🔭 astrophysics

The impact of cosmic filaments on starburst galaxies across cosmic times

This study combines cosmological simulations (SIMBA) and observational data from the COSMOS field to demonstrate that starburst galaxies are preferentially located closer to cosmic filaments at high redshifts (z>1) but shift to larger distances at low redshifts, a trend that confirms the detectable imprint of large-scale environmental effects on galaxy star formation activity across cosmic time.

Original authors: Baptiste Jego, Matthieu Béthermin, Katarina Kraljic, Clotilde Laigle, Lingyu Wang, Antonio La Marca, Olivier Ilbert, Hollis B. Akins, Caitlin M. Casey, Gavin Leroy, Ali Hadi, Jeyhan S. Kartaltepe, Ant

Published 2026-06-03

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Baptiste Jego, Matthieu Béthermin, Katarina Kraljic, Clotilde Laigle, Lingyu Wang, Antonio La Marca, Olivier Ilbert, Hollis B. Akins, Caitlin M. Casey, Gavin Leroy, Ali Hadi, Jeyhan S. Kartaltepe, Anton M. Koekemoer, Henry Joy McCracken, Louise Paquereau, Jason Rhodes, Brant E. Robertson, Marko Shuntov, Greta Toni, Can Xu

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: Galaxies and the Cosmic Web

Imagine the universe not as a random scattering of stars, but as a giant, three-dimensional spiderweb made of invisible threads. Astronomers call this the Cosmic Web. It is made of:

Filaments: Long, thin strands of dark matter and gas (the threads).
Nodes: The heavy knots where threads cross (like galaxy clusters).
Voids: The empty spaces between the threads.

For a long time, scientists have wondered: Does where a galaxy sits on this web change how it behaves? Specifically, does being close to a "thread" make a galaxy form stars faster or slower?

This paper investigates that question by looking at three types of galaxies:

Starbursts: Galaxies having a wild, short-lived party, forming stars at a furious rate.
Main-Sequence (MS): The "normal" galaxies, steadily forming stars like a steady job.
Quenched: The "retired" galaxies that have stopped making stars entirely.

The researchers wanted to see if the distance to the cosmic web threads changes for these different groups as the universe gets older (from the distant past to today).

The Two-Pronged Approach: The Simulation and The Reality Check

To answer this, the team used two methods: The Simulation and The Real Data.

1. The Simulation (The "Virtual Universe")

First, they used a supercomputer simulation called Simba. Think of this as a video game where they programmed the laws of physics to build a universe from scratch.

They tracked billions of particles to see how galaxies formed.
They measured how far different types of galaxies were from the nearest cosmic web thread at different times in the universe's history.

What the Simulation Predicted:

In the Early Universe (High Redshift): Starburst galaxies (the party animals) were found very close to the cosmic threads. It seems the threads act like highways, delivering fresh gas fuel that triggers these star-forming parties.
In the Late Universe (Low Redshift): As time went on, the Starbursts moved away from the threads. Meanwhile, the "retired" (Quenched) galaxies moved closer to the threads.
The Main-Sequence galaxies stayed in the middle, not changing their location much.

2. The Real Data (The "Cosmic Detective Work")

Simulations are great, but we need to check if they match reality. The team looked at the COSMOS field, a patch of sky observed by powerful telescopes (including the James Webb Space Telescope and Herschel).

The Challenge: We can't see the 3D web perfectly because we only see a 2D projection (like looking at a shadow). Also, starbursts are often hidden behind thick dust, making them invisible to optical telescopes.
The Solution: They used Far-Infrared (FIR) light. Dust absorbs starlight and re-emits it as heat (infrared). By looking at this heat, they could find the dusty, star-forming galaxies that other telescopes missed.
They mapped the positions of thousands of galaxies and reconstructed the 2D "shadow" of the cosmic web.

What the Real Data Showed:
The real universe agreed with the simulation!

At high redshift (young universe): Starburst galaxies were indeed closer to the cosmic threads than normal galaxies.
At low redshift (older universe): The trend flipped. Starbursts were found further away, while "retired" galaxies huddled closer to the threads.
The "Crossing Point": Around a specific time in the universe's history (when the universe was about half its current age), the average distance of Starbursts and Normal galaxies swapped.

The statistical certainty of this finding in the new JWST data (COSMOS-Web) was extremely high (over 5 sigma), meaning it is almost certainly not a fluke.

Why Does This Happen? (The "Toy Model" Explanation)

The authors built a simple "toy model" to explain why this happens. They didn't need to invent complex new physics; they just used a simple rule found in their simulation: The environment changes how efficiently a galaxy forms stars.

The Analogy: Imagine the cosmic threads are like water pipes.
- Early Universe: The pipes are full of fresh water. If you are a "Starburst" (a thirsty plant), you want to be right next to the pipe to get a drink. Being close to the thread means you get the fuel you need to throw a party.
- Late Universe: The pipes are still there, but the environment has changed. Maybe the "water" is too hot, or the "soil" is toxic. Now, being right next to the pipe is bad for a Starburst. They move away to find a better spot. Meanwhile, the "retired" galaxies (which don't need water) are fine being close to the pipes, perhaps because the environment there helps them stay quiet.

The study suggests that the link between a galaxy's location and its star formation isn't a direct "push" or "pull" on individual galaxies. Instead, it's a statistical trend: the probability of a galaxy having a starburst episode changes depending on how far it is from the cosmic web, and this probability shifts as the universe ages.

The "Shadow" Problem (Projection Effects)

The paper also notes a technical detail: because we are looking at a 2D slice of a 3D universe, our view is a bit distorted (like looking at a 3D object through a flat window).

In the 3D simulation, the trends are even sharper.
In the 2D observation, the signal is slightly "smeared out," but the main pattern (the flip-flop of where Starbursts live) remains clear.

Conclusion

This paper provides the first solid observational evidence that where a galaxy lives on the cosmic web matters.

Young Universe: Star-forming galaxies hug the cosmic threads.
Old Universe: They drift away, while dead galaxies move closer.

It confirms that the large-scale structure of the universe acts like a stage director, subtly influencing the script of how galaxies grow and change over billions of years. The "Cosmic Web" isn't just a background; it's an active participant in the life story of every galaxy.

Technical Summary: The Impact of Cosmic Filaments on Starburst Galaxies Across Cosmic Times

Problem Statement
Cosmological simulations suggest that galaxy properties, including star formation activity, depend on their location within the cosmic web (voids, walls, filaments, and nodes). While observational evidence links galaxy quenching and mass to proximity to filaments, direct evidence regarding the dependence of star formation activity (specifically distinguishing between starburst, main-sequence, and quenched populations) on filament distance remains scarce, particularly at redshifts $z \gtrsim 1$ . A key challenge is disentangling environmental effects from stellar mass dependencies and addressing the limitations of current datasets in reconstructing the 3D cosmic web and identifying heavily dust-obscured starbursts.

Methodology
The study employs a dual approach, combining cosmological hydrodynamical simulations with deep multi-wavelength observational data:

Simulations (Simba):
- The authors utilize the Simba simulation suite (Davé et al. 2019) to predict the redshift evolution of the mean distance to the nearest cosmic filament ( $\langle d_{\text{fil}} \rangle$ ) for starburst, main-sequence (MS), and quenched galaxies from $z=3$ to $z=0$ .
- To isolate environmental effects, they construct mass-matched samples, comparing starbursts to MS galaxies and MS galaxies to quenched galaxies, ensuring differences in $\langle d_{\text{fil}} \rangle$ are not driven by stellar mass.
- Filaments are reconstructed using the DisPerSE algorithm on projected 2D galaxy positions (to mimic observational conditions) and full 3D positions for interpretation.
- Starbursts are defined by specific thresholds in Star Formation Rate (SFR) relative to the MS and gas depletion time ( $\tau$ ).
Observations (COSMOS Field):
- Data is drawn from the COSMOS2020 and COSMOS-Web catalogs, utilizing accurate photometric redshifts to reconstruct the projected 2D cosmic web from $z=2$ to $z=0.5$ .
- Starburst Identification: Unlike optical/UV studies, this work identifies starbursts via Far-Infrared (FIR) Spectral Energy Distribution (SED) fitting using the COSMOS-XID+ catalog (Herschel data). This allows for the robust detection of dust-obscured star formation. Starbursts are selected based on $SFR > 4 \times SFR_{\text{MS}}$ .
- Comparison Samples: MS and quenched galaxies are selected using rest-frame color criteria and mass-matched to the starburst sample.
- Cosmic Web Reconstruction: The cosmic web is reconstructed in 2D slices using DisPerSE. To handle photometric redshift uncertainties and survey masks, the authors generate 100 realizations of the web per slice and apply random galaxy filling within masked regions to prevent artificial filament generation.
- Statistical Analysis: The study focuses on the ratio of mean filament distances between populations (e.g., $\langle d_{\text{fil}} \rangle_{\text{SB}} / \langle d_{\text{fil}} \rangle_{\text{MS}}$ ) to minimize systematic errors related to absolute distance scales and redshift uncertainties.

Key Contributions and Results

Simulation Predictions: In Simba, starburst galaxies are found closer to filaments at high redshift ( $z > 1$ ) but at larger distances than MS galaxies at low redshift ( $z < 1$ ). Quenched galaxies show progressively shorter distances to filaments toward low redshift. The crossing point between starburst and MS populations occurs around $z \sim 1$ . These trends are robust to variations in starburst selection criteria.
Observational Detection: In the COSMOS field, the observational data confirms the simulation predictions.
- At $z > 1$ , starbursts are, on average, closer to filaments than mass-matched MS galaxies.
- At $z < 1$ , this trend reverses, with starbursts located further away.
- The relative evolution between starburst and MS populations is detected with a significance of at least $5\sigma$ (specifically $6\sigma$ for linear fits and $6.6\sigma$ for power-law fits in COSMOS-Web).
- Quenched galaxies exhibit a trend of moving closer to filaments as redshift decreases, consistent with simulations.
Projection Effects: The authors demonstrate that reconstructing the cosmic web in 2D (projected) versus 3D significantly affects the measured amplitudes and redshift dependencies. While the relative ordering of populations is preserved, 2D projections artificially flatten the evolution of MS galaxies and reduce the absolute distance values compared to 3D reconstructions.
Toy Model Validation: A minimal toy model, which applies the sSFR-filament distance modulation measured in Simba to a synthetic population, successfully reproduces the observed differential evolution between starburst and MS galaxies. This suggests that the observed trends can be explained by the environment-dependent modulation of star formation activity without requiring direct one-to-one causal links (e.g., mergers) for every starburst event.

Significance and Claims
The paper claims to provide the first direct observational evidence, across cosmic time ( $z=2$ to $z=0.5$ ), linking the large-scale environment (specifically cosmic filaments) to the star formation activity of galaxies, distinguishing between starburst, main-sequence, and quenched populations.

The authors emphasize that their results support the simulation-based prediction that the large-scale environment modulates star formation. They argue that the observed trends reflect statistical correlations between the likelihood of experiencing a starburst phase and the long-term environmental history of galaxies, rather than instantaneous triggering by local filamentary structures. The study highlights the necessity of FIR data for identifying dust-obscured starbursts and the importance of 3D cosmic web reconstructions (future goal) to fully characterize these environmental dependencies. The work establishes that current deep multi-wavelength surveys are capable of detecting these subtle environmental imprints predicted by cosmological simulations.