Residual Abundances in GALAH DR3: Implications for Nucleosynthesis and Identification of Unique Stellar Populations

TL;DR

This study analyzes elemental abundances in GALAH DR3 stars using a two-process model to understand nucleosynthesis, identifying unique stellar populations and the impact of data quality on abundance residuals.

Contribution

It introduces a detailed residual analysis of stellar abundances with a two- and three-process model, revealing new insights into stellar populations and nucleosynthesis sources.

Findings

Residuals are generally small (<0.07 dex) for well-measured elements.

Correlated residuals suggest common enrichment sources among some elements.

Identification of stars with unique abundance patterns and the impact of data issues.

Abstract

We investigate the [X/Mg] abundances of 16 elements for 82,910 Galactic disk stars from GALAH+ DR3. We fit the median trends of low-Ia and high-Ia populations with a two-process model, which describes stellar abundances in terms of a prompt core-collapse and delayed Type-Ia supernova component. For each sample star, we fit the amplitudes of these two components and compute the residual $\Delta$[X/H] abundances from this two-parameter fit. We find RMS residuals $\lesssim 0.07$ dex for well-measured elements and correlated residuals among some elements (such as Ba, Y, and Zn) that indicate common enrichment sources. From a detailed investigation of stars with large residuals, we infer that roughly $40\%$ of the large deviations are physical and $60\%$ are caused by problematic data such as unflagged binarity, poor wavelength solutions, and poor telluric subtraction. As one example of a population with distinctive abundance patterns, we identify 15 stars that have 0.3-0.6 dex enhancements of Na but normal abundances of other elements from O to Ni and positive average residuals of Cu, Zn, Y, and Ba. We measure the median elemental residuals of 14 open clusters, finding systematic $\sim0.1-0.4$ dex enhancements of O, Ca, K, Y, and Ba and $\sim0.2$ dex depletion of Cu in young clusters. Finally, we present a restricted three-process model where we add an asymptotic giant branch star (AGB) component to better fit Ba and Y. With the addition of the third process, we identify a population of stars, preferentially young, that have much higher AGB enrichment than expected from their SNIa enrichment.