🔭 astrophysics

Forbidden Formation Histories: The Binary Black Hole Merger Rate Disfavors Long Delay Times

By deconvolving the observed binary black hole merger rate evolution, this study demonstrates that long delay time distributions are physically forbidden due to low-redshift overpredictions, thereby constraining binary evolution channels and revealing a steeper decline in progenitor formation rates at low redshifts compared to the global star formation history.

Original authors: Aryanna Schiebelbein-Zwack, Maya Fishbach

Published 2026-05-14

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Aryanna Schiebelbein-Zwack, Maya Fishbach

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the universe as a giant, bustling factory that has been running for billions of years. This factory produces "black hole couples" (binary black holes). Sometimes, these couples are born, and then, after a long wait, they crash into each other.

This paper is about figuring out how long that wait usually is and how many couples were born in the first place, based on the crashes we see happening right now.

Here is the story of the paper, broken down into simple concepts:

1. The "Echo" Problem

Think of the black hole crashes we detect today as echoes.

If a black hole couple was born 1 billion years ago and crashed immediately, we hear that "echo" today.
If a couple was born 10 billion years ago but waited 9 billion years to crash, we also hear that "echo" today.

The problem is: We only hear the crash. We don't know exactly when the couple was born or how long they waited. The paper tries to work backward from the "crash sound" to figure out the "birth history."

2. The Mathematical "Un-mixing"

The authors use a clever math trick (called deconvolution) to "un-mix" the signal.

The Mix: The number of crashes we see today is a mix of all the couples born in the past, waiting different amounts of time.
The Goal: They take the data of crashes we actually see (from the LIGO/Virgo/KAGRA observatories) and try to separate it into two parts:
1. The Birth Rate: How many couples were formed at different times in the past.
2. The Wait Time: How long those couples typically waited before crashing.

3. The "Forbidden" Histories

This is the most exciting part of the paper. The authors realized that some combinations of "Wait Time" and "Birth Rate" are impossible.

Imagine you are trying to fill a bucket with water (the crashes we see today).

Scenario A: You have a hose that drips very slowly (long wait times). If you assume the hose dripped slowly for a very long time, you would expect the bucket to be overflowing with water by now.
The Conflict: But the bucket isn't overflowing; it only has a little bit of water.
The Conclusion: If your "slow drip" theory predicts too much water, that theory is forbidden. To make the math work with that slow drip, you would have to assume the hose was sucking water out of the bucket in the past. Since you can't have "negative water" (or negative star births), that theory is physically impossible.

The paper says: "If your theory predicts that the universe should have had a huge number of black hole crashes in the past, but we don't see them, your theory is wrong."

4. What They Found

Using data from the latest gravitational wave catalog (GWTC-4.0), they tested many different theories about how long black holes wait before crashing.

The "Long Wait" Theories are Out: Many theories suggest black holes form and then wait a very long time (billions of years) before crashing. The authors found that these theories are disfavored. If black holes waited that long, we would see way more crashes today than we actually do.
The "Short Wait" Theories Win: The data fits best with theories where black holes form and crash relatively quickly (in cosmic terms).
The "Birth Rate" Surprise: When they calculated how many black hole couples were actually born in the past, they found the number drops off much faster as we get closer to "today" than the general rate of star formation does. It's like the factory stopped making these specific couples much sooner than it stopped making stars in general.

5. A Simple Analogy: The Party

Imagine a party that started 10 years ago.

The Observation: You walk in today and see 10 people leaving the party.
Theory 1 (Long Wait): You assume everyone who came to the party stayed for 9 years before leaving. If this were true, you would expect to see thousands of people leaving today (because so many people came over the last 9 years). But you only see 10. So, this theory is wrong.
Theory 2 (Short Wait): You assume everyone who came left within 1 year. This fits the data better. You only need a small number of people to have arrived recently to explain the 10 people leaving today.

Summary of the Main Takeaway

The paper doesn't just guess how black holes form; it uses the "rules of physics" (you can't have negative births) to rule out theories that don't fit the data.

They found that the universe is not full of black hole couples that waited billions of years to crash. Instead, the couples that are crashing today likely formed and crashed much sooner. This helps scientists narrow down the specific physics of how these cosmic couples are made, ruling out the "slow and steady" scenarios in favor of "fast and furious" ones.

Technical Summary: Forbidden Formation Histories: The Binary Black Hole Merger Rate Disfavours Long Delay Times

Problem Statement
The binary black hole (BBH) merger rate, $R_{\text{merge}}(t)$ , is the convolution of the progenitor formation rate, $R_{\text{form}}(t)$ , and the distribution of time delays between formation and merger, $p(\tau)$ . Traditional approaches to understanding BBH origins often rely on forward modeling: assuming specific functional forms for $R_{\text{form}}$ (e.g., tied to the global star formation rate) and $p(\tau)$ (derived from population synthesis), then comparing the predicted merger rate to gravitational wave (GW) data. This approach suffers from inherent degeneracies; different combinations of formation histories and delay times can produce identical merger rates, and rigid parametric assumptions may obscure data features or force physically inconsistent results. Furthermore, strong models for $R_{\text{form}}$ and $p(\tau)$ may fundamentally mismatch the observed redshift evolution of the merger rate.

