[Paper Review] POSYDON Data Release 2

Jeff J. Andrews, Simone S. Bavera+22 more|arXiv (Cornell University)|Nov 4, 2024

Advanced Chemical Physics Studies4 citations

TL;DR

POSYDON v2 introduces a cosmologically extensive binary population synthesis framework using detailed MESA simulations across metallicities from 10⁻⁴ Z☉ to 2 Z☉, including dedicated grids for the Small and Large Magellanic Clouds. It enables high-fidelity interpolation of binary evolution across mass, metallicity, and evolutionary phases, significantly improving modeling of compact object populations and their progenitors across cosmic time.

ABSTRACT

The POSYDON Data Release 2, corresponding to the binary population synthesis code POSYDON version 2.0.0. Refer to https://posydon.org/ for a detailed description, download scripts, and installation instructions. This dataset includes the downsampled single- and binary-star model grids (two single star grids and five binary star grids) along with their post-processed quantities and the trained classification and interpolation models, one tarball for each of our eight modeled metallicities (POSYDON_data_v2_grids_XXX). The auxiliary data file (POSYDON_data_auxiliary.tar.gz) contains ancillary data to aid in post-processing grids and generating binary populations, including pre-loaded supernova prescriptions, star-formation histories, and gravitational wave sensitivities. We additionally provide our randomly sampled validation (POSYDON_validation_v2_grids_XXX) and test (POSYDON_random_v2_grids_XXX) binary star model grids for each of our metallicities.

Motivation & Objective

• To extend binary population synthesis to cover a cosmologically relevant range of metallicities, from 10⁻⁴ Z☉ to 2 Z☉, including key environments like the SMC and LMC.
• To improve modeling of compact object formation and evolution by incorporating detailed, resolved MESA simulations of binary evolution across multiple phases.
• To enable accurate interpolation across a four-dimensional parameter space (mass, metallicity, mass ratio, initial period) for synthetic population generation.
• To address complex evolutionary pathways such as reverse-mass transfer and stellar mergers, which are critical for understanding compact binary populations.
• To support future modeling of low-mass binaries, white dwarfs, XRBs, and pulsar behavior through planned enhancements in upcoming versions.

Proposed method

• Constructing dense, multi-dimensional grids of MESA simulations covering initial masses from 5.5 M☉ to 286 M☉ and metallicities from 10⁻⁴ Z☉ to 2 Z☉.
• Using advanced interpolation techniques to generate synthetic binary populations from the high-resolution MESA grids without re-running simulations.
• Incorporating single-star evolution models into the population synthesis framework to improve consistency with observed stellar populations.
• Implementing a treatment for stellar mergers and reverse-mass transfer binaries, where a star that previously accreted later becomes a donor.
• Extending the model to include metallicity-dependent mass loss, core-collapse supernova prescriptions, and compact object formation with natal kicks.
• Leveraging machine learning techniques such as active learning and emulators in future versions to reduce computational cost and improve grid coverage.

Figure 1: The evolution, as a function of age, of two binary systems experiencing reverse MT. The left panels show a stable reverse mass-transfer phase with the initial conditions: $M_{\mathrm{1,ZAMS}}=28.2\,M_{\odot}$ , $q=0.95$ , $P=268$ days at $Z=Z_{\odot}$ , while the right panels show a binary

Experimental results

Research questions

• RQ1 How do binary evolution outcomes, particularly compact object formation, vary across a cosmological range of metallicities?
• RQ2 What is the impact of metallicity on the formation rates and properties of double neutron stars and black hole binaries?
• RQ3 How accurately can synthetic populations be generated via interpolation across a four-dimensional parameter space including metallicity?
• RQ4 What role do reverse-mass transfer and stellar mergers play in shaping the final mass function and merger rate of compact binaries?
• RQ5 How can detailed MESA simulations be efficiently scaled to model large-scale binary populations across cosmic time?

Key findings

• POSYDON v2 includes approximately 300,000 MESA simulations across multiple metallicities, expanding the v1 grid from three to four dimensions.
• The code now supports metallicity coverage from 10⁻⁴ Z☉ to 2 Z☉, with dedicated grids at 0.2 Z☉ (SMC) and 0.45 Z☉ (LMC), enabling realistic modeling of nearby dwarf galaxies.
• The inclusion of reverse-mass transfer and merger treatments improves the fidelity of compact binary population synthesis, especially for systems with complex mass transfer histories.
• The model grids are limited to primary masses >5.5 M☉ and extend up to 286 M☉, focusing on massive binary evolution relevant to gravitational wave sources.
• The interpolation framework enables efficient generation of synthetic populations without re-running full MESA simulations, significantly reducing computational cost.
• Future enhancements will include modeling of eccentric mass transfer, XRB behavior, improved common envelope treatment, and α-element enhanced models for low-metallicity stars.

Figure 2: Schematic summary of the wind prescriptions used for our grids. The bottom panel shows an Hertzsprung–Russell diagram with evolutionary tracks for $M_{\mathrm{ZAMS}}/M_{\odot}\in\{1.2,3,6,12,30,60,120,300\}$ at $Z_{\odot}$ with solid, black lines and the ZAMS as a dotted, gray line. The up