1: % MNRAS article
2: \documentclass[useAMS,usenatbib]{mn2e}
3: \usepackage{graphicx}
4:
5: \newcommand{\kms}{km\,s$^{-1}$}
6: \newcommand{\sqcm}{cm$^{-2}$}
7: \newcommand{\lya}{Ly$\alpha$}
8: \newcommand{\hi}{\mbox{H\,{\sc i}}}
9: \newcommand{\hw}{\mbox{H\,{\sc ii}}}
10: \newcommand{\os}{\mbox{O\,{\sc vi}}}
11: \newcommand{\cf}{\mbox{C\,{\sc iv}}}
12: \newcommand{\ct}{\mbox{C\,{\sc iii}}}
13: \newcommand{\sif}{\mbox{Si\,{\sc iv}}}
14: \newcommand{\nf}{\mbox{N\,{\sc v}}}
15: \newcommand{\mgii}{\mbox{Mg\,{\sc ii}}}
16: \newcommand{\nsys}{35} % Num O VI systems %*
17: \newcommand{\ncomp}{101} % Num O VI comp
18: \newcommand{\Nstr}{Nine} % Num Strong
19: \newcommand{\nstr}{9} % Num Strong
20: \newcommand{\nwea}{26} % Num Weak %*
21: \newcommand{\zabs}{$z_{\rm abs}$}
22: \newcommand{\zqso}{$z_{\rm qso}$}
23:
24: \title[High-$z$ Proximate \os]
25: {A Study of Quasar Proximity in \os\ absorbers at
26: $z$=2--3\thanks{Based on observations taken under ESO
27: Programme ID 166.A-0106(A), using the Ultraviolet and Visual Echelle
28: Spectrograph (UVES) on the Very Large Telescope, Unit 2 at
29: Paranal, Chile.}}
30: \author[Fox, Bergeron, \& Petitjean]
31: {Andrew J. Fox$^{1,2}$\thanks{E-mail: afox@eso.org},
32: Jacqueline Bergeron$^2$, and Patrick Petitjean$^{2,3}$\\
33: $^1$European Southern Observatory, Alonso de C\'ordova 3107, Casilla
34: 19001, Vitacura, Santiago 19, Chile\\
35: $^2$Institut d'Astrophysique de Paris, UMR 7095 CNRS,
36: Universit\'e Pierre et Marie Curie, 98bis Boulevard Arago, 75014
37: Paris, France\\
38: $^3$LERMA, Observatoire de Paris, 61 Avenue de l'Observatoire, 75014
39: Paris, France}
40:
41: \begin{document}
42: \maketitle
43:
44: \begin{abstract}
45: With the goal of investigating the nature of \os\ absorbers at high
46: redshifts, we study the effects of proximity to the background quasar.
47: In a sample of sixteen quasars at \zqso\ between 2.14 and 2.87
48: observed at high signal-to-noise and 6.6~\kms\ resolution with VLT/UVES,
49: we detect \nsys\ \os\ absorption-line systems (comprising over 100
50: individual \os\ components) lying within 8\,000~\kms\ of \zqso.
51: We present component fits to the \os\ absorption and the accompanying
52: \hi, \cf, and \nf. The systems can be categorized into \nstr\ strong
53: and \nwea\ weak \os\ absorbers.
54: The strong (intrinsic) absorbers
55: are defined by the presence of either broad, fully saturated %mini-BAL
56: \os\ absorption or partial coverage of the continuum source, and in
57: practice all have log\,$N$(\os)$\ga$15.0; these systems are interpreted as
58: representing either QSO-driven outflows or gas close the central engine
59: of the AGN.
60: The weak (also known as narrow) systems show no partial coverage or
61: saturation, and are characterized by log\,$N$(\os)$<$14.5 and a median
62: total velocity width of only 42~\kms. %*
63: The incidence d$N$/d$z$ of weak \os\ systems within 2\,000~\kms\ of
64: the quasar down to a limiting equivalent width of 8~m\AA\ is %*
65: 42$\pm$12. Between 2\,000 and 8\,000~\kms, d$N$/d$z$ falls to 14$\pm$4, %**
66: equal to the incidence of intervening \os\ absorbers measured in
67: the same spectra.
68: Whereas the accompanying \hi\ and \cf\ column densities are
69: significantly lower (by a mean of $\sim$1~dex) %*
70: in the weak \os\ absorbers within 2\,000~\kms\ of \zqso\
71: than in those at larger velocities, %*
72: the \os\ column densities display no dependence on proximity.
73: Furthermore, significant offsets between the \hi\ and \os\ centroids
74: in $\approx$50 per cent of the weak absorbers imply that (at least in
75: these cases) the \hi\ and \os\ lines are not formed in the same phase of
76: gas, preventing us from making reliable metallicity determinations,
77: and ruling out single-phase photoionization-model solutions.
78: In summary, we find no firm evidence that quasar radiation influences
79: the weak \os\ absorbers, suggesting they are collisionally ionized rather than
80: photoionized, possibly in the multi-phase halos of foreground galaxies.
81: Non-equilibrium collisional ionization models are needed to explain
82: the low temperatures in the absorbing gas, which are implied by narrow
83: line widths ($b\!<\!14$~\kms) in over half of the observed \os\
84: components. %* non-solar abundances
85: \end{abstract}
86:
87: \begin{keywords}
88: cosmology: observations -- quasars: absorption lines -- intergalactic
89: medium.
90: \end{keywords}
91:
92: \section{Introduction}
93: Quasars are ideal backlights for absorption-line spectroscopy, thanks
94: to their high luminosities and flat continua.
95: Yet in addition to allowing us to detect foreground absorbers, quasars
96: also create and influence them, through the effects of outflows and
97: ionizing radiation. The enhanced level of ionizing radiation in the
98: vicinity of quasars gives rise to the line-of-sight proximity effect,
99: the observed decrease in the number density of
100: \lya\ forest lines (equivalent to a decrease in the mean optical
101: depth, and an increase in the level of hydrogen ionization) at velocities
102: approaching the quasar \citep{Ca82, Mu86, Ty87, Ba88, Lu91, Bh94,
103: Ra98, Sc00, Gu07, FG08}.
104:
105: To obtain an unbiased view of absorbing structures in the high-$z$
106: Universe, one needs to remove the effects of proximity from samples of
107: metal-line systems.
108: The broad absorption line (BAL) systems, easy to identify in QSO
109: spectra because of their optically thick absorption extending over
110: thousands of \kms, are clearly connected to the
111: quasar. BALs trace high-velocity QSO outflows at speeds of up to 0.1$c$
112: \citep[e.g.][]{Tu84, Tu88}, and are seen in 10--15\% of quasar
113: spectra at $z\!>\!1.5$ \citep{We91, Tr06}.
114: A separate category of mini-BAL systems, with fully saturated absorption
115: profiles and total velocity width $<$ 2\,000~\kms, has also been
116: identified \citep{Tu88, Ch99, Yu02, Mi07b}.
117: As with BALs, mini-BALs are interpreted as tracing outflows ejected by QSOs.
118: A more difficult task is to decide whether {\it narrow}
119: absorbers in quasar spectra are either created by or ionized by the
120: QSO. The traditional way to do this employs the displacement velocity
121: from the quasar,
122: $\delta v\!\equiv\!v_{\rm qso}\!-\!v_{\rm abs}$.
123: So-called `associated' or `$z_{\rm abs}\!\approx\! z_{\rm qso}$' or
124: `proximate' absorbers, typically defined as narrow absorbers at
125: $\delta v\!<\!5\,000$~\kms\ \citep{We79, Fo86, An87, HF99}
126: are often removed from intervening
127: samples because of the possibility they may be intrinsic\footnote{We
128: favour the use of the term `proximate' since it involves no
129: assumption about the absorber origin.}.
130: Arguments in favour of an intrinsic nature for many proximate
131: absorbers include time-variability of absorption,
132: partial coverage of the continuum source,
133: detection of excited ionic states implying high gas densities,
134: profiles that are smoother than intervening absorbers,
135: and super-solar metallicities
136: \citep{Wa93, Mo94, Pe94, Sv94, Tr96, BS97, Ha97,
137: Ha97a, Ha97c, Ha01, Ha00, PS99, SP00, Ga99, Ga01, Ga06, Na04}.
138:
139: Using a velocity cutoff to differentiate intervening and
140: intrinsic absorbers has two main problems:
141: quasar-ejected intrinsic systems can appear at higher $\delta v$
142: \citep{Ha97b, Ri99, Ri01, Mi07a, Ne08}, and intervening systems
143: can appear at lower $\delta v$ \citep{Mo98, Se04}.
144: So are many genuine intervening systems lost when they are excluded
145: from absorber samples just because of their proximity to the quasar?
146: In this paper we investigate the transition between proximate and
147: intervening absorbers. We focus on \os\ absorbers, which
148: trace either warm-hot ($T\!\ga\!10^5$~K)
149: collisionally-ionized plasma, or photoionized gas
150: subject to a hard ionizing spectrum extending to energies above
151: 113.9~eV (the ionization potential to create O$^{+5}$).
152: \os\ absorbers are of interest
153: for many reasons, including the significant role they play in the
154: baryon and metal budgets, the window they provide on intergalactic
155: metal enrichment, and their ability to trace energetic galaxy/IGM
156: interactions such as accretion and galactic winds. After the first
157: detection of \os\ absorbers at $z$=2--3 by \citet{LS93}, much work has
158: been done in the era of 10m-class telescopes to characterize their properties
159: \citep*{Sy00, Ca02, Be02, Si02, Si04, Si06, Be05, Lo07, Ag08, Go08}.
160:
161: Here we ask a simple question: is there a proximity effect in \os?
162: In other words, can we see the signature of quasar photoionization, in
163: terms of correlations between the ionization properties of absorbers
164: and their proximity to the quasar?
165: To address this issue, we form a sample of
166: proximate \os\ absorption systems at $z$=2--3 using a homogeneous set
167: of high-resolution VLT/UVES quasar spectra, and we then compare their
168: properties to a sample of intervening (i.e., non-proximate) \os\
169: absorbers observed in the same set of spectra;
170: the intervening sample has been published by \citet[][hereafter BH05]{Be05},
171: but it has since been enlarged from ten to twelve sight lines. %*
172: We also briefly compare our results on \os\ absorbers at $z$=2--3 to those
173: obtained in the low-redshift Universe with space-based ultraviolet
174: spectrographs. This paper is thus a study of both the proximity effect and
175: the nature of \os\ absorbers in general, and is
176: structured as follows. In \S2 we describe the data
177: acquisition and reduction, and our absorber identification and
178: measurement processes. In \S3 we discuss the observational
179: properties of proximate \os\ absorbers. In \S4 we discuss the
180: implications of our results, and we then present a summary in \S5.
181: Throughout this paper we adopt a WMAP 3-year cosmology with
182: $\Omega_{\rm M}$=0.27, $\Omega_\Lambda$=0.73, and $h_{\rm 70}$=1 \citep{Sp07}.
183:
184: \section{Data Acquisition and Handling}
185: \subsection{Observations}
186: The ESO-VLT Large Programme {\it The Cosmic Evolution of the
187: Intergalactic Medium} (IGM) has built a homogeneous sample of
188: high resolution, high signal-to-noise quasar spectra for use in
189: studying the high-redshift IGM.
190: Twenty bright quasars (most with $V\!<\!17$) in the redshift range
191: 2.1--3.8 (median \zqso=2.44) were selected for observation,
192: deliberately avoiding sight lines that contain high-column density damped
193: \lya\ absorbers. The observations were carried out with the
194: Ultraviolet and Visual Echelle Spectrograph
195: \citep[UVES;][]{De00} on VLT UT2 at Paranal, Chile, using a
196: total of thirty nights of service-mode telescope time
197: between June 2001 and September 2002.
198: The observations were taken in good seeing conditions ($<$0.8\arcsec)
199: and at airmass$<$1.4, with both blue and red dichroics
200: (Dic1 346+580, Dic1 390+564, and Dic2 437+860 settings),
201: giving spectra complete from 3\,000
202: to 9\,000~\AA. The data reduction was conducted using
203: the {\sc MIDAS} pipeline described in \citet{Ba00}, and then continua
204: were fitted using a third-degree spline function interpolated between
205: regions free from absorption. Full details of the reduction process
206: are given in \citet{Ch04} and \citet{Ar04}.
207: The 2$\times$2 binning mode was used, giving a rebinned
208: pixel size of 2.0--2.4~\kms.
209: The spectral resolution of the data, $R\approx45\,000$, corresponds to
210: a velocity resolution of $\approx$6.6~\kms\ (FWHM). Typical
211: signal-to-noise ratios per pixel in the spectra are $\approx$35 at
212: 3\,500~\AA\ and $\approx$60 at 6\,000~\AA.
213: Among the twenty quasars, sixteen are at $z\!<\!3$.
214: These form the basic sample for this paper. The four cases at
215: $z\!>\!3$ were excluded, because the density of the \lya\ forest at these
216: redshifts renders futile searches for \os.
217:
218: The emission-line redshifts of ten of the sixteen quasars
219: in our sample have been carefully determined by \citet[][and
220: references therein]{Ro05},
221: and we adopt their values. For the remaining six quasars we take
222: \zqso\ from \citet{Sc06}. Each of these quasar redshifts is based on
223: a combination of the observed H$\alpha$, \mgii, \cf, and \sif\
224: emission-line redshifts, allowing for systematic shifts in the
225: displacement velocities of the various lines
226: \citep[e.g.][]{TF92}. The uncertainty on each \zqso\ is of the order
227: 500~\kms\ \citep{Ro05}.
