The D/H ratio in interstellar gas toward G191-B2B

Recent analysis of Goddard High-Resolution Spectrograph (GHRS) echelle spectra suggests ∼30% variations in the D/H abundance ratio along the line of sight to the nearby (69 pc) hot white dwarf (WD) G191-B2B (Vidal-Madjar et al.). Variations in the D/H ratio on such short length scales imply nonuniform production/destruction of deuterium and an inefficient mixing of gas in the local interstellar medium (LISM). We reinvestigate the question of the spatial variation of the local D/H abundance using both archival GHRS spectra and new echelle spectra of G191-B2B obtained with the Space Telescope Imaging Spectrograph (STIS) aboard the Hubble Space Telescope. The STIS spectra were obtained in the high-resolution (E140H) mode and cover the wavelength region ranging from 1140 to 1700 Å. Our analysis uses stratified line-blanketed non-LTE model atmosphere calculations to determine the shape of the intrinsic WD Lyα profile and to estimate the WD photospheric contamination of the interstellar lines. Although three velocity components were reported previously toward G191-B2B, we deduce only two velocity components. The first component is at v_he1 ∼ 8.6 km s^-1, and the second is at v_he1 ∼ 19.3 km s^-1, which we identify with the local interstellar cloud (LIC). From the STIS data, we derive D/H = 1.60^+0.39_-0.27 × 10^-5 for the LIC component and D/H > 1.26 × 10^-5 for the 8.6 km s^-1 component (uncertainties denote 2 σ or 95% confidence limits). The derived D/H values in both components are consistent with (D/H)_LIC = (1.5 ± 0.1) × 10^-5, which was determined by Linsky in 1998. The STIS data provide no evidence for local or component-to-component variation in the D/H ratio. Our reanalysis of the GHRS data gives essentially the same results as Vidal-Madjar et al., despite using two velocity components for the profile fitting (vs. three by Vidal-Madjar et al.) and a more physically realistic WD Lyα profile for G191-B2B. The GHRS data indicate a component-to-component variation as well as a variation of the D/H ratio in the LISM, neither of which are supported by the newer STIS data. The D I absorption in the GHRS spectrum is shallower than in the STIS spectrum. The most probable cause for this difference in the two data sets is the characterization of the background due to scattered light in the GHRS and STIS spectrographs. The D/H ratios derived are sensitive to the background-subtraction procedures employed. The two-dimensional MAMA detectors of STIS measure both the spatial and wavelength dependences of scattered light, allowing more accurate scattered-light corrections than was possible with GHRS.