Bayesian model selection of Primordial Black Holes and Dressed Primordial Black Holes with lensed Gravitational Waves

This paper demonstrates that third-generation ground-based gravitational wave detectors, such as the Einstein Telescope and Cosmic Explorer, can effectively distinguish between bare primordial black holes and dressed primordial black holes surrounded by dark matter using Bayesian model selection on lensed gravitational wave signals.

The Gist

Here is an explanation of the paper, translated into everyday language with some creative analogies.

The Big Picture: Finding Invisible Ghosts in a Cosmic Storm

Imagine the universe is a giant, dark ocean. For decades, scientists have been trying to figure out what makes up the "dark matter" that holds galaxies together. One popular theory is that it's made of Particle Dark Matter (tiny, invisible dust). Another theory suggests it's made of Primordial Black Holes (PBHs)—black holes formed right after the Big Bang, like ancient, heavy anchors dropped in the ocean.

But what if both exist at the same time?

This paper asks a fascinating question: What happens if a Primordial Black Hole gets "dressed" in a coat of particle dark matter?

Think of a PBH as a lonely, naked black hole. Now, imagine it gets surrounded by a thick, fuzzy cloud of dark matter particles. We call this a "Dressed PBH" (dPBH). It's like a black hole wearing a heavy winter coat.

The Problem: The "Look-Alike" Confusion

The scientists in this paper are trying to tell the difference between a Naked PBH and a Dressed PBH.

They use Gravitational Waves (ripples in space-time caused by colliding black holes) as their flashlight. When these ripples pass near a black hole, the black hole acts like a magnifying glass, bending and amplifying the waves. This is called Gravitational Lensing.

Here is the tricky part:

• If the black hole is Naked, it bends the light in a specific way.
• If the black hole is Dressed (wearing that fuzzy coat), it also bends the light.

In the high-frequency range (where our current and future detectors listen), the "Naked" and "Dressed" versions look almost identical. It's like trying to tell the difference between a person wearing a heavy coat and a person who is just naturally very large. From a distance, they look the same size.

The Solution: The "Super-Listening" Ears

The authors propose using the next generation of gravitational wave detectors: the Einstein Telescope (ET) and Cosmic Explorer (CE). Think of these not just as microphones, but as super-sensitive ears that can hear the faintest whispers of the universe.

They used a mathematical tool called Bayesian Inference.

• The Analogy: Imagine you hear a sound in the dark. You have two guesses: "It's a cat" or "It's a small dog."
  • Bayesian Inference is the process of weighing the evidence. You look at the pitch, the volume, and the pattern of the sound.
  • If the sound matches the "cat" pattern 99% of the time, you bet on the cat.
  • If the sound has a tiny, unique quirk that only a "dog" would make, you bet on the dog.

The scientists ran simulations to see if these super-detectors could hear the tiny "quirks" that distinguish a Dressed Black Hole from a Naked one.

The Results: Who Won the Bet?

The study found that yes, we can tell them apart, but it depends on a few things:

1. The Size of the Coat: If the "coat" of dark matter around the black hole is heavy (meaning the black hole is surrounded by a lot of dark matter), the difference is easier to spot. It's like a heavy winter coat making a person look much bigger than they are.
2. The Frequency: The detectors need to listen to a wide range of frequencies. If the black hole collision is "loud" and happens over a long time (low mass black holes), the detectors get more data points, making it easier to spot the difference.
3. The "Smoking Gun": When the scientists simulated the data, the math (the Bayes Factor) screamed that the signal was definitely coming from a Dressed Black Hole, not a Naked one. The confidence level was incredibly high (over 74 times stronger than the alternative).

Why This Matters

This isn't just about math; it's about solving a cosmic mystery.

• If we can prove that black holes are "dressed" in dark matter, it means both particle dark matter and black holes exist together.
• It tells us how the universe grew up: Did black holes eat the dark matter around them as they formed?
• It proves that our future telescopes (ET and CE) are powerful enough to see details we couldn't even imagine before.

The Takeaway

Think of the universe as a giant puzzle. For a long time, we thought the pieces were either "Dark Matter Dust" OR "Black Hole Anchors." This paper suggests that sometimes, the anchors are covered in dust.

By using the most sensitive ears we can build, we can finally hear the difference between a naked anchor and a dusty one. If we can hear that difference, we unlock a new chapter in understanding what the universe is actually made of.

Technical Summary

Here is a detailed technical summary of the paper "Bayesian model selection of Primordial Black Holes and Dressed Primordial Black Holes with lensed Gravitational Waves."

1. Problem Statement

The paper addresses the challenge of distinguishing between two competing astrophysical models for dark matter candidates using gravitational wave (GW) lensing:

• Bare Primordial Black Holes (PBHs): Black holes formed in the early universe, modeled as point-mass lenses.
• Dressed Primordial Black Holes (dPBHs): PBHs surrounded by a halo of particle dark matter (DM), formed via accretion during the radiation- and matter-dominated eras.

The Core Challenge:
While dPBHs and bare PBHs exhibit distinct lensing signatures in the low-frequency (wave-optics/diffraction) regime, their amplification factors become nearly identical in the high-frequency geometric-optics regime. Since ground-based GW detectors (like LIGO, Virgo, and future third-generation detectors) are most sensitive to high-frequency signals (stellar-mass range, $10^{-1} \sim 10^2 M_\odot$), traditional visual inspection of waveforms is insufficient to differentiate these models. The authors aim to determine if Bayesian model selection can statistically distinguish between a bare PBH and a dPBH lens using third-generation detector data.

2. Methodology

A. Theoretical Modeling of Lensing

The authors model the lensed GW waveform $h_L(f)$ as the unlensed waveform $h_0(f)$ multiplied by an amplification factor $F(f)$.

