Presentation #102.19 in the session Poster Session.
The TRAPPIST-1 (T-1) planets are among the most promising candidates for the first detailed study of temperate terrestrial exoplanets with the James Webb Space Telescope (JWST). CEA is leading a guaranteed-time observation (GTO) program that is dedicated to the observation of 5 occultations of T-1b with the Mid-Infrared Instrument (MIRI) at 12.8 μm (ID: 1279), paired with a similar program at 15 μm (ID: 1177). The reason that motivates these GTOs lays in the fact that T-1b is believed to have a day-side thermal emission large enough to be detected through the observation of a few occultations with MIRI, which could bring key insights into the existence of an atmosphere. In the past, detection of T-1b’s occultations was attempted with Spitzer/IRAC, but inconclusive (Ducrot el al. 2020). In this talk, we (1) introduce our pipeline specially designed to analyse photometric time series of faint and sparse sources, (2) show its outcomes on existing Spitzer observations of T-1b occultations to discuss conclusions from Ducrot el al. 2020, and (3) present its results on realistic simulations of JWST/MIRI for different planet properties. To detect low significance signals such as occultations (expected ≃ 150 ppm) our pipeline is based on a Blind source separation (BSS) method. First, we analyse 28 occultations of T-1b observed with Spitzer at 4.5 μm. This consists of applying aperture photometry, followed by a wavelet transform to get time and frequency observations scales, and then an Independent Component Analysis (ICA) in the wavelet domain to get rid of the high-frequency scatter (Waldmann et al., 2014, Morello et al., 2015). ICA maximizes the different components non-gaussianity (Stone et al., 2004) (ie. minimizes the negentropy) which is convenient for Spitzer as systematics are highly non-gaussian therefore separation is optimal. Although we do not detect the occultation signal of T-1b (confirming results from Ducrot et al.,2020) we derive an updated upper-constraint on the brightness temperature of the planet. Then, to prepare for GTOs 1279&1177 observations, we test our pipeline on 5 simulated occultations of T-1b at 12.8 and 15 μm. These are made by creating realistic time series observations (Martin-Lagarde et al., 2020), adding detector noises, non-linearities (Klaassen et al., 2021) and temporal drifts (Martin-Lagarde et al.,2020). Data reduction is done using the STScI pipeline. Finally, we compare the outcomes from our BSS approach to a more standard method using parametric noise models. We show that we detect the occultation of T-1b at least at 3σ in both wavelengths for a day-side temperature of 400 K.