<p>In the current theories of planet formation, the amount of energy that a forming gas giant retains from its&#160;accretion flow is still unknown. This unconstrained parameter has a large impact on the post-formation evolution&#160;of the new planet, as it defines its initial temperature and luminosity. Models have been developed, ranging from&#160;&#8220;hot-start&#8221; models assuming that all the energy is retained internally, to &#8220;cold-start&#8221; ones assuming that&#160;everything is radiated away, and "warm-start" ones in between. Their coexistence introduces large degeneracies&#160;on the determination of age and mass in direct imaging observations, as these studies use the cold or hot-start&#160;models to infer these parameters from the observed luminosity of a planet. A promising way of solving this&#160;problem is the study of atomic emission lines originating from the hot gas shocked by the accretion flow.&#160;Recently, Aoyama et al. (2018, 2020) presented simulations of hydrogen lines emitted by the accretion shock&#160;onto the circumplanetary disk and the planetary surface. They showed that the line luminosity and width can be&#160;used to infer the protoplanet mass, thus giving an estimation that is independent from the evolution models. They&#160;applied it to the case of PDS70 b and c (Aoyama & Ikoma 2019, Hashimoto et al. 2020), but were ultimately&#160;limited by the spectral resolution of the MUSE observations they used (R ~ 2500). In this context, our team&#160;recently proposed and carried out a pilot program using the VLT/ESPRESSO fiber-fed spectrograph, equipped&#160;with very high resolution (R = 190 000), to characterize the H&#945; line of the young substellar companion GQ Lup b.&#160;We will present in this poster how these observations were conducted, the methods used to remove the&#160;contamination from the host star, and the results we obtained.</p>