From cosmological simulations to binary black hole mergers: The impact of using analytical star formation history models on gravitational-wave source populations

arXiv:2601.20202 · astro-ph.HE, astro-ph.GA, astro-ph.SR · Submitted 2026-01-28 · Read on arXiv

Listen

Transcript

Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.

Vera: Today's paper: "From cosmological simulations to binary black hole mergers".

Jocelyn: The impact of using analytical star formation history models on gravitational-wave source populations is investigated by coupling cosmological simulations with population synthesis codes to assess how these analytical prescriptions affect…

Vera: First, who's behind it and why it matters.

Title and authors: Jocelyn: So they are diving into what this paper actually does. They take analytical prescriptions for S(Z, z)—which is just their way of approximating how much star formation happened at different redshifts and metallicities—and they compare that against what the simulations themselves generated.

Vera: And they found that these analytical fits can cause some pretty big changes in the predicted merger rates depending on which model you use. They're looking at how different S(Z, z) prescriptions affect the BBH merger rates.

Subrahmanyan: This is important because if we rely too much on simple analytical models without checking them against the simulation data, our predictions for when and where these mergers happen could be skewed by a lot.

Jocelyn: They specifically point out that at high redshift, around z greater than six the analytical fits give merger rates that are systematically larger than what you get directly from the TNG simulation itself <ref:2601.20202#pg1>.

Vera: How big of a difference are we talking here? The paper gives some pretty striking numbers when they compare different simulations.

The paper's summary: Subrahmanyan: The researchers summarize that they constructed their analytical model for S(Z, z) by combining the total star formation rate density, S(z), with a skewed log-normal metallicity distribution, dP/dZ. This is the specific setup they used to test things.

Jocelyn: So it’s not just one simple fit; they are using a combination of elements to build their analytical model for the star formation history that they then plug into their merger rate calculations using COMPAS.

Vera: The summary highlights that these analytical prescriptions can cause significant changes in BBH merger rates, and this change depends heavily on which specific model they chose to use for S(Z, z).

Subrahmanyan: They found that at high redshift, specifically z greater than six the analytical fits yield systematically larger BBH merger rates compared to those derived directly from the IllustrisTNG simulation <ref:2601.20202#pg1>.

Jocelyn: And it’s not just a slight difference; they found these factors can be as large as one hundred four times bigger for the TNG300-one simulation, and around ten times bigger for the TNG50-one and TNG100-one simulations.

Vera: That factor of one hundred four is a big one, and it shows that you can get really far off if you aren't careful about your star formation history approximations.

The paper's improvements: Vera: So what do the authors suggest we should do differently? They look at how to fix these issues in their analytical models for S(Z, z). The main critique is that simple analytical models fail to capture some key features of the underlying data.

Subrahmanyan: Specifically, they say that these simple fits don't manage to capture the bimodality of S(Z), which is how star formation usually shows up across different redshifts.

Jocelyn: And they also point out that these simple models fail to account for the truncation of the low-metallicity tail in their metallicity distribution, which is another important part of that history.

Vera: They suggest future analytical fits need to be much more flexible. They should allow for a secondary peak in the metallicity distribution and a flatter low-metallicity tail instead of just one simple log-normal distribution.

Subrahmanyan: That flexibility would help them better capture the structure of S(Z, z) that they see in their simulations, which is crucial because those features affect the resulting BBH mass distribution too.

Jocelyn: They also suggest improving how we optimize these fits to balance getting an accurate picture of the main peak of star formation with accurately describing that low-metallicity tail.

Conclusion: Vera: So to wrap up, the central message from this paper is a caution about using simple analytical models for star formation history when predicting black hole mergers. They show that these simple models can lead to serious overestimations of merger rates at high redshift.

Subrahmanyan: The implication is that we need more complex modeling tools for S(Z, z) if we want robust predictions about the BBH population, because those simple approximations miss important physical details about how galaxies evolve.

Jocelyn: They emphasize that caution is needed when using these simple analytical fits to represent the more complex data coming out of cosmological simulations.

Vera: And they suggest that future work should look at a direct comparison with observational S(Z, z) distributions, to see if the features where these analytical fits fail in simulations actually exist in what we observe.

Subrahmanyan: I think the paper also hints that binary evolution uncertainties often dominate over the uncertainties in S(Z, z) when it comes to how much mergers happen.

Jocelyn: And they also noted that resolution matters; lower resolution simulations show a bigger disagreement between the analytical fit and the simulation's actual star formation history.

