1: %
2: %% Beginning of file 'sample.tex'
3: %%
4: %% Modified 2004 January 9
5: %%
6: %% This is a sample manuscript marked up using the
7: %% AASTeX v5.x LaTeX 2e macros.
8:
9: %% The first piece of markup in an AASTeX v5.x document
10: %% is the \documentclass command. LaTeX will ignore
11: %% any data that comes before this command.
12:
13: %% The command below calls the preprint style
14: %% which will produce a one-column, single-spaced document.
15: %% Examples of commands for other substyles follow. Use
16: %% whichever is most appropriate for your purposes.
17: %%
18: %\documentclass[12pt,preprint]{aastex}
19:
20: %% manuscript produces a one-column, double-spaced document:
21:
22: %% \documentclass[manuscript]{aastex}
23:
24: %% preprint2 produces a double-column, single-spaced document:
25:
26: %\documentclass[preprint2,longabstract]{aastex}
27: %\documentclass[12pt,preprint]{aastex}
28: %\documentclass[12pt,manuscript]{aastex}
29: \documentclass[apjl]{emulateapj}
30:
31: %\usepackage{apjfonts}
32: %\usepackage{CT_macros}
33: \def\spose#1{\hbox to 0pt{#1\hss}}
34: \def\simlt{\mathrel{\spose{\lower 3pt\hbox{$\mathchar"218$}}
35: \raise 2.0pt\hbox{$\mathchar"13C$}}}
36: \def\simgt{\mathrel{\spose{\lower 3pt\hbox{$\mathchar"218$}}
37: \raise 2.0pt\hbox{$\mathchar"13E$}}}
38: \providecommand\scription[2]{\scriptsize#1$\;${\scriptsize\uppercase\expandafter{\romannumeral #2}}\relax}%
39: \providecommand\tabion[2]{#1$\;${\tiny\uppercase\expandafter{\romannumeral #2}}\relax}%
40: \providecommand{\FeH}{\ensuremath{\mbox{[Fe/H]}}}
41: \providecommand{\bnt}{\ensuremath{b_{\mathrm nt}}}
42: \providecommand{\btherm}{\ensuremath{b_{\mathrm therm}}}
43: \providecommand{\MH}{\ensuremath{\mbox{[M/H]}}}
44: \providecommand{\logT}{\ensuremath{\mbox{log}\,T}}
45: \providecommand{\logb}{\ensuremath{\mbox{log}\,b}}
46: \providecommand{\logU}{\ensuremath{\mbox{log}\,U}}
47: \providecommand{\Msun}{\,\ensuremath{\mbox{M}_{\odot}}}
48: \providecommand{\kms}{\,\ensuremath{\rm{km\,s}^{-1}}}
49: \providecommand{\cm}{\,\ensuremath{\mbox{cm}^{-2}}}
50: \providecommand{\A}{\,\ensuremath{\mbox{\AA}}}
51: \providecommand{\mA}{\,\ensuremath{\mbox{m\AA}}}
52: \providecommand{\kpc}{\,\ensuremath{\mbox{kpc}}}
53: \providecommand{\Mpc}{\,\ensuremath{\mbox{Mpc}}}
54: \providecommand{\eV}{\,\ensuremath{\mbox{eV}}}
55: \providecommand{\Zsun}{\,\ensuremath{Z_{\odot}}}
56: \providecommand{\HI}{\ensuremath{\mbox{\ion{H}{1}}}}
57: \providecommand{\Lya}{\ensuremath{\mbox{Ly\,}\alpha}}
58: \providecommand{\Lyb}{\ensuremath{\mbox{Ly\,}\beta}}
59: \providecommand{\Lyab}{\ensuremath{\mbox{Ly\,}\alpha, \beta}}
60: \providecommand{\Lycd}{\ensuremath{\mbox{Ly\,}\gamma, \delta}}
61: \providecommand{\Lyc}{\ensuremath{\mbox{Ly\,}\gamma}}
62: \providecommand{\Lyd}{\ensuremath{\mbox{Ly\,}\delta}}
63: \providecommand{\NeVIII}{\ensuremath{\mbox{\ion{Ne}{8}}}}
64: \providecommand{\NeIX}{\ensuremath{\mbox{\ion{Ne}{9}}}}
65: \providecommand{\OI}{\ensuremath{\mbox{\ion{O}{1}}}}
66: \providecommand{\OIII}{\ensuremath{\mbox{\ion{O}{3}}}}
67: \providecommand{\OIV}{\ensuremath{\mbox{\ion{O}{4}}}}
68: \providecommand{\OVI}{\ensuremath{\mbox{\ion{O}{6}}}}
69: \providecommand{\OVII}{\ensuremath{\mbox{\ion{O}{7}}}}
70: \providecommand{\OVIII}{\ensuremath{\mbox{\ion{O}{8}}}}
71: \providecommand{\CI}{\ensuremath{\mbox{\ion{C}{1}}}}
72: \providecommand{\CII}{\ensuremath{\mbox{\ion{C}{2}}}}
73: \providecommand{\CIII}{\ensuremath{\mbox{\ion{C}{3}}}}
74: \providecommand{\CIV}{\ensuremath{\mbox{\ion{C}{4}}}}
75: \providecommand{\NII}{\ensuremath{\mbox{\ion{N}{2}}}}
76: \providecommand{\NV}{\ensuremath{\mbox{\ion{N}{5}}}}
77: \providecommand{\SiII}{\ensuremath{\mbox{\ion{Si}{2}}}}
78: \providecommand{\SiIII}{\ensuremath{\mbox{\ion{Si}{3}}}}
79: \providecommand{\SiIV}{\ensuremath{\mbox{\ion{Si}{4}}}}
80: \providecommand{\SII}{\ensuremath{\mbox{\ion{S}{2}}}}
81: \providecommand{\scHI}{\ensuremath{\mbox{\scription{H}{1}}}}
82: \providecommand{\scOVI}{\ensuremath{\mbox{\scription{O}{6}}}}
83: \providecommand{\scSiIII}{\ensuremath{\mbox{\scription{Si}{3}}}}
84: \providecommand{\scCIII}{\ensuremath{\mbox{\scription{C}{3}}}}
85: \providecommand{\tabHI}{\ensuremath{\mbox{\tabion{H}{1}}}}
86: \providecommand{\tabOVI}{\ensuremath{\mbox{\tabion{O}{6}}}}
87: \providecommand{\tabOI}{\ensuremath{\mbox{\tabion{O}{1}}}}
88: \providecommand{\tabOIII}{\ensuremath{\mbox{\tabion{O}{3}}}}
89: \providecommand{\tabOIV}{\ensuremath{\mbox{\tabion{O}{4}}}}
90: \providecommand{\tabOVII}{\ensuremath{\mbox{\tabion{O}{7}}}}
91: \providecommand{\tabCI}{\ensuremath{\mbox{\tabion{C}{1}}}}
92: \providecommand{\tabCII}{\ensuremath{\mbox{\tabion{C}{2}}}}
93: \providecommand{\tabCIII}{\ensuremath{\mbox{\tabion{C}{3}}}}
94: \providecommand{\tabCIV}{\ensuremath{\mbox{\tabion{C}{4}}}}
95: \providecommand{\tabNII}{\ensuremath{\mbox{\tabion{N}{2}}}}
96: \providecommand{\tabNV}{\ensuremath{\mbox{\tabion{N}{5}}}}
97: \providecommand{\tabSiII}{\ensuremath{\mbox{\tabion{Si}{2}}}}
98: \providecommand{\tabSiIII}{\ensuremath{\mbox{\tabion{Si}{3}}}}
99: \providecommand{\tabSiIV}{\ensuremath{\mbox{\tabion{Si}{4}}}}
100: \providecommand{\tabSII}{\ensuremath{\mbox{\tabion{S}{2}}}}
101: \providecommand{\hden}{\ensuremath{n_{\mathrm H}}}
102: \providecommand{\Oden}{\ensuremath{n_{\mathrm O}}}
103: \providecommand{\loghden}{\ensuremath{\mbox{log}\,\hden}}
104: \providecommand{\NHI}{\ensuremath{N(\HI)}}
105: \providecommand{\NOVI}{\ensuremath{N(\OVI)}}
106: \providecommand{\NCIII}{\ensuremath{N(\CIII)}}
107: \providecommand{\logN}{\ensuremath{\mbox{log\,}N}}
108: \providecommand{\logNOVI}{\ensuremath{\mbox{log\,}N(\OVI)}}
109: \providecommand{\logNHI}{\ensuremath{\mbox{log\,N(\HI)}}}
110: \providecommand{\Wr}{\ensuremath{W_r}}
111: \providecommand{\K}{\,\ensuremath{\mbox{K}}}
112: \providecommand{\Oion}{\ensuremath{{\mathrm O}^{5+}}}
113: \providecommand{\ROVI}{\ensuremath{R_{\scOVI}}}
114: \providecommand{\bOVI}{\ensuremath{b_{\scOVI}}}
115: \providecommand{\bHI}{\ensuremath{b_{\scHI}}}
116: \providecommand{\bSiIII}{\ensuremath{b_{\scSiIII}}}
117: \providecommand{\bCIII}{\ensuremath{b_{\scCIII}}}
118: \providecommand{\pI}{\ensuremath{\mbox{paper~I}}}
119: \providecommand{\etal}{\ensuremath{\mbox{et~al.}}}
120: \providecommand{\dex}{\,\ensuremath{\mbox{dex}}}
121: \providecommand{\OmegaOVI}{\ensuremath{\Omega_{{\rm O}^{5+}}}}
122: \providecommand{\zqso}{\ensuremath{z_{\mathrm qso}}}
123: \providecommand{\zabs}{\ensuremath{z_{\mathrm abs}}}
124:
125: \shorttitle{Physical Conditions of low-$z$ \OVI\ Absorbers}
126: \shortauthors{Thom \& Chen}
127:
128: \begin{document}
129:
130: \title{A STIS Survey for \OVI\ Absorption Systems at $0.12 < z\lesssim
131: 0.5$ II.: Physical Conditions of the Ionized Gas}
132:
133: \author{C. Thom and Hsiao-Wen Chen} \affil{Dept.\ of Astronomy \& Astrophysics and Kavli
134: Institute for Cosmological Physics \\
135: University of Chicago, Chicago, IL, 60637, U.S.A. \\
136: {\tt cthom, hchen@oddjob.uchicago.edu}}
137:
138: \altaffiltext{1}{Based in part on observations with the NASA/ESA Hubble Space
139: Telescope, obtained at the Space Telescope Science Institute, which is operated
140: by the Association of Universities for Research in Astronomy, Inc., under NASA
141: contract NAS5--26555.}
142:
143: \begin{abstract}
144:
145: We present a complete catalogue of 27 \OVI\ absorbers at low
146: redshift ($0.12 < z < 0.5$) from a blind survey of 16 QSO echelle
147: spectra in the HST/STIS data archive. These absorbers are
148: identified based only upon matching line profiles and the expected
149: doublet ratio between the $\lambda\lambda\,1031, 1037$ transitions.
150: Subsequent searches are carried out to identify their associated
151: transitions. Here we present all relevant absorption properties.
152: By considering absorption components of different species which are
153: well-aligned in velocity-space, we derive gas temperatures and
154: non-thermal broadening values, $\bnt$. We show that in all 16 cases
155: considered the observed line width is dominated by non-thermal
156: motion and that gas temperatures are well below those expected for
157: \Oion\ in collisional ionization equilibrium. This result reaffirms
158: previous findings from studies of individual lines of sight, but are
159: at odds with expectations for a WHIM origin. At least half of the
160: absorbers can be explained by a simple photoionization model. In
161: addition, in some absorbers we find evidence for large variation in
162: gas density/metallicity across components in individual absorbers.
163: Comparisons of multiple associated metal species further show that
164: under the assumption of the gas being photoionized by the
165: metagalacitic background radiation field, the absorbing clouds have
166: gas densities $<\hden> < -2.9$ and sizes $L > 1\kpc$. Finally, we
167: compare our absorber selection with the results of other independent
168: studies.
169:
170: \end{abstract}
171:
172: \keywords{cosmology: observations---intergalactic medium---quasars: absorption lines}
173:
174: \section{Introduction}
175: \label{sec: introduction}
176: One of the current key questions in observational cosmology is the location of the missing
177: baryons. The {\it total} baryon content of the universe is well constrained, with various
178: measurements in relatively good agreement \citep[e.g.][]{burles-etal-01-BBN-omega,
179: spergel-etal-03-WMAP-yr1, omeara-etal-06-deuterium}. In the high-redshift universe, the \Lya\
180: forest dominates the baryon census \citep{rauch-etal-97-Lya-Omegab}, but in the present day
181: universe, only 1/3 of the baryons have been identified in known components
182: \citep{fukugita-peebles-04-mean-densities}. Cosmological simulations indicate that up to 50\% of
183: the baryons exist in a warm-hot intergalactic medium (WHIM), in which gas is shock-heated to $\sim
184: 10^5 - 10^7\K$ by the accretion onto large-scale structures, and remains hot owing to the low gas
185: density and inefficient cooling \citep{cen-ostriker-99-missing-baryons, dave-etal-01-WHIM-baryons,
186: cen-ostriker-06-II-winds}. This gas is a result of accretion onto large-scale filamentary
187: structures, where cooling is inefficient due to the low densities, and (in some models) the
188: operation of large scale winds that shock-heat outflowing gas to $\sim10^6\K$
189: \citep[e.g.][]{cen-ostriker-06-II-winds}. It is therefore important to identify this gas, and
190: constrain its contribution to the baryon fraction.
191:
192: At the temperature range in question ($\logT = 5.0 - 7.0$), the best observational window for this
193: hot gas is X-ray absorption lines \citep[e.g.\ \OVII\,K$\alpha$, \OVIII\,K$\alpha$,
194: \NeIX\,K$\alpha$; ][]{gnat-sternberg-07-non-equilibrium-cooling}, but the resolutions of current
195: X-ray spectrographs (e.g.\ Chandra and XMM/Newton) are an order of magnitude too coarse to be
196: useful, and the sensitivities are more than an order of magnitude too low
197: \citep[e.g.][]{fang-etal-06-galactic-OVII}. The best tool currently available, therefore, is UV
198: absorption spectroscopy. The \OVI\ doublet ($\lambda\lambda\,1031.9261, 1037.617\A$) offers the best
199: hot gas tracer for a number of reasons: the transition has a large oscillator strength; Oxygen is a
200: relatively common metal; the abundance of the \Oion\ ion peaks in collisional ionization equilibrium
201: (CIE) at $\logT \sim 5.3$\footnote{It also has an ionization fraction greater than $10^{-2}$ up to
202: $\logT \sim 5.7$; see e.g.\ \citet{gnat-sternberg-07-non-equilibrium-cooling}.}; the transitions
203: occurs longward of the Lyman limit; and absorption can be detected down to limiting column densities
204: $\logN \approx 13.5 (\Wr = 30\mA)$ with current instruments such as STIS.
205:
206: If the absorption lines of an element are resolved and unsaturated, we can measure directly the
207: column density ($N$) and Doppler parameter ($b$) of the absorbing gas by voigt profile
208: fitting\footnote{In cases where multiple transitions of the same species are available, a
209: curve-of-growth analysis can also be used.}. The Doppler parameter is an oft-used measure of the
210: temperature of the gas, through the well-known relation $b^2 = \bnt^2 + 2kT/m$, where \bnt\ accounts
211: for non-thermal broadening of the line due to e.g.\ turbulence. The column-density, meanwhile, may
212: be used in conjunction with the path-length to derive the contribution to the cosmological mass density of
213: the O$^{5+}$ ions, \OmegaOVI\ \citep[see][hereafter \pI]{thom-chen-08-I-OVI-statistics}.
214:
215: In the optical band, ground-based high resolution (echelle) spectra have made \OVI\ detection
216: possible at $2.3 \simlt z \simlt 3$ \citep[e.g.][]{norris-etal-83-OVI, carswell-etal-02-OVI,
217: simcoe-etal-02-OVI-search}. The lower limit is set by the plummeting transmissivity of the
218: atmosphere to UV photons below $\sim 3500\A$, while the upper limit is set by confusion with the
219: \Lya\ forest. In the high-redshift regime, \Oion\ is predominantly photoionized by the UV background
220: radiation field, much the same as the \Lya\ clouds \citep{carswell-etal-02-OVI, simcoe-etal-04}. To
221: push to lower redshifts, space-based UV spectrographs are required. Using near-UV spectra from the
222: Faint Object Spectrograph on-board HST, \citet{burles-tytler-96-OVI-Omega_b} identified 12 \OVI\
223: doublets with $W_{r}(1031) > 0.21\AA$ in the range $0.5 \simlt z \simlt 2$ and derived a
224: cosmological mass density $\OmegaOVI\,h \geq 7\times10^{-8}$, providing the first constraint on the
225: mass density of this highly ionized gas.
226:
227: In the low-redshift universe ($z \simlt 0.5$), the \OVI\ doublet transition remains in the far-UV
228: ($\lambda_{obs} \simlt 1600\A$). The terminated Far Ultraviolet Spectroscopic Explorer \citep[{\it
229: FUSE}; ][]{moos-etal-00-FUSE} could probe \OVI\ absorption out to $z \sim 0.15$ with good
230: signal-to-noise, but at a resolution of only $\sim 20\kms$
231: \citep[e.g.][]{danforth-etal-05-OVI_baryon_census}. This compares poorly with the thermal line width
232: of \OVI\ ($\sim 3\kms$ at $\logT = 4.0$ and $\sim 10\kms$ at $\logT = 5.0$). The (currently
233: suspended) Space Telescope Imaging Spectrograph \citep[{\it STIS}; ][]{woodgate-etal-98-STIS}
234: on-board the Hubble Space Telescope ({\it HST}) offers a resolution of $\sim 6-7\kms$ and its data
235: are useful for studies of \OVI\ absorption at slightly higher redshifts than FUSE ($0.12 < z < 0.5$).
236:
237: Early work on \OVI\ absorption in the low-$z$ universe typically concentrated on single lines of
238: sight, due to the paucity of data. Single-sightline or single-absorber analyses have been conducted
239: on the various lines of sight, beginning with the QSO H\,1821$+$643
240: \citep{tripp-etal-98-Lya-galaxies, tripp-etal-00-H1821+643, tripp-etal-01-H1821+643_z0.1212,
241: oegerle-etal-00-H1821+643-FUSE}. As more data become available, further sightline were analysed:
242: PG0953$+$415 \citep{savage-etal-02-PG0953+415}, PG1259$+$593 \citep{richter-etal-04-PG1259+593},
243: PG1116$+$215 821\citep{sembach-etal-04-PG1116+215}, PKS0405$-$123
244: \citep{prochaska-etal-04-PKS0405-12}, HE0226$-$4110 \citep{lehner-etal-06-HE0226-4110} and
245: PKS\,1302$-$102 \citep{cooksey-etal-08-PKS1302-102}. Now, with a database of UV data
246: available\footnote{More data are likely soon to be available with the upcoming HST servicing
247: mission.}, statistical approaches have become possible \citep{danforth-etal-05-OVI_baryon_census,
248: danforth-etal-06-OVI_survey, thom-chen-08-I-OVI-statistics, tripp-etal-08-OVI,
249: danforth-shull-08-OVI}.
250:
251:
252: This is the second in a series of papers reporting the results of our search for \OVI\ absorption
253: systems in the STIS E140M archive. Unlike other searches, we employ a {\it blind} search for \OVI\
254: doublets, which is independent of {\it a priori} knowledge of the presence of other transitions such
255: as \Lya. In \pI\ we reported results on the statistics of the \OVI\ absorbers. The major results
256: were: a) a measurements of the number of absorbers per unit redshift, $d\,{\cal N}(W\ge 30\,\mA)/dz
257: = 10.4 \pm 2.2$; b) a measurement of the cosmological mass density of the \Oion\ gas, $\OmegaOVI\,h
258: = (1.7 \pm 0.3) \times10^{-7}$; c) $<5$\% of \OVI\ absorbers originate in underdense regions that do
259: not show a significant trace of \HI; d) \HI\ column densities of \OVI\ absorbers span more than 5
260: orders of magnitude, and a moderate correlation exists between \NHI\ and \NOVI; and e) the number
261: density of \OVI\ absorbers along a given line of sight appears to be inversely correlated with the
262: number density of \HI\ absorbers. In this paper we present our catalogue of \OVI\ absorbers upon
263: which the results of \pI are based. We also address the physical conditions of the \Oion\ bearing
264: gas. The nature of the IGM, the \OVI\ absorbers in particular, and their relation to the WHIM is an
265: area that has seen much recent progress. At least two other groups have contemporaneously reported
266: results of similar analyses. We refer the interested reader to the works of
267: \citet{tripp-etal-08-OVI} and \citet{danforth-shull-08-OVI}. Specifically, see
268: \citet{tripp-etal-08-OVI} Sec~\ref{sec: not_recovered} for comments on the differences between both
269: works.
270:
271: We recall the description of our search and selection technique from \pI\ in Sec~\ref{sec: data},
272: presenting the full table of absorbers. We discuss individual lines-of-sight, and present the
273: measured quantities for each system, in Sec~\ref{sec: los}. In Sec~\ref{sec: not_recovered} we
274: discuss those systems reported in \citet{tripp-etal-08-OVI} that are not accepted by our selection
275: criteria. The physical properties of the ionized gas selected via \OVI\ absorption are discussed in
276: Sec~\ref{sec: properties}. Sec~\ref{sec: discussion} contains a summary and concluding remarks.
277:
278:
279: \section{Data and Catalogue}
280: \label{sec: data}
281: \subsection{STIS Data}
282: Our data were drawn from the STIS data archive\footnote{CALSTIS v2.23 (2006 Oct 06)}. The data and
283: search technique is described in \pI\ (see in particular, Sec~2.1 for details). For completeness, we
284: repeat parts of that description here. We chose all data with sufficient signal-to-noise ratio ($S/N
285: \geq 5$ per pixel), and resolution, as to be able to detect weak ($\Wr > 30\mA$) \OVI\
286: absorbers. This selection yielded 16 lines of sight with STIS E140M data. Table~\ref{tab:
287: journal_observations} describes these lines of sight.
288:
289: \OVI\ absorbers were selected on the basis of the equivalent width ratio of the doublet lines
290: alone. In order not to bias our search by the presence of other transitions, we do not consider
291: other associated lines until a later stage. This differs from the traditional technique, which
292: relies on {\it a priori} knowledge of absorber positions (usually from \Lya), and then searches for
293: possibly associated species. We began our search by Hanning smoothing the spectra, and identifying
294: all deviations $> 1.5\,\sigma$ from the continuum level. A gaussian profile was fit to each feature,
295: with the width restricted to $ 6 < \sigma < 300\kms$, where the lower limit is taken from the
296: spectrograph resolution, and the upper limit from consideration of the line width distribution of
297: known \OVI\ systems \citep[e.g.][]{heckman-etal-02, danforth-etal-05-OVI_baryon_census}. An
298: equivalent width was determined by directly integrating the data, with integration limits determined
299: from the gaussian width. All features with $<2\,\sigma$ significance were rejected.
300:
301: We consider each feature a putative \OVI\,1031 line and fit a doublet absorption model to the data
302: to determine whether the spectrum is consistent with the presence of both \OVI\,1031, 1037
303: lines. The doublet model requires both transitions to have the same line width, and have a line
304: strength ratio of $2:1$ (i.e.\ the ratio of the oscillator strength--wavelength product,
305: $f\,\lambda^2$, for the two transitions). Each system was also visually inspected to determine
306: whether the spectrum is consistent with the presence of a doublet. We accepted candidates if (a) the
307: \OVI\,1031 member has $> 3\,\sigma$ significance, and; (b) the ratio of rest-frame line
308: strengths---$\ROVI \equiv W_r(1031)/W_r(1037)$---lies between $1-\sigma_{\ROVI}$ and
309: $2+2\,\sigma_{\ROVI}$.
310:
311: Finally, we visually identified other species associated with each absorber, typically searching for
312: transitions of the ions H$^{0}$, Si$^{+, 2+, 3+}$, S$^{+, 2+}$, O$^{0, 2+, 3+}$, C$^{+, 2+}$, N$^{+,
313: 4+}$ and Ne$^{7+}$. The {\it vpfit}\footnote{http://www.ast.cam.ac.uk/\~{ }rfc/vpfit.html}
314: software was used to fit Voigt profiles to all components of all detected species in each
315: absorber. {\it vpfit} convolves the voigt profile with a gaussian line-spread function (LSF), whose
316: width is set by the instrument resolution. The STIS LSF has significant broad wings in some
317: configurations. We have tested that this difference does not affect our measurements of $N$ and $b$
318: (i.e. the differences are much smaller than the error in the measured values). The number, position
319: and initial values for the absorption components was assessed initially by eye. We performed a
320: minimum-$\chi^2$ analysis that includes multiple components; new components were added and fit
321: iteratively until either the normalized $chi^2$ did not decrease, or the newly added component
322: became ill-constrained (error-bars for the best-fit parameters were greater than the best-fit values)
323: by the data. For \OVI\ and \HI, this process is typically facilitated by the presence of multiple
324: transitions. Section~\ref{sec: los} has details of the component structure for each absorber. The
325: fits were used to evaluate the total column density for each absorber, and are given in
326: Table~\ref{tab: ovi_absorbers}. Fit results for the individual components that comprise each
327: absorber are given in Tables~\ref{tab: 3C249.1_transitions}--\ref{tab: Ton28_transitions}. In our
328: fitting, we employed the latest version of the standard atomic data distributed with {\it
329: vpfit}. These data are primarily from the compilation of \citet{morton-03}, with some more recent
330: updates included.
331:
332:
333: \subsection{A Catalogue of Random \OVI\ Absorbers at $0.12 < z < 0.50$}
334:
335: The final catalogue of \OVI\ doublet systems is given in Table~\ref{tab: ovi_absorbers}. The table
336: lists the line-of-sight and redshift of the absorber, typically the redshift of the strongest \OVI\
337: component (columns 1 \& 2). Rest-frame equivalent widths ($W_r$), and errors ($\sigma_{W_{r}}$), are
338: reported in units of \mA\ (columns 3--6), for both lines of the \OVI\ doublet. The ratio of
339: equivalent widths, \ROVI, and associated error are listed in columns (7) and (8). Finally, columns
340: (9) and (10) give the total \OVI\ column density for the absorber, which is the sum of the
341: individual components. Individual component fitting results are given in the following section.
342:
343: \placetable{tab: ovi_absorbers}
344:
345: \section{Individual Lines of Sight}
346: \label{sec: los}
347: For each absorber in the following sub-sections, we present the results of our profile fitting in
348: the associated figures and tables. Spectra are unbinned, and profile fits are overlayed in solid
349: (blue). The error spectrum is plotted as a solid line at the bottom of each panel (red), also
350: unbinned. The dot-dashed (green) lines indicate the continuum and zero flux levels. Component
351: positions are marked above the spectra with the (red) solid ticks. The tables list, for each
352: absorber along each line of sight, the ion, component position in redshift ($z$) and velocity offset
353: ($v$) (from the absorber redshift), and the profile fit Doppler parameter ($b$) and column-density
354: (\logN), and their associated errors. The flags denote upper (U) and lower (L) limits, or uncertain
355: features (Z). Note that errors are not reported for upper and lower limits.
356:
357:
358: \subsection{3C\,249.1}
359: The QSO 3C\,249.1 lies at $\zqso = 0.3115$, and we searched for \OVI\ absorption along the line of
360: sight from $0.122 < \zabs < 0.2885$. We identify a single \OVI\ doublet system along this line of
361: sight.
362:
363: \noindent $\zabs = 0.24676$ (Figure~\ref{fig: 3C249.1_z0.24676}; Table~\ref{tab:
364: 3C249.1_transitions})---This system shows a clear detection of both \OVI\ lines, strong \Lyab, and
365: tentative \SiIII\,1206. The \OVI\,1037 line shows warm pixels on the very edges of the line wings,
366: which do not affect our line fits. The blue wing of the \Lyb\ line partially blends with another
367: strong line, which we tentatively identify as \Lya\ at $z = 0.0517$. We cannot check this
368: assignment, as the corresponding \Lyb\ line for this putative assignment is below our wavelength
369: range. In the unblended region, the model fit to the data is good. The \Lyc\ line is on the edge of
370: the Galactic \Lya\ trough, and was excluded from our fits. The \SiIII\ line is very weak and offset
371: from the \OVI\ and \HI\ absorption ($\Delta\,v = 11\pm3\kms$); its identification is thus
372: uncertain. With the well aligned \OVI\ and \HI\ components, we derive $\bnt=26.1\kms$ and $\logT =
373: 4.7$; see Sec~\ref{sec: gas_temperature} for details.
374:
375: \subsection{3C\,273}
376: The STIS data for 3C\,273 ($\zqso = 0.1583$) are of excellent quality, but offer only a short
377: path-length over which to detect \OVI\ doublet systems ($0.1144 < \zabs < 0.1417$). The lowest
378: redshift absorber in our sample, at $\zabs = 0.12003$ is detected along this line of sight.
379:
380: \noindent $\zabs = 0.12003$ (Figure~\ref{fig: 3C273_z0.12003}; Table~\ref{tab:
381: 3C273_transitions})---This narrow, weak, single-component absorber is the lowest redshift absorber
382: in our sample, and is detectable at such a low wavelength ($\lambda_{obs} = 1156\A$) only due to the
383: high quality data for the 3C\,273 sightline. For both \OVI\ lines, while noisy, the data show
384: corresponding profiles, and the model fits are satisfactory for both lines. The line strengths from
385: direct integration of the data, are mismatched (the equivalent width ratio is $\sim 1.0 \pm 0.4$),
386: but neither line can be weak \Lya\ (the system is blueward of the Galactic \Lya\ line). Of the \HI\
387: lines, only \Lya\ is present in our data at such low redshifts, and the line is unsaturated and well
388: fit by a single component. Due to the simple structure, we are able to derive $\bnt = 6.6\kms$ and
389: the gas temperature, $\logT = 4.5$; see Sec~\ref{sec: gas_temperature} for details.
390:
391: \subsection{3C\,351.0}
392: 3C\,351.0 lies at $\zqso = 0.3716$, giving a usable path-length for detecting \OVI\ systems $0.1309
393: < \zabs < 0.3483$. We detect only a single \OVI\ doublet along this line of sight.
394:
395: \noindent $\zabs = 0.31659$ (Figure~\ref{fig: 3C351.0_z0.31659}; Table~\ref{tab:
396: 3C351.0_transitions})---This system shows a complex structure, with three well-defined \OVI\
397: components present, and corresponding \HI\ profiles. The \Lya\ line is saturated, but the \Lyb\
398: line shows three components that are well aligned with the \OVI. The weak \Lyc\ line is noisier and
399: less well fit by this \HI\ model; the \Lyd\ line is contaminated by Galactic \SII\,1250 absorption.
