Pulsating stars allow a direct investigation of their structure - and therefore of their evolutionary history - from the evaluation of the pulsation modes. Pulsations produce photometric and also line profile variations. For the sdB variable PG1605+072 we have demonstrated that simultaneous time resolved spectroscopy and multi-channel photometry allows the main periods and amplitudes of radial velocity and light curves to be measured reliably. We propose a multi-site spectroscopic and photometric campaign to increase the time resolution in order to resolve closely spaced periods and to measure the amplitudes of both light and velocity variations. These will be the key signatures from which to identify the pulsation modes. Once modes are identified, the powerful tools of asteroseismology can be applied to derive stellar parameters such as total mass and envelope mass.
SdB stars dominate the populations of faint blue stars and are found in both the old disk (field sdBs) and halo populations (globular cluster members) of our own Galaxy. Considerable evidence has accumulated that these stars are sufficiently common to be the most likely source for the "UV upturn phenomenon" observed in elliptical galaxies and galaxy bulges (Brown et al., 1997). However, important questions remain about their evolutionary paths and the appropriate timescales. SdB stars can be identified with models for extreme Horizontal Branch (EHB) stars burning He in their core (Heber 1986).
Kilkenny et al. (1997) were the first to detect photometric variations due to radial and non-radial pulsations in some of the sdB stars (sdBV, two dozen are known today). These pulsating sdB stars have periods in the range 100--600 s with low amplitudes (a few mmag). Log g and Teff for the pulsators are estimated to be in the ranges 5.2-6.1 and 30000-36000 K. The pulsations in these stars are driven by an opacity bump due to Fe and other metallic species (Charpinet et al. 1997) at a temperature of ~2 x 10^5 K in the sdB envelope. However it is a puzzle that some sdBs pulsate while others of identical log g and Teff do not.
Pulsating stars allow a direct investigation of their structure - and therefore of the evolutionary history - from the evaluation of the pulsation modes which are eigenfrequencies of a self-gravitating gaseous sphere. The periods of the pulsation modes are determined by the chemical stratification and the mass. These are otherwise impossible or difficult to measure from spectral analyses. Asteroseismology was successfully applied to white dwarfs (e.g. Winget et al. 1991). A prerequisite for stellar seismology is the identification of the modes, i.e. the number and position of nodes on the stellar surface (described by spherical harmonics with indices l and m) and in radial direction. This is possible from photometric observations when the different pulsation modes are grouped very regularly, a slow rotation removes the m-degeneracy and produces e.g. triplets (for l=1 modes) in the fourier spectrum. In the sun and in pulsating hot white dwarfs, modes with high radial number are excited which produce a nearly equidistant mode pattern for consecutive overtones. This is not true in sdBV stars. So far it has been impossible to identify the pulsation modes of the observed periods. The interiors of sdB stars therefore remain to be explored.
The primary aim of this proposal is a mode identification for an sdBV star, namely PG1605+072, through an alternative approach. Pulsations not only produce photometric variations but also spectroscopic variations. PG1605+072 is the ideal target, it has the longest pulsation periods (~500s) which allows a reasonable S/N within each pulsation period. PG1605+075 also has the largest variations of sdBV stars (0.2 mag in the optical) and with more than 50 modes the richest pulsation spectrum. However, PG1605+072 is rotating rather rapidly (v sin i = 39 km/s, Heber et al. 1999) which makes the mode identification more complicated since non-linear effects have to be accounted for. Therefore it was up to now impossible to identify the modes from optical light curves.
O'Toole et al. (2000) were the first to detect radial velocity shifts of 14 km/s (at H beta) from time-resolved spectroscopy at medium resolution. Time-resolved spectroscopy has been carried out at the AAT, the WHT (14h in 2000, Woolf et al., 2001) and the Calar Alto 3.5m telescope (5h in May 2001, Falter et al. 2002) at much better spectral and temporal resolution (< 1 A). In both runs three pulsation periods have been resolved. Woolf et al. suspect a splitting of the two dominant modes which they tentatively attribute to rotation (about 12h rotation period). They find, however, that the amplitude ratios of the detected modes differ from those derived from photometry in 1998. They also find evidence of line profile variations. Observations in May 2001 indicate that the amplitudes are consistent with those of the 1998 photometry again. This means that there has been a switch in power over the years. Falter et al. (2002) were also able to obtain photometry in four filters simultaneous with the spectroscopy for the first time. This is important since it provides information on the continuum variations.
We aim to measure accurate amplitudes for oscillation modes down to ~2 km/s and line-profile variations. We therefore require spectra with high spectral and temporal resolution down to ~2 km/s, using nearly continuous spectroscopy for 4 nights on 4m telescopes. To resolve the individual frequencies, we require spectroscopy of slightly lower resolution using 2m telescopes for up to to 2 weeks. We also require simultaneous photometry in order to measure light/velocity amplitude ratios. We have found that, for small l, this quantity is dependent on the mode of oscillation.