🔭 astrophysics

Towards precision cosmology with Void x CMB correlations (II): Impact of mock catalogs on the Void x CMB lensing signal

Using validated Roman mock catalogs, this paper demonstrates that the Void x CMB lensing signal is robust against methodological choices and forecasts a significant signal-to-noise ratio (up to 31 $\sigma$ ) with future CMB surveys, paving the way for the first direct cosmological constraints from this observable.

Original authors: Mar Pérez Sar, Carlos Hernández Monteagudo, András Kovács, Alice Pisani

Published 2026-05-22

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Mar Pérez Sar, Carlos Hernández Monteagudo, András Kovács, Alice Pisani

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

The Big Picture: Finding "Empty Rooms" in the Universe

Imagine the universe as a giant, three-dimensional city. Most of the "buildings" in this city are galaxies, but there are also massive, empty spaces between them called cosmic voids. These are like giant, empty plazas or vacant lots in the cosmic city.

Scientists have known about these empty spaces for decades, but they are now trying to use them as a new tool to measure how the universe works. Specifically, they want to see how these empty spaces bend light from the very beginning of time (the Cosmic Microwave Background, or CMB).

Think of the CMB as a giant, glowing wallpaper covering the entire universe. When light from this wallpaper passes through a cosmic void, the emptiness of the void acts like a weak lens, slightly stretching the wallpaper pattern. By measuring how much the wallpaper is stretched, scientists can learn about the "weight" and structure of the universe.

The Problem: Are We Using the Right Map?

To test this idea, scientists use computer simulations (called "mock catalogs") to pretend they are looking at the universe. They build a fake universe on a computer, find the voids, and see how the light bends.

However, there is a catch: How you build the fake universe matters.

Do you fill the fake voids with the same number of stars as the real universe?
Do you place the stars in the same patterns?
Do you use different rules to define what a "void" is?

The authors of this paper were worried: If we change the rules of our computer simulation, does the result change too much? If the answer is "yes," then our measurements of the real universe might be wrong because our computer models are flawed.

The Experiment: Testing the "Recipe"

The authors used a very advanced computer simulation (based on the upcoming Roman Space Telescope data) to test this. They created 16 different versions of a fake universe. Some versions were simple (just matching the number of galaxies), while others were complex (matching galaxy mass, environment, and type).

They then ran their "void lensing" test on all 16 versions to see if the results changed.

The Surprise Finding:
They found that the results were surprisingly stable.

The Analogy: Imagine you are trying to measure the shadow of a tree. You can draw the tree using a stick, a pencil, or a thick marker. Even though the drawings look different, the shadow they cast on the ground is almost exactly the same.
The Result: Whether they used a simple or complex computer model, the "shadow" (the lensing signal) stayed the same. The differences between the models were so small (less than 10%) that they wouldn't confuse the scientists, even with very precise future telescopes.

The Best Tools: 2D vs. 3D and "Rescaling"

The paper also tested different ways to measure these voids:

2D vs. 3D:
- 3D Voids: Like looking at a hollow ball in space.
- 2D Voids: Like looking at a shadow of that ball on a flat wall.
- The Winner: The 2D (flat) approach worked better. It gave a clearer, stronger signal (about twice as strong) than the 3D approach. It's like looking at a shadow on a wall is often easier to measure than trying to measure the volume of the object casting it.
Rescaling vs. Not Rescaling:
- Rescaling: Imagine you have a pile of different-sized rubber bands. To compare them, you stretch them all to the same size before measuring. This makes the signal very clear.
- Not Rescaling: You measure them exactly as they are, big and small mixed together.
- The Winner: Rescaling (stretching them to a common size) generally gave the best results, especially when combining data from many voids.

The Future: What Can We Expect?

The authors looked ahead to what will happen when the Roman Space Telescope launches (planned for 2026) and teams up with powerful CMB telescopes like Planck, Simons Observatory (SO), and CMB-S4.

Current Status: Right now, with the Planck telescope, the signal is detectable but a bit "fuzzy" (like trying to hear a whisper in a noisy room).
The Future: With the new, ultra-sensitive telescopes (SO and CMB-S4), the "noise" will disappear.
- They predict they will be able to detect the signal with extreme confidence (up to 31 times stronger than the background noise).
- This is strong enough to finally use cosmic voids to measure the fundamental rules of the universe, such as the nature of dark energy or the weight of neutrinos.

Summary

This paper is a "quality control" check. The authors wanted to make sure that the tools scientists use to study cosmic voids aren't broken. They found that:

The tools are robust: Changing how you build the computer models doesn't break the results.
2D voids are the most effective way to measure the signal.
With the upcoming Roman Space Telescope and next-gen CMB telescopes, we are on the verge of using cosmic voids as a precision ruler to measure the universe's secrets.

