1: \documentclass[useAMS,usenatbib]{mn2e}
2: \usepackage{times}
3: \usepackage{graphicx, amssymb}
4:
5: %%%%% AUTHORS - PLACE YOUR OWN MACROS HERE %%%%%
6: \newcommand{\km}{${\rm km\,s}^{-1}$}
7: \newcommand\nodata{ ~$\cdots$~ }%
8: \newcommand{\hst}{{\em HST}}
9: \newcommand{\fuse}{{\em FUSE}}
10:
11: \newcommand{\hi}{H$\;${\small\rm I}\relax}
12: \newcommand{\hii}{H$\;${\small\rm II}\relax}
13: \newcommand{\heii}{He$\;${\small\rm II}\relax}
14: \newcommand{\di}{D$\;${\small\rm I}\relax}
15: \newcommand{\nni}{N$\;${\small\rm I}\relax}
16: \newcommand{\nai}{Na$\;${\small\rm I}\relax}
17: \newcommand{\ari}{Ar$\;${\small\rm I}\relax}
18: \newcommand{\cli}{Cl$\;${\small\rm I}\relax}
19: \newcommand{\ci}{C$\;${\small\rm I}\relax}
20: \newcommand{\cii}{C$\;${\small\rm II}\relax}
21: \newcommand{\ciii}{C$\;${\small\rm III}\relax}
22: \newcommand{\civ}{C$\;${\small\rm IV}\relax}
23: \newcommand{\nii}{N$\;${\small\rm II}\relax}
24: \newcommand{\niii}{Ni$\;${\small\rm II}\relax}
25: \newcommand{\nv}{N$\;${\small\rm V}\relax}
26: \newcommand{\oi}{O$\;${\small\rm I}\relax}
27: \newcommand{\oii}{O$\;${\small\rm II}\relax}
28: \newcommand{\oiii}{O$\;${\small\rm III}\relax}
29: \newcommand{\oiv}{O$\;${\small\rm IV}\relax}
30: \newcommand{\ov}{O$\;${\small\rm V}\relax}
31: \newcommand{\ovi}{O$\;${\small\rm VI}\relax}
32: \newcommand{\ovii}{O$\;${\small\rm VII}\relax}
33: \newcommand{\oviii}{O$\;${\small\rm VIII}\relax}
34: \newcommand{\pii}{P$\;${\small\rm II}\relax}
35: \newcommand{\si}{S$\;${\small\rm I}\relax}
36: \newcommand{\sii}{S$\;${\small\rm II}\relax}
37: \newcommand{\siii}{Si$\;${\small\rm II}\relax}
38: \newcommand{\siiii}{Si$\;${\small\rm III}\relax}
39: \newcommand{\siiv}{Si$\;${\small\rm IV}\relax}
40: \newcommand{\siv}{S$\;${\small\rm IV}\relax}
41: \newcommand{\mgii}{Mg$\;${\small\rm II}\relax}
42: \newcommand{\svi}{S$\;${\small\rm VI}\relax}
43: \newcommand{\feii}{Fe$\;${\small\rm II}\relax}
44: \newcommand{\Niii}{Ni$\;${\small\rm II}\relax}
45: \newcommand{\alii}{Al$\;${\small\rm II}\relax}
46: \newcommand{\aliii}{Al$\;${\small\rm III}\relax}
47: \newcommand{\feiii}{Fe$\;${\small\rm III}\relax}
48: \newcommand{\piv}{P$\;${\small\rm IV}\relax}
49: \newcommand{\pv}{P$\;${\small\rm V}\relax}
50:
51:
52: %%%%%%%%%%%%%%%%%%%%%%%%%%
53: %%Beginning of the paper%% Lehner et al.:
54: %%%%%%%%%%%%%%%%%%%%%%%%%%
55:
56: \title[Multiphase gas in a $z\approx 2.4$ DLA]{Probing feedback in protogalaxies: Multiphase gas in a DLA at $z\approx 2.4$}
57:
58: \author[N. Lehner et al.]{N.~Lehner,$^1$ J.C.~Howk,$^1$ J.X.~Prochaska,$^2$ and A.M.~Wolfe$^3$
59: \\
60: $^1$Department of Physics, University of Notre Dame,
61: 225 Nieuwland Science Hall, Notre Dame, IN, USA\\
62: $^2$UCO/Lick Observatory, University of California, Santa Cruz, CA, USA \\
63: $^3$ Department of Physics and Center for Astrophysics and Space Sciences,
64: University of California, La Jolla, CA, USA
65: }
66: \date{Accepted XXX.
