Hunting Exoplanets with Single-Mode Optical Interferometry

GRAVITY Science DOI: 10.18727/0722-6691/5177

GRAVITY Collaboration (see page 20)

2000 The GRAVITY instrument was primarily conceived for imaging and astrometry of the Galactic centre. However, its 1000

sensitivity and astrometric capabilities ) 40

have also enabled interferometry to as 0

reach a new domain of astrophysics: (m 20 0 exoplanetology. In March 2019, the ion 0 GRAVITY collaboration published the first spectrum and astrometry of an 0

exoplanet obtained by optical interfer- Declin at –2 ometry. In this article, we show how –1000 00 this observation is paving the way to –4 Fringe tracker even more exciting discoveries — find- 00 ing new planets, and characterising their atmospheres. –2000 Science spectrometer 2000 1000 0 New opportunities, new challenges 20 Right ascension (mas –) 0 –2 With the 2019 Nobel Prize, jointly awarded 10 0 ﬁbre @ 70 mas to Michel Mayor and Didier Queloz, the

s) 20 ﬁbre @ 90 mas

field of exoplanet research received te –0 fibre @ 120 mas worldwide recognition. It is true that for 10 –3 40 fibre @ 150 mas the first 20 years, the domain was more minu 0 fibre @ 200 mas 5 fibre @ 250 mas

akin to a giant search for Easter eggs. /1 The rarity of each discovery meant it had σ (5 a major impact. The field was later signi 10 –4 ficantly boosted by the space-based mis- sions CoRoT (Convection, Rotation and planetary Transits), Kepler and now TESS rast ratio (Transiting Exoplanet Survey Satellite), 10 –5 resulting in the discovery of thousands of Cont exoplanets, and the development of a large community including many young 050 100150 200 250300 scientists. The success of transit photom- Angular separation (mas) etry, accompanied by a steady increase in the capabilities of stable high-precision search for M dwarfs with Exoearths with Figure 1. Upper panels: SPHERE observations of radial velocity instruments — in which Near-infrared and optical Échelle Spec- AU Mic (data from the InfraRed Dual-band Imager and Spectrograph [IRDIS] and the infrared Integral ESO has invested significantly; for exam- trographs), for example, it is possible Field Spectrograph [IFS]; Boccaletti et al. 2018). ple, the High Accuracy Radial velocity to constrain the level of atmospheric Over-plotted on the IFS observation are the GRAV- Planet Searcher (HARPS) and the Echelle evaporation from the observation of a He I ITY single-mode fibres. The sizes of the circles SPectrograph for Rocky Exoplanet and line (Alonso-Floriano et al., 2019). How- correspond to the field of view of GRAVITY. The fringe-tracker fibre is situated on the star, while the Stable Spectroscopic Observations ever, transit spectroscopy is limited to spectrometer’s fibre is positioned at separations (ESPRESSO) — now allows the analysis probing the upper atmosphere, and is between 70 and 250 mas to the south-east of the of the mass-separation distribution of inherently constrained by the duration of star. From each position of the fibre, a 5-s dynamic planets, revealing gaps such as the hot the transit. In the long term, the most range is extracted and is plotted on the contrast curve. At 120 mas, a dynamic range of 4 × 10–5 is Neptune desert (Neptune-sized planets promising technique is direct spectros- achieved. At 250 mas, the dynamic range is 4 × 10–6 within ~ 1 astronomical unit). copy, where the light of the planet (13.5 magnitudes). (either reflected light or thermal emission) The next challenge in the field is the is directly imaged on a spectrograph. characterisation of exoplanetary atmos- and a numerical capability to remove pheres through spectroscopy. Until now, This is where GRAVITY enters the field. A the stellar speckles (to distinguish the the technique has been dominated by good exoplanet imager must have two planet). On an instrument like the Spectro- high-resolution spectroscopy of evapo- main characteristics: an instrumental Polarimetric High-contrast Exoplanet rating atmospheres. With the CARMENES capability to remove the stellar diffraction REsearch (SPHERE) instrument, the for- instrument (Calar Alto high-Resolution pattern (to decrease the photon noise), mer is done by means of a coronagraph,

The Messenger 178 – Quarter 4 | 2019 47 GRAVITY Science GRAVITY Collaboration, Hunting Exoplanets with Single-Mode Optical Interferometry

while the latter uses angular or spectral c differential imaging techniques (ADI b and SDI). On GRAVITY, an off-axis single- mode fibre plays the role of a coronagraph: placed on the planet, its limited field of view filters out the stellar light. The GRAVITY interferometer surpasses single-dish instruments in post-detection Fringe e speckle removal: the angular resolution tracker of about 3 milliarcseconds (mas) yields an Science spectrometer unprecedented capability to distinguish speckles from planetary photons. 500 mas d GRAVITY as a planet hunter

0.03 GRAVITY’s high dynamic range at Luhman 16A Exo-REM T = 1150 K / log(g) = 4.3 GPI data

angular separations as small as 100 mas ) is obtained thanks to this exquisite post- –1 0.02 SPHERE photmetry

µm GRAVITY data

processing. Figure 1 shows GRAVITY –2 Wm observations of the star AU Mic. The disc 4

