Multi-Generational Black Hole Population Analysis with an Astrophysically Informed Mass Function

arXiv:2406.06109 · astro-ph.HE, astro-ph.CO, gr-qc, hep-ph · Submitted 2024-06-10 · Read on arXiv

Listen

Radio episode about this paper

Transcript

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

Vera: We're looking at the paper: "Multi-Generational Black Hole Population Analysis with an Astrophysically Informed Mass Function".

Jocelyn: Here's the gist of the paper. Multi-Generational Black Hole Population Analysis with an Astrophysically Informed Mass Function analyzes black hole population statistics from LIGO/Virgo/KAGRA data using a parametric mass function derived from pulsational pair-instability supernovae simulations to separate black holes formed through stellar collapse from those formed via prior mergers, providing constraints on formation pathways and cosmology Mass Function Modeling The study employs a parametric mass function that accounts for different generations of black holes, modeling the total probability distribution as a sum of terms representing various binary formation channels.

Vera: The mass function for first-generation black holes is derived from one-dimensional stellar structure simulations and incorporates the effects of pulsational pair-instability supernovae (PPISN).

Jocelyn: This function includes a "turn-on" function at the low mass end, which is derived from analysis in a prior work.

Vera: The authors compare their PPISN-inspired mass function to the phenomenological power law+peak (PLP) model, finding that their model provides a better fit to the data with a Bayes factor of log ZPPISN − log ZPLP = 9.7 ± 0.1.

Title and authors: Vera: So, we're looking at the paper "Multi-Generational Black Hole Population Analysis with an Astrophysically Informed Mass Function." We discussed how this paper uses a mass function from pulsational pair-instability supernovae simulations to sort black holes into different generations.

Jocelyn: And it also incorporates spin information to separate the first-generation and second-generation black holes, which is really interesting for understanding their formation history.

Subrahmanyan: The title itself signals that this isn't just about counting black holes; it's about understanding the different astrophysical pathways they take, right?

Vera: It’s about separating those formation pathways—stellar collapse versus mergers—and seeing how those differences show up in the observed statistics.

Jocelyn: They are using this framework to see if the statistical fingerprints of stellar collapse and dynamical mergers match what we observe in the gravitational wave catalog >

Subrahmanyan: That’s what they are doing, trying to use these statistical tools to test which physical processes dominate black hole growth in the early universe >

Vera: The main implication is that by separating these populations, we get more precise constraints on how black holes grow over time.

Jocelyn: And they're using the mass function comparison against a phenomenological power law plus peak model to show their PPISN-inspired model is statistically better, with a Bayes factor of log ZPPISN minus log ZPLP equals nine point seven plus or minus zero point one > <ref:2406.06109#pg2>

Subrahmanyan: That Bayes factor is strong evidence supporting the new mass function formalism over the older phenomenological one, which means we should trust their separation of formation channels more when interpreting data >

Vera: It’s a significant statistical improvement for fitting these complex gravitational wave signals, especially since they’re dealing with the intricacies of different black hole generations.

Jocelyn: And they are not just relying on one aspect; they're layering in spin information to make the separation even clearer.

Subrahmanyan: Layering in spin information is crucial because it directly relates the mass function to the underlying physics of how those black holes were born >

Vera: So, we’re essentially using two different observational features—mass and spin—to get a much richer picture of black hole formation history.

Jocelyn: And they are not just looking at the statistics; they are trying to link those statistics back to the simulations that predict those initial conditions >

The paper's summary: Vera: So, let’s look at what the paper actually summarized about their methodology. They explain how they use a parametric mass function derived from PPISN simulations to separate black holes into two subpopulations.

Jocelyn: They describe how they define the terms lambda one two and lambda two one to quantify the number of black holes in each population, which helps with normalization >

Subrahmanyan: That quantification is necessary because you have to make sure you correctly normalize these counts so that the different contributions to the total are properly accounted for >

Vera: And they’re using those terms, like lambda one two and lambda two one, to count the number of primary and secondary black holes in the second generation >

Jocelyn: They also introduce N0 to count events where both black holes are first generation, which helps avoid dependence on Ntot for normalization >

Subrahmanyan: That approach is clever because it lets them handle the complexity of counting events involving different generations without getting stuck on a single total number upfront >

Vera: So, it’s about building a statistically sound way to handle these overlapping formation scenarios in a gravitational wave catalog, which is what they are doing here.