67: Received XXX.}
68:
69: \pagerange{\pageref{firstpage}--\pageref{lastpage}}
70: \pubyear{2006}
71:
72: \begin{document}
73:
74: \maketitle
75:
76: \label{firstpage}
77:
78: \begin{abstract}
79: We investigate the physical processes occuring in the multiphase gas of a damped Ly$\alpha$ system (DLA).
80: We base our analysis on a high quality Keck HIRES spectrum of the QSO J1211+0422 in which a DLA
81: is detected at $z\simeq 2.377$. There is little contamination of the high-ion (\ovi, \nv, \civ, \siiv)
82: absorption, allowing us to explore the properties of the highly ionized gas and its connection
83: to other gas-phases. The metallicity ($[{\rm Z/H}] = -1.41 \pm 0.08$), \hi\ column density
84: ($\log N($\hi$)= 20.80 \pm 0.10$), full-width velocity ($\Delta v_{\rm neut} \simeq 70$ \km) and relative
85: abundances ($[{\rm Si/Fe}] = +0.23 \pm 0.05$ and $[{\rm N/Si}] = -0.88 \pm 0.07$) of this DLA are not unusual.
86: However, we derive the lowest \cii*\ cooling rate in a DLA, $l_c < 10^{-27.8}$ erg\,s$^{-1}$ per H atom
87: (3$\sigma$). Using this stringent limit, we show that the neutral gas (confined at $|v| < +39$ \km) must be warm and
88: the star formation rate is $<7.1\times 10^{-3}$ M$_\odot$\,yr$^{-1}$\,kpc$^{-2}$.
89: Surprisingly, the gas shows strong, complex absorption profiles from highly ionized gas
90: whose kinematics appear connected to each other and the low ions.
91: The total amount of highly and weakly ionized gas is very large with
92: $N_{\rm tot}($\hii$)/N_{\rm tot}($\hi$) \ga 1.5$. At $|v| \ga +39$ \km,
93: the gas is fully and highly ionized (H$^+/$H\,$\sim 1$,
94: $N($\civ$)\gg N($\cii$)$, $N($\siiv$)\gg N($\siii$)$).
95: Based on ionization models, \ovi\ and \nv\ are generally difficult
96: to produce by hard photons, while \siiv\ and \civ\ can be photoionized to a large extent.
97: There is, however, no evidence of \ovi-bearing gas at $T\sim 10^6$~K
98: associated with this DLA. In contrast, there is some evidence for narrow \ovi, \nv, and \civ\
99: components (unexplained by photoionization), implying too low temperatures ($T < 10^5$~K)
100: for simple collisional ionization models to produce their observed
101: column densities. Stellar feedback is a possible source for producing the high ions, but we cannot rule out
102: accretion of non-pristine material onto the protogalaxy.
103: \end{abstract}
104:
105: \begin{keywords}
106: cosmology: observations -- quasars: absorption lines -- galaxies: high-redshift --
107: galaxies: haloes -- galaxies: kinematics and dynamics
108: \end{keywords}
109:
110: \section{Introduction}
111: Damped Ly$\alpha$ absorbers (DLAs) are defined as absorbers with
112: $\log N($\hi$) \ge 20.3$ \citep[][and references therein]{wolfe05}.
113: As such these absorbers are often treated as dominantly neutral entities.
114: This characteristic separates the DLAs from the other intervening
115: absorbers seen in the QSO sightlines, the so-called Ly$\alpha$ forest ($\log N($\hi$) \le
116: 17$) and Lyman limit systems ($17 < \log N($\hi$) < 20.3$ ) where the neutral
117: gas is a minor or non-existent phase. The presence of neutral, cold, and molecular gas is crucial
118: to link the DLAs to star-forming galaxies \citep{wolfe03,wolfe05,howk05,ledoux03,noterdaeme08},
119: although the exact nature of the DLAs and their relation to present-day galaxies is still under
120: debate and study \citep{wolfe05,wolfe06,rauch08,meiksin08}.
121:
122: Surveys of DLAs have also shown systematic absorption from weakly
123: (e.g., \aliii) and highly (\siiv, \civ, \nv, \ovi) ionized species in
124: DLAs \citep{lu96,wolfe00a,fox07a,fox07}, demonstrating a complex interstellar
125: structure as observed in present-day galaxies.