–1 0.01 0

of AU Mic has prominent structures, which (1

λ are resolved with SPHERE (Boccaletti, F Thalmann & Lagrange, 2015). GRAVITY 0.00 looks for point-like sources, of size 1.9 2.02.1 2.2 2.32.4 2.5 smaller than its interferometric resolution Wavelength (µm) (< 3 mas). Larger objects are not seen in Figure 2. Upper panel: SPHERE/IRDIS image of the coherent flux of the interferometer. CH4 absorption is detected. This gives HR8799 acquired with a broadband H-filter (from clues that help to characterise the atmos- Wertz et al., 2017). As in Figure 1, we put the fringe- The disadvantage of the single-mode tracker fibre on the star. The science spectrometer phere. The difficulty in interpreting the interferometer is its field of view, which is fibre is on HR8799e. The sizes of the circles corre- data lies in the complex physical pro- given by the diffraction limit of the tele- spond to the GRAVITY field of view. Below: GRAVITY cesses at work. Radiative transfer is used scope (60 mas in the K-band for the VLT K-band spectrum of HR8799e at spectral resolution to derive the pressure-temperature 500 (grey points) after 2 hours of integration. The Unit Telescopes). Therefore, while the dashed curve is the K-band Gemini Planet Imager curves, but clouds at different altitudes, fringe tracker fibre stays on the star, (GPI) spectrum from Greenbaum et al. (2018), show- with various compositions and possibly the science fibre (which feeds the spec- ing speckle contamination. (GRAVITY Collaboration also heterogeneous, modify the tempera- trograph) is placed at different positions et al., 2019). ture distribution. Chemical disequilibrium across the disc. This is how GRAVITY also adds complexity, with the necessity hunts for exoplanets; it dithers the posi- GRAVITY as a way to characterise to add chemical timescales and mixing tion of the fibre to cover a large area. In exoplanet atmospheres coefficients. In short, models need to the case of AU Mic, we took advantage be challenged by observations, and of the fact that we are only looking for a In addition to its dynamic range, GRA GRAVITY data is meeting that challenge. planet along an edge-on disc, so we only VITY’s angular resolution yields i) precise had to scan one line, thereby minimising astrometry (between 10 and 100 μas) In the near future, following the recent the required telescope time. and ii) K-band spectra mostly unbiased upgrade of GRAVITY’s high-resolution by stellar light. Fortunately, such near- grism, a resolution of 4000 will be achiev- In the resulting dataset, which covers only infrared spectra are rich in many molecu- able on exoplanets — a significant the south-eastern part of the disc, no lar absorption lines: for example, H2O, increase compared to the previous reso- detection was made. The dynamic range CO, CO2, CH4, N2O. We applied this to lution of 500 (because of limited sensitiv- achieved by GRAVITY, with 15-minute HR8799e in GRAVITY Collaboration et al. ity). GRAVITY will therefore continue to exposures, is 11 magnitudes at 120 mas (2019). HR8799e is the innermost object challenge models of exoplanetary atmos- (5-s). At 250 mas the dynamic range is in a multi-planetary system. The angular pheres, requiring simulations with more even higher, reaching 13.5 magnitudes. separation to its host star is 380 mas, resolution and more complex chemical This is several magnitudes fainter than and the contrast is close to 11 magni- processes. One exciting prospect, for what was achieved with aperture masking tudes in the K-band. The young planet example, is the detection of C13 isotopes (Gauchet et al., 2016), and a completely has an effective temperature of 1150 K, (Mollière & Snellen, 2019). In parallel, new domain compared to what could be still hot from its formation. The spectra, the recent development of atmospheric done with ADI and SDI techniques on a shown in the bottom panel of Figure 2, parameter retrieval is an exciting new single 8-metre telescope. show the CO absorption bands — no technique, which performs better than

48 The Messenger 178 – Quarter 4 | 2019 fitting a grid of models. The aim is to Acknowledgements Gauchet, L. et al. 2016, A&A, 595A, 31G obtain direct estimates of atomic ratios. GRAVITY Collaboration et al. 2019, A&A, 623, 11 See page 23. Greenbaum, A. Z. et al. 2018, AJ, 155, 226G One of them, the atmospheric C:O ratio, Mollière, P. & Snellen, I. A. G. 2019, A&A, 622A, 139 is currently believed to be a key tracer Öberg, K. I., Murray-Clay, R. & Bergin, E. A. 2011, of an exoplanet’s formation history References ApJ, 743L, 16O (Öberg, Murray-Clay & Bergin, 2011). Wertz, O. et al. 2017, A&A, 598, A83 Alonso-Floriano, F. J. et al. 2019, A&A, 629A, 110 With GRAVITY, we will soon show that Boccaletti, A., Thalmann, C. & Lagrange, A. M. 2015, we are able to measure this C:O ratio Nature, 526, 230 (GRAVITY Collaboration et al., in press). Boccaletti, A. et al. 2018, A&A, 614A, 52B ESO/Digitized Sky Survey Acknowledgement: 2. Davide de Martin

Wide-field image showing the field in the constellation of Pegasus centred on HR8799.

The Messenger 178 – Quarter 4 | 2019 49