The GAPS Programme at TNG -- XXIII. HD 164922 d: a close-in super-Earth discovered with HARPS-N in a system with a long-period Saturn mass companion

In the framework of the Global Architecture of Planetary Systems (GAPS) project we collected more than 300 spectra with HARPS-N at the TNG for the bright G9V star HD164922. This target is known to host one gas giant planet in a wide orbit (Pb~1200 days, semi-major axis ~2 au) and a Neptune-mass planet with a period Pc ~76 days. We searched for additional low-mass companions in the inner region of the system. We compared the radial velocities (RV) and the activity indices derived from the HARPS-N time series to measure the rotation period of the star and used a Gaussian process regression to describe the behaviour of the stellar activity. We exploited this information in a combined model of planetary and stellar activity signals in an RV time-series composed of almost 700 high-precision RVs, both from HARPS-N and literature data. We performed a dynamical analysis to evaluate the stability of the system and the allowed regions for additional potential companions. Thanks to the high sensitivity of the HARPS-N dataset, we detect an additional inner super-Earth with an RV semi-amplitude of 1.3+/-0.2 m/s, a minimum mass of ~4+/-1 M_E and a period of 12.458+/-0.003 days. We disentangle the planetary signal from activity and measure a stellar rotation period of ~42 days. The dynamical analysis shows the long term stability of the orbits of the three-planet system and allows us to identify the permitted regions for additional planets in the semi-major axis ranges 0.18-0.21 au and 0.6-1.4 au. The latter partially includes the habitable zone of the system. We did not detect any planet in these regions, down to minimum detectable masses of 5 and 18 M_E, respectively. A larger region of allowed planets is expected beyond the orbit of planet b, where our sampling rules-out bodies with minimum mass > 50 M_E.