Stellar irradiation and particles forcing strongly affect the immediate environment of extrasolar giant planets orbiting near their parent stars .
However , it is not clear how the energy is deposited over the planetary atmosphere , nor how the momentum and energy spaces of the different species that populate the system are modified .
Here , we use far-ultraviolet emission spectra from HD209458 in the wavelength range \left ( 1180 - 1710 \right ) Å to bring new insight to the composition and energetic processes in play in the gas nebula around the transiting planetary companion .
In that frame , we consider up-to-date atmospheric models of the giant exoplanet where we implement non-thermal line broadening to simulate the impact on the transit absorption of superthermal atoms ( HI , OI , and CII ) populating the upper layers of the nebula .
Our sensitivity study shows that for all existing models , a significant line broadening is required for OI and probably for CII lines in order to fit the observed transit absorptions .
In that frame , we show that OI and CII are preferentially heated compared to the background gas with effective temperatures as large as T _ { OI } / T _ { B } \sim 10 for OI and T _ { CII } / T _ { B } \sim 5 for CII .
By contrast , the situation is much less clear for HI because several models could fit the Ly- \alpha observations including either thermal HI in an atmosphere that has a dayside vertical column [ HI ] \sim 1.05 \times 10 ^ { 21 } { cm ^ { -2 } } , or a less extended thermal atmosphere but with hot HI atoms populating the upper layers of the nebula .
If the energetic HI atoms are either of stellar origin or populations lost from the planet and energized in the outer layers of the nebula , our finding is that most models should converge toward one hot population that has an HI vertical column in the range [ HI ] _ { hot } \sim \left ( 2 - 4 \right ) \times 10 ^ { 13 } { cm ^ { -2 } } and an effective temperature in the range T _ { HI } \sim \left ( 1 - 1.3 \right ) \times 10 ^ { 6 } K , but with a bulk velocity that should be rather slow .