BROAD H I ABSORBERS AS METALLICITY-INDEPENDENT TRACERS OF THE WARM-HOT INTERGALACTIC MEDIUM

Thermally broadened Lya absorbers (BLAs) offer an alternative method to highly ionized metal lines for tracing the warm-hot intergalactic medium (WHIM) at T > 10(5) K. However, observing BLAs requires data of high quality and accurate continuum definition to detect the low-contrast features, and a good knowledge of the velocity structure to differentiate multiple blended components from a single broad line. Even for well-characterized absorption profiles, disentangling the thermal line width from the various thermal and non-thermal contributors to the observed line width is ambiguous. We compile a catalog of reliable BLA candidates along seven active galactic nucleus sight lines from a larger set of Ly alpha absorbers observed by the Space Telescope Imaging Spectrograph on the Hubble Space Telescope (HST). We compare our measurements based on independent reduction and analysis of the data to those published by other research groups. We examine the detailed structure of each absorber and determine a reliable line width and column density. Purported BLAs are grouped into probable (15), possible (48), and non-BLA (56) categories. Combining the first two categories, we infer a line frequency (dN/dz)(BLA) = 18 +/- 11, comparable to observed O VI absorbers, also thought to trace the WHIM. We discuss the overlap between BLA and O VI absorbers (20%-40%) and the distribution of BLAs in relation to nearby galaxies (O VI detections in BLAs are found closer to galaxies than O VI nondetections). We assume that the line width determined through a multi-line curve of growth (COG) is a close approximation to the thermal line width. Based on 164 measured COG H I line measurements, we statistically correct the observed line widths via a Monte Carlo simulation. Gas temperature and neutral fraction f(H I) are inferred from these statistically corrected line widths and lead to a distribution of total hydrogen columns. Summing the total column density over the total observed path length, we find a BLA contribution to the closure density of Omega(BLA) = 6.3(-0.8)(+1.1) x 10(-3) h(70)(-1) based on 10(4) Monte Carlo simulations of each BLA system. There are a number of critical systematic assumptions implicit in this calculation, and we discuss how each affects our results and those of previously published work. In particular, the most comparable previous study by Lehner et al. gave Omega(BLA) = 3.6 x 10(-3) h(70)(-1) or 9.1 x 10(-3) h(70)(-1), depending on which assumptions were made about hydrogen neutral fraction. Taking our value, current O VI and BLA surveys can account for similar to 20% of the baryons in the local universe while an additional similar to 29% can be accounted for in the photoionized Ly alpha forest; about half of all baryons in the low-z universe are found in the intergalactic medium. Finally, we present new, high signal-to-noise ratio observations of several of the BLA candidate lines from Early Release Observations made by the Cosmic Origins Spectrograph on HST.