Recovering cosmological parameters from the mock gravitational wave data of the Einstein Telescope

This paper presents a fast and effective technique using the intrinsic chirp mass spectrum of binary black holes to demonstrate that one year of Einstein Telescope observations can constrain the Hubble constant to 1% and the matter density parameter to 4% using gravitational wave spectral sirens.

The Gist

The Big Picture: Listening to the Universe's "Chirps"

Imagine the universe is a giant, dark concert hall. For a long time, we couldn't hear the music because our ears (our telescopes) weren't sensitive enough. Now, we are building a super-sensitive set of ears called the Einstein Telescope (ET). This new telescope will be ten times better at hearing than our current ones.

When two heavy objects like black holes crash into each other, they make a sound—a "chirp"—that ripples through space. These ripples are called gravitational waves. The Einstein Telescope will hear millions of these chirps every year.

The goal of this paper is to see if we can use these millions of "songs" to measure two very important things about our universe:

1. How fast the universe is expanding (The Hubble Constant, or $H_0$).
2. How much "stuff" (matter) is in the universe (The Matter Density, or $\Omega_m$).

The Problem: The "Volume Knob" Mystery

Here is the tricky part. When we hear a chirp, we can tell how loud it is. But in space, a loud sound could mean two things:

1. The source is close to us but quiet.
2. The source is far away but very loud.

This is like hearing a car horn. If you hear a faint horn, is it a quiet car nearby, or a loud truck far away? In astronomy, this is called a "degeneracy." We can't tell the distance just by listening to one sound.

Usually, astronomers solve this by looking for a visual flash of light (like a camera flash) to see exactly where the sound came from. But most black hole collisions don't make a flash. They are "dark sirens."

The Solution: The "Spectral Siren" Method

The authors of this paper came up with a clever trick called the Spectral Siren method. Instead of looking at one sound, they look at the entire library of sounds the telescope hears.

The Analogy: The Orchestra of Mass
Imagine you have a massive orchestra playing instruments of different sizes. You know the "standard" distribution of instrument sizes in this orchestra (e.g., there are many small violins, fewer medium cellos, and very few giant tubas). This is the intrinsic chirp mass spectrum.

When the sound travels through the expanding universe, it gets stretched. A small instrument might sound like a medium one because of the stretch.

• If you assume the universe is expanding at Speed A, the small instruments will look like medium ones.
• If you assume the universe is expanding at Speed B, the small instruments will look like giant ones.

By comparing the "stretched" sounds we hear against the "standard" distribution of instruments we expect, we can figure out exactly how much the sound was stretched. This tells us the distance and, consequently, how fast the universe is expanding.

What They Did (The Experiment)

Since we don't have the Einstein Telescope running yet, the authors built a virtual simulation (a "mock" universe).

1. They used a computer program to create 1 million fake binary star systems (pairs of black holes and neutron stars).
2. They simulated the Einstein Telescope listening to these systems for a year.
3. They "injected" specific values for the expansion speed and matter density into the simulation.
4. They then tried to "recover" those values using only the sound data, pretending they didn't know the answers beforehand.

The Results: How Well Did It Work?

They ran the simulation many times with different scenarios. Here is what they found:

• Measuring the Expansion Speed ($H_0$):
  If they only wanted to measure the expansion speed, they found that after one year of listening, they could pin down the speed with 1% accuracy. That is incredibly precise!

  • Analogy: It's like listening to a symphony for a year and being able to say, "The conductor is beating time at exactly 60 beats per minute, plus or minus 0.6."

• Measuring the Matter Density ($\Omega_m$):
  If they wanted to measure how much matter is in the universe, they could get within 4% accuracy with the same amount of data.

  • Analogy: They could estimate the total weight of the orchestra with a 4% margin of error.

• The "Systematic Error" Catch:
  The paper also tested what happens if we aren't 100% sure about the "standard" distribution of instruments (the mass spectrum).

