Planets around stellar remnants V, and summary

The final day of the Planets around stellar remnants conference included some of the more interesting, but also more speculative, talks. Andre Maeder opened with a general talk on life in the universe, concerning whose existence he takes a somewhat pessimistic view. There are many obstacles that stand in the way of the evolution of life, and complex life in particular. One of these is the problem of timescales: although it took only a few hundred million years for bacteria to develop on Earth, it took around 1.5 billion each for eukarya (complex single-celled organisms such as amoebae) and multicellular life to evolve. (Although, as pointed out by Lynn Rothschild in the discussion after the talk, complex life may have arisen earlier but left no trace in the fossil record, and in the lab multicellularity can be bred through artificial selection in just a few months.) By this arguement, a star must live for several billion years in order for complex life to develop. Since higher mass stars have shorter lifetimes, any star more massive than about 1.2 Solar masses would probably not have a long enough lifetime for complex life to develop. Similar arguments make the development of life in habitable zones around evolved stars, as described by William Danchi earlier in the week, less likely. Low mass M-dwarfs, although they live for a long time, are also unsuitable, since planets with an average temperature suitable for life are likely to be tidally locked, presenting one face permanently to the star, resulting in a planet with one parched and one frozen hemisphere. Other obstacles include perturbations from other planets in the system, volcanic and asteroidal hazards, the existence or not of plate tectonics, lack of water, or too much water. Most of these were debated after the talk or later in the day. Since we have a sample size of one known life-bearing planet to work from, any generalisation can be seen as somewhat hasty.

Lynn Rothschild then spoke about extremophiles: organisms that can survive extreme environmental conditions. The existence of such beings on Earth is a reminder that we should not take too anthropocentric a view when evaluating the prospects for another planet’s hosting life. “Extreme” is of course a relative term, and we ourselves tolerate levels of oxygen (a highly reactive chemical) that would be death to many microbes. The easiest environmental quality to measure astrophysically is a planet’s temperature, and while the traditional “habitable zone” around a star is the region which gives a planet a surface temperature that allows liquid water to exist, known species can actually survive at between -40 and 120 degrees celsius. The range for complex life is narrower, however, and many extremophiles (such as the tardigrade arthropods) are actually only “extremo-tolerant”: they can survive extremes such as dessication, but are only metabolically active in more comfortable environments.

Eric Agol next decribed the habitable zone around white dwarfs. This is located very close to the star, at around 0.01 AU. Although the WD cools as it ages, the cooling is sufficiently slow that the habitable zone can remain habitable for several billion years, so life may have ample time to evolve. Although such a planet would be tidally locked to the star, the extremely short spin period would induce atmospheric currents that could warm the night side, hence distributing heat more evenly over the planet’s surface. Such planets should be easy to detect: because the WD is so small, a planet a little larger than Earth could block out 100% of the star’s light, similar to a Solar eclipse by the Moon. Quite how such a planet would form is another question: it would have to have survived the star’s giant evolution, and then been moved close to the star at a later time.

Lisa Kaltenegger the described the habitability of exoplanets in more detail. The criterion that liquid water must exist on the planet’s surface means that the planet’s temperature as calculated simply from black-body absorption of the star’s light must be about -100 to 0 degrees celsius: it is raised above zero by a combination of clouds and the greenhouse effect. On Earth, these raise the average temperature from about -10 to +14. Proper modelling of exoplanets’ atmospheres must take into account geological and biological models as well as simple atmospheric chemistry: on Earth, the carbonate–silicate cycle driven by plate tectonics regulates the level of atmospheric CO2 and prevents the greenhouse effect from becoming too severe. The study of terrestrial planet atmospheres will be difficult, but not impossible, although targets close to the Sun will be required: the Earth-sized planets so far found by Kepler are all too distant.

Abel Méndez next attempted to quantify habitability. Instead of being a simple there-is-life/there-is-no-life dichotomy, he borrowed measures from ecology such as the Habitat Suitability Index that evaluate ecosystem productivity compared to some optimum. Another metric he proposed was an Earth Similarity Index, combining not just temperature but other planetary properties such as density to evaluate the similarity of a planet to Earth. There was some skepticism expressed as to how to calibrate such metrics, however.