400: No \CIII\,977 is observed; the strong, putative \CIII\,977 line which aligns with the red-most \OVI\
401: component is \HI\,937 at $z = 0.37193$, associated with the QSO host. As all three \OVI\ components
402: are well-aligned with the three \HI\ components, we attempt to derive the gas temperature and
403: non-thermal broadening. For the components at $v = -3, -53\kms$ we derive $\bnt = 23.0, 19.5\kms$
404: and $\logT = 4.1, 4.8$ respectively. For the component at $v = +62\kms$, we cannot find a solution
405: since $\bOVI > \bHI$; at the $1\,\sigma$ level this component is consistent with a system whose
406: Doppler parameter is entirely dominated by non-thermal broadening. Sec~\ref{sec: gas_temperature}
407: contains further discussion of this case.
408:
409:
410: \subsection{H\,1821+643}
411: \label{subsec: H1821+643}
412: The H\,1821+643 ($\zqso = 0.297$) line has been studied extensively in terms of intervening
413: absorbers. \citet{tripp-etal-98-Lya-galaxies} used GHRS and galaxy redshifts to study the \Lya\
414: absorbers. \citet{tripp-etal-00-H1821+643} followed up with STIS observations, focusing on \OVI\
415: absorption, complemented by \citet{oegerle-etal-00-H1821+643-FUSE} with FUSE observations. The STIS
416: data allow us to search for \OVI\ absorbers between $0.1144 < \zabs < 0.2741$. We uncover four
417: absorbers at $\zabs = 0.22496, 0.22638, 0.24532, 0.26656$.
418:
419: \noindent $\zabs = 0.22496$ (Figure~\ref{fig: H1821+643_z0.22496}; Table~\ref{tab:
420: H1821+643_transitions})---This absorber and the next (H\,1821+643; $\zabs = 0.22638$) are
421: separated by only $\sim350\kms$, but the two absorbers appear to be physically distinct systems (as
422: opposed to components of the same absorption system). The \OVI\ absorption consists of a broad,
423: strong component, with a very weak, narrow component on the red edge. The associated \Lyab\ lines
424: are strongly saturated, while the \Lycd\ lines show some saturation. Three separate \HI\ components
425: are included in the fit, but \NHI\ is a lower limit; the two strong components are saturated in the
426: three lowest order Lyman lines, while the \Lyd\ line is very noisy. The \CIII\,977 profile is
427: similarly complex, with several possible weak components evident around the three main, strong
428: components, (at least two of which show signs of saturation). There are three well-defined but
429: unsaturated \SiIII\,1206 components corresponding to the strong \CIII\,977 absorption lines. The
430: $v=0\kms$ component of \SiIII\ and the $v = -6\kms$ \CIII\ absorption component, while aligned with
431: the main \OVI\ absorption, cannot arise in the same gas phase, since e.g.\ $\bSiIII \ll \bOVI$, and
432: we require $0.8 < \bSiIII/\bOVI < 1.0$. These limits are discussed in more detail in Sec~\ref{sec:
433: gas_temperature}. At least two weak \SiIV\,1393 components are also present, which align with the
434: two strongest \CIII\ components, but are too weak to be detected in the weaker \SiIV\,1402 doublet
435: transition.
436:
437:
438: \noindent $\zabs = 0.22638$ (Figure~\ref{fig: H1821+643_z0.22638}; Table~\ref{tab:
439: H1821+643_transitions})---The weak absorber at $z=0.22638$ is separated from the strong system at
440: $z=0.22496$ by only $\sim350\kms$. The \OVI\ doublet is well fit by a weak, narrow, single component
441: absorption profile. \Lya\ absorption at $v = -53\kms$ shows a similar single component profile.
442: There is some evidence of weak \CIII\,977 at the same velocity as the \Lya, but we could not obtain
443: a satisfactory fit, and better quality data are needed to confirm this claim.
444:
445: \noindent $\zabs = 0.24532$ (Figure~\ref{fig: H1821+643_z0.24532}; Table~\ref{tab:
446: H1821+643_transitions})---There are two weak \OVI\ components in this system, with a single weak,
447: broad \HI\ component. There is possible \NV\ present, corresponding the red-most \OVI\ component, but
448: the weak \NV\,1242 line is totally obscured by Galactic \CIV\,1548, and we regard this
449: identification as uncertain.
450:
451:
452: \noindent $\zabs = 0.26656$ (Figure~\ref{fig: H1821+643_z0.26656}; Table~\ref{tab:
453: H1821+643_transitions})---This absorber is relatively simple, with well-aligned, single-component
454: \OVI\ and \HI\ lines. The \OVI\ and \HI\ line centroids differ by only 4\kms, and the good alignment
455: permits us to derive gas temperature and non-thermal broadening of $\logT = 4.9$ and $\bnt =
456: 24.4\kms$ respectively. Sec~\ref{sec: gas_temperature} has the details of this derivation.
457:
458:
459: \subsection{HE\,0226$-$4110}
460: The STIS spectrum of HE\,0226$-$4110, at $\zabs = 0.495$, allows us to search for intervening \OVI\
461: absorbers in the interval $0.1154 < \zabs < 0.4707$. We identified \OVI\ doublets at $\zabs =
462: 0.20702, 0.32639 0.34034, 0.35525$. This line of sight has also been studied by
463: \citet{savage-etal-05-HE0226-4110-NeVIII}, who focused on the $\zabs = 0.20702$ system, and
464: \citet{lehner-etal-06-HE0226-4110} who studied the full path length using STIS and FUSE data.
465:
466:
467: \label{subsec: HE0226-4110}
468: \noindent $\zabs = 0.20702$ (Figure~\ref{fig: HE0226-4110_z0.20702}; Table~\ref{tab:
469: HE0226-4110_transitions})---The strong absorber at $\zabs = 0.20702$ has associated \HI,
470: \CIII\,977, \SiIII\,1206 and \NV. The \OVI\,1037 transition differs from the single-component
471: structure seen in \OVI\,1031, which is probably a result of bad pixels in the \OVI\,1037 region. We
472: measure only a lower limit on \NHI; the \Lya\ and \Lyb\ transitions are saturated, while the \Lyc\
473: line is contaminated by hot pixels. \SiIII\,1206 has two components blueward of the fiducial \OVI\
474: position, which roughly correspond to the saturated \CIII\,977 line. The \NV\ lines are weak, and
475: uncertain. \citep{savage-etal-05-HE0226-4110-NeVIII} detected \NeVIII\ aligned with the \OVI\
476: absorption in FUSE data. They show this system is likely collisionally ionized, which is consistent
477: with the single broad-component in \OVI\ that we observe.
478:
479:
480: \noindent $\zabs = 0.32639$ (Figure~\ref{fig: HE0226-4110_z0.32639}; Table~\ref{tab:
481: HE0226-4110_transitions})---We tentatively identify an \OVI\ doublet at $\zabs = 0.32639$ as the
482: only system in our sample that shows no sign of \HI\ absorption. Both \OVI\ lines are well fit by a
483: single doublet model, while the position of any putative \Lya\ line is in the red-most portion of
484: the STIS wavelength coverage, making the data noisier than the \OVI\ region, as can be seen in
485: Figure~\ref{fig: HE0226-4110_z0.32639}. Fixing the redshift and expected Doppler parameter from the
486: \OVI\ profile, we set an upper limit on the \HI\ column density $\NHI < 12.5$. No other transitions
487: are present in this system. We note that \citet{lehner-etal-06-HE0226-4110} do not report
488: identifications for either lines (see e.g.\ their Figure~3). We also note that the \OVI\,1037 line
489: is detected at low significance (only $2\,\sigma$), and emphasize that caution is required
490: interpreting this system as \HI\ free. We suggest that further observations would be very valuable
491: to confirm this system, and whether it is \HI\ free.
492:
493: \noindent $\zabs = 0.34034$ (Figure~\ref{fig: HE0226-4110_z0.34034}; Table~\ref{tab:
494: HE0226-4110_transitions})---The \OVI\,1037 line in this system is contaminated by an unidentified
495: metal line at $v = -26\kms$, which is excluded from the fit. The \Lya\ line aligns well with the
496: \OVI\ absorption, but is in a poor-quality region of the spectrum, and the agreement between the
497: \Lya\ and \Lyb\ profiles is poor. A weak, narrow \CIII\,977 line is also detected, which aligns
498: well with the \OVI\ and main \HI\ component. The Doppler parameter of the \CIII\ line, \bCIII, is
499: significantly smaller than that of the \OVI\ component. This may be explained in several ways: the
500: \CIII\ line is not real---possible but unlikely, given its precise alignment with the \OVI\ and \HI\
501: positions; the \OVI\ Doppler width is over-estimated, either by unresolved components, or due to
502: noise in the profile; finally, the \OVI\ and \CIII\ absorption may arise in physically distinct gas
503: clouds. For the well-aligned \OVI\ and \HI\ components, we calculate $\logT = 4.0$ and $\bnt =
504: 16.5\kms$. Sec~\ref{sec: gas_temperature} has the details of this calculation.
505:
506:
507: \noindent $\zabs = 0.35525$ (Figure~\ref{fig: HE0226-4110_z0.35525}; Table~\ref{tab:
508: HE0226-4110_transitions})---There is only weak \OVI\ and \HI\ at $\zabs = 0.35525$, with both
509: transitions precisely aligned. The \OVI\,1031 line shows a weak contaminating component or feature,
510: although it is too weak to be detected in the \OVI\,1037 transition if it is real. We fit this
511: component as a contaminating \HI\ line, and the resulting model is a good fit to both \OVI\
512: transitions. As with the $z=0.34034$ system, the \Lya\ profile is quite noisy, but the \HI\,1215,
513: 1025 regions are fit to within the noise. The resulting \HI\ position is closely matched to the
514: \OVI\ redshift. Using this close match, we derive gas temperature $\logT = 4.3$ and non-thermal
515: broadening $\bnt = 22.2\kms$; see Sec~\ref{sec: gas_temperature}.
516:
517: \subsection{HS\,0624+6907}
518: The QSO HS\,0624+6907 at $\zqso = 0.370$ offers a path-length for \OVI\ absorption $0.1222 < \zabs <
519: 0.3464$. We detect two \OVI\ absorbers, at $\zabs = 0.31796, 0.33984$.
520:
521: \noindent $\zabs = 0.31796$ (Figure~\ref{fig: HS0624+6907_z0.31796}; Table~\ref{tab:
522: HS0624+6907_transitions})---This absorber is seen in only three lines---\OVI\,1031, 1037 and \Lya;
523: \Lyb\ is not present. The \Lya\ and \OVI\ positions are offset ($15 \pm 5\kms$). The \OVI\,1031
524: line shows evidence of some warm or noisy pixels.
525:
526: \noindent $\zabs = 0.33984$ (Figure~\ref{fig: HS0624+6907_z0.33984}; Table~\ref{tab:
527: HS0624+6907_transitions})---The significance of the \OVI\ in this system is very weak. It meets
528: our formal definition of a significant \OVI\,1031 line, and an \OVI\,1037 profile that is consistent
529: with the \OVI\,1031 transition. We thus accept this system, noting that the \OVI\ doublet is
530: uncertain. Other transitions are clearly present, with strong \HI, and possibly weak \NV. The \Lya\
531: line is saturated and the fit is sub-optimal in the red wing, but good fits are possible to the
532: \Lyb, \Lyc\ and \Lyd\ lines. The \Lyb\ profile is partly blended with another line, whose
533: identification is uncertain---if this contaminant is \Lya, the expected \Lyb\ line is at a low
534: enough wavelength that it would not be detected in the noisy STIS data. The blended portion of the
535: spectrum is excluded from our fits. There is a very weak, broad line at the expected position of
536: \NV\,1238, but is too weak to be confirmed in the \NV\,1242 transition.
537:
538:
539: \subsection{PG\,0953+415}
540: Along the line of sight to PG\,0953+415 ($\zqso = 0.239$), we can detect \OVI\ absorption in the
541: range $0.1144 < \zabs < 0.2163$. A single system is detected, at $\zabs = 0.14232$, which has been
542: previously discussed by \citet{tripp-savage-00-PG0953+415_z0.1423}.
543:
544: \noindent $\zabs = 0.14232$ (Figure~\ref{fig: PG0953+415_z0.14232}; Table~\ref{tab:
545: PG0953+415_transitions})---This strong \OVI\ absorber falls in a noisy part of the STIS spectrum
546: of PG\,0953+415. The \OVI\,1037 line is stronger than expected, based on the \OVI\,1031 line
547: strength, but this may be due to noise in the spectrum. \Lya\ is present, and well aligned with the
548: \OVI, but we are unable to obtain the gas temperature, \logT, and \bnt, since $\bOVI > \bHI$; see
549: Sec~\ref{sec: gas_temperature} for further discussion. Due to noise in the \Lyb\ portion of the
550: spectrum, and blending of the \Lyb\ profile with the QSO host \HI\,949 line at $z=0.2335$, we fit
551: only the zero-velocity \HI\ component. There is a hint of \CII\,1334, but it is very weak and
552: uncertain. This system has been discussed by \citet{tripp-savage-00-PG0953+415_z0.1423}.
553:
554: \subsection{PG\,1116+215}
555: The PG\,1116+215 ($\zqso = 0.1765$) sightline has a single \OVI\ absorber detected at $\zabs =
556: 0.13847$ from the available path-length between $0.1144 < \zabs < 0.1536$.
557:
558:
559: \noindent $\zabs = 0.13847$ (Figure~\ref{fig: PG1116+215_z0.13847}; Table~\ref{tab:
560: PG1116+215_transitions})---Previously analysed by \citet{sembach-etal-04-PG1116+215} this absorber
561: has noisy \OVI\ due to the low STIS efficiency at the observed wavelength ($\lambda_{\scOVI\,1031} =
562: 1174.8\A$). Strong, saturated \HI\ absorption is observed. Well-aligned metal lines of \SiII,
563: \SiIII, \SiIV, \CII, \NII\ and possibly \NV, are all observed. The \Lyab\ lines in our STIS data are
564: heavily saturated, and we cannot simultaneously provide a good fit to both \HI\ lines. Using a
565: curve-of-growth analysis and the weak Lyman-limit, \citet{sembach-etal-04-PG1116+215} measured
566: $\logNHI \sim 16.2$ for this absorber. They also concluded that photoionization models at a single
567: ionization parameter, or collisionally ionization models at a single temperature, cannot explain all
568: the metal lines observed.
569:
570:
571: \subsection{PG\,1216+069}
572: The STIS data for PG\,1216+069 ($\zqso = 0.3313$) allow us to search for \OVI\ doublets in the range
573: $0.1309 < \zabs < 0.3078$. We detect only a single intervening \OVI\ absorber, at $\zabs = 0.28232$.
574:
575: \noindent $\zabs = 0.28232$ (Figure~\ref{fig: PG1216+069_z0.28232}; Table~\ref{tab:
576: PG1216+069_transitions})---The only \OVI\ system we detect towards PG\,1216+069, this absorber has
577: very weak \OVI\ absorption, along with very strong, saturated, multi-component \HI, allowing us to
578: set only a lower limit on \logNHI. We also detect saturated \CIII\,977, and strong \SiIII\,1206. In
579: the \CIII\ region, we fit only the zero velocity component. Other components may be {\it bona fide}
580: \CIII\ associated with the blue \HI\ component, but this is not clear. Precise knowledge of the
581: \HI\ component positions would be valuable for determining this.
