🔭 astrophysics

Charting circumstellar chemistry of carbon-rich asymptotic giant branch stars. III. SiO and SiS abundances

This study utilizes ALMA and single-dish observations combined with non-LTE radiative transfer modeling to derive well-constrained SiO and SiS abundances for five carbon-rich AGB stars, revealing that SiO abundances are significantly higher than in the archetype IRC+10216 and highlighting the need to revise SiS photodissociation rates in chemical models.

Original authors: R. Unnikrishnan, E. De Beck, L. -Å. Nyman, H. Olofsson, W. H. T. Vlemmings, M. Maercker, M. Van de Sande, T. J. Millar, T. Danilovich, M. Andriantsaralaza, S. B. Charnley, M. G. Rawlings

Published 2026-05-18

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: R. Unnikrishnan, E. De Beck, L. -Å. Nyman, H. Olofsson, W. H. T. Vlemmings, M. Maercker, M. Van de Sande, T. J. Millar, T. Danilovich, M. Andriantsaralaza, S. B. Charnley, M. G. Rawlings

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 dying star as a giant, blooming flower made of gas and dust. As this star (called an AGB star) runs out of fuel, it puffs up and blows a massive, expanding shell of material into space. This shell is like a cosmic greenhouse where new molecules are constantly being cooked up.

This paper is the third part of a series where astronomers act like "cosmic chefs" trying to figure out exactly what ingredients are in this shell and how much of each there is. Specifically, they are looking for two key ingredients: Silicon Monoxide (SiO) and Silicon Sulfide (SiS).

Here is a breakdown of what they did and what they found, using simple analogies:

The Recipe Book vs. The Actual Dish

For a long time, scientists thought they understood the chemistry of these dying stars because they studied one famous "archetype" star called IRC+10 216. It was like studying one specific apple to understand the entire orchard.

In this study, the team looked at five different carbon-rich stars to see if they all taste the same or if they are unique recipes. They used two types of "tasting spoons":

Single-dish telescopes: These give a blurry, overall taste of the whole star shell.
ALMA (The Atacama Large Millimeter/submillimeter Array): This is a super-sharp camera that can see the fine details and structure of the shell, like seeing the individual grains of sugar rather than just the whole cake.

The Main Findings

1. The "Silicon Sulfide" (SiS) Surprise
The team found that the amount of SiS is surprisingly consistent across all the stars they studied. It's like if you went to five different bakeries and found they all used exactly the same amount of vanilla in their cakes. The peak amount of SiS was very similar for every star.

2. The "Silicon Monoxide" (SiO) Difference
However, the SiO was a different story. For four of the stars, there was about five times more SiO than there was in the famous archetype star (IRC+10 216). This suggests that the famous star isn't the perfect "average" for all carbon stars; it's actually a bit of an outlier with less SiO than its neighbors.

3. The "Fading Light" Analogy
The molecules don't stay in the shell forever. As they drift away from the star, they get hit by ultraviolet light from the star and the galaxy, which breaks them apart (photodissociation).

The team measured how far out the molecules survive before fading away.
They found that in stars with denser, thicker gas shells, the molecules survive further out. Think of it like a thick fog: if the air is dense, the light has a harder time penetrating, so the molecules can survive further away from the source before being destroyed.

The Modeling Struggle: The "One-Dimensional" Problem

To figure out the exact amounts, the astronomers used computer models. Imagine trying to describe a 3D object (like a basketball) using only a 2D shadow. That's what they were doing: using a 1D model to describe a complex, 3D star shell.

The Glitch: For the famous star (IRC+10 216), the models hit a wall. They could fit the data from the sharp ALMA telescope or the data from the single-dish telescopes, but not both at the same time.
The Analogy: It's like trying to tune a radio. If you tune it to get a clear picture of the station's logo (the detailed ALMA data), the music sounds staticky (the single-dish data). If you tune for clear music, the logo blurs.
The Conclusion: This mismatch suggests that the simple 1D model is too basic. The real star shell likely has bumps, spirals, or clumps that a simple "smooth shell" model can't capture.

The Chemical Kitchen

The team also compared their observations with chemical recipes (computer simulations of how molecules form and break).

SiO: The chemical recipes predicted the SiO amounts very well. The models and the observations agreed.
SiS: The chemical recipes were wrong. They predicted that SiS should survive much further out than it actually does. This tells the scientists that the "destruction rate" (how fast SiS gets broken by light) used in the recipes is too slow. They need to update the recipe to make SiS break down faster.

The Bottom Line

This study confirms that while we have a good general idea of how these dying stars work, the famous "archetype" star isn't the only rulebook.

SiS is a reliable ingredient found in similar amounts everywhere.
SiO varies wildly, with some stars having much more than others.
Our models need work: We need better 3D maps and updated chemical recipes to fully understand how these stars cook up the elements that will eventually become new stars and planets.

In short, the universe is more diverse than our single "famous example" suggested, and we need sharper tools and better recipes to understand the full menu of cosmic chemistry.

Technical Summary: Charting circumstellar chemistry of carbon-rich asymptotic giant branch stars III. SiO and SiS abundances

Problem and Context
The circumstellar envelopes (CSEs) of carbon-rich (C-type) asymptotic giant branch (AGB) stars are sites of complex molecular chemistry. While the archetypal carbon star IRC+10 216 has been extensively studied, current understanding of molecular abundances in the broader population of carbon stars largely relies on single-dish (SD) spectra or limited interferometric mapping. These methods often fail to simultaneously constrain spatial morphology and excitation conditions across a wide range of radii. Specifically, the abundances of silicon-bearing refractory species, SiO and SiS, remain uncertain. Unlike CS, which forms readily in C-rich environments, SiO and SiS formation may depend on factors such as freeze-out onto dust or shock chemistry. This study aims to derive robust circumstellar abundance profiles for SiO, SiS, and their major isotopologues ( $^{29}$ SiO, $^{30}$ SiO, $^{29}$ SiS, $^{30}$ SiS, and Si $^{34}$ S) for a sample of five carbon stars, testing whether IRC+10 216 is representative of the population and evaluating chemical modeling predictions.

Methodology
The authors analyzed a sample of five high mass-loss rate (MLR) C-type AGB stars: IRAS 15194−5115, IRAS 15082−4808, IRAS 07454−7112, AFGL 3068, and IRC+10 216. The study employed a comprehensive dataset combining:

Spatially Resolved Data: ALMA band 3 spectral surveys (85–116 GHz) and, for IRC+10 216, combined ALMA/IRAM 30m on-the-fly maps. Multi-aperture spectra were extracted from convolved cubes to trace line fluxes across different spatial scales.
Single-Dish Data: Spectra from APEX, Herschel/HIFI, IRAM 30m, and Yebes 40m telescopes to cover a broad range of excitation conditions (low- to high-J transitions).

The authors utilized detailed 1D non-local thermodynamic equilibrium (non-LTE) radiative transfer (RT) modeling using an accelerated lambda iteration (ALI) code. The models incorporated:

Physical Structure: Spherically symmetric, smooth CSE models with constant MLR and isotropic expansion, using input parameters (stellar luminosity, dust properties, gas density) derived from previous dust and CO modeling (Paper II).
Molecular Data: Radiative and collisional excitation rates for SiO (including $v=0, 1$ ) and SiS (including $v=0, 1, 2$ and ro-vibrational transitions).
Abundance Profiles: Gaussian profiles defined by a peak abundance ( $f_0$ ) and an e-folding radius ( $R_e$ ) as free parameters.
Constraints: The models were constrained by minimizing $\chi^2$ between modeled and observed integrated line intensities, utilizing both the spatial constraints from ALMA multi-aperture spectra and the excitation constraints from SD lines.

Key Results

Abundance Profiles: Robust abundance profiles were derived for all sources except AFGL 3068, which lacked spatially resolved data, leading to unconstrained $f_0$ and $R_e$ values.
- SiO: Peak abundances for the four constrained sources ranged from $1.0 \times 10^{-7}$ to $7.1 \times 10^{-7}$ . Notably, the SiO peak abundances for IRAS 15194−5115, IRAS 15082−4808, and IRAS 07454−7112 were found to be a factor of $\sim$ 5 larger than that of IRC+10 216 ( $2.2 \times 10^{-7}$ ).
- SiS: Peak abundances were more uniform across the sample ( $2.0 \times 10^{-6}$ to $4.7 \times 10^{-6}$ ), consistently higher than SiO abundances by a factor of 2–15.
- Radial Extents: The e-folding radii ( $R_e$ ) for SiO ranged from $1.3 \times 10^{16}$ to $7.0 \times 10^{16}$ cm, and for SiS from $6.0 \times 10^{15}$ to $1.0 \times 10^{17}$ cm. $R_e$ was found to increase with gas density for both species.
Isotopic Ratios: Derived isotopic ratios ( $^{28}$ Si/ $^{29}$ Si, $^{28}$ Si/ $^{30}$ Si, $^{29}$ Si/ $^{30}$ Si, and $^{32}$ S/ $^{34}$ S) were consistent with solar values for all sources, indicating no significant nucleosynthetic alteration of these elements in the AGB phase, with the exception of possible minor s-process enhancements.
Modeling Challenges:
- IRC+10 216 SiO: The authors could not simultaneously fit the low-J ALMA lines and high-J SD lines for SiO in IRC+10 216. Models fitting the ALMA data underestimated the SD lines, and vice versa. Attempts to resolve this using modified inner abundance profiles (step functions) or varying stellar/dust parameters were unsuccessful, highlighting limitations in 1D RT modeling with simplified stellar assumptions.
- SiS J=6-5: Similar to SiO, models could not simultaneously fit the SiS $J=5-4$ and $J=6-5$ lines. The authors attribute this to potential IR pumping overlaps with other molecules (e.g., C $_2$ H $_2$ , HCN) affecting the $v=1$ levels, which are not included in their standard RT models. Consequently, the $J=6-5$ lines were excluded from the final abundance constraints.
Comparison with Chemical Models:
- SiO: Chemical models reproduced the derived SiO abundance profiles and radial extents well.
- SiS: Chemical models consistently over-predicted the $R_e$ values for SiS, suggesting that the current photodissociation rate for SiS used in chemical models is underestimated.

Significance and Conclusions
The paper demonstrates that robust constraints on circumstellar molecular abundances require spatially resolved observations across a broad range of excitation conditions. The study confirms that IRC+10 216 is not fully representative of the C-type AGB population, as other sources exhibit significantly higher SiO abundances. The findings highlight the critical role of infrared pumping in the excitation of SiO and SiS lines. Furthermore, the discrepancy between observed SiS radial extents and chemical model predictions suggests that the photodissociation rate of SiS needs revision. The authors conclude that while 1D RT modeling is effective, it has inherent limitations when applied to complex, non-smooth circumstellar environments, and that future comprehensive overviews of AGB chemistry require larger samples, including lower mass-loss rate sources, combined with updated chemical models.

The Recipe Book vs. The Actual Dish

The Main Findings

The Modeling Struggle: The "One-Dimensional" Problem

The Chemical Kitchen

The Bottom Line

More like this