Improoving stellar libraries for stellar population synthesis

Abstract

Libraries of stellar spectra are fundamental tools for the study of stellarpopulations. Despite the significant improvement of recent libraries, there is still alot to be done in order to have libraries that are able to generate models that canreproduce the observed stellar population spectra of real galaxies. Theoretical stellarlibraries are particularly interesting for stellar population models, since it is impossibleto reproduce the spectrum of galaxies with chemical histories that differ from our solarneighborhood using only the spectra of observed stars. On the other hand, theoreticalstellar spectra are still very imprecise. One of the major difficulties of theoreticalspectra is the atomic and molecular line list used to generate it. Very few of the atomicand molecular parameters of these lists were measured in laboratory, and because ofthat, most of them are very uncertain. We propose in this work an innovative methodto calibrate atomic lines lists - ALiCCE (Atomic Lines Calibration using Cross-Entropyalgorithm). This method allows the the calibration to be done in a robust and efficientway. The improved line list can then be used to generate theoretical stellar modelswith unprecedented accuracy. (AU)