582:
583: \subsection{PG\,1259+593}
584: The good quality STIS data for the line-of-sight PG\,1259+593 ($\zqso = 0.4778$) allow us to search
585: for weak absorbers over a large path-length $0.1144 < \zqso < 0.4533$. We detect two \OVI\ systems
586: at $\zabs = 0.21950, 0.25981$. This sightline has also been studied by
587: \citet{richter-etal-04-PG1259+593}.
588:
589:
590: \noindent $\zabs = 0.21950$ (Figure~\ref{fig: PG1259+593_z0.21950}; Table~\ref{tab:
591: PG1259+593_transitions})---This absorber contains two principle \OVI\ components, both of which
592: are well resolved. The blue component suffers from several bad pixels, which were excluded from the
593: fitting regions. The \HI\ components are aligned with the \OVI, although the stronger component is
594: saturated even in the noisy, higher order \HI\,949 transition. The blue edge of the \Lyb\ profile
595: blends with Galactic \SII\,1250, but does not affect our fitting. We detect weak \SiIII\,1206 and
596: saturated \CIII\,977, both aligned with the zero velocity \OVI\
597: component. \citet{richter-etal-04-PG1259+593}, who studied all the systems on this sightline,
598: measured $\logNHI = 15.2$.
599:
600: \noindent $\zabs = 0.25981$ (Figure~\ref{fig: PG1259+593_z0.25981}; Table~\ref{tab:
601: PG1259+593_transitions})---The multi-component absorber at $z=0.25981$ has two main \OVI\
602: components: a narrow component at $-44\kms$ separation from the broad zero-velocity component. The
603: best fit \HI\ profile has two components, well aligned with the \OVI\ components. We discount the
604: reality of a putative weak \SiIII\,1206 transition at $\sim +40\kms$. We attempted to derive \logT\
605: and \bnt\ for both components (Sec~\ref{sec: gas_temperature}), but were only successful only for
606: the \OVI/\HI\ pair at $-44\kms$. The zero-velocity \OVI/\HI\ component pair has $\bOVI > \bHI$, and
607: is discussed further in Sec~\ref{sec: gas_temperature}. For the component at $v = -44\kms$, we
608: measure $\bnt = 13.9\kms$ and $\logT = 4.5$.
609:
610:
611: \subsection{PHL\,1811}
612: The PHL\,1811 ($\zqso = 0.1917$) sightline contains a single \OVI\ absorber at $\zabs = 0.15786$ in
613: the available range $0.1144 < \zabs < 0.1690$. \citet{jenkins-etal-05-PHL1811_z0.0809} have studied
614: this sightline, focusing on the LLS at $\zabs = 0.0809$.
615:
616: \label{sec: phl1811}
617: \noindent $\zabs = 0.15786$ (Figure~\ref{fig: PHL1811_z0.15786}; ---This \OVI\ system is very
618: weak. The $\OVI\,1031$ line is obvious, but the $\OVI\,1037$ line is detected at only $\sim
619: 2\,\sigma$. Nevertheless, the data show a profile very consistent with the stronger line of the
620: doublet, and the fit to both lines is good. Further, this system is blueward of the Galactic \Lya\
621: line, so the \OVI\,1031 line cannot be ascribed to \Lya\ absorption. The \HI\ in this system is
622: harder to quantify. There are two components at the expected position for \Lya\ at this
623: redshift. Both are well fit, but the narrow component is conspicuous given the broad nature of the
624: \OVI. Our fits are consistent with those of \citet{tripp-etal-08-OVI}, but further investigation
625: reveals the narrow component to be \OI\,1302 in the $z=0.0809$ Lyman-limit system (LLS)
626: \citep{jenkins-etal-05-PHL1811_z0.0809}. Our fit parameters for the narrow component agree well
627: with measurements of \citeauthor{jenkins-etal-05-PHL1811_z0.0809}. Further, there is no
628: corresponding \Lyb\ feature, which would be expected if the narrow component were \HI. The broad
629: \Lya\ component is too weak for \Lyb\ absorption to be detectable. The broad \OVI\ and \HI\
630: components are well aligned, and we derive $\logT = 5.1$ and $\bnt = 40.5\kms$ (Sec~\ref{sec:
631: gas_temperature}). This is the only system for which we derive a temperature greater that
632: $10^5\K$, and even at this temperature, \OIII\ and \OIV\ dominate the ionization state of Oxygen by
633: several orders of magnitude in CIE models. Both species, however, have absorption lines which are
634: shortward of the Lyman limit for the $z=0.0809$ LLS, and so cannot be observed.
635:
636:
637: \subsection{PKS\,0312$-$77}
638: The data for the PKS\,0312$-$77 ($\zqso = 0.223$) allow a path-length for our doublet search $0.1241
639: < \zabs < 0.1999$. We detect the strongest \OVI\ absorber in our survey in these data, at $\zabs =
640: 0.20275$, which is part of a partial LLS.
641:
642: \noindent $\zabs = 0.20275$ (Figure~\ref{fig: PKS0312-77_z0.20275}; Table~\ref{tab:
643: PKS0312-77_transitions})---The STIS spectrum of the sightline toward PKS\,0312$-$77 was obtained
644: under proposal id 8651 (PI Kobulnicky). This line of sight contains the strongest \OVI\ absorber of
645: our entire sample: a partial Lyman Limit System at $\zabs = 0.20275$ with $15.9 < \logNHI < 18.4$,
646: which is confirmed by a brief inspection of a FUSE spectrum. This system has two groups of
647: components, each having a complex structure. The main component group has strong, broad \OVI\
648: absorption and covers the range $-100 \simlt v \simlt 100\kms$, while the weaker group shows
649: evidence of weak \OVI\ ($-300 \simlt v \simlt -100\kms$).
650:
651: Fitting for this absorber was complicated by the fact that many components are saturated,
652: particularly in stronger transitions such as \Lyab, \CIII\,977 and \SiIII\,1206. In general, we used
653: weaker, low ion transitions (typically \NII\ and \CII) to fix the redshifts of the components
654: (excluding \OVI), and initialized fits with fixed redshift, and initial $b$-value and column density
655: guided by weaker absorbers. Degeneracies and poorly constrained parameters naturally result from
656: fitting saturated absorbers; these are indicated by large formal parameter errors, and the fit
657: values are typically indicative of lower limits only.
658:
659: For the main component group, we fit each species separately. Five components were first identified
660: from low the ion transitions \CII\,1036, \NII\,1038 and \SiII\,1193, 1260, which all show the same
661: structure. These redshifts were fixed and fits performed. The \OVI\, line, while probably requiring
662: more than one component for the measured breadth, shows little evidence of individual component
663: structure, and a single component provided adequate fits. Only three \HI\ transitions are present in
664: the STIS data, all of which are strongly saturated, resulting in mostly \NHI\ lower limits. \SiII\
665: transitions are well fit, except for \SiII\,1190, which shows possible contamination from Galactic
666: \CI*\,1190 at the red side, and Galactic \CI\,1190 at the blue side. The stronger \SiIII\,1206 and
667: \SiIV\ transitions show saturation, with data quality at \SiIV\ deteriorating markedly due to
668: dropping STIS sensitivity (\SiIV\ is at the extreme red end of the wavelength coverage). \SII\
669: shows weak absorption with a different profile than other low ions---only the two strongest
670: components in the main body of the absorber are present; better quality data would be valuable to
671: confirm that this detection is correct, since the weaker \SII\,1259 transition is barely detected.
672: For \CII, the $\lambda\,1036$ line is well fit, but the $\lambda\,1334$ line suffers from the same
673: decrease in sensitivity as \SiIV, and is poorly fit, as is the strong \CIII\,977 transition. \NII\
674: is well fit; in the blueshifted part it is strongly saturated. Blueward of this saturated \NII\
675: absorption we see Galactic \OI\,1301 at $\sim -380\kms$, and at $\sim 140\kms$ we see Galactic
676: \SiII\,1304. Finally, \NV\ is present in multiple components, but very weak, and we use only the
677: stronger \NV\,1238 region for the fitting. Figure~\ref{fig: PKS0312-77_z0.20275_2} shows the predicted
678: profile of the \NV\,1242 region, based on the fits.
679:
680: \subsection{PKS\,0405$-$12}
681: PKS\,0405$-$12 ($\zqso = 0.5723$) is the highest redshift QSO in our sample, allowing a path-length
682: to search for \OVI\ absorbers $0.1241 < \zabs < 0.5478$. We detect four \OVI\ systems, at $\zabs =
683: 0.15597, 0.18291, 0.36333, 0.49514$. This line of sight has also been studied previously by
684: \citet{chen-prochaska-00-PKS0405-12_z0.167, prochaska-etal-04-PKS0405-12,
685: prochaska-etal-06-PKS0405-12_galaxies}
686:
687: \noindent $\zabs = 0.16697$ (Figure~\ref{fig: PKS0405-12_z0.16703}; Table~\ref{tab:
688: PKS0405-12_transitions})---The PKS\,0405$-$12 sightline has been studied by
689: \citet{prochaska-etal-04-PKS0405-12}, and the partial Lyman-limit system at $z=0.167$ by
690: \citet{chen-prochaska-00-PKS0405-12_z0.167}. The system has strong, saturated \HI, and two \OVI\
691: components---one broad and strong; one weak and narrow. \citet{prochaska-etal-04-PKS0405-12}
692: estimate $\logNHI = 16.45$ from an analysis of the flux decrement shortward of the 912\A\
693: Lyman-limit in FUSE data for this sightline.
694:
695: The STIS data exhibit strong lines of \SiII, \SiIII, \SiIV, \CII, \NII\ and \OI, some of which
696: appear saturated. In general, the low ion transitions are well aligned, and details are given in
697: Table~\ref{tab: PKS0405-12_transitions}. Due to the heavily saturated \HI\ lines, only the weak
698: \OVI\ feature at $-110\kms$ can be matched well with \HI\ absorption components, and the resulting
699: parameters $\bnt = 7.0\kms$ and $\logT = 3.7$ are derived in Sec~\ref{sec: gas_temperature}.
700:
701: \noindent $\zabs = 0.18291$ (Figure~\ref{fig: PKS0405-12_z0.18291}; Table~\ref{tab:
702: PKS0405-12_transitions})---The absorber at $\zabs = 0.18291$ consists of two separate components,
703: separated by $\sim 90\kms$, both of which are reported in \citet{prochaska-etal-04-PKS0405-12}. In
704: our STIS data, we have only the \Lya\ line corresponding to these absorbers (\Lyb\ blends into the
705: Galactic \Lya\ trough), which is heavily saturated. We thus derive only lower limits to \logNHI,
706: consistent with \citet{prochaska-etal-04-PKS0405-12}. The two \OVI\ components are also well fit,
707: despite the \OVI\,1031 line appearing on the red edge of the Galactic \Lya\ absorption.
708:
709: \noindent $\zabs = 0.36333$ (Figure~\ref{fig: PKS0405-12_z0.36333}; Table~\ref{tab:
710: PKS0405-12_transitions})---This system has a well-fit \OVI\ doublet, and corresponding noisy \HI\
711: profile. \Lya\ is the only \HI\ transition we detect, but it is marred by hot pixels in the data, so
712: our resulting measurements are uncertain. Further, the \Lya\ line lies very close to the Galactic
713: \CI\,1656, blending with the fine structure \CI*\,1657.38 line. Lacking the detection of \Lyb, we
714: cannot obtain reliable \logNHI. We also detect weak \CIII.
715:
716: \noindent $\zabs = 0.49514$ (Figure~\ref{fig: PKS0405-12_z0.49514}; Table~\ref{tab:
717: PKS0405-12_transitions})---\citet{prochaska-etal-04-PKS0405-12} present a brief analysis of the
718: absorber at $\zabs = 0.49514$. At such a high redshift, the \Lya\ line is redshifted out of the STIS
719: bandpass, and the \Lyb\ and \Lyc\ transitions are weak and noisy. Both \OVI\ lines are contaminated
720: by hot pixels. In the region of the \OVI\,1037 line, Galactic \CIV\,1538 falls at $\sim
721: -90\kms$. Inspection of the \CIV\,1542 line shows that it does not significantly affect the \OVI\
722: line. \CIII\ is detected at $v = 0 \kms$, as are offset \OIII\,832 and \OIV\,787. Both the \OIII\
723: and \OIV\ transitions also show a component at $\sim 90\kms$. While there is no associated \HI\ with
724: these components at $v = 90\kms$, the exact alignment of both transitions argues that the absorption
725: is real (as opposed to, say, weak \HI). Due the noise in the data, we could not obtain an acceptable
726: fit to the \HI\ lines and we were forced to fix the \HI\ doppler-parameter. Higher quality data
727: would be very valuable for this system, both to obtain an accurate \HI\ column density, and confirm
728: the assumed doppler parameter for \HI.
729:
730:
731: \subsection{PKS\,1302$-$102}
732: We searched the STIS data for PKS\,1302$-$102 ($\zqso = 0.2784$) for \OVI\ absorbers in the range
733: $0.1183 < \zabs < 0.2558$, detecting two close absorbers at $\zabs = 0.22565, 0.22744$. This
734: sightline has also been studied by \citet{cooksey-etal-08-PKS1302-102}.
735:
736:
737: \noindent $\zabs = 0.22565$ (Figure~\ref{fig: PKS1302-102_z0.22565}; Table~\ref{tab:
738: PKS1302-102_transitions})---As with the H\,1821+643 absorber pair, these absorbers at $\zabs = 0.22565$
739: and $\zabs = 0.22744$ along the PKS\,1302$-$102 sightline are separated by only $\Delta v = 440\kms$. Also
740: similar to the previous example, this absorber pair consists of a strong and weak system, with the
741: weaker system at a slightly higher redshift (although in this case, the difference is far less
742: dramatic). The system at $\zabs = 0.22565$ consists of two, well defined \OVI\ components separated by
743: 28\kms. Both \OVI\ components have matching \HI\ absorption, detected primarily in \Lya. The two
744: strong lines at the expected positions of the \NV\ doublet are obviously unrelated \Lya\
745: absorbers. The temperature and non-thermal broadening can be determined for only the $v=0\kms$
746: component ($\logT = 4.4; \bnt = 12.5\kms$), since the \HI\ position for the component at $v=-28\kms$
747: has been fixed to that of \OVI\ in our fitting; see Sec~\ref{sec: gas_temperature} for details.
748:
749: \noindent $\zabs = 0.22744$ (Figure~\ref{fig: PKS1302-102_z0.22744}; Table~\ref{tab:
750: PKS1302-102_transitions})---The weaker of the two close \OVI\ absorbers towards PKS\,1302$-$102, this
751: system has only very weak transitions of \OVI\,1031, 1037 and \Lya. The \OVI\,1031 line is close to,
752: but unaffected by, a strong \Lya\ absorption line at $z=0.4224$. Both species are well aligned, and
753: suitable for temperature analysis in Sec~\ref{sec: gas_temperature}. We derive $\logT = 4.2$ and
754: $\bnt = 10.7\kms$.
755:
756: \subsection{Ton\,28}
757: The Ton\,28 ($\zqso = 0.3297$) line of sight hosts only one \OVI\ absorber at $\zabs = 0.27340$ in
758: the available path-length $0.1231 < \zabs < 0.3059$.
759:
760: \noindent $\zabs = 0.27340$ (Figure~\ref{fig: Ton28_z0.27340}; Table~\ref{tab:
761: Ton28_transitions})---The system towards Ton\,28 contains strong \HI\ absorption, and only weak
762: \OVI. The \OVI\,1031 line is detected at $>3\,\sigma$ significance, while the \OVI\,1037 is only
763: $\sim 2.5\,\sigma$. The \Lya\ absorption line is saturated, and blended with Galactic
764: \CIV\,1548. Comparison of the Galactic \CIV\,1548 and 1550 lines shows that the full \Lya\ profile
765: is blended, not just the red wing (which is apparent in Figure~\ref{fig: Ton28_z0.27340}). The \Lyb\
766: and \Lyc\ lines allow us to obtain a good \HI\ fit, and we include the \CIV\ absorption component
767: that is obvious in the \Lya\ wing. The discrepancy between the \Lya\ data and fit profile is can be
768: ascribed to the \CIV\ blending. No other associated transitions are detected. The alignment of the
769: \HI\ and \OVI\ positions makes this system useful for our temperature analysis in Sec~\ref{sec:
770: gas_temperature}. We have derived $\logT = 4.5$ and $\bnt = 20.2$.
771:
772:
773: \section{Comparison With Other Work}
774: \label{sec: not_recovered}
775: As noted in Sec~\ref{sec: introduction}, two similar contemporaneous studies have been conducted by
776: other groups \citep{tripp-etal-08-OVI, danforth-shull-08-OVI}. While differences exist in the
777: selection of \OVI\ systems in the different samples there is typically good agreement in measured
778: quantities for common systems. Here we specifically address systems reported by
779: \citep{tripp-etal-08-OVI} that are not accepted by our selection criteria. In general, this is a
780: difference between performing a {\it blind} search for the \OVI\ doublet, and looking for one of the
781: doublet lines in previously identified systems (e.g.\ \Lya\ absorbers). This associated \Lya\
782: technique accepts cases in which one or other of the \OVI\ doublet lines is masked by strong
783: absorption from ISM or IGM lines at a different redshift. Other systems simply fall outside our
784: wavelength range or are identified in different data (e.g.\ the \OVI\ doublet is identified in FUSE
785: spectra).
786:
787: We only consider absorbers from the \citet{tripp-etal-08-OVI} compilation that are within our
788: wavelength range, but are not recovered in our search. Table~\ref{tab: unconfirmed_systems}
789: summarizes these systems, and we comment on each system below. The \OVI\,1031 (top) and 1037
790: (bottom) spectral regions for each system are shown in Figures~\ref{fig:
791: OVI_rejected_1}~\&~\ref{fig: OVI_rejected_2}. Note that \citet{danforth-shull-08-OVI} have also
792: recently conducted a similar survey using both STIS and FUSE data. Finally, where our sample has
793: common systems with the samples of \citet{tripp-etal-08-OVI} and \citet{danforth-shull-08-OVI}, we
794: see no marked differences in the measured total column densities, or the component $b$-values and
795: column densities where similar component structures are fitted.
796:
797: \subsection{3C\,351.0}
798: \noindent $z = 0.21811$---This system shows broad absorption at the expected position of both \OVI\
799: doublet members, but is rejected because acceptable fits could not be obtained; the line strengths
800: of the putative doublet members are inconsistent at the $3-4\,\sigma$ level, with \OVI\,1037
801: stronger than \OVI\,1031. There is a putative \Lya\ absorber at this redshift, but its velocity is
802: inconsistent with the \OVI\ position ($\Delta\,v \sim 50\kms$). \CIII\ is obscured by Galactic
803: \SiII\,1190 and no \SiIII\ is detectable. It is possible that the \OVI,1037 line in this system is a
804: broad \Lya\ absorber \citep[see e.g.\ ][]{sembach-etal-04-PG1116+215}. We concur with
805: \citet{tripp-etal-08-OVI} that higher S/N observations would be very valuable to firmly establish
806: the nature of this system.
807:
808: \noindent $z = 0.22111$---This system is not considered because the \OVI\,1031 position is clearly
809: blended with heavily saturated Galactic \SiII\,1260 absorption. \Lya\ and \SiIII,1206 are present at
810: the expected wavelengths; \CIII\,977 blends with \SiII\,1193 at $z = 0$.
811:
812:
813: \subsection{H\,1821+643}
814: \noindent $z = 0.12143$---There is strong \Lya\ absorption at $\sim -80\kms$ from the nominal
815: position of this system, but our STIS data are far too noisy to detect \OVI\ absorption ($\rm{S/N}
816: \approx 2-3$ in this region), which is reported by \citet{tripp-etal-08-OVI} in FUSE data. The
817: strong absorption at the \OVI\,1037 position for this redshift is \Lyd\ at z=0.225
818: \citep{tripp-etal-01-H1821+643_z0.1212}.
819:
820: \noindent $z = 0.21331$---We initially considered this system, but rejected it due to mis-matched
821: line strength ratio. The \OVI\,1037 is stronger than the \OVI\,1031 line, and the strength ratio is
822: $R_{\scOVI} = 0.77 \pm 0.14$. As noted by \citet{tripp-etal-00-H1821+643}, the region of the
823: \OVI\,1037 line contains a strong Galactic \SII\,1259 at a velocity $v \simeq +100\kms$ with respect
824: to the absorber redshift, and a weaker Galactic \SII\,1259 feature at $\sim+30\kms$. This weaker
825: component partially blends with the putative \OVI\,1037 line. The weaker component is also seen in
826: \SII\,1253, and is likely associated with the intermediate-velocity cloud (IVC) in the Milky Way
827: known as the IV arch \citep{kuntz-danly-96-ComplexM}.
828:
829:
830: \subsection{HE\,0226$-$4110}
831: \noindent $z = 0.42670$---The data show a strong \OVI\,1031 line in this candidate, but there is no
832: corresponding \OVI\,1037 line (the $\lambda\,1037$ position does contain a narrow absorption spike
833: due to noise, which is reflected in the error array). \Lya\ is redshifted out of the STIS bandpass
834: and there is no \Lyb. Galactic \SiIV\,1393 blends with the expected position of \CIII\,977, and
835: there is also no \SiIII\,1206. We conclude that this system is not real.
836:
837:
838: \subsection{PG\,1216+069}
839: \noindent $z = 0.26768$---\OVI\,1037 blends with the heavily saturated \Lyb\ line from the absorber
840: at $z = 0.28232$, and hence this system cannot meet our criteria for a doublet search.
841:
842:
843: \subsection{PG\,1259+593}
844: \noindent $z = 0.31972$---We do not detect the \OVI\,1037 line in this system, although there is a
845: single-pixel negative noise spike (which is reflected in the error array). A model based on the
846: putative \OVI\,1031 line is not consistent with the data in the $\lambda\,1037$ region. The red wing
847: of the \OVI\,1037 region blends with a weak ($\sim 20\mA$) unidentified line
848:
849:
850: \subsection{PG\,1444+407}
851: \noindent $z = 0.22032$---The \OVI\,1031 line blends with Galactic \SII\,1259. The $z = 0$ \SII\
852: line exhibits the same structure in the \SII\,1253 transition, confirming its Galactic nature. Shallow,
853: broad \Lya\ may be seen, but the \OVI\,1037 is very weak and uncertain.
854:
855:
856: \subsection{PHL\,1811}
857: \noindent $z = 0.13240$---At the observed wavelength of this system ($\lambda_{\rm obs} =
858: 1168.5\A$), the spectrum is quite noisy. We detect a strong \OVI\,1031 line, and at higher
859: wavelengths the saturated, offset \Lya\ line is easily detected in higher S/N data. We do not,
860: however, see evidence of an \OVI\,1037 line, and thus reject this system.
861:
862:
863: \subsection{PKS\,0312$-$77}
864: \noindent $z = 0.15890$---The data for the PKS\,0312$-$77 sightline is quite noisy, especially in the
865: low-wavelength region about the \OVI\,1031 position. We see a well-fit \OVI\,1031, but this same
866: profile bears no resemblance to the data in the expected region of \OVI\,1037. Multi component \Lya\
867: is also seen, but the lack of \OVI\,1037 absorption leads us to reject this system.
868:
869: \noindent $z = 0.19827$---As with the lower redshift system described above, we detected \Lya\
870: (saturated in this case) and broad \OVI\,1031, but no evidence of \OVI\,1037. There is no
871: correspondence in the data between \OVI\,1031 and \OVI\,1037 regions, and the profile fit from the
872: \OVI\,1031 region is a poor fit to \OVI\,1037 region.
873:
874:
875: \subsection{PKS\,0405$-$12}
876: \noindent $z = 0.36156$---Initially detected in our doublet search, and published by
877: \citet{prochaska-etal-04-PKS0405-12}, there is considerable mis-match between the \OVI\,1031 and
878: \OVI\,1037 regions. \Lya\ is uncertain, but there may be a weak line in the wing of the strong
879: (\NHI\ = 15.1) \Lya\ line at $z = 0.3608$. Due to the profile mismatch, we are unable to confirm
880: this system.
881:
882: \subsection{PKS\,1302$-$102}
883: \noindent $z = 0.19159$---This system is another example that we are unable to confirm due to lack
884: of \OVI\,1037 absorption. A line at the position of \OVI\,1031 is present, as is saturated \Lya,
885: and possibly \SiIII\,1206. Given the strength of the apparent \OVI\,1031 line ($\sim 62\mA$), we
886: should expect to detect any \OVI\,1037 line, but no significant absorption exists, and the
887: \OVI\,1031 profile fit is a poor descriptor of the data in the \OVI\,1037 region. The independent
888: work of \cite{cooksey-etal-08-PKS1302-102} confirms this result; those authors report \Lya\ and
889: \OVI,1031, but do not detect \OVI,1037 (see their Table~3).
890:
891:
892: \subsection{Ton\,28}
893: \noindent $z = 0.13783$---A strong line at the position of \OVI\,1031 is seen in this absorber, but
894: no corresponding \OVI\,1037 is seen. At such low wavelengths, however, this is hardly
895: surprising---our STIS spectra have ${\rm S/N} < 3$ per pixel at the position of \OVI\,1037. A
896: profile fit to the \OVI\,1031 line does not follow the data in the \OVI\,1037 region. \Lya\ is
897: detected as part of a stronger, saturated \Lya\ system.
898:
899:
900: \section{Physical Properties of Ionized Gas Selected by OVI Absorption}
901: \label{sec: properties}
902: Having identified a sample of \OVI\ absorbers, and related transitions, we now consider these
903: absorbers in detail. We begin by matching individual absorption components of \OVI\ and \HI\ within
904: an absorber. These components trace physically distinct gas clouds, and we consider first the global
905: properties of these components. With the sample of absorption components well-aligned in velocity
906: space, we derive the gas temperature and non-thermal broadening, and examine whether the
907: temperatures are consistent with a collisional ionization origin. We also consider the ionization
908: state of the \Oion-bearing gas using photoionization models. We close with a brief comment on the
909: lack of \OVI-only systems.
910:
911: \subsection{\NOVI\ vs \NHI\ for Individual Components}
912: \label{subsec: components}
913: To consider any interpretation of the \Oion-bearing gas, we must determine whether the different
914: species observed in the absorbers arise from the same gas. It is well established from very high
915: resolution studies of the ISM that absorbers can be resolved into multiple (even many) individual
916: absorption components ($\Delta v < 1\kms$) \citep[e.g.][]{welty-etal-96-CaII-ISM,
917: welty-etal-99-SN1987A}. Since these components will trace physically distinct gas clouds, it is at
918: this component level that we must assess the physical nature of the gas. To define well-matched
919: components, we compare the positions of the \OVI\ and \HI\ components in each absorber. The
920: motivation for this matching is to constrain the temperature of the \Oion-bearing gas, a subject to
921: which we will return in the following sections. In principle, highly ionized species such as \NV\
922: would be the ideal candidate to compare to \OVI\ in deriving physical conditions, in particular the
923: gas temperature. \NV\ peaks in CIE at $\logT = 5.25$ \citep{sutherland-dopita-93-gas-cooling}, so
924: occupies the same temperature range as \OVI. Conversely, \HI\ is orders of magnitude more abundant,
925: and so is much more likely to result in a good component match. We therefore attempt to match all
926: \OVI\ components with corresponding \HI\ components. If the two component positions are equal to
927: within their $2\,\sigma$ positional uncertainties (to a maximum of the STIS resolution---$6-7 \kms$)
928: we consider them related.
929:
930: In Figure~\ref{fig: column_density_component_plots} we plot, for all well-matched components, the
931: \OVI\ column density as a function of the neutral hydrogen column density (c.f. Figure~19 of
932: \citealt{tripp-etal-08-OVI}). It is not surprising that no points lie in the upper left part of
933: this plot since: they would require high metallicity {\it and} optimum ionization parameter, and; we
934: see little evidence of a class of systems lacking (or with weak) \HI\ (See Section~\ref{subsec:
935: noHI}). The lower portion of the plot is empty due to a selection effect; in most of the survey,
936: our detection limit is $\Wr \gtrsim 30\mA$ which corresponds to $\logNOVI \gtrsim 13.5$.
937:
938: A range of physical conditions may be used to explain any particular point on this plot, since the
939: ionization parameter and metallicity are degenerate. Nevertheless, it is worthwhile considering the
940: range of conditions which are allowed by the data. In this vein, we have overplotted two sets of
941: model photoionization curves: models for a variable ionization parameter at fixed metallicity; and,
942: varying metallicity at fixed \logU. The dot-dashed line shows our "fiducial" condition of gas at
943: peak ionization parameter and $1/10^{th}$ solar metallicity. Since \NOVI\ peaks for fixed density
944: and metallicity at $\logU = -0.2$, any increase or decrease in \logU\ will act to shift this line to
945: the lower right (i.e. \NHI\ will be larger at fixed \NOVI), an effect that is clear in
946: Figure~\ref{fig: photo_model}. What, then, of the points above this line? Under the assumption of
947: photoionization equilibrium, we then infer that these systems {\it must} have higher metallicity
948: than our canonical 1/10$^{th}$ solar value. The dashed line shows the model curve for $\logU = -1.2;
949: \MH\ = 1.0$ and hence, at fixed $\MH\ = 1.0$, the data allow a range of ionization parameters of
950: $\sim 1\dex$. If we instead consider the variation in the points as solely a function of
951: metallicity at fixed (optimal) ionization parameter, we obtain the bounds delimited by the crosses
952: and square-dots, and a range of metallicities of $\sim 1.7\dex$.
953:
954: It is immediately apparent that the data do not follow the same global trend as the model curves
955: shown. Although we cannot determine a zero-point for the model curves, due to the degeneracy between
956: metallicity and \logU, the trend is nevertheless valid, and we should expect a variety of
957: metallicities and ionization conditions in the absorbers. The scatter in the observations of \NOVI\
958: vs \NHI\ can thus be understood as due to scatter in the metallicity and ionization state of the
959: gas\footnote{Other factors such as length scale, and varying redshift will also play a role here.}.
960:
961: The case of the 3C\,351.0 absorber at $z = 0.31659$ (open squares) is interesting, since the \OVI\
962: and \HI\ clearly separate into three individual components. No other species are detected in the
963: STIS band. If we posit a physical connection, based on the proximity of the components, we might
964: expect similar conditions for each gas cloud. The three components are shown as open squares in
965: Figure~\ref{fig: column_density_component_plots}. It is clear that no single set of ionization and
966: metallicity conditions can explain all the components, yet the temperatures do not admit of a
967: collisional origin (see Sec~\ref{sec: gas_temperature}). Likely we are simply seeing, for a constant
968: flux of ionizing photons, density variations along the line of sight. By contrast, we would require
969: variations of $\sim 1.0\dex$ to explain the observations by metallicity variations alone.
