2022A&A...660A.124S

Aims. We use the recently published database of radial velocities (RVs) that were derived from fifteen years of HARPS/ESO observations to search for planet candidates.
Methods. For targets with sufficient RV data, we applied an automated algorithm to identify significant periodic signals and fit a Keplerian model for orbital estimates. We also searched the auxiliary data of stellar-activity indices and compared our findings with existing literature, to detect periodic RV signals that have no counterpart in the activity timeseries. The most convincing signals were then manually inspected to designate additional false planet detection, focusing the search on long-period (P> 1000 days) massive candidates around FGK dwarf stars.
Results. We identified two Jupiter analogs, in orbit around the slightly evolved F8V star HD 103891 and the Solar-like star HD 105779. We used nested sampling to derive their orbital parameters, and found their orbital periods to be 1919 ± 16 days and 2412 ± 54 days, while their minimum masses are 1.44 ± 0.02 M_Jup and 0.64 ± 0.06 M_Jup, respectively. While the orbit of HD 103891 b is slightly eccentric (e = 0.31 ± 0.03), that of HD 105779 b is likely circular (e < 0.16).
Conclusions. With minimum astrometric signatures of ∼59 and ∼42 µas, HD 103891 b and HD 105779 b join the growing sample of planets whose exact masses may soon be derived with Gaia astrometry. This finding also highlights the importance of long-term RV surveys to study planetary occurrence beyond the snow line of Solar-like stars.