The birthplace of hot-Jupiter exoplanets

The WASP-discovered Jupiter-sized exoplanets all orbit very close to their star, having orbital periods of typically a few days. Yet they are not thought to form there because it is generally too hot. It is likely that they form further out, where it is colder, where ices can form and start sticking together to form the planetesimals that then clump together to form planets.

The ALMA array is designed to observe the disks of cold material around young stars, disks that planets form out of. Here’s an artist’s impression, released by the National Astronomical Observatory of Japan, of a Jupiter-sized exoplanet forming out of a protoplanetary disk. Note that the planet sweeps up all the disk material near its orbit, creating a hole in the disk.

Artist's illustration of planet formation in TW Hya

The image below, from work led by Takashi Tsukagoshi and described in a NAOJ press release, then shows the real proto-planetary disk surrounding the young star TW Hydrae, as imaged by ALMA.

ALMA image of TW Hya

The image shows gaps in the disk, just as expected if planets are forming there. The planets themselves are too small to be seen directly, but this is among the best images yet of the likely birthplace of giant exoplanets.

Long-lasting starspots on exoplanet-host Qatar-2

Planets transiting their star can cross a starspot, and that — since the spot is dimmer than its surrounding — causes an upward blip in the light-curve of the transit. The same starspot can be occulted in consecutive transits, and so is seen later in phase each time because the star has rotated between the transits.

An illustration of a starspot feature in consecutive transits.  Image by Klaus Felix Huber.

A starspot feature in three consecutive transits. Image by Klaus Felix Huber.

Keele University PhD student Teo Močnik has looked at the Kepler K2 lightcurve of Qatar-2, a star known to host a hot Jupiter in a 1.34-day orbit. The lightcurve records 59 consecutive transits over a 79-day period and Močnik finds that most of the observed transits are affected by starspots (link to paper).

In the plot below each numbered lightcurve is from a transit, which occurs between the vertical dashed lines. The transit profile itself, however, has been subtracted in order to better show the starspot features.

Starspots in transits of exoplanet host Qatar-2

The starspots occur in groups, shown by red ellipses, and each group is the same starspot being seen in consecutive transits. Interestingly, though, the groups of spots themselves recur. Thus the starspots are lasting long enough that they pass behind the limb of the star, and then re-appear to be transited again one stellar rotation cycle later!

One particular starspot first causes the features in transits 20 to 22, then comes round again to produce the features in transits 33 to 36, and then comes round once again to produce the features in transits 46 to 50. Thus the starspot must have lasted for at least 40 days.

We thus have one of the best observations yet of a starspot on a star other than our sun. From this information we can calculate the rotation period of the star, place limits on the size, position and longevity of the spots, and also show that the planet’s orbit is closely aligned with the spin axis of the star.

Using the stellar rotation to trace a planet’s orbit

As a transiting exoplanet tracks across its star it progressively blocks out different regions of the face of the star. Since the star will be rotating, one limb of the star will be moving towards us (and so its light will be blueshifted) while the other limb recedes (producing a redshift). The blocking of light by the planet thus changes the spectral lines from the star. This is called the Rossiter–McLaughlin effect, and it can be used to discern the track of the planet’s orbit.

Brett Addison and Jonti Horner have written a nice introduction to such techniques on the widely read The Conversation website. Since large numbers of WASP planets orbit stars bright enough to enable a detection of the Rossiter–McLaughlin effect, around half of the planets with measured orbits are WASP planets.

Addison and Horner illustrate their piece with an artist’s conception of the polar orbit of WASP-79b:

Hot Jupiter exoplanet WASP-79b polar orbit

KELT-16b and sub-1-day hot-Jupiter exoplanets

Until recently the only hot-Jupiter exoplanets known with orbital periods below one day were the four discovered by WASP-South (WASP-18b, WASP-19b, WASP-43b and WASP-103b). But last month HATSouth reported that HATS-18b has a 0.84-day period and now KELT have announced KELT-16b at 0.97 days.

The KELT team, lead by Thomas Oberst, have produced this figure showing planetary masses against orbital separation (semi-major axis):

Short-period hot Jupiter exoplanets

One can see that all the planets just mentioned are Jupiter-mass or heavier. There are relatively few planets in the blue-shaded region, where they would have both Neptune-like masses and very short orbital periods. There are, though, Earth-mass planets known at these orbital periods. The paucity of short-period Neptunes cannot just be a selection effect, since they would have been readily found in the Kepler mission.

Instead, the currently favoured explanation is that planets in the blue-shaded region would rapidly be evaporated and be stripped down to their cores. At such short separations from their stars planets are subject to high irradiation and tidal forces. The combination can inflate the planets to the point that their atmospheres “boil off” and overflow the planet’s Roche lobe.

They avoid this fate only if the planet has enough mass, and thus gravity, to hold on to its atmosphere. Thus, at these very short orbital periods, we see either large, Jupiter-mass planets, or small, dense, rocky planets (possibly remnant cores of evaporated larger planets) — but not any in-between planets the size of Neptune.