Technical Summary: Towards precision cosmology with Void × CMB correlations (II): Impact of mock catalogs on the Void × CMB lensing signal

Problem Statement
While the cross-correlation between cosmic voids and Cosmic Microwave Background (CMB) lensing convergence ( $\kappa$ ) has been robustly measured, its utility as a precision cosmological probe is currently limited. The primary obstacles are the lack of a comprehensive analytical model for the observable and the uncertainty regarding how methodological choices—specifically the construction of mock catalogs, void definitions, and analysis techniques—affect the measured signal. Previous studies have reported tensions (deviations in the amplitude parameter $A_\kappa$ ) with $\Lambda$ CDM expectations, but it remains unclear whether these arise from new physics or artifacts introduced by the simulation and analysis pipeline. Furthermore, the dependence of this signal on cosmological parameters (e.g., neutrino mass) requires a rigorous understanding of systematic errors before it can be used for constraints.

Methodology
The authors utilize a suite of validated mock catalogs derived from the Agora full-sky simulation, which includes self-consistent CMB secondary anisotropies. To mimic the Nancy Grace Roman Space Telescope's High Latitude Wide Area Survey (HLWAS), they generate 16 distinct mock galaxy catalogs using an analog matching algorithm. These mocks vary in complexity, ranging from simple number-density matching to those incorporating halo mass, local environment, and galaxy-type proxies (stellar mass, specific star formation rate).

The study employs two void-finding approaches:

3D Voids: Identified using the REVOLVER algorithm on the 3D matter distribution.
2D Voids: Identified from projected density fields using a 2D void finder with varying smoothing scales ( $s_{mVF} = 5$ and $10 \, h^{-1}$ Mpc).

The methodology involves stacking CMB $\kappa$ maps at the positions of these voids to measure the lensing profile. The authors compare two stacking strategies:

Rescaled: Profiles are normalized by the void radius ( $R_v$ ) before stacking to maximize statistical power.
Non-rescaled: Voids are stacked by their intrinsic angular size to preserve physical accuracy.

Signal-to-noise ratios (S/N) are estimated using a likelihood approach that incorporates the full data vector and covariance matrices derived from 1,000 random CMB realizations. The analysis forecasts S/N for Roman combined with current (Planck) and future (Simons Observatory, CMB-S4-like) CMB experiments, testing various noise levels and map smoothing scales.

Key Contributions and Results

Robustness to Mock Construction: A central finding is that the Void $\times$ CMB lensing signal is significantly less sensitive to the specific details of mock catalog construction than galaxy or void clustering statistics. While variations in one- and two-point statistics across mocks can be substantial, the resulting lensing profiles show deviations typically below $2\sigma$ relative to the mean. For mocks that successfully reproduce the reference catalog's statistics, the profile variations drop to $\sim1-3\%$ for 2D voids and $\sim7-9\%$ for 3D voids. This suggests that future tensions with observational data are unlikely to stem solely from mock inaccuracies.
2D vs. 3D Performance: 2D voids consistently yield higher signal-to-noise ratios than 3D voids. When normalized by the number of voids, 2D voids are approximately twice as efficient at tracing the CMB lensing signal. This is attributed to the fact that 2D voids are defined from the same projected field that contributes to the lensing signal, leading to more stable profiles.
Stacking Strategies: The rescaled methodology generally provides the highest total S/N for the full sample by aligning features of different-sized voids. However, the non-rescaled approach can offer advantages in specific noise regimes (e.g., Planck noise with 3D voids) or when separating voids by type, though binning often reduces total S/N due to statistical losses.
Forecasts for Roman: The paper provides the first S/N forecasts for Roman voids combined with various CMB surveys.
- Planck: Roman 2D voids ( $s_{mVF}=5 \, h^{-1}$ Mpc) combined with Planck yield $S/N \approx 13\sigma$ .
- Simons Observatory (SO): The S/N increases to $\approx 22\sigma$ .
- CMB-S4-like: The S/N reaches $\approx 31\sigma$ .
- For 3D voids, the corresponding forecasts are $8\sigma$ (Planck), $13\sigma$ (SO), and $18\sigma$ (CMB-S4-like).
- These forecasts assume a Roman survey area of $\sim2000$ deg $^2$ , which is smaller than current surveys (e.g., DES, DESI) but offers higher tracer density and resolution.

Significance and Claims
The authors claim that this work establishes the robustness of the Void $\times$ CMB lensing observable against methodological systematics, a prerequisite for its use in precision cosmology. They argue that because the signal is less sensitive to the nuances of galaxy bias and mock construction than other probes, it is a reliable tool for future surveys.

The paper posits that while current Planck-based measurements are noise-limited and do not provide competitive constraints for the Roman footprint, the combination of Roman with next-generation CMB experiments (SO, CMB-S4) will enable the first direct cosmological constraints from Void $\times$ CMB lensing. The authors emphasize that the divergent behavior of 2D voids—where increasing smoothing scales reduces sensitivity to mock-building nuances while increasing physical sensitivity to neutrino mass—suggests that 2D void-lensing is particularly well-suited for breaking degeneracies between systematics and new physics.

The study concludes that while Roman's smaller sky coverage limits standalone constraints compared to wide-area surveys, its high tracer density and angular resolution will allow for the detection of smaller voids and improved mapping of their internal structure, opening a path toward precision cosmology when combined with other Large Scale Structure (LSS) and CMB datasets.