Detection and populations
The meeting kicked off with an overview of the (Kepler) K2 mission and its datasets from Christina Hedges (NASA Kepler), with new tools (lightkurve) and different ways of approaching the data, more planets are being found, even in 'old' data. The dataset keeps on giving. The remainder of the first morning session was devoted to non-standard detection methods. Chris Manser (Warwick) has detected a planetesimal orbiting a white dwarf in a gap in its gaseous debris(?) disk. Gaseous disks around WDs are quite rare. The planetesimal is thought to be first generation, not a remnant from the red giant stage. Giovanni Rosotti (Cambridge) talked about detecting very young planets around young stars by searching for substructure in protoplanetary disks (with ALMA, SPHERE), and comparing them to simulations. This technique is sensitive to >10 M_E planets at > 10 au. Finally, George King (Warwick) has detected a planetary transit in X-rays, using the X-ray bright stellar corona as an illuminating source. This has been done before at XUV/Lyman-alpha, but not with an X-ray telescope. Perhaps the combination of XUV and X-ray transit data can be combined to say something about the composition of the planet.
Following coffee there was a session on telescopes and surveys, including TESS, CHEOPS and JWST... all of which are expected to launch soon (perhaps some sooner than others). The future is looking bright for exoplanets, especially with ESA's endorsement of the Ariel mission the day before the conference.
Amaury Triaud stood in for Laetitia Delrez (Birmingham) to talk about the TRAPPIST-1 system. Not much is known about planet formation around low-mass stars (which are the most common kind of star), so TRAPPIST-1 is getting a lot of attention, including several papers in the last few months (Van Grootel et al. 2017, Delrez et al. 2018, Grimm et al. 2018). Planetary radius, density, incident flux have all been constrained more tightly. The TRAPPIST-1 planets are mostly under-dense compared to the Solar System (excluding 1e). An important point was raised about star spots affecting particularly the TRAPPIST-1 (transmission spectra) data... see recent papers by Rackham et al. 2017 and Zhang et al. 2018. However, the Triaud, Delrez et al. team performed extensive tests, and found that there are no issues regarding star spots in the data that they are using. The next planned observations of TRAPPIST-1 will be UV observations (Hubble?) led by Julian de Wit.
On the subject of atmospheres, Jessica Spake (Exeter) reported on a detection of excited helium being stripped from the atmosphere of a hot Jupiter(?) WASP-107b, creating a comet-like tail pointed away from the planet. This is a key observation because it could be the first evidence of a planet in the gap in planet size distribution shown by Fulton et al. 2017. Also on atmospheres, Matteo Brogi (Warwick) detailed a new method of doing high resolution spectroscopy of exoplanet atmospheres in the NIR, by cross-correlating observed spectra with synthetic spectra. Doing this causes masses and orbital inclinations to drop out. With a bit of extra work you can also get the rotation and winds of the planet, and you can even say something about inversion layers in the atmosphere. This technique is being used in a new survey, MEASURE (Jayne Birkby, Amsterdam), of a handful of planets.
Several people talked about their preparations for JWST. Only 7.5% of the proposals for Early Release Science time were on exoplanets, but exoplanets were actually awarded 27% of the time, showing the priorities of the selection committee. Hannah Wakeford (STScI) and Joanna Barstow (UCL) talked about tools for transmission spectroscopy. Currently HST is leading the way for the detection of clouds & hazes (see Sing+16, Nature), but JWST will step into its place. There are GTO programmes to do transit and eclipse spectroscopy on a range of planets, from hot Jupiters to warm Neptunes to TRAPPIST-1e. Jo Barstow is looking at modelling clouds, because although transparent atmospheres are great, they're not ubiquitous. Clouds mean that you only see the top of the atmosphere, and understanding how clouds affect spectra can help to get more information from them. Barstow was another person to mention the issues with starspots. Cool stars have molecular features in their starspots, which can be confused with cloud features, or make spectra completely unfittable with synthetic spectra.
