Abstract of PhD thesis of N. Christlieb

Abstract

This thesis deals with the development of quantitative object selection methods and their application to the digital database of the Hamburg/ESO objective prism survey (HES), which covers the total southern extragalactic sky in the magnitude range 12<BJ<17.5. The aim is a systematic exploitation of the stellar content of the HES.

A flexible, robust algorithm for detection of stellar absorption and emission lines in HES spectra was implemented. Equivalent widths for the strongest stellar absorption lines were derived for all 3437630 spectra (87% of the HES) used in this work. Broad band (U-B, B-V) and narrow band (Strömgren c1) colours were also derived directly from HES spectra, with precisions of sigmaU-B=0.092m; sigmaB-V=0.095m; sigmac1=0.15m. Together with continuum parameters derived by principal component analysis, the above features are used for quantitative object selection.

For selection of extremely metal-poor halo stars, automatic spectral classification was employed. Spectroscopic follow-up observations of 58 stars showed that the selection in the HES has a more than three times higher efficiency then the selection in the the so-called HK survey of Beers et al., the up to now largest survey for metal-poor stars. The effective yield of turnoff stars with [Fe/H]<-2.0 is 80% in the HES, but only 22% in the HK survey on average. This is very remarkable considering the fact that the spectral resolution of the HES (10Å at Ca K) is two times lower than in the HK survey (5Å). In spectroscopic follow-up campaigns of metal-poor stars carried out so far, 90 metal-poor stars were discovered; 11 are unevolved stars with [Fe/H]<-3.0. Since in the HK survey 37 stars with [Fe/H]<-3.0 and 0.3<(B-V)0<0.5 were found, the sample of unevolved, extremely metal-poor stars was increased noticably.

Again by automatic spectral classification, it was searched for Field Horizontal-Branch A-type stars (FHB/A) in the HES. In a sample of 104 stars for which follow-up observations were obtained, 91 (or 88%) turned out to be A-type stars. The HES FHB/A sample contains stars down to V=17.5, so that distances of almost 25kpc from the sun can be reached.

Other object types can be selected efficiently by simpler selection methods. For the selection of carbon stars, cutoff lines in two-dimensional feature spaces were constructed to separate stars with either two strong C2 or two CN absorption bands from ``normal'' stars. Application of this procedure to the spectra present on 329 HES plates (effective area 6400deg2) led to the identification of 351 carbon stars. The mean surface density detected by the HES hence is 0.055deg-2, which is almost a factor three higher than the surface density found by Green et al. (1994) in their photometric CCD survey. Moreover, the survey of Green et al. is 1.5m deeper than the HES (Vlim˜16.5 in the HES; Vlim ˜18.0 for the Green et al. survey). This indicates that photometric carbon star surveys are highly incomplete.

We started to obtain recent-epoch CCD images for HES carbon stars, in order to measure their proper motions (p.m.) by comparison of these images with archival plate material (DSS-I, POSS, USNO-catalog 2.0). The aim of this project is to increase the sample of dwarf carbon stars (dCs). Up to now, one dC was rediscovered; another star is a likely candidate for a dC, but it can not be ruled out by its p.m. that it is a subgiant. Unfortunately, the selection probability for high p.m. objects in the HES is reduced due to an epoch difference between the direct plates of the DSS-I, used for object detection, and the HES spectral plates. Simulations show that nevertheless 20 new dCs are expected to be found in the HES. This would triple the dC sample.

The epoch difference problem is also relevant for searching white dwarfs (WDs) in the HES. A rough estimation indicates that only in the order of 50% of all WDs can be found in the HES, without employing special techniques to correct for for their p.m. However, it was found that even without using such techniques, we detect a higher surface density of WDs than the PG survey. This is most likely due to the fact that samples of DA white dwarfs drawn from UV-excess surveys are incomplete at the cool end, if too strict selection criteria are employed and/or colour measurements are inccurate. In the HES, cool DAs with temperatures above 9000K can readily be identified by their broad Balmer lines.

Further projects aimed at the selection of hot subdwarfs (sdO, sdB), magnetic DB white dwarfs, DZ white dwarfs, and cataclysmic variable stars in the HES.

click here Back to PhD thesis of Norbert Christlieb