Jocelyn: And then they move into calculating those normalisation integrals, Ntot, N1, and N2 to get the actual statistics for the different populations >

Subrahmanyan: Those integrals are complex because they can be split into terms that can be calculated analytically in terms of hypergeometric functions and some that require numerical evaluation >

Vera: It shows how hard it is to calculate these things, which means they’re using a mix of analytical math and numerical methods for the inference process >

Jocelyn: And the machine-readable version of these integrals allows other researchers to run the inference process themselves if they want to test their own priors >

Subrahmanyan: That’s a good feature because it opens up the modeling space for others to build on their own theoretical predictions about black hole populations >

The paper's improvements: Vera: So, the authors suggest some key improvements to this analysis. One major suggestion is incorporating spin information, which they use to separate the spin distributions of stellar remnant black holes from those formed via prior mergers >

Jocelyn: That’s where they find that the 1g population prefers smaller values of the dimensionless spin parameter chi than the 2g population > <ref:2406.06109#pg1>

Subrahmanyan: That is consistent with stellar structure theory for first-generation black holes, and dynamical merger scenarios for higher-generation ones >

Vera: So, it’s suggesting a way to use spin as a discriminator between those two formation channels.

Jocelyn: They are also saying that black holes formed in the dynamical merger scenario should have a normalized spin magnitude tending to zero point seven, whereas those from isolated stellar evolution are predicted to be born with negligible dimensionless spin magnitudes >

Subrahmanyan: That specific zero point seven normalization for the dynamical mergers is a concrete prediction that connects the spin measurement directly to their formation scenarios >

Vera: So they are suggesting that measuring that spin distribution is a direct way to confirm or rule out different formation hypotheses about these black holes >

Jocelyn: It’s an experimental path, because if we can measure those spins, we get a much clearer picture of the physics happening at the source >

Subrahmanyan: And if the measurement holds up, it really solidifies the link between spin and generation and formation physics >

Conclusion: Vera: So, to wrap up this paper "Multi-Generational Black Hole Population Analysis with an Astrophysically Informed Mass Function." The core message is that they’ve successfully distinguished between black holes from stellar collapse and those formed via prior mergers >

Jocelyn: They’ve also achieved a better fit to the LVK O3 data than the phenomenological power law plus peak model >

Subrahmanyan: Overall, this new framework gives us a solid way to interpret gravitational wave catalog data based on astrophysical theory rather than just phenomenological fits >

Vera: But we have that measured H0 value equal to thirty-six point one nine plus seventeen point five zero minus ten point nine one kilometers per second per megaparsec, which is discrepant with SH0ES and Planck measurements at a level of two sigma >

Subrahmanyan: This paper provides a framework for interpreting gravitational wave catalog data to provide information about black hole formation and cosmology, but we need to keep looking into why that measured H0 value is lower than expected >

Vera: We're done with this one, the "Multi-Generational Black Hole Population Analysis with an Astrophysically Informed Mass Function." We’re ready for whatever comes next.

Yannick Ulrich, Djuna Croon, Jeremy Sakstein, Samuel D. McDermott

Institute for Particle Physics Phenomenology at Durham University · Institut f¨ur Theoretische Physik & AEC at Universität Bern, Switzerland · Department of Physics & Astronomy at the University of Hawai’i · Department of Astronomy and Astrophysics at the University of Chicago

astro-ph.HE, astro-ph.CO, gr-qc, hep-ph

Submitted: 2024-06-10

Updated: 2026-10-28

Comments: Note added, 9 pages, 4 figures, 2 tables. Code available at https://github.com/multipass-black-holes/multipass, supplementary material at https://zenodo.org/records/11242245

Code: https://github.com/multipass-black-holes/multipass

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 88/100

The gist: Multi-Generational Black Hole Population Analysis with an Astrophysically Informed Mass Function analyzes black hole population statistics from LIGO/Virgo/KAGRA data using a parametric mass function…

Key concepts

Mass Function Modeling
This involves creating a mathematical formula that describes how many black holes of different masses are expected based on their formation history. The authors used a specific parametric model derived from stellar structure simulations, which accounts for both the initial collapse and subsequent mergers to better fit the observed gravitational wave data.
Black Hole Mass Gap (MBHMG)
This is a predicted range of black hole masses where formation via standard stellar processes is theoretically suppressed. The study measured this gap's lower edge at approximately 84.05+17.19−12.88 solar masses, suggesting that black holes in the 50M⊙ to 120M⊙ range are less likely to form through the specific stellar collapse mechanism modeled.
Spin Distribution Analysis
This technique separates black holes based on how they were formed—either from isolated stellar evolution or prior mergers. The analysis showed that first-generation black holes tend to have smaller spin values, consistent with theory, while merger-formed black holes show a different characteristic spin magnitude.
Spectral Siren Method
This method uses features in the gravitational wave mass function to estimate cosmological parameters like the Hubble constant (H0). By analyzing these features, researchers found an H0 value of 36.19+17.50−10.91 km/s/Mpc, which serves as a measurement tool for the expansion rate of the universe.