228:
229: \subsection{Identification of \os\ systems}
230: We searched systematically through the sixteen quasar spectra in our sample
231: for \os\ absorbers within 8\,000~\kms\ of the quasar redshift.
232: This velocity range corresponds to redshifts between
233: $z_{\rm min}$=\zqso--[(1+\zqso)\,8\,000~\kms/c]
234: and \zqso.
235: Our choice of velocities up to 8\,000~\kms\ from the quasar was
236: made in order to investigate whether a transition occurs at
237: 5\,000~\kms, the conventional cutoff between proximate and
238: intervening systems. In practise, we also searched for
239: (and found) \os\ absorbers at up to 2\,500~\kms\ {\it beyond}
240: the quasar redshift, indicating the presence of high-velocity inflows
241: toward the quasar.
242:
243: 32 \os\ absorbers were identified by their presence in both lines of the
244: \os\ $\lambda\lambda$1031.926, 1037.617 doublet in the correct ratio
245: through the line profile, i.e.
246: $\tau_a(v, \lambda 1031)/\tau_a(v, \lambda 1037)=2$,
247: where the apparent optical depth
248: $\tau_a(v)={\rm ln}\,[F_c(v)/F(v)]$, and where $F(v)$ and $F_c(v)$ are the
249: observed flux level and estimated continuum flux level, respectively,
250: as a function of velocity.
251: In addition, three \os\ absorbers were identified in one line of the
252: \os\ doublet only, with partial blending in the other \os\ line, but
253: with corresponding absorption seen in \cf\ and \hi\
254: at exactly the same redshift. %*
255: Together the sample consists of the \nsys\ \os\ systems listed in Table 1.
256: We define the system redshift \zabs\ by the position of the
257: strongest \os\ component in the system, i.e. the wavelength of maximum
258: optical depth. In each system, we looked for accompanying absorption
259: in \hi\ (Ly$\alpha$, Ly$\beta$, and Ly$\gamma$), \ct\ $\lambda977$,
260: \cf\ $\lambda\lambda1548,1550$, and \nf\ $\lambda\lambda1238,1242$ at
261: the same redshift \zabs.
262:
263: We denote the velocity displacement of each absorber from the quasar as
264: $\delta v\equiv c|z_{\rm qso}\!-\!z_{\rm abs}|/(1\!+\!z_{\rm abs})$.
265: %For simplicity, we refer to $\delta v$ as the proximity.
266: A negative value corresponds to an absorber moving toward the quasar.
267: The \os\ systems almost always contain multiple components clustered
268: together. We thus define $\delta v$ both for each absorber
269: and for each individual \os\ component.
270: The uncertainties on all velocity displacements are dominated by the
271: $\approx$500~\kms\ uncertainty of the quasar emission redshifts.
272:
273: \subsection{Classification of \os\ systems}
274: After identifying \nsys\ proximate \os\ systems, we
275: found that they could be classified into:
276:
277: (a) Seven {\bf strong} \os\ systems, which are defined by showing
278: both fully saturated \os\ lines extending over tens to hundreds of
279: \kms, and strong (often saturated) \nf\ and \cf\ absorption.
280: In practice we find that these seven systems all show
281: log\,$N$(\os)$>$15.0 and are at $\delta v\!<\!3\,000$~\kms, with four
282: of the seven at $\delta v<0$ (\zabs$>$\zqso), i.e. the strong
283: systems are strongly clustered around \zqso. %*
284: The strong systems are interpreted as being {\bf intrinsic} to the
285: quasar. Three of the strong systems are classified as
286: mini-BALs\footnote{The mini-BALs are the absorbers at \zabs=2.4426
287: toward HE~1158-1843, 2.1169 toward HE~1341-1020, and \zabs=2.9041
288: toward HE2347-4342.}, but no full (classical) BAL systems are
289: present in our sample. Partial coverage of the continuum source is
290: present in many of the strong systems, as evidenced by the failure
291: to find a successful Voigt profile fit to the \os\ profiles, even
292: though the velocity structure is common among the two members of the
293: doublet. This confirms an intrinsic origin, because partial coverage
294: of the continuum source only occurs when the absorbing gas is physically
295: close to the AGN. The strong systems contain multiple components, with
296: a mean of 5.9 components per system. %*
297:
298: (b) \nwea\ {\bf weak} \os\ systems, also known as {\bf narrow} systems,
299: which are unsaturated and show no evidence for partial
300: coverage, i.e. we were able to conduct
301: successful Voigt profile fits to the two lines of the \os\ doublet.
302: The weak systems all have 12.80$<$log\,$N$(\os)$<$14.50, clearly
303: separated from the strong systems, but coincident with the
304: range of \os\ column densities
305: observed in intervening systems (BH05). %*
306: Their total line widths are narrow, with a median and standard
307: deviation of $\Delta v_{90}$=42$\pm$33~\kms, %*
308: where $\Delta v_{90}$ is the line width encompassing the central 90\%
309: of the integrated optical depth.
310: The weak \os\ systems are accompanied by \cf\
311: absorption in 24/26 cases, and in 9/26 cases by \nf. %*
312: An mean of 1.9 components per system is required to
313: fit these weak absorbers.%*
314:
315: There were two intermediate cases that presented difficulties in their
316: classification: the systems at $z$=2.6998 and 2.7134 toward HE~0151-4326.
317: Neither of these systems shows a saturated \os\ profile, and their
318: \os\ column densities lie in-between the values shown by the strong
319: ($>$15.0) and weak ($<$14.5 populations). However, differences between
320: the profiles of the two doublet lines of \os, and between the two
321: lines of \nf, in similar velocity ranges are strongly suggestive of
322: partial coverage. This implies they should be classified as intrinsic.
323: For this reason we place these systems in the strong category, and our
324: final sample consists of \nstr\ strong systems and \nwea\ weak systems.
325:
326: In addition to the different \os\ (and \nf\ and \cf) column densities,
327: there are several lines of evidence that support the notion that the
328: strong and weak samples trace physically different populations.
329: The strong population have velocities (relative to \zqso) that
330: distribute differently, with 4/9 at \zabs$>$\zqso, whereas 22/26 weak
331: absorbers are at \zabs$<$\zqso. The strong population are show
332: markedly different \hi/\os\ ratios than the weak population (see
333: \S3.5). Finally absorbers in the strong population frequently show
334: evidence for partial coverage of the continuum source.
335: Whereas the strong absorbers are clearly associated with the
336: quasar \citep[e.g.][]{Tu84, We91, HF99}, either tracing QSO outflows
337: or inflows near the central engine of the AGN, the
338: origin of the weak \os\ absorbers is less clear. They are often called
339: `associated' because of their proximity in velocity to the quasar, but
340: there is not necessarily any physical connection.
341: %between the absorber and the AGN.
342: In the rest of this
343: paper we investigate the properties of this weak, proximate sample.
344:
345: \subsection{Component fitting}
346: We used the {\sc VPFIT} software
347: package\footnote{Available at http://www.ast.cam.ac.uk/$\sim$rfc/vpfit.html.}
348: to determine the column densities, line widths, and central velocities
349: of the \os\ components in each system.
350: The number of components to be fit was usually
351: self-evident from inspection of the line profiles, but in case of doubt we
352: used the minimum number of components necessary.
353: Our sample of \nsys\ \os\ systems comprises \ncomp\ components.
354: We also fit, when possible, the \hi, \nf, \ct, and \cf\ absorption,
355: independently of the \os\ fits, i.e. we did not fix any component
356: velocity centroids or line widths when fitting the various ions.
357: In the case of the \hi\ lines, we simultaneously
358: fit \lya\ with whichever of Ly$\beta$, Ly$\gamma$, and
359: Ly$\delta$ were unblended.
360: The total column densities in each ion in each system were obtained by
361: summing the component column densities, with the individual column
362: density errors added in quadrature to produce the overall error estimate.
363: All the Voigt profile fits are included in Figure 1, which shows the
364: absorption-line spectra for each system in our sample.
365: The total column density of each high ion and of \hi\ in each system
366: is given in Table 1.
367: The detailed results of the component fits to \os, \nf,
368: \cf, and \hi\ in the weak sample are given in Tables 2 and 3.
369: All velocities in Tables 2 and 3 are given on a scale relative to the
370: system redshift \zabs.
371:
372: In addition to the component fits, we measured the total column
373: densities in each system using the apparent
374: optical depth (AOD) method of \citet{SS91}. This method has the
375: advantage of requiring no assumptions about the component structure:
376: it simply requires a choice of minimum and maximum velocities for the
377: integration. It will return accurate column densities so long as
378: saturated structure is not present.
379: Whereas saturation is obvious if resolved, because the flux goes to zero,
380: unresolved saturation can be difficult to detect.
381: In the case of a non-detection of \cf\ or \nf, we measured
382: the (null) equivalent width in this ion over the velocity range
383: in which \os\ is detected, and found the
384: 3$\sigma$ error on this measurement. We then converted this maximum
385: allowed equivalent width into a column density limit using a linear
386: curve-of-growth. For example, if the equivalent width measurement was
387: 0$\pm$10~m\AA, we derived a 3$\sigma$ column density limit
388: $N_{\rm lim}$ using ${\rm EW_{lim}}$=30~m\AA\ and the relation
389: $N_{\rm lim}=1.13\times10^{17}{\rm EW_{lim}}/\lambda^2f$
390: (with $N_{\rm lim}$ in \sqcm\ and $\lambda$ in \AA). Atomic data were
391: taken from \citet{Mo03}.
392:
393: In many of the strong absorbers, partial coverage of the continuum
394: source is present. Indeed, we use this as a criterion to classify an
395: absorber as intrinsic. Partial coverage is indicated by a similar
396: velocity structure in the two members of a doublet, but with
397: inconsistent optical depths in each ion.
398: In the case of partial coverage the equation
399: $F(v)=F_0e^{-\tau(v)}$ is no longer valid \citep[e.g.][]{HF99}, and a
400: velocity-dependent coverage fraction must be introduced. For these
401: cases, the column density derived from an AOD integration should be
402: considered as a lower limit.
403:
404: \section{Results}
405: In this section we present a discussion of the overall properties of
406: the population of proximate \os\ absorbers at $z$=2--3. Although we
407: discuss ionization processes, a detailed discussion of the physical
408: conditions in the individual systems that make up our sample is beyond
409: the scope of this paper. Some of the systems in our sample are
410: discussed in more detail elsewhere. These include the proximate
411: absorbers toward HE~2347-4342 \citep{Fe04}, and toward HE~1158-1843,
412: PKS~0329-255, and Q0453-423 \citep{DO04}.
413:
414: \subsection{Location of absorbers in velocity space}
415: The number of proximate \os\ absorption line systems seen in each
416: quasar spectrum varies between zero and five.
417: There are two quasar spectra in our sample without \os\ absorption
418: detected within 8\,000~\kms\ of \zqso:
419: HE~2217-2818 (\zqso=2.41) and Q0002-422 (\zqso=2.77).
420: There are two quasars with five proximate \os\ absorbers in their
421: spectra: PKS~0237-23 (\zqso=2.233) and HE~1341-1020 (\zqso=2.135).
422: To illustrate the location in velocity space of the absorbers, we show
423: in Figure 2 the velocity displacement of all absorbers, strong and weak, for
424: each quasar. Over the whole sample (16 spectra) we find a mean of
425: 0.6 strong proximate \os\ systems per sight line, and
426: 1.6 weak proximate \os\ absorbers per sight line. %*
427: However, the \os\ sample is incomplete, because of regions where
428: \os\ systems are missed due to blending by the Ly$\alpha$
429: forest. As we show in \S3.7, %*
430: when we calculate the corrected redshift path for the \os\ search, our
431: estimated \os\ completeness is 84\% down to a limiting equivalent
432: width of 8~m\AA. %*
433:
434: As a useful gauge of the velocity interval in which we expect to see
435: proximity effects, we can derive the distance from each quasar at
436: which the flux of
437: extragalactic background (EGB) radiation equals the flux of quasar
438: radiation, i.e. we can derive an estimate of the size of the quasar
439: sphere-of-influence. The monochromatic quasar ionizing luminosities
440: at the Lyman Limit
441: have been derived for ten of the sixteen quasars by \citet{Ro05},
442: who calculated $L_{912}$ by extrapolating the QSO
443: $B$-magnitudes assuming a spectral slope of $\alpha$=$-$0.5
444: (where $F_\lambda\!\sim\!\lambda^\alpha$).
445: The EGB flux at the Lyman Limit at $z$=2.5 has been calculated to be
446: $F_{\rm EGB}\!\approx\!6\!\times\!10^{-21}$~erg~cm$^{-2}$~s$^{-1}$~Hz$^{-1}$
447: \citep[][corresponding to
448: log\,$J_\nu$=$-$21.3~erg~cm$^{-2}$~s$^{-1}$~Hz$^{-1}$~sr$^{-1}$;
449: the values for the $z$=2 and 3 cases are not significantly different]{HM96}.
450: The distance $D$ at which $F_{\rm QSO}$=$F_{\rm EGB}$ is given by
451: $D$=$\sqrt(L_{912}/4\pi F_{EGB}$).
452: In the case where the velocities are Hubble flow-dominated, we can
453: convert our calculated values of $D$ to velocity displacement from the
454: quasar, using a value for the Hubble parameter calculated at the
455: redshift of each
456: quasar using $H^2(z)=H^2_0[\Omega_{\rm M}(1+z)^3+\Omega_\Lambda]$.
457: With our adopted cosmology, $H(2.5)$=246~\kms~Mpc$^{-1}$.
458: These results are shown in Table 4, and are included in Figure 2, %*
459: where it can be seen that if the velocity displacements from the QSO
460: are due to the Hubble flow, then the proximity effect should only
461: extend to $\sim$1\,000--2\,500~\kms. While peculiar velocities are to be
462: expected, this is still a useful order-of-magnitude estimate for the
463: size of the quasars' ionizing sphere-of-influence.
