Assessing Planetary Stability and Long-Term Habitability in Nearby Stellar Binaries: 70 Oph, 36 Oph, gamma Leo

This paper assesses the potential for habitable planets in three nearby stellar binary systems: 36 Ophiuchi, 70 Ophiuchi, and γ Leonis. The authors use N-body simulations with the IAS15 integrator from REBOUND to test the long-term dynamical stability and habitability of test particles placed in the habitable zones of the component stars.

For the 36 Oph and 70 Oph systems, the simulations show that both stellar components can support planets in permanently habitable zones with low ejection rates and moderate eccentricity oscillations. The paper states: We find that for the 36 Oph A/B system and for the 70 Oph A/B system, the stars can support planets residing in permanently habitable zones with low ejection rates and moderate eccentricity oscillations. Specifically, for these two systems, "planets in the habitable zone with orbits coplanar to that of the binary become uninhabitable due to interactions with the binary only 1.5%−1.8% of the time, while planets with orbits 45 degrees misaligned to the plane of the binary experience larger oscillations in orbital eccentricity and as a result become uninhabitable 4.8% − 5.4% of the time."

In contrast, the γ Leo system, which contains two red giants with a highly eccentric binary orbit (e = 0.90), shows severe dynamical instability. The paper states: The habitable zones around the red giants in the γ Leo system exhibit severe dynamical instability due to the high binary eccentricity, eliminating the habitable zones around both stars. In this system, all test particles in the habitable zones were ejected, and the authors also find that the unconfirmed 1340-day planet candidate around γ Leo A is dynamically unstable and likely does not exist.

The paper concludes that 36 Oph and 70 Oph are promising targets for future missions such as the Habitable Worlds Observatory and SHERA, while γ Leo is unlikely to host any habitable planets. The methods are presented as a general framework for assessing habitability in other binary systems.

Improvements to AI Systems Based on This Paper:

1. Orbital Dynamics Prediction with High-Eccentricity Robustness

• Improvement: Enhance AI models (e.g., transformer-based time-series predictors) to handle chaotic, high-eccentricity (e > 0.8) multi-body systems by integrating physics-informed loss functions that penalize unphysical energy drift, using the IAS15 integrator as a ground-truth simulator.

• What the improved AI can do: Accurately forecast long-term (10 6–10 7 year) orbital stability and ejection probabilities for planets in binary systems without needing full N-body re-simulation, even for extreme eccentricities like γ Leo.

2. Habitability Classification Under Dynamical Perturbations

• Improvement: Train a classifier (e.g., gradient-boosted trees or graph neural networks) on simulation outputs (eccentricity oscillations, ejection rates, insolation flux variability) to predict the fraction of time a planet remains in the habitable zone, with explicit uncertainty quantification.

• What the improved AI can do: Given a binary system's parameters (masses, separation, eccentricity, inclination), instantly output a probability map of permanently habitable vs. intermittently habitable vs. uninhabitable zones, reducing the need for costly Monte Carlo runs.

3. Automated Target Selection for Exoplanet Missions

• Improvement: Build a recommendation engine that ingests stellar binary parameters and simulation-derived habitability metrics (e.g., 1.5%–1.8% uninhabitability for coplanar orbits) to rank target stars for the Habitable Worlds Observatory and SHERA.

• What the improved AI can do: Automatically filter and prioritize binary systems where habitable planets are dynamically stable, flagging systems like 36 Oph and 70 Oph as high-priority while deprioritizing γ Leo, with a confidence score based on simulation variance.

4. Validation of Planet Candidates via Stability Priors

• Improvement: Incorporate a stability prior into AI-based exoplanet confirmation pipelines (e.g., radial velocity or transit validation networks) that rejects candidates with orbital parameters leading to rapid ejection in N-body simulations.

• What the improved AI can do: Automatically flag unconfirmed candidates (like the 1340-day planet around γ Leo A) as dynamically implausible, preventing false positives and guiding follow-up observations.

5. Generalizable Framework for Arbitrary Binary Systems

• Improvement: Develop a meta-learning AI that, after training on simulations from 36 Oph, 70 Oph, and γ Leo, can generalize to unseen binary configurations (varying mass ratios, eccentricities, inclinations) by learning a latent representation of stability boundaries.

• What the improved AI can do: Provide instant habitability assessments for any newly discovered binary system, outputting not just yes/no but also the dominant instability mechanisms (e.g., eccentricity-driven ejection vs. inclination-induced oscillations), enabling rapid triage for follow-up studies.

Binary stars are common and have the potential to host habitable planets, which may reside in more complex habitable zones as compared to planets orbiting single stars. In this work, we use numerical simulations to assess the possibility that bright, nearby stellar multiples 36 Oph, 70 Oph, and gamma Leo could host habitable planets. We find that for the 36 Oph A/B system and for the 70 Oph A/B system, the stars can support planets residing in permanently habitable zones with low ejection rates and moderate eccentricity oscillations. The habitable zones around the red giants in the gamma Leo system exhibit severe dynamical instability due to the high binary eccentricity, eliminating the habitable zones around both stars. In these two systems, we find that planets in the habitable zone with orbits coplanar to that of the binary become uninhabitable due to interactions with the binary only 1.5% - 1.8% of the time, while planets with orbits 45 degrees misaligned to the plane of the binary experience larger oscillations in orbital eccentricity and as a result become uninhabitable 4.8% - 5.4% of the time. Our results identify 36 Oph and 70 Oph as promising targets for future missions such as the Habitable Worlds Observatory and SHERA, while suggesting that the stars in gamma Leo are unlikely to host any habitable planets. Our methods can be applied more generally to other binary stellar systems to refine target lists for upcoming habitable planet searches.