (who needs it anyway? ;-) ).

    I have been browsing through similar lists (Hipparcos, NStars and other NASA, Simbad/Vizier, etc.), but frustratingly metallicity data can vary a great deal, and reliable, consistent metallicity data seem to be hard to come by. If anyone has good suggestions on this, please let me know.

  • andy August 24, 2011, 16:42

    are they really A stars?, their B-V color index is lower

    These seem to be mainly G and K-class subgiants. Hence the term “retired A-stars” – they were A-stars on the main sequence but have now evolved to the cooler temperatures and slower rotation rates at which precision RV becomes more feasible.

  • Erik Anderson August 25, 2011, 22:36

    Ronald said: “…metallicity data can vary a great deal, and reliable, consistent metallicity data seem to be hard to come by. If anyone has good suggestions on this, please let me know.”

    Have a look at Section 5 of our paper: http://arxiv.org/abs/1108.4971

  • Ronald August 26, 2011, 5:44

    Erik Andersen: I have browsed through the Extended Hipparcos database, impressive and very useful. I want to do some more analysis using some of these data, maybe in the weekend.
    Am I correct that there are no stellar mass data here? And in most cases no other designation than Hipparcos nr (i.e. no HD catalogue nr.)? Pity.

    Do you know other high quality metallicity data on Simbad/Vizier or elsewhere?

    I want to try and correllate metallicity (and to a lessser extent stellar mass), if possible even different elements (Fe, Si, O, C, …), to various types of planets and planetary systems.

  • Erik Anderson August 26, 2011, 12:32

    Dear Ronald,

    Yes, a stellar mass model goes beyond the scope of this project.

    Thanks for the feedback on HD designations. I regret their omission, but it would be easy for you to batch-convert the few hundred exoplanet host stars from HIP to HD using SIMBAD if you know how.

    Frankly, I don’t think we’ve overlooked any useful source of [Fe/H] data published in the last twenty years when we constructed this compilation.

    Getting the abundances of other elements would be an even more challenging scavenger hunt!

    Cheers, – Erik