Next, Yutaka Abe described modelling the habitability of dry planets: those with some water but not enough to form global oceans as on Earth (which he called “aqua planets”). On these planets, precipitation and evaporation are very local, so small wet regions can have very different conditions to dry ones. The liquid water habitable zone can then be much broader than for an aqua planet: such planets could have liquid water rather close to or distant from their star.

Marc Kuchner then talked about “carbon planets”, where carbon is more common than oxygen. The rocks on such planets would be carbides, rather than oxides and silicates as on Earth. The bulk composition of such planets could be inferred from spectroscopic observations of their atmospheres (cf the next talk). Such planets may not be uncommon, since the C/O ratios of stars are often greater than unity. And such planets could host life: as Marc said, “you don’t have to look very far to find organisms that metabolise carbon”. Although the biology of creatures on these planets could be rather different: instead of breathing an oxygen atmosphere and hunting carbon-rich food, the roles could be reversed, with carbon-breathing predators hunting smaller creatures for their valuable oxygen.

Finally, Nikku Madhusudhan described observations of WASP-12b, a transiting gas giant planet whose transmission spectrum can be modelled by a carbon-rich but oxygen-poor atmosphere. Although a giant planet, it shows that terrestrial carbon-rich planets may exist too.

This ended the contributed talks at the conference. Here is a list of what I personally felt to be the important and interesting topics raised over the week. In no particular order:

  1. The statistics of planets around old and dead stars is still uncertain, due to the difficulty of detecting them. In particular, the first pulsar planets were discovered when only four or five millisecond pulsars were known, but now that hundreds are known there are still only a few planets. Did we just get lucky early on? See talks by Alex Wolszczan, Scott Ransom, and Bettina Posselt for Neutron Stars; Roberto Silvotti, Stephan Geier for subdwarfs; Matt Burleigh, Hans Zinnecker, JJ Hermes, Wei Wang for White Dwarfs; Andrzej Niedzielski, Johny Setiawan, Eva Villaver for subgiants and giants.
  2. The existence of some planetary detections is strongly disputed, particularly those detected by the timing of stellar oscillations or eclipses of binaries. See talks by Richard Wade, JJ Hermes, and Steven Parsons, in particular. There is also a growing body of literature disputing some detections on grounds of dynamical stability (see here for the most recent example).
  3. What happens when planets are engulfed by tides and their host stars’ expansion? (Eva Villaver, David Spiegel) Can they survive evaporation and the drag forces moving them into the stellar core? Can they unbind stellar envelopes to form subdwarf stars? (Stephane Charpinet, Stephane Greier)
  4. How do planets form in the harsh environment around neutron stars? Are they captured from other systems (Steinn Sigurdsson), do they form in discs after supernova fallback or collision with another star (Brad Hansen), or are they the stirpped-down remnants of stars partially consumed by the neutron star (Sam Bates)? How can planets survive in the extreme radiation environment (Cole Miller), and could we detect discs around neutron stars (Geoff Bryden)?
  5. There is now a growing consensus that the pollution of white dwarfs, and discs of dust and gas around them, are the result of planets or planetesimals being flung close to the star when planetary systems become unstable after the star becomes a white dwarf. John Debes and Shane Frewen modelled the delivery of particles to the star by planetary perturbations, and Kaitlin Kratter by binary perturbations, while Roman Rafikov modelled the evolution of the discs they form when tidally disrupted. How sensitive are such delivery mechanisms to the largely unknown architecture of extra-Solar planetary systems?
  6. Despite the conditions for life to emerge and survive being very poorly known, the existance of a habitable zone where liquid water can survive for several hundred million years may be possible after a star has passed the red giant stage (William Danchi) and for several billion when a white dwarf (Eric Agol). Hence, searches for life around other planets should not be restricted only to Solar-type stars. The conditions for life to emerge and survive are highly uncertain, however, (Andre Maeder and Lynn Rothschild), and the chances of success of such a search are impossible to predict.

Thus ended a fruitful conference. Something I really liked was that, since Arecibo observatory is a radio-quiet zone, there was no wifi in the conference auditorium. That meant that all the audience had to pay attention to every single talk, instead of playing on their laptops. It’s no surprise then that the discussion sessions were among the most interactive I’ve ever seen. For those interested, the abstract booklet is available for download here, and the slides from the talks may be uploaded in future.

After the conference it was back on the plane to Madrid, via a scheduled stop, and unscheduled extra delay to fix the air conditioning, in the Dominican Republic. In the parched plain of Castilla I’ll not see anything so green again until my next holiday to the UK later this year.

Anisotropic frequency-dependent scattering of visible light from a G2V dwarf. Interesting fact: if you entitle a picture of a rainbow "rainbow.jpg", WordPress will censor it.


Planets around stellar remnants IV

Thursday’s sessions were largely devoted to theory and modelling rather than observations, and began with my new supervisor, Eva Villaver, reviewing various aspects of the effects of giant stars on any planets they host. Particularly during the Asymptotic Giant Branch, stellar radii attain very large values, of around 1AU for a star like the Sun. Tidal torques acting on the planet’s orbit, which are heavily dependent on the stellar radius, are therefore very strong, and any Jovian planet within about 3AU of a Solar-mass star will be engulfed. What happens then is a difficult question. Although we heard talks earlier in the week about the potential of planets to unbind stellar envelopes to form hot subdwarfs, this may not be possible for planets less than 10 Jovian masses or so. Planets may also be evaporated by the high energy flux, losing about 1 Jovian mass of matter every million years. The prospects for planets surviving from the main sequence to the white dwarf phase seem bleak.

Harold Yorke then discussed planet formation around massive O stars. The discs surrounding these stars when young are very massive but very short-lived, so planet formation by core accretion–colliding dust grains and rocks to slowly build up planets–is not likely, but planet formation by gravitational instabilities in the gas disc is possible. However, the final mass of these `planets’ is uncertain, as they would accrete a great deal of material in the massive gas disc, and may well end up as small stars.

Kaitlin Kratter then discussed the stability of planets in orbits around one component of a binary system, as the stars evolve and lose mass. The regions where orbits are stable on the main sequence change when a star in a binary system loses mass, and previously stable planets can be captured into orbits around the other star, or collide with one of the stars. This may provide another way of delivering material to polluted white dwarfs, although few polluted WDs are known to have binary companions.

Dimitri Veras next described how the orbits of circumbinary planets–in which the planet orbits both stars of a binary system–change under stellar evolution. Here the orbit of the planet expands as mass is lost from the binary system, and if mass loss is very rapid the planet can be expelled from the system entirely. Planets around binaries of stars of one to two Solar masses are however usually safe, unless they orbit at very large distances.

Stein Sigurdsson then talked about the planet orbiting the binary pulsar PSR 1620-26, in the globular cluster M4. Due to the high stellar densities in globular clusters, capture of a planet from another star is a possible explanation for the planet’s origin. However, planets are not found around main sequence stars in globular clusters, so the origin may still be a puzzle.

Eduardo Martín presented a novel mechanism for creating Hot Jupiter planets, on orbits very close to their host stars. Rather than forming in the primordial circumstellar disc, he proposed that they arise from a merger of a contact binary (W UMa star), during which a fresh disc of material is thrown out. This may offer an explanation for the large radii of the so-called inflated hot Jupiters.

Roman Rafikov discussed models for the evolution of discs around WDs. Following the disruption of an asteroid, the material must be brought in from a disc at around a Solar radius to the surface of the White Dwarf, in order to cause observed metal pollution. The nature of a dust disc is similar to Saturn’s rings, and the timescale for such rings to spread is very long: Saturn has not accreted its ring material despite having several Gyr in which to do so. However, around WDs, two effects enhance the movement of dust towards the star. One is the Poynting–Robertson effect, a drag force caused by the starlight. The other is gas drag from dust that gets close to the WD and sublimates. This latter effect can trigger rapid bursts of accretion which can move significant quantities of material onto the WD quickly.

Noam Soker discussed transient events arising from the destruction of planets and planetesimals. Such events can be very violent, with the merger of a Brwon Dwarf and a Jovian planet for example casuing the BD to brighten by up to 8 magnitudes, and the destruction of an asteroid by a neutron star offering an explanation for an unusual gamma ray burst in 2010.

Shane Frewen then presented work on the dynamics of planetesimals perturbed by eccentric planets, attempting to explain how to scatter asteroids close to a White Dwarf in order to provide metal pollution. The most effective planets for sending bodies close to the star are highly eccentric and of mass somewhat less than Jupiter’s. However, many planetesimals orbiting close to planets are destabilised on the main sequence, and after the star loses mass on its way to beoming a white dwarf not many more planetesimals are destabilised. A larger source population could however be provided if the planetesimals experience gas drag in the planetary nebula, and migrate in to the previously depopulated region.

Brad Hansen described models of the formation of rocky planets around pulsars. The idea is to integrate the orbits of a collection of large protoplanets, assumed to have formed from a disc, in the same way as is done for terrestrial planet formation around Solar type stars. A variety of protoplanet configurations, corresponding to the discs expected from different formation mechanisms, such as supernova fallback and WD merger, were tried. The configuration yielding the configuration most like the planets of PSR B1257+12 corresponded to a supernova fallback disc, although since these may not be rare this raises the question of why there are not many more pulsar planets.

To end Thursday’s talks, Cole Miller gave a grimmer assessment of the prospects for forming planets around pulsars, since the environment in which they form is so harsh: heating from the pulsar’s radiation and ablation by the wind could easily destroy planetesimals of up to a kilometer in size. This suggests that planets must have formed quickly, and the planet formation process would be all-or-nothing: there could be no asteroid belts surviving as in main sequence planetary systems such as our own.

After Thursday’s talks, the conference dinner took place, in a seaside restaurant with views of the old Arecibo lighthouse:

Late 19th Century Arecibo lighthouse. The rocks in the water around here looked pretty vicious.

Planets around stellar remnants III

Wednesday saw a crammed schedule of talks. We first moved away from planets to dusty discs around stellar remnants, with Geoff Bryden opening with a review of dust discs around Main Sequence stars, and then a description of the different physical processes influencing dust discs around Neutron Stars. Notably, pulsar winds can ablate dust grains and significantly alter the size distribution of particles. There is at least one neutron star with a dust disc: 4U 2259+586, an anomalous X-ray pulsar (these pulsars are not accreting gas; the X-rays come from magnetospheric processes) which has mid IR emission from dust but no sign of gas.

Next, Ben Zuckerman reviewed the study of White Dwarf pollution. Due to the very strong gravitational fields of WDs, any metals in their atmospheres sink on a timescale of years to megayears depending  on atomic mass, so any metals present in their atmosphere must have been delivered at astronomically recent times. The best candidate for the pollution is asteroids which are tidally disrupted when they pass close to the WD, and then are accreted onto it. Indeed, for the polluted WDs whose metal content is known in detail, the composition of the pollution is very similar to the composition of rocky bodies in the inner Solar System. At least 25% of WDs show pollution, suggesting that bombardment of them by asteroids is common. Ben closed with an intriguing peek at an upcoming result showing that a dust disc around a Main Sequence star appears to have disappeared in 2009/10…

Next John Debes described the sizes of these WD dust discs in more vivid terms: roughly, they are similar in size to Saturn’s rings, and vary from thin belts to very wide discs. John then described how asteroids in a Solar-System type asteroid belt can be destabilised by a Jovian planet as the star loses mass just before becoming a WD–the Kirkwood gaps in the asteroid belt where orbits are unstable grow and more asteroids are encompassed by the unstable region. A few per cent of the unstable asteroids get hurled onto orbits that take them sufficicently close to the WD to be tidally disrupted and form a disc and provide a pollution source.

Jay Farihi next gave more physical details about the dust in these discs. They typically show evidence for silicate rocks in their infrared spectra, suggesting an origin from terrestrial planets or asteroids. Estimated disc masses are around the size of large asteroids in the Solar System, as are the estimated masses of accreted material providing the atmospheric pollution. Together, these talks gave a very strong case for the idea that WD pollution and dust discs are caused by asteroids passing close to the WD.

Next, Boris Gaensicke described how some WDs have gaseous discs. These can be hydrogen dominated, but several are only composed of gaseous metals. The gas emission lines are split, allowing the Keplerian velocity of the gas to be determined. Several show time variability, hinting at non-circular discs. Furthermore, the emission lines can be highly asymmetric, which is naturally explained as us seeing a tidal stream from a recent disruption event.

Stephan Hartmann then described explicitly the modelling of asymmetric gaseous emission lines towards the WD SDSS J1228+1040, where a simple viscous disc model can reproduce the observations. Such a model is however unrealistic as it ignores illumination from the WD, actually the dominant source of heating.

Patrick Dufour then described another particular WD, SDSS J0738+1835, which has accreted a body at least as big as Ceres, the largest asteroid, and also hosts a disc with gas and dust components. The elemental composition of the accreted matter is rather deficient in refractories, in comparison to most other polluted WDs, so the composition of extra-Solar asteroids and planets is clearly somewhat variable between systems.

Kate Su and Jana Bilikova then teamed up to talk about dust discs around hot, young WDs stil surrounded by planetary nebulae. Spitzer has found discs around 9 such WDs, well inside the large planetary nebula, a typical example being the WD and disc at the centre of the Helix Nebula. 6 of the discs are similar to the Helix disc, explicable as Kuiper Belt type discs that have survived the star’s giant stages and mass loss. The other three are more complicated, as there appears to be a close binary companion to the WD in each case, as well as the disc.

We then moved back to planets, Andrzej Niedzielski talking about planets around sub-giant and giant stars. Several dozen of these are now known, with none having been found within about 0.5 AU of the host star, in contrast to planets around Main Sequence stars where many are on close orbits. It is not clear whether this is the result of tidal engulfment of the close-in planets, or a signpost of the formation of the planets, since the masses of the giants targetted tend to be higher than the masses of the main sequence stars. He also pointed out that some giant stars have far more Lithium than they should. Lithium is quickly destroyed inside stars and should not persist until the giant stages, so perhaps these stars have had their lithium replenished by swallowing a planet or two.

Johny Setiawan then described some giant stars and their planets in some detail. Many of these giants are of very low metalicity, somewhat challenging for conventional theories of planet formation. He also showed an RV curve for a 20 Solar mass O-type star, hinting at a substellar companion. If confirmed, this will be the most massive host of an exoplanet or low-mass brown dwarf known.

David Spiegel next talked about the survival of planets to tidal forces as stars expand, and what happens if they get engulfed by the expanding stellar envelope. The huge uncertainties in tidal theory make it very hard to predict a planet at a radius of a few AU will survive its host’s expansion or not. It is also hard to produce planets that end up on intermediate orbits of a few tenths of an AU, or low mass planets on any orbit less than about 1AU, with our current understanding of what happens to planets that enter the envelope. Yet, these objects are seen.

Frederic Hessman finished the day with a description of the three circumbinary planets discovered by Kepler (two in this week’s Nature). He then described the previously reported circumbinary planets detected by timing the eclipses of the binary, drawing attention to many problems in the analyses. These include varying eclipse durations as well as times (not expected from planetary perturbations), inconsistency in the dyanical models (typically the influence of the binary on the putative planet’s orbit is neglected, so the solution is not self-consistent), to trivial dynamical instability (a notorious example, to which already at least two refutation papers have been published, actually claims two planets whose orbits overlap…). He made an explicit call for “stricter referees”. Of the 10 or so timing-based circumbinary planet claims, only NN Serpentis looks robust.

Planets around stellar remnants II

The morning of the second day was dedicated to planets orbiting white dwarfs (WDs). Matt Burleigh opened with an overview of recent attempts to detect substellar objects around WDs with direct imaging. So far there are few positive results, with the UKIDSS survey having turned up a few candidates, showing maybe a half of a per cent of white dwarfs host wide L/T dwarf companions, and a possible Y dwarf companion to GJ3483 in a very wide (~2500 AU) orbit. Meanwhile, constraints from the DODO survey show that <5% of WDs host companions heavier than 13 Jovian masses on orbits between about 10 and 60 AU, and less than a third host such companions heavier than 6 Jovian masses. A great number of such companions is perhaps not to be expected, given that they are not seen around many main sequence stars.

Hans Zinnecker then gave another negative result, from a small NICMOS survey of 7 Hyades WDs, none of them having companions heavier than 10 Jovian masses. WDs in cludters are very useful to study, since their total age can be determined with reasonable accuracy, and the age is needed to convert a brown dwarf flux into an actual mass.

JJ Holmes then talked about the use of the timing of WD oscillation modes to detect planets. Periodic signals that could arise from acceleration due to planets have been observed from several pulsating WDs. Among them is GD66 (aka V361 Aurigae), where a planet on a ~6 year orbit was announced in 2007. Since then, another “orbital” cycle has been observed, which would seem to strengthen the attribution of the signal to a planet rather than some quasi-periodic effect. However, when the timing variations of a separate mode were analysed, it was found to vary with the same period but exactly the opposite phase. Hence, the changes cannot be due to the acceleration of the whole star due to the planet, and the signals must be due to some as-yet unidentified asteroseismological mechanism. This underlines the importance of confirming “planetary” signals from timing analysis with independent measurements.

Wei Wang next presented evidence for the existence of a brown dwarf or low-mass red dwarf orbiting the young WD at the centre of the planetary nebula NGC 246. A radial velocity signal with a period of 4 days and amplitude of ~9 km/s was seen, which could be due to an object of mass at least 60 times Jovian. There also appears to be a an IR excess, indicative of hot dust or a companion radiating in the infrared. There was some criticism from the audience, since planetary nebulae are very confusing environments and contamination of the spectra could be a serious issue.

To end tuesday morning’s talks, Steven Parsons discussed the detection of companions by timing eclipsing WD binaries. He went through a similar list of caveats as mentioned by Richard Wade on Monday: timing variations may be due to angular momentum loss from winds, the Applegate mechanism, star spots, apsidal precession, and of course planets. Of the eclipsing binaries with proposed planets, NN Serpentis seems the most secure detection, particularly as newer data has vindicated the original ephemeris.

On Tuesday afternoon we were dedicated to a tour of the Observatory. This was somewhat rushed, since the megawatt RADAR transmitter was needed urgently for tracking a Near-Earth Object. While the health risks posed by 20W mobile phone towers are not credible, having a 4MW RADAR pulse aimed at you is probably a bad idea. So first we hurried to cross the narrow catwalk to the central antenna platform, suspended 450 feet over the main mirror:

The world's most handsome 'blog author prepares to cross the catwalk to the central antenna platform.

We set out…

We set out across the bridge.


Looking down from the central platform. You can see the rectangular waveguide heading back to the control room on the left.

The views from the top were pretty spectacular:

The view back down the bridge to the visitor/conference centre.

After going over the dish, we went under it. Rather than resting on the ground, the mirror is suspended about 4m above it, allowing its shape to be controlled to milimetric precision.

Underneath the dish. It feels like being in a really big athletics stadium.

From the very centre, you can look right up to the receiver structure:

This thing is 100m long.

The tour concluded with a brief look at the control centre, which was a series of rooms full of computer hardware. Not so impressive, after seeing the telescope itself!

The day ended with a drinks reception at the observatory pool, where Marc Kuchner chatted about his book Marketing for Scientists, aiming to help scientists to argue more forcefully for the value of science in society, and give practical careers advice for young scientists.