464:
465: \setcounter{figure}{1}
466: \begin{figure}
467: \includegraphics[width=9.2cm]{figures/f2.eps}
468: \caption{Graphical display of the location in velocity space of the
469: \os\ absorbers in this study, relative to their background quasar.
470: Weak absorbers are shown in red, and strong absorbers in blue.
471: The location of $\delta v$=0, i.e. \zabs=\zqso, is
472: shown with a solid black line. Negative values for
473: $v_{\rm qso}\!-\!v_{\rm abs}$ imply the absorber is approaching the
474: quasar. Dashed green lines show the velocity at which
475: $F_{\rm QSO}$=$F_{\rm EGB}$ assuming Hubble flow (see text).}
476: \end{figure}
477:
478: \setcounter{table}{3}
479: \begin{table}
480: \begin{minipage}{\columnwidth}
481: \caption{Size of quasar spheres-of-influence at 912~\AA.}
482: \begin{tabular}{lccc}
483: \hline
484: QSO & log\,$L_{912}$\footnote{Monochromatic luminosity at Lyman Limit,
485: derived from the $B$ magnitude and $F_\lambda\sim\lambda^{-0.5}$
486: \citep{Ro05}.}
487: & $D$\footnote{Distance at which $F_{\rm QSO}$=$F_{\rm EGB}$=
488: $6\!\times\!10^{-21}$~erg~cm$^{-2}$~s$^{-1}$~Hz$^{-1}$
489: at $z$=2.5 \citep{HM96}.}
490: & $\delta v$\footnote{Velocity separation from QSO at $D$ assuming Hubble
491: flow, with $H$($z$) calculated at \zqso. $H$($z$=2.5)=246~\kms~Mpc$^{-1}$.}\\
492: & (erg~s$^{-1}$~Hz$^{-1}$) & (Mpc) & (\kms)\\
493: \hline
494: HE~0001-2340 & 31.65 & 7.7 & 1700 \\
495: HE~2217-2818 & 31.99 & 11.4 & 2600 \\
496: PKS~0237-23 & 31.67 & 7.8 & 1700 \\
497: PKS~0329-255 & 31.58 & 7.1 & 1800 \\
498: PKS~1448-232 & 31.53 & 6.7 & 1400 \\
499: Q~0002-422 & 31.72 & 8.3 & 2200 \\
500: Q~0109-3518 & 31.82 & 9.3 & 2200 \\
501: Q~0122-380 & 31.63 & 7.5 & 1600 \\
502: Q~0329-385 & 31.28 & 5.0 & 1100 \\
503: Q~0453-423 & 31.71 & 8.2 & 2100 \\
504: \hline
505: \end{tabular}
506: \end{minipage}
507: \end{table}
508:
509: \subsection{Presence of other ions}
510: The presence and strength of absorption in other ionic species in the
511: proximate \os\ absorbers provides important information on the
512: ionization conditions. \hi\ absorption is detected in each of the
513: \nsys\ \os\ systems, always in \lya\ and occasionally up to Ly$\delta$.
514: Higher-order Lyman lines may also be present, but they lie deep in the
515: \lya\ forest where blending and low signal-to-noise aggravate the
516: search for detections. All but two of the
517: \os\ systems show \cf\ detections (the exceptions are the absorbers at
518: \zabs=2.2378 toward PKS~0237-23 and \zabs=2.6610
519: toward PKS~0329-255, both of which are weak \os\ systems with low
520: \os\ column densities).
521: \nstr/\nstr\ strong \os\ systems contain \nf,
522: whereas 9/\nwea\ weak \os\ systems contain \nf. %*.
523: Therefore while the presence of \nf\ is a feature of
524: all strong systems, it does not imply that an \os\ system is strong.
525: Finally, \ct\ absorption is seen in 13/\nwea\ weak systems, and 7/\nstr\ %*
526: strong systems. However, because \ct\ $\lambda$977 lies in the
527: \lya\ forest, it is often blended and hence not detectable
528: even if present, so these fractions should be treated as lower limits.
529:
530: \subsection{\os\ system column densities}
531: A plot of the total \os\ column densities in each absorber versus
532: proximity (Figure 3, top panel) shows the distinction
533: between the strong and weak populations.
534: The strong population all show log\,$N\!\ga\!15.0$
535: and are(with two exceptions) clustered around \zabs=\zqso,
536: whereas the weak population lies between the
537: detection limit of log\,$N\!\approx\!12.80$
538: (corresponding to a limiting equivalent width of 8~m\AA) and 14.5.
539: Once the strong systems have been removed, the underlying weak
540: population shows no dependence of $N$(\os) with proximity to the quasar.
541: To investigate this further, we split the sample into \os\ absorbers
542: at $\delta v\!<\!2\,000$~\kms\ and those at
543: $2\,000\!<\!\delta v\!<\!8\,000$~\kms. This division is motivated by
544: an observed enhancement in d$N$/d$z$ below 2\,000~\kms\ (\S3.7), as
545: well as this representing the expected size of the proximity zone (\S3.1).
546: A two-sided Kolmogorov-Smirnov (K-S) test shows that the probability
547: that the distribution of log\,$N$(\os) for absorbers at $<$2\,000~\kms\ is
548: different from those at $>$2\,000~\kms\ is only 17\%. %*
549: There is also no clear correlation between the quasar redshift
550: and the total \os\ column, as shown in the lower panel of Figure 3.
551: We do note that the ratio of weak systems to strong systems changes
552: from 6.3 (19/3) at $z\!<\!2.5$ to 1.2 (7/6) at $z\!>\!2.5$, but the
553: statistics are too small for us to investigate this further. %*
554:
555: \begin{figure}
556: \includegraphics[width=9.2cm]{figures/f3.eps}
557: \caption{Scatter plots comparing the {\it integrated} column density
558: (over all components) of the \os\ absorbers with their proximity
559: to the background quasar $\delta v\equiv v_{\rm qso}\!-\!v_{\rm abs}$
560: and with their redshift \zabs. An error of 500~\kms\ in $z_{\rm
561: qso}$ is assumed. Strong absorbers are shown in blue (lower limits),
562: with weak absorbers in red.
563: The distinction between the weak and strong populations is evident.
564: The dashed line shows the detection limit.
565: Among the weak absorbers there is no correlation between
566: $N$(\os) and $\delta v$, or between $N$(\os) and $z$.}
567: \end{figure}
568:
569: \subsection{\os\ component column densities and $b$-values}
570: Proximity effects may occur at the component level,
571: as well as the system level. To investigate this, we compare in Figure
572: 4 the component properties returned by {\sc VPFIT}
573: (column densities and line widths) of the
574: weak \os\ absorbers with $\delta v$. We include for comparison the
575: \os\ components from the intervening sample of BH05.
576: As with the system column densities, there is no trend for the
577: component \os\ column densities to correlate with proximity to the
578: quasar. %*
579:
580: \begin{figure}
581: \includegraphics[width=9.2cm]{figures/f4.eps}
582: \caption{Scatter plots comparing the column densities and line widths
583: of the \os\ {\it components} with their proximity to the background
584: quasar. The errors shown are returned by the {\sc VPFIT}
585: profile-fitting code. Included in green is the BH05 sample of
586: intervening \os\ components at $z$=2--2.5, plotted for convenience
587: at $\delta v>9\,000$~\kms, with arbitrary shifts in the x-direction
588: applied for clarity. Two proximate components with $b$=34$\pm$14 and
589: 45$\pm$18~\kms\ are not shown because of their large errors.} %*
590: \end{figure}
591:
592: The distributions of log\,$N$ and $b$ for the \os\ components
593: in our sample are shown in Figure 5. We include the following
594: component distributions as separate curves: the components at $\delta
595: v\!<\!2\,000$~\kms, those at $2\,000\!<\!\delta v\!<\!8\,000$~\kms,
596: and the intervening absorbers from BH05.
597: A two-sided K-S test finds the probability that the log\,$N$ distributions
598: for the samples at $<$2\,000~\kms\ and
599: 2\,000--8\,000~\kms\ are different (i.e. drawn from
600: distinct parent populations) is only 23\%. %*
601: Comparing the $\delta v\!<\!2\,000$~\kms\ components to the
602: intervening sample, we do not find any difference valid at more than
603: the 1.0$\sigma$ level, %*
604: with the intervening sample showing a median log\,$N$(\os) of 13.54,
605: and the $\delta v\!<\!2\,000$~\kms\ sample showing a median of 13.45.
606: This is an important result considering the intervening sample was
607: measured independently in BH05, not in this analysis. %*
608:
609: \begin{figure}
610: \includegraphics[width=9.2cm]{figures/f5.eps}
611: \caption{
612: Top two panels: normalized column density- and line width
613: distributions of the proximate \os\ components
614: at $\delta v\!<\!2\,000$~\kms, at $2\,000\!<\!\delta v\!<\!8\,000$~\kms,
615: and in the intervening sample.
616: No statistically significant differences are found, with
617: the exception of a slight excess of systems with $b$=20--30~\kms\
618: in the weak sample in the 2\,000--8\,000~\kms\ bin. %*
619: The bottom panel compares the line width distributions for \os, \cf,
620: and \nf\ in the weak, proximate sample.}
621: \end{figure}
622:
623: Turning to the component line widths, we find that
624: among the proximate weak sample, 29 out of 48 components
625: (60\%) are narrow ($b\!\le\!14$~\kms)\footnote{We
626: define `narrow' as $b\!<\!14$~\kms\ since this corresponds to
627: the thermal line width of an \os\ line in gas at 188\,000~K, and
628: negligible \os\ is produced in gas in collisional ionization
629: equilibrium (CIE) below this temperature \citep{SD93, GS07}.}. %*
630: Comparing the $b$-value distributions, we find that the
631: probability that the population at $\delta v<2\,000$~\kms\
632: is different from the intervening distribution is less than 50 per cent. %*
633: The median $b$-value in the intervening sample is 12.7~\kms; in the
634: sample at $\delta v\!<\!2\,000$~\kms, it is 12.3~\kms. %*
635: However, an excess of systems with $b$=20--30~\kms\
636: is observed for components in the 2\,000--8\,000~\kms\ range, as can
637: be seen in the lower panel of Figure 4. %*
638:
639: Finally, we compare the line width distribution of the \os\ components
640: in the weak proximate sample with the \cf\ and \nf\ line width
641: distributions (in the same sample). The line widths of \cf\ and \nf\
642: distribute differently than those of \os, without the strong tail
643: extending between 10 and 25~\kms\ seen in the \os\ distribution.
644: 77\% of \cf\ components and 90\% of \nf\ components show %*
645: $b\!<\!14$~\kms. These differences indicate that it cannot be assumed
646: that the \cf, \nf, and \os\ arise in the same volumes of gas.
647:
648: In summary, among the weak \os\ absorbers at $z$=2--3, there is no
649: compelling evidence for any difference in the column density and line width
650: distributions between the intervening and proximate \os\ populations,
651: but within the proximate sample, the \os\ line widths distribute
652: differently than those of \nf\ and \cf. %*
653:
654: \subsection{\hi\ column density}
655: The median and standard deviation of the logarithmic \hi\
656: column density in the weak proximate \os\ sample is
657: log\,$N$(\hi)=14.66$\pm$0.95, with a total range covered in
658: log\,$N$(\hi) of over four decades. The \hi/\os\ ratio
659: (integrated over all components in each system)
660: varies enormously, taking values between 0.02 and 1500. %*
661: In Figure 6 (top panel) we plot the \hi/\os\ column density ratio
662: against the \hi\ column density. Within both the strong and weak
663: samples, a trend is evident in which \hi/\os\ rises with increasing
664: $N$(\hi). However, the two populations are clearly distinct, in that
665: the weak systems show much higher \hi/\os\ ratios at a given
666: $N$(\hi). The clear segregation of the weak and strong systems on this plot
667: supports our contention that they represent two separate
668: populations.
669:
670: \begin{figure}
671: \includegraphics[width=9.2cm]{figures/f6.eps}
672: \caption{Analysis of $N$(\hi) and the \hi/\os\ ratio in the proximate
673: \os\ sample. The column densities are integrated over all components
674: in each absorber. In the top panel, a dashed line delineates the locations
675: of the strong (blue) and weak (red) samples in a plot of \hi/\os\ vs
676: \hi. In the central two panels we show that both $N$(\hi) and \hi/\os\
677: show a dependence on proximity. The bottom panel shows the histogram
678: of $N$(\hi) for the weak \os\ absorbers, split into two sub-samples:
679: those at $\delta v\!<\!2\,000$ and $\delta v\!>\!2\,000$~\kms.
680: A significant difference exists between these two histograms, with
681: the absorbers closer in velocity to the quasar showing lower $N$(\hi).}
682: \end{figure}
683:
684: In the lower panels of Figure 6, we explore whether the
685: \hi\ column density in the weak proximate sample depends on $\delta v$.
686: We find that $N$(\hi) distributes very differently
687: at $\delta v\!<\!2\,000$~\kms\ than at $\delta v\!>\!2\,000$~\kms.
688: The $N$(\hi) distribution for the $\delta v\!>\!2\,000$~\kms\ sub-sample is
689: strongly clustered around log\,$N$(\hi)$\approx$15, whereas the distribution
690: for the $\delta v<2\,000$~\kms\ sub-sample is centered near
691: log\,$N$(\hi)$\approx$14 (see bottom panel). A K-S test shows the two
692: distributions differ at the 99.97\% level. %*
693: Consequently, there is also a difference in the mean \hi/\os\ ratio
694: above and below 2\,000~\kms, but this is due to the proximity effect in
695: $N$(\hi), not in $N$(\os), which shows no correlation with proximity
696: to the quasar.
