🔭 astrophysics

Reaching the Metallicity Floor at $z\sim 10$ : Lensed Star Clusters at Cosmic Dawn and Cosmic Noon

This study utilizes gravitational lensing to analyze star clusters in two systems spanning Cosmic Dawn and Cosmic Noon, revealing that the high-redshift GEMS clusters formed at $z \sim 10$ with extremely low metallicities consistent with a metallicity floor, while the lower-redshift Sparkler clusters formed later with significantly higher enrichment, together tracing the evolution of cluster-forming environments from the early universe to Cosmic Noon.

Original authors: Raul Jimenez, Elena Tomasetti, Carmela Lardo, Licia Verde

Published 2026-07-29

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Raul Jimenez, Elena Tomasetti, Carmela Lardo, Licia Verde

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

Imagine the universe as a giant, cosmic kitchen where stars are the chefs and galaxies are the restaurants. For a long time, astronomers have been trying to figure out the recipe for the universe's oldest, most stubborn dishes: globular clusters. These are massive, tightly packed balls of stars that have been around since the dawn of time, acting like cosmic fossils. Just as a geologist can tell the history of the Earth by looking at layers of rock, astronomers can read the history of the universe by studying these star clusters. The key ingredient they look for is "metallicity." In astronomy, "metals" don't just mean iron or gold; they mean any element heavier than hydrogen and helium. When the universe began, it was made almost entirely of the lightest ingredients. As stars lived and died, they cooked up heavier elements and scattered them into space, making the next generation of stars slightly more "seasoned." By measuring how metal-rich a star cluster is, scientists can tell when it was born: the older and more metal-poor it is, the closer to the beginning of time it formed.

The big question has always been: how did this seasoning process start? Did the very first star clusters form in a completely pristine, unseasoned kitchen, or were they born in a place that had already been stirred by previous cooking? To answer this, we need to look at two very different moments in the universe's life: the "Cosmic Dawn," when the very first galaxies were flickering into existence, and "Cosmic Noon," a later era when the universe was bustling with star formation. Until recently, we could only guess what happened at the Cosmic Dawn because those early galaxies were too faint and far away to see clearly. But with the powerful new eyes of the James Webb Space Telescope (JWST), combined with the natural magnifying glasses of the universe (gravitational lensing), we can finally peek into these ancient kitchens.

The Paper's Story: A Tale of Two Star Clusters

This paper takes a deep dive into two specific cosmic "dishes" that were caught on camera at these two different times. The first is a system called the "Cosmic Gems arc" (or GEMS), which we see as it was when the universe was only about 500 million years old. The second is a system called the "Sparkler," which we see from a much later time, when the universe was roughly 4.6 billion years old. Because the Sparkler is closer to us, we can see it more clearly, but the GEMS system is the real star of the show because it is so ancient.

The researchers used a special technique to untangle the light from the GEMS system, which had been stretched and magnified by a giant galaxy cluster sitting between us and the arc. They found ten unique star clusters in the GEMS system, each appearing twice because of the lensing effect. They then compared these ancient clusters to the five star clusters found in the Sparkler system.

What They Found

The results are like finding a perfectly unseasoned soup at the beginning of a feast and a rich, savory stew at the middle of it.

The GEMS Clusters (Cosmic Dawn): The star clusters in the GEMS system are incredibly young when we see them (only about 38 million years old) and they are extremely metal-poor. The paper calculates that their metallicity is around [Z/H] = -2.3. To put that in perspective, this means they have only about 0.5% of the heavy elements found in our Sun. This is right at the "metallicity floor"—the lowest limit we see for the oldest star clusters in our own Milky Way galaxy. The data strongly suggests that these clusters formed in a "closed box" environment. Imagine a small, isolated pot of water that has never been touched by anything else. The gas in these tiny, early galaxies was so pristine that it had barely been touched by any previous stars. The paper rules out the idea that these clusters were formed in a metal-rich environment; the data explicitly excludes metallicity levels as high as [Z/H] = -0.5. Instead, they formed in the very first halos of dark matter that were massive enough to cool down and make stars, essentially being the "first generation" of stellar chefs.
The Sparkler Clusters (Cosmic Noon): In contrast, the star clusters in the Sparkler system formed much later, about 2.5 billion years after the GEMS clusters, during the peak epoch of cosmic star formation known as "Cosmic Noon." They are significantly more "seasoned," with a metallicity of about [Z/H] = -0.5. This is still less than the Sun, but it is much richer than the GEMS clusters. These clusters formed in a more common type of dwarf galaxy that was growing steadily. The paper suggests this galaxy was an "open system," meaning it was constantly being fed fresh, metal-poor gas from the cosmic web while also losing gas through winds. This constant mixing and dilution kept the metal levels at a moderate, intermediate level, rather than letting them get too high or too low.

Why It Matters

The paper puts these two systems on a single timeline to show the evolution of the universe's chemical makeup. It confirms that the universe really did start with a "metallicity floor" and slowly got richer over time. The GEMS clusters represent the moment when the universe was just starting to cook, forming stars in rare, isolated pockets of pristine gas. The Sparkler clusters represent a later stage where the universe was a busy, interconnected kitchen, with gas flowing in and out, creating a more complex chemical environment.

The authors are very careful to note that while they are confident about the difference in metallicity between the two groups, the exact number for the GEMS clusters has some uncertainty because the light from such young, distant objects is hard to measure perfectly. However, they are sure that the GEMS clusters are definitely very metal-poor and that the Sparkler clusters are definitely more metal-rich. They also clarify that while these clusters look like the ancestors of the star clusters in our own Milky Way, we can't say for sure if they are the exact same ones that ended up here; they are just analogues that show us what the conditions were like at those different times.

In short, this paper uses the power of JWST to take a snapshot of the universe's "cooking" process at two different stages. It shows us that the first star clusters were born in a nearly pure, unseasoned universe, while later clusters formed in a galaxy that had already been stirred by billions of years of cosmic history. It's a vivid confirmation that the universe has been getting more complex and "flavored" since the very beginning.

Technical Summary: Reaching the Metallicity Floor at z ∼10: Lensed Star Clusters at Cosmic Dawn and Cosmic Noon

Problem and Motivation
The origins of globular clusters (GCs) are intrinsically linked to the assembly of their host galaxies. While local GC systems serve as archaeological probes of hierarchical assembly, their present-day properties have been altered by over ten billion years of dynamical processing, tidal shocks, and accretion. To directly probe the conditions of early star formation, this study analyzes two strongly lensed systems that bracket the epochs of "Cosmic Dawn" and "Cosmic Noon": the Cosmic Gems arc (GEMS) at $z_{obs} = 9.625$ and the Sparkler galaxy at $z_{obs} = 1.378$ . The central problem addressed is whether the chemical enrichment of cluster-forming gas follows the expected cosmic timeline and how the formation environments of these high-redshift clusters compare to theoretical regimes of atomic-cooling halos and accretion-regulated dwarf galaxies.

Methodology
The analysis relies on a homogeneous compilation of compact star-cluster candidates derived from JWST/NIRCam imaging.

Data Processing: For GEMS, new deconvolution photometry was performed using the STARRED package (Millon et al. 2024) on eight bands (F090W–F444W). This separated compact point sources from diffuse arc emission, identifying ten unique cluster candidates (each doubly imaged), including five bright clusters, four fainter tail sources, and a newly discovered pair. For the Sparkler, the study adopted stellar parameters derived by Tomasetti et al. (2025) for five GC candidates, correcting for gravitational magnification.
SED Fitting: Physical parameters (mass, age, metallicity, extinction) were derived using the Bayesian SED-fitting code BAGPIPES. The analysis employed Bruzual & Charlot (2003, BC03) stellar-population models with a Kroupa IMF and a Calzetti dust law. Crucially, no cosmological age prior was imposed, ensuring ages were independent of the assumed background cosmology.
Model Variations: To test systematics, the entire catalogue was refit using BPASS models (Eldridge et al. 2017), which include interacting binaries and produce harder UV spectra. Star-formation histories (SFH) were tested under single-burst, exponentially declining, and delayed- $\tau$ assumptions, with the single-burst model adopted as the reference case.
Chemical Evolution Frameworks: The results were interpreted against two limiting chemical-evolution models:
1. A closed-box model for GEMS, assuming formation in rare atomic-cooling halos ( $M_h \sim 10^8 M_\odot$ ) near the metallicity floor.
2. An open-system gas-regulator model for the Sparkler, assuming formation in a dwarf galaxy where ISM metallicity is set by the balance of enrichment, metal-poor inflow, and outflow.
Cosmological Context: Formation redshifts ( $z_{form}$ ) were calculated by subtracting the measured cluster age from the age of the Universe at $z_{obs}$ , utilizing a flat $\Lambda$ CDM cosmology (Planck Collaboration et al. 2020) only for this conversion step.