Changeable weather on exoplanet WASP-43b?

WASP-43b is the “hot Jupiter” exoplanet with the orbit closest-in to its star, producing an ultra-short orbital period of only 20 hours. The dayside face is thus strongly heated, making it a prime system for studying exoplanet atmospheres.

Kevin Stevenson et al have pointed NASA’s Spitzer Space Telescope at WASP-43, covering the full orbit of the planet on three different occasions. Spitzer observed the infrared light from the heated face in two bands around 3.6 microns and 4.5 microns.

The three resulting “phase curves” are shown in the figure:

Spitzer phase curves of exoplanet WASP-43b

The 4.5-micron data from one visit are shown in red in the lower panel; the 3.6-micron data from the two other visits are in the upper panel. The transit (when the planet passes in front of the star) is at phase 1.0, and drops below the plotted figure. The planet occultation (when it passes behind the star) is at phase 0.5. The sinusoidal variation results from the heated face of the planet facing towards us (near phase 0.5) or away (near phase 1.0).

Intriguingly, the depth of the variation in the 3.6-micron data is clearly different between the two visits. Why is this? Well, Stevenson et al are not sure. One possibility is that the data are not well calibrated and that the difference results from systematic errors in the observations. After all, such observations are pushing the instruments to their very limits, beyond what they had been designed to do (back when no exoplanets were known and such observations were not conceived of).

More intriguingly, the planet might genuinely have been different on the different occasions. The authors report that, in order to model the spectra of the planet as it appears to be during the “blue” Visit 2 in the figure, the night-time face needs to be predominantly cloudy. But, if the clouds cleared, more heat would be let out and the infrared emission would be stronger. That might explain the higher flux during the “yellow” Visit 1. Here on Earth the sky regularly turns from cloudy to clear; is the same happening on WASP-43b?

HATS-18b and short-period hot Jupiters

Congratulations to the HATSouth project for the discovery of HATS-18b, a hot Jupiter with the very short orbital period of only 0.84 days. The other known hot Jupiters with periods below 1 day are all WASP-South discoveries (WASP-19b at 0.79 d, WASP-43b at 0.81 d, WASP-103b at 0.93 d and WASP-18b at 0.94 d).

Since such short-period systems are the easiest to find in transit surveys (owing to lots of transits!) they must be very rare, presumably because tidal forces are causing the orbits to decay, so that the planets spiral into their stars on relatively short timescales of tens of millions of years.

The HATSouth team note that the rotational periods of the host stars of HATS-18b and WASP-19b are much shorter than expected given the ages of the stars, and suggest that the stars have been spun up by the same tidal interaction that caused the planet’s orbit to decay. By modelling the in-spiral process Penev et al arrive at constraints on the “quality factor” Q* of the star. This is a measure of how efficient the star is at dissipating the tidal energy resulting from the planet’s gravitational tug on the star, and this sets the timescale for the tidal decay. Penev et al argue that the log of Q* is between 6.5 and 7, one of the tightest constraints yet estimated.

Stellar tidal decay quality factor

Estimates of the tidal quality factor, from modelling the HATS-18b and WASP-19b systems. The different models use different assumptions and are explained in the text. Figure by Penev et al.

New HATSouth planets gives us at WASP a check on our methods, since we can look for them in our own data (and if we don’t see them we can ask why not). At V = 14.1, HATS-18 is fainter than any of the WASP host stars, and fainter than we would adopt as a candidate (HATSouth is optimised to get better photometry on a slightly fainter magnitude range, whereas WASP-South is optimised for a wider field). Nevertheless, 26 000 data points from WASP-South do detect the transit of HATS-18b, giving a detected signal at the 0.837-day period and its first harmonic (1.67-d) in the period search:


There is then a clear detection of the transit when the data are folded on the transit period:


This is thus the faintest detection of a planet yet by WASP-South and so is reassuring about WASP data quality.

Cloudy Days on Exoplanets May Hide Atmospheric Water

NASA’s Jet Propulsion Laboratory have put out a press release suggesting that clouds in exoplanet atmospheres might be preventing the detection of water that lies beneath the clouds, thus explaining why some hot Jupiters show signs of water while others don’t.

The release is based on work by Aishwarya Iyer et al, published in the Astrophysical Journal in June. Iyer et al made a comprehensive study of Hubble/WFC3 data for 19 transiting hot Jupiters, including many WASP planets.

Cloud or haze layers in the atmospheres of hot Jupiters  may prevent space telescopes from detecting atmospheric water that lies beneath the clouds, according to a study in the Astrophysical Journal.

Clouds in Hot-Jupiter atmospheres might be preventing space telescopes from detecting atmospheric water. Image credit: NASA/JPL-Caltech

The press release has been extensively reported, being carried on over 40 news websites. In the UK the Daily Mail covered the story, and included a note about the recent Keele University-led discovery of five new hot Jupiters, WASP-119b, WASP-124b, WASP-126b, WASP-129b and WASP-133b.