Abundance sensitive points of line profiles in the stellar spectra

TL;DR

This paper introduces a simple simulation-based method to identify spectral regions in stellar spectra that are most sensitive to elemental abundances, emphasizing the importance of weak lines and caution with very weak features.

Contribution

It presents a straightforward approach to locate abundance-sensitive spectral features, highlighting the utility of weak lines and the need for accurate broadening mechanisms.

Findings

Weak spectral features are most sensitive to abundances.

Equivalent widths of weak lines are ideal for abundance analysis.

Strong line wings require accurate broadening information.

Abstract

Many abundance studies are based on spectrum synthesis and $\chi$-squared differences between the synthesized and an observed spectrum. Much of the spectra so compared depend only weakly on the elemental abundances. Logarithmic plots of line depths rather than relative flux make this more apparent. We present simulations that illustrate a simple method for finding regions of the spectrum most sensitive to abundance, and also some caveats for using such information. As expected, we find that weak features are the most sensitive. Equivalent widths of weak lines are ideal features, because of their sensitivity to abundances, and insensitivity to factors that broaden the line profiles. The wings of strong lines can also be useful, but it is essential that the broadening mechanisms be accurately known. The very weakest features, though sensitive to abundance, should be avoided or used with great caution because of uncertainty of continuum placement as well as numerical uncertainties associated with the subtraction of similar numbers.