Empirical Stability Boundary for Hierarchical Triples

Abstract: The three-body problem is famously chaotic, with no closed-form analytical solutions. However, hierarchical systems of three or more bodies can be stable over indefinite timescales. A system is considered hierarchical if the bodies can be divided into separate two-body orbits with distinct time- and length-scales, such that one orbit is only mildly affected by the gravitation of the other bodies. Previous work has mapped the stability of such systems at varying resolutions over a limited range of parameters, and attempts have been made to derive analytic and semi-analytic stability boundary fits to explain the observed phenomena. Certain regimes are understood relatively well. However, there are large regions of the parameter space which remain un-mapped, and for which the stability boundary is poorly understood. We present a comprehensive numerical study of the stability boundary of hierarchical triples over a range of initial parameters. Specifically, we consider the mass ratio of the inner binary to the outer third body ($q_{\rm out}$), mutual inclination ($i$), initial mean anomaly and eccentricity of both the inner and outer binaries ($e_{\rm in}$ and $e_{\rm out}$ respectively). We fit the dependence of the stability boundary on $q_{\rm out}$ as a threshold on the ratio of the inner binary's semi-major axis to the outer binary's pericentre separation $a_{\rm in}/R_{\rm p, out} \leq 10^{-0.6 + 0.04q_{\rm out}}q_{\rm out}^{0.32+0.1q_{\rm out}}$ for coplanar prograde systems. We develop an additional factor to account for mutual inclination. The resulting fit predicts the stability of $10^4$ orbits randomly initialised close to the stability boundary with $87.7\%$ accuracy.