Fe-bump instability: the excitation of pulsations in subdwarf B and other low-mass stars

Abstract

We consider the excitation of radial and non-radial oscillations in low-mass B stars by the iron-bump opacity mechanism. The results are significant for the interpretation of pulsations in subdwarf B stars, helium-rich subdwarfs and extreme helium stars, including the EC14026 and PG1716 variables. We demonstrate that, for radial oscillations, the driving mechanism becomes effective by increasing the contrast between the iron-bump opacity and the opacity from other sources. This can be achieved either by increasing the iron abundance or by decreasing the hydrogen abundance. The location of the iron-bump instability boundary is found to depend on the mean molecular weight in the envelope and also on the radial order of the oscillation. A bluer instability boundary is provided by increasing the iron abundance alone, rather than the entire metal component, and is required to explain the observed EC14026 variables. A bluer instability boundary is also provided by higher radial order oscillations. Using data for observed and theoretical period ranges, we show that the coolest EC14026 variables may vary in the fundamental radial mode, but the hottest variables must vary in modes of higher radial order. In considering non-radial oscillations, we demonstrate that g-modes of high radial order and low spherical degree (l < 4) may be excited in some blue horizontal branch stars with near-normal composition (Z= 0.02). Additional iron enhancement extends the g-mode instability zone to higher effective temperatures and also creates a p-mode instability zone. The latter is essentially contiguous with the radial instability zone. With sufficient iron, the p- and g-mode instability zones overlap, allowing a small region where the EC14026- and PG1716-type variability can be excited simultaneously. The overlap zone is principally a function of effective temperature and only weakly a function of luminosity. However, its location is roughly 5000 K, too low compared with the observed boundary between EC14026 and PG1716 variables. The discrepancy cannot be resolved by simply increasing the iron abundance.