Terminology

Summary

Multi-Generational Black Hole Population Analysis with an Astrophysically Informed Mass Function analyzes black hole population statistics from LIGO/Virgo/KAGRA data using a parametric mass function derived from pulsational pair-instability supernovae simulations to separate black holes formed through stellar collapse from those formed via prior mergers, providing constraints on formation pathways and cosmology

Mass Function Modeling

The study employs a parametric mass function that accounts for different generations of black holes, modeling the total probability distribution as a sum of terms representing various binary formation channels. The mass function for first-generation black holes is derived from one-dimensional stellar structure simulations and incorporates the effects of pulsational pair-instability supernovae (PPISN). p. 1 This function includes a turn-on function at the low mass end, which is derived from analysis in a prior work. The authors compare their PPISN-inspired mass function to the phenomenological power law+peak (PLP) model, finding that their model provides a better fit to the data with a Bayes factor of log ZPPISN − log ZPLP = 9. p. 17 ± 0.1.

Constraints on Black Hole Mass Gap

A key result is the measurement of the lower edge of the upper black hole mass gap, which is found to be MBHMG = 84.05+17.19 −12.88M⊙. This location is consistent with predictions from stellar structure theory and results from prior work. p. 7 The authors argue that this feature is a prediction of stellar structure theory, resulting pairinstability supernova, which prevent BHs with masses in the approximate range 50M⊙–120M⊙ from forming. p. 3 Furthermore, the posterior for MBHMG is bimodal and weakly correlated with λ12, λ21. p. 4

Spin Distribution Analysis

The analysis separates the spin distributions of stellar remnant black holes from those formed via prior mergers using different mass functions. The results indicate that the 1g population prefers smaller values of the dimensionless spin parameter χ than the 2g population. p. 5 This is consistent with stellar structure theory for first-generation black holes and dynamical merger scenarios for higher-generation black holes. p. 0

Cosmological Measurements via Spectral Sirens

The spectral siren method was applied to measure the Hubble constant H0 using features in the BHMF, which breaks the degeneracy between BH mass and source redshift. The model yields a peaked distribution at H0 = 36. p. 119+17.50 −10.91 km/s/Mpc, which is discrepant with SH0ES and Planck measurements at around 2σ. p. 6 The model also shows a correlation between H0 and MBHMG, preferring larger MBHMG and smaller values of H0. p. 5

Discussion on Discrepancies

The discrepancy in the measured H0 value suggests that the model may be incomplete, with possibilities including redshift-dependence of parameters or a dependence on the ratio between stellar and dynamically formed black holes. The authors suggest that future modifications to their formalism may shed light on this issue. p. 6 The model is preferred over the canonical PLP model, but the low H0 value is noted as being discrepant with cosmological measurements. p. 7 The authors also suggest that a physical origin for a potential peak at M ∼ 35M⊙ could be explored in future work. p. 6

Conclusions

The framework successfully distinguishes between black hole subpopulations originating from stellar collapse and prior mergers, yielding results compatible with theoretical expectations regarding spin distributions and the mass gap location. The model provides a better fit to the LVK O3 data than the phenomenological power law+peak model. p. 5 However, the measured H0 value is noted as being discrepant with SH0ES and Planck measurements at a level of 2σ, suggesting that their model may be incomplete. p. 7 The study concludes by suggesting followup investigations to understand the origins of this discrepancy. p. 7

Software

The analysis was implemented using a new code called MULTIPASS — MULTIPopulation Analysis of Stellar and Secondary black holes — which uses the FORTRAN library Multinest for Bayesian inference. The code's modular design allows for easy addition of new mass functions or features.

Acknowledgements

The authors thank Eric Baxter and Reed Essick for useful discussions, and the work is supported by the STFC under Grant No. ST/T001011/1. p. 7 The simulations were run on the University of Hawai‘i’s high-performance supercomputer KOA. p. 7 The technical support and advanced computing resources from University of Hawai‘i Information Technology Services – Cyberinfrastructure, funded in part by the National Science Foundation MRI award 1920304, are gratefully acknowledged. p. 7

References

The paper cites numerous references supporting its methodology and findings, including work on stellar structure theory, black hole merger scenarios, and cosmological measurements. The list of references is extensive, covering topics from gravitational wave observations to theoretical astrophysics. p. 7 The paper includes references to specific mass function derivations and spin distributions used in the analysis. p. 1 The citations cover a wide range of astrophysical literature relevant to black hole formation and cosmology. p. 0 p. 0 The references are listed at the end of the paper, providing a comprehensive bibliography for the research presented. The paper includes references to specific data catalogs used in the analysis, such as GWTC-3. p. 1 p. 1 The citations demonstrate a strong grounding in established astrophysical principles and observational constraints.