970:
971: \subsection{Gas Temperature}
972: \label{sec: gas_temperature}
973:
974: In the previous discussion, we have assumed that the matched components arise from photoionized
975: gas. Is this assumption justified? Having identified components of two different species that
976: likely arise from the same gas phase, we can use the Doppler parameters to derive an estimate of the
977: gas temperature, which are related through the relation $b^2 = \bnt^2 + 2kT/m$, where $\bnt^2$ is
978: the non-thermal component to the Doppler parameter (i.e. turbulence, bulk motion etc), and $k$, $T$
979: and $m$ are the Boltzmann constant, temperature and atomic mass as usual. The two equations thus
980: determine both the gas temperature and the non-thermal line broadening, and this well-established
981: fact has been used in the past to determine gas temperatures of absorbers \citep[e.g.\
982: ][]{chen-prochaska-00-PKS0405-12_z0.167, tripp-savage-00-PG0953+415_z0.1423}. It is also clear that,
983: in the absence of a turbulent component, the Doppler parameters will be related through the
984: square-root of the ratio of atomic masses. The distribution of \OVI\ and \HI\ $b$-values for
985: well-aligned absorption components is shown in Figure~\ref{fig:
986: doppler_parameter_component_plots}. The solid lines bound two limiting cases. At one extreme, for
987: $\bnt = 0$, we have simply that $\bHI = 4\times \bOVI$. The upper boundary in Figure~\ref{fig:
988: doppler_parameter_component_plots} shows the other limiting case, where $\bHI = \bOVI$. Formally,
989: we should not see absorbers above this limit, since that regime is non-physical. In practice, any
990: absorbers in this area imply either unresolved components, or that the \Oion\ and $\mbox{H}^0$ gas
991: is unrelated---either physically distinct\footnote{In which case the component alignment is purely
992: coincidental.}, or in a separate phase of a multi-phase system. While we have three components in
993: our sample that lie above this limit, all are consistent with $\bHI = \bOVI$ at the 1$\,\sigma$
994: level.
995:
996: Gas in collisional ionization equilibrium with $T \simgt 2\times10^5$ implies $\bHI \simgt
997: 60\kms$. It is immediately apparent that few of our components satisfy this criteria. Including a
998: non-thermal contribution to the $b$-value will further lower the implied temperature (or conversely,
999: increase the implied Doppler parameter). This simple argument indicates that collisional ionization
1000: is not the dominant ionization mechanism. It does not, however, rule out collisionally ionized
1001: systems, since it is conceivable that weak, collisionally ionized systems may not contain a
1002: detectable amount of neutral hydrogen (i.e. by considering {\it only} systems with well-aligned
1003: \OVI/\HI\ components, we may be biased against systems in CIE). To quantify this statement, we
1004: examined with {\it Cloudy} (version 07.02.01) a simple model of gas in collisional ionization at
1005: $\logT = 5.5$ and $1/10^{th}$ solar metallicity. For \OVI\ column densities in the range $\logN =
1006: 13.5 - 14.0$ the corresponding \HI\ column density is $\logN \simlt 13.0$, which will be difficult
1007: to detect in our data. We note that while we see no evidence of a population of \OVI\ {\it
1008: absorbers} lacking \HI\ absorption, we certainly see {\it components} that are not well matched
1009: with \HI.
1010:
1011: Table~\ref{tab: ovi_components_temperature} gives the temperature and \bnt\ for the well-matched
1012: components defined above. The first three columns identify the absorber and component. The measured
1013: \HI\ and \OVI\ Doppler parameters are recalled here for reference. The derived gas temperature and
1014: error are given, along with the non-thermal contribution to the Doppler parameter, and the resulting
1015: thermal line-width for both \HI\ and \OVI. Finally, for comparison, we give the gas temperature we
1016: would have derived from \bOVI\ had we assumed $\bnt = 0$, which is strictly a formal upper-limit.
1017: Only one system lies above $\logT = 5.0$: the absorber towards PHL\,1811 at $z=0.15786$, and even in
1018: this case, the data are ambiguous (see Sec~\ref{sec: phl1811}). It is clear, however, that for all
1019: cases in which we can derive a gas temperature, the temperature is well below the $\logT = 5.5$
1020: expected for CIE conditions. It may also be possible that we are observing gas that is radiatively
1021: cooling out of equilibrium. \citet{gnat-sternberg-07-non-equilibrium-cooling} have shown that \OVI\
1022: can be present at temperatures well below those expected in equilibrium models, since the cooling
1023: time is much shorter than the recombination time. These non-equilibrium effects are only important
1024: for enriched gas (solar metallicity or higher), however, and are unlikely to be a dominant effect
1025: for our sample.
1026:
1027: In three cases we could not determine \logT\ and \bnt, because $\bOVI > \bHI$; these cases are
1028: listed in Table~\ref{tab: ovi_components_temperature} by ``...''. Note also, that these are the same
1029: three components in Fig~\ref{fig: doppler_parameter_component_plots} which lie above the $\bOVI =
1030: \bHI$ boundary. All are consistent with \bnt-dominated systems. Alternatively, this lack of
1031: correspondence between the \HI\ and \OVI\ line widths may indicate that either the assumption of the
1032: absorption being physically related is not correct for these systems, or there are unresolved
1033: absorption components. In other words, {\it if} both absorption components arise from the same gas,
1034: then we require that $1/4 < \bOVI/\bHI < 1$, since $m_{\rm O} = 16 \times m_{\rm H}$. The upper
1035: limit comes from a state where the $b$-values are dominated entirely by \bnt (i.e. $\bHI = \bOVI$),
1036: while the lower limit is the pure thermal case, and the b-values are simply related through the
1037: square-root of the ratio of their respective atomic masses. Similar limits can be trivially written
1038: down for all other pairs of species.
1039:
1040: It is worth emphasizing, at this stage, that these measurements are dependent on several critical
1041: factors: a) the resolution of the spectrograph must be sufficient to resolve the lines in question,
1042: reinforcing our decision to rely solely on STIS E140M data with superior resolution (e.g.\ the FUSE
1043: FWHM is $\sim 20-25\kms$, but our {\it median} \bOVI\ is $\sim 21\kms$); b) Further, the resolution
1044: must be sufficient to resolve multiple absorption components and possible blending; c) the
1045: non-thermal contribution to the Doppler parameter {\it must} be considered, since it dominates
1046: $b_{tot}$ in most cases. As Table~\ref{tab: ovi_components_temperature} clearly demonstrates, if we
1047: do not take this into account, most of our absorbers would be consistent with the coronal
1048: temperature range $10^5 - 10^7\K$, and we could easily mis-interpret these as WHIM absorbers. This
1049: point has also been made by \citep{tripp-etal-08-OVI}, while \citet{danforth-shull-08-OVI} derive
1050: only upper limits to the temperature using \bHI.
1051:
1052: \subsection{Ionization State of \Oion\ Bearing Gas}
1053: \label{subsec: ionization}
1054: Before considering the ionization conditions of the \Oion-bearing gas, we first inspect the results
1055: of a typical photoionization model of low-density gas. We use similar models when considering
1056: individual systems below, each tailored to the specific conditions of the individual absorber.
1057: Figure~\ref{fig: photo_model} shows the ionization state for the most commonly observed species as a
1058: function of the ionization parameter, $\logU \equiv {\rm log}\,(n_{\gamma}/\hden)$, where
1059: $n_{\gamma}$ and \hden\ are the volume densities of ionizing photons (${\rm h}\nu > 13.6\eV$) and
1060: hydrogen, respectively. The model was calculated for a slab of gas at with neutral hydrogen column
1061: density $\logNHI = 15.0$ and metal abundance $1/10^{th}$ the solar value. The gas temperature was
1062: constrained to be $\logT > 4.0$, although in practice this constraint has little
1063: effect\footnote{Collisional ionization is not yet significant at these temperatures.}. Ionizing
1064: radiation comes from an updated version of the \citet{haardt-madau-96-UV-background} spectrum
1065: included with {\it cloudy}\footnote{Cloudy denotes this ``HM05''.}, at redshift $z = 0.25$ (the
1066: median redshift of our sample). Since the spectrum of the UV background is fixed, and the relative
1067: intensity is set at a fixed redshift, only changes in the density affect the ionization parameter.
1068: Thus at high densities, there are fewer ionizing photons per atom, and lower ionization states
1069: predominate. At lower densities (higher \logU), there are more ionizing photons for a given atom,
1070: and higher ionization states may be ionized. Inspection of the curves in Figure~\ref{fig:
1071: photo_model} shows that \OVI\ can be observed over a wide range of conditions, especially if one
1072: considers that metallicity may vary over a $1-2\dex$ range. Thus, unlike with CIE where \OVI\
1073: occupies a very narrow temperature range, simply detecting \OVI\ is not enough to determine the
1074: ionization conditions.
1075:
1076: The column density ratios predicted by the {\it Cloudy} models are not sensitive to the adopted
1077: value of \logNHI\ in the optically thin regime, although the absolute column densities would
1078: obviously be lower for a lower gas column. A comparison between models with solar and $1/10^{th}$
1079: solar metallicity shows that the column density ratios are unaffected by metallicity effects in the
1080: regime $\logU > -4.0$, which corresponds to densities $\loghden < -1.6$. This is well away
1081: from $\logU = -0.2$, where \OVI\ peaks for photoionization models. None of the systems we consider
1082: fall into this region of parameter space, so the adopted $\MH = -1.0$ for our models should not
1083: affect our conclusions.
1084:
1085: The detection of multiple metal transitions in an absorber offers a powerful diagnostic of the
1086: ionization conditions of the gas, assuming that the metals responsible for the absorption are
1087: co-spatial. By comparing the column density ratios for well-aligned transitions with predictions of
1088: a model like that in Figure~\ref{fig: photo_model}, we can place limits on \logU; even saturated
1089: lines can provide limits in \logU. With the ionization parameter determined, we obtain the average
1090: volume density of the cloud and using the ionization fraction (from the Cloudy output) we can then
1091: calculate a length scale for the cloud. These numbers are, however, dependent on the strength of the
1092: UV background ionization field, which is uncertain. When relevant in the following analysis, we give
1093: a range of parameters for a range of UV background intensities $J_{\nu, 912\,\mathrm{\AA}} = 2 - 7
1094: \times 10^{-23}\,\mathrm{erg\,s^{-1}\,cm^{-2}\,Hz^{-1}\,sr^{-1}}$
1095: \citep{shull-etal-99-UV-background, scott-etal-02-UV-background}.
1096:
1097: We address several systems below. Many of these systems have already been studied in detail in terms
1098: of their ionization conditions, characteristic sizes and/or average densities. In these cases, we do
1099: not comment unless we have additional information to add, beyond what has been published. The
1100: interested reader is encouraged to see the following references for more information:
1101: \citet[][HE\,0226$-$4110 $z = 0.20702$]{savage-etal-05-HE0226-4110-NeVIII}; \citet[][HE\,0226$-$4110
1102: $z = 0.34034$]{lehner-etal-06-HE0226-4110}; \citet[][PG\,0953+415 $z =
1103: 0.14232$]{tripp-savage-00-PG0953+415_z0.1423}; \citet[][PG\,1116+215 $z =
1104: 0.13847$]{sembach-etal-04-PG1116+215}; \citet[][PKS\,0405$-$12 $z = 0.16703,
1105: 0.36333$]{chen-prochaska-00-PKS0405-12_z0.167}. Note that all our {\it Cloudy} models assume a
1106: solar abundance pattern. If there are significant deviations from this pattern (which may be
1107: particularly problematic for C or N), then we will deduce incorrect limits. Hence, the more metal
1108: transitions we have for a system, the better our results will be.
1109:
1110: \noindent H\,1821+643; $z=0.22496$---This system was first examined by
1111: \citet{tripp-etal-00-H1821+643}, who report $N$ for \OVI\ absorbers on this sightline, but were
1112: primarily interested in the cosmological mass density of \OVI, and did not discuss the ionization
1113: condition of individual absorbers. As noted in Sec~\ref{subsec: H1821+643} the \OVI\ and \SiIII\
1114: Doppler parameters for the $v = 0\kms$ component are inconsistent with a single phase, since $\bOVI
1115: \gg \bSiIII$ and $\bOVI \gg \bCIII$. For this $v=0\kms$ component, we also detect weak \SiIV, but no
1116: \HI\ components are aligned with the \OVI\ position. The \SiIII\ to \SiIV\ column density ratio is
1117: consistent with the \SiIII\ to \CIII\ column density limit, and $\logU \approx -2.1$ ($\loghden
1118: \approx -3.5$). The weaker component at $v \simeq 25\kms$ also contains saturated \CIII, \SiIII\ and
1119: \SiIV. This component also falls within the width of the \OVI\ line, but the centroids are
1120: significantly different. The nearest \HI\ component to this low-ionization gas is at $v = 15\kms$,
1121: and the relation between the two is unclear. For this low-ion $v \simeq 25\kms$ component, \SiIII\
1122: and \SiIV\ have the same column density, while log\,N(\CIII)/N(\SiIII) is $\geq 1.0\dex$, both of
1123: which are consistent with $\logU \approx -1.4$ ($\loghden \approx -4.2$). There are no obvious
1124: matches to the weak \OVI\ component at $v = 60\kms$. The mismatch of the broad \OVI\ and low ion
1125: Doppler parameters, in addition to the multiple low-ion components falling within the \OVI\
1126: absorption range, argues for a core-halo cloud model, with a hot gaseous halo traced by \OVI,
1127: surrounding cooler gas traced by lower ionization species.
1128:
1129: \noindent PG\,1216+069; $z = 0.28232$---\citet{tripp-etal-05-PG1216+069_NGC4261} were the first to
1130: publish STIS data of PG\,1216+069, but concentrate on the DLA at $z = 0.00632$ associated with the
1131: NGC\,4261 group, and do not consider this absorber in detail. We detect \OVI\, \SiIII\ and
1132: saturated \CIII. The \CIII, \SiIII\ and \HI\ are offset from the \OVI\ position by $\Delta\,v =
1133: 6-9\kms$ to the blue, and its association is unclear. We measure $\log\,N(\CIII)/N(\SiIII) > 1.3$
1134: which gives the limit $\logU > -2.6$ ($\loghden \simlt -2.9; L > 5\kpc$, where we have folded in the
1135: range in $J_{\nu}$ into the quoted limits).
1136:
1137: \noindent PG\,1259+593; $z = 0.21950$---Data for the PG\,1259+593 sightline was published by
1138: \citet{richter-etal-04-PG1259+593}, who report the detection of \OIII\ in FUSE data, in addition to
1139: \OVI, \SiIII\ and (saturated) \CIII\ metal lines seen in the STIS data. They conclude that the
1140: absorber is a multi-phase medium, based on the inconsistency of the \SiIII\ data with an ionization
1141: parameter derived using \OIII\ and \OVI. This conclusion is emphasised by our discussion above of
1142: the limits placed by the ratio of Doppler parameters. In our STIS data we measure
1143: $\bOVI/\bSiIII = 1.6 \pm 0.6$ which is inconsistent at the $1\,\sigma$ level with a single
1144: gas phase, since $m_{\scOVI}/m_{\scSiIII} = 0.6$ i.e. for these species to arise from the same gas
1145: phase, we must have $0.6 < \bOVI/\bSiIII < 1.0$, irrespective of the relative
1146: contributions of \bnt\ and \btherm.