An evening lecture was given by John Sutherland (Cambridge) who provided the view of a molecular biologist on the origins of life. Unfortunately geoscience/geochemistry can not provide us with rigorous enough constraints to help us to understand how life arose, so one must proceed via molecular biology. Very interesting to see how a few key (and simple) molecules drive prebiotic chemistry... HCN, H2CO, a source of electrons (Cu or something else), a solvent (H2O) can produce many of the amino acids, etc. needed for life as we know it.
Lena Noack (Berlin) opened the day with a discussion of planetary habitability. We can actually learn quite a lot from a stellar spectrum... this gives planet composition, which constrains interior structures of planets, which influences melting temperatures and densities, which influence the convection speed of the mantle. The redox state of the mantle influences outgassing, which is key to early atmospheres. However there are lots of unknowns: the amount of water and carbon in the mantle; the redox state of the mantle; the iron fraction; the accretion time; radioactive heat sources; plate tectonics; the survival of the primordial atmosphere. So she runs Monte Carlo simulations for likely range of ~20 parameters. Her conclusions are that low-mass planets (0.5-4 M_E) should be habitable because volcanism should provide outgassing (even in stagnant lid planets). Tidal heating will also help. Induction heating may melt the crust. Fortunately a number of our favourite targets fall into this category: the TRAPPIST-1 planets, Proxima Cen b. Other beneficial properties: a water-rich mantle, with a low-medium iron content. Planets where the primordial atmosphere remains. Planets around less active stars.
Following on from this Tim Lichtenberg (ETH Zurich) talked about the dessication of rocky protoplanets. Water worlds depend on the water content of their constituent planetesimals... but models should take into account the fact that 26Al controls planetesimal heating in the early stages of planet formation. This process can dry out planetesimals and heat water to 1000C (see Fu+15, 17 and Fu & Elkins-Tanton 14). Taking this into account can reduce available water, reducing planetesimals with >30% water content to sub-20% content. This should also lead to a decrease in expected radius, something that may be statistically detectable in surveys.
Rachel Booth (Queen's University Belfast) talked about refining the age-activity relationship (see Jackson+12) for older stars where it is not so well known. To do this they leverage asteroseismic ages for a number of stars, and use the Ca II K line as an activity indicator (caveat: Ca II could also come from ablating close-in planets).
Eric Hebrard (Exeter) talked about the habitability of the early Earth, turning a pale orange dot (because of haze formation) into a pale blue dot. Theories about the early Earth are changing... the modelling of aerosol seeds as fractal structures rather than spherical (Wolf & Toon 10) changed Earth from a snowball to temperate. Hazes should be visible in reflection, emission and transmission spectra in the mid-IR. Jack Yates (Edinburgh) also talked about habitability from the viewpoint of forming an ozone layer. Around M dwarfs this may be tricky because they don't have the UV needed to kick off the ozone chemistry. Paul Rimmer (Cambridge) took the place of Sarah Rugheimer to talk about biosignatures, saying that we need to be aware that they change over time. Molecular oxygen is a good biomarker, but has a very limited window of usefulness. HCN might be a better bet... HCN is produced by volcanism and impacts and lightning and volcanic lightning. The atmosphere must be the main source, and for that UV is required to break N2 molecules (but too much UV also breaks water, which will reduce HCN). Angelica Mariani (Cambridge) talked about turning the building blocks of life into the polymers of life, particularly turning HCN in ferrocyanide. Polymerisation needs two things: an activating agent and imidazole. Water is not helpful for polymerisation. Methyl isocyanide (CH3NC) is a great candidate. So we need energy (impacts/lightning), water (but not deep, to keep UV levels up), UV irradiation (but with day/night cycles because some reactions don't work in the presence of UV), Fe-Ni meteorites in solution. Sam Roberts (UCL) talked about making RNA... but different. ANA, arabino nucleic acids. In this regime you can make all four arabino nucleoside derivatives in one pot.
