Statistical analysis of the first results of the SPHERE GTO Survey

This dissertation reports about a statistical analysis of the first 88 targets observed by the science channel IFS (Integral-Field-Spectrograph) during the SPHERE Guaranteed Time Survey (GTO). The final goal of this work is to put some initial constraints on the frequency of giant planets in wide orbits, on their mass distribution, on their semi-major axis distribution and possibly on the formation mechanism. In the first chapter we briefly present the two theories aimed at explaining the planet formation mechanism: the “core accretion” and the “disk instability”. In the second chapter we introduce the direct-imaging technique to discover young and self-luminous exoplanets and in the same chapter we present a new instrument optimized to perform direct imaging: SPHERE a Spectro-Polarimetric High-contrast-Exoplanet- REsearch. In the third chapter we describe the Guaranteed Time of Observation (GTO) survey SHINE, which is currently ongoing on SPHERE, the target selection and finally the results of the first two semesters of the survey. In the fourth outline we describe the statistical formalism used for the analysis and in particular the Quick-MESS code (Quick Multipurpose Exoplanet Simulation System): a fast alternative code to the classic Monte-Carlo tools for the statistical analysis of exoplanet direct imaging surveys. The results of our analysis are given in the fifth chapter and finally we compare our data with results from other surveys in sixth chapter. Although still exploratory, because the candidates so far found with SPHERE still require confirmation and only about 1=4 of the targets have been observed, this analysis will provide a first test of the methodology we plan to use once the survey is completed and some very early results. We show that current data are compatible with distributions, from the radial velocities, with only few planets beyound 10 - 20 AU. The peak of the giant planet distribution should then be at a separation not much larger than the snow-line, in agreement with the very recent result obtained by Bryan et al. (2016) from a combination of a radial velocity and the direct imaging data. This is interesting because the selection criteria used in our survey is very different, focusing on young objects and is not biased versus system with closer planets. On the other hand, the number of planets so far detected in our survey, while still compatible with an extrapolation of the results by Bryan et al., is at the lower limit of the acceptable range. Completion of the SPHERE survey will roughly reduce at half the current error bar in the frequency of planets at large separations (> 10 AU). This will be enough to show if this frequency is as high as expected from Bryan et al. analysis, or lower as suggested by our preliminary data. Finally, we propose a distribution of the frequency of giant planets versus the separations. The peak of distribution, in agreement with RV and SHINE data, should be slightly out of the snow-line as predicted by the core accretion scenario. Furthermore we note that the positions of Jupiter and Saturn are compatible with the peak of the overall distribution of giant planets, showing that on this respect the Solar System does not represent an exception.