697: The finding that $N$(\hi) correlates with $\delta v$ has an important
698: implication: it implies that $\delta v$ does correlate with physical
699: distance to the quasar. If $\delta v$ was dominated by peculiar
700: motions rather than Hubble flow, and so was uncorrelated to physical
701: proximity to the quasar, then the two $N$(\hi) distributions shown in the
702: bottom panel of Figure 6 would not be distinct.
703: The fact that they are so different allows us to make the
704: general assumption that proximity in velocity does imply proximity in
705: distance.%*
706:
707: \subsection{\cf\ and \nf\ column densities}
708: The logarithmic \cf\ column densities in the proximate \os\ sample
709: cover over 2~dex, with a median value and standard deviation of %*
710: log\,$N$(\cf)=12.64$\pm$0.60. \nf\ is detected in nine
711: weak absorbers, with log\,$N$(\nf) between 11.54 and 13.60. %*
712: %Among the weak proximate sample, we measure the median and standard
713: %deviation of the \cf/\os\ and \nf/\os\ ratios to be
714: %log\,[$N$(\cf)/$N$(\os)]=$-$0.81$\pm$0.66 and %*
715: %log\,[$N$(\nf)/$N$(\os)]=$-$1.81$\pm$0.70 %*
716: %(where the upper limits were excluded in forming these values, and the
717: %column densities are integrated over all components in each system).
718: %These high-ion column density ratios are of use in diagnosing physical
719: %conditions and ionization mechanisms, although their use depends on
720: %whether the various ions co-exist in the same volumes of gas (see \S4).
721: Comparisons between the \cf\ column density and $\delta v$, and
722: between \nf\ column density and $\delta v$, as well as the dependence
723: of the high-ion column density ratios with $\delta v$, are presented in
724: Figure 7. In a similar fashion to \hi, $N$(\cf) in the
725: weak \os\ absorbers depends strongly on $\delta v$, with the absorbers
726: at $\delta v\!<\!2\,000$~\kms\ showing column densities clustered
727: around log\,$N$(\cf)$\approx$12.4, and those at $\delta v\!>\!2\,000$~\kms\
728: showing values centered at log\,$N$(\cf)$\approx$13.3. A two-sided K-S %*
729: test shows that the $N$(\cf) distributions in the two velocity intervals
730: differ at a significance level of 93\%. %*
731: This leads to a difference in the mean \cf/\os\ ratio between the two
732: velocity bins, with lower ratios at lower velocity.
733: There is no indication that either $N$(\nf) or \nf/\os\
734: depends on $\delta v$, though the smaller sample size of \nf\
735: (because of the non-detections) prevents us from making statistically
736: meaningful conclusions.
737:
738: \begin{figure*}
739: \includegraphics[width=18cm]{figures/f7.eps}
740: \caption{Dependence of $N$(\os), $N$(\cf), $N$(\cf)/$N$(\os),
741: $N$(\nf), and $N$(\nf)/$N$(\os) on $\delta v$, for the weak
742: proximate \os\ sample. Downward-pointing arrows denote
743: non-detections. The right column shows the histograms of each
744: quantity plotted on the left, split into sub-samples at
745: $\delta v\!<\!2\,000$ and $>$2\,000~\kms\ from the quasar.
746: In the lower two panels, the nine \nf\ detections are shown with
747: open circles to distinguish them from the non-detections. Only
748: detections are included in forming the histograms.
749: Whereas there is no discernable effect for $N$(\os) to
750: depend on $\delta v$, $N$(\cf) distributes very differently at $<$2\,000
751: than at $>$2\,000~\kms. The \nf\ samples are too small to draw
752: conclusions about whether $N$(\nf) depends on $\delta v$.}
753: \end{figure*}
754:
755: In order to study the high-ion velocity structure more closely, we
756: present in Figures 8a and 8b apparent column density profiles as a
757: function of velocity, for each weak proximate \os\ absorber.
758: The apparent column density per unit velocity in units of
759: ions~\sqcm~(\kms)$^{-1}$ is given by
760: $N_a(v)=3.768\times10^{14}\tau_a(v)/f\lambda$ \citep{SS91},
761: where $\lambda$ is the
762: transition rest wavelength in \AA, $f$ is the oscillator strength
763: \citep[taken from][]{Mo03}, and $\tau_a(v)$ is defined in \S2.2. %*.
764: These figures allow the \os, \cf, \hi, and (in the nine cases where
765: present) \nf\ profiles to be closely compared. %*
766: Note that for \hi, we plot the apparent column density profile of our
767: best-fit model to the Lyman series absorption lines, because several
768: lines are used to derive our best-fit model. For the high
769: ions, the actual data from one of the two doublet lines are shown,
770: rather than the model.
771:
772: In 11/24 absorbers where both \os\ and \cf\ are detected, %*
773: the profiles are similar enough to suggest co-spatiality of the ions
774: C$^{+3}$ and O$^{+5}$. In the remaining 13 cases there are significant %*
775: differences in the line profiles, indicating that \os\ and \cf\ trace
776: different gas phases. Here a `significant difference' is either
777: a centroid offset of $>$10~\kms, or a case where
778: $b$(\cf)$<$\,$b$(\os), for which there is no single-phase solution
779: [since an oxygen atom is heavier than a carbon atom, $b$(\os) must be
780: less than $b$(\cf) for lines arising in the same gas].
781: % no matter what the non-thermal broadening. %*
782: In only one of the nine cases with \nf, its profile is similar in
783: centroid and width to the \os. However, detailed comparisons of the
784: \nf\ and \os\ profiles are complicated by the fact that low-optical
785: depth \nf\ absorption is lost in the noise, even in this high-quality
786: dataset, so we do not read into this result any further.
787: Most significantly, offsets of order 10--30~\kms\ %*
788: between the centroid of the \hi\ absorption and
789: the centroid of \os\ absorption are seen in 14/26 weak proximate cases. %*
790: These offsets are highly important, since they imply that the H$^0$
791: atoms do \emph{not} live in the same volume of gas as the O$^{+5}$
792: ions in at least half of the weak \os\ absorbers.
793: In other words, at least half of the weak \os\ systems are multi-phase.
794: Consequently, it is not possible to derive a metallicity estimate in
795: these multi-phase cases from the \os/\hi\ ratio, even with an
796: ionization correction, and single-phase photoionization models cannot
797: be applied. The dangers of treating multi-phase quasar
798: absorbers as single-phase have been pointed out before \citep{Gi94, Re01}.
799:
800: \begin{figure*}
801: \includegraphics[width=18cm]{figures/f8.0.eps}
802: \caption{High-ion and \hi\ apparent column density profiles for each
803: weak proximate \os\ absorber. The \cf\ (blue), \hi\ (black), and
804: \nf\ (green, where present) profiles have been scaled to allow their
805: shape to be compared with \os. For the high ions the AOD profiles
806: are formed from the data themselves. For \hi, we show the AOD
807: profile of the best-fit model to the Lyman series absorption
808: lines. The green vertical lines show the velocity range over which
809: the AOD measurement is made. Annotated on the top-right
810: of each panel is the proximity to the quasar. In cases where the
811: \os\ and \cf\ profiles show similar line centers and line widths,
812: we add a tick to the \cf\ label.
813: In cases where the \os\ and \hi\ profiles show similar
814: velocity centroids (but not necessarily the same width),
815: we add a tick mark to the \hi\ label.
816: Cross marks next to the \cf\ and \hi\ labels indicate significant
817: differences between the profiles of these ions and \os,
818: e.g. centroid offsets of $>$10~\kms, indicating
819: that the lines do not form in the same phase of gas as the \os.}
820: \end{figure*}
821:
822: \addtocounter{figure}{-1}
823: \begin{figure*}
824: \includegraphics[width=18cm]{figures/f8.1.eps}
825: \caption{(continued). High-ion and \hi\ apparent column density
826: profiles in the weak \os\ absorbers.}
827: \end{figure*}
828:
829: \subsection{Calculation of d$N$/d$z$ for proximate \os}
830: In this section we derive the incidence of proximate \os\
831: absorption, in terms of the traditional d$N$/d$z$,
832: and also in terms of the number of absorbers detected per 2\,000~\kms\
833: interval in $\delta v$. Note that because we are working at $z\!\gg\!0$,
834: relativistic effects cause the relationship between velocity and
835: redshift intervals to be d$v\!\approx\!c$\,d$z$/(1+$z$).
836: Before calculating the incidence of proximate \os\ absorption,
837: we correct the redshift path for blending with the
838: \lya\ forest, since \os\ is only identifiable in
839: unblended regions of the spectrum. To estimate the number of
840: \os\ absorbers that were missed due to \lya\ forest contamination, we
841: looked for \cf\ absorbers within 8\,000~\kms\ of \zqso\ in which
842: the accompanying \os\ is blended, and found five.
843: These are listed in Table 5.
844: This method should provide a good estimate for the \os\ incompleteness,
845: since firstly \os\ is detected in all proximate \cf\ absorbers in which the
846: \os\ data are unblended, and secondly \cf\ is unblended
847: by the \lya\ forest, so no \cf\ systems should be missed down to a
848: limiting equivalent width of $\approx$2~m\AA. %*.
849: The limiting equivalent width for the \cf\ search is smaller
850: than the limiting equivalent width for the \os\ search, because \cf\
851: lies in a higher S/N region of the spectrum.
852:
853: \begin{table}
854: \caption{Proximate \cf\ with blended \os}
855: \begin{tabular}{lccc}
856: \hline
857: QSO & \zqso\ & \zabs\ & $\delta v$ (\kms) \\ %& Notes\\
858: \hline
859: HE~0001-2340 & 2.2670 & 2.1870 & 7530 \\ %& ...\\
860: HE~0151-4326 & 2.7890 & 2.6956 & 7580 \\ %& ...\\
861: HE~1341-1020 & 2.1350 & 2.1065 & 2750 \\ %& ...\\
862: HE~2347-4342 & 2.8710 & 2.8781 & $-$550 \\ %& ...\\
863: Q~0122-380 & 2.2030 & 2.1472 & 5320 \\ %& ...\\
864: \hline
865: \end{tabular}
866: \end{table}
867:
868: The total redshift path within 8\,000~\kms\
869: of sixteen quasars (at $\langle z_{\rm qso}\rangle$=2.44) is 1.48.
870: Because there are five `missing' \os\ systems, we estimate our
871: \os\ completeness to be \nsys/(\nsys+5)=84\%, %*
872: implying that a path equal to 0.25 is blended near \os, %*
873: so that the corrected redshift interval for the proximate
874: \os\ search is 1.24. %*
875: %Finally, including the path between $-$1\,000 and 0~\kms\
876: %(in which four weak systems are found) increases the total
877: %unblended redshift interval to 1.39, in which \nwea\ weak systems
878: In this path \nwea\ weak systems (comprising 48 components) are detected. %*
879: so overall d$N$/d$z$ of weak \os\ systems
880: at $\delta v\!<\!8\,000$~\kms\ is 21$\pm$5, %**
881: where the errors on the counts are assumed to be Poissonian.
882: However, this is an average value, and when we break down the absorbers
883: into those at $\delta v\!<\!2\,000$ and those at 2\,000--8\,000~\kms,
884: we find a marked change in the incidence: d$N$/d$z$ \emph{trebles}
885: within 2\,000~\kms, %*
886: from 14$\pm$4 to 42$\pm$12\footnote{We have included the four weak \os\ %**
887: absorbers at $\delta v\!<\!0$ (\zabs$>$\zqso) in the 0--2\,000~\kms\ bin,
888: since regardless of their peculiar velocity
889: they must lie in front of the quasar.
890: Even without these, d$N$/d$z$ in the 0--2\,000\kms\ bin is 28$\pm$8, %*
891: twice the incidence in the 2\,000--8\,000~\kms\ bin.}. %*.
892: When counting components rather than systems, a similar increase in the
893: incidence is observed within 2\,000~\kms\ of \zqso.
894: These results are shown in Figure 9, and are also summarized in Table~6.
895: Note how the intervening incidence of BH05 is recovered in
896: the range 2\,000--8\,000~\kms.
897: % i.e. an enhancement in d$N$/d$z$
898: %due to quasar proximity is only seen within 2\,000~\kms.
899:
900: \begin{figure}
901: \includegraphics[width=9.2cm]{figures/f9.eps}
902: \caption{Number of weak \os\ absorbers as a function of proximity to
903: the quasar. Plotted are the number of \os\ systems per 2\,000~\kms\ bin of
904: $\delta v$ over our sample of sixteen quasars.
905: The corresponding values of d$N$/d$z$ are shown on the right,
906: corrected for the blended redshift path. The errors on the counts are
907: assumed to be Poissonian; plotted are the 1$\sigma$ errors.
