🔭 astrophysics

Retrieving the hot circumgalactic medium physics from the X-ray radial profile from eROSITA with an IlustrisTNG-based forward model

The authors develop a forward modeling approach using IllustrisTNG simulations to interpret eROSITA X-ray radial profiles, demonstrating that the observed X-ray emission from the hot circumgalactic medium in Milky Way-mass galaxies is primarily driven by the underlying halo mass distribution.

Original authors: Soumya Shreeram, Johan Comparat, Andrea Merloni, Gabriele Ponti, Paola Popesso, Yi Zhang, Kirpal Nandra, Mara Salvato, Ilaria Marini, Johannes Buchner, Nicola Locatelli, Zsofi Igo

Published 2026-02-11

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Soumya Shreeram, Johan Comparat, Andrea Merloni, Gabriele Ponti, Paola Popesso, Yi Zhang, Kirpal Nandra, Mara Salvato, Ilaria Marini, Johannes Buchner, Nicola Locatelli, Zsofi Igo

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 Cosmic "Fog" Around Galaxies: A Detective Story

Imagine you are looking at a distant city at night from an airplane. You see bright, sharp points of light—those are the streetlamps and building windows. But between those bright lights, there is a faint, hazy glow. That glow isn't just empty space; it’s the light reflecting off the mist, smog, and humidity in the air.

In astronomy, galaxies are like those cities. The stars and black holes are the bright "streetlamps." But surrounding these galaxies is a massive, invisible, and incredibly hot "fog" of gas called the Circumgalactic Medium (CGM).

This paper is about a team of scientists trying to figure out exactly how much of that "glow" comes from the actual cosmic fog versus the bright "streetlamps" (like black holes and star clusters) that might be tricking our eyes.

The Problem: The "Flashlight in the Fog" Dilemma

The scientists used data from a powerful X-ray telescope called eROSITA. This telescope sees the universe in X-rays, which is how we detect super-hot gas.

However, they ran into a classic detective problem: The Signal vs. The Noise.

Imagine you are trying to measure how thick the fog is in a park by looking at how much light is scattered. But there’s a problem: there are also bright flashlights being waved around in the park. If you see a bright spot, is it because the fog is really thick there, or is it just someone pointing a powerful flashlight directly at you?

In space, those "flashlights" are:

AGNs (Active Galactic Nuclei): Hungry supermassive black holes that shine like cosmic searchlights.
XRBs (X-ray Binaries): Pairs of stars that act like smaller, flickering lamps.
Satellite Galaxies: "Small towns" orbiting a "big city" (the main galaxy), which add their own extra glow to the mix.

If the scientists don't account for these "flashlights," they will overestimate how much hot gas (the fog) is actually there.

The Solution: The "Digital Twin" Method (Forward Modeling)

To solve this, the researchers didn't just look at the real sky; they built a "Digital Universe" using a massive supercomputer simulation called IllustrisTNG.

Think of this like a flight simulator. Instead of just guessing how a pilot reacts to wind, they build a perfect digital model of the wind, the plane, and the pilot. They then "fly" their digital galaxies through their digital simulation to see what the X-ray glow should look like.

They created three different "Digital Universes" (Models 1, 2, and 3) where they changed the "neighborhood" of the galaxies. In some models, the galaxies lived in massive, crowded clusters; in others, they lived in more lonely, quiet suburbs.

The Discovery: Finding the Right Neighborhood

By comparing their "Digital Universe" to the real data from eROSITA, they found the winner: Model 3.

Model 3 revealed that the galaxies they were looking at lived in a specific kind of "suburban" environment—not too crowded, but not totally empty. This allowed them to finally separate the components:

The Inner Circle (The "Lamp" Zone): Close to the center of the galaxy (within 40 kiloparsecs), the light is a messy mix. About half the glow comes from the hot gas (the fog), and the other half comes from the bright "flashlights" (black holes and stars).
The Outer Reach (The "Fog" Zone): Once you move further out, the "flashlights" fade away, and the glow is dominated by the massive halos of surrounding galaxies.

Why does this matter?

Understanding this "cosmic fog" is like understanding the atmosphere of Earth. The gas in the CGM acts as a reservoir—it breathes gas into galaxies to help them make stars, and it breathes gas out when black holes explode with energy.

By learning how to accurately "see through the flashlights," these scientists have given us a new pair of glasses. We can now use these techniques to study how galaxies grow, breathe, and eventually die, helping us understand the history of the entire universe.

Technical Summary: Retrieving the Hot Circumgalactic Medium Physics from eROSITA X-ray Radial Profiles

Paper Reference: Shreeram et al. (2026), Astronomy & Astrophysics (Manuscript no. aa54508-25).

1. The Problem: Disentangling the CGM Signal

The Circumgalactic Medium (CGM) is a critical component in galaxy evolution, acting as a reservoir for gas inflows and outflows driven by stellar and Active Galactic Nucleus (AGN) feedback. In Milky Way (MW)-mass galaxies, the hot phase of the CGM ( $T \gtrsim 10^6$ K) emits soft X-rays. However, extracting the intrinsic CGM signal from X-ray observations is extremely challenging due to two primary sources of "contamination":

Point Sources: Unresolved X-ray emission from X-ray Binaries (XRBs) and AGNs.
Satellite Contamination: In large-scale photometric galaxy surveys (like the DESI Legacy Survey), galaxies cannot be easily classified as "centrals" or "satellites." When stacking X-ray data, the signal from satellite galaxies is dominated by the much more massive host halos they reside in, which artificially flattens and boosts the observed radial surface brightness profile.

Current models struggle to reconcile hydrodynamical simulations (like IllustrisTNG) with observations because the predicted X-ray signal is highly sensitive to the underlying halo mass distribution at a fixed stellar mass.

2. Methodology: The TNG-Based Forward Model

The authors developed a novel forward-modeling framework to disentangle these components. Instead of trying to "clean" the data, they created mock observations from the IllustrisTNG300 simulation to match the observational properties of the eROSITA/DESI Legacy Survey (Fullphot) sample.

Key components of the model:

Hot Gas Component: Modeled using TNG300 lightcones. The model accounts for both the intrinsic emission from central galaxies ( $S_{X, \text{cen}}$ ) and the emission from the massive host halos of satellites ( $S_{X, \text{sat}}$ ).
Point Source Component: Since simulations often struggle to predict instantaneous AGN/XRB activity accurately, the authors did not use TNG for this. Instead, they used a normalization parameter ( $N_{ps}$ ) and fixed the radial shape to the eROSITA Point Spread Function (PSF).
Halo Mass Distribution Testing: They generated three distinct mock catalogs (Models 1, 2, and 3) by varying the underlying halo mass distribution (e.g., by discarding the most massive halos) to see which best matched the observed X-ray profile.
Statistical Fitting: They employed a Markov chain Monte Carlo (MCMC) approach to fit the model to the eROSITA surface brightness profiles, using two free parameters: the satellite normalization ( $N_{sat}$ ) and the point source normalization ( $N_{ps}$ ).

3. Key Results

Selection of Model 3: By comparing the resulting point source luminosities against independent empirical models for XRBs and AGNs, the authors found that Model 3 (which assumes a lower mean halo mass, $M_{200m} \approx 3.5 \times 10^{12} M_\odot$ ) was the only model consistent with known AGN/XRB populations. This implies that the "true" halo mass distribution for this stellar-mass-selected sample is lower than what the standard TNG300 distribution suggests.
Emission Budget at Small Radii ( $\lesssim 40$ kpc): The total X-ray emission is a nearly even split: $\sim 40\text{--}50\%$ comes from the hot CGM of central galaxies, and $\sim 40\text{--}50\%$ comes from point sources (AGNs and XRBs).
Emission Budget at Large Radii ( $> 40$ kpc): The signal is dominated by the hot gas in the massive host halos of satellite galaxies. This explains the "flattening" of the radial profile observed in eROSITA data.
CGM Luminosity: The integrated X-ray luminosity of the hot CGM for central galaxies was found to be $L_{X, \text{CGM}} = 1.69^{+0.28}_{-0.91} \times 10^{40}$ erg/s.
Validation: The TNG-based Model 3 shows excellent agreement with previous spectroscopic studies (the SDSS-based "CEN" sample) at radii $\gtrsim 60$ kpc.

4. Significance and Impact

New Diagnostic Tool: This work introduces a technique to jointly constrain AGN X-ray luminosity and the radial distribution of hot CGM gas using stacking experiments.
Addressing Simulation Biases: The study highlights that comparing simulations to observations requires a rigorous treatment of the Stellar-to-Halo Mass Relation (SHMR). A small shift in the assumed halo mass distribution can lead to a factor of $\sim 4$ difference in predicted X-ray luminosity.
Future Calibration: The framework provides a roadmap for testing other state-of-the-art simulations (e.g., EAGLE, SIMBA, FLAMINGO). By comparing the predicted AGN luminosity from a simulation with the retrieved luminosity from this method, researchers can better calibrate sub-grid feedback models in cosmological simulations.