Clouds and chemistry
In the afternoon session Vivien Parmentier (Marseille) talked about cloud modelling. H- has fairly recently been identified as a valuable source of opacity, and is being added to models. Together with molecular gradients, these control the appearance of spectra. Atmospheres can be probed in some detail now, for example, with Hubble recently observing 14 planets in secondary eclipses and being able to distinguish inversion layers in planetary atmospheres. It is interesting to compare Spitzer photometry (IRAC 3.6 and 4.5um) with Hubble spectra: atmospheres that look like blackbodies with HST do not necessarily look like blackbodies with Spitzer. Stefan Lines (Exeter) continued the discussion of exonephology (the study of exoplanet clouds) by talking about his 3D models of cloudy atmospheres, which is a combination of the UK Met Offices' unified model with a cloud formation/nucleation model. Despite its sophistication, its synthetic spectra still struggle to match real exoplanet spectra... there are many unknowns. But clouds are important, since they affect local temperatures and densities; e.g., silicate clouds scatter strongly and can cool an atmosphere by 250K or so. Arazi Pinhas (Cambridge) has been thinking about water depletion. The famous (and well referenced at this conference) Sing et al. (2016, Nature) cloud progression assumes no water depletion, but observations/models show (e.g., Barstow et al. 2017) a variety of sub-solar water abundances, implying water depletion. So Pinhas has been re-doing the Barstow et al. effort, with a huge increase in the number of models, finding that the majority of the Sing et al. sample have sub-solar water abundances in their atmospheres. This can be used to constrain migration and formation models for planets, e.g., Pinhas found that sub-solar water abundances are only possible in disk-free migration models in core accretion scenarios.
Ben Drummond (Exeter) has also been using the UK Met Office atmosphere models, adding non-equilibrium chemistry in the form of photoprocesses and radiative transfer. He finds that there is also a need for advection because, for instance, hot Jupiters are seen to have significant dayside/nightside transport (e.g., Amundsen et al. 2016). Adding this seems to boost methane abundances by several orders of magnitude, something which is discernable in predicted transmission and emission spectra. Ryan McDonald (Cambridge) talked about nitrogen chemistry in exoplanet atmospheres. So far no nitrogen-bearing molecules have been convincing detected, but Kilpatrick et al. 2017 have tentatively seen HCN and MacDonald & Madhusan 2017 have detected NH3 at low significance. Progress on these fronts might be made in the K-band (NH3) and Spitzer bands (HCN). Finally Alex Cridland (Leiden) showed his working on determining bulk elemental abundances of exoplanets from their discs, e.g., the nitrogen chemistry of an atmosphere is determined by when its gas was accreted: early accretion means an NH3-dominated chemistry; late accretion implies N2 dominance. He also stressed the need for considering planetesimal accretion rather than gas accretion alone, since planetesimal accretion can add much carbon to a planet's atmosphere.
Dynamics and discs
Stefan Kraus (Exeter) talked about the modelling and observation of structures within planet-forming discs. In particular, he has found features in the disc around V1247 Orionis that are probably dust traps and spiral arms, features which might slow the accretion of dust into the central star. He also has observed a HAeBe star, MWC614, which has sharply defined rings in the mid-infrared. The inner part of the disc can be seen in the near-infrared, and dust modelling gives a surprisingly high dust temperature of 1800K... could it be populated by quantum-heated particles (akin to PAHs? Kluska et al. 2018; Banzatti & Pontoppidan 2015). Soko Matsumura (Dundee) is trying to model pebble accretion in order to reproduce the distributions of planetary systems. The results of the models are similar to those of Coleman & Nelson (2016), who made used a planetismal accretion model. Mass-eccentricity distributions for giant planets are fairly consistent with observations, but both models have big issues with planetary migration, though. Giant planets at ~100au are hard to form... they are either scattered or formed in a very massive or metal-rich disc. Colin McNally (Queen Mary) is thinking about the next generation of protoplanetary disc models. Observations of discs indicate that turbulence is low (Pinte et al. 16, Flaherty et al. 2015, 2017) so rather than the traditional viscous disc models, a new regime of wind-driven discs is needed (e.g., McNally et al. 2017) that will be able to reproduce spiral arms and gaps and planet-driven torques. Adrian Barker (Leeds) discussed the effect of tidal forces on the structure of giant planets, particularly a density instability that occurs when hot material enriched in heavy metals, say, is below cooler, depleted material. The instability causes a density staircase structure which enhances tidal dissipation in the planet. Caroline Terquem (Oxford) presented her (analytical?) model of the TRAPPIST-1 system, which shows that the planets in the system must have migrated inward beyond the inner edge of the system's gas disk without stalling. However, there was no comparison to other, recent models in the literature. Jean Teyssandier (Cornell/Cambridge) discussed orbital evolution in planet-disc systems, particularly the eccentricity. Eccentricity is increased by Lindblad resonances in the disc, and damped by corotation resonances (information contributed by studies of galactic discs), but it is unclear which effect wins. Teyssandier finds that eccentricity can grow during a disc's lifetime, especially when it contains planets more massive than three Jupiters. The effect of this on inclinations is under study. Rebecca Nealon (Leicester) continued the theme of warped discs, using an SPH code to show that misaligned planets can tilt their disc (as seen observationally: Walsh et al. 2017, Loomis et al. 2017, Benisty et al. 17). Indeed, once this happens the disc interior to the planet will tilt even more. It is not clear whether this could actually be seen in scattered light observations.