126: The detection of ionized gas, and more specifically highly ionized gas,
127: is important because while high ions can be produced by hard photoionization,
128: their ionization potentials are also high enough to be directly associated with galactic feedback
129: following supernovae and stellar winds from massive stars \citep[e.g.,][]{kawata07}.
130:
131: Stellar feedback is important since it may produce galactic outflows that are the dominant
132: mechanisms for enriching the intergalactic medium (IGM) with metals, as other
133: dynamical processes, including ram-pressure stripping or tidal interactions between
134: galaxies, are rather inefficient at polluting the IGM on large scales \citep[e.g.,][]{aguirre01}.
135: Energetic feedback from stellar winds and supernovae not only pollutes the IGM but is
136: also believed to serve as a regulator of star formation, even cutting it off altogether
137: by super heating or removing (almost) all the gas from a galaxy. Hence stellar
138: feedback is an essential ingredient for cosmological models of the formation
139: and evolution of galaxies \citep[e.g.,][]{mashchenko08,dalla08}.
140: For example, \citet{white91} showed that without
141: feedback, star formation is far too efficient compared to the observations. However,
142: feedback is extremely difficult to model because it requires enough resolution
143: to resolve individual supernova and the ability to model the complex hydrodynamic
144: processes in a highly multiphase interstellar gas \citep[e.g.][]{dalla08}.
145: Characterizing feedback signatures via observations of the low and high $z$ Universe
146: is therefore critical.
147:
148: At all redshifts, evidence for outflows resulting from feedback processes
149: has been discovered. High-$z$ Lyman break galaxies (LBGs) show signs of
150: outflows \citep{pettini01,shapley03}, analogous to the complex,
151: multiphase outflows seen in low- and intermediate-$z$ starburst galaxies
152: \citep{heckman00,martin06,tremonti07,weiner08}. At low redshift,
153: winds have been found in even some modestly star forming galaxies such as
154: the Large Magellanic Cloud \citep{lehner07} and the Milky Way
155: \citep{zech07,yao07,bland03}.
156:
157: At $z\ga 2$, two
158: important DLA surveys undertaken by \citet{fox07a,fox07} have revealed the statistical
159: properties of the high ions associated with DLAs. They presented circumstantial
160: evidence that \civ\ and \ovi\ absorption at high velocities relative to the
161: DLAs is associated with protogalactic outflows. However, there is not yet a detailed analysis
162: of the high ions (\ovi, \nv, \civ, and \siiv) in DLAs in order to test their origin(s)
163: \citep[although for detailed analysis of a DLA with \civ\ and \siiv\ absorption, see][]{prochaska02}.
164: Part of the difficulty is that, while \civ\
165: and \siiv\ have generally little contamination, \nv\ absorption is weak
166: and \ovi\ is often mixed up in the Ly$\alpha$ forest. Among the
167: 9 QSOs with intervening DLAs presented by \citet{fox07}, there is not a single intervening DLA
168: where these four ions and both transitions of each doublet are simultaneously detected,
169: complicating the physical interpretation of their profiles. In particular,
170: it is not clear if the apparent differences in breadth between the \ovi\ and \civ\ component-profiles
171: are because \ovi\ probes truly hot gas or the
172: \ovi\ profiles are significantly blended with intervening Ly$\alpha$ absorbers.
173:
174: In this work, we analyze a Keck/HIRES spectrum of the QSO J1211+0422
175: (SDSS\,J121117.59+042222.3, $z_{\rm em} = 2.5412$) in order to study
176: the DLA at $z\simeq 2.377$ observed along this sightline.
177: This DLA, not surprisingly, shows high-ion absorption. Remarkably,
178: however, the profiles of the high ions, as well as other tracers of neutral and weakly ionized
179: gas, have little contamination from other intervening absorbers,
180: allowing us to pursue detailed modeling of their profiles.
181: Notably, both lines of the \nv\ doublet, while weak, are well detected.
182: The \ovi\ doublet is partially tainted by Ly$\alpha$ forest, but there is enough
183: information in each transition to interpret the \ovi.
184: With our detailed modeling, the main issues we address
185: in this work are the connection of the highly ionized with
186: the neutral and weakly ionized gas, the origin(s) of the highly ionized
187: gas in DLAs at $z\sim 2.4$, and a search for possible feedback signatures.
188:
189: The outline of this paper is as follows: in \S\ref{sec-obs} we describe
190: the observation, give an overview of the velocity structure in the DLA,
191: present the analysis techniques and results (kinematics, column densities,
192: and $b$-values of the high ions). In
193: \S\ref{sec-prop} we present and discuss the results for the
194: low-velocity gas ($|v| < +39$ \km), while in \S\ref{sec-ion}
195: the higher-velocity ($|v|\ge +39$ \km), fully ionized gas is presented.
196: In \S\ref{sec-disc} we discuss some of the properties of this DLA
197: in the context of other studies. Our main results are summarized in \S\ref{sec-sum}.
198:
199:
200: \section{Observations and Analysis}\label{sec-obs}
201: \subsection{Data reduction}
202:
203: The data presented here were collected using the upgraded High
204: Resolution Echelle Spectrometer (HIRES; Vogt et al. 1994) on the Keck
205: I telescope on 3 May 2005. The total exposure time was 2.5 hours.
206: The HIRES upgrade that took place in 2004 replaced the previous single
207: CCD with a new three CCD mosaic, including two CCDs with enhanced blue
208: sensitivity compared with the earlier instrument. The
209: $2048\times4096$ MIT/Lincoln Labs CCDs (15 \micron\ pixels) were
210: binned by 2 pixels ($0\farcs24$) in the spatial direction and have 1.3
211: \km\ pixels in the dispersion direction. All data were collected
212: using the UV cross disperser and the C1 decker, which has a
213: $0\farcs86$ slit width, giving a spectral resolution of $R\approx
214: 45,000$ or $\approx$6.6 km/s for a source filling the slit. The
215: seeing was approximately $0\farcs8$.
216:
217: The reduction and extraction of the raw data used the HIRedux package
218: (v2.2) of J.X. Prochaska, which is distributed with the XIDL
219: routines.\footnote{Available through:
220: http://www.ucolick.org/$\sim$xavier/IDL/.} We refer the reader to
221: the documentation for details. In short, the two-dimensional echelle
222: images were bias-subtracted, flat-fielded, and wavelength-calibrated
223: using the HIRES ThAr and quartz (flat field) lamps. One-dimensional
224: spectra are extracted using an optimal extraction routine, and
225: individual exposures and orders were co-added with an inverse variance
226: weighting. A continuum was fit to each order before coaddition to
227: remove the effects of the blaze. The signal-to-noise per resolution
228: element is between 35 and 45. For the figures showing the profiles
229: and apparent column densities, the spectra were binned by 3 pixels
230: (however, the profile fitting and apparent optical depth measurements
231: described below were realized with the fully sampled spectra).
232:
233:
234: \subsection{Kinematics overview}
235: The normalized profiles of \civ, \siiv, \nv, \ovi\ are shown in Fig.~\ref{spec1}
236: against the rest-frame velocity at $z = 2.37656$ (determined from \siii\ $\lambda$1808
237: and \oi\ $\lambda$$\lambda$1039,1302), while in Fig.~\ref{spec2},
238: we show the normalized profiles of the neutral and weakly ionized species.
239: The profiles show complex kinematics that we separate into four main absorption
240: components according to their velocity spread: the negative high-velocity
241: component (NHVC) with $-120 \la v \la -40$ \km, the low velocity component (LVC)
242: with $-40\la v \la +35$ \km, the $+39$ \km\ component, and the
243: positive high-velocity component (PHVC) with $+60 \la v \la +140$ \km.
244:
245: The LVC is associated with the main DLA component, which is in part neutral (strong saturated
246: \oi), but has also weakly and highly ionized species present. The NHVC is
247: more ionized than the LVC since no \oi\ or \nni\ absorption is detected and
248: is also highly ionized since \civ\ absorption is saturated and the
249: strong transition of \cii\ $\lambda$1334 is not. For similar reasons,
250: the $+39$ \km\ component also appears nearly fully and highly ionized.
251: Finally, the PHVC is very highly ionized
252: with only detection of \civ\ and \ovi\ (and possibly \nv). In this work, we
253: would like to understand the relationship (if any) of the observed highly ionized
254: species with the other neutral and low ionized species (in particular, if relatively
255: simple ionization models can reproduce the observed properties of the ionized gas), and
256: discern the origin(s) of these different components and the relationship with each other (if
257: any).
