2006ApJ...643.1198R

We present the results of an extensive multisite campaign on the long-period variable subdwarf B star PG 1627+017. We gathered 300 hr of useful R-band and ∼50 hr of simultaneous U/R differential photometry. From the R-band data we were able to extract 23 periodicities in the 4500-9000 s range with amplitudes between 0.05% and 0.5% of the star's mean brightness. The oscillations with the highest amplitudes cluster between 6300 and 7050 s and are thought to exhibit frequency splitting due to binary-synchronous stellar rotation. Interestingly, we find the observed period distribution to be extremely nonuniform, with dense frequency multiplets occurring in several narrow band passes. In order to compare the observed period spectrum to theoretical predictions, we constructed a set of newly updated and improved subdwarf B star models. We find that by invoking degree indices of l=2, 3, and 4, nonadiabatic calculations can qualitatively reproduce the range of periodicities measured for PG 1627+017 if its atmospheric parameters are pushed to the lower end of their spectroscopic temperature uncertainties. However, the exploitation of rotational splitting and the U/R photometry, as well as the mean spacing between periodicities, indicate that at least the four highest amplitude peaks probably correspond to modes with l=1. While this points to deficiencies in our models at the nonadiabatic level, the resulting constraints on mode identification are invaluable to first attempts at asteroseismology. Indeed, we identify only a few families of models that can closely reproduce the main periodicities observed in terms of dipole modes. This leaves us hopeful that, given a larger number of partially identified observed frequencies, asteroseismology may be achieved for long-period variable subdwarf B stars.