[Paper Review] The Cycle of Dust in the Milky Way: Clues from the High-Redshift and Local Universe

E. Dwek, F. Galliano+1 more|ArXiv.org|Feb 27, 2009

Astronomy and Astrophysical Research1 references6 citations

TL;DR

This paper proposes that dust evolution models, integrating stellar evolution and interstellar processing, explain the metallicity-dependent abundance of polycyclic aromatic hydrocarbons (PAHs) in galaxies and the origin of massive dust reservoirs at high redshift. It shows that delayed injection of carbon dust from low-mass AGB stars explains the PAH-metallicity correlation, while high-redshift dust masses require supernovae to produce 0.1–1 M☉ of dust per explosion—far exceeding observed yields unless accretion in molecular clouds enhances grain growth.

ABSTRACT

Models for the evolution of dust are used to show that the observed trend of the abundance of polycyclic aromatic hydrocarbons (PAHs) with metallicity is the result of the delayed injection of carbon dust that formed in low mass asymptotic giant branch (AGB) stars into the interstellar medium. We also use our dust evolution models to examine the origin of dust at redshifts > 6, when only supernovae and their remnants could have been, respectively, their sources of production and destruction. Unless an average supernova (or its progenitor) produces between 0.1 and 1 Msun of dust, alternative sources will need to be invoked to account for the massive amount of dust observed at these redshifts.

Motivation & Objective

• Explain the observed correlation between polycyclic aromatic hydrocarbon (PAH) abundance and metallicity in galaxies.
• Determine the dominant dust production and destruction mechanisms in the early universe, particularly at redshift z > 6.
• Assess whether supernovae alone can account for the massive dust masses observed in high-redshift galaxies like J1148+5251.
• Integrate dust evolution into population synthesis models to self-consistently link stellar and dust emission in spectral energy distributions (SEDs).

Proposed method

• Construct dust evolution models that track the injection of dust from asymptotic giant branch (AGB) stars and Type II supernovae (SNe) into the interstellar medium (ISM), accounting for finite main-sequence lifetimes.
• Use chemical evolution models with delayed injection to simulate the time-dependent release of carbon-rich dust from low-mass stars and silicate dust from SNe.
• Incorporate grain destruction by supernova blast waves and growth via accretion in molecular clouds, using the instantaneous recycling approximation for simplicity in high-redshift analysis.
• Apply SED fitting to 35 nearby galaxies, using free-free, mid-IR (PAH), and far-IR (dust heating) emission to constrain dust and gas components.
• Model the dust-to-gas mass ratio in high-redshift galaxies by varying SN dust yields and ISM grain destruction efficiency, using the IMF-averaged dust yield $\widehat{Y}_d$ as a key parameter.
• Compare theoretical dust mass fractions with observations of J1148+5251 (z = 6.4), requiring Z_d = 0.0067 in a 3×10¹⁰ M☉ ISM.

Experimental results

Research questions

• RQ1 Why does the abundance of polycyclic aromatic hydrocarbons (PAHs) correlate with metallicity in galaxies?
• RQ2 What is the relative contribution of AGB stars versus supernovae to the interstellar dust budget over cosmic time?
• RQ3 Can supernovae alone produce enough dust to account for the observed dust masses in high-redshift galaxies at z > 6?
• RQ4 How does grain destruction in the ISM affect the required dust yield per supernova to explain high-redshift dust reservoirs?
• RQ5 What role does accretion of refractory elements onto pre-existing grains play in growing dust mass in molecular clouds?

Key findings

• The observed correlation between PAH abundance and metallicity arises from the delayed injection of carbon dust from low-mass AGB stars into the ISM, not from direct metallicity dependence of PAH formation.
• Supernovae must produce between 0.1 and 1 M☉ of dust per explosion to account for the dust mass fraction Z_d = 0.0067 in the high-redshift quasar J1148+5251, assuming no grain destruction.
• The largest observed SN dust yield to date (0.054 M☉ in Cas A) is insufficient to explain high-z dust masses unless condensation efficiency is near unity or accretion in molecular clouds enhances grain growth.
• Grain destruction in the ISM significantly increases the required SN dust yield; even with modest destruction (⟨m_ISM⟩ = 100 M☉), the required yield rises to 1–2 M☉ per SN.
• Dust evolution models must be integrated into population synthesis models to self-consistently link stellar and dust emission components in the spectral energy distribution (SED) of galaxies.
• Accretion in molecular clouds is likely essential to grow dust mass beyond SN yields, requiring SNRs to enhance nucleation centers via non-evaporative grain-grain collisions.