🔭 astrophysics

Simba Simulation: The Effect of Feedback Physics on Matter Distribution in the Cosmic Web

Using the Simba hydrodynamical simulation suite and the T-web classification method, this study demonstrates that while feedback models only slightly alter the overall baryon fractions across cosmic web structures, jet feedback significantly redistributes gas to the outskirts of filaments and knots, offering critical insights for refining Fast Radio Burst foreground models.

Original authors: Chenze Dong, Florian Dedieu, Daniela Galárraga-Espinosa, Khee-Gan Lee, Daniele Sorini, Romeel Davé

Published 2026-04-30

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Chenze Dong, Florian Dedieu, Daniela Galárraga-Espinosa, Khee-Gan Lee, Daniele Sorini, Romeel Davé

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 Mystery: Where Did the Missing Baryons Go?

Imagine the universe as a giant house. We know exactly how much "stuff" (matter) should be in the house based on how the house was built billions of years ago (the Big Bang). However, when we look around the house today, we can only find about 10% of that stuff in the rooms we can see (galaxies, stars, and gas clouds).

Where is the other 90%? This is the "Missing Baryon Problem."

Scientists suspect the missing stuff is floating around in the empty spaces between the galaxies, like invisible mist. This mist is called the Intergalactic Medium (IGM). To find it, astronomers use "Fast Radio Bursts" (FRBs)—brief, bright flashes of radio waves from deep space. As these flashes travel to us, they get slightly delayed by the amount of "stuff" they pass through. By measuring this delay, scientists can count the missing particles.

The Problem: The "Feedback" Mess

The problem is that the universe isn't a quiet, empty house. It's a chaotic construction site. Galaxies are constantly blowing things around. Stars explode, and supermassive black holes shoot out powerful jets of energy. This is called feedback.

Think of feedback like a powerful leaf blower in a garden. If you have a pile of leaves (gas) in a corner, a leaf blower can blow them out into the rest of the yard.

The Question: Does this "leaf blower" just move the leaves around locally, or does it scatter them so widely that it changes the entire layout of the garden?
The Goal: This paper wanted to see if different types of "leaf blowers" (feedback models) change where the missing gas ends up in the cosmic "garden."

The Tool: Mapping the Cosmic Web

To study this, the researchers used a supercomputer simulation called Simba. They didn't just look at the gas; they looked at the shape of the universe itself.

Imagine the universe as a giant spiderweb made of:

Knots: The busy intersections where galaxies cluster (like a city center).
Filaments: The long strands connecting the knots (like highways).
Sheets: The flat walls between the strands.
Voids: The huge empty spaces in the middle (like the empty fields between towns).

The researchers used a method called T-web to sort every bit of gas in their simulation into one of these four categories.

What They Found

They ran the simulation five times, turning different "feedback" features on and off (like turning off the black hole jets or the stellar winds). Here is what they discovered:

1. The Big Picture Doesn't Change Much
When they looked at the total amount of gas in the Knots, Filaments, Sheets, and Voids, the different feedback models barely changed the numbers.

Analogy: Imagine you have a bucket of water divided into four cups. Whether you stir the water gently or violently, the total amount of water in each cup remains almost exactly the same. The "leaf blower" didn't move the water from one cup to another in a way that changed the overall balance.

2. But the "Local" Details Do Change
While the total amounts stayed the same, the location of the gas inside those cups changed significantly, especially when Jet Feedback (the powerful black hole jets) was active.

The Effect: The jets acted like a high-pressure hose. They blew gas out of the dense "Knots" (galaxy clusters) and pushed it into the surrounding "Filaments" and "Sheets."
The Result: The gas didn't just disappear; it got pushed to the edges of the structures. It created a "halo" of extra gas around the dense areas.

3. The Impact on "Measuring" the Universe
This is the most important part for the FRB astronomers. When scientists use FRBs to measure the universe, they usually assume the gas is spread out evenly, like sugar dissolved perfectly in a cup of tea.

The Reality: Because of the jets, the gas is actually clumped up in specific ways (like sugar that has settled at the bottom or floated to the top).
The Consequence: If you assume the gas is evenly spread out, your measurement will be wrong. Specifically, the paper found that if you ignore these jets, you might underestimate how much gas is actually there. You might think there is less "stuff" in the universe than there really is because the gas is hiding in the "edges" of the structures where your simple math doesn't expect it.

The Conclusion

The paper concludes that while the "big map" of the universe (the cosmic web) looks the same regardless of how galaxies blow their "leaves" around, the local details are very different.

The powerful jets from black holes act like a redistribution crew, moving gas from the crowded city centers (halos) out to the suburbs (the diffuse IGM). If we want to accurately count the missing baryons using Fast Radio Bursts, we can't just use a simple, average map. We need to account for these "leaf blowers" to avoid miscounting the invisible stuff floating in the cosmic web.

1. Problem Statement

The paper addresses the "missing baryon problem," the discrepancy between the cosmic baryon density ( $\Omega_b$ ) measured in the early universe (via CMB and Big Bang Nucleosynthesis) and the lower density observed in the low-redshift universe. While Fast Radio Bursts (FRBs) have recently provided evidence that these baryons exist in the diffuse Intergalactic Medium (IGM), the precise spatial distribution of these baryons remains uncertain.

A critical gap in current understanding is how galaxy feedback physics (specifically from supernovae and Active Galactic Nuclei/AGN) redistributes baryons between galactic haloes, the circumgalactic medium (CGM), and the IGM. This redistribution affects the Dispersion Measure (DM) of FRBs, which is used to probe cosmic baryons. Current FRB foreground mapping models (e.g., the FLIMFLAM survey) often assume a simple linear relationship between matter density and electron density. This study investigates whether different feedback models significantly alter the baryon partition within cosmic web structures (voids, sheets, filaments, knots) and if this necessitates a refined data model for FRB analysis.

2. Methodology

Simulation Suite:
The authors utilized the Simba cosmological hydrodynamical simulation suite, which employs the GIZMO code (Meshless Finite Mass). They analyzed five specific feedback variants, all starting from identical initial conditions in a $(50 \text{ cMpc}/h)^3$ box at redshift $z=0$ :

Simba-50: Full physics (Stellar feedback, AGN radiative winds, AGN jets, X-ray heating).
Simba-nox: No X-ray heating.
Simba-nojet: No AGN jets (only radiative winds).
Simba-noagn: No AGN feedback (only stellar feedback).
Simba-nofb: No feedback at all.

Cosmic Web Classification (T-web):
To classify the large-scale structure, the authors applied the T-web method.

They computed the deformation tensor ( $\psi_{ij}$ ) from the Hessian of the rescaled gravitational potential.
The matter density field was smoothed with a Gaussian kernel of $R_{sm} = 2 \text{ cMpc}/h$ (motivated by the resolution limits of current galaxy surveys used in FRB foreground mapping).
Structures were classified based on the eigenvalues ( $\lambda_1 \ge \lambda_2 \ge \lambda_3$ $λ_{1} \geq λ_{2} \geq λ_{3}$ ) of the deformation tensor relative to a threshold $\lambda_{th} = 0.16$ $λ_{t h} = 0.16$ :
- Knots (Nodes): 3 eigenvalues $> \lambda_{th}$ .
- Filaments: 2 eigenvalues $> \lambda_{th}$ .
- Sheets: 1 eigenvalue $> \lambda_{th}$ .
- Voids: 0 eigenvalues $> \lambda_{th}$ .

Analysis Metrics:

Global Partition: Calculated the mass fractions of total matter, gas, and free electrons within each cosmic web structure for all feedback variants.
Local Gas Fraction ( $f_{gas}$ ): Defined as $f_{gas} = \rho_{gas} / (f_b \rho_{tot})$ , where $f_b$ is the cosmic baryon fraction. This measures local gas excess ( $f_{gas} > 1$ ) or deficit ( $f_{gas} < 1$ ) relative to the cosmic mean.
FRB DM Modeling: Simulated FRB sightlines to compare the predicted Dispersion Measure (based on a global $f_{IGM}$ assumption) against the measured DM from the simulation, testing for biases introduced by feedback.

3. Key Contributions

Quantification of Feedback Impact on Cosmic Web Partition: The study provides a rigorous comparison of how different feedback mechanisms redistribute baryons across the four cosmic web environments (voids, sheets, filaments, knots) at $z=0$ .
Decoupling of Scales: It distinguishes between the global distribution of baryons (which is robust against feedback changes) and the local distribution within structures (which is highly sensitive to feedback, particularly AGN jets).
Refinement of FRB Foreground Models: The paper challenges the standard assumption in FRB foreground mapping that the IGM gas density strictly traces the total matter density ( $\rho_{gas} \propto \rho_{tot}$ ). It demonstrates that feedback creates a non-linear, environment-dependent relationship.
Identification of Systematic Bias: It identifies a specific systematic bias in FRB DM estimation: the underestimation of DM in lines of sight that do not intersect massive haloes when strong jet feedback is present but unmodeled.

4. Key Results

A. Stability of Large-Scale Structure Classification

The T-web classification of the cosmic web (based on total matter density) is highly insensitive to feedback models. The volume and mass fractions of voids, sheets, filaments, and knots vary by less than 0.1% to a few percent across all feedback variants.
This is because large-scale structure is dominated by dark matter, which is unaffected by baryonic feedback.

B. Global vs. Local Baryon Partition

Global Partition: When looking at the total fraction of IGM gas residing in each structure, feedback models cause only minor variations (typically $< 2\%$ ). Jet feedback slightly increases gas in sheets and decreases it in knots, but the overall partition remains stable.
Local Distribution ( $f_{gas}$ ): In contrast, the local gas fraction shows dramatic differences:
- Jet Feedback: Creates a significant gas excess ( $f_{gas} > 1$ , up to $f_{gas} \sim 3$ ) in the diffuse IGM surrounding filaments and knots (specifically in the range $0.1 < 1+\delta_m < 10$ ). This is caused by AGN jets ejecting gas from haloes into the surrounding IGM.
- Halo Depletion: Conversely, within haloes (high overdensity), jet feedback causes a severe gas deficit ( $f_{gas} < 1$ ), removing up to 50% of baryons from haloes compared to weaker feedback models.
- No-Jet Models: Without jets, the gas fraction follows a monotonic decline with overdensity, driven primarily by star formation consumption.

C. Impact on FRB Dispersion Measure (DM)

Bias Detection: When simulating FRB sightlines, models that assume a global, constant $f_{IGM}$ (linear tracing) underestimate the true DM for sightlines passing through low-to-intermediate density regions ( $\int (1+\delta_m) ds < 100$ ) in simulations with strong jet feedback.
Magnitude: The deviation can reach $\sim 5 \text{ pc cm}^{-3}$ , which is significant for precision cosmology.
Cause: The underestimation occurs because jet feedback pumps extra gas into the diffuse IGM, increasing the electron density beyond what is predicted by the total matter density alone.

5. Significance and Implications

For FRB Cosmology: The study argues that while incorporating cosmic web classification into FRB foreground mapping (e.g., separating DM contributions by structure type) offers limited improvement at current resolution scales ( $\sim 2 \text{ cMpc}/h$ ), modeling the local gas fraction ( $f_{gas}$ ) as a function of overdensity is crucial.
Refined Data Models: Future FRB analyses (like FLIMFLAM) must move beyond the assumption that $\rho_{gas} = f_b \rho_{tot}$ . Instead, they should adopt a density-dependent $f_{gas}$ correction, particularly to account for the "baryon excess" in the outskirts of filaments caused by AGN jets.
Feedback Constraints: The results suggest that the "baryon excess" in the diffuse IGM is a potential observational signature of AGN jet feedback. If future FRB surveys can detect this specific DM excess in low-density sightlines, it could provide a new, independent constraint on AGN feedback physics.
Future Directions: The authors propose using machine learning emulators (trained on simulations like CAMELS) to model the complex $f_{gas}$ vs. overdensity relationship, which could be integrated into FRB data models to reduce systematic errors in measuring the cosmic baryon budget.

In summary, the paper demonstrates that while the skeleton of the cosmic web is robust to feedback, the flesh (baryon distribution) is highly sensitive. Ignoring this sensitivity in FRB modeling leads to systematic biases in measuring the cosmic baryon density, necessitating a shift toward environment-dependent gas fraction models.