Vera: So, this paper on "From cosmological simulations to binary black hole mergers: The impact of using analytical star formation history models on gravitational-wave source populations" is a reminder that the complexity of galaxy evolution needs complex modeling to get accurate results for these kinds of astrophysical populations.

Department of Astronomy & Astrophysics, University of California, San Diego · William H. Miller III Department of Physics and Astronomy, Johns Hopkins University · Radboud University, Faculty of Science · Nicolaus Copernicus Astronomical Center, Polish Academy of Sciences · Universit¨at Heidelberg, Zentrum f¨ur Astronomie (ZAH), Institut f¨ur Theoretische Astrophysik, Albert Ueberle Str. 2, Heidelberg · Dipartimento di Fisica e Astronomia Galileo Galilei, Universit`a di Padova · Max Planck Institute for Astrophysics · Department of Astronomy, Columbia University

astro-ph.HE, astro-ph.GA, astro-ph.SR

Submitted: 2026-01-28

Updated: 2026-10-08

Comments: 26 pages, 12 figures. For associated code, see https://zenodo.org/records/18384088. Published in ApJ 09/30/26

Journal ref: The Astrophysical Journal, Volume 1009, Issue 2, id.240, 20 pp.(2026)

DOI: 10.3847/1538-4357/ae9c2f

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 81/100

The gist: The impact of using analytical star formation history models on gravitational-wave source populations is investigated by coupling cosmological simulations with population synthesis codes to assess

Key concepts

S(Z, z): This represents the cosmic star formation rate density as a function of metallicity (Z) and redshift (z). It describes how much stars are being formed in different chemical environments across cosmic time. The paper tests how well simple analytical formulas capture this complex reality derived from simulations.
IllustrisTNG Simulations: These are large-scale cosmological simulations that model the evolution of galaxies, including their star formation and chemical enrichment histories. They provide the 'ground truth' data used to test the accuracy of simpler analytical models for S(Z, z).
Population Synthesis Code (COMPAS): This is a tool used to connect the star formation history (S(Z, z)) with binary evolution physics. It takes the simulated stellar populations and calculates the resulting rates and properties of binary black hole systems that could eventually merge.
BBH Merger Rates/Mass Distribution: These are the key observables in gravitational-wave astronomy—how often black holes merge (rate) and what their masses are (distribution). The paper investigates how inaccuracies in modeling star formation history directly corrupt these observable properties.

Terminology

Metallicity — 4× in this paper · explained in 3 episodes
Population Synthesis — 2× in this paper · explained in 2 episodes
Accuracy — 1× in this paper · explained in 1 episode

Summary

The impact of using analytical star formation history models on gravitational-wave source populations is investigated by coupling cosmological simulations with population synthesis codes to assess how these analytical prescriptions affect BBH merger rates and mass distributions

Modeling Star Formation History

The paper evaluates the accuracy of analytical star formation history models, denoted as S(Z, z), when compared against the underlying distributions derived from IllustrisTNG cosmological simulations The authors construct S(Z, z) in two ways: directly based on the IllustrisTNG simulations and using an analytical fit that is often used in the field to represent S(Z, z) for the TNG simulation

Analytical Model Construction

The analytical model for S(Z, z) is constructed by combining the total star formation rate density, S(z), with a skewed log-normal metallicity distribution, dP/dZ (Z, z)

Impact on Merger Rates

The analytical S(Z, z) prescriptions can lead to significant changes in BBH merger rates depending on the model used Specifically, at high redshift (z ≳ 6), the analytical fits yield systematically larger BBH merger rates than those derived directly from the TNG simulation S(Z, z)—by factors of up to 104 for TNG300-1 and ∼ 10 for TNG50-1 and TNG100-1 (Figure 4)>

Impact on Mass Distributions

The analytical fit introduces spurious features in the BBH mass distribution, such as an artificial feature near 8 M⊙ in the primary mass distribution at z ≲ 2, which is absent when using the full simulation-based S(Z, z), while simultaneously suppressing high-mass features

Resolution and Simulation Effects

The numerical resolution of the TNG simulations significantly affects both the amplitude and the redshift evolution of the expected BBH merger rate Furthermore, the merger rate peaks at z ∼ 2 for TNG50-1 and at z ∼ 2.5 for TNG100-1 and TNG300-1, similarly to the behavior of S(Z, z) (see Section 3.1)>

Conclusion on Analytical Models

The paper concludes that simple analytical models fail to capture the bimodality of S(Z) and the truncation of the low-metallicity tail; features that both affect the resulting BBH mass distribution (see Section 3.3.1) The study emphasizes that caution is needed when using simple analytical fits to represent more complex underlying data from cosmological simulations

Future Directions

Future analytical fits should allow for more flexible modeling of the metallicity distribution—potentially incorporating a secondary high-metallicity peak and a flatter low-metallicity tail—to better capture the structure of S(Z, z) Improvements in optimization methods should also be explored to balance between an accurate fit to the peak of star formation and the low-metallicity tail of the metallicity distribution The study suggests that a direct comparison with observational S(Z, z) distributions would identify whether features where the analytical fit fails for TNG also exist in observations

Simulation-Based Comparison

Cosmological simulations can be used directly in population synthesis studies by associating BBH systems with newly formed star particles of similar metallicity The findings suggest that binary evolution uncertainties often dominate over S(Z, z) uncertainties in their impact on BBH merger rates

Resolution Effects on Normalization

The cosmological simulation resolution primarily impacts the normalization of S(Z, z) and the resulting BBH population properties At lower resolution (TNG300-1), the rates are lower and the disagreement between analytical fit and the simulation S(Z, z) is greatest The analytical model for TNG50-1 (high resolution) most accurately represents the shape of its S(Z, z) The paper suggests that a more complete treatment would move beyond galaxy-averaged S(Z, z) to include spatial gradients and scatter in metallicity and SFR within galaxies The study also suggests that future work should explore the dependence of results on both resolution and simulation box volume

Final Summary Finding

The analytical S(Z, z) model fails to accurately represent S(Z, z) of the TNG simulations by missing a secondary bump in the metallicity distribution near 3Z⊙ When using the analytical fit to model S(Z, z), the high-redshift (z ≳ 4) BBH merger rate is overestimated by up to a factor of 10 − 104, depending on simulation resolution<ref:5pg1,When using the analytical fit to model S(Z,

Improvements for AI systems

The AI system can accurately predict BBH merger rates by incorporating resolution-dependent biases from analytical star formation history models, as the numerical resolution of the TNG simulations significantly affects both the amplitude and the redshift evolution of the expected BBH merger rate.
The system can identify spurious features in predicted black hole mass distributions by detecting where analytical S(Z, z) prescriptions can overestimate BBH merger rates at high redshift (z ≳ 6) by up to a factor of 10–104, depending on cosmological simulation resolution.
The AI system can constrain the necessary functional form for star formation history models by evaluating the limitations of widely used analytical S(Z, z) fits and suggesting improvements such as allowing for a secondary, redshift-dependent peak in the metallicity distribution and greater flexibility for changes in slope within the low-metallicity tail.
The system can distinguish between physical features and modeling artifacts in stellar mass distributions by analyzing the artificial bump feature at MBH,1 ∼ 8 M⊙ that arises when using the analytical fit, which may primarily originate from the CE channel.
The AI system can assess the reliability of different modeling approaches by quantifying the discrepancy between methods using metrics like the fit-to-simulation ratio for S(Z, z), (7) and the fit-to-simulation ratio for the BBH merger properties, (8).

Abstract

Observations of binary black hole (BBH) mergers provide a unique window into the lives of massive stars across cosmic time. Connecting redshift-dependent merger properties to massive star progenitors requires accurate models of cosmic star formation and chemical enrichment histories. Analytical fits for the metallicity-specific cosmic star formation rate density S(Z, z) are commonly used as proxies for the complex underlying star formation history, yet they remain unconstrained. Using the IllustrisTNG cosmological simulations, we evaluate the accuracy of these analytical S(Z, z) prescriptions. By coupling the simulated and analytical S(Z, z) to the population synthesis code COMPAS, we investigate the resulting BBH merger rates and mass distributions. We find that analytical S(Z, z) prescriptions can overestimate BBH merger rates at high redshift (z 6) by up to a factor of 10 - 10 4, depending on cosmological simulation resolution, and can introduce spurious features in the BBH mass distribution. For example, they can produce an artificial feature near 8,M in the primary mass distribution at z 2, which is absent when using the full simulation-based S(Z, z), while simultaneously suppressing high-mass features. These discrepancies arise because simple analytical models fail to capture a high-metallicity bump and a more flattened low-metallicity tail in the simulated S(Z, z) metallicity distribution. Our results highlight the importance of accurate star formation histories for modeling BBH populations, demonstrate the limitation of widely used analytical S(Z, z) fits, and underscore the need for careful integration of cosmological simulations, analytical fits, and population synthesis when interpreting gravitational-wave observations.