A pulsation mechanism for GW Virginis variables

The mechanism that produces pulsations in the hottest pre-white dwarfs has been uncertain since the early work indicated that helium is a poison that smooths opacity bumps in the opacity-temperature plane caused by the ionizations of the large observed amounts of carbon and oxygen. Very little helium seemed to be needed to prevent the kappa effect pulsation driving, but helium amounts of almost half of the mass in the surface composition are observed in the pulsating PG 1159-035 stars called the GW Virginis variables. Rather little change in the C and O surface abundances is observed from the hottest (RX J2117.1+3412 at 170, 000 K) to the coolest (PG 0122+200 at 80, 000 K) GW Vir variables. Actually the shortest observed periods (300-400 s) of these variables are generally predicted to be unstable in all models, but the longest observed periods ( up to 1000 s) are difficult to excite. Three recent investigations differ in their conclusions, with two finding that helium and even a slight amount of hydrogen does not prevent the kappa effect of C and O ionizations. A more detailed study reported here confirms the poisoning effect of helium. However, the ionization K- and L-edge opacity of the original iron, whose global abundance is unaffected by all previous evolution, especially if enhanced by radiation absorption levitation, can give different, previously unexplored, opacity driving that can explain the observed pulsations. But even this iron ionization driving can be somewhat poisoned by bump smoothing if the C and O abundances are large. Nonvariable GW Vir stars in the observed instability strip could be the result of small composition variations in the pulsation driving layers.