One we missed: EPIC 228735255b

At WASP we routinely “reverse engineer” transiting exoplanets announced from other surveys to see whether we could have found them. Since the K2 mission has vastly better photometry it will find anything we’ve missed in K2 fields.

An interesting case is EPIC 228735255b, a transiting hot Jupiter in a 6.57-day orbit around a V = 12.5, G5 star, newly announced by a team led by Helen Giles, a PhD student at the University of Geneva.

In principle this planet should be within the reach of the WASP survey. However, at V = 12.5 it is at the faint end of the survey, and with a period of 6.57 days (fairly long for hot Jupiters) fewer transits get covered. Further, the WASP camera use large pixels, in order to get wide-field coverage, and for this object there is another star on the edge of our photometric aperture (see left), which degrades our photometry. Lastly, at a declination of −09 it is just below the sky covered by SuperWASP-North and so we have data only from WASP-South, principally 4600 data points from 2009 and 5700 data points from 2010.

Nevertheless, the transit was detected in WASP data, found by our standard transit-search algorithms (the WASP transit period is 6.5692 days, which compares with the Giles et al period of 6.5693 days, where the match affirms that our detection is real).

The plots show the search periodogram, showing a clear “spike” at the transit period and at twice the transit period, and (below) the WASP data folded on the transit period (transit is at phase 0).

The problem is that there is always a lot of “red noise” in WASP data, and picking candidates always involves a judgement call as to whether the signal is real. This one was just not quite convincing enough for us. The folded light curve looks pretty ratty, and the individual transit lightcurves are not particularly convincing. It had been flagged as a possible candidate, but rated as not secure enough a detection to send to the radial-velocity follow-up teams. Perhaps WASP detections might be more reliable than we thought!

While the WASP data are now superseded by the K2 photometry, it is worth recording the WASP transit ephemeris, which is period = 6.56919 (+/− 0.00036) days, epoch HJD = 2455151.1052 (+/− 0.0084), and transit width 3.56 hrs (which results from transit features spanning HJD 2454914 to 2455348).

Since these observations are from March 2009 to May 2010, they greatly extend the baseline of the Giles et al photometry, which covers 2016 July to 2017 March, and so will help refine the ephemeris to assist future observations.

The imminent TESS mission will find all the hot Jupiters that we’ve missed over the whole sky (whereas K2 is confined to the ecliptic plane), but will observe regions of sky for only a limited period and so give poor ephemerides. The above comparison suggests that WASP data will still be of valuable in being able to greatly improve the ephemerides for many TESS finds.

What happens to short-period hot-Jupiter planets?

Hot-Jupiter planets close to their host star will arouse tides in the host star, and the gravitational pull from tidal bulges will cause the planet to gradually spiral inwards. What happens to them? One possibility is that they end up spiralling into the star and are engulfed.

Another possibility is that strong irradiation from the star blasts off the planet’s atmosphere. Over time, all that would be left might be the small, rocky core of the original Jupiter-size gas giant. Maybe, then, the small, rocky, Earth-size planets seen by Kepler at ultra-short orbital periods are the remnants of hot Jupiters?

The figure shows the planetary radius versus the orbital period for a sample of Kepler planets. The Earth-size, ultra-short-period planets are in red, the hot Jupiters are in orange.

Planet radius versus orbital period for Kepler planets

Now, a team led by Joshua Winn of Princeton has tested this idea. The looked at the host stars of both the small, rocky ultra-short-period planets and of the hot Jupiters, and measured their metallicity (the fraction of elements heavier than hydrogen and helium, for which the iron abundance, denoted [Fe/H], is a good proxy).

They ended up with the following figure, which shows the metallicity distribution for the small, rocky planets (red histogram), for medium, sub-Neptune-size planets (blue histogram) and for hosts of hot Jupiters (orange histogram).

Metallicity distribution for hot Jupiter hosts versus hosts of small, rocky, ultra-short-period planets

The distribution for the rocky-planet hosts is significantly different from that for the hot-Jupiter hosts, showing that they cannot be part of the same population. This means that it is unlikely that hot Jupiters turn into rocky, ultra-short-period planets. Such planets might, however, be descended from hot Neptune-sized planets, for which the host-star metallicity does have the right distribution.

Super-Neptune WASP-107b has an oblique orbit

WASP-107b is only twice the mass of Neptune but nearly the radius of Jupiter. It is thus a hugely bloated and fluffy exoplanet and one of the more important of the recent WASP discoveries, being a prime target for atmospheric characterisation (see the discovery paper by Anderson et al 2017).

WASP-107b was also in the Campaign-10 field of the K2 mission, leading to a Kepler-quality photometric lightcurve. Recent papers by two teams, led by Teo Močnik and Fei Dai, have arrived at a similar conclusion: WASP-107b seems to be in an oblique orbit, rather than in an orbit aligned with the rotation axis of the host star.

spot_tran

The conclusion comes from star spots. If the orbit is aligned, consecutive transits will repeatedly cross the same star spot, producing a “bump” in the lightcurve each time, whereas if the orbit is oblique this will not happen.

Thus one can play the game of looking for transit bumps and seeing if they repeat. But spots can change, by growing or shrinking, so is a smaller bump in the next transit the same spot, or a different one? Also, if there is some uncertainty in the rotational period of the star, then we’re not fully sure exactly where in the next transit the spot will recur.

Star spots in transits of exoplanet WASP-107b

In the figure at left (in which the transit itself, between the dashed lines, has been removed, leaving only the starspot bumps), obvious spots are circled in red, while possible spots are marked with a lighter red. The rotational period of the star is nearly three times the orbital period of the planet, and so, if the spots recurred, they would be seen every three transits. (The gap, and thus the missing of transits 3, 4 and 5, arose from a spacecraft malfunction.)

