🔭 astrophysics

Ly{\alpha} Intensity Mapping in HETDEX: Galaxy-Ly{\alpha} Intensity Cross-Power Spectrum

This paper presents the first measurement of the Lyman- $\alpha$ intensity mapping cross-power spectrum with Ly $\alpha$ -emitting galaxies using HETDEX data, yielding bias and intensity constraints in the redshift range $1.9 < z < 3.5$ that are consistent with hydrodynamical simulations but significantly lower than previous quasar-based results.

Original authors: Maja Lujan Niemeyer, Eiichiro Komatsu, José Luis Bernal, Chris Byrohl, Robin Ciardullo, Olivia Curtis, Daniel J. Farrow, Steven L. Finkelstein, Karl Gebhardt, Caryl Gronwall, Gary J. Hill, Matt J. Jar

Published 2026-03-18

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Maja Lujan Niemeyer, Eiichiro Komatsu, José Luis Bernal, Chris Byrohl, Robin Ciardullo, Olivia Curtis, Daniel J. Farrow, Steven L. Finkelstein, Karl Gebhardt, Caryl Gronwall, Gary J. Hill, Matt J. Jarvis, Donghui Jeong, Erin Mentuch Cooper, Deeshani Mitra, Shiro Mukae, Julian B. Muñoz, Masami Ouchi, Shun Saito, Donald P. Schneider, Lutz Wisotzki

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 not as a collection of distinct islands (galaxies), but as a vast, foggy ocean. For decades, astronomers have been excellent at spotting the bright lighthouses (the brightest galaxies) in this ocean. But what about the fog itself? What about the faint, glowing mist that exists between the lighthouses?

This paper is about a new way of mapping that "fog" using a giant telescope called HETDEX (Hobby-Eberly Telescope Dark Energy Experiment).

Here is the story of how they did it, explained simply:

1. The Problem: Too Much Noise, Too Little Signal

The HETDEX telescope is like a massive net made of thousands of tiny fiber-optic strands, casting a wide net over the sky. It catches light from the early universe (when the universe was about 2 to 3 billion years old).

The problem is that most of the light the telescope catches is noise. It's like trying to hear a whisper in a stadium full of people shouting. The telescope sees:

Bright Galaxies: The "lighthouses" (detected galaxies).
The Sky: The background glow of our own atmosphere and the Milky Way.
The Fog: The faint, diffuse glow of hydrogen gas from billions of tiny, invisible galaxies and the space between them.

If you just look at the raw data, the "fog" is completely drowned out by the "shouting" of the bright galaxies and the "noise" of the sky.

2. The Solution: The "Silent" Map

The team decided to do something clever. Instead of trying to find every single galaxy, they decided to ignore the loud ones.

Step 1: The Mute Button. They identified the bright galaxies (the "LAEs") that were loud enough to be seen clearly. They then digitally "masked" or covered them up in their data, like putting a piece of black tape over a bright lightbulb.
Step 2: Listening to the Silence. Once the bright lights were covered, they looked at what was left. This remaining light is the "fog"—the combined glow of millions of tiny, invisible galaxies and the gas between them that no single telescope could ever see individually.

3. The Analogy: The Crowd at a Concert

Imagine a massive concert where you can't see the stage.

The Old Way: You try to count the famous singers on stage.
The New Way (Intensity Mapping): You turn off the microphones of the famous singers. You then listen to the hum of the crowd. Even though you can't see individual fans, the collective sound of the crowd tells you how big the crowd is, how excited they are, and how they are moving.

This paper is about measuring that "crowd hum" of the early universe.

4. The Challenge: Cleaning the Data

The data was messy. It had "static" from the Earth's atmosphere and glitches from the camera. To fix this, the scientists used a technique called Principal Component Analysis (PCA).

Think of this like noise-canceling headphones.

The scientists analyzed the "static" patterns that appeared in the data (like the hum of a fluorescent light or the wind).
They identified these patterns mathematically and subtracted them out, leaving behind only the true cosmic signal.
It was a delicate balance: if they removed too much, they accidentally deleted the cosmic signal too. If they removed too little, the noise remained. They had to find the "Goldilocks" zone.

5. The Discovery: A Faint but Real Signal

After all the cleaning, they compared their "fog map" with the locations of the bright galaxies they had masked out.

They found a statistical connection. The "fog" wasn't just random; it was clustered around the bright galaxies, just like a cloud of dust gathers around a streetlamp.

What they measured: They didn't measure the exact shape of the fog, but they measured its strength and clumpiness.
The Result: They successfully detected the cross-signal between the bright galaxies and the invisible fog. It's like hearing the echo of a shout in a canyon and realizing, "Ah, the canyon walls are there, even if I can't see them."

6. Why This Matters

It's a New Tool: This proves we can map the universe by looking at the "glow" of the invisible stuff, not just the bright stuff. This is crucial for understanding how the universe grew from a smooth soup into the clumpy web of galaxies we see today.
It's Different from Before: Previous studies tried to do this by looking at quasars (super-bright black holes). Those studies found a lot of light, but it might have been "contaminated" by the intense radiation of the black holes themselves. This study used normal star-forming galaxies, giving a cleaner picture of the universe's background glow.
The "Fog" is Fainter than Expected: Surprisingly, the "fog" they found is dimmer than some previous theories predicted. This tells us that the invisible galaxies are either very faint, or the light from them gets scattered and absorbed by gas before it reaches us.

Summary

The scientists took a noisy, chaotic picture of the early universe, covered up the bright spots, cleaned out the static, and found a faint, glowing pattern in the background. They proved that we can map the "invisible" universe by listening to its collective hum, opening a new window to study how galaxies and gas evolved over billions of years.

1. Problem and Motivation

Line Intensity Mapping (LIM) is a technique designed to measure the aggregate emission of unresolved sources (galaxies and diffuse gas) to trace the large-scale structure of the universe, rather than detecting individual galaxies. While previous studies have attempted to measure the Lyman-alpha (Lyα) intensity field, a significant discrepancy exists in the literature:

The Conflict: Previous cross-correlations between Quasars (QSOs) and Lyα intensity (e.g., Croft et al. 2016, 2018; Lin et al. 2022) inferred a high cosmic Lyα luminosity density. However, these results were inconsistent with upper limits derived from the Lyα forest and were potentially biased by the intense radiation fields surrounding QSOs.
The Gap: There was a lack of high-sensitivity measurements cross-correlating Lyα intensity with star-forming galaxies (LAEs) on large scales, which would provide a cleaner probe of the intergalactic medium (IGM) and circumgalactic medium (CGM) without QSO contamination.

Objective: This paper presents the first detection of the LAE-Lyα intensity cross-power spectrum using data from the Hobby-Eberly Telescope Dark Energy Experiment (HETDEX) to constrain the mean intensity of undetected sources and test cosmological models.

2. Methodology

Data Source and Selection

Dataset: The authors used the internal HETDEX data release (HDR5), covering 90 deg² of sky (Fall, Spring, and NEP fields) with the VIRUS spectrograph.
Redshift Range: $1.9 \le z \le 3.5$ .
LAE Catalog: They selected high-confidence Lyman-alpha Emitters (LAEs) with Signal-to-Noise Ratio (SNR) $> 5.5$ . Strict quality cuts (machine learning classifiers, aperture corrections) were applied to ensure a purity of $\sim 96\%$ .
Intensity Map Construction:
- Masking: Spectra within $10''$ of detected LAEs were masked to ensure the intensity map contained only undetected sources.
- Contaminant Removal: Continuum sources (stars, foreground galaxies), bad fibers, and sky emission lines were masked.
- Systematic Removal: A Principal Component Analysis (PCA) was performed on the spectra to remove residual sky emission and detector systematics. The number of Principal Components (PCs) removed was optimized ( $N_{PC} \approx 100-170$ ) to balance signal retention with systematic suppression.
- Resolution: Data was binned into $2 h^{-1}$ Mpc voxels.

Simulation and Modeling

Mock Generation: The authors used the Simple code to generate 100 lognormal mocks per redshift bin. These mocks included:
- Nonlinear Redshift-Space Distortions (RSD).
- HETDEX sensitivity, selection functions, and masks.
- Lyα radiative transfer (RT) effects (modeled implicitly via the transfer function).
Power Spectrum Estimator: They calculated the cross-power spectrum monopole ( $P_0(k)$ ) between the LAE overdensity ( $\delta_g$ ) and the Lyα intensity fluctuation ( $\delta_I$ ) using Fast Fourier Transforms (FFT).
Covariance: The covariance matrix was derived from a combination of mocks (capturing cosmic variance) and shuffled data (capturing noise and systematics).

Fitting Strategy

The measured power spectrum was fitted to a theoretical model:
$P_{model}(k) = A \times P_{fiducial}(k)$
Where the free parameter $A$ represents the scaling factor:
$A = \frac{b_g b_I \langle I \rangle \bar{F}_{RSD}}{b_g^{fid} b_I^{fid} \langle I \rangle^{fid} \bar{F}_{RSD}^{fid}}$
Here, $b_g$ is the LAE bias, $b_I$ is the intensity bias, $\langle I \rangle$ is the mean intensity of undetected sources, and $\bar{F}_{RSD}$ accounts for redshift-space distortions.

3. Key Contributions

First Detection: This is the first statistically significant detection of the cross-power spectrum between LAEs and the Lyα intensity field from undetected sources.
Methodological Rigor: The paper establishes a robust pipeline for LIM using HETDEX data, specifically addressing the challenges of sky subtraction residuals and systematics through PCA and extensive masking.
Resolution of Discrepancy: It provides an independent measurement of Lyα intensity that avoids the QSO-dominated bias of previous studies.
Forward Modeling: The use of lognormal mocks to forward-model the data processing effects (including PCA signal loss) allows for accurate parameter inference without needing to reconstruct a "PCA-free" spectrum.

4. Key Results

The study measured the product $b_g b_I \langle I \rangle \bar{F}_{RSD} / \bar{F}_{RSD}^{fid}$ in three redshift bins:

Redshift Bin ( $\bar{z}$ )	Measured Value ( $10^{-22}$ erg s $^{-1}$ cm $^{-2}$ arcsec $^{-2}$ Å $^{-1}$ )	Significance
2.1	$6.7 \pm 3.1$	$\sim 2.1\sigma$
2.6	$11.7 \pm 1.4$	$\sim 8.4\sigma$
3.2	$8.3 \pm 1.5$	$\sim 5.5\sigma$

Comparison with QSO Studies: The measured intensity is significantly lower than values inferred from QSO-Lyα cross-correlations. This suggests that previous QSO-based measurements may have been contaminated by localized, high-intensity emission around quasars.
Comparison with Lyα Forest: The results are consistent with (or slightly higher than) the upper limits derived from Lyα forest-Lyα intensity cross-correlations.
Physical Interpretation:
- Assuming a fiducial LAE bias $b_g = 2$ , the derived intensity bias $b_I$ and mean intensity $\langle I \rangle$ are consistent with cosmological hydrodynamical simulations (e.g., Byrohl & Nelson 2023) that include Lyα radiative transfer.
- The results imply that the Lyα intensity is dominated by photons from faint, undetected galaxies and diffuse emission in the CGM/IGM, rather than just bright galaxies.
- The data is consistent with a star-formation-driven origin for Lyα emission with an escape fraction $f_{esc} < 1$ .

5. Significance and Future Outlook

Cosmological Constraints: This measurement provides a new, independent constraint on the Lyα luminosity density of the universe, crucial for understanding galaxy formation and the reionization epoch.
Validation of Simulations: The agreement with radiative transfer simulations validates current models of how Lyα photons scatter in the CGM and IGM.
HETDEX Potential: The results demonstrate the statistical power of HETDEX for intensity mapping, paving the way for future analyses with larger samples and improved data processing (e.g., better sky subtraction to reduce signal loss from masking).
Future Work: The authors note that future improvements will focus on refining the modeling of Lyα absorption features (which currently dampen the signal) and incorporating higher-fidelity hydrodynamical simulations to better constrain the intensity bias $b_I$ .

In conclusion, this paper successfully demonstrates that HETDEX can detect the Lyα intensity field from undetected sources, providing a cleaner and more consistent picture of the cosmic Lyα background than previous QSO-based studies.