Hunting exoplanets around ultracool dwarfs with near-infrared radial velocity spectrographs

The ultracool dwarf stars are frequent in the solar neighborhood, but so far only four systems with exoplanets have been confirmed, TRAPPIST-1, SPECULOOS-2 and 3, and Teegarden’s star. There are several reasons for the low number of detections. These stars are faint with their flux peaking in the near-infrared, but the RV information available in this part of the spectrum is lower than in the visible. Thus, we need more stable NIR spectrographs, such is the achievement of the CARMENES-PLUS project, and telescopes with a larger collecting area, such is the goal of MARCOT.  Additionally, ultracool dwarfs are usually fast rotators, resulting in broadened spectral lines, what lowers our capacity for detecting exoplanets.  Selecting slow rotators and correctly handling the fast ones becomes crucial. Regarding the planets, the ones around these stars are usually small, showing low RV amplitudes, few m/s or even in the cm/s regime.  The detectability expected for transits is around 2% whereas it increases to 55% with radial velocities. The strategy used to find periodic signals in these stars is still a problem to be addressed, but we can find it by looking into the successful cases. We evaluate the improvement on the detectability of exoplanets after the upgrades of CARMENES-PLUS in the near-infrared channel, that has taken the instrument’s RV precision below the 1 m/s frontier.  We also analyse the combination of the RV content of the different CARMENES spectral orders to optimize the precision and detectability.

We have a sample of 20 ultracool dwarfs (M6.0 V and later) observed with CARMENES. We analyse the radial velocity precision achieved in each object and for every spectral order of CARMENES, both of the near-infrared and the visible channels. We define two way to evaluate the precision of each spectral order: (i) by the RV error obtained and (ii) by how far the RV estimation of the order is from the final RV value. By means of an iterative process, we obtain the best combination of orders based on the previous order ranking and the detectability obtained based on the estimated RVs. With this, we are able to improve the radial velocity precision more than a 15%.

We also analyse the impact of the telluric absorption correction developed by Nagel et al. 2023. The effect on each spectral order is not the same, not even on different stars, which makes challenging to come to a clear conclusion about the improvement of these corrections. Addtionalluy, we study how to handle the fast rotators, as well as the effect of this rotation on the results. We have computed the projected rotational velocity of all the stars of our sample, and estimated most of the rotational periods. We also update the status of the Teegarden’s star system, now with more than 400 observations with CARMENES over 8 years.