Michael Cretignier

Abstract:

Abstract The PLATO Stellar Variability Mitigation Working Group designed a radial-velocity (RV) data challenge to assess the precision and accuracy achievable on RV semi-amplitudes in the context of the future PLATO follow-up. Using Sun-as-a-star observations obtained with HARPS-N, we generated three multi-planet systems with realistic sampling, an effective signal-to-noise ratio of S/N ~ 450, an exposure time of Texp = 15 minutes, and a total of N ~ 1000 observations. The resulting data set is publicly available and can serve as a benchmark for the community to test and compare RV analysis methods. Eight independent teams participated in the challenge, aiming to recover the semi-amplitudes of both transiting and non-transiting planets. A meta-analysis of the submitted results shows that, with current data analysis methods, HARPS-N-like stability and the observing strategy studied in this paper, achieving a precision on semi-amplitudes better than 10% for habitable-zone signals around a solar-type star is challenging, and is typically limited to planets with masses of ~8–10 $\rm{$\mathrm{M}_{\oplus }$}$. This performance is expected to improve for K dwarfs, where planets down to ~2–5 $\rm{$\mathrm{M}_{\oplus }$}$ may be characterized at a similar level of precision. Finally, based on the results of this challenge and the occurrence of false positives, we derive an empirical detection threshold for RV surveys, which can be used to distinguish robust planetary detections from candidates.

Abstract:

Context . Detecting the tiny Doppler shifts induced by Earth-mass planets in stellar radial-velocity measurements remains extremely challenging due to stellar activity. Despite substantial progress in statistical and machine-learning techniques, many deep-learning methods performing well on simulated data remain difficult to apply reliably on real stellar spectra. Aims . The aim of this work is to develop a deep-learning framework that generalizes to real, unseen spectra and improves the detectability of Earth-mass planets in radial-velocity data. Methods . We train artificial neural networks on HARPS-N solar spectra with injected planetary signals, using physics-motivated spectral representations based on flux and line-formation temperature, together with their velocity gradients. Two training strategies are explored: hold-out testing, which provides a direct assessment of generalization to unseen spectra, and cross-validation, which evaluates performance across multiple folds of the dataset. The robustness of the model is enhanced by optimizing the hyperparameters based on genetic-algorithms, and the predictive uncertainty is quantified using the Monte Carlo dropout. Results . Our most precise neural network model reliably retrieves, under the cross-validation strategy, the amplitudes, phases, and orbital periods of planetary signals with amplitudes greater than or equal to 25 cm/s and periods between 10 and 550 days. In addition, in all cases tested here, the successfully recovered signals correspond to the most significant peaks in the periodograms of the Doppler-shift predictions. Temperature-based spectral-shell representations consistently outperform flux-based shells, particularly in terms of predictive uncertainty and generalization to unseen data. As a byproduct, we release doppleriann , a Python package that implements the proposed framework. Conclusions . Our results demonstrate that combining physically motivated spectral representations with deep learning provides a promising pathway toward the detection of Earth-mass planets in radial-velocity data from real observations, supported by a modeling framework that is both physically grounded and statistically rigorous, incorporating uncertainty quantification and optimized training strategies.

Abstract:

Abstract We revisit the TOI-2134 planetary system with three new high-cadence TESS sectors and 98 more spectra. This new analysis confirms the two orbiting planets by simultaneously modelling a total of eight sectors of corrected TESS photometry and 280 HARPS-N and SOPHIE radial velocities: an inner mini-Neptune in a near-circular 9.229198 ± 0.000003 days orbit, and an outer temperate sub-Saturn orbiting with a 95.852840 ± 0.000042 days period and eccentricity of 0.31 ± 0.01. The masses and radii of the planets were computed to be 9.37 ± 0.54 M⊕ and 2.735 ± 0.068 R⊕ for planet b, and 58.3 ± 1.9 M⊕ and 7.35 ± 0.18 R⊕ for planet c. The new data not only improves the detection significance and precisions on the planetary orbits, but also breaks the original multimodality in the eccentricity solution for the outer planet. We also detect a long-term trend in the radial velocity data, which we attribute to a stellar magnetic cycle. We investigate the spin-orbit alignment of the system via observations of the Rossiter-McLaughlin effect for TOI-2134 b with EXPRES and TOI-2134 c with PARAS-2. No RM effect was detected for planet b, but we find a 4.7σ detection of a 59 ± 31○ obliquity for planet c. Finally, we examine the architecture of the system, assess its completeness, investigate the planetary interior, and their suitability for follow-up atmospheric analysis.

Abstract:

Young planets offer a unique window into the early stages of planetary evolution. AU Mic is one of the nearest (9.8 pc) pre-main-sequence stars (∼20 Myr), hosting two transiting Neptune-sized planets and a debris disk. Previous studies have shown that the rotations of the central star, the debris disk, and the inner planet b are all aligned, suggesting that the system has not undergone violent evolution. Here we report new Rossiter–McLaughlin measurements for both AU Mic b and c, which happened to transit back-to-back on 2024 August 24 and 25, using the Magellan Planet Finder Spectrograph, together with contemporaneous photometry from LCOGT and CHEOPS. We confirm the aligned orbit of AU Mic b (λb = 1∘ ± 12∘) and find two possible solutions for AU Mic c: we slightly favor an aligned solution (λc = −10∘ ± 16∘) but cannot rule out a polar solution ( λc=87∘−29∘+36∘ ). Broader considerations, including dynamical stability and transit possibility, also support the mutually aligned scenario. An unexpected stellar signal during ingress and the poor transit-timing variation predictions of AU Mic c prevent a precise constraint on its obliquity, and various attempts using chromatic spectral analyses fail to outperform simple data exclusion in mitigating stellar contamination. Our observation highlights the importance of understanding stellar activity across multiple timescales and channels when characterizing young, active systems. A robust solution for the AU Mic architecture will require either a better understanding of stellar activity or future observations fortuitously free from strong stellar contamination.

Abstract:

Abstract Many areas of astrophysics, including exoplanetary studies, rely on precise and accurate stellar parameters. This demands that uncertainties on these parameters truly reflect all biases and systematics. Within this second work of the gr8stars collaboration, we take a set of 585 bright FGK dwarfs with high resolution, high signal-to-noise ratio spectra from the SOPHIE spectrograph. We determine stellar effective temperature, surface gravity, and metallicity using five different spectroscopic methods for each star, with an additional method used for comparisons. We find a typical scatter of 76 K in Teff, 0.14 dex in log g, and 0.07 dex in $\rm [Fe/H]$. These deviations are significantly larger than the average precision error on these parameters. We furthermore use isochrone fitting to determine mass, radius, and age for all 585 stars, using input from all results. We use the radii determined by SED fitting in the first gr8stars paper as a comparison to our isochronal radii from this work, in addition to comparing the isochronal log g to spectroscopic log g. The scatter in mass and radius from the use of different spectroscopic methods is investigated and propagated to exoplanetary parameters. The induced fractional uncertainties in planetary radius (≲ 3 %) and mass (≲ 5 %) are found to be below those typically found in the literature. We estimate a lower limit on planetary equilibrium temperature fractional uncertainty of ≈ 4 %, a noise floor that is currently not sufficiently represented in the literature.