  • If we have a little bit of uncertainty about the instruments, the accuracy drops.
  • Interestingly, if we just keep listening longer (more data), the accuracy doesn't improve as fast as we might hope if that initial uncertainty exists. It's like trying to tune a radio: if the station is slightly off-frequency, turning up the volume (getting more data) doesn't fix the static as well as it would if the station were perfectly tuned.

The Bottom Line

The authors conclude that the Einstein Telescope, acting alone, will be a powerful tool for cosmology. By using the "Spectral Siren" method—comparing the sounds of millions of colliding black holes against a known pattern of masses—we can measure the expansion of the universe with high precision, even without seeing any light.

Key Takeaways from the paper:

• 1 year of data = 1% accuracy on the universe's expansion speed.
• 1 year of data = 4% accuracy on the amount of matter in the universe.
• The method relies on the statistical pattern of black hole masses, not on finding individual host galaxies.
• The accuracy depends heavily on how well we understand the "standard" distribution of black hole masses. If our understanding of that distribution is fuzzy, our measurements of the universe will be fuzzier too.

Technical Summary

1. Problem Statement

The primary challenge in using gravitational waves (GWs) for cosmography is the degeneracy between the source redshift ($z$) and the intrinsic chirp mass ($\mathcal{M}$) of merging compact binaries.

• The Issue: GW detectors measure the redshifted chirp mass ($\mathcal{M}_z = \mathcal{M}(1+z)$) and the luminosity distance ($d_L$). Without an electromagnetic counterpart (a "bright siren") to identify the host galaxy and provide an independent redshift, the intrinsic mass and redshift cannot be uniquely determined from a single event ("dark siren").
• The Goal: The authors aim to break this degeneracy statistically using the Einstein Telescope (ET), a proposed third-generation GW detector. By leveraging the known shape of the intrinsic chirp mass distribution of binary black holes (BBHs) and binary neutron stars (BNSs), they seek to infer cosmological parameters—specifically the Hubble constant ($H_0$) and the matter density parameter ($\Omega_m$)—without relying on electromagnetic counterparts.

2. Methodology

A. Mock Data Generation

The authors generated a synthetic catalog of GW events to simulate ET observations:

• Population Synthesis: Used the COMPAS code to evolve 1 million zero-age main sequence (ZAMS) binaries across 76 metallicity values ($Z = 0.0001 - 0.03$).
• Event Selection: Identified compact objects (NS if $<2.5 M_\odot$, BH if $>2.5 M_\odot$) and applied a binary fraction of 0.5.
• Redshift Distribution: Merging rates were calculated based on a metallicity-dependent star formation rate (SFR) model (Chruślińska et al. 2021) up to a redshift of $z \sim 10$.
• Detection Criteria: Events were filtered based on ET-D (triangular design) sensitivity. An event was considered "detected" if the signal-to-noise ratio (SNR) in any single interferometer was $\ge 2$ and the effective SNR ($\rho_{eff} = \sqrt{\rho_1^2 + \rho_2^2 + \rho_3^2}$) was $\ge 5$.
• Observables: For detected events, the authors simulated the recovery of:
  • Redshifted chirp mass ($\mathcal{M}_z$) with Gaussian errors.
  • Luminosity distance ($d_L$) derived from SNR ratios and phase differences between the three ET interferometers to constrain sky location and inclination.

B. Cosmological Inference (The Spectral Siren Method)

The core of the analysis is a Monte Carlo approach using the Kullback-Leibler (KL) Divergence:

1. Iterative Sampling: For a given iteration, the authors sample $d_L$ and $\mathcal{M}_z$ from their posterior distributions for all $N$ detected events.
2. Redshift Conversion: Assuming a trial cosmological parameter (e.g., a specific $H_0$), they convert $d_L$ to redshift $z$.
3. Mass Reconstruction: They calculate the intrinsic chirp mass for each event: $\mathcal{M} = \mathcal{M}_z / (1+z)$.
4. Distribution Comparison: They construct the empirical probability distribution $P_{ij}(\mathcal{M})$ of the reconstructed masses and compare it to the true injected intrinsic mass distribution $P(\mathcal{M})$ (derived from the COMPAS simulation).
5. Optimization: The KL divergence ($D_{KL}$) is calculated between the empirical and true distributions. The value of the cosmological parameter that minimizes the KL divergence is selected as the best estimate for that iteration.
6. Aggregation: This process is repeated 10,000 times to build a posterior distribution for the cosmological parameters.

