SPECTROSCOPIC DETERMINATIONS OF SURFACE GRAVITIES OF GIANT STARS AND ULTRAVIOLET OBSERVATIONS OF RR LYRAE AND X ARIETIS

Abstract

The surface gravities of four Κ giant stars, including the well-known standard stars Arcturus (a Bootis) and Aldebaran (a Tauri) are determined from an analysis of the molecular dissociation equilibrium of OH. High-resolution FTS spectra of the infrared ground state vibration-rotation lines of the OH Δυ = 1 and Δϋ = 2 sequences were used in the analysis. The oxygen abundances were derived from the [O i] lines at 6300 A and 6363 A. Lifetimes for the OH Δ ν = 2 sequence lines are derived based on the Werner et al. (1983) electric dipole moment function and rotationless Einstein A values. Additionally, the Δυ = 2 sequence OH lines are identified in the solar spectrum using the infrared atlas of Delbouille et al. (1981) and analyzed to establish the accuracy of the calculated lifetimes. Masses for the stars based on the derived gravities, measured parallaxes and angular diameters are also discussed. The results indicated by the OH/[O i] solutions for Arcturus are logg = 1.7 ± 0.2 for an effective temperature of 4375 ± 50 Κ and a mass of 0.8 3ΚΘ. The corresponding results for Aldebaran are log g = 1.0 ± 0.5 using a temperature of 3860 ± 100 Κ and amass of0.73Jio. The surface gravities of five further G and Κ giants were determined from an analysis of the molecular equilibrium of MgH, the magnesium abundances being derived from visible and near-infrared high-excitation potential Mg I lines. The MgH features used, near 5135 A, are from the A-X system (0,0) band. The effects of C2 Swan band blends are included in the MgH analysis. The sensitivity of the derived gravities to the adopted model-atmosphere parameters is described. As a test for non-LTE effects, the MgH/Mg I gravities are compared to surface gravities derived based on an analysis of the ionization equilibrium of iron. Masses derived using these gravities, measured parallaxes, and angular diameters determined from theoretical surface brightness vs. color relationships are given. It is found that the giant star masses are not significantly subsolar. Ultraviolet photometric observations of RR Lyrae obtained with the Astronomical Netherlands Satellite (ANS) are also discussed. These observations are compared with light curves derived using model atmospheres and synthetic spectrum calculations in conjunction with angular diameters determined by Manduca et al. (1981) from photometry at longer wavelengths. A good agreement is found. These results represent an important independent confirmation of the distance and obtained for RR Lyr by Manduca et al. using a theoretically calibrated surface-brightness relation. As an extension of this work, seven low-dispersion spectra of RR Lyr and eleven of X Arietis were obtained with the LWR camera of the International Ultraviolet Explorer (IUE) satellite. RR Lyr was also observed in the low-dispersion mode with the SWP camera (9 spectra) and in the high-dispersion mode with the LWR camera (3 spectra). These observed fluxes are also compared with fluxes calculated using synthetic spectra and the Manduca et al. angular diameters. The IUE fluxes for RR Lyr are found to be significantly fainter than the computed ones. RR Lyr is also found to be 0T4 fainter at 2200 A and 2400 A during minimum light than it was during the previous set of observations made with the ANS satellite. It is suggested that both these differences are due to the combined effect of the secondary cycle of the star (the Blazhko effect) and line blocking not included in the synthetic flux calculations. One of the high-dispersion exposures made near maximum light is deep enough to reveal the cores of the Mg H h and k lines. No emission is observed in this exposure and a strong Mg π interstellar absorption component is present. Emission features are also found to be generally absent from the SWP spectra. An excellent agreement between observed and calculated fluxes is found for X Ari, an extremely metal-poor RR Lyrae star which has no observed secondary cycle. This result for X Ari strongly supports the Manduca et al. distance and .