1147:
1148: \noindent PKS\,0312$-$77; $z = 0.20275$---While the Lyman limit system at $z=0.20275$ in the sightline
1149: towards PKS\,0312$-$77 shows many metal absorption lines, particularly those of low ions, none can be
1150: unambiguously matched with the \OVI\ absorption. The main \OVI\ absorption component is very broad
1151: compared to other ions; even the highly ionized species like \SiIV\ and \NV\ exhibit a complex
1152: multi-component structure that is seen in the low ionization species. It is therefore likely that
1153: \OVI\ arises in a different gas phase, or that there are unresolved \OVI\ absorption components.
1154:
1155: \noindent PKS\,0405$-$12; $z = 0.49514$---The highest redshift \OVI\ absorber in our survey, this
1156: absorber is fit with a single \OVI\ component. We detect \OIII\ and \OIV\ at $v = 90\kms$, but no
1157: associated \OVI\ or \HI, and we do not consider this component in detail. At $v = -10\kms$, the data
1158: clearly show \CIII\ and \OIV\,787. We discount the \OIII\,832 absorption at $v = -29\pm8\kms$ as too
1159: far from the \OVI\ centroid. \citet{prochaska-etal-04-PKS0405-12} do not provide a full analysis of
1160: this system, referring instead to a future paper which was not forthcoming. From our \OVI, \OIV\ and
1161: \CIII\ measurements, we find $\logU \simeq -0.8 - -1.3$ ($\loghden \approx -4.0 - -4.5; L \approx 12
1162: - 140\kpc$).
1163:
1164:
1165: \subsection{Systems lacking \HI}
1166: \label{subsec: noHI}
1167: In a survey for \OVI\ absorbers, we might expect to find a class of systems that do not contain
1168: associated \HI. If large-scale galactic winds and superwinds shock-heat metals to a WHIM phase,
1169: they may also ionize hydrogen to a level that is not detectable for weak systems (as in the above
1170: discussion on CIE systems). Such hot bubbles may exist even without strong winds
1171: \citep{kawata-rauch-07-galactic-winds}. Since we conduct a {\it blind} search for the \OVI\ doublet,
1172: we should be sensitive to the existence of such systems, down to our detection limit, which is
1173: approximately 30\mA\ with good data (but see \pI\ for a better quantification of this, since not all
1174: data are of equal quality).
1175:
1176: Given the above discussion, it may be somewhat telling that we find no strong evidence for such a
1177: class of systems. Our survey contains two system that we suggest are \HI\ free. The $z=0.32639$
1178: absorber toward HE\,0226$-$4110 (Figure~\ref{fig: HE0226-4110_z0.32639}) shows no evidence of \Lya\
1179: or \Lyb\ absorption, although the \Lya\ region is quite noisy. The upper limit on \NHI\ for this
1180: system implies $\log[NOVI/NHI] \simgt 1.1\dex$. The $z = 0.22638$ system toward H\,1821+643
1181: (Figure~\ref{fig: H1821+643_z0.22638}) shows a single broad \HI\ component significantly offset
1182: ($\Delta\,v = -53\pm2\kms$) from the \OVI\ position. If we posit that this component could hide a
1183: weak, aligned \HI\ component, and attempt to force-fit an \HI\ component aligned with the \OVI\
1184: absorption, we obtain $\NHI < 12.3$, implying $\log[NOVI/NHI] \simgt 1.2\dex$. Such large column
1185: density ratios are well outside the bounds of what is found for the well-aligned components ($-0.75
1186: < log[\NOVI / \NHI] < 0.67$), and imply either strongly enriched gas (super-solar assuming
1187: photoionization; See Figure~\ref{fig: column_density_component_plots}), or that the gas is
1188: collisionally ionized. Expanded samples of low-$z$ \OVI\ absorbers will be very valuable for
1189: determining the fraction of such systems, and their potential as tracers of hot gas.
1190:
1191:
1192:
1193: \section{Summary}
1194: \label{sec: discussion}
1195:
1196: We have presented a catalogue of \OVI\ absorbers selected from high-resolution HST/STIS echelle
1197: data. Our selection technique followed a blind-search for the \OVI\,1031, 1037 doublet
1198: feature. Relying solely on the presence of the \OVI\ doublet, and the ratio of doublet line
1199: strength, we detect 27 \OVI\ absorption systems, independent of other transitions. The statistics
1200: of these absorbers were presented in \pI. We note that 16 systems reported in a contemporaneous
1201: work by \citet{tripp-etal-08-OVI} do not satisfy our selection criteria and therefore are not
1202: included in our sample of 27 absorbers. In cases where only one transition is found or the doublet
1203: ratio appears to be inconsistent with model expectations, these authors include the presence of
1204: other transitions, such as \Lya\ absorption, for justifying the identifications of \OVI.
1205:
1206: In our absorbers, it is common to find multiple absorption components in any given system,
1207: corresponding to physically distinct gas structures. By matching these absorption components from
1208: different species, we can identify different transitions likely due to the same gas. Under this
1209: assumption, we can then derive both the temperature of the gas, and the non-thermal contribution to
1210: the Doppler parameter, \bnt. This analysis demonstrates that, for well-matched \OVI/\HI\ components,
1211: gas temperatures are in the range $\logT = 3.7 - 5.0$, well below the temperature range at which
1212: \OVI\ is expected to be found in collisional ionization equilibrium ($\logT = 5.5$). We thus advise
1213: caution in identifying the \OVI\ absorbers with the hot WHIM gas that simulations predict will
1214: contain a significant fraction of the baryons. Our finding based on all available absorbers in the
1215: current HST/STIS data archive reaffirms previous findings from studies of individual lines of sight.
1216: Future generations of X-ray spectrographs are likely to be necessary to solve this question
1217: conclusively.
1218:
1219: If galactic winds are the dominant source of \OVI\ formation, shock heating the gas to the WHIM
1220: regime, and pushing it out to $\sim 1\Mpc$ from galaxies, we may expect to see this in a variety of
1221: ways. A class of \HI\ free \OVI\ absorbers may arise, as both the hydrogen and oxygen are ionized in
1222: strong winds. We see no evidence of a large number of \HI-free \OVI\ absorbers, with only two
1223: systems present in our survey. Reasoning that the \HI\ absorption will best trace the local
1224: over-density, we compared the velocity difference between our \OVI\ components and the nearest \HI\
1225: component. A wind origin may leave an imprint on the kinematics of the absorbers, manifest as a
1226: systematic offset between the \HI\ and \OVI\ positions. We see no such effect (see also comparisons
1227: in T08), but this test is likely a weak one. Clearly systematic galaxy redshift surveys around the
1228: QSO lines of sight are required to properly address this.
1229:
1230: The location of the missing baryons in the low-redshift is clearly an important issue in modern
1231: cosmology and astronomy. The WHIM is a leading candidate reservoir for this mass with simulations
1232: and suggesting that it contains as much as 50\% of all $z = 0$ baryons. While some have claimed that
1233: the observations support this view, we argue that caution is required. High-resolution spectra that
1234: cover a broad wavelength range are crucial for constraining the ionization, metallicity, and the
1235: temperature of the gas involved, and care must be taken to deduce non-thermal broadening
1236: mechanisms. If the baryons are to be unambiguously located in the WHIM, further observations are
1237: crucial.
1238:
1239:
1240:
1241: \acknowledgements
1242:
1243: This research has made use of the NASA/IPAC Extragalactic Database (NED) which is operated by the
1244: Jet Propulsion Laboratory, California Institute of Technology, under contract with the National
1245: Aeronautics and Space Administration. C.~T. and H.-W.~C.\ acknowledge support from NASA grant
1246: NNG06GC36G. H.-W.~C. acknowledges partial support from an NSF grant AST-0607510.
1247:
1248:
1249: \facility{HST(STIS)}
1250:
1251: \bibliography{MASTER}
1252: \bibliographystyle{apj}
1253:
1254: \clearpage
1255:
1256: % Figures and tables go here
1257:
1258:
1259: \begin{deluxetable}{p{1in}rrrcr}
1260: \tabletypesize{\scriptsize}
1261: \tablecaption{\label{tab: journal_observations}Summary of the STIS Echelle Spectra}
1262: \tablewidth{0pt}
1263: \tablehead{
1264: \colhead{QSO} &
1265: \colhead{$z_{\rm QSO}$} &
1266: \colhead{$z_{\rm min}$} &
1267: \colhead{$z_{\rm max}$\tablenotemark{a}} &
1268: \colhead{$t_{\rm exp}$} &
1269: \colhead{PID} \\
1270: \colhead{(1)} &
1271: \colhead{(2)} &
1272: \colhead{(3)} &
1273: \colhead{(4)} &
1274: \colhead{(5)} &
1275: \colhead{(6)}
1276: }
1277: \startdata
1278: 3C\,249.1 \dotfill & 0.3115 & 0.1222 & 0.2885 & 68776 & 9184 \\
1279: 3C\,273 \dotfill & 0.1580 & 0.1144 & 0.1417 & 18671 & 8017 \\
1280: 3C\,351.0 \dotfill & 0.3719 & 0.1309 & 0.3483 & 73198 & 8015 \\
1281: HE\,0226$-$4110 \dotfill & 0.4950 & 0.1154 & 0.4707 & 43772 & 9184 \\
1282: HS\,0624+6907 \dotfill & 0.3700 & 0.1222 & 0.3464 & 61950 & 9184 \\
1283: H\,1821+643 \dotfill & 0.2970 & 0.1144 & 0.2741 & 50932 & 8165 \\
1284: PG\,0953+415 \dotfill & 0.2390 & 0.1144 & 0.2163 & 24478 & 7747 \\
1285: PG\,1116+215 \dotfill & 0.1765 & 0.1144 & 0.1536 & 39836 & 8165/8097 \\
1286: PG\,1216+069 \dotfill & 0.3313 & 0.1309 & 0.3078 & 69804 & 9184 \\
1287: PG\,1259+593 \dotfill & 0.4778 & 0.1144 & 0.4533 & 95760 & 8695 \\
1288: PG\,1444+407 \dotfill & 0.2673 & 0.1222 & 0.2442 & 48624 & 9184 \\
1289: PHL\,1811 \dotfill & 0.1917 & 0.1144 & 0.1690 & 33919 & 9418 \\
1290: PKS\,0312$-$77 \dotfill & 0.2230 & 0.1241 & 0.1999 & 37908 & 8651 \\
1291: PKS\,0405$-$12 \dotfill & 0.5726 & 0.1241 & 0.5478 & 27208 & 7576 \\
1292: PKS\,1302$-$102 \dotfill & 0.2784 & 0.1183 & 0.2558 & 22119 & 8306 \\
1293: Ton\,28 \dotfill & 0.3297 & 0.1231 & 0.3059 & 48401 & 9184 \\
1294: \enddata
1295: \tablecomments{This summary of STIS echelle lines of sight is taken from \pI. It is repeated
1296: here for completeness.}
1297: \tablenotetext{a}{The maximum redshift is defined for O\,VI absorbers at velocity separation $>
1298: 5000$ \kms\ from the background QSO, but the line search is conducted through the emission
1299: redshift of the QSO. In this paper we do not consider system inside this 5000\kms limit.}
1300: \end{deluxetable}
1301:
1302: \input{tab2}
1303:
1304:
1305: \begin{figure}
1306: \epsscale{0.6}
1307: \plotone{f1.eps}
1308: \caption{3C\,249.1 $z = 0.24676$---\OVI\ and strong \HI\ are detected in this system. \Lyb\ is blended
1309: with what we tentatively assign as \Lya\ line at $z = 0.0517$. There is a weak and uncertain feature
1310: which may be \SiIII\,1206.}
1311: \label{fig: 3C249.1_z0.24676}
1312: \end{figure}
1313:
1314: \input{tab3.tex}
1315:
1316:
1317: \begin{figure}
1318: \epsscale{1.0}
1319: \plotone{f2.eps}
1320: \caption{3C\,273 $z = 0.12003$---This weak system is detected in only \OVI\ and \Lya. The \OVI\ lines
1321: are mismatched in their strengths, but neither can be weak \Lya, since they both lie to the blue
1322: of the $z=0$ \Lya\,1215 line.}
1323: \label{fig: 3C273_z0.12003}
1324: \end{figure}
1325:
1326: \input{tab4}
1327:
1328:
1329: \begin{figure}
1330: \epsscale{0.6}
1331: \plotone{f3.eps}
1332: \caption{3C\,351.0 $z = 0.31659$---Three obvious \OVI\ components are seen in this
1333: absorber. Corresponding \Lya\ absorption is heavily saturated, but \Lyb\ and \Lyc\ allow us to
1334: derive \HI\ column densities and positions. The \Lya\ components are well matched in position
1335: with the \OVI\ absorption. No associated metals are detected in our wavelength range.}
1336: \label{fig: 3C351.0_z0.31659}
1337: \end{figure}
1338:
1339: \clearpage
1340:
1341: \input{tab5}
1342:
1343: \begin{figure*}
1344: \epsscale{1.0}
1345: \plottwo{f4a.eps}{f4b.eps}
1346: \caption{H\,1821+643 $z = 0.22496$---A strong main \OVI\ absorption component and weaker, offset
1347: component are both obvious in this system. \Lya\ is heavily saturated, but higher order Lyman
1348: lines are available. Strong \CIII\ and \SiIII\ can also be seen. }
1349: \label{fig: H1821+643_z0.22496}
1350: \end{figure*}
1351:
1352: \begin{figure}
1353: \epsscale{0.8}
1354: \plotone{f5.eps}
1355: \caption{H\,1821+643 $z = 0.22638$---This system comprises weak \OVI\ absorption, and very broad \HI
1356: absorption offset from the \OVI\ by $-53\kms$. This absorber is only $\sim350\kms$ from the
1357: system at $z=0.22496$.}
1358: \label{fig: H1821+643_z0.22638}
1359: \end{figure}
1360:
1361: \begin{figure}
1362: \epsscale{0.6}
1363: \plotone{f6.eps}
1364: \caption{H\,1821+643 $z = 0.24532$---Two weak \OVI\ components and a corresponding very weak \HI\
1365: absorber are seen in this system. \NV\,1238 is possibly present, but the \NV\,1242 line is lost
1366: in the Galactic \CIV\,1548 line.}
1367: \label{fig: H1821+643_z0.24532}
1368: \end{figure}
1369:
1370:
1371: \begin{figure}
1372: \epsscale{1.0}
1373: \plotone{f7.eps}
1374: \caption{H\,1821+643 $z = 0.26656$---This weak system consists of a single \OVI\ and \HI\ component,
