The RoPES project with HARPS and HARPS-N II. A third planet in the multi-planet system HD 176986

TL;DR

This study leverages the RoPES program’s long-baseline HARPS/HARPS-N RV data and advanced spectral/activity correction (YARARA) to search for Earth-like planets around HD 176986. Using a multidimensional Gaussian process that jointly models RVs with activity indicators, the authors detect a new planet HD 176986 d at P ≈ 61.376 d, adding to the two known planets and yielding minimum masses all in the super-Earth to mini-Neptune range. They demonstrate that the three-planet, circular-orbit model is strongly favored by Bayesian evidence, while Keplerian eccentric solutions are not, and they validate the signal through apodized tests and single-instrument checks. The work also reports stellar magnetic cycle and rotation periods and defines the system’s detection limits, underscoring the continued value of long-term RV surveys for characterizing compact, low-mass planetary systems in Sun-like neighborhoods.

Abstract

Earth-like planets orbiting in the habitable zone of K- to G-type stars create an RV effect in amplitude of less than 1 \si{\meter\per\second} and have orbital periods of hundreds of days. Only long-term RV surveys with sub-meter per second precision instruments can explore the outer regions of Sun-like stars and look for Earth-like planets and super-Earths. We present the analysis of the K-type star HD 176986. It has a brightness of V=8.45 mag and a distance from the Sun of d = 27.88 pc. This star hosts a known planetary system of two super-Earths. We joined historical datasets with new data collected in an ongoing blind search program. We took advantage of recently developed tools for RV extraction and stellar activity filtering. The analysis of activity indicators permits us to determine the period of the magnetic cycle of the star alongside its rotation period. We performed a joint analysis of RVs and activity indicators through multidimensional GPs to better constrain the activity model in RVs and avoid overfitting. We detected a new planet orbiting the star and retrieved the two known planets. HD 176986 b has an orbital period of 6.49164$^{+0.00030}_{-0.00029}$ d and a minimum mass of 5.36 $\pm$ 0.44 M$\oplus$. HD 176986 c has an orbital period of P$_c$ = 16.8124 $\pm$ 0.0015 d and a minimum mass of 9.75$_{-0.64}^{+0.65}$ M$\oplus$. HD 176986 d has an orbital period of 61.376$^{+0.051}_{-0.049}$ d and a minimum mass of 6.76$_{-0.92}^{+0.91}$ M$\oplus$. From the analysis of activity indicators, we find evidence of a magnetic cycle with a period of 2432$_{-59}^{+64}$ d, along with a rotation period of 36.05 $_{-0.71}^{+0.67}$ d. We discover a new planet in the multi-planet system orbiting the K-type star HD 176986. All the planets have minimum masses compatible with super-Earths or mini-Neptunes.

Figures (16)

• Figure 1: FIP periodogram for HD 176986. The dashed red line represents the 1 % probability threshold for the periodic signal to be a false positive. We see three peaks exceeding this threshold, at period of 6.5 d, 16.8 d, and 61.4 d. The FIP level of 6.5 d and 16.8 d is cut at 11 because the numeric result gives an unphysical infinity. The peak at 61.4 d was not retrieved in the previous analysis and is an original result of our work.
• Figure 2: Time series of RV and activity indicators. Panel (a): RV time series. Panel (b): GLS periodogram related to the RV time series. We see two prominent peaks in correspondence with the two planets detected in 2018_ropes_alejandro. Panels (c), (e), (g), (i), and (k): Activity indicators time series. Panels (d), (f), (h), (j), and (l): Corresponding GLS periodograms. The red, green, and blue lines in the periodogram indicate a power corresponding to a 10%, 1%, and 0.1% FAP, respectively. Panels (c) and (d): FWHM. We see a strong peak in the GLS periodogram at 2331.89 d. Panels (e) and (f): Bisector. The strongest peak in the GLS periodogram is at 17.9 d period. This peak is likely related to stellar rotation. A weaker peak at FAP $\sim$ 10 % in correspondence with the magnetic cycle is also visible at a long period. Panels (g) and (h): S index. In panel (h) we see again a strong peak at 2331.89 d, caused by the magnetic cycle. In panels (i) and (j), we see the time series and the GLS periodogram of H$\alpha$. We see again a peak at 2331.89 d due to the magnetic cycle of the star. Panels (k) and (l): Time series and the GLS periodogram of Contrast. We can also find in this dataset the signature of the magnetic cycle in the peak at 2331.89 d in the GLS periodogram. FWHM, S index, H$\alpha$, and Contrast all show an additional peak at $\sim$ 35 d. This peak is related to the rotation period of the star as shown in the analysis of stellar activity.
• Figure 3: Apodized test for the three planets orbiting around HD 176986. Color-coded dots represent the epochs of observations of different instruments. Panel (a): Apodized signal for the 6.5 d planet found in 2018_ropes_alejandro. The signal is stable throughout the time series, but it shows a different amplitude at different epochs. Panel (b): Apodized signal for the 16.8 d planet. The signal is present in the full dataset and it is stable over time. Panel (c): Apodized signal for the 61.4 d candidate. The signal is stable over time.
• Figure 4: Apodized test for the 6d signal with the CCF RVs. The signal is stable throughout the full time series and it is not localized in any specific epoch.
• Figure 5: Phase-folded plot of HD 176986 b, HD 176986 c, and HD 176986 d.