Physical Input for the Determination of Stellar Abundances

Abstract

ABSTRACT.The fundamental parameters and the physical modelling and data used in the analysis of high-resolution high S/N spectra of stars are discussed. Particular emphasis is led on recent developments in these respects and the importance of further improvements is stressed. INTRODUCTIONFor weak stellar spectral line the equivalent width is approximately proportional to the chemical abundance of the element in question and the oscillator strength. The factor of proportionality is, however, dependent on stellar parameters, atmospheric structure and other data and processes that directly affect the line formation. The equivalent widths can today, in high S/N spectroscopy at sufficient resolution, be measured with an accuracy of 10% or better. The f values can in many cases be determined with an even higher accuracy (cf. Huber's review in the present volume). Thus, one could hope that abundances could be estimeted with a corresponding accuracy. This is, in general, not possible; typical external errors in abundances are in good cases about 30%, and often greater than that. The reason for this is our lack of precise knowledge about the fundamental parameters, atmospheric structures and line formation. A clear-cut separation between the errors in chemical abundances caused by uncertainties in these different respects is rather artificial. In particular, the approximation of the atmospheric structures by a grid of standard model atmospheres also affects the choice of fundamental parameters (if made spectroscopically or photometrically * Present address: