A constant upper luminosity limit of cool supergiant stars down to the extremely low metallicity of I Zw 18

The paper argues that cool supergiants at extremely low metallicity are limited by a constant upper luminosity due to a metallicity-independent late-phase mass loss mechanism that strips hydrogen layers, allowing massive stars to evolve into hot, helium-rich objects capable of producing hard ionizing radiation and contributing to early Universe nitrogen enrichment.

The Gist

Imagine the universe as a giant, bustling construction site. The "bricks" of this site are stars, and the "foremen" are massive stars that live fast, burn bright, and die young. For decades, astronomers had a specific rulebook for how these massive stars behave as they near the end of their lives.

The rulebook said: "If a star is born in a metal-poor environment (like the early universe), it should hold onto its hydrogen 'jacket' longer and grow into a giant, cool, red supergiant. The bigger the star, the brighter and redder it gets."

However, a new study by Abel Schootemeijer and his team has just torn up that rulebook. They found that no matter how metal-poor the environment is—even in the most primitive, "metal-starved" galaxies—the universe has a strict brightness cap on these cool, red giants.

Here is the story of their discovery, explained simply.

1. The "Red Giant" Ceiling

Think of a cool supergiant star like a balloon being inflated. As it gets bigger and brighter, it gets cooler and redder.

• The Old Theory: In galaxies with very little "metal" (elements heavier than hydrogen and helium, like gold or iron), the wind blowing off the star is weak. So, the star should keep inflating, getting brighter and brighter, until it becomes a super-massive, super-bright red giant.
• The New Reality: The team looked at galaxies ranging from our own neighborhood (the Milky Way) to the extremely metal-poor dwarf galaxy I Zw 18 (which has only 1/40th the metal of our Sun). They found a "ceiling." No matter how metal-poor the galaxy, no cool red giant ever gets brighter than a specific limit.

It's as if there is an invisible glass ceiling in the sky. Once a star tries to grow brighter than this limit, something snaps, and it stops being a cool, red giant.

2. The "Great Disappearance"

The researchers used the James Webb Space Telescope (JWST)—the most powerful camera in space—to look at I Zw 18. They expected to see a few super-bright red giants there, because the low metal content should have let them grow huge.

Instead, they saw nothing.

• The Analogy: Imagine you are looking at a forest. You expect to see giant oak trees. But when you look at the most remote, untouched part of the forest, you see only small saplings and medium-sized trees. The giant oaks are missing.
• The Twist: The stars didn't just vanish; they changed costumes. The massive stars that should have been giant red supergiants are actually showing up as Wolf-Rayet stars. These are the "naked" versions of the stars—they have lost their hydrogen "jacket" and are burning hot, blue, and helium-rich.

3. The "Stripping" Mystery

Why do these stars lose their jackets so early in metal-poor environments?

• The Wind Theory: Usually, stars lose mass because of strong winds, and these winds depend on metal (like a sail catching the wind). Less metal = weaker wind = less mass loss. But this doesn't explain why the red giants disappear in metal-poor places.
• The New Idea: The authors suggest that these massive stars have a self-destruct mechanism that doesn't care about metal. Once they get too bright (crossing that "ceiling"), they violently shed their outer layers, regardless of the environment. It's like a pressure valve that pops open automatically when the pressure gets too high, blowing off the star's skin and turning it into a hot, naked helium star.

4. Why Does This Matter?

This discovery changes how we understand the early universe and the black holes we see today.

• The Black Hole Limit: If massive stars shed their heavy outer layers before they die, they end up lighter. This means the black holes they leave behind might be smaller than we thought. It puts a "speed limit" on how heavy the black holes in the early universe could be.
• The Cosmic Glow: These hot, naked stars are like powerful UV lamps. Because they are so hot and have weak winds (due to low metal), their intense radiation can escape and light up the surrounding gas. This helps explain why we see certain chemical signatures (like helium and carbon) in the light from the very first galaxies.
• The Nitrogen Puzzle: High-redshift galaxies (very old ones) have surprisingly high levels of nitrogen. The authors suggest that these "stripped" massive stars are the culprits, spewing out processed nitrogen before they explode.

The Bottom Line

The universe has a strict dress code for massive stars. No matter how "primitive" the environment, if a star gets too massive, it is forced to shed its cool, red outer layers and transform into a hot, blue, helium-burning star.

The "cool supergiant" phase has a hard stop. The universe simply doesn't allow a cool, red giant to get too bright. This discovery forces us to rewrite the biography of how massive stars live, die, and shape the cosmos.

Technical Summary

Here is a detailed technical summary of the paper "A constant upper luminosity limit of cool supergiant stars down to the extremely low metallicity of I Zw 18" by Schootemeijer et al. (2025).

1. Problem Statement

Massive stars are expected to lose their hydrogen-rich outer layers via stellar winds before ending their lives. Standard evolutionary models predict that in low-metallicity ($Z$) environments, radiation-driven winds are weaker. Consequently, massive stars in low-$Z$ galaxies should retain more of their hydrogen envelopes, evolve into brighter cool supergiants (cool SGs; $T_{\text{eff}} < 7$ kK), and potentially form more massive black holes.

However, previous studies in the metallicity range $0.2 \lesssim Z/Z_{\odot} \lesssim 1.5$found that the upper luminosity limit ($L_{\text{max}}$) of cool SGs is constant at$\log(L/L_{\odot}) \approx 5.6$, regardless of metallicity. This contradicts the expectation that lower metallicity leads to brighter cool SGs. The authors aim to test if this "Humphreys-Davidson limit" holds true at **extremely low metallicities** ($Z/Z_{\odot} \approx 1/40$), specifically in the dwarf galaxy **I Zw 18**, and to determine the evolutionary fate of stars born with$M_{\text{init}} \gtrsim 30 M_{\odot}$ in such environments.

2. Methodology

Data Collection:

• Target Galaxies: The study analyzed four dwarf galaxies using archival Hubble Space Telescope (HST) data (NGC 300, NGC 5253, NGC 4395) and new James Webb Space Telescope (JWST) infrared data for I Zw 18.
• Comparison Sample: The sample was augmented with literature data for the Small Magellanic Cloud (SMC), Large Magellanic Cloud (LMC), and M 31, covering a metallicity range of $1/40 \lesssim Z/Z_{\odot} \lesssim 1.5$.
• Photometry: Used filters such as F115W/F200W (JWST) and optical HST filters (F555W/F814W).

Source Classification (Cool SG Identification):

• Neural Network (NN): For galaxies with well-populated Red Supergiant (RSG) branches, the authors trained a Neural Network on Small Magellanic Cloud (SMC) data (using 2MASS and SkyMapper data) to identify cool SGs in color-magnitude diagrams (CMDs). The NN was applied to optical CMDs of NGC 300, NGC 5253, and NGC 4395.
• Evolutionary Tracks: For I Zw 18, where the RSG branch is poorly defined due to low metallicity, the authors used MIST evolutionary tracks for $8 M_{\odot}$stars at$[Fe/H] = -1.7$to select evolved sources more massive than$8 M_{\odot}$.
• Exclusion: Foreground sources were removed using Gaia DR3 astrometry. Sources with $T_{\text{eff}} > 7$ kK were excluded from the "cool SG" count.

Luminosity Calculation:

• Effective temperatures and bolometric corrections (BC) were derived using MIST model grids.
• Luminosities were calculated using the formula:
  $$\log(L/L_{\odot}) = -0.4 (M_{\text{abs},0} - M_{\text{bol},\odot} + \text{BC}(\text{color}))$$
• The method was validated against existing SMC luminosity catalogs (Davies et al. 2018), showing high accuracy (standard deviation of 0.03–0.05 dex).

Theoretical Comparison:

• Observed luminosity distributions were compared against synthetic populations from BoOST (Bonn Optimized Stellar Tracks) models, which include rotation and metallicity-dependent mass loss.
• Simulations were weighted to match the observed number of stars in the range $5.0 \leq \log(L/L_{\odot}) \leq 5.4$.

3. Key Results

Constant Upper Luminosity Limit:

• Across all studied galaxies, including the extremely metal-poor I Zw 18 ($Z \approx 1/40 Z_{\odot}$), no cool SGs were found with $\log(L/L_{\odot}) > 5.6$.
• The third brightest cool SG in each galaxy has an average luminosity of $\log(L/L_{\odot}) = 5.46 \pm 0.06$, showing no correlation with metallicity.
• This confirms the limit is constant down to $Z \approx 1/40 Z_{\odot}$.

Discrepancy with Models:

• BoOST models predict a significant population of bright cool SGs ($\log(L/L_{\odot}) > 5.6$) in low-metallicity environments because they assume metallicity-dependent mass loss (weaker winds at low $Z$).
• The observed absence of these stars yields extremely low Poisson probabilities ($p < 10^{-7}$) for all galaxies, indicating the discrepancy is not due to chance.
• In I Zw 18, the observed number of cool SGs in the range $4.6 < \log(L/L_{\odot}) < 5.0$ is also lower than predicted, suggesting potential evolutionary effects or lifetime variations.

Nature of Missing Stars:

• Not Warm SGs: Analysis of the CMD of I Zw 18 shows the region for "warm SGs" ($7 < T_{\text{eff}} < 20$kK,$\log L > 5.6$) is also empty.
• Presence of WR Stars: Spectroscopic evidence indicates the presence of Wolf-Rayet (WR) stars in I Zw 18 and other target galaxies.
• Conclusion: Stars born with $M_{\text{init}} \gtrsim 30 M_{\odot}$ in low-metallicity environments do not appear as cool or warm supergiants; instead, they evolve into hot, helium-rich stars (likely WR stars).

4. Astrophysical Implications

Metallicity-Independent Mass Loss:

• The most likely mechanism explaining the constant $L_{\text{max}}$ is metallicity-independent mass loss during the late evolutionary phases (e.g., LBV phase or cool SG phase).
• This suggests that mechanisms other than radiation-driven winds (such as pulsational instability or binary interaction) strip the hydrogen envelopes of massive stars even at extremely low metallicities.

Black Hole Mass Limits:

• If massive stars ($>30 M_{\odot}$) lose their hydrogen envelopes before core collapse regardless of metallicity, the resulting black hole (BH) masses will be limited. This challenges models predicting very massive BHs from low-metallicity progenitors via direct collapse without mass loss.

Hard Ionizing Radiation and Nitrogen Enrichment:

• Ionizing Radiation: The authors propose a scenario where single stars undergo self-stripping to become hot helium-rich stars. Below SMC metallicity, a "window of opportunity" opens where these stars have weak winds (insufficient to absorb photons) but are hot enough to emit hard ionizing radiation (He II, C IV). This could explain the strong ionizing signatures observed in high-redshift galaxies.
• Nitrogen Abundance: The loss of CNO-processed hydrogen layers from stars with $30 \lesssim M_{\text{init}}/M_{\odot} \lesssim 100$ could contribute significantly to the high nitrogen abundances observed in high-redshift, low-metallicity galaxies.

5. Significance

This study extends the known metallicity range of the cool supergiant luminosity limit by a factor of 5 (from $Z \approx 1/5 Z_{\odot}$ to $Z \approx 1/40 Z_{\odot}$). It provides robust observational evidence that the evolutionary path of massive stars ($>30 M_{\odot}$) is fundamentally different from standard single-star wind models at low metallicity. The findings necessitate a revision of stellar population synthesis models (like BPASS and Starburst99) regarding the production of ionizing radiation and the chemical enrichment of the early Universe, suggesting that single stars can produce hard ionizing photons without requiring chemically homogeneous evolution (CHE) or extreme binary interactions.