Data Analysis Details

The data analysis was performed using the injection campaign of [26] with a false alarm rate (FAR) filter of FAR < 1. The selection process involved selecting all events up to and including GWTC-3 with FAR < 1 except for several specific events like GW170817, GW190814, and others. p. 3 This selection leaves the researchers with 69 merger events for detailed analysis. p. 3 The parameters and prior ranges used in the analysis are summarized in Table I. p. 3

Parameter Ranges

Table I provides the prior ranges for various parameters used in the data analysis, including mass function parameters, spin distribution parameters, and spectral siren analysis priors. The parameter ranges are detailed across several tables in the paper. p. 3

Results Summary

Table II presents the best-fit values and error estimates of the mass function analysis, summarizing the inferred parameters for both mass and spin functions. The results show good correspondence between the mass function parameters when spin information is included compared to when it is not. p. 4

Figure Analysis

Figure 1 displays the 95% confidence intervals of the mass function for the primary black hole in both models, showing a better fit for the PPISN model. Figure 2 presents a corner plot of parameters in the mass function, comparing fits without and with spin information. p. 4 Figure 3 shows the marginalized posterior probability distributions for H0 for both the PLP and their astrophysically-informed scenario. p. 5 Figure 4 illustrates how H0 correlates with MBHMG in the PPISN model, preferring larger MBHMG and smaller values of H0. p. 5

Final Conclusions

The study concludes that the spin distributions are compatible with theoretical expectations for first-generation and second-generation black holes. The model is preferred over the canonical LIGO/Virgo/KAGRA power law+peak model, but the value of H0 measured is discrepant with SH0ES and Planck measurements at a level of 2σ. p. 7 The paper suggests followup investigations to understand the origins of this discrepancy. p. 7 The authors also suggest that the lack of a peak feature in their BHMF might be due to redshift-dependence or other factors. p. 6 The study provides a framework for interpreting gravitational wave catalog data to provide information about black hole formation and cosmology. p. 0

The software used for the analysis includes Multinest 3.12, Getdist 1. p. 7 p. 7 p. 74.7, Astropy 5. p. 7 p. 7 p. 73.4, Matplotlib 3. p. 7 p. 7 p. 78.0, Numpy 1. p. 7 p. 7 p. 726.1, Scipy 1. p. 7 p. 711.3, and h5py 3.9

The paper includes a comprehensive list of references supporting its methodology and findings, covering topics from gravitational wave observations to theoretical astrophysics. The bibliography is extensive, covering specific mass function derivations and spin distributions used in the analysis. p. 0 The citations cover a wide range of astrophysical literature relevant to black hole formation and cosmology. The references are listed at the end of the paper, providing a comprehensive bibliography for the research presented. The paper includes references to specific data catalogs used in the analysis, such as GWTC-3. The citations demonstrate a strong grounding in established astrophysical principles and observational constraints.

The software used for the analysis includes Multinest 3.12, Getdist 1.4.7, Astropy 5.3.4, Matplotlib 3.8.0, Numpy 1.26.1, Scipy 1.11.

The software used for the analysis includes Multinest 3.12, Getdist 1.4.7, Astropy 5.3.4, Matplotlib 3.8.0, Numpy 1.26.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved black hole population modeling allows AI systems to distinguish between first-generation and higher-generation black holes by analyzing spin distributions, as the 1g population prefers larger values of beta1 than alpha1, which translates into a preference for small values of the spin parameter chi close to zero.

  2. Enhanced mass function fitting enables AI to determine the location of the upper black hole mass gap with high confidence, specifically measuring MBHMG = 84.05+17.19 −12.88M⊙, which provides evidence that the 35M⊙ Gaussian peak detected in the data using other models is not associated with the PPISN pile-up predicted to precede this gap.

  3. The spectral siren method, when integrated into an AI framework, allows for cosmological parameter estimation by utilizing features in the black hole mass function to find H0 = 36.1917.50 −10.91 km/s/Mpc, which is used to measure the Hubble constant using the spectral sirens method.

  4. AI systems can perform Bayesian inference on gravitational wave catalogs by implementing complex parametric models, such as the one described in equation (2), yielding a better fit than phenomenological models with a Bayes evidence factor of log ZPPISN − log ZPLP = 9.7 ±0.1.

  5. The system can analyze spin distributions to separately measure the spin of different black hole generations, noting that "black holes formed in the dynamical merger scenario should have a normalised spin magnitude tending to 0.7, whereas black holes formed in isolated stellar evolution are predicted to be born with negligible dimensionless spin magnitudes."

Sources

Related papers