On the Astrophysical Origin of Binary Black Hole Subpopulations: A Tale of Three Channels?

Abstract

There is increasing evidence for multiple binary black hole~(BBH) subpopulations in the cumulative gravitational wave catalog by the LIGO-Virgo-KAGRA Collaboration. The astrophysical interpretation of this complex underlying population is subject to theoretical uncertainties in treatments of binary stellar evolution, core collapse, and host environments. In this \textit{Letter}, using parametrized mixture models, we show that the BBH detection sample comprises three astrophysical subpopulations that are likely dominated by specific formation channels. In particular, we show that the $10M_{\odot}$ peak and the $35M_{\odot}$ feature in the BBH mass spectrum correspond to distinct mass-ratio, spin alignment, spin precession, and redshift evolution properties. We show that mass-based transitions reported in the distribution of BBH parameters naturally emerge from our inferred distributions without explicit modeling. Our results are consistent with the current observed population arising from specific relative abundances of isolated binary evolution, dynamical formation in globular clusters, and higher-generation BBH mergers. Under this interpretation, we constrain the relative underlying fraction of these channels to be $79.0^{+11.5}_{-10.9}\%$, $14.5^{+11.6}_{-8.0}\%$, and, $2.5^{+5.5}_{-1.8}\%$, respectively, and find these relative fractions to be evolving over cosmic time with more than $1σ$ confidence. Our interpretation relies on simple theoretical predictions that are mostly robust against uncertainties in BBH formation, with more definite conclusions expected in the near future.

Figures (13)

• Figure 1: Distributions of primary mass (weighted by the branching fractions, top left), mass-ratio (top center), effective aligned (top right) and preccesing (bottom left) spins for each component. The redshift evolution of the merger rate (bottom center) and branching fractions (bottom right) are also shown. See supplemental material for comparisons with prior-predictive draws.
• Figure 2: Constraints on the S factor of $^{12}\mathrm{C}(\alpha, \gamma)^{16}O$ at $300~\mathrm{kev}$ obtained using the stellar models of Ref. Farag:2022jcc, depending on the chosen rate of the $3\alpha$ process, which were taken to be either equal to or $37.8\%$ larger than the values reported in the NCARE compilation 1999NuPhA.656....3AFarag:2022jcc.
• Figure 3: Variations of the metrics $P_3(q>0.7),\sigma_{\chi_{eff}}^3,$ and $P(\chi_3({eff}<0)$ for Comp. 3 with functional forms and hyperpriors. It can be clearly seen that flexible modeling assumptions do not lead to any qualitative changes in the inferred values and only lead to a broadenning of measurement uncertainties.
• Figure 4: Variations of the metrics $P(q>0.7)$ for Comp. 2 with functional forms (model I vs model IV).
• Figure 5: Posterior vs Prior. In the top two panels, the solid lines represent the posterior density where as the dashed lines represent the prior. For the other panels, the shaded region represents the $90\%$ credible interval of the posterior where as dashed lines enclose the $90\%$ credible interval of the prior.