Accelerated size evolution in the FirstLight simulations from z=14 to z=5

Using the FirstLight simulations and synthetic JWST images, this study reveals that galaxy sizes at cosmic dawn ($z=14$ to $z=5$) evolve rapidly due to increasing galaxy efficiency and complex dust attenuation effects, establishing a diverse size-mass relation early on where compact galaxies exhibit higher specific star-formation rates.

The Gist

Imagine the early Universe as a bustling, chaotic construction site just a few hundred million years after the Big Bang. This paper is like a detailed blueprint and a time-lapse video of how the very first "cities" (galaxies) were built, grew, and changed shape during that frantic first billion years.

Here is the story of the paper, broken down into simple concepts:

1. The Big Question: How fast do baby galaxies grow?

For a long time, astronomers thought galaxies grew slowly, like a child growing a few inches a year. But the new James Webb Space Telescope (JWST) has started spotting massive, bright galaxies very early in the Universe's history. This was a shock! It's like finding a fully grown oak tree in a garden that was only planted yesterday.

The authors wanted to know: Are these galaxies just growing taller (more mass), or are they also getting wider (larger in size)? And if they are getting wider, how fast is that happening?

2. The Tool: A Cosmic "Time Machine"

To answer this, the team didn't just look at the sky; they built a massive digital universe. They ran 430 separate, high-definition simulations (think of them as 430 different "what-if" movies) of galaxy formation.

• The Resolution: They zoomed in so closely that they could see details as small as a few light-years across. It's like looking at a city from space and being able to count the cars on the streets.
• The Output: They didn't just look at the raw data; they used a special program (SKIRT) to paint "synthetic photos" of these galaxies, exactly as the JWST telescope would see them, including the effects of dust and the telescope's own blur.

3. The Discovery: The "Size-Mass" Dance

The team found that the relationship between a galaxy's mass (how much stuff is in it) and its size (how big it looks) was already established at the very beginning of time (when the Universe was only about 300 million years old).

However, there was a wild twist: Galaxies of the same weight could look very different.

• The "Party Animals": Some galaxies were huge and spread out. These were the ones having a massive "party," creating stars at a frantic pace (high star-formation rate).
• The "Compact Clusters": Others were tiny, dense, and compact. These were the quiet ones, or perhaps the ones that had just finished a burst of activity and were shrinking back down.

Analogy: Imagine two people weighing exactly 150 lbs. One is a tall, lanky basketball player (extended galaxy), and the other is a short, muscular weightlifter (compact galaxy). In the early Universe, both types existed side-by-side.

4. The Secret Ingredient: Dust as a "Makeup Artist"

One of the most important findings is about dust. In the early Universe, galaxies were full of cosmic dust (like soot from a fire).

• The Problem: Dust is great at blocking light. It acts like a heavy curtain in the center of a galaxy, hiding the bright, young stars there.
• The Effect: Because the center is darkened by dust, the "brightest" part of the galaxy looks like it's further out on the edges. This tricks the telescope into thinking the galaxy is larger than it actually is.
• The Result: The paper shows that if you remove the dust in the simulation, the galaxies look much smaller and more compact. The dust is essentially "inflating" the size of the galaxy in our eyes.

5. The Main Conclusion: An Accelerated Growth Spurt

The biggest headline of the paper is that galaxies grew in size incredibly fast between redshift 14 (very early) and redshift 5 (a bit later).

• The Speed: In just 600 million years (which is a blink of an eye in cosmic time), the "normal" size of a galaxy increased by a factor of about 3 (0.5 dex).
• Why? It's driven by efficiency. In the early Universe, gas was denser and fell into galaxies faster. It was like a construction crew that had unlimited materials and no traffic jams; they could build the whole city in a week. As the Universe got older and gas became harder to gather, the construction slowed down.

6. Why This Matters

This paper helps us understand why the early Universe looked so different from today.