The conclusion is that the star spots do not seem to recur and thus that WASP-107b is in an oblique orbit.

WASP-118 is pulsating

The K2 spacecraft is monitoring a series of fields along the ecliptic and so producing Kepler-quality photometry on some of the exoplanet systems previously discovered by WASP.

WASP-118b is an inflated hot-Jupiter planet (0.5 Jupiter masses but 1.4 Jupiter radii) on a 4-day orbit around a bright F-type star of V = 11. It was observed for 75 days in K2‘s Campaign 8. Teo Močnik et al have now analysed the data and see transits of the planet, as expected:

WASP-118 transits as observed with K2

The upper black curve is the raw data, while the lower red curve has been corrected for artefacts caused by drifts in K2‘s pointing. Nineteen transits are seen, recurring with the 4-day orbital period.

But Teo Močnik noticed that the out-of-transit photometry was not as flat as expected. After further investigation he deduced that the host star is pulsating:

WASP-118 is a pulsating star

The pulsations have a timescale of 1.9 days and a very low amplitude of 2 parts in 10 000, only discernable given a lightcurve with Kepler‘s photometric accuracy. Thus WASP-118 appears to be a γ-Doradus pulsator, possibly the first γ-Dor variable known to host a transiting exoplanet.

WASP-43b has an aligned orbit

WASP-43b is the hot Jupiter that is closest to its parent star, around which it orbits in only 19 hours. At such a close location, tidal interactions between the planet and the star will be intense. That means that we expect the planet to be phase locked (with its rotation period equalling the orbital period, so that the same side always faces the star), and we expect the orbit to be circular (any eccentricity having been damped by tides), and we expect the orbit to be aligned with the rotation axis of the star.

Tidal damping of the alignment of the orbit is the subject of much investigation. It seems to be most efficient if the planet orbits cooler stars, and much less efficient if the planet orbits a hotter star. This might be because cooler stars have large “convective zones” in their outer layers, which can efficiently dissipate tidal energy, whereas hotter stars have only very shallow convective zones with little mass in them.

Since WASP-43b orbits a cool star, a K7 star with a surface temperature of only 4400 Kelvin, that’s another reason for expecting its orbit to be aligned. This has now been confirmed by observations with the Italian Telescopio Nazionale Galileo. The way to measure the orbital alignment of a transiting exoplanet is by the Rossiter–McLaughlin effect. As the planet transits a rotating star, it first obscures one limb and then the other, and since the different limbs will be either blue-shifted or red-shifted, according to how the star is spinning, the effect on the overall light of star will reveal the path of the orbit.

Rossiter-McLaughlin effect

A new paper by Esposito et al reports R–M measurements for three planets including WASP-43b. The data show the classic R–M signature of an aligned planet.

Rossiter-McLaughlin effect for exoplanet WASP-43b

The upper panel shows the change in stellar radial-velocity around the planet’s orbit, caused by the gravitational tug of the planet. The lowest panel highlights the data through transit, showing the expected excursion first to a redder light (when blue-shifted light on the approaching limb is occulted) and then to blue light (when the red-shifted receding limb is occulted).

Gaia detects transits of WASP exoplanets

ESA’s Gaia satellite is a €740-million mission to map a billion stars in our galaxy. By observing repeatedly with unprecedented astrometric precision it is measuring the parallaxes, and thus the distances, of hundreds of millions of stars, and so mapping out the 3-D structure of our galaxy.

Gaia can detect exoplanets in two ways, first by astrometry (measuring the position of a star), so detecting the wobble in the star’s location caused by an orbiting massive planet, and secondly by the transit method, detecting the dip in the light of a star caused by a transiting planet.

The Gaia team have just announced the first detections of exoplanet transits, by looking at the accumulated Gaia data on two already-known WASP planets.

ESA's Gaia satellite detects its first exoplanet transit

The plot shows a year’s worth of Gaia data of the star WASP-19, folded on the 0.79-day orbital period of the planet WASP-19b (the three different panels are the star’s magnitude in three different colours). The coverage is sparse — it is designed for astrometric measurements, not for recording lightcurves — but one observation was made in-transit, demonstrating that Gaia can indeed detect exoplanet transits.

The ESA/Gaia team have also looked at the data on WASP-98, and again detect the transit of WASP-98b.

ESA's Gaia satellite detects exoplanet transit of WASP-98b

Is there potassium in WASP-31b’s atmosphere?

Many forefront facilities such as the Hubble Space Telescope and ESO’s Very Large Telescope are being pointed at exoplanets to try to find out what their atmospheres are made of. Yet such work is right at the limit of what can currently be done (though we hope that the James Webb Space Telescope will soon change that). So to what extent can we trust the results?

Here is an interesting puzzle. A new paper by Neale Gibson et al reports a spectrum of the atmosphere of WASP-31b, obtained with the FORS2 instrument on the VLT.

VLT/FORS2 Spectrum of the atmosphere of exoplanet WASP-31b

The spectrum is mostly flat, implying that the planet has a fairly cloudy atmosphere, but towards the right-hand side the orange line (a computed model) shows a strong emission line owing to potassium. The problem is that while one data point from previous HST data (small grey circle) indicates the presence of a strong potassium line, the new data from the VLT (the green-square data point) is incompatible with the HST data and would mean that there is no strong potassium line.

Gibson and co-authors put a lot of effort into trying to resolve the discrepancy, and consider whether Earth’s atmosphere might be contaminating the ground-based data, or whether unknown systematic uncertainties might be affecting the Hubble data. Overall they can only “highlight the need for caution” in interpreting such features. This illustrates that science at the cutting edge is never easy, and that much of an astronomer’s time is spent investigating whether one can trust the data one is working with.