1375: both well-aligned.}
1376: \label{fig: H1821+643_z0.26656}
1377: \end{figure}
1378:
1379: \input{tab6}
1380: \clearpage
1381:
1382: \begin{figure*}
1383: \epsscale{1.0}
1384: \plottwo{f8a.eps}{f8b.eps}
1385: \caption{HE\,0226$-$4110 $z=0.20702$---This strong \OVI\ absorber is also seen in saturated \Lyab\
1386: absorption. The \Lyc\ line is contaminated by hot pixels, making the \NHI\ determination
1387: uncertain. We see associated \SiIII, \CIII\ and possibly \NV\ absorption.}
1388: \label{fig: HE0226-4110_z0.20702}
1389: \end{figure*}
1390:
1391: \begin{figure}
1392: \plotone{f9.eps}
1393: \caption{HE\,0226$-$4110 $z=0.32639$---This system is weak and uncertain. It is the only system we
1394: detect without corresponding \HI\ absorption. No other transitions are present and correspond
1395: to the \OVI\ absorption.}
1396: \label{fig: HE0226-4110_z0.32639}
1397: \end{figure}
1398:
1399: \begin{figure}
1400: \epsscale{0.7}
1401: \plotone{f10.eps}
1402: \caption{HE\,0226$-$4110 $z=0.34034$---The \OVI\,1037 line in this system is contaminated by an
1403: unidentified metal line. \HI\ is present, but noisy, and the fit is poor. Weak \CIII\ is also
1404: detected.}
1405: \label{fig: HE0226-4110_z0.34034}
1406: \end{figure}
1407:
1408: \begin{figure}
1409: \epsscale{1.0}
1410: \plotone{f11.eps}
1411: \caption{HE\,0226$-$4110 $z=0.35525$---\OVI\,1031 in this system shows some evidence of contamination
1412: in the line wing, but this does not unduly affect our fits. Only associated \HI\ lines are
1413: detected; no metal lines are seen in our STIS data.}
1414: \label{fig: HE0226-4110_z0.35525}
1415: \end{figure}
1416: \clearpage
1417:
1418: \input{tab7}
1419:
1420: \begin{figure}
1421: \epsscale{1.0}
1422: \plotone{f12.eps}
1423: \caption{HS\,0624+6907 $z=0.31796$---This weak absorber is detected only in \OVI\ and \Lya, with a
1424: 15\kms\ offset between the two species.}
1425: \label{fig: HS0624+6907_z0.31796}
1426: \end{figure}
1427:
1428: \begin{figure}
1429: \epsscale{0.5}
1430: \plotone{f13.eps}
1431: \caption{HS\,0624+6907 $z=0.33984$---While the \Lya\ line is heavily saturated in this absorber,
1432: the \OVI\ lines are very weak. The \Lyb\ profile is partly blended, but the \Lyc\ line is free
1433: from contamination. No significant absorption from other species are detected.}
1434: \label{fig: HS0624+6907_z0.33984}
1435: \end{figure}
1436:
1437: \input{tab8}
1438:
1439: \begin{figure}
1440: \epsscale{0.8}
1441: \plotone{f14.eps}
1442: \caption{PG\,0953+415 $z=0.14232$---Strong \OVI\ is present in this system, although the \OVI\,1037
1443: line is stronger than expected. \HI\ is well fit in the \Lya\ line, and is not affected by the
1444: blending seen to the red of the line. \Lyb\ in this absorber is contaminated by \HI\,949 line
1445: from the QSO host at $z=0.2335$. \CII\,1334 may be present, but is extremely weak and we flag it
1446: as uncertain.}
1447: \label{fig: PG0953+415_z0.14232}
1448: \end{figure}
1449:
1450: \input{tab9}
1451:
1452:
1453: \begin{figure*}
1454: \epsscale{1.0}
1455: \plottwo{f15a.eps}{f15b.eps}
1456: \caption{PG\,1116+215 $z=0.13847$---While the \OVI\ transitions in this absorber are quite noisy, we
1457: have good data for \Lyab, both of which are saturated. We cannot obtain satisfactory fits to
1458: \Lya\ and \Lyb\ simultaneously. \SiII, \SiIII\, \SiIV, \NII\ and \CII\ are all clearly detected,
1459: often in multiple transitions (e.g.\ \SiII\ is detected in the 1190, 1193, 1260 and 1304
1460: transitions).}
1461: \label{fig: PG1116+215_z0.13847}
1462: \end{figure*}
1463:
1464: \input{tab10}
1465:
1466: \begin{figure*}
1467: \epsscale{1.0}
1468: \plottwo{f16a.eps}{f16b.eps}
1469: \caption{PG\,1216+069 $z=0.28232$---This system is another instance of very strong, saturated \HI,
1470: strong low ion metal species, but only weak \OVI. All the Lyman series lines in our spectrum are
1471: saturated, or are too noisy to be useful. \CIII\,977 is also saturated. It is not clear whether
1472: the absorption to the blue of the central \CIII\ component is \CIII\ associated with blue \HI\
1473: components, or a contaminating line.}
1474: \label{fig: PG1216+069_z0.28232}
1475: \end{figure*}
1476: \clearpage
1477:
1478: \input{tab11}
1479:
1480: \begin{figure}
1481: \epsscale{0.5}
1482: \plotone{f17.eps}
1483: \caption{PG\,1259+593 $z=0.21950$---Multi-component \OVI\ and \HI\ are both present in this system,
1484: with the \HI\ lines saturated. The blue wing of the \Lyb\ line blends with Galactic
1485: \SII\,1250. \CIII\ and \SiIII\ are both detected; \CIII\ is saturated, while the \SiIII\ line is
1486: weak. Both species align with the $v=0\kms$ component.}
1487: \label{fig: PG1259+593_z0.21950}
1488: \end{figure}
1489:
1490: \begin{figure}
1491: \epsscale{0.8}
1492: \plotone{f18.eps}
1493: \caption{PG\,1259+593 $z=0.25981$---This double-component absorber shows only two \OVI\ and \HI\
1494: components.}
1495: \label{fig: PG1259+593_z0.25981}
1496: \end{figure}
1497:
1498: \input{tab12}
1499:
1500:
1501: \begin{figure}
1502: \epsscale{1.0}
1503:
1504: \plotone{f19.eps}
1505: \caption{PHL\,1811 $z=0.15786$---Both \OVI\ and \HI\ in this system are weak. The putative narrow
1506: \HI\ component is {\it not} \HI, but rather \OI\,1302 in the $z=0.0809$ Lyman-limit system. We
1507: include this line in the fit, detecting only broad, weak \HI\ that is too weak to be detected in
1508: the \Lyb\ transition.}
1509: \label{fig: PHL1811_z0.15786}
1510: \end{figure}
1511:
1512: \input{tab13}
1513: \clearpage
1514:
1515: \begin{figure*}
1516: \epsscale{1.0}
1517: \plottwo{f20a.eps}{f20b.eps}
1518: \caption{PKS\,0312$-$77 $z=0.20275$---This Lyman-limit system shows strong, saturated \OVI\, \HI,
1519: \SiII, \SiIII\, \SiIV and \CIII. We also see weak \NV. We are not able to resolve multiple
1520: absorption components in the main \OVI\ absorber, but their presence is strongly suggested by the
1521: complicated component structure exhibited in other transitions.}
1522: \label{fig: PKS0312-77_z0.20275}
1523: \end{figure*}
1524: \begin{figure}
1525: \epsscale{1.0}
1526: \plottwo{f21a.eps}{f21b.eps}
1527: \caption{Further transitions of the PKS\,0312$-$77 $z=0.20275$ system, as for Figure~\ref{fig: PKS0312-77_z0.20275}.}
1528: \label{fig: PKS0312-77_z0.20275_2}
1529: \end{figure}
1530:
1531: \input{tab14}
1532:
1533:
1534: \begin{figure*}
1535: \epsscale{1.0}
1536: \plottwo{f22a.eps}{f22b.eps}
1537: \caption{PKS\,0405$-$12 $z=0.16703$---This partial Lyman-limit system has strong \OVI\ and
1538: corresponding saturated \HI, and low ion metals with a complicated structure.}
1539: \label{fig: PKS0405-12_z0.16703}
1540: \end{figure*}
1541:
1542: \begin{figure}
1543: \epsscale{1.0}
1544: \plotone{f23.eps}
1545: \caption{PKS\,0405$-$12 $z=0.18291$---This is system has two strong \OVI\ components separated by some
1546: 90\kms. The \OVI\,1031 data are quite noisy. Both \OVI\ components have corresponding \HI\
1547: absorption, but the \Lya\ line is heavily saturated and higher order lines are outside the
1548: wavelength coverage of our data. No metal absorption is seen.}
1549: \label{fig: PKS0405-12_z0.18291}
1550: \end{figure}
1551: \clearpage
1552:
1553:
1554: \begin{figure}
1555: \epsscale{1.0}
1556: \plotone{f24.eps}
1557: \caption{PKS\,0405$-$12 $z=0.36333$---This system shows weak \OVI\ with a well-fit doublet. The \Lya\
1558: line shows contamination from bad pixels, and is also possibly blended with Galactic \CI*\,1657;
1559: \Lyb\ is not detected. Weak \CIII\ is also present.}
1560: \label{fig: PKS0405-12_z0.36333}
1561: \end{figure}
1562:
1563: \begin{figure}
1564: \epsscale{0.6}
1565: \plotone{f25.eps}
1566: \caption{PKS\,0405$-$12 $z=0.49514$---The highest redshift absorber in our sample, both \OVI\ lines
1567: are clearly detected, but both suffer from bad pixels, complicating the fitting. \Lya\ is
1568: redshifted out of our wavelength range, and \Lyb\ shows a very wide profile. Strong \CIII\ is
1569: detected, as are \OIII\ and \OIV.}
1570: \label{fig: PKS0405-12_z0.49514}
1571: \end{figure}
1572:
1573: \input{tab15}
1574:
1575: \begin{figure}
1576: \epsscale{1.0}
1577: \plotone{f26.eps}
1578: \caption{PKS\,1302$-$102 $z=0.22565$---The lopsided, two-component structure seen in \OVI\ in this
1579: system is mirrored in the \HI\ absorption. No other transitions are detected.}
1580: \label{fig: PKS1302-102_z0.22565}
1581: \end{figure}
1582:
1583: \begin{figure}
1584: \plotone{f27.eps}
1585: \caption{PKS\,1302$-$102 $z=0.22744$---A mere 440\kms from the absorber at $z=0.22565$, this absorber
1586: is weaker than the lower redshift system. The \OVI\,1037 line is detected with low significance,
1587: and the corresponding \HI\ line is very weak. The strong line to the red of the \OVI\,1031 line
1588: is \Lya.}
1589: \label{fig: PKS1302-102_z0.22744}
1590: \end{figure}
1591:
1592: \input{tab16}
1593:
1594: \begin{figure}
1595: \epsscale{0.8}
1596: \plotone{f28.eps}
1597: \caption{Ton\,28 $z=0.27340$---Only very weak \OVI\ absorption is seen to correspond with the
1598: saturated \Lya\ in this system. \Lyb\ is present and unsaturated, as is \Lyc. The \Lya\ line is
1599: blended with Galactic \CIV\,1548, which we fit simultaneously.}
1600: \label{fig: Ton28_z0.27340}
1601: \end{figure}
1602:
1603: \input{tab17}
1604:
1605:
1606:
1607: \begin{deluxetable}{p{1in}rrr}
1608: \tabletypesize{\scriptsize}
1609: \tablecaption{\label{tab: unconfirmed_systems}Systems not confirmed by our search}
1610: \tablewidth{0pt}
1611: \tablehead{
1612: \colhead{QSO} &
1613: \colhead{$z_{abs}$} &
1614: \colhead{\NOVI} &
1615: \colhead{\NHI} \\
1616: \colhead{(1)} &
1617: \colhead{(2)} &
1618: \colhead{(3)} &
1619: \colhead{(4)}
1620: }
1621: \startdata
1622: 3C\,249.1 \dotfill & 0.23641 & 13.99 & 13.23 \\
1623: 3C\,351.0 \dotfill & 0.21811 & 13.96 & 13.50 \\
1624: 3C\,351.0 \dotfill & 0.22111 & 14.25 & $>$14.56 \\
1625: H\,1821+643 \dotfill & 0.12143 & 13.94 & 14.31 \\
1626: H\,1821+643 \dotfill & 0.21331 & 13.57 & $>$14.29 \\
1627: HE\,0226$-$4110 \dotfill & 0.42670 & 14.34 & \nodata\tablenotemark{a} \\
1628: PG\,1216+069 \dotfill & 0.26768 & 13.30 & $>$13.86 \\
1629: PG\,1259+593 \dotfill & 0.31972 & 13.92 & 14.04 \\
1630: PG\,1444+407 \dotfill & 0.22032 & 13.95 & 13.63 \\
1631: PHL\,1811 \dotfill & 0.13240 & 14.39 & $>$14.30 \\
1632: PKS\,0312$-$77 \dotfill & 0.15890 & 13.94 & 13.90 \\
1633: PKS\,0312$-$77 \dotfill & 0.19827 & 13.84 & $>$14.26 \\
1634: PKS\,0405$-$12 \dotfill & 0.36156 & 14.00 & $>$14.14 \\
1635: PKS\,1302$-$102 \dotfill & 0.19159 & 13.93 & $>$14.17 \\
1636: Ton\,28 \dotfill & 0.13783 & 13.99 & $>$14.37 \\
1637: Ton\,28 \dotfill & 0.20524 & 13.69 & 13.20 \\
1638: \enddata
1639: \tablecomments{Summary of \scOVI\ candidates which we do not recover. See text for details on
1640: individual systems.}
1641:
1642: \tablenotetext{a}{\citet{tripp-etal-08-OVI} report no \scHI\ with this system. \Lya\ is redshifted out
1643: of the STIS band, and \Lyb\ is not detected.}
1644:
1645: \end{deluxetable}
1646:
1647: \begin{figure}
1648: \plotone{f29.eps}
1649: \caption{Spectra of the \OVI\,1031 (top) and \OVI\,1037 (bottom) spectral regions for absorbers
1650: we do not confirm. Blended regions are shown by dotted spectral regions.}
1651: \label{fig: OVI_rejected_1}
1652: \end{figure}
1653:
1654: \begin{figure}
1655: \includegraphics[angle=0, scale=0.8]{f30.eps}
1656: \caption{As for Figure~\ref{fig: OVI_rejected_1}, but remaining systems not shown in that figure.}
1657: \label{fig: OVI_rejected_2}
1658: \end{figure}
1659:
1660:
1661: \begin{figure}
1662: \plotone{f31.eps}
1663: \caption{Distribution of \OVI\ and \HI\ column densities for well aligned absorption components.
1664: Full symbols are regular components, while open circles designate components whose alignment is
1665: flagged as uncertain.}
1666: \label{fig: column_density_component_plots}
1667: \end{figure}
1668:
1669: \begin{figure}
1670: \epsscale{1.0}
1671: \plotone{f32.eps}
1672: \caption{Example of column densities as a function of ionization parameter derived from {\it
1673: Cloudy} photoionization models. With increasing ionization parameter, species become more
1674: highly ionized, as expected. The models were computed for \logN = 15.0, and a nominal abundance
1675: $[{\rm M}/{\rm H}] = -1.0$. \OVI\ dominates the Oxygen ionization at $\logU > -0.8$ and peaks at
1676: $\logU = -0.2$.}
1677: \label{fig: photo_model}
1678: \end{figure}
1679:
1680:
1681: \begin{figure}
1682: \plotone{f33.eps}
1683: \caption{Distribution of Doppler parameters for aligned components. Symbols are the same as for
1684: Fig~\ref{fig: column_density_component_plots}. The lower line shows the boundary for $\bnt = 0$,
1685: below which we should not see any components. The upper line shows $\bOVI = \bHI$, where the
1686: Doppler parameter is dominated by turbulent motions.}
1687: \label{fig: doppler_parameter_component_plots}
1688: \end{figure}
1689:
1690:
1691: \input{tab19}
1692:
1693:
1694: \end{document}
1695: