Spectroscopic analyses of subluminous B stars: observational constraints for the theory of stellar evolution, pulsation, and diffusion

This thesis deals with quantitative spectroscopic analyses of large samples of subluminous B stars in order to find constraints the theory of stellar evolution, pulsation, and diffusion.

Subluminous B stars, also known as subdwarf B (sdB) stars, are very important in several respects: They dominate the population of faint blue stars in high galactic latitudes, and are found both in the field and in globular clusters. Therefore, sdB stars are important to understand the structure and evolution of our galaxy. From the cosmological point of view, they are candidate progenitors of supernovae of type Ia due to their membership in close binary systems. In the context of stellar astrophysics, subdwarf B stars play an important role because several of them are discovered to show non-radial pulsations, which allows to probe their interior by asteroseismology. Last but not least, sdB stars show very peculiar element abundance patterns, probably caused by diffusion processes.

Subluminous B stars are generally considered to be core helium-burning stars with extremely thin hydrogen envelopes (< 0.02M) and masses around 0.5M. In the Hertzsprung-Russell diagram they populate a very narrow area which lies on a blue-ward extension of the horizontal branch (HB), the so called "Extreme Horizontal Branch" (EHB). Due to their thin hydrogen-rich envelope, they cannot sustain a hydrogen-burning shell. This means that the sdB stars cannot ascend the asymptotic giant branch after the end of the helium-core burning, but should evolve directly to the white dwarf graveyard. However, according to standard stellar evolution theory, subdwarf B stars should not exist. Their evolution is still unclear and several scenarios are under debate. For all these investigations, knowledge of the stellar parameters (effective temperature, gravity and chemical composition) is very important to verify or discard theoretical models and predictions.

Numerous observing runs have been performed mostly by myself at the German-Spanish Astronomical Center on Calar Alto, Spain, and at the European Southern Observatory on La Silla, Chile, in order to perform detailed analyses of large samples of subluminous B stars. To obtain optical spectra with high S/N, 1.5m to 8m class telescopes equipped with up-to-date long-slit low-resolution as well as echelle high-resolution spectrographs together with state-of-the-art CCD detectors were used.

Optical low-resolution long-slit spectra of 146 subdwarf B stars (a sample twice as large as that of previous investigations), drawn from follow-up observations of candidates from the Hamburg Quasar and Hamburg ESO surveys have been analyzed to determine their fundamental parameters (effective temperature, gravity, and helium abundance). The parameters are determined by matching synthetic line profiles calculated from LTE and NLTE model atmospheres to all hydrogen and helium absorption lines present in the observed spectra. The results determined from the LTE and NLTE fits show only slight differences. We compared our analysis results with data from the literature. Systematic differences occur at a 10% level in temperatures, and at the level of ∼0.2 dex in gravity, which cannot be explained by NLTE effects alone. Extensive investigations on a star by star basis are needed to resolve the cause, which are beyond the scope of this thesis. Due to the remaining uncertainties in the atmospheric parameters we had to refrain from a discussion of the details of EHB evolution. This has to be postponed until the reason for the systematic differences have been clarified.

Recently, several sdB stars have been found to show non-radial pulsations. We initiated a collaboration with two groups in Norway and Italy in 1999 to search for pulsating sdB stars in our sample. About one pulsator within ten observed sdB stars were found. With this discovery we enhanced the number of known pulsating sdB stars by about 50%.

The surface metal abundance patterns of 16 sdB stars have been determined from high resolution, high S/N, optical spectra using equivalent widths measurements. This analysis almost quadruples the number of detailed metal abundance analyses of sdB stars. As typical for early B type stars, the metal lines are few and very weak. Three peculiar sdB stars have been found which show in addition to the absorption lines common in sdB stars many lines due to iron group elements (calcium, scandium, titanium, vanadium, manganese, and nickel) which have never been found before in the optical spectra of sdB stars. Surprisingly, almost all program stars show very similar abundance patterns: carbon, oxygen, magnesium, and aluminum are strongly depleted (with respect to the solar composition), nitrogen is mostly slightly under-abundant, silicon and sulfur are subsolar, iron is always about solar, and argon is even suprasolar. For the peculiar sdB stars we determine an enormous enrichment of the iron group elements which are found to be 1000 to ∼32000 times the solar values. However, in contrast to the "normal" sdB stars, no lines due to iron itself occur in the spectra of the peculiar stars. From curve-of-growth measurements we determine vanishing microturbulent velocities for almost all program stars, except for two which have values of 3 ± 1 km/s and 6 ± 1 km/s. The projected rotational velocities determined for all program stars indicate that all are (very) slowly rotating stars.

For the first time, for sdB stars it was possible to correlate element (especially metal) abundance patterns with the atmospheric parameters to search for possible trends, which are predicted by diffusion theory. We discover a correlation for some elements with the effective temperature: the larger the temperature, the larger the abundances for helium, oxygen, and magnesium. On the other hand it is remarkable that the abundances for nitrogen, silicon, aluminum, argon, and iron are constant, irrespective of the stellar parameters. In addition, a separation into two sequences of sdB stars may exist: a minority having much lower helium abundances at the same temperatures than the bulk of the sdB stars. The same possibly holds for oxygen, magnesium, and silicon too. An isotopic anomaly of helium has been found in two sdB stars. In these stars 4He is largely replaced by the isotope 3He which is also due to diffusion processes.

Recently, observational evidence is accumulating that close binary evolution is crucial to understand the origin of sdB stars. Therefore, we embarked on a search program to discover single-lined spectroscopic binaries from radial velocity variations. 47 bright hot subluminous stars have been observed for radial velocity variations from high resolution optical spectra. For nine single lined stars we can definitely rule out them to be RV variable stars. Neither short-term (a few hours), nor long-term (weeks or years) variations could be found. 22 stars remained with unknown status (more observations are necessary). However, for 13 hot subdwarfs short period radial velocity variations have been detected. For nine of them radial velocity curves have been measured. The periods determined range from about half a day up to almost nine days. The radial velocity semi amplitudes found are in the order of 22 km/s to 188 km/s. By use of the mass function, the masses of the unseen companions have been constricted. Lower limits and most probable values of 0.07 up to 0.73 M, and 0.09 up to 1.20 M, respectively, have been determined. We conclude that the invisible companions are most likely late type main sequence stars, or white dwarfs with a C/O core. However, for one star a neutron star cannot be ruled out as a secondary.

As spin off result, an apparently normal main sequence B star with an enormous absolute radial velocity of v_rad = 723 ± 3 km/s has been discovered which will be emitted into the intergalactic space.