Search for RV-Variations in active and young stars
Planet-induced RV-variations can be detected most easily on slowly rotating stars with undisturbed line profiles. This is why RV searches have been targeting mostly planets of old and inactive stars. Yet, all theories suggest the simultaneous formation of stars and planets. Therefore, planets should also exist around the young, more active stars. For our efforts to determine differential rotation on stars either by comparing Doppler images or via their Fourier-transformed line profiles, we have developed sophisticated techniques to extract line profiles from high signal-to-noise, high-spectral resolution data. For template spectra, we use either suitable comparison stars or theoretical (PHOENIX) model spectra. The observed high-resolution spectrum is modeled as a convolution between the template spectrum and the line profile. We have shown these algorithms to be extremely powerful, and complex line profiles (for example, the spectrum of a triple system) can be obtained from heavily blended spectra. We now plan to extend our methods to RV data. RV measurements achieve high velocity precision through the use of superimposed calibration spectra, but are currently limited to 3 m/sec by the imperfect modeling of the (variable) instrument profile in the spectral analysis. The simultaneous modeling of superimposed reference spectra can easily be accomplished by our technique. In oscillation studies extending over it short times, i.e., without the long term stability problems of RV planet searches, an RV accuracy of 10 cm/sec has been demonstrated. Our line profile analysis thus applied to RV measurements is therefore expected to improve the accuracy of RV measurements to the cm/sec level or that permitted by the photon noise of the bright stars. The second factor limiting the RV accuracy can be the transformation of the measured 2-D spectral frame into the 1-D spectrum used in the analysis. Temporal and spectral variability can cause substantial errors (at the m/sec accuracy level). We will investigate this systematically with high-S/N stellar data and expect to improve the accuracy substantially. The techniques for the 2-D line profile analysis are almost identical to the ones used in spectro-astrometry for direct IR planet searches. We expect substantial cross-fertilization between these two research programs of our RTG. The research plan foresees the detailed study of the techniques, the incorporation of simultaneous reference spectrum modeling into the deconvolution algorithms and codes, and the development of the 2D spectral PSF analysis. The latter will be applied in parallel to data from the spectroastrometry project of the RTG.
Metallicity of planet bearing stars
The characterization of the spectroscopic properties of stars continues to be an important task - for the investigation of exoplanet host stars it is essential. Spectroscopy is not only an analytical tool, but may be required for the planet detection or identification (when direct imaging is not possible or ambiguous). Important stellar quantities in this context are: rotation, atmosphere kinematics and the metal abundance. The 150 known extrasolar planet host stars are metal rich compared to field stars. Detailed studies show that this difference is significant and a real physical property of the stars rather than an artifact of the planet search or abundance analysis method. The observed metal enhancement of extrasolar planet host stars can either be explained by a metal enhancement of the parent cloud and, hence, a higher efficiency in the production of proto-planetary dust and proto-planetary bodies, or by a subsequent metal enrichment due to infall of planets onto the stars. Once the star has arrived on or at least near the main sequence, any infalling material can only be mixed with the convection zone and not with the radiative core. Therefore, one expects an inverse relation between convection zone mass and metal abundance enhancement. Observations seem to support the alternative explanation, i.e., high metallicity of parent cloud, with important implications for the way planets actually form in the disk, i.e., via core-accretion rather than a disk instability mechanism. (Fischer & Valenti 2005). More and better observational data will be required to settle the isse. We intend to address planet host star metallicities by heavy element abundance measurements. High-quality optical spectra will be available from observations with public telescopes (ESO, Calar Alto) accessible through the normal application process, observations with the echelle spectrograph at the robotic telescope HRT owned by the Hamburg Observatory, and VLT/ESO & HST archive spectra. We expect improvements in the accuracy of abundance measurements from our advanced modeling capabilities which include non-LTE and 3-D radiation transfer. As a novel feature we will introduce the supplementary abundance measurements via IR spectroscopy of molecules in the cooler stars (F5 and later, which coincides with the range of stars accessible to the RV technique). The molecular abundances provide a critical independent test of the element abundances determined from atomic lines in the optical. This technique (conceived originally for isotopic CO lines in red giant stars) is particularly powerful, if reliable atmosphere models at photospheric/chromospheric heights are available (such as those supplied by our theory branch).
Atmospheres of planet host stars
The analysis of chemical abundances and atmospheric structure parameters is of particular importance to the parent stars of extrasolar planets. Accurate stellar atmosphere parameters are important, first for target selection of RV and other planet searches, and second in any kind of correlation with planetary or planetary orbital parameters relevant to the formation and evolution. Standard analyses use LTE for the coupled problem of differential abundances and atmosphere structure determinations. We plan to improve this methodologically by using detailed non-LTE models. This will include a systematic investigation of the effects of stellar activity on the derived models. Observations: IR spectroscopy will be carried out at large telescopes (VLT, CRIRES). The research plan further include optical spectroscopy, the use of archive data and the observations of active stars with our own telescopes. Observations and analyses are suitable for and will be carried out on a wide range of host stars. Our claim to top level analyses stems from: 1) the multi-wavelength stellar spectroscopy, 2) the most advanced modeling capabilities developed by and available to us and 3) the close collaboration of observers and theoreticians. The RTG host institutes are well positioned for this effort, having considerable experience with non-LTE modeling. Hamburg further has experience with chromospheric models (Schmitt and Hauschildt) resp. bifurcated and chromosphereless models apparent in molecular diagnostics (Wiedemann). Our analyses will use a homogeneous set of model atmospheres (e.g., derived from the current Hamburg GAIA grids) and can span a large stellar parameter range (from M dwarf to A stars and beyond).