Key Results

Formation Epochs: The two populations occupy distinct epochs. The GEMS clusters formed at $z_{form} \approx 10–11$ (median $z_{form} \simeq 10.2$ ), within the first $\sim 500$ Myr of cosmic history. The Sparkler clusters formed $\sim 2.5$ Gyr later, at $z_{form} \approx 2–3.5$ (median $2.3$), during the peak of cosmic star formation.
Metallicity Floor at Cosmic Dawn: The GEMS clusters are extremely metal-poor. The inverse-variance-weighted population metallicity is $[Z/H] = -2.34 \pm 0.27$ $[Z / H] = - 2.34 \pm 0.27$ (corresponding to $Z \simeq 0.5\% Z_\odot$ $Z ≃ 0.5% Z_{⊙}$ ).
- Individual clusters are consistent with $[Z/H] \lesssim -1.2$ at 68% confidence.
- The data exclude $[Z/H] \ge -0.5$ .
- While the photometry cannot reliably distinguish among values below $[Z/H] \simeq -1.5$ , the formal estimate is consistent with the metallicity floor of Milky Way GCs ( $[Fe/H] \simeq -2.5$ to $-2.3$).
- This conclusion holds for both BC03 and BPASS templates, with BPASS refits yielding similarly low metallicities ( $[Z/H] = -2.2$ to $-2.7$).
Enrichment at Cosmic Noon: The Sparkler clusters are substantially more enriched, with a median metallicity of $[Z/H] \approx -0.48$ . This intermediate, sub-solar metallicity is consistent with a gas-regulator equilibrium in a dwarf galaxy environment.
Halo Mass and Efficiency:
- GEMS: The formation epoch matches the extended Press–Schechter prediction for the assembly of halos at the atomic-cooling threshold ( $M_h \sim 10^8 M_\odot$ ). The implied star-formation efficiencies ( $\epsilon$ ) range from a few percent to formally $\sim 100\%$ for the most massive clusters if they formed in single threshold halos, suggesting the clusters formed in distinct progenitor halos that merged rapidly.
- Sparkler: The host halo mass is estimated at $M_{halo} \sim 10^{10.7}–10^{11} M_\odot$ , two orders of magnitude above the atomic-cooling threshold, consistent with an ordinary dwarf galaxy.

Significance and Claims
The paper claims to provide a direct empirical bridge between young massive clusters at high redshift and old GC populations observed today.

Chronological Enrichment: The study establishes that cluster-forming gas grows progressively more enriched with cosmic time, with the earliest clusters (GEMS) forming in near-pristine gas and later clusters (Sparkler) forming in a chemically evolved, regulated environment.
Two Modes of Formation: The systems are identified as fossils of two distinct, epoch-dependent modes of galaxy assembly:
1. A primordial burst mode ( $z \gtrsim 10$ ) in rare, atomic-cooling halos, where clusters form before significant self-enrichment, resulting in metallicity-floor values.
2. A secular accretion mode ( $z \approx 2–3.5$ ) in common dwarf galaxies, where continuous inflow of metal-poor gas regulates the ISM metallicity to intermediate levels.
Comparison to Milky Way: When placed on the Milky Way age–metallicity plane, GEMS coincides with the enrichment onset of the earliest accreted progenitors (e.g., Gaia-Sausage-Enceladus), while the Sparkler overlaps with the enriched in-situ population. The authors caution that while these systems resemble specific MW components, the comparison identifies similarities in formation conditions rather than uniquely determining an in-situ or ex-situ assembly channel for the high-redshift systems.

Caveats and Limitations
The authors explicitly state that:

The specific metallicity value for GEMS ( $[Z/H] \simeq -2.3$ ) is prior-dependent, as broadband photometry cannot distinguish values below $[Z/H] \simeq -1.5$ .
The identification of the two systems with specific formation regimes is an internally consistent interpretation subject to cosmic variance, given the sample size of one system per epoch.
The survival of these clusters to $z=0$ is untested; the analysis focuses on birth conditions rather than the final survivor population.
Current photometry constrains nebular gas (via existing PRISM data) rather than stellar photospheres; deeper spectroscopy is required to directly measure birth metallicities and resolve the metallicity floor.

More like this