🔭 astrophysics

Tiny galaxies and dark substructures: exploring the "dark" subhaloes in TNG50

Using the TNG50-1 simulation, this study reveals that while no haloes above $10^{9.7} M_{\odot}$ are entirely dark under standard formation models, realistic surface brightness limits allow for massive dark subhaloes that are ubiquitous in galaxy groups and clusters, with those in the mass range of $4.5 \times 10^{7}$ to $2.1 \times 10^{8} M_{\odot}$ being the most likely candidates for detection via gravitational lensing.

Original authors: Jessica E. Doppel, Mathilde Jauzac, David J. Lagattuta, Azadeh Fattahi, Guillaume Mahler

Published 2026-03-25

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jessica E. Doppel, Mathilde Jauzac, David J. Lagattuta, Azadeh Fattahi, Guillaume Mahler

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 Question: Are There "Ghost" Galaxies?

Imagine the universe as a giant, invisible scaffolding made of Dark Matter. This scaffolding is so heavy and strong that it holds everything together. According to our best theories (called the Lambda-CDM model), this scaffolding should be covered in tiny "nests" or clumps.

Usually, these nests are big enough to catch gas and dust, which then collapse to form stars and galaxies. But the theory also predicts that there should be some tiny, weak nests that are too small to catch any gas. These would be "ghost" nests: they have the mass of a galaxy, but they are completely empty of stars. They are invisible to our telescopes.

The big question this paper asks is: Do these "ghost" galaxies actually exist, or is our theory wrong?

The Tool: A Cosmic Video Game

To find the answer, the scientists didn't look through a telescope first. Instead, they used a super-powerful computer simulation called TNG50.

Think of TNG50 as a hyper-realistic video game of the universe.

It has a "physics engine" that knows exactly how gravity works.
It has a "chemistry engine" that knows how stars are born and die.
It has a resolution so high that it can see tiny clumps of dark matter that other simulations miss.

The researchers ran this simulation and looked at three massive "cities" of galaxies (galaxy clusters) and their smaller neighbors (groups) to see what kind of "ghosts" were hiding there.

The Findings: The "Invisible" vs. The "Too Faint"

The researchers looked for these ghost galaxies in two ways, using two different "magnifying glasses":

1. The "Perfect Vision" Check (The Simulation's Eye)
First, they asked the computer: "Is there a clump of dark matter with absolutely zero stars inside it?"

The Result: They found almost none.
The Analogy: Imagine looking for a house with no lights on. In the simulation, almost every house, no matter how small, had at least one tiny light bulb (a star) inside it.
The Limit: The simulation suggests that if a dark matter clump is heavier than a certain amount (about 10 million times the mass of our Sun), it always manages to form at least a few stars. So, truly "empty" ghosts are incredibly rare.

2. The "Human Eye" Check (The Real World)
Next, they asked: "What if we look at these clumps with a real telescope? Would we see the stars?"

The Result: Suddenly, the "ghosts" appeared!
The Analogy: Imagine a house with a single, tiny candle in a huge, dark room. From far away, or if the candle is behind a curtain, the house looks pitch black.
The Discovery: The simulation showed that many of these tiny galaxies are so dim and spread out that even our best telescopes would miss them. They aren't "empty" ghosts; they are just ultra-faint dwarfs that are too dim to see.

Where Do They Hide? (The Neighborhood Effect)

The paper also looked at where these objects live within a galaxy cluster.

The Heavyweights: Big, bright galaxies tend to live near the center of the cluster, like the rich people living in the city center.
The Lightweights: The tiny, dim galaxies (the ones we can't see) are scattered everywhere, but they are surprisingly common in the outer edges of the cluster.
The "Strong Lensing" Zone: This is a specific area around a galaxy cluster where gravity acts like a magnifying glass, bending light from objects behind it. The researchers found that the tiny, dim galaxies are most likely to be found right in this "magnifying glass" zone.

Why Does This Matter?

This is crucial for astronomers who are trying to find these dark clumps using gravitational lensing (looking for distortions in light).

The Problem: If astronomers find a "ghost" clump that seems to have a mass of 10 billion suns, they might think, "Aha! A dark matter ghost!"
The Paper's Warning: The simulation says, "Wait a minute. If it's that heavy, it probably has stars in it. You just can't see them yet."
The Conclusion: If we find a massive "dark" object, it's likely not a ghost at all. It's probably a tiny, dim dwarf galaxy that our current telescopes aren't powerful enough to see.

The Takeaway

The universe is likely full of these tiny, dim galaxies that we can't see yet. They aren't "dark" because they are empty; they are "dark" because they are too faint.

The paper suggests that future, super-powerful telescopes (like the upcoming Habitable Worlds Observatory) will be able to turn on the lights in these "ghost houses." When we do, we might find that the "ghosts" were just shy neighbors all along.

In short: The "dark" subhaloes predicted by theory probably aren't empty. They are just tiny, dim galaxies that are hiding in plain sight, waiting for a better telescope to spot them.

1. Problem Statement

Within the $\Lambda$ CDM cosmological paradigm, hierarchical structure formation predicts the existence of dark matter haloes and subhaloes that do not host luminous galaxies ("dark subhaloes"). While gravitational lensing has provided potential detections of such substructures in galaxy groups and clusters, a critical ambiguity remains: Are these objects truly dark (containing no stars), or do they host faint, low-surface-brightness dwarf galaxies that are currently undetectable?

Current observational limits make it difficult to distinguish between truly dark subhaloes and those containing ultra-faint satellites. Furthermore, cosmological simulations often lack the resolution to resolve low-mass dwarfs in massive cluster environments, or conversely, lack the volume to contain massive lensing clusters. This paper aims to resolve this ambiguity by utilizing high-resolution simulations to determine the mass limits and spatial distribution of truly dark versus observationally dark subhaloes.

2. Methodology

The authors utilized the IllustrisTNG50-1 simulation, chosen for its ability to simultaneously resolve low-mass objects and massive galaxy cluster environments.

Simulation Parameters:
- Volume: 51.7 Mpc box size.
- Resolution: Dark matter particle mass $m_{DM} = 4.5 \times 10^5 M_\odot$ ; baryonic (star/gas) particle mass $m_{baryon} = 8.4 \times 10^4 M_\odot$ .
- Physics: Includes hydrodynamics (AREPO), star formation, feedback (AGN and supernovae), and magnetic fields.
Sample Selection:
- Focused on the three most massive galaxy groups/clusters at $z=0$ (masses ranging from $\sim 6 \times 10^{13}$ to $1.8 \times 10^{14} M_\odot$ ) and their progenitors from $z=0$ to $z=2$ .
- Analyzed subhaloes with total dark matter masses ( $M_{DM}$ ) between $4.54 \times 10^7 M_\odot$ and $2 \times 10^{10} M_\odot$ (ensuring resolution by at least 100 particles).
Classification Criteria:
1. Particle Count Limit: A subhalo is "resolved" if it contains $\ge 100$ star particles ( $M_* \gtrsim 8 \times 10^6 M_\odot$ ).
2. Observational Limit: A subhalo is "observationally resolved" if its average surface brightness within the stellar half-mass radius is brighter than 29 mag arcsec $^{-2}$ (mimicking deep Hubble imaging).
3. Truly Dark: Subhaloes with $M_* = 0$ and $M_{gas} = 0$ .
Analysis:
- Constructed mass functions and occupation fractions ( $f_{occ}$ ) for different redshifts.
- Analyzed radial distributions (number density profiles) of subhaloes relative to the cluster center ( $r/r_{200,c}$ ).
- Performed projected analysis by simulating 36 lines of sight through the simulation volume to account for foreground/background subhaloes affecting strong lensing regions.

3. Key Contributions

Quantification of "Darkness": The study provides rigorous mass limits for when subhaloes transition from being "observationally dark" to "truly dark" within the TNG50 galaxy formation model.
Radial Distribution in Lensing Regions: It maps the specific radial locations where dark subhaloes are most likely to be found relative to strong lensing regions ( $R \sim 10-250$ kpc).
Resolution vs. Volume Trade-off: The paper explicitly addresses the limitations of current simulations, noting that TNG50's high resolution allows for the study of low-mass objects in clusters, whereas lower-resolution simulations (like TNG100) fail to resolve these populations.

4. Key Results

Mass Limits for Dark Subhaloes:
- Truly Dark: Under the TNG50 model, there are no subhaloes with $M_{DM} > 10^{9.7} M_\odot$ that are entirely devoid of stars and gas. The model predicts that haloes above this mass threshold cool gas and form stars.
- Observationally Dark: When applying realistic surface brightness cuts ( $\le 29$ mag arcsec $^{-2}$ ), the most massive "dark" subhaloes have masses $M_{DM} \gtrsim 2 \times 10^{10} M_\odot$ . These are predominantly found at higher redshifts ( $z \gtrsim 1$ ) where low-surface-brightness galaxies are harder to detect.
Occupation Fraction:
- The fraction of subhaloes hosting a galaxy ( $f_{occ}$ ) drops significantly for low-mass objects when applying surface brightness cuts compared to simple particle counts.
- At $z=0$ , the difference between particle-count resolution and surface-brightness resolution is small, but it widens at higher redshifts due to observational challenges.
Spatial Distribution:
- Mass Segregation: Low-mass subhaloes ( $4.5 \times 10^7 \le M_{DM}/M_\odot \le 2.1 \times 10^8$ ) are the most numerous in the strong lensing regions of galaxy groups and clusters.
- Radial Profiles: Truly dark subhaloes tend to reside further from the cluster centers ( $0.8-0.9 r_{200,c}$ ) compared to luminous subhaloes ( $0.2-0.5 r_{200,c}$ ), likely due to dynamical friction and tidal stripping affecting massive, luminous satellites more severely.
- Projected Distribution: When accounting for line-of-sight effects, the distribution of dark subhaloes becomes flatter, but massive dark subhaloes ( $M_{DM} \gtrsim 10^9 M_\odot$ ) remain exceptionally rare in strong lensing regions.

5. Significance and Implications

Interpretation of Lensing Detections: Many subhaloes detected via gravitational lensing with masses $> 10^9 M_\odot$ are likely not truly dark. Instead, they probably host faint dwarf galaxies that are currently below detection thresholds.
Future Observations: The paper argues that future searches for dark subhaloes should focus on the mass range $4.5 \times 10^7 \le M_{DM}/M_\odot \le 2.1 \times 10^8$ , as these are the most abundant in lensing regions and are the most likely to be truly dark (or contain only unresolved baryons).
Technological Requirements: To confirm whether massive lensing subhaloes are truly dark or host ultra-faint dwarfs, next-generation telescopes with extreme sensitivity (e.g., the Habitable Worlds Observatory) capable of detecting surface brightnesses fainter than 30 mag arcsec $^{-2}$ are required.
Model Constraints: The results serve as a conservative upper limit on the number of dark subhaloes. If future observations find more massive dark subhaloes than predicted by TNG50, it may indicate issues with the galaxy formation model (specifically the suppression of star formation in low-mass haloes) or the need for alternative dark matter physics.

In conclusion, the study suggests that while "truly" dark subhaloes exist, they are generally low-mass. Massive subhaloes detected in lensing studies are likely to contain faint, unresolved galaxies, challenging the assumption that lensing anomalies directly probe dark-matter-only structures at high masses.