3. Key Contributions

• Novel Application of KL Divergence: The paper introduces the use of KL divergence as a metric to statistically break the mass-redshift degeneracy for dark sirens, a method not previously applied to this specific cosmography problem.
• Inclusion of Low-SNR and High-Redshift Events: Unlike some previous studies that filter for high-SNR events, this analysis includes a broad range of events (including low-SNR) to maximize the statistical power of the ET catalog.
• Systematic Uncertainty Analysis: The authors explicitly model the impact of systematic uncertainties in the "fixed" cosmological parameters (e.g., if $\Omega_m$ is known only to within 3%) on the precision of the inferred parameter.
• Scaling Laws: They derive a general formula for net uncertainty ($\varsigma_{net}$) that accounts for both statistical noise and systematic errors, showing that the standard $1/\sqrt{N}$ scaling breaks down when systematic errors are present.

4. Key Results

A. Parameter Recovery (Statistical Only)

Using $\sim 5.5 \times 10^4$ detected events (simulating roughly 1 year of ET observation):

• Hubble Constant ($H_0$):
  • Recovered with ~3.5% uncertainty (90% credible interval).
  • Example: Injected $H_0 = 67.3$ km/s/Mpc $\rightarrow$ Recovered $67.8^{+2.6}_{-2.2}$.
  • Example: Injected $H_0 = 73.5$ km/s/Mpc $\rightarrow$ Recovered $73.3^{+3.2}_{-2.2}$.
• Matter Density ($\Omega_m$):
  • Recovered with ~14% uncertainty when $H_0$ is fixed.
  • Example: Injected $\Omega_m = 0.27$ $\rightarrow$ Recovered $0.262^{+0.032}_{-0.042}$.

B. Scaling to 1% Precision

• To achieve 1% uncertainty on $H_0$, the authors estimate that $7 \times 10^5$ events (approximately one year of continuous ET observation) are required.
• Under the same conditions (1 year of data), $\Omega_m$ can be constrained to within 4% uncertainty.

C. Impact of Systematic Uncertainties

• When the "known" parameter (e.g., $\Omega_m$) has a systematic uncertainty, the improvement in the precision of the target parameter ($H_0$) deviates from the ideal $1/\sqrt{N}$ rule.
• The authors define a scaling factor $\kappa$ to quantify this:
  • For $H_0$, $\kappa \approx 0.44$.
  • For $\Omega_m$, $\kappa \approx 6.0$.
• This indicates that $\Omega_m$ is significantly more sensitive to systematic errors in $H_0$ than vice versa.

5. Significance and Conclusion

• Standalone Cosmography: The study demonstrates that the Einstein Telescope, operating as a standalone instrument (without electromagnetic follow-up), can effectively constrain the Hubble constant and matter density using the "spectral siren" method.
• Resolving the Hubble Tension: The ability to constrain $H_0$ to 1% precision with one year of data suggests ET could play a crucial role in resolving the current tension between early-universe (CMB) and late-universe (supernova) measurements of the Hubble constant.
• Limitations: The method relies heavily on the accuracy of the intrinsic chirp mass distribution model ($P(\mathcal{M})$). If the true astrophysical mass spectrum differs from the model (e.g., due to different formation channels like dynamical capture in clusters), the cosmological inference could be biased.
• Future Work: The authors note that accounting for Earth's rotation (which improves localization) would further reduce distance errors, and future studies will investigate the robustness of the method against variations in the intrinsic mass spectrum.

In summary, this paper provides a robust statistical framework for using the Einstein Telescope to perform precision cosmography, proving that the sheer volume of dark siren events will allow for the independent measurement of fundamental cosmological parameters.