Dust Mass Loss Rates in Evolved Stars

Updated 5 August 2025

Dust mass loss rates are a measure of how evolved stars and disks eject dust, playing a key role in ISM enrichment and galactic chemical evolution.
Advanced SED fitting and radiative transfer modeling, which incorporate gas-to-dust ratios and grain properties, are used to derive these rates.
Variations in stellar type, metallicity, and pulsation influence dust mass loss rates, affecting stellar evolution outcomes and supernova progenitor profiles.

Dust mass loss rates quantify the rate at which astrophysical systems—most notably evolved stars and protoplanetary disks—expel solid (dust) particles into their surroundings. These rates are central to understanding the lifecycle of dust in the interstellar medium, the feedback of metal-enriched material into galaxies, and the chemical evolution and observational signatures of late stellar evolution. The dust mass loss rate, typically expressed in units of solar masses per year (M $_\odot$ yr $^{-1}$ ), is often distinguished from the total (gas + dust) mass loss rate; conversion between the two requires assumptions about the local gas-to-dust ratio, which can vary by environment and stellar type.

1. Methods for Measuring and Modeling Dust Mass Loss Rates

Radiative transfer modeling of the circumstellar environment is the principal methodology for deriving dust mass loss rates. This approach combines multi-wavelength photometry—optical to far-infrared—with mid-infrared spectroscopy to constrain stellar and dust shell parameters, such as luminosity ( $L$ ), effective temperature ( $T_\mathrm{eff}$ ), optical depth ( $\tau_\lambda$ ), composition, and grain size distribution.

Key tools and procedures include:

SED fitting codes: 2Dust, GRAMS, DUSTY, and MCMax are commonly applied to model broadband spectral energy distributions (SEDs) and match spectral features specific to dust (e.g., silicate or SiC emission).
Calibration against indices: Empirical relations link mid-IR colors or emission contrasts to mass loss. For example, in Spitzer IRS observations, the [7]–[15] color and Dust Emission Contrast (DEC) are robust quantifiers of circumstellar dust content. Calibrated equations, such as

$\log \dot{M}_\mathrm{dust} (M_\odot\,\mathrm{yr}^{-1}) = 1.759([7]-[15])-8.664,$

connect IR photometric colors to dust mass loss rates (Sloan et al., 2010).

Gas-to-dust ratio: Converting from dust mass loss to total mass loss requires an adopted gas-to-dust ratio ( $\Psi$ ), typically assumed to be 200 for Galactic AGB stars and varying up to 700 in extreme mass-losing carbon stars in the Magellanic Clouds (Nanni et al., 2019).
Dynamical and grain growth models: Detailed grain growth calculations and dynamical wind models, such as those implemented in DARWIN, give access to time-dependent grain size, outflow density, and velocity—integral factors in determining mass-loss rates, especially for carbon-rich environments (Mattsson et al., 2011, Bladh et al., 2019).

2. Physical Dependencies: Stellar Type, Luminosity, and Pulsation

Dust mass loss rates are not uniform but depend on a variety of stellar parameters and evolutionary states:

AGB Stars: The mass loss rate correlates strongly with luminosity and, for many formulations, the luminosity-to-mass ratio ( $L/M$ ), with a trend $\dot{M}_\mathrm{dust}\propto L^{2.3-3.0}$ or $\dot{M}\propto(L_*/M_*)^{2.5}$ for oxygen-rich M-type AGBs (Bladh et al., 2019). For C-rich AGBs, bolometric luminosity, pulsation period, and color indices can all serve as proxies for dust loss (Sloan et al., 2010, Riebel et al., 2012). AGB stars with longer periods and larger luminosities display higher dust production, especially among "extreme" sources (Sloan et al., 2010, Nanni et al., 2019, Wen et al., 15 Sep 2024).
Red Supergiants (RSGs): Mass loss rates in RSGs follow an empirical trend of increasing steeply beyond a luminosity threshold of $\log(L/L_\odot) \approx 4.4$ (Wen et al., 8 Jan 2024, Wen et al., 15 Sep 2024). Below this point, increases with $L$ are modest, but above it, mass-loss rates may escalate by orders of magnitude, attributed to enhanced efficiency of dust-driven winds in the more luminous regime.
Role of Dust Chemistry and Grain Size: The nature of the dust (alumina-rich versus silicate-rich in O-rich stars, amorphous carbon with SiC inclusions in C-rich stars) is set by atmospheric chemistry and influences the mid-IR features used to derive $\dot{M}_\mathrm{dust}$ (Sloan et al., 2010, Srinivasan et al., 2010, Sargent et al., 2014). Grain size is also pivotal: in critical cases near the threshold for wind driving, larger grains (comparable to the IR wavelength) significantly enhance radiative acceleration and thus $\dot{M}_\mathrm{dust}$ (Mattsson et al., 2011).
Pulsation and Variability: For both AGB and RSG stars, strong pulsation (long periods, high amplitudes) correlates with increased dust mass loss; for AGBs, DPR sees a sharp uptick at $\sim300$ day periods and I-band amplitudes above 0.5 mag (Wen et al., 15 Sep 2024).

