🔭 astrophysics

Evidence of Supernova Between Formation of Stellar Populations in a Globular Cluster

This paper presents evidence that the globular cluster M92 retained supernova ejecta between the formation of its first and second stellar populations, as demonstrated by higher iron abundances in the sodium-enriched second generation compared to the first.

Original authors: Evan N. Kirby, Roman Gerasimov, Alice Cai, Benjamin Coco, Pranav Nalamwar, Lauren Henderson

Published 2026-06-30

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Evan N. Kirby, Roman Gerasimov, Alice Cai, Benjamin Coco, Pranav Nalamwar, Lauren Henderson

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 a globular cluster like M92 as a massive, ancient apartment building where stars live together. For a long time, astronomers believed that once the first group of tenants (the "First Population" or 1P) moved in, the building's "kitchen" (the gas cloud) was sealed off. No new ingredients could get in, and no waste could get out. The second group of tenants (the "Second Population" or 2P) was thought to be made from the exact same leftover ingredients, just rearranged.

However, this new paper argues that the kitchen wasn't actually sealed. It suggests that between the arrival of the first tenants and the second, a giant cosmic explosion (a supernova) happened right next door, and the building managed to catch some of the falling debris.

Here is the story of how they found this out, using simple analogies:

1. The Two Groups of Tenants

In this star cluster, there are two distinct groups of stars, like two different generations of residents:

Generation 1 (1P): These stars have "normal" amounts of Sodium (Na). Think of them as the original, plain vanilla residents.
Generation 2 (2P): These stars are "enriched." They have a lot more Sodium and Aluminum. They are the spicy, seasoned residents.

For decades, scientists thought the only difference between these two groups was how the "spices" (light elements like Sodium) were cooked up inside stars before the second group was born. They assumed the "heavy ingredients" (like Iron, which comes from supernova explosions) were the same for everyone because the cluster was too small to hold onto the heavy debris from an explosion.

2. The "Line-by-Line" Detective Work

The authors didn't just take a blurry photo of the stars; they acted like forensic scientists looking at a single fingerprint.

The Problem: If you compare stars that are different temperatures, it's hard to tell if a difference in their chemical makeup is real or just because one star is hotter than the other. It's like trying to compare the flavor of two soups when one is boiling hot and the other is cold; the heat changes how you taste it.
The Solution: The team picked stars that are almost identical twins in terms of temperature and size. Because the "cooking temperature" is the same for all of them, any difference in the chemical lines they see in the starlight must be a real difference in the ingredients.

3. The Smoking Gun: Iron

When they compared the "twins," they found something surprising.

The Vanilla Group (1P): Had a baseline amount of Iron.
The Spicy Group (2P): Had more Iron than the Vanilla group.

The Analogy: Imagine baking two batches of cookies.

Batch A (1P) is made with a standard recipe.
Batch B (2P) is made with the same recipe, but someone sneaked in a handful of chocolate chips (Iron) from a nearby explosion.
If Batch B has more chocolate chips than Batch A, it means the explosion happened after Batch A was baked but before Batch B was mixed.

4. What This Tells Us About Time

This discovery changes the timeline of how the cluster formed:

The "Vanilla" Batch: Formed in a very quick burst (less than 1 million years). The gas was well-mixed for Iron, but not for the rarest elements (like the r-process elements), suggesting the gas clouds hadn't fully blended yet.
The Explosion: A massive star exploded (a supernova) roughly 3 million years after the first batch formed.
The "Spicy" Batch: The cluster was lucky enough to catch some of the debris from that explosion. This debris added extra Iron to the gas cloud. Then, the second generation of stars formed from this "polluted" gas.

5. Why This Matters

Usually, globular clusters are too small to hold onto the heavy stuff from supernovae; it all flies away into space. Finding that M92 kept some of this debris is like finding a tiny, fragile house that managed to catch a falling brick from a nearby construction site without the wind blowing it away.

It proves that:

The second generation of stars formed at least 3 million years after the first.
The cluster did retain supernova ejecta, contradicting the old idea that these clusters are too small to do so.
The "spicy" stars (2P) are chemically distinct not just in light elements, but in heavy elements too.

Summary

The paper is essentially saying: "We looked at two groups of stars that are twins in every way except their chemical makeup. We found that the younger group has more Iron than the older group. This means a supernova exploded in between the birth of the two groups, and the cluster caught the debris. The cluster didn't just rearrange old ingredients; it added new, heavy ones from a cosmic explosion."

Technical Summary: Evidence of Supernova Between Formation of Stellar Populations in a Globular Cluster

Problem Statement
Globular clusters (GCs) exhibit unique abundance patterns of light elements (e.g., Na–O and Mg–Al anti-correlations) that cannot be fully explained by conventional chemical evolution driven by supernova enrichment. While "normal" low-sodium stars constitute the first stellar population (1P), "enriched" high-sodium stars constitute a second population (2P). The origin of these light element variations remains debated, with proposed sources including asymptotic giant branch (AGB) stars, fast-rotating massive stars (FRMSs), and extremely massive stars (EMSs). A critical constraint for distinguishing these mechanisms is the timescale of production relative to iron-peak element synthesis, which occurs in core-collapse supernovae. Most GCs, including M92, have historically been thought to lack significant dispersion in iron (Fe) abundances, implying they did not retain supernova ejecta. However, recent photometric hints and the classification of M92 as a "metal-complex" cluster suggest that subtle variations in heavy elements might exist but have been obscured by systematic errors in previous spectroscopic studies.

Methodology
To investigate the origin of abundance patterns in M92, the authors conducted a high-resolution spectroscopic study of 11 stars in M92 and 9 stars in M15 (used as a control case). The study employed a differential, line-by-line analysis strategy designed to minimize systematic uncertainties:

Sample Selection: Stars were selected from a very narrow range of effective temperature ( $T_{\text{eff}}$ ) and surface gravity ( $\log g$ ) within the red giant branch. This minimizes differences in atmospheric parameters that could mimic abundance variations.
Observations: High-resolution spectra were obtained using Keck/HIRES with a resolving power of 45,000.
Differential Analysis: Instead of calculating absolute abundances, the authors computed differential abundances for each spectral line relative to the cluster average. This approach cancels out correlated errors arising from atmospheric parameters, oscillator strengths, and the assumption of local thermodynamic equilibrium (LTE).
Statistical Rigor: The study utilized 1,000 Monte Carlo (MC) trials to propagate uncertainties in atmospheric parameters and equivalent widths. An exact permutation test was used to determine the statistical significance of differences between populations.
Independent Validation: An independent analysis using forward modeling of the entire spectrum (including non-LTE corrections for Sodium) was performed to verify the primary findings.

Key Contributions and Results
The study presents the first robust evidence that the second stellar population in M92 is enhanced in iron-peak elements compared to the first population:

Iron Enhancement in 2P: The second population (2P), defined by high Na and Al abundances, exhibits a mean iron abundance that is $0.082 \pm 0.016$ dex higher than the first population (1P). This difference is statistically significant ( $p = 0.002$ ).
Population Separation: The two populations are well-separated in Na, Al, and Fe abundances. The 1P stars show an immeasurably small scatter in Fe ( $\sigma < 0.038$ at 95% confidence), while 2P stars show a clear enhancement.
Elemental Correlations:
- Fe-peak Elements: All measured Fe-peak elements (Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Zn) are enhanced in 2P relative to 1P.
- Light Elements: The Na–Mg anti-correlation in M92 cannot be explained by a simple linear dilution of two sources; the Mg depletion in 2P is more extreme than predicted by a single enrichment source, suggesting continuous evolution over time.
- r-process Elements: The study corroborates previous findings that 1P stars in M92 show a dispersion in main r-process elements, whereas 2P stars do not. This suggests the r-process event occurred in the proto-cluster gas before the formation of 1P, while the gas remained well-mixed in Fe.
Control Case (M15): In contrast to M92, the control cluster M15 shows no significant difference in Fe abundance between its stellar populations, confirming that the result in M92 is not a systematic artifact of the analysis method.

Significance and Interpretation
The paper claims that the retention of supernova ejecta by M92 provides a critical lower limit on the time delay between the formation of the first and second stellar populations.

Supernova Retention: The presence of enhanced Fe in 2P implies that M92 retained at least some ejecta from core-collapse supernovae. Since the most massive supernovae take $\gtrsim 3$ Myr to explode, the formation of 2P must have begun at least 3 Myr after the formation of 1P.
Timescale Constraints: The results disfavor FRMSs and EMSs as the sole sources of light element enrichment in M92. FRMSs produce light elements on the same timescale as Fe, which contradicts the observed small Fe dispersion in 1P versus the large light element dispersion in 2P. Similarly, EMSs evolve too quickly (<3 Myr) to account for the observed Mg evolution within 2P.
Formation Scenario: The authors propose a formation scenario where:
1. 1P formed in a nearly instantaneous burst ( $\lesssim 1$ Myr) from gas that was well-mixed in Fe but not in main r-process elements.
2. High-temperature hydrogen burning was inactive during 1P formation.
3. 2P began forming $\gtrsim 3$ Myr later, incorporating Fe-peak elements from supernovae that exploded after 1P finished forming.
4. 2P formed over an extended duration, sampling hydrogen burning at a range of temperatures.

The paper concludes that while M92 appears chemically homogeneous in Fe at a coarse level, high-precision differential analysis reveals that it retained supernova ejecta, offering a new constraint on the chemical evolution and formation timescales of globular clusters.