🔭 astrophysics

Reionization in HESTIA: Studying reionization in the LG through zoom simulations

Using high-resolution HESTIA zoom simulations, this study reveals that Local Group reionization proceeds primarily via an inside-out process driven by internal sources rather than external fronts, with progenitors of the Milky Way and M31 ionizing significantly earlier than the cosmic average and resulting in a weak correlation between present-day satellite distances and their reionization redshifts.

Original authors: David Attard, Luke Conaboy, Noam Libeskind, Sergey Pillipenko, Keri Dixon, Ilian T. Iliev

Published 2026-03-25

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: David Attard, Luke Conaboy, Noam Libeskind, Sergey Pillipenko, Keri Dixon, Ilian T. Iliev

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: The Universe's "Morning Light"

Imagine the early Universe as a giant, pitch-black room filled with thick, cold fog. This was the "Dark Ages." Suddenly, the first stars and galaxies began to turn on like lightbulbs. Their intense ultraviolet light started to burn away the fog, turning the cold, neutral gas into hot, ionized plasma. This process is called Reionization.

For a long time, scientists have known when this happened globally (somewhere between 13 and 12 billion years ago), but they didn't know exactly how it happened in our own cosmic neighborhood, the Local Group (which includes our Milky Way, the Andromeda galaxy, and their smaller satellite galaxies).

Did the light come from outside, washing over us like a sunrise? Or did our own local galaxies turn on their lights first, clearing the fog around them from the inside out?

The Experiment: A Cosmic "Zoom-In"

To answer this, the researchers used a supercomputer simulation called HESTIA. Think of this simulation as a massive, 3D video game of the universe.

The Wide Shot: First, they ran a simulation of a huge chunk of the universe (100 million light-years across) with lower detail. This helped them calibrate their "lightbulbs" (the galaxies) to make sure the simulation matched real-world observations.
The Zoom-In: Then, they took a "zoom lens" and focused intensely on just our neighborhood (the Local Group). They cranked up the resolution so high that they could see tiny, dwarf galaxies that would otherwise be invisible. This is like switching from a blurry satellite photo of a city to a high-definition street view where you can see individual houses.

They ran this simulation four different times, tweaking the rules of physics slightly each time to see how different types of "feedback" (like supernova explosions blowing gas away) would change the story.

The Key Findings

Here is what they discovered, explained with analogies:

1. The "Inside-Out" Party

The Question: Did the Milky Way and Andromeda wait for the rest of the universe to clear the fog, or did they clear it themselves?
The Result: They cleared it themselves, from the inside out.
The Analogy: Imagine a group of friends in a dark room. You might expect the lights to turn on from the ceiling (external sources). But in this simulation, the friends (the Milky Way and Andromeda) turned on their own flashlights first. They burned away the fog around them before the "sunrise" from the rest of the universe even arrived.
The Takeaway: Even in the most extreme scenarios where they tried to make external light stronger, the local galaxies were so bright and dense that they ionized their own surroundings first. The "external front" from the rest of the universe played almost no role.

2. The "Rich Get Richer" (and Light Up First)

The Question: Did the heavy, massive galaxies get ionized earlier than the small, weak ones?
The Result: Yes.
The Analogy: Think of the Milky Way and Andromeda as heavyweights in a boxing ring. Because they are so massive, they gathered their "fighters" (gas and stars) together very early. They started glowing and burning away the fog around them when the universe was still very young (around 500 million years old).
The Timing: The global universe took until about 700–800 million years old to reach the halfway point of reionization. But our local giants reached that halfway point much earlier, around 500–600 million years old. They were essentially "early risers."

3. The Satellite Mystery

The Question: What about the tiny dwarf galaxies orbiting the Milky Way and Andromeda? Did they get cleared out at the same time?
The Result: It's a bit of a mess, but there's a pattern.
The Analogy: Imagine the Milky Way is a lighthouse. The satellites are small boats nearby. You'd expect the boats closest to the lighthouse to get lit up first.
The Reality: The data shows a very weak connection. The boats closest to the lighthouse did get lit up slightly earlier, but there was a lot of scatter. Some far-away boats got lit up early because they had their own small flashlights (they were massive enough to ionize themselves), while some close boats stayed in the dark longer.
The Twist: The pattern was slightly clearer around Andromeda (M31) than the Milky Way, suggesting Andromeda's neighborhood might be a bit more orderly.

4. The Oldest Stars Live in the Biggest Houses

The Question: Where do the oldest stars in our neighborhood live today?
The Result: They live in the biggest, most massive dwarf galaxies.
The Analogy: Think of the "fog" (reionization) as a flood. If you build a house (a galaxy) before the flood hits, you have time to build a strong foundation and gather furniture (stars). If you try to build after the flood, the water washes away your materials, and you can't build much.
The Conclusion: The simulation showed that the dwarf galaxies that formed before the fog cleared (reionization) managed to grow into massive, star-filled galaxies. The ones that tried to form after the fog cleared were starved of gas and remained small and dark. Therefore, if you want to find the oldest stars in the Local Group, look in the biggest dwarf galaxies, not the tiny ones.

Why Does This Matter?

This paper solves a puzzle about our cosmic home. It tells us that the Milky Way and Andromeda are largely self-sufficient when it comes to their early history. They didn't need the rest of the universe to "save" them from the Dark Ages; they saved themselves.

It also gives us a roadmap for future telescopes. If we want to find the oldest, most pristine stars in our neighborhood, we shouldn't look at the tiny, faint dwarf galaxies. We should look at the massive ones, because those are the ones that were "born" before the cosmic fog cleared, allowing them to grow up strong and old.

In short: Our local galaxies were the early birds that cleared their own nests, and the biggest of the little galaxies are the ones that kept the oldest family heirlooms (stars).

1. Problem Statement

While the global history of cosmic reionization (EoR) is constrained by observables like the Cosmic Microwave Background (CMB) and quasar absorption spectra, the specific interplay between internal ionizing sources (galaxies within the Local Group) and external ionizing fronts (from large-scale structure) remains poorly understood.

Key Questions: Did the Local Group (LG) reionize via an "inside-out" mechanism driven by the Milky Way (MW) and Andromeda (M31), or were external fronts dominant? How do feedback mechanisms (radiative and mechanical) affect the timing of reionization in low-mass haloes? What is the correlation between a satellite galaxy's reionization redshift ( $z_{\text{reion}}$ ) and its present-day properties (mass, distance from host)?
Challenge: Full-volume simulations capable of resolving the LG's complex structure often lack the resolution to model low-mass haloes ( $10^8 - 10^9 M_\odot$ ) where reionization physics is critical. Conversely, zoom-in simulations lack the large-scale context to capture external ionizing fronts.

2. Methodology

The authors employed a hybrid approach combining high-resolution $N$ -body simulations with radiative transfer (RT) post-processing.

Simulations:
- HESTIA-1024: A uniform $1024^3$ dark-matter-only (DMO) simulation used to calibrate source models and subgrid physics.
- HESTIA-Zoom: A high-resolution zoom-in simulation centered on the MW-M31 pair with an effective resolution of $4096^3$ particles in the zoom region, resolving haloes down to $10^8 M_\odot$ .
- S1 (Schwandt et al. in prep): A high-resolution $4096^3$ simulation used to calibrate the subgrid collapse fraction model.
Subgrid Modeling:
- Since the $1024^3$ run cannot resolve haloes between $10^8$ and $10^{9.4} M_\odot$ , the authors coupled the simulation with an instantaneous halo bias subgrid model. This model predicts the collapse fraction ( $f_{\text{coll}}$ ) of unresolved Low-Mass Atomic Cooling Haloes (LMACHs) based on local overdensity ( $\delta$ ), calibrated against the high-resolution S1 simulation.
- Sources were categorized into LMACHs ( $10^8 < M < 10^9 M_\odot$ ) and High-Mass Atomic Cooling Haloes (HMACHs) ( $M > 10^9 M_\odot$ ).
Radiative Transfer (RT) & Source Models:
- Used PyC2Ray for RT post-processing on a $256^3$ grid.
- Tested four distinct scenarios varying two factors:
  1. Suppression Models: Full Suppression (LMACHs in ionized regions produce zero photons) vs. Partial Suppression (LMACHs in ionized regions produce photons at HMACH efficiency).
  2. Source Efficiencies ( $g_\gamma$ ): Varied photon production rates for HMACHs and LMACHs to match global constraints.
- Calibration: Runs 1 & 2 were calibrated to match the same global reionization history but with different feedback strengths. Runs 3 & 4 used identical efficiencies but different suppression models to isolate feedback effects.
Reionization Time Definition: $z_{\text{reion}}$ for a halo is defined as the redshift where the mass-weighted ionization fraction of its constituent dark matter particles exceeds 50%.

3. Key Contributions

Validation of Zoom-RT for LG Studies: Demonstrated that zoom-in simulations, when properly calibrated against global constraints, accurately reproduce the reionization history of the MW and M31 without needing explicit external radiation maps from a parent box.
Subgrid Calibration: Successfully integrated a subgrid collapse-fraction model to account for unresolved low-mass haloes, ensuring the ionizing photon budget is accurate even in lower-resolution calibration runs.
Systematic Feedback Analysis: Provided a controlled comparison of "Full" vs. "Partial" suppression models to quantify how radiative feedback impacts the topology and timing of reionization specifically within the Local Group context.

4. Key Results

Inside-Out Reionization: In all four scenarios, reionization within the LG proceeds in an inside-out manner. The progenitors of the MW and M31 reionized their immediate surroundings significantly earlier ( $z \approx 8.6 - 9.3$ ) than the global cosmic midpoint ( $z \approx 7.1 - 7.7$ ).
Negligible External Influence: External ionizing fronts from large-scale structure played a negligible role, even in the most permissive feedback models. The LG is effectively isolated in its reionization history.
Satellite Correlations:
- There is a weak negative correlation between a satellite's $z_{\text{reion}}$ and its present-day distance from the host (MW or M31). Satellites closer to the host tend to reionize slightly earlier.
- The correlation is slightly stronger for M31 satellites than for MW satellites, with less scatter.
- Most satellites reionized much earlier than the global average because their Lagrangian volumes reside in overdense regions.
Mass-Dependent Assembly: Satellites that assembled before their reionization epoch are systematically more massive at $z=0$ than those assembling after. This suggests that the oldest stellar populations preferentially reside in the most massive subhaloes.
Feedback Impact: While feedback models (Full vs. Partial suppression) altered the total ionized volume and the specific timing of the global reionization midpoint, they did not fundamentally change the inside-out topology of the LG.

5. Significance

Methodological Efficiency: The study validates a computationally efficient pipeline for studying LG reionization. It confirms that high-resolution zoom simulations, calibrated against global constraints, are sufficient to study the internal reionization of the Local Group without the prohibitive cost of full-box RT simulations.
Understanding the "Missing Satellites" Problem: The results reinforce the idea that radiative feedback (photo-heating) suppresses star formation in low-mass haloes. The finding that older populations are in more massive haloes helps explain the distribution of dark matter subhaloes and the "missing satellites" problem.
Local Group Specificity: The confirmation that the LG reionized earlier than the cosmic mean due to its overdense environment highlights the importance of local environmental factors in galaxy evolution, distinct from global cosmological averages.
Future Directions: The authors propose applying this calibrated pipeline to the full suite of HESTIA constrained realizations to quantify the variance in reionization histories caused by different large-scale environments.