3. Metallicity Effects

The dependence of dust mass loss rates on metallicity is nuanced:

Oxygen-rich Evolved Stars: Wind expansion velocity decreases with lower metallicity due to reduced dust (v $_\mathrm{exp}\propto ZL^{0.4}$ ), but empirically, the total mass-loss rate ( $\dot{M}$ ) is nearly independent of metallicity between $0.5-2\,Z_\odot$ (Goldman et al., 2016, Loon, 21 Jul 2025). Dust mass loss becomes less prevalent at lower metallicity primarily because gas-to-dust ratios increase, reducing dust-driven wind efficiency in the less massive or warmer stars (Loon, 21 Jul 2025).
Carbon-rich AGBs: In carbon-rich environments, the metallicity dependence is weaker since carbon is synthesized internally and dredged up. However, in the SMC, observed carbon-dust production rates are over-predicted by models that match the LMC, implying a stronger metallicity dependence in mass loss during the carbon star stage (Schneider et al., 2014).
Population Effects: Integrated for populations, lower metallicity yields lower mean DPR and a larger fraction of highly optically thin stars, although the most extreme stars (luminous, cool, strongly pulsating) can still achieve high dust production (Wen et al., 15 Sep 2024).

4. Observational Proxies and Empirical Relations

Empirical determinations of $\dot{M}_\mathrm{dust}$ rely on specific diagnostics:

Mid-IR Color Relations: A variety of relations link infrared colors directly to the dust mass-loss rate, enabling rapid estimation for large samples. For example:

$\log \dot{M}_\mathrm{dust} = \frac{-18.90}{(K_s-[8.0])+3.37} - 5.93$

for C-rich AGBs in the LMC (Riebel et al., 2012).

Piecewise Luminosity–MLR Relations: For RSGs, piecewise linear relations demarcate a mild slope below $\log(L/L_\odot)\approx4.4$ , with a much steeper increase above this threshold:

$\log\dot{M}_\mathrm{dust} = \begin{cases} m_1 \log (L/L_\odot) + c_1, & \text{if } \log(L/L_\odot) < 4.4 \ m_2 \log (L/L_\odot) + c_2, & \text{if } \log(L/L_\odot) \geq 4.4 \end{cases}$

with empirically derived $m_i$ , $c_i$ (Wen et al., 8 Jan 2024, Wen et al., 15 Sep 2024).

SED Fitting and Population Synthesis: The SED-fitting approach applied uniformly to large samples (e.g., $\sim40,000$ evolved stars in the Magellanic Clouds) provides consistent catalogs of DPR and highlights the dominance of a small fraction of highly dusty stars in the integrated dust return (Wen et al., 15 Sep 2024).

5. Global Dust Production and Interstellar Impact

The integrated dust injection rates from evolved stars are now well-characterized for galactic environments with comprehensive surveys:

Host	Total DPR (M $_\odot$ /yr)	All evolved stars	Carbon-/Oxygen-rich breakdown
LMC	$9.69\times10^{-6}$	RSGs + AGBs	No significant C-/O-rich DPR bias
LMC (RSGs only)	$1.14\times10^{-6}$	RSGs	O-rich RSGs: $97.2\%$ of DPR
SMC	$1.75\times10^{-6}$	RSGs + AGBs	O-rich stars less efficient at low Z
M31 (RSGs only)	$1.1\times10^{-3}$	RSGs	16% C-dust, rest silicates (Wang et al., 2021)
M33 (RSGs only)	$6.0\times10^{-4}$	RSGs	13% C-dust, rest silicates (Wang et al., 2021)

These measurements facilitate comparisons with dust masses observed in the ISM and enable assessments of whether stellar sources alone can account for galactic dust budgets (Schneider et al., 2014, Wen et al., 15 Sep 2024).

6. Challenges and Uncertainties

Determining dust mass loss rates encounters several systematic challenges:

Model Dependencies: Uncertainties in the gas-to-dust ratio, dust composition, and grain size distribution propagate directly into estimated $\dot{M}_\mathrm{dust}$ .
Geometric and Variability Effects: Asymmetric dust shell morphology, episodic versus steady mass loss, and line-of-sight effects can bias SED-based determinations (Loon, 21 Jul 2025, O'Gorman et al., 2014).
Chromospheres and Episodic Events: Dense chromospheres by themselves are insufficient for high-mass loss. Instead, it is the cool extended layers (often pulsation-induced) where effective dust condensation—and thus high rates—can be sustained (Loon, 21 Jul 2025).
Calibration Limits: Infrared excess and maser-based methods require careful calibration against wind velocity, grain properties, and circumstellar optical depth, which are often uncertain in low-metallicity or peculiar stars.

7. Broader Relevance and Implications

Dust mass loss rates impact stellar evolution, galactic chemical enrichment, and star formation.

AGB and RSG contributions determine the replenishment of dust in star-forming regions and consequent ISM properties, with evolving observational support that a minority of the most extreme stars dominate the total dust injection (Wen et al., 15 Sep 2024).
Binary Interaction and Population Effects: On a population scale, binary interaction acts as a stochastic agent producing major mass-loss events, influencing the circumstellar dust environment and ultimately affecting the fate of massive stars prior to supernova (Loon, 21 Jul 2025).
Stellar Evolution and Supernovae: In the context of core-collapse progenitors, sufficiently high mass loss near the tip of the AGB or in the most massive RSGs can lead to envelope stripping, with direct consequences for supernova outcomes and observed transients (Loon, 21 Jul 2025).
Extragalactic Context and Early Universe: Comprehensive mass loss and dust yield prescriptions derived from local group galaxies underpin models of ISM evolution at high redshift, where dust condensation by evolved stars may have dominated dust formation channels.

This synopsis captures the principal mechanisms, measurement strategies, dependencies, and implications of dust mass loss rates as established in recent astrophysical research, with quantitative formulations, observed systematics, and leading methodologies tied directly to empirical and theoretical advances.