1 Introduction

Subluminous B (sdB) stars dominate the populations of faint blue stars of our own Galaxy and are found in the disk (field sdBs) as well as in globular clusters (Moehler et al. 1997). Furthermore, these stars play an important role trying to explain the "UV upturn phenomenon'' observed in elliptical galaxies and galaxy bulges (Greggio & Renzini 1990, 1999). According to observations with the Ultraviolet Imaging Telescope (Brown et al. 1997) and the Hubble Space Telescope (Brown et al. 2000) sdB stars are sufficiently numerous to be responsible for the excessive UV flux.

It is generally accepted that sdB stars can be identified with models of the extended Horizontal Branch (EHB) burning He in their core (Heber 1986; Saffer et al. 1994). The hydrogen envelope surrounding the core of about half a solar mass is very thin (<2 % by mass) and therefore inert. These EHB stars will continue their evolution directly towards the white dwarf graveyard avoiding the AGB and planetary nebula phases (Dorman et al. 1993).

How sdB stars evolve towards the EHB with effective temperatures of up to 40 000 K remains a puzzle. The star must have lost all but a tiny fraction of the hydrogen envelope at the same time as the He core has attained the minimum mass ($\approx$0.5 $M_\odot$) required for the He flash. This challenges every mass loss mechanism in aspects of timing and effectivity. Recent findings (Maxted et al. 2001; Saffer et al. 2001; Heber et al. 2002a) emphasize the significance of close binary evolution.

Some of the sdB stars were recently found to exhibit rapid multi-periodic light variations ($P \approx$ 80-600 s) of low amplitudes (a few mmag). They form a new class of pulsating stars named after the prototype EC 14026 stars (Kilkenny et al. 1997). Since then a relatively large number of new sdB pulsators has been discovered. 31 are known today (Charpinet 2001; Piccioni et al. 2000; Silvotti et al. 2000). The observed brightness variations are caused by radial and non-radial, low degree and low order acoustic pulsation modes. 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 $\approx$2 $\times$ 10⁵ K in the sdB envelope.

Stellar pulsations allow a direct insight into the structure of such stars and therefore into the evolutionary history. The frequencies or periods of the pulsation modes probe the chemical stratification and the mass which otherwise are difficult or even impossible to determine. The power of asteroseismological tools has been demonstrated in the field of pulsating white dwarfs for which stellar parameters like mass, luminosity or thickness of the envelope were derived (e.g. Winget et al. 1991). In the case of variable sdB stars these parameters will constrain the evolutionary history and consequently shed more light on the origin of these stars.

Identification of pulsation modes (characterized by spherical harmonics with the indices l and m) is a prerequisite for asteroseismology. Brassard et al. (2001) have successfully carried out an asteroseismological analysis for PG 0014+067. For the first time they were able to determine the stellar mass ( $M_{\star}/M_{\odot} = 0.490\pm 0.019$) as well as the envelope mass ( $\log\left(M_{\rm env}/M_{\odot}\right) = -4.31 \pm 0.22$) and both are in excellent agreement with predictions from evolutionary models (Dorman et al. 1993).

Pulsations produce not only photometric variations but also line profile variations that offer an alternative approach towards mode identification. PG 1605+072 is the ideal target for this application: it has the longest pulsation periods ($\approx$500 s) which enables spectra with reasonable S/N to be obtained within each pulsation period. Moreover, this star has the largest variations of all known sdBVs (0.2 mag in the optical) and by far the richest frequency spectrum (>50 modes, Kilkenny et al. 1999). A recent spectroscopic study revealed this star to be a rather rapid rotator ( $v \sin i = 39$ km s^-1, Heber et al. 1999) which will complicate the identification of modes due to non-linear effects on mode splitting.

This work presented here serves as a feasibility study in order to find out whether an asteroseismologic analysis of PG 1605+072 is possible. For this purpose we have done simultaneously time-series spectroscopy and multi-band photometry (observations and reductions in Sect. 2). The analysis of the results is presented in Sect. 3. Previously, other groups have done radial velocity studies (O'Toole et al. 2000, 2002; Woolf et al. 2002) or photometric campaigns (e.g. Kilkenny et al. 1999) on PG 1605+072. Simultaneous multi-band photometry has not been observed before. This enables us to study the temporal evolution of the frequencies and amplitudes of the pulsation modes (Sect. 4). Finally, we discuss our results and give a brief outlook to future work in Sect. 5.