908: We use a smaller bin for the range $-$1\,000 to
909: 0~\kms; these absorbers lie in front of the quasar and so are included
910: in the 0--2\,000~\kms\ bin in our final d$N$/d$z$ calculations. %**
911: The incidence of intervening \os\ systems is shown as the green data
912: point; there is no significant difference in d$N$/d$z$ between the
913: absorbers at 2\,000--8\,000~\kms\ and the intervening sample.}
914: \end{figure}
915:
916: \begin{table}
917: \begin{minipage}{\columnwidth}
918: \caption{Weak \os\ absorbers at $z$=2--3:
919: Summary of d$N$/d$z$ statistics (8~m\AA\ sensitivity)}
920: \begin{tabular}{lccc}
921: \hline
922: Category &
923: %$0\!<\!\delta v\!<\!2\,000$\footnote{Range in velocity
924: $|\delta v|\!<\!2\,000$\footnote{Range in velocity
925: relative to \zqso\ (in \kms).} &
926: $2\,000\!<\!\delta v\!<\!8\,000^{\rm a}$
927: & Intervening\footnote{Based on BH05; in the updated intervening sample
928: 58 systems are seen over a total $\Delta z$ of 4.09, complete down to
929: log\,$N$(\os)=13.11.}\\
930: \hline
931: Systems & 42$\pm$12 & 14$\pm$4 & 14$\pm$2 \\ %**
932: Components & 87$\pm$17 & 23$\pm$5 & 35$\pm$6 \\ %**
933: \hline
934: \end{tabular}
935: \end{minipage}
936: \end{table} %*
937:
938: \subsection{Comparison of $z\approx2.5$ and $z\approx0$ absorbers}
939: A considerable amount of work has been invested in studying \os\ in
940: the low-redshift IGM using the \emph{FUSE} and \emph{HST} orbiting
941: observatories \citep{Be94, BT96, Tr96, Tr00, Sa02, Se04, Ri04, Pr04,
942: DS05, DS08, Le06, Tr08, Co08, TC08a, TC08b}.
943: By comparing our results obtained
944: from proximate \os\ absorbers at $z$=2--3 with the $z\!\approx\!0$ results,
945: we can search for evolution of the highly ionized gas content of the
946: Universe. Since the intensity of the extragalactic UV background rises
947: with redshift
948: \citep{HM96}, one might expect to see higher photoionized \os\ column
949: densities at higher redshift. Simultaneously, the fraction of baryons
950: in warm-hot gas is predicted to decrease with increasing redshift
951: \citep{Da01}, since at early cosmic epochs there has not been adequate
952: time for gas to accrete and virialize into dark matter-dominated
953: potential wells.
954:
955: Based on an analysis of low-redshift ($0.15\!<\!z\!<\!0.5$) \os\ absorption
956: line systems and with a sensitivity limit of $W_{\rm r}\!>\!30$~m\AA,
957: \citet{Tr08} report that d$N$/d$z$ is enhanced by a factor of $\approx$3 %*
958: at velocities $<$2\,500~\kms\ from the quasar compared to the
959: intervening sample, with
960: d$N$(\os)/d$z$=16$^{+3}_{-2}$ for the intervening systems and
961: d$N$(\os)/d$z$=51$^{+21}_{-15}$ for those at $\delta v<2\,500$~\kms.
962: Our result for weak \os\ systems at $z$=2--3
963: with $W_{\rm r}\!>\!30$~m\AA\ and $\delta v\!<\!2\,000$~\kms\ is
964: d$N$/d$z$ is 32$\pm$10, %**
965: lower than the low-redshift result by a factor of $\approx$1.6. %**
966: However, comparing the high- and low-redshift results for d$N$/d$z$
967: can be misleading, because redshift has a highly non-linear
968: relationship with spatial dimensions, that depends on cosmology.
969: With our adopted cosmology, then in the interval $z$=0.15 to $z$=0.5,
970: d$x$(Gpc)=3.63\,d$z$, whereas for the interval $z$=2 to $z$=3,
971: d$x$(Gpc)=1.22\,d$z$, %*
972: where d$x$ is the comoving radial distance, and where
973: we have used the online cosmology calculator of \citet{Wr06}.
974: Correcting for these factors to calculate the number density of weak \os\
975: absorbers per unit co-moving distance, a better indicator of the space
976: density of \os, we derive the results listed in Table 7.
977: Here we see evidence that the space density of \os\ absorbers
978: \emph{increases} with redshift, by a factor of two between
979: $z$=0.15--0.5 and $z$=2--3. This increase is observed independently
980: for both the proximate and intervening absorbers. %**
981:
982: \begin{table}
983: \begin{minipage}{\columnwidth}
984: \caption{d$N$(\os)/d$x$ statistics (30~m\AA\ sensitivity)}
985: \begin{tabular}{lccc}
986: \hline
987: Category\footnote{The entries in this table list the incidence of weak \os\
988: absorbers per unit comoving distance, in units of Gpc$^{-1}$;
989: the entries were derived by correcting d$N$/d$z$ for cosmology, and
990: are presented for both proximate and intervening
991: absorbers, at low and high redshift, and for systems and components.}
992: & Sample & $z$=0--0.5\footnote{Derived using results from \citet{Tr08}.}
993: & $z$=2--3 \\
994: & & (Gpc$^{-1}$) & (Gpc$^{-1}$)\\
995: \hline
996: Systems & Proximate ($\delta v\!<\!2\,000$)
997: & 14$^{+6}_{-4}$ %Tr08
998: & 26$\pm$8\\ %**
999: & Intervening ($\delta v\!>\!5\,000$)
1000: & 4$\pm$1 %Tr08
1001: & 8$\pm$2\footnote{Derived from BH05 result.} \\ %*
1002: \hline
1003: Components & Proximate ($\delta v\!<\!2\,000$)
1004: & 30$^{+8}_{-6}$ %Tr08
1005: & 71$\pm$14\\ %**
1006: & Intervening ($\delta v\!>\!5\,000$)
1007: & 6$\pm$1 %Tr08
1008: & 21$\pm$3$^{\rm c}$\\ %*
1009: \hline
1010: \end{tabular}
1011: \end{minipage}
1012: \end{table}
1013:
1014: \section{Discussion}
1015: The BAL, mini-BAL, and intrinsic systems, which together we classify
1016: as strong systems based on their fully saturated \os\ absorption, strong
1017: accompanying \os\ and \cf\ absorption, and frequent evidence for
1018: partial coverage, are well-understood as being formed in either
1019: QSO-driven outflows or inflows in the immediate vicinity of
1020: the AGN central engine \citep[see review by][]{HF99}.
1021: Here we focus on the weak (narrow) proximate systems, which represent
1022: a separate population. Is there any evidence that these weak systems
1023: are directly photoionized by the quasar?
1024:
1025: We report two results that %could be referred to as `\os\ proximity effects',
1026: address this question, though their interpretation is not straightforward.
1027: First, in our sample of proximate \os\ absorbers at $z$=2--3, there is an
1028: enhancement by a factor of three %*
1029: in the incidence d$N$/d$z$ of weak \os\ absorbers within 2\,000~\kms\ of the
1030: quasar versus those in the interval 2\,000 to 8\,000~\kms.
1031: A similar enhancement is seen near low-redshift quasars \citep{Tr08}.
1032: While at face value this could be interpreted as a proximity effect,
1033: in which quasars preferentially ionize nearby
1034: clouds more often than they photoionize more distant clouds, this is
1035: not the only explanation. The enhancement in d$N$/d$z$
1036: could also be explained by an over-density
1037: of galaxies near quasars, and where the \os\ absorbers are located in
1038: the gaseous halos of these galaxies \citep{Yo82}.
1039: It is well-known that quasars are preferentially formed in cluster
1040: environments, particularly at high redshift \citep[][]{Cr05, Sh07}.
1041: This idea is supported by the observation that the internal properties
1042: of the weak \os\ absorbers are not dependent on redshift or proximity to
1043: the quasar.
1044:
1045: Second, there are statistically significant differences %*
1046: between the \hi/\os\ and \cf/\os\ ratios measured in the weak \os\
1047: populations above and below 2\,000~\kms, with the weak
1048: absorbers at $\delta v\!<\!2\,000$~\kms\ showing a median \hi/\os\ ratio
1049: lower by a factor of $\approx$10, and a median \cf/\os\ ratio lower by a %*
1050: factor of $\approx$7 than those at $2\,000\!<\!\delta v\!<\!8\,000$~\kms. %*
1051: Again, at face value, these results could be interpreted as
1052: proximity effects, in which gas closer to the quasar shows a higher
1053: ionization level than intervening gas.
1054: However, closer inspection finds that while
1055: $N$(\hi) and $N$(\cf) show a tendency to decrease as $\delta v$
1056: decreases, $N$(\os) is uncorrelated with $\delta v$, i.e.
1057: it is the behaviour of $N$(\hi) and $N$(\cf) alone that is driving the
1058: trends seen in the column density ratios, not the behaviour of $N$(\os).
1059: Importantly, the \hi\ and \cf\ proximity effects imply that proximity
1060: in velocity does correlate with proximity in distance, which supports
1061: the idea that the velocities of the weak absorbers are dominated by
1062: the Hubble flow.
1063:
1064: %Because the \os\ column density distribution is independent of
1065: %proximity, and
1066: Because \os-\hi-\cf\ velocity centroid offsets are
1067: observed directly in $\approx$50 per cent of the weak systems,
1068: single-phase photoionization models for the \os, \cf, and \hi\ are
1069: inadequate for at least half the absorbers in our sample.
1070: Indeed, the non-dependence of $N$(\os) on proximity casts doubt on
1071: whether photoionization by the background QSO creates the \os\ at all.
1072: Photoionization by nearby stellar sources of radiation is
1073: also unlikely, since such sources do not emit sufficient fluxes of
1074: photons above 54\,eV (the {\mbox{He\,{\sc ii}}} ionization edge)
1075: to produce the observed quantities of \os.
1076: We cannot rule out photoionization by the quasar in every
1077: individual case, for example the absorber at $z$=2.4183 toward
1078: HE~1122-1649, which shows strong \os\ but barely detectable \cf, as
1079: do several proximate \os\ absorbers reported by \citet{Go08}.
1080: Nonetheless, $N$(\os) does not depend on proximity in the way that $N$(\hi)
1081: and $N$(\cf) do. Thus we infer that the sizes of spheres-of-influence around
1082: quasars, which are implied the shapes of Gunn-Petersen troughs
1083: \citep{ZD95, Sm02}, and from studies of proximate absorption in lower
1084: ionization species, such as \mgii\ \citep{Va08, Wi08} and \cf\
1085: \citep*{Fo86, Ve03, Ne08}, are dependent on photon energy, and
1086: the presence of a sphere of photons at energies above 113.9~eV (capable of
1087: ionizing O$^{+4}$ to O$^{+5}$) has yet to be demonstrated. %*
1088: %We ran a series of CLOUDY \citep[v96.01;][]{Fe98} photoionization
1089: %models to several test cases where the \os\ and \hi\ line profiles are
1090: %consistent with a single-phase model. These models assume the gas
1091: %exists in a uniform density slab illuminated with an input QSO
1092: %ionizing spectrum. We found the values of [O/H], [N/O], and ionization
1093: %parameter log\,$U$ that best reproduce the observed column densities
1094: %of \os, \cf, and \hi, assuming [O/C]=0 and taking the solar elemental
1095: %abundances from \citet{Gr07}.
1096: %For the system at \zabs=2.4183 toward HE~1122-1649, we find
1097: %[O/H]$\approx-0.5$, [N/O]$\la-0.3$, log\,$U\approx-0.2$.
1098: %For the system at \zabs=2.2298 toward PKS~0237-23, we find
1099: %[O/H]$\approx-0.2$, [N/O]$\la-0.1$, log\,$U\approx-0.7$.
1100: %For the system at \zabs=2.2363 toward PKS~0237-23, we find
1101: %[O/H]$\approx-1.0$, [N/O]$\la-1.0$, log\,$U\approx-0.5$.
1102: %These values of log\.$U$ are consistent with those found for
1103: %intervening systems by BH05
1104:
1105: Thus we turn to collisional ionization models.
1106: Since the line widths of a significant fraction
1107: ($\approx$60\%) of the components in the weak \os\ absorbers %*
1108: are low enough ($b\!<\!14$~\kms) to imply gas temperatures below
1109: 188\,000~K, collisional ionization {\it equilibrium} can be ruled out,
1110: because essentially no \os\ is produced in gas in CIE at these
1111: temperatures \citep{SD93, GS07}.
1112: Indeed, the narrow line widths of many intergalactic \os\
1113: absorbers at $z\!\approx\!2$ have led various authors to conclude that
1114: photoionization is the origin mechanism \citep{Ca02, Be02, Lv03, Be05,
1115: Re06, Lo07}.
1116: However, \emph{non-equilibrium} collisional ionization models cannot
1117: be ruled out so easily. Indeed one expects that collisionally ionized gas
1118: at `coronal' temperatures of a few $\times10^5$~K, where \os\
1119: is formed through collisions, will be in a non-equilibrium state.
1120: This is because the peak of the interstellar cooling curve exists at
1121: these temperatures, and so the cooling
1122: timescales are short. When the cooling times are shorter than the
1123: recombination timescales, `frozen-in' ionization can result at
1124: temperatures well below those at which the ions exist in equilibrium
1125: \citep{Ka73, SM76, EC86}, provided that there is a source of
1126: $\sim10^6$~K gas in the first place.
1127:
1128: There are at least two physical reasons why collisionally-ionized,
1129: million-degree regions of interstellar and intergalactic gas
1130: could arise in the high-$z$ Universe.
1131: The first is the (hot-mode) accretion and shock heating of gas falling into
1132: potential wells \citep{BD03, Ke05, DB06},
1133: a process which is incorporated into cosmological
1134: hydrodynamical simulations \citep{CO99, Da01, FB03, Ka05, CF06},
1135: and creates what is referred to as the Warm-Hot Intergalactic
1136: Medium.
1137: However, these models generically predict that the fraction of
1138: all baryons that exist in the WHIM rises from essentially zero at
1139: $z$=3 to 30--50\% at $z$=0, so \emph{little WHIM is expected at the
1140: redshifts under study here}.
1141: The second reason is the presence of galactic-scale outflows, which
1142: due to the energy input from supernovae are likely to contain (or even
1143: be dominated by) hot, highly ionized gas
1144: \citep[see recent models by][]{OD06, Fa07, KR07, Sa08}.
1145: There is strong observational evidence for outflows at redshifts of
1146: $\approx$2--3 \citep{He02}, including blueshifted absorption
1147: in the spectra of Lyman break galaxies \citep{Pe00, Pe02, Sh03},
1148: the presence of metals in the low-density,
1149: photoionized IGM \citep[the \lya\ forest; e.g.][]{Ar04, Ag05, Ag08},
1150: though see \citet{Sy07}, and the presence of super-escape velocity
1151: \cf\ components in the spectra of damped \lya\ (DLA) galaxies \citep{Fo07b}.
1152: In addition, collisionally ionized gas in
1153: galactic halos at high redshift has been seen directly through
1154: detections of \os\ and \nf\ components in DLAs \citep{Fo07a},
1155: albeit with much broader system velocity widths than in the proximate
1156: absorbers discussed here. \os\ absorbers with lower \hi\ column
1157: density may probe the outer reaches of such halos or `feedback zones'
1158: \citep[BH05;][]{Si02}.
1159:
1160: We explore the ability of non-equilibrium collisional ionization
1161: models to reproduce the data in our proximate \os\ sample in
1162: Figure 10, which shows the \cf/\os\ vs
1163: \nf/\os\ ratio-ratio plane. We take the isobaric non-equilibrium model
1164: at log\,$T$=5.00 (consistent with the observed \os\ component line
1165: widths) from \citet{GS07}, computed using solar abundances, and then find the
1166: values of [N/O] and [C/H] that are required to reproduce the
1167: observations. The models for 0.1 and 0.01 solar absolute abundances,
1168: and for the isochoric case, predict similar values.
1169: The non-equilibrium models are not capable of reproducing the observed
1170: ratios when solar relative elemental abundances are used.
1171: However, if [N/O] takes values between $-$1.8 and 0.4, %*
1172: and [C/O] between $-$1.9 and 0.6, it is possible to explain
1173: the observed column densities in each weak proximate system
1174: with a non-equilibrium collisional ionization model.
1175: Note that we have not corrected these model predictions for the effect
1176: of photoionization by the extragalactic background; the model
1177: represents the pure collisional ionization case.
1178: Hybrid collisional+photo-ionization models would help in making
1179: progress in this area \citep[see discussion in][]{Tr08}.
1180:
1181: \begin{figure}
1182: \includegraphics[width=9.2cm]{figures/f10.eps}
1183: \caption{Comparison of predictions of non-equilibrium collisional
1184: ionization models with the observed column density ratios in our weak
1185: proximate sample. Non-detections of \nf\ (\cf) lead to upper limits
1186: on the x- (y-) axis. We take the predictions from the isobaric
1187: non-equilibrium collisional ionization model of \citet{GS07} for an
1188: absorber at $T$=$10^5$~K (higher temperatures are ruled out by the component
1189: line widths). The model is fairly insensitive to changes in the
1190: overall abundance level, but is sensitive to non-solar {\it
1191: relative} abundances. The dashed green lines show the range of
1192: values of [N/O] and [C/O] that are required to fit the proximate
1193: data with this collisional ionization model.}
1194: \end{figure}
1195:
1196: \section{Summary}
1197: We have presented a study of proximate ($\delta v<8\,000$~\kms)
1198: \os\ absorbers in the spectra of sixteen quasars observed at high
1199: signal-to-noise and 6.6~\kms\ resolution with VLT/UVES.
1200: The quasars are at redshifts between 2.14 and 2.87. %*
1201: We found \nsys\ proximate \os\ absorbers
1202: comprising \ncomp\ components. We used component fitting
1203: software to determine the properties of the absorption,
1204: studied the statistical properties of the sample, and
1205: addressed the ionization conditions in the gas. Our study has
1206: produced the following results.
1207:
1208: \begin{enumerate}
1209: \item \Nstr\ of the \nsys\ \os\ systems are strong, with log\,$N(\os)\ga15.0$.
1210: These absorbers all show detections of \cf\ and \nf,
1211: and show either broad, fully saturated, mini-BAL-type
1212: \os\ absorption troughs extending over tens to hundreds of \kms,
1213: or evidence for partial coverage in the high-ion profiles.
1214: A mean of 6.0 \os\ components are seen in the strong
1215: absorbers. The strong absorbers are formed either in QSO-driven
1216: outflows or in the immediate vicinity of the quasar.
1217:
1218: \item \nwea\ of the \nsys\ \os\ systems are weak, and show
1219: log\,$N(\os)\!<\!14.5$. These systems are narrow, with a median
1220: velocity width $\Delta v_{90}$ of 42~\kms. %*
1221: The weak systems all contain detectable \hi, all but two show \cf,
1222: and 9/\nwea\ show \nf. We find a mean of 1.9 \os\ components per %*
1223: weak system. Approximately 60 per cent of %*
1224: these \os\ components have Doppler $b$-values $<$14~\kms.
1225: The gas in these narrow components is constrained by the line widths
1226: to be at $T\!<\!188\,000$~K.
1227:
1228: \item Among the weak sample there is no correlation between the \os\
1229: column density and the proximity to the quasar. This is true when
1230: using either the component-level or the system-level column
1231: densities. Two-sided K-S tests confirm that the column density
1232: distributions of the intervening and weak proximate \os\ samples are
1233: statistically indistinguishable. %*
1234: The \os\ column density also does not correlate with redshift.
1235:
1236: \item There is a difference between the line width distributions
1237: of the \os\ components versus the distributions of \cf\ and \nf, in
1238: the weak proximate sample. The \cf\ and \nf\ distributions are
1239: strongly peaked below $b$=10~\kms\ and do not show the tail of
1240: absorbers with $b$-values between 10 and 25~\kms\ seen in the
1241: \os\ distribution. %*
1242:
1243: \item There is a significantly lower $N$(\hi) in the weak \os\
1244: absorbers with $\delta v\!<\!2\,000$~\kms\ than in those with
1245: $\delta v\!>\!2\,000$~\kms.
1246: The difference in the median $N$(\hi) between these two samples is
1247: $\sim$1.0~dex. We interpret this as a manifestation of
1248: the \hi\ proximity effect, and it leads to a trend in which the \hi/\os\
1249: ratio is lower for absorbers within 2\,000~\kms\ of the QSO than for
1250: those at higher $\delta v$. An identical trend is observed for \cf,
1251: with the median $N$(\cf) lower by $\approx$0.8~dex at $\delta
1252: v\!<\!2\,000$~\kms\ than at $\delta v\!>\!2\,000$~\kms. %*
1253: Thus \cf\ and \hi\ behave differently than \os.
1254:
1255: \item Apparent column density profile comparisons show that \os\ and
1256: \cf\ show significantly different velocity structure in 13 of the 24
1257: weak \os\ absorbers where \cf\ is detected. %*
1258: Offsets between the centroid of the \hi\ model and the \os\
1259: components are seen in 14/26 weak proximate cases. %*
1260: \emph{Therefore at least half of the weak \os\ systems are multi-phase.}
1261:
1262: \item Down to a limiting rest equivalent width of 8~m\AA, the number
1263: density of weak \os\ systems within 2\,000~\kms\ of the background
1264: quasar is d$N$/d$z$=42$\pm$12 (87$\pm$17 for components). %**
1265: Between 2\,000 and 8\,000~\kms\ in $\delta v$, d$N$/d$z$ falls to
1266: 14$\pm$4 (23$\pm$5 for components), equal to the incidence of
1267: intervening \os\ measured by BH05. %**
1268: Thus an enhancement in d$N$/d$z$ by a factor of three %*
1269: is seen at $\delta v\!<\!2\,000$~\kms.
1270:
1271: \item We compare the incidence d$N$/d$z$ of \os\ absorbers at
1272: high ($z$=2--3) and low ($z$=0--0.5) redshift.
1273: After correcting for cosmology, the incidence of
1274: \os\ systems with $W_{\rm r}\!>\!30$~m\AA\ \emph{per unit comoving
1275: distance} at $z$=2--3 is twice the value at $z$=0--0.5. %**
1276: This increase is seen for both proximate and intervening absorbers.
1277:
1278: \item
1279: %$Because the \os\ column densities in the weak \os\ sample at
1280: %$z$=2--3 show no dependence on proximity to the quasar, and
1281: Because \os-\cf-\hi\ velocity offsets are observed in half the sample,
1282: single-phase photoionization models are (at least for these cases)
1283: unable to explain the data. Indeed, the non-dependence of $N$(\os) on
1284: proximity casts doubt on whether photoionization creates the \os\ at
1285: all. Instead, we propose that the weak \os\ absorbers
1286: trace collisionally-ionized regions that exist as part of
1287: multi-phase galactic or protogalactic structures.
1288: In order to reconcile the collisional
1289: ionization hypothesis with the narrow line widths measured in
1290: approximately 60 per cent of the \os\ components, the gas must be
1291: out of CIE. This can be explained by a scenario in which initially
1292: hot gas is cooling down through the coronal regime at $T\!\sim\!10^5$~K.
1293: The observed high-ion column density ratios and line widths in the
1294: proximate \os\ absorbers can be explained by models of non-equilibrium
1295: collisionally ionized gas at $T\!\approx\!10^5$~K, but only if the
1296: relative elemental abundances are non-solar: we require
1297: [N/O] to be in the range $-$1.8 to 0.4 and
1298: [C/O] between $-$1.9 and 0.6, to explain the data with these models. %*
1299:
1300: \item Future \os\ surveys do not need to remove absorbers (and thus
1301: compromise the sample size) that happen to lie at
1302: 2\,000--8\,000~\kms\ relative to the quasar. A better approach would
1303: be to remove those showing either $N$(\os)$\ga$15.0 or evidence for
1304: partial coverage.
1305:
1306: \end{enumerate}
1307:
1308: \section*{Acknowledgments}
1309: AJF gratefully acknowledges the support of a Marie Curie
1310: Intra-European Fellowship (contract MEIF-CT-2005-023720)
1311: awarded under the European Union's Sixth
1312: Framework Programme, which partially funded his contribution to this paper.
1313: We thank Bastien Aracil for his continuum fits to the UVES data, and
1314: Bob Carswell for making his {\sc VPFIT} software publically available.
1315: We thank the anonymous referee for useful comments.
1316: % and Todd Tripp for comments on the manuscript.
1317:
1318: \begin{thebibliography}{}
1319: \bibitem[Aguirre et al.(2005)]{Ag05} % IGM wind Z enrichment
1320: Aguirre A., Schaye J., Hernquist L., Kay S., Springel V.,
1321: Theuns T. 2005, ApJ, 620, L13
1322:
1323: \bibitem[Aguirre et al.(2008)]{Ag08} % IGM OVI POD
1324: Aguirre A., Dow-Hygelund C., Schaye J., Theuns T. 2008, ApJ,
1325: submitted (astro-ph/0712.1239)
1326:
1327: \bibitem[Anderson et al.(1987)]{An87}
1328: % Traditional def of ``associated'' systems
1329: Anderson S. F., Weymann R. J., Foltz C. B., Chaffee F. H. Jr
1330: 1987, AJ, 94, 278
1331:
1332: \bibitem[Aracil et al.(2004)]{Ar04}
1333: % pixel optical depth, 19 quasars,OVI/CIV/HI, log[tau(OVI)/tau(HI)]~-2.0
1334: Aracil B., Petitjean P., Pichon C., Bergeron J. 2004, A\&A,
1335: 419, 811
1336:
1337: \bibitem[Bajtlik, Duncan, \& Ostriker(1988)]{Ba88} % Prox eff
1338: Bajtlik S., Duncan R. C., Ostriker J. P. 1988, ApJ, 327, 570
1339:
1340: \bibitem[Ballester et al.(2000)]{Ba00} % UVES reduction pipeline
1341: Ballester P., Modigliani A., Boitquin O., Cristiani S., Hanuschik
1342: R., Kaufer A., Wolf S. 2000, The Messenger, 101, 31
1343:
1344: \bibitem[Barlow \& Sargent(1997)]{BS97}
1345: % AAL in radio-loud quasar
1346: Barlow T. A., Sargent W. L. W. 1997, AJ, 113, 136
1347:
1348: \bibitem[Bechtold(1994)]{Bh94} % Prox Eff
1349: Bechtold J. 1994, ApJS, 91, 1
1350:
1351: \bibitem[Bergeron et al.(1994)]{Be94}
1352: Bergeron J., et al. 1994, ApJ, 436, 33
1353:
1354: \bibitem[Bergeron et al.(2002)]{Be02}
1355: Bergeron J., Aracil B., Petitjean P., Pichon C. 2002, A\&A, 396, L11
1356:
1357: \bibitem[Bergeron \& Herbert-Fort(2005)]{Be05}
1358: % VLT O VI systems at 2.0--2.6 in 10 quasars
1359: Bergeron J., Herbert-Fort S. 2005, `Probing Galaxies through
1360: Quasar Absorption Lines', Proc. IAU Coll. No. 199, ed. Williams, Shu,
1361: M\'enard (BH05)
1362:
1363: \bibitem[Birnboim \& Dekel(2003)]{BD03} %Hot mode accretion
1364: Birnboim Y., Dekel A. 2003, MNRAS, 345, 349
1365:
1366: \bibitem[Burles \& Tytler(1996)]{BT96}
1367: Burles S., \& Tytler D. 1996, ApJ, 460, 584
1368:
1369: \bibitem[Carswell et al.(1982)]{Ca82}
1370: % Original discussion of the proximity effect
1371: Carswell R. F., Whelan J., Smith M., Boksenberg A., Tytler D., 1982,
1372: MNRAS, 198, 91
1373:
1374: \bibitem[Carswell et al.(2002)Carswell, Schaye, \& Kim]{Ca02}
1375: % O VI in Ly a forest at z~2 in two QSO spectra, large omega_b in systems
1376: Carswell B., Schaye J., Kim T.-S. 2002, ApJ, 578, 43
1377:
1378: \bibitem[Chand et al.(2004)]{Ch04} % Data reductoin
1379: Chand H., Srianand R., Petitjean P., Aracil B 2004, A\&A, 417, 853
1380:
1381: \bibitem[Churchill et al.(1999)]{Ch99}% Mini-BAL @ z>4
1382: Churchill C. W., Schneider D. P., Schmidt M., Gunn J. E. 1999,
1383: AJ, 117, 2573
1384:
1385: \bibitem[Cen \& Ostriker(1999)]{CO99} % Hydro code
1386: Cen R., Ostriker J. P. 1999, ApJ, 514, 1
1387:
1388: \bibitem[Cen \& Fang(2006)]{CF06}
1389: Cen R., Fang T. 2006, ApJ, 650, 573
1390:
1391: \bibitem[Cooksey et al.(2008)]{Co08}
1392: Cooksey K. L., Prochaska J. X., Chen H.-W., Mulchaey J. S.,
1393: Weiner B. J. 2008, ApJ, 676, 262
1394:
1395: \bibitem[Croom et al.(2005)]{Cr05}
1396: Croom S. M., et al. 2005, MNRAS, 356, 415
1397:
1398: \bibitem[Danforth \& Shull(2005)]{DS05}
1399: % OVI baryon census
1400: Danforth C. W., Shull J. M. 2005, ApJ, 624, 555
1401:
1402: \bibitem[Danforth \& Shull(2008)]{DS08}
1403: % OVI baryon census
1404: Danforth C. W., Shull J. M. 2008, ApJ, 679, 194
1405:
1406: \bibitem[Dav\'e et al.(2001)]{Da01}
1407: % Hydro code,30%of z=0baryons in WHIGM
1408: Dav\'e R. et al. 2001, ApJ, 552, 473
1409:
1410: \bibitem[Dekel \& Birnboim(2006)]{DB06} %Hot mode accretion
1411: Dekel A., Birnboim Y. 2006, MNRAS, 368, 2
1412:
1413: \bibitem[Dekker et al.(2000)]{De00}
1414: % UVES design
1415: Dekker H., D'Odorico S., Kaufer A., Delabre B., Kotzlowski
1416: H. 2000, SPIE, 4008, 534
1417:
1418: \bibitem[D'Odorico et al.(2004)]{DO04}
1419: % Intrinsic absorbers in UVES LP dataset
1420: D'Odorico V., Cristiani S., Romano D., Granato G. L., Danese
1421: L. 2004, MNRAS, 351, 976
1422:
1423: \bibitem[Edgar \& Chevalier(1986)]{EC86}
1424: Edgar R. J., Chevalier R. A. 1986, ApJ, 310, L27
1425:
1426: \bibitem[Fang \& Bryan(2003)]{FB03} % WHIM SIM
1427: Fang T., Bryan G. L. 2003, ApJ, 561, L31
1428:
1429: \bibitem[Fangano, Ferrara, \& Richter(2007)]{Fa07}
1430: % Wind Signatrues in Absorption
1431: Fangano A. P. M., Ferrara A., Richter P. 2007, MNRAS, 381, 469
1432:
1433: \bibitem[Faucher-Gigu\`ere et al.(2008)]{FG08}%Prox Effect
1434: Faucher-Gigu\`ere C.-A., Lidz A., Zaldarriaga M., Hernquist
1435: L. 2008, ApJ, 673, 39
1436:
1437: \bibitem[Fechner, Baade, \& Reimers(2004)]{Fe04}
1438: % HE 2347 associated system paper
1439: Fechner C., Baade R., Reimers D. 2004, A\&A, 418, 857
1440:
1441: %\bibitem[Ferland et al.(1998)]{Fe98}
1442: %Ferland, G. J., Korista, K. T., Verner, D. A., Ferguson, J. W., Kingdon,
1443: %J. B., \& Verner, E. M. 1998, \pasp, 110, 761
1444:
1445: \bibitem[Foltz et al.(1986)]{Fo86}
1446: % Definintion of 5000 km\s cutoff for assoc systems
1447: Foltz C. B, Weymann R. J., Peterson B. M., Sun L., Malkan M. A.,
1448: Chaffee F. H. 1986, ApJ, 307, 504
1449:
1450: \bibitem[Fox et al.(2007a)]{Fo07a} % OVI in DLAs
1451: Fox A. J., Petitjean P., Ledoux C., Srianand R. 2007a, A\&A,
1452: 465, 171
1453:
1454: \bibitem[Fox et al.(2007b)]{Fo07b} % CIV in DLAs
1455: Fox A. J., Ledoux C., Petitjean P., Srianand R. 2007b, A\&A,
1456: 473, 791
1457:
1458: \bibitem[Ganguly et al.(1999)]{Ga99}
1459: Ganguly R., Eracleous M., Charlton J. C., Churchill
1460: C. W. 1999, AJ, 117, 2594
1461:
1462: \bibitem[Ganguly et al.(2001)]{Ga01}
1463: Ganguly R., Bond N. A., Charlton J. C., Eracleous M., Brandt W. N.,
1464: Churchill C. W. 2001, ApJ, 549, 133
1465:
1466: \bibitem[Ganguly et al.(2006)]{Ga06}
1467: Ganguly R., Sembach K. R., Tripp T. M., Savage B. D., Wakker
1468: B. P. 2006, ApJ, 645, 868
1469:
1470: \bibitem[Giroux, Sutherland, \& Shull(1994)]{Gi94} %not photoionized!
1471: Giroux M. L., Sutherland R. S., Shull J. M. 1994, ApJ, 435, L97
1472:
1473: \bibitem[Gnat \& Sternberg(2007)]{GS07}
1474: % New CIE/rad cooling Models
1475: Gnat O., Sternberg A. 2007, ApJS, 168, 213
1476:
1477: \bibitem[Gon\c{c}alves et al.(2008)Gon\c{c}alves, Steidel, \& Pettini]{Go08}
1478: Gon\c{c}alves T. S., Steidel C. C., Pettini M. 2008, ApJ, 676, 816
1479:
1480: %\bibitem[Grevesse, Asplund, \& Sauval(2007)]{Gr07} Latest Abundances Solar
1481: %Grevesse N., Asplund M., Sauval A. J. 2007, Space Sci. Rev., 130, 105
1482:
1483: \bibitem[Guimar\~aes et al.(2007)]{Gu07}
1484: % Proximity effect and references
1485: Guimar\~aes R., Petitjean P., Rollinde E., de Carvalho R. R.,
1486: Djorgovski S. G., Srianand R., Aghaee A. , Castro S. 2007,
1487: MNRAS, 377, 657
1488:
1489: \bibitem[Haardt \& Madau(1996)]{HM96} %EGB models
1490: Haardt F., Madau P. 1996, ApJ, 461, 20
1491:
1492: \bibitem[Hamann(1997)]{Ha97}
1493: % Ionization in Associated Systems, also see intro
1494: Hamann F. 1997, ApJS, 109, 279
1495:
1496: \bibitem[Hamann et al.(1997a)]{Ha97a}
1497: % Keck spectra of UM 675 intrinsic absorbers
1498: % intro: discussion of why many proximate absorbers are intrinsic
1499: Hamann F., Barlow T. A., Junkkarinen V., Burbidge E. M. 1997a,
1500: ApJ, 478, 80
1501:
1502: \bibitem[Hamann, Barlow, \& Junkkarinen (1997b)]{Ha97b}
1503: % High v Intrinsic system (24 000 km/s)
1504: Hamann F., Barlow T. A., Junkkarinen V., 1997b,
1505: ApJ, 478, 87
1506:
1507: \bibitem[Hamann et al.(1997c)]{Ha97c}
1508: % High ions in z_ab ~ z_em system, also see intro
1509: Hamann F., Beaver E. A., Cohen R. D., Junkkarinen V., Lyons
1510: R. W., Burbidge E. M. 1997c, ApJ, 488, 155
1511:
1512: \bibitem[Hamann \& Ferland(1999)]{HF99}
1513: %ARAA on chemical abundances in QSOs
1514: Hamann F., Ferland G. 1999, ARA\&A, 37, 487
1515:
1516: \bibitem[Hamann, Netzer, \& Shields(2000)]{Ha00}
1517: Hamann F., Netzer H., Shields J. C. 2000, ApJ, 536, 101
1518:
1519: \bibitem[Hamann et al.(2001)]{Ha01}
1520: % Keck spectrum of z_ab ~ z_em system ** NICE INTRO
1521: Hamann F., Barlow T. A., Chaffee F. C., Foltz C. B.,
1522: Weymann R. J. 2001, ApJ, 550, 142
1523:
1524: \bibitem[Heckman(2002)]{He02}
1525: %Gal superwinds
1526: Heckman T. 2002, ASPC, 254, 292
1527:
1528: \bibitem[Kafatos(1973)]{Ka73}
1529: % Overionized gas that has cooled
1530: Kafatos M. 1973, ApJ, 182, 433
1531:
1532: \bibitem[Kang et al.(2005)]{Ka05} % New WHIM models
1533: Kang H., Ryu D., Cen R., Song D. 2005, ApJ, 620, 21
1534:
1535: \bibitem[Kawata \& Rauch(2007)]{KR07}
1536: %Galactic Wind Signatures around High Redshift Galaxies
1537: Kawata D., Rauch M. 2007, ApJ, 663, 38
1538:
1539: \bibitem[Kere{\v s} et al.(2005)]{Ke05} % Hot accretion
1540: Kere{\v s} D., Katz N., Weinberg D. H., Dave R. 2005, MNRAS, 363, 2
1541:
1542: %\bibitem[Kirkman \& Tytler(1997)]{KT97}
1543: % OVI and CIV in LLS I
1544: %Kirkman D., Tytler D. 1997, ApJ, 489, L123
1545:
1546: \bibitem[Lehner et al.(2006)]{Le06} %DN/dz HE0226 OVI
1547: Lehner N., Savage B. D., Wakker B. P., Sembach K. R., Tripp
1548: T. M. 2006, ApJS, 164, 1
1549:
1550: \bibitem[Levshakov et al.(2003)]{Lv03}
1551: % Photoionized O VI toward HE 0515-4414 at z<1.7
1552: Levshakov S. A., Agafonova I. I., Reimers D., Baade R. 2003, A\&A,
1553: 404, 449
1554:
1555: \bibitem[Lopez et al.(2007)]{Lo07} % IGM O VI IN LENSED QSOs
1556: Lopez S., Ellison S., D'Odorico S., Kim T.-S. 2007, A\&A, 469, 61
1557:
1558: \bibitem[Lu \& Savage(1993)]{LS93}
1559: Lu L., Savage B. D. 1993, ApJ, 403, 127
1560:
1561: \bibitem[Lu, Wolfe, \& Turnshek(1991)]{Lu91}
1562: Lu L., Wolfe A. M., Turnshek D. A. 1991, ApJ, 367, 19
1563:
1564: \bibitem[Misawa et al.(2007a)]{Mi07a}
1565: %10-17% of C IV absorbers at 5000-70,000 km s-1 relative to QSO are intrinsic
1566: Misawa T., Charlton J. C., Eracleous M., Ganguly R., Tytler D.,
1567: Kirkman D., Suzuki N., Lubin D. 2007a, ApJS, 171, 1
1568:
1569: \bibitem[Misawa et al.(2007b)]{Mi07b} %Mini BAL
1570: Misawa T., Eracleous M., Charlton J. C., Kashikawa N. 2007b, ApJ, 660,
1571: 152
1572:
1573: \bibitem[M\o ller, Jakobsen, \& Perryman(1994)]{Mo94} % AAL
1574: M\o ller P., Jakobsen P., Perryman M. A. C. 1994, A\&A, 287, 719
1575:
1576: \bibitem[M\o ller, Warren, \& Fynbo(1998)]{Mo98} % Proximate Sys
1577: M\o ller P., Warren S. J., Fynbo J. U. 1998, A\&A, 330, 19
1578:
1579: \bibitem[Morton(2003)]{Mo03} % Atomic Data
1580: Morton D. C. 2003, ApJS, 149, 205
1581:
1582: \bibitem[Murdoch et al.(1986)]{Mu86}
1583: % Original discussion of the proximity effect
1584: Murdoch H. S., Hunstead R. W., Pettini M., Blades J. C., 1986,
1585: ApJ, 309, 19
1586:
1587: \bibitem[Narayanan et al.(2004)]{Na04} %AAL
1588: Narayanan D., Hamann F., Barlow T., Burbidge E. M.,
1589: Cohen R. D., Junkkarinen V., Lyons R. 2004, ApJ, 601, 715
1590:
1591: \bibitem[Nestor, Hamann, \& Rodr\'iguez Hidalgo(2008)Nestor et al.]{Ne08}
1592: % C IV proximity
1593: Nestor D., Hamann F., Rodr\'iguez Hidalgo P. 2008, MNRAS, 386, 2055
1594:
1595: \bibitem[Oppenheimer \& Dav\'e(2006)]{OD06}
1596: % Wind simulations of IGM metal enrichment
1597: Oppenheimer B., Dav\'e R. 2006, MNRAS, 373, 1265
1598:
1599: \bibitem[Petitjean, Rauch, \& Carswell(1994)]{Pe94}
1600: Petitjean P., Rauch M., Carswell R. F. 1994, A\&A, 291, 29
1601:
1602: \bibitem[Petitjean \& Srianand(1999)]{PS99}
1603: Petitjean P., Srianand R. 1999, A\&A, 345, 73
1604:
1605: \bibitem[Pettini et al.(2000)]{Pe00}
1606: % Wind from MS1512 Lyman Break Galaxy Paper I
1607: Pettini M., Steidel C. C., Adelberger K. L., Dickinson M.,
1608: Giavalisco M. 2000, ApJ, 528, 96
1609:
1610: \bibitem[Pettini et al.(2002)]{Pe02}
1611: % Wind from MS1512 Lyman Break Galaxy Paper II
1612: Pettini M., Rix S. A., Steidel C. C., Adelberger K. L., Hunt
1613: M. P., Shapley A. E. 2002, ApJ, 569, 742
1614:
1615: \bibitem[Prochaska et al.(2004)]{Pr04}
1616: %PKS0405 OVI IN IGM
1617: Prochaska J. X., Chen H., Howk J. C., Weiner B. J., Mulchaey
1618: J. 2004, ApJ, 617, 718
1619:
1620: \bibitem[Rauch(1998)]{Ra98}
1621: Rauch M. 1998, ARA\&A, 36, 267
1622:
1623: \bibitem[Reimers et al.(2001)]{Re01}
1624: % O VI toward HE 0515 (z=1.2--1.7) not found in the same phase as C IV
1625: Reimers D., Baade R., Hagen H.-J., Lopez S. 2001, A\&A, 374, 871
1626:
1627: \bibitem[Reimers et al.(2006)]{Re06}
1628: % Spectral shape of the UV bck and O VI absorbers at z1.5 to HS 0747+4259
1629: Reimers D., Agafonova I. I., Levshakov S. A., Hagen H.-J.,
1630: Fechner C., Tytler D., Kirkman D., Lopez S. 2006, A\&A, 449, 9
1631:
1632: \bibitem[Richter et al.(2004)]{Ri04} %PG1259 paper
1633: Richter P., Savage B. D., Tripp T. M., Sembach K. R., ApJS, 153, 165
1634:
1635: \bibitem[Richards(2001)]{Ri01} % C IV at v_displ > 5000 may be assoc
1636: Richards G. 2001, ApJS, 133, 53
1637:
1638: \bibitem[Richards et al.(1999)]{Ri99} % C IV at v_displ > 5000 may be assoc
1639: Richards G., York D. G., Yanny B., Kollgaard R. I.,
1640: Laurent-Muehleisen S. A., Vanden Berk D. E., 1999, ApJ, 513, 576
1641:
1642: \bibitem[Rollinde et al.(2005)]{Ro05}
1643: % Large Program: Determined z_QSOs
1644: Rollinde E., Srianand R., Theuns T., Petitjean P., Chand
1645: H. 2005, MNRAS, 361, 1015
1646:
1647: \bibitem[Samui, Subramanian, \& Srianand(2008)]{Sa08}
1648: Samui S., Subramanian K., Srianand R. 2008, MNRAS, 385, 783
1649:
1650: \bibitem[Savage \& Sembach(1991)]{SS91} % AOD I
1651: Savage B. D., Sembach K. R. 1991, ApJ, 379, 245
1652:
1653: \bibitem[Savage et al.(2002)]{Sa02} %PG0953 dn/dz
1654: Savage B. D., Sembach K. R., Tripp T. M., Richter P. 2002, ApJ,
1655: 564, 631
1656:
1657: \bibitem[Savaglio, D'Odorico, \& M\o ller(1994)]{Sv94}
1658: Savaglio S., D'Odorico S., M\o ller P. 1994, A\&A, 281, 331
1659:
1660: \bibitem[Scannapieco et al.(2006)]{Sc06}% z_QSOs
1661: Scannapieco E., Pichon C., Aracil B., Petitjean P., Thacker
1662: R. J., Pogosyan D., Bergeron J. Couchman H. M. P. 2006 MNRAS,
1663: 365, 615
1664:
1665: \bibitem[Schaye et al.(2000)]{Sy00}
1666: Schaye J., Rauch M., Sargent W. L. W., Kim T.-S 2000, ApJ, 541, L1
1667:
1668: \bibitem[Schaye, Carswell, \& Kim(2007)]{Sy07}
1669: Schaye J., Carswell R. F., Kim T.-S 2007, MNRAS, 379, 1169
1670:
1671: \bibitem[Scott et al.(2000)]{Sc00} % Proc Effect
1672: Scott J., Bechtold J., Dobrzycki A., Kulkarni V. P. 2000, ApJS, 130, 67
1673:
1674: \bibitem[Sembach et al.(2004)]{Se04}
1675: Sembach K. R., Tripp T. M., Savage B. D., Richter P. 2004,
1676: ApJS, 155, 351
1677:
1678: \bibitem[Shapiro \& Moore(1976)]{SM76}
1679: Shapiro P. R., Moore R. T. 1976, ApJ, 207, 460
1680:
1681: \bibitem[Shapley et al.(2003)]{Sh03}
1682: % Rest-Frame Ultraviolet Spectra of z~3 Lyman Break Galaxies
1683: Shapley A. E., Steidel C. C., Pettini M., Adelberger
1684: K. L. 2003, ApJ, 588, 65
1685:
1686: \bibitem[Shen et al.(2007)]{Sh07} %QSO clustering
1687: Shen Y., et al. 2007, AJ, 133, 2222
1688:
1689: \bibitem[Simcoe et al.(2002)Simcoe, Sargent, \& Rauch]{Si02}
1690: % Keck O VI at z=2.5 N(HI)>15.2 =in 5 SLs
1691: Simcoe R. A., Sargent W. L. W., Rauch M. 2002, ApJ, 578, 737
1692:
1693: \bibitem[Simcoe et al.(2004)Simcoe, Sargent, \& Rauch]{Si04}
1694: % Keck O VI/CIV at z=2.5 N(HI)>13.6 in 7 SLs
1695: Simcoe R. A., Sargent W. L. W., Rauch M. 2004, ApJ, 606, 92
1696:
1697: \bibitem[Simcoe et al.(2006)]{Si06}
1698: % Keck O VI/CIV and imaging in 1 SL (IGM high ion ratios) + G wind discussion
1699: Simcoe R. A., Sargent W. L. W., Rauch M., Becker G.
1700: 2006, ApJ 637, 648
1701:
1702: \bibitem[Smette et al.(2002)]{Sm02}
1703: % QSO sphere of influence
1704: Smette A., Heap S. R., Williger S. M., Tripp T. M., Jenkins
1705: E. B., Songaila A. 2002, ApJ, 564, 542
1706:
1707: \bibitem[Spergel et al.(2007)]{Sp07} %WMAP3
1708: Spergel D. N., et al. 2007, ApJS, 170, 377
1709:
1710: \bibitem[Srianand \& Petitjean(2000)]{SP00}
1711: Srianand R., Petitjean P. 2000, A\&A, 357, 414
1712:
1713: \bibitem[Sutherland \& Dopita(1993)]{SD93} % CIE
1714: Sutherland R. S., Dopita M. A. 1993, ApJS, 88, 253
1715:
1716: \bibitem[Thom \& Chen(2008a)]{TC08a} % Statistical props
1717: Thom C, Chen H.-W. 2008a, ApJ, submitted (astro-ph/0801.2380)
1718:
1719: \bibitem[Thom \& Chen(2008b)]{TC08b} % Physical Conditions
1720: Thom C, Chen H.-W. 2008b, ApJS, submitted (astro-ph/0801.2381)
1721:
1722: \bibitem[Tripp et al.(1996)Tripp, Lu, Savage]{Tr96}
1723: % Assoc systems with CLOUDY photoionization models
1724: Tripp T. M., Lu L., Savage B. D. 1996, ApJS, 102, 239
1725:
1726: \bibitem[Tripp, Savage, \& Jenkins(2000)]{Tr00} % dn/dz for low-z O VI
1727: Tripp T. M., Savage B. D., Jenkins E. B. 2000, ApJ, 534, L1
1728:
1729: \bibitem[Tripp et al.(2008)]{Tr08}
1730: % Full survey of low-z O VI systems
1731: Tripp T. M., Sembach K. R., Bowen D. V., Savage B. D., Jenkins
1732: E. B., Lehner N., Richter P. 2008, ApJS, in press (astro-ph/0706.1214)
1733:
1734: \bibitem[Trump et al.(2006)]{Tr06} % BAL Hall P. B., Reichard T. A.,
1735: Trump J. R. et al. 2006, ApJS, 165, 1
1736:
1737: \bibitem[Turnshek(1984)]{Tu84}% BAL systems
1738: Turnshek D. A. 1984, ApJ, 280, 51
1739:
1740: \bibitem[Turnshek(1988)]{Tu88}% BAL systems
1741: Turnshek D. A. 1988, in QSO Absorption Lines: Probing the Universe,
1742: ed. J. C. Blades, C. Norman, D. A. Turnshek (Cambridge: Cambridge
1743: Univ. Press), 17
1744:
1745: \bibitem[Tytler(1987)]{Ty87} % Prox effect
1746: Tytler D. 1987, ApJ, 321, 69
1747:
1748: \bibitem[Tytler \& Fan(1992)]{TF92}
1749: Tytler D., Fan X.-M. 1992, ApJS, 79, 1
1750:
1751: \bibitem[Vanden Berk et al.(2008)]{Va08} %Mg II effect
1752: Vanden Berk D., et al. 2008, ApJ, 679, 239
1753:
1754: \bibitem[Vestergaard(2003)]{Ve03}
1755: Vestergaard M. 2003, ApJ, 599, 116
1756:
1757: \bibitem[Wampler, Bergeron, \& Petitjean(1993)]{Wa93}
1758: Wampler E. J., Bergeron J., Petitjean P. 1993, A\&A, 273, 15
1759:
1760: \bibitem[Weymann et al.(1979)]{We79}
1761: % Definintion of 5000 km\s cutoff for assoc systems
1762: Weymann R. J., Williams R. E., Peterson B. M., Turnshek
1763: D. A. 1979, ApJ, 234, 33
1764:
1765: \bibitem[Weymann et al.(1991)]{We91} % BALs
1766: Weymann R. J., Morris S. L., Foltz C. B., Hewett P. C. 1991,
1767: ApJ, 373, 23
1768:
1769: \bibitem[Wild et al.(2008)]{Wi08} %Kauffmann G., White S. M.,
1770: Wild V. et al. 2008, MNRAS, submitted (astro-ph/0802.4100)
1771:
1772: \bibitem[Wright(2006)]{Wr06} % Cosmology Calculator
1773: Wright E. L. 2006, PASP, 118, 1711
1774:
1775: \bibitem[Young, Sargent, \& Boksenberg(1982)]{Yo82}
1776: Young P., Sargent W. L. W., Boksenberg A. 1982, ApJS, 48, 455
1777:
1778: \bibitem[Yuan et al.(2002)]{Yu02} %Mini BAL
1779: Yuan Q., Green R. F., Brotherton M., Tripp T. M., Kaiser M. E.,
1780: Kriss G. A. 2002, ApJ, 575, 687
1781:
1782: \bibitem[Zheng \& Davidsen(1995)]{ZD95}% QSO sphere of influence
1783: Zheng W., Davidsen A. 1995, ApJ, 440, L53
1784:
1785: \end{thebibliography}
1786:
1787: % TABLES
1788: \clearpage
1789: \setcounter{table}{0}
1790: \input{tab1.tex}
1791: \input{tab2.tex}
1792: \clearpage
1793:
1794: % FIGURES
1795: \setcounter{figure}{0}
1796: \begin{figure*}
1797: \includegraphics[width=185mm]{figures/f1.0.eps}
1798: \caption{Absorption-line spectra of all proximate \os\ systems in the
1799: UVES Large Programme.
1800: Strong absorbers are shown first, followed by the weak absorbers.
1801: Our Voigt profile fits are shown in red.
1802: In many of the \cf\ and \nf\ panels, the data
1803: are shown over a limited range in the y-axis, for clarity.
1804: Green vertical lines indicate the velocity range of the AOD integration.
1805: The label `PC' indicates regions where partial coverage of the
1806: continuum source affects the line profiles.
1807: At the top of each column we annotate our classification and the
1808: proximity of the absorber to the quasar.}
1809: \end{figure*}
1810: \addtocounter{figure}{-1}
1811: \begin{figure*}
1812: \includegraphics[width=195mm]{figures/f1.1.eps}
1813: \caption{(cont.) }
1814: \end{figure*}
1815: \addtocounter{figure}{-1}
1816: \begin{figure*}
1817: \includegraphics[width=195mm]{figures/f1.2.eps}
1818: \caption{(cont.) }
1819: \end{figure*}
1820:
1821: \addtocounter{figure}{-1}
1822: \begin{figure*}
1823: \includegraphics[width=195mm]{figures/f1.3.eps}
1824: \caption{(cont.) }
1825: \end{figure*}
1826: \addtocounter{figure}{-1}
1827: \begin{figure*}
1828: \includegraphics[width=195mm]{figures/f1.4.eps}
1829: \caption{(cont.) }
1830: \end{figure*}
1831: \addtocounter{figure}{-1}
1832: \begin{figure*}
1833: \includegraphics[width=195mm]{figures/f1.5.eps}
1834: \caption{(cont.) }
1835: \end{figure*}
1836:
1837: \addtocounter{figure}{-1}
1838: \begin{figure*}
1839: \includegraphics[width=195mm]{figures/f1.6.eps}
1840: \caption{(cont.) }
1841: \end{figure*}
1842:
1843: \addtocounter{figure}{-1}
1844: \begin{figure*}
1845: \includegraphics[width=195mm]{figures/f1.7.eps}
1846: \caption{(cont.) }
1847: \end{figure*}
1848: \addtocounter{figure}{-1}
1849: \begin{figure*}
1850: \includegraphics[width=195mm]{figures/f1.8.eps}
1851: \caption{(cont.) }
1852: \end{figure*}
1853:
1854: \end{document}
1855: