🔭 astrophysics

X-Shooting ULLYSES: Massive stars at low metallicity XIV. Properties of SMC late-O and B supergiants reveal the metallicity dependence of winds in the Magellanic Clouds

This study analyzes 20 late-O and B supergiants in the Small Magellanic Cloud to confirm the metallicity dependence of stellar wind momentum, identify a discrepancy between evolutionary and spectroscopic masses, and conclude that binary interactions or luminous blue variable eruptions are required to explain the formation of Wolf-Rayet stars in low-metallicity environments since standard mass-loss rates are insufficient.

Original authors: T. Alkousa, P. A. Crowther, J. M. Bestenlehner, H. Sana, F. Tramper, J. S. Vink, F. Najarro, A. A. C. Sander, M. Bernini-Peron, L. Oskinova, J. Th. van Loon, R. Kuiper, The XShootU collaboration

Published 2026-02-25

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: T. Alkousa, P. A. Crowther, J. M. Bestenlehner, H. Sana, F. Tramper, J. S. Vink, F. Najarro, A. A. C. Sander, M. Bernini-Peron, L. Oskinova, J. Th. van Loon, R. Kuiper, The XShootU collaboration

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 massive stars as giant, fiery furnaces in the sky. They are so bright and hot that they don't just sit there; they constantly blow a "wind" of particles away from themselves. This isn't a gentle breeze like on Earth; it's a hurricane of charged particles driven by the star's own intense light.

This paper is like a detailed weather report for 20 of these massive stars, but with a twist: these stars live in the Small Magellanic Cloud (SMC), a small galaxy neighbor to our own. The key difference? The SMC is "metal-poor." In astronomy, "metals" are just a fancy word for elements heavier than hydrogen and helium (like carbon, oxygen, and iron). Think of the SMC as a kitchen with very few spices, while our Milky Way is a kitchen packed with a full spice rack.

Here is the story of what the astronomers found, broken down simply:

1. The "Spice" Problem (Metallicity)

The scientists wanted to know: Does the amount of "spices" (metals) in a star change how hard it blows its wind?

The Theory: Scientists have long believed that metals act like "sails" for the star's wind. The star's light hits these metal atoms, pushes them, and that push drags the rest of the wind along. So, fewer metals (like in the SMC) should mean weaker sails and a weaker wind.
The Experiment: The team used powerful telescopes (like the Hubble Space Telescope and the Very Large Telescope) to take high-definition "photos" and "videos" (spectra) of these stars. They analyzed the light to measure how fast the wind was blowing and how much mass the star was losing.

2. The Big Discovery: The Wind is Weaker, but the Speed is the Same

The team confirmed the theory, but with a specific nuance:

The Rate: The stars in the metal-poor SMC are indeed losing mass much slower than similar stars in our metal-rich Milky Way. It's like a sprinkler with a clogged nozzle; less water (mass) is coming out.
The Speed: Surprisingly, the speed of the wind didn't change much. Even though there was less "spray," the water that did come out was just as fast.

They created a new "recipe" (a mathematical formula) to predict how strong a star's wind will be based on its brightness and how "spicy" (metal-rich) it is. This recipe is now a better tool for predicting how these stars will evolve.

3. The Mystery of the "Jump" (The Bi-Stability Jump)

There was a famous theory that predicted a "cliff" in the wind behavior. The theory said that when a star cools down to a certain temperature (around 21,000 to 25,000 degrees), its wind should suddenly get much stronger, like a car hitting a speed bump and accelerating.

The Reality Check: The astronomers looked for this "speed bump" in their data. It wasn't there. The wind didn't suddenly get stronger as the stars cooled down. The wind just flowed steadily. This suggests that the old theories about this specific "jump" might need to be rewritten.

4. The "Mass Discrepancy" (The Weight Problem)

The team also weighed the stars in two different ways:

Spectroscopic Mass: Weighing them by looking at how their light bends (like weighing a car by how much it squishes a spring).
Evolutionary Mass: Weighing them by guessing how heavy they must have been to get to their current age and size (like guessing a person's weight by looking at their growth chart).

The Result: For 40% of the stars, the "growth chart" said they were much heavier than the "spring" measurement. It's like a teenager who looks like they should weigh 200 pounds, but when you put them on a scale, they only weigh 150. This suggests our models of how stars grow and age might be slightly off, or that these stars are losing mass faster than we thought in the past.

5. The Great Escape: How Do Stars Become Wolf-Rayet Stars?

This is the most dramatic part of the story.

The Goal: Some massive stars are supposed to shed their entire outer skin (their hydrogen layers) to become Wolf-Rayet stars—hot, naked cores that are the "naked" version of a star.
The Problem: The winds these stars are blowing right now are too weak to strip off that skin. It's like trying to peel an orange with a feather; you just can't do it in time before the star dies.
The Solution: Since the SMC has many Wolf-Rayet stars, something else must be happening. The authors suggest that these stars must have had violent eruptions (like a massive volcanic explosion) or binary interactions (where a partner star steals their skin) to get rid of the outer layers. The steady wind isn't enough; you need a "sledgehammer" to get the job done.

Summary

In short, this paper tells us that stars in metal-poor galaxies blow weaker winds, but they don't blow them any faster. We also learned that the "magic speed bump" in wind theory doesn't exist, and that the heavy stars we see today probably needed a dramatic, violent event to become the naked stars we see in the future.

It's a bit like realizing that a car in a dusty, low-oxygen environment runs on less fuel, but the engine doesn't rev any differently—and if you want to strip the car down to its frame, you can't just drive it; you need to crash it or take it apart piece by piece.

1. Problem and Context

Massive stars ( $M > 8M_\odot$ ) lose mass through radiation-driven winds, a process governed by line-driving mechanisms that depend heavily on the metallicity ( $Z$ ) of the stellar atmosphere. While theoretical models predict a specific dependence of wind properties (mass-loss rates $\dot{M}$ and terminal velocities $v_\infty$ ) on metallicity, there is a persistent discrepancy between theoretical predictions and empirical observations, particularly in low-metallicity environments like the Small Magellanic Cloud (SMC).

Key issues addressed in this study include:

The Bi-stability Jump: Theoretical models predict a sudden increase in mass-loss rates and a decrease in wind velocity for stars with effective temperatures ( $T_{\text{eff}}$ ) between 21–25 kK due to changes in ionization states (the bi-stability jump). Empirical evidence for this jump in low- $Z$ environments remains controversial.
Evolutionary Pathways: Understanding mass loss is crucial for determining stellar evolution, specifically whether massive stars can shed their hydrogen envelopes to become Wolf-Rayet (WR) stars via single-star evolution or if binary interactions/LBV eruptions are required.
Methodological Inconsistencies: Previous studies often combined data from different instruments and used varying radiative transfer codes, leading to large variances in derived parameters.

2. Methodology

The authors performed a homogeneous spectroscopic analysis of 20 late-O and B supergiants in the SMC, utilizing a multi-wavelength dataset to ensure consistency.

Data Sources:
- UV: High-resolution spectra from the HST ULLYSES program (using STIS and COS instruments) and archival FUSE data.
- Optical: Spectra from the XShootU legacy library (VLT/X-shooter) and high-resolution Magellan/MIKE data for rotational broadening.
Modeling Code: The analysis employed CMFGEN, a 1D non-LTE radiative transfer code.
- Grid Construction: A grid of models was generated with varying $T_{\text{eff}}$ , surface gravity ( $\log g$ ), and mass-loss rates ( $\log \dot{M}$ ).
- Clumping: Wind inhomogeneities were modeled using an "optically thin clumping" approximation with an exponential clumping law.
- Velocity Law: A modified $\beta$ -velocity law was used, with terminal velocities ( $v_\infty$ ) initially constrained by an empirical recipe dependent on $T_{\text{eff}}$ and $Z$ .
Fitting Procedure:
- Stellar Parameters: $T_{\text{eff}}$ was determined via ionization equilibrium (He I/He II, Si III/IV, Mg II); $\log g$ via Balmer line wings; and luminosity ( $L_{\text{bol}}$ ) via SED fitting.
- Wind Parameters: $\dot{M}$ and clumping factors ( $f_{\text{vol},\infty}$ ) were constrained by fitting H $\alpha$ profiles and UV P Cygni resonance lines (e.g., C IV, Si IV).
- Evolutionary Masses: An updated Bayesian inference method (similar to BONNSAI) was applied to Brott et al. (2011) rotating single-star evolutionary tracks to derive evolutionary masses ( $M_{\text{evo}}$ ) and ages.
Comparison: Results were directly compared with a previous study of LMC supergiants (Paper XIII) using identical methods, as well as with literature values and numerical mass-loss recipes (Vink & Sander 2021; Björklund et al. 2023; Krtiˇcka et al. 2024).

3. Key Contributions

Homogeneous Analysis: This is the first study to apply the exact same modeling techniques (CMFGEN, ULLYSES/XShootU data) to a large sample of SMC supergiants, allowing for a direct, bias-free comparison with LMC counterparts.
New Metallicity-Dependent Recipe: The authors derived a new empirical formula for wind momentum ( $D_{\text{mom}}$ ) that explicitly accounts for metallicity.
Resolution of Mass-Discrepancy: A systematic comparison of spectroscopic vs. evolutionary masses was performed, highlighting the prevalence of the "mass discrepancy" in low- $Z$ environments.
Bi-stability Jump Investigation: A rigorous test of the bi-stability jump hypothesis in the SMC using a sample spanning the critical temperature range.

4. Key Results

A. Stellar and Wind Properties

Sample: 20 stars with spectral types O9 to B8 and luminosities $5.4 \leq \log(L_{\text{bol}}/L_\odot) \leq 6.1$ .
Mass-Loss Rates: Derived $\dot{M}$ values range from $\approx 10^{-8.2}$ to $10^{-5.9} M_\odot \text{yr}^{-1}$ .
Metallicity Dependence:
- Wind Momentum: A clear dependence on metallicity was found. The derived recipe for wind momentum is:
  $\log D_{\text{mom}} = (1.64 - 0.75 \log \frac{Z}{Z_\odot}) \log \frac{L_{\text{bol}}}{10^6 L_\odot} + 1.38 \log \frac{Z}{Z_\odot} + 29.17$
  (Applicable for $5.4 \leq \log L_{\text{bol}}/L_\odot \leq 6.1$ and $14 \leq T_{\text{eff}}/\text{kK} \leq 32$ ).
- Mass-Loss vs. Velocity: The metallicity dependence of wind momentum is driven almost entirely by the mass-loss rate ( $\dot{M}$ ). The terminal wind velocity ( $v_\infty$ ) shows no significant metallicity dependence when compared to LMC counterparts.
Comparison with Recipes:
- For $T_{\text{eff}} \gtrsim 22$ kK, numerical recipes generally match observations.
- For $T_{\text{eff}} \lesssim 22$ kK, the Vink & Sander (2021) recipe overpredicts $\dot{M}$ by $\approx 1.5$ dex. The Krtiˇcka et al. (2024) recipe performs better but still deviates.
- The authors' derived $\dot{M}$ values are generally lower than previous literature estimates (e.g., Bernini-Peron et al. 2024).

B. The Bi-stability Jump

No Evidence Found: The study found no discontinuity in mass-loss rates or the ratio $v_\infty/v_{\text{esc}}$ in the 21–25 kK range.
Transformed Mass-Loss Rate: When analyzing the transformed mass-loss rate ( $\dot{M}_t$ , which corrects for luminosity and clumping), the rate appears to increase with temperature, contrary to the predicted drop associated with the bi-stability jump. This trend is attributed to the authors' derived low clumping factors ( $f_{\text{vol},\infty} < 0.2$ ).

C. Evolutionary Masses and Chemical Composition

Mass Discrepancy: In 40% of the sample, the evolutionary mass ( $M_{\text{evo}}$ ) is systematically higher than the spectroscopic mass ( $M_{\text{spec}}$ ).
Nitrogen Enrichment: All stars show significant nitrogen enrichment. However, the observed rotational velocities ( $v \sin i$ ) are generally too low ( $\approx 30-100$ km/s) to explain this enrichment via rotational mixing alone, suggesting binary interaction or undetected companions are the primary drivers.
WR Formation: The derived mass-loss rates for blue supergiants are far too low to strip the hydrogen envelope within the star's remaining lifetime. This implies that single-star evolution cannot explain the observed population of Wolf-Rayet stars in the SMC.

5. Significance and Conclusions

Evolutionary Implications: The inability of radiation-driven winds to strip the envelopes of SMC supergiants suggests that the formation of classical Wolf-Rayet stars in low-metallicity environments requires Luminous Blue Variable (LBV) eruptions or binary interactions (mass transfer/common envelope).
Theoretical Validation: The lack of a bi-stability jump in the data challenges current theoretical prescriptions for line-driven winds in the B-supergiant regime, suggesting that the physics of wind clumping or ionization balance in these models needs revision.
Future Work: The authors plan to analyze Galactic blue supergiants using the same methodology to derive a universal mass-loss recipe free from the inconsistencies of mixed datasets and codes.

In summary, this paper provides a robust, homogeneous dataset that refines our understanding of wind physics in low-metallicity environments, confirming a strong metallicity dependence in mass-loss rates but refuting the existence of the bi-stability jump in the observed SMC population.