• Wave-Optics Regime ($w < 6$): They employ the asymptotic expansion method to solve the diffraction integral numerically. This is crucial for dPBHs, which lack an analytical solution for the amplification factor due to their complex density profiles ($\rho \propto r^{-9/4}$).
• Geometric-Optics Regime ($w > 10$): They use the geometric optics approximation, summing contributions from multiple images.
• Post-Geometric Corrections: To improve accuracy in the intermediate-to-high frequency range, they incorporate post-geometric optics corrections ($\delta F_m$ and $\delta F_c$) to account for diffraction effects near the lens center.
• Specific Models:
  • PBH: Modeled as a point mass ($\rho \propto \delta(r)$).
  • dPBH: Modeled with a central point mass plus a surrounding dark matter halo with a specific density profile derived from cosmological accretion theories.

B. Bayesian Inference Framework

The study utilizes Bayesian model selection to compare the likelihood of the data ($d$) under two hypotheses:

1. $H_{dPBH}$: The lens is a dressed PBH.
2. $H_{PBH}$: The lens is a bare PBH.

• Likelihood Function: Calculated using the inner product of the observed strain data and the theoretical waveform, weighted by the detector's Power Spectral Density (PSD).
• Evidence Calculation: The authors use the nested sampling algorithm (implemented via the dynesty package) to compute the Bayesian evidence ($Z$) for both models.
• Bayes Factor ($BF$): Defined as $BF = Z_{dPBH} / Z_{PBH}$. The logarithmic Bayes factor ($\log BF$) is used as the decision metric. A threshold of $\log BF \geq 8$ is adopted to claim strong evidence favoring one model over the other.
• Waveform Model: The intrinsic GW signal is modeled using the IMRPhenomD phenomenological waveform, covering inspiral, merger, and ringdown phases.

C. Simulation Setup

• Detectors: Third-generation detectors Einstein Telescope (ET) and Cosmic Explorer (CE).
• Parameters: The study injects simulated signals with a dPBH lens and attempts to recover them under both $H_{dPBH}$ and $H_{PBH}$ hypotheses.
• Priors: Flat priors are used for intrinsic parameters (mass ratio $\eta$, total mass $M$, redshift $z_S$) and extrinsic parameters (angles). The lens mass ($m_{PBH}$) prior ranges are adjusted based on the hypothesis to ensure efficient sampling (e.g., $10-50 M_\odot$for dPBH vs.$10-300 M_\odot$ for PBH to account for the "dressing" effect mimicking a heavier bare mass).

3. Key Contributions

1. Development of dPBH Lensing Formalism: The paper provides a robust computational framework for calculating the amplification factor of dPBHs across all frequency regimes, combining asymptotic expansions for wave optics and geometric optics with post-geometric corrections.
2. Demonstration of Distinguishability: It proves that despite the similarity in high-frequency amplification factors, the full frequency-dependent phase and amplitude structure contains enough information for Bayesian inference to distinguish dPBHs from bare PBHs.
3. Parameter Dependence Analysis: The study systematically analyzes how the distinguishability (Bayes Factor) depends on:
   • Source Total Mass ($M$): Lower total masses yield wider frequency bands, improving distinguishability.
   • Symmetric Mass Ratio ($\eta$): Systems with equal masses ($\eta \approx 0.25$) provide the widest frequency range and highest Bayes factors.
   • Halo Mass ($m_{PBH}$): Larger halo masses (and thus larger effective lensing potentials) lead to higher Bayes factors, making the distinction easier.

4. Key Results

• Model Selection Success: In a baseline simulation with a dPBH lens ($m_{PBH} = 30 M_\odot$) and a binary source ($M=50 M_\odot, \eta=0.23$), the Bayesian analysis yielded a $\log BF = 74.43$. This is far above the threshold of 8, indicating that the data overwhelmingly favors the dPBH model over the bare PBH model.
• Parameter Recovery: When assuming the correct dPBH hypothesis, the posterior distributions for $\eta$, $M$, $z_S$, and $m_{PBH}$ closely matched the injected values. However, assuming the wrong hypothesis (bare PBH) resulted in significant biases in parameter estimation (e.g., inferring a much lower total mass and higher redshift).
• Mass Dependence:
  • For small total source masses ($M < 200 M_\odot$), the Bayes factor remains high ($>8$) even for smaller halo masses ($10 M_\odot$).
  • As $M$ increases, the frequency range narrows, causing the Bayes factor to drop. For $M > 200 M_\odot$, distinguishing the models becomes difficult unless the halo mass is large ($m_{PBH} \geq 50 M_\odot$).
• Symmetry Dependence: The ability to distinguish models is maximized when the binary components have equal masses ($\eta = 0.25$). Asymmetry ($\eta < 0.25$) reduces the frequency bandwidth, lowering the Bayes factor, though models with large halo masses ($30-50 M_\odot$) remain distinguishable even at low$\eta$.

5. Significance

• Probing Dark Matter Nature: This work establishes a viable pathway to test the nature of dark matter. If GW lensing events are detected by ET or CE, Bayesian analysis can determine if the lens is a "naked" PBH or a PBH surrounded by a DM halo. A detection of dPBHs would provide strong evidence for the coexistence of particle dark matter and PBHs.
• Third-Generation Detector Potential: The results highlight the critical importance of third-generation detectors (ET/CE). Their superior sensitivity in the high-frequency regime allows for the accumulation of sufficient signal-to-noise ratio and frequency coverage to break the degeneracy between these two models, which is impossible with current second-generation detectors.
• Methodological Advancement: The paper demonstrates the necessity of combining wave-optics calculations with geometric optics and post-geometric corrections to accurately model lensing in the transition regime, providing a template for future GW lensing studies involving complex lens profiles.