• Then: Galaxies were tiny, dense, and growing at breakneck speed, often hidden behind thick curtains of dust.
• Now: They are larger, more spread out, and growing more slowly.

The Takeaway Metaphor:
Think of the early Universe as a spray-paint can. When you first press the nozzle, the paint comes out in a tiny, incredibly dense, high-pressure burst (the compact, fast-growing galaxies). As you keep spraying, the stream widens, the pressure drops, and the paint spreads out over a larger area (the slower, larger galaxies we see today).

The FirstLight simulations confirm that this "spray" was much more intense and rapid in the beginning than we previously thought, driven by the sheer efficiency of the early cosmos.

Technical Summary

Here is a detailed technical summary of the paper "Accelerated size evolution in the FirstLight simulations from z=14 to z=5" by Ceverino et al.

1. Problem Statement and Context

The James Webb Space Telescope (JWST) has revealed a population of bright, massive galaxies at extremely high redshifts ($z = 9-15$), suggesting accelerated mass growth during the first Gyr of the Universe. A critical open question is how the physical sizes of these galaxies evolve during "cosmic dawn" ($z \gtrsim 10$) and what drives this evolution.

Previous observations (HST) established a size-mass relation at $z \le 3$ with a slope $\alpha \approx 0.2$ and a redshift evolution slope $\alpha_z \approx -1$. However, early JWST data presents conflicting trends: some studies suggest a slower evolution ($\alpha_z > -1$), while others suggest steeper evolution ($\alpha_z \approx -1.5$) or extreme compactness (e.g., $R_e \approx 260$ pc at $z=14$).

Current cosmological simulations struggle to reproduce these size-mass relations, often predicting flat or negative slopes due to resolution limits and the complex interplay between dust attenuation and stellar light. The authors aim to resolve these discrepancies by analyzing the FirstLight (FL) simulation suite to understand the drivers of galaxy size evolution from $z=14$ to $z=5$.

2. Methodology

Simulation Suite (FirstLight)

The study utilizes the FirstLight database, comprising 430 zoom-in cosmological simulations of the first galaxies.

• Resolution: High spatial resolution (down to $\sim 8.7$ pc) and mass resolution ($10^4 M_\odot$ for dark matter).
• Physics: The simulations use the ART code (N-body + Hydro) with adaptive mesh refinement (AMR). Key physics include:
  • Gas cooling (H, He, metals).
  • Stochastic star formation (threshold density $1 \text{ cm}^{-3}$,$T < 10^4$ K).
  • Feedback models: Thermal, kinetic (supernovae/stellar winds), and radiative feedback (RadPre_IR, modeling radiation pressure and IR photon trapping).
  • Metal enrichment from Type II and Ia supernovae.

Synthetic Image Generation

To directly compare simulations with JWST observations, the authors generated synthetic images using SKIRT (radiative transfer code).

• Stellar/Nebular Emission: Intrinsic emission calculated using BPASS (stellar populations) and TODDLERS (nebular emission).
• Dust Modeling: Dust distribution derived from metal density with a constant dust-to-metal ratio ($DMR=0.4$). A Weingartner & Draine (2001) dust mix was used.
• Observational Effects:
  • Images generated in 7 JWST wide bands (F090W to F444W).
  • 9 different viewing angles (face-on, edge-on, random).
  • Addition of background noise (mimicking JADES-GDS survey, $5\sigma \approx 30$ AB mag) and Point Spread Function (PSF) convolution.
• Size Measurement: A non-parametric, pixel-based method was used to calculate the effective radius ($R_e$) containing 50% of the flux above a $4\sigma$surface brightness limit, avoiding biases from Sersic fitting on irregular high-$z$ galaxies.

3. Key Contributions

1. High-Resolution Statistical Sample: The largest suite of zoom-in simulations ($\sim 400$ independent regions) covering the epoch from $z \approx 14$ to $z \approx 5$ with sufficient resolution to resolve sub-kpc structures.
2. Radiative Transfer Integration: The first application of full radiative transfer (SKIRT) to a statistical sample of high-$z$ galaxies to generate multi-band synthetic images, explicitly accounting for differential dust attenuation.
3. Decoupling Intrinsic vs. Observed Sizes: The study distinguishes between the intrinsic stellar distribution and the observed size, demonstrating how dust significantly alters the measured size-mass relation, particularly in the rest-frame optical.

4. Key Results

A. Size-Mass Relation and Diversity

• Existence at High-$z$: The size-mass relation is already established at $z \simeq 14$.
• Scatter and sSFR: There is a large intrinsic scatter at fixed mass.
  • Extended galaxies (larger $R_e$) tend to have higher specific Star Formation Rates (sSFR) and are often undergoing mergers or starbursts.
  • Compact galaxies tend to have lower sSFR, often associated with "mini-quenching" or post-burst phases where gas is concentrated centrally.
• Slope Stability: The mass-dependent slope ($\alpha$) remains roughly constant ($\alpha \approx 0.2$) from $z=5$ to $z=14$, consistent with some JWST observations.

B. Accelerated Size Evolution

• Rapid Normalization Change: While the slope is constant, the normalization of the size-mass relation evolves rapidly. Between $z \simeq 14$ and $z \simeq 6$ (a span of $\sim 600$ Myr), the normalization increases by 0.5 dex.
• Evolution Rate: The redshift evolution slope is steep, $\alpha_z \approx -1.8$ to $-2.0$ for fixed stellar mass ($10^9 M_\odot$). This is significantly faster than previous simulations ($\alpha_z \approx -1.3$) and some JWST observations ($\alpha_z \approx -0.2$).
• Driver: This acceleration is driven by increasing galaxy efficiency (stellar-to-halo mass ratio) at $z \ge 5$ and efficient loss of angular momentum (wet compaction) in the early Universe.

C. The Role of Dust Attenuation

• Differential Attenuation: Dust preferentially dims the central, dense star-forming regions. This makes galaxies appear larger in observed images than their intrinsic stellar distributions would suggest.
• Slope Modification: In the rest-frame optical (F356W/F444W), the observed slope is positive ($\alpha \approx 0.12$). However, without dust, the slope becomes slightly negative ($\alpha \approx -0.02$). This implies that massive galaxies are intrinsically more compact (due to compaction events), but dust attenuation inflates their observed sizes, creating a positive correlation.
• Compaction Threshold: The transition to negative slopes in dust-free images coincides with the "turn-on mass" ($M_* \approx 10^9 M_\odot$) where wet compaction becomes efficient.

D. Star Formation Rate Surface Density ($\Sigma_{SFR}$)

• $\Sigma_{SFR}$ increases with redshift, reaching extreme values ($\sim 100 M_\odot \text{ yr}^{-1} \text{ kpc}^{-2}$) at $z \approx 14$.
• This is driven by high SFRs and smaller physical sizes, consistent with high galaxy formation efficiencies.

5. Significance and Conclusions

• Accelerated Growth: The paper confirms that galaxy size evolution accelerates at cosmic dawn ($z \ge 5$), driven by high galaxy formation efficiencies and rapid angular momentum loss.
• Dust is Critical: The study highlights that dust attenuation is not a minor correction but a fundamental driver of observed size trends. It can mask intrinsic compactness and artificially flatten or invert the size-mass slope. Models comparing to observations must include radiative transfer to be valid.
• Resolution of Discrepancies: The steep evolution found in FirstLight ($\alpha_z \approx -2$) suggests that current JWST observations at $z > 10$ might be missing the faintest, most compact galaxies, or that the evolution slows down significantly after $z=5$.
• Future Directions: The authors argue that future simulations require parsec-scale resolution and improved dust production/destruction models to fully resolve the complex mix of dust and stars in the first galaxies.

In summary, this work provides a robust theoretical framework linking the rapid structural evolution of early galaxies to their high formation efficiencies and the critical role of dust in shaping our observational view of the early Universe.