A Diverse Distribution of Black Hole Spins from Stable Mass Transfer

Gravitational wave observations have found over 300 merging binary black holes, yet their origins remain uncertain. Recent work showed that many may come from isolated stellar binaries whose orbits shrink through stable mass transfer. If true, their spins may help to distinguish this channel from other formation pathways. We investigate the tidal spin up of black hole progenitor stars with detailed modeling of binaries undergoing stable mass transfer. We calculate the tidal torques by solving tidally excited oscillation modes and predict the resulting black hole spins. We find a diverse spin distribution strongly affected by the mass transfer histories of the progenitors. Binaries can form black holes with moderate spins (0.1 ≲ chieff ≲ 0.3) if they only go through case A or case B mass transfer. In the former case, they can become super-synchronized upon detachment, while in the latter case, the donor is usually only partially stripped, leaving a puffy envelope where strong tides are excited. If both case A and case AB mass transfer occur, the resulting black hole spins are almost always negligible. As the mass transfer history is jointly determined by mass ratio and initial binary period, our results predict an anti-correlation between black hole spins and mass ratio, consistent with limited evidence from data. Our results can also potentially explain the case of GW190412, a moderately-spinning binary with a high mass ratio. We discuss the limitations of our methods and additional physics (e.g., nonlinear tides, case C, and L2 mass transfer) that need to be incorporated in future work.

Improvements to AI Systems:

1. Physics-Informed Tidal Spin Evolution Model

• Enhance binary black hole population synthesis codes by embedding the paper’s tidal torque calculations (solving tidally excited oscillation modes) as a submodule. The AI can then predict spin distributions (χeff) directly from mass transfer history (case A, B, AB) without expensive hydrodynamical simulations.

• Capability: Rapidly generate synthetic spin–mass-ratio correlations for arbitrary initial binary parameters, enabling real-time comparison with LIGO/Virgo catalogs.

2. Bayesian Inference with Mass-Transfer-Aware Priors

• Replace flat or simplistic spin priors in gravitational wave parameter estimation with priors derived from this paper’s predicted anti-correlation between χeff and mass ratio. The AI can then re-analyze events like GW190412, yielding more physically motivated posterior distributions.

• Capability: Automatically flag events whose measured spins are inconsistent with stable mass transfer, prioritizing them for alternative formation channel analysis (e.g., dynamical capture).

3. Data-Driven Classification of Formation Channels

• Train a classifier (e.g., random forest or neural network) on the paper’s simulated spin outputs (moderate spins for case A/B, negligible for case AB) to distinguish isolated binary origins from other channels. Input features: χeff, mass ratio, orbital period, and mass transfer duration.

• Capability: Given a new gravitational wave event, output a probability that it originated from stable mass transfer, with uncertainty estimates.

4. Predictive Surrogate for Tidal Mode Excitation

• Build a fast surrogate model (e.g., Gaussian process or neural ODE) that maps (donor mass, radius, orbital separation, mass transfer rate) to the resulting black hole spin, trained on the paper’s detailed oscillation-mode solutions.

• Capability: Enable Monte Carlo sampling of millions of binary evolution pathways in minutes, allowing AI to explore the full parameter space of spin outcomes and identify rare high-spin events.

5. Automated Literature-to-Simulation Translation

• Use the paper’s limitations (nonlinear tides, case C, L2 transfer) to create an AI-driven gap analysis tool that suggests which physics to add next. The AI can then generate synthetic datasets with those missing effects and compare against observed spin distributions to quantify their necessity.

• Capability: Prioritize future simulation efforts by predicting which neglected physics most affects observable spin signatures, reducing wasted computational resources.

6. Spin–Mass-Ratio Correlation Detector

• Implement a statistical test in the AI that automatically searches LIGO/Virgo data for the predicted anti-correlation, using the paper’s exact functional form as a template. The AI can then report a significance level and update it as new events are added.

• Capability: Provide an early-warning system for whether the isolated binary channel dominates, or if alternative channels are needed to explain observed spins.