Loosely tied to the theme of dynamics, Amy Bonsor (Cambridge) discussed the pollution of white dwarves by circumstellar planetesimals that are scattered inwards and accrete onto the WD. This gives a direct measure of the composition of planetesimals (since WDs should have 'clean' H/He spectra), and even planet(esimal) geology. Derived abundances from WD spectra can give ideas of formation temperatures and compositions. Aurora Sicilia-Aguilar (Dundee) has been taking spectra of T Tauri stars, but finding it hard to get any information about stars and planets because emission lines from the disc(?) veil it. She has been using rotational modulation (i.e., variability) to get around this.
James Owen (Princeton) wants to explain the 'photoevaporation valley' seen in the distribution of planet sizes shown by Fulton et al. (2017) (and Owen & Wu 2013; Lopez & Fortney 2013). It turns out that a 1% envelope mass fraction is the cutoff (this is where the envelope's radius is two times the core radius). Less massive than this, the planet becomes a stripped-bare core; more than this, the planet will accrete a 1% H/He atmosphere. This photoevaporation valley will sit at a different position with a different core composition. Mark Hammond (Oxford) talked about tidally locked close-in planet atmospheres, saying that lava planets are not a given in this situation. He has looked at different atmospheric compositions, finding that different kinds of atmosphere have different amounts of day side/night side temperature differences and hotspot shifts). A predominantly CO or N2 atmosphere produces large day/night temperature differences and a shift in the planet's hotspot from the substellar point. However, a H2 + N2 atmosphere with a strong CO component makes for a small day side/night side contrast with a shifted hotspot. i.e., molecular weight affects hotspot position. Angelos Tsiaras (UCL) has been performing a population study of hot Jupiter atmospheres (with open source data and code). Key to this is their own consistently reduced HST/WFC3 data for 30 planets. They detected water in 15 atmospheres.
The meeting closed with the discussion of missions and instrumentation. Giovanna Tinetti (UCL) introduced ESA's new Ariel mission, which will study the chemical diversity of exoplanets. We live in a Solar System where O, C, N and S are solid. In exoplanet systems things can be warmer... O, C, N, S are gas and there are appreciable amounts of Ti, Si, V. Ariel will develop on current missions like TESS. Sasha Hinkley (Exeter) talked about METIS, a first-generation imaging and spectroscopic instrument for the ELT which will cover 3 to 14 microns (perhaps 17 microns). It has a similar sensitivity to Spitzer, but will be on a 40m telescope rather than 85cm (but also on the ground). Its mid-IR window will make it sensitive to younger planets, perhaps even distinguishing between 'cold start' and 'warm start' planets (Spiegel & Burrows 2012). It will be great for imaging Earth-like planets because the contrast between star and planet drops at longer wavelengths (10^6 at 10 microns compared to 10^10 in the visible). Niranjan Thatte (Oxford) also talked about a first-generation ELT instrument, HARMONI, an integral field spectrograph. It needs more money to incorporate a high contrast mode. Finally David Brown (Warwick) gave an overview of Plato, an ESA M3 mission at L2, which will probe the habitable zone of planetary systems, aiming to increase the accuracy of mass estimates for planets. It consists of 24 (currently) telescopes on one spacecraft, and will fly at the end of 2026.