Methodology
The authors propose a "backward-modeling" framework to deconvolve the observed BBH merger rate into its constituent components without assuming a specific functional form for the progenitor formation rate.

Mathematical Framework: Utilizing the convolution theorem, the authors transform the integral equation $R_{\text{merge}}(t) = \int R_{\text{form}}(t-\tau)p(\tau)d\tau$ into the frequency domain via the Fast Fourier Transform (FFT). This allows for the direct reconstruction of $R_{\text{form}}$ given an observed $R_{\text{merge}}$ and a known $p(\tau)$ :
$R_{\text{form}} = \mathcal{F}^{-1} \left\{ \frac{\mathcal{F}\{R_{\text{merge}}\}}{\mathcal{F}\{p(\tau)\}} \right\}$
Numerical Implementation: To handle the non-periodic nature of the data and the finite observation horizon (GWTC-4.0 data up to $z \approx 1.5$ ), the authors employ zero-padding to prevent wrap-around artifacts and apply cosine tapering to mitigate Gibbs phenomenon noise at the boundaries. They assume a conservative boundary condition where $R_{\text{merge}}(z > 1.5) = 0$ , though they test asymptotic values up to 500 Gpc $^{-3}$ yr $^{-1}$ .
Regularization: To ensure numerical stability when dividing by the Fourier transform of the delay time distribution (which may approach zero), they utilize Wiener deconvolution, adding a regularization term $\lambda$ to the denominator.
Data Inputs: The method is applied using the non-parametric B-spline fit to the BBH merger rate from GWTC-4.0 and a suite of delay time distributions generated by the COMPAS population synthesis code, which vary parameters related to mass transfer, supernovae, and remnant masses.

Key Contributions and Results

Reconstruction of Formation Histories: By deconvolving GWTC-4.0 merger rates with COMPAS delay time distributions, the authors reconstruct the implied progenitor formation rate. They find that the inferred $R_{\text{form}}$ declines more steeply toward low redshift than the global star formation rate (SFR), consistent with metallicity-dependent BBH formation efficiency.
Identification of "Forbidden" Histories: The study demonstrates that certain combinations of merger rates and delay time distributions are physically impossible because they require $R_{\text{form}}$ to become negative at late times. This occurs when a delay time distribution with a long tail (shallow power-law index) predicts more mergers from ancient high-redshift formation than are actually observed at low redshift.
Constraints on Delay Time Distributions: Using the requirement that $R_{\text{form}} > 0$ $R_{form} > 0$ , the authors constrain the delay time distribution power-law index $\alpha$ $α$ (where $p(\tau) \propto \tau^\alpha$ $p (τ) \propto τ^{α}$ ).
- Delay time distributions with shallow indices ( $\alpha \gtrsim -0.7$ ), characteristic of stable mass transfer channels, are strongly disfavoured.
- Specifically, for the GWTC-4.0 data, $\alpha < -1.08$ is preferred over $\alpha > -0.67$ by a factor of 19.
- COMPAS simulations with steep distributions (favoring short delays, $\alpha < -1.2$ ) are consistent with nearly all GWTC-4.0 merger rate posterior draws, whereas shallow distributions are compatible with only $\sim30\%$ of the draws.
Constraints on Binary Evolution Physics: As a proof of concept, the authors constrain the mass transfer efficiency parameter $\beta$ in COMPAS. Lower $\beta$ values (non-conservative mass transfer, leading to orbital shrinkage and short delays) show higher compatibility with GW data. High $\beta$ values (conservative mass transfer, long delays) are disfavoured by a factor of $\approx 3.7$ .
Impact on Merger Rate Inference: Imposing the physical constraint of a positive formation rate as a prior on the GWTC-4.0 B-spline model reduces the median inferred merger rate at $z=1.5$ by approximately 10% (from 334 to 299 Gpc $^{-3}$ yr $^{-1}$ ), favoring a shallower merger rate evolution than the standard data-driven inference.
Parameter Space Mapping: The authors map the forbidden region in the $\alpha$ – $\kappa$ parameter space (where $\kappa$ describes the merger rate evolution $R_{\text{merge}} \propto (1+z)^\kappa$ ). They show that for a given $\kappa$ , there is a critical $\alpha_{\text{crit}}$ above which the formation rate becomes unphysical. Current data excludes power-law delay time distributions with $\alpha \gtrsim -0.7$ .

Significance
The paper presents a model-independent framework for ruling out regions of binary evolution and merger rate parameter space. By directly evaluating the compatibility of population synthesis parameters with gravitational wave observations through the lens of causality (i.e., the formation rate cannot be negative), the method bypasses the need for high-dimensional priors on star formation and metallicity evolution. The authors argue that this approach exposes "forbidden" formation histories that forward modeling might miss, providing a robust consistency check for stellar binary evolution theories. The results suggest that formation channels predicting long delay times (such as stable mass transfer) are in tension with the observed steep decline in the BBH merger rate at low redshift, favoring channels that produce shorter delays.