258: This requires that we first model the high-ion absorption, which we detail in the next section.
259:
260: \begin{figure}
261: \includegraphics[width = 8.truecm]{f1.eps}
262: \caption{Normalized profiles (histograms) of the highly ionized species observed
263: at $z = 2.37656$ with their profile fits (smooth red curves). The vertical dotted
264: lines show the centroid-velocities obtained from the profile fitting for each species.
265: Each doublet was fitted simultaneously, but each species was fitted independently.
266: Blue portions indicate some blending with unrelated lines.
267: \label{spec1}}
268: \end{figure}
269:
270: \subsection{Kinematics and column densities of the high ions}\label{sec-fit}
271: To determine the basic properties of the high ions, our approach is twofold:
272: (1) a detailed modeling of the absorption assuming the gas can be modeled by Voigt
273: profiles,
274: (2) a simpler approach using the apparent optical depth method.\footnote{To estimate the
275: column densities in this work,
276: we use the atomic parameters compiled by Morton (2003), except \Niii\ $\lambda$1370 that
277: is from \citet{jenkins06}. }
278:
279: The absorption profiles of the high ions have several components, which
280: we model using the Voigt component fitting software of
281: Fitzpatrick \& Spitzer (1997) where the instrumental line spread function was
282: modeled with a Gaussian with a full-width at half maximum of 6.6 \km.
283: Each doublet was fitted simultaneously,
284: but each ionic species was fitted independently (except for two strong \civ\ components
285: where we fix their velocities from \siiv, see below). We started by fitting the \siiv\
286: doublet because it is strong enough to see most of the components but not
287: so strong that saturation is a problem. Our
288: strategy was first to determine the obvious components
289: ($\sim -100,-85,-65, -15,+20,+40$ \km)
290: and then add additional components, keeping them if the $\chi^2$ of the fit
291: improved significantly. This resulted in the first 7 components summarized
292: in Table~\ref{t1}. We note that the fit has some discrepancies in the core
293: of the NHVC components in the \siiv\ $\lambda$1393 profile but
294: not observed in \siiv\ $\lambda$1402, which we attribute to possible contamination
295: from another intervening absorber (we therefore put more weight on \siiv\
296: $\lambda$1402 profile for the fit, but see below). There is also some
297: discrepancy between the fit and the data near $-50$ and $-10$ \km, but adding
298: more components did not improve the fit to the data. Yet since the main component
299: of the DLA is near 0 \km, this could suggest the need to add an additional component
300: near this velocity;
301: doing so the fit produced a component at 3.6 \km, but with very uncertain
302: $b$-value and column density ($12.6 \pm 0.4$ dex) (note that adding
303: this component did not change significantly the parameters for the other components).
304: The reduced-$\chi^2$ also only changed marginally from 1.08 to 1.04. Thus, we adopt
305: the result
306: in Table~\ref{t1} as our best model. While the results of this fit may not be unique,
307: they are robust since if we modify the input $b,N$ parameters, the solution of the
308: fit converges to the one summarized in Table~\ref{t1}. However, it is important
309: to bear in mind that the results from fitting complex profiles assume no hidden
310: components (e.g., narrow or broad components) and the errors produced from the fitting code
311: are only valid for the adopted number of components. That is a reason it is important to
312: also investigate the profiles with another method (see below).
313:
314: The profiles of the \civ\ doublet are more complicated
315: because the absorption is very strong in both lines. In particular, the flux from $-30$ to
316: $0$ \km\ for both lines reaches zero, and therefore we
317: force the centroid-velocities found from \siiv\ in this velocity range.
318: The column densities and $b$-values in this velocity range
319: remain, however, largely unknown. Except for those
320: two velocities, all the other parameters ($v,b,N$) were allowed to vary.
321: The \civ\ profiles reveal an additional component not observed in \siiv, the so-called PHVC at
322: about $+100$ \km. We fitted the PHVC with one component because while
323: the over sampled profiles present some asymmetry, there is no real evidence for
324: more than one component (see below). The overall fit to \civ\ is not as good as the model
325: of \siiv\ because the \civ\ profiles are stronger. The fit reveals,
326: however, a very close correspondence in the kinematics between \civ\ and \siiv\
327: at least for the NHVC and the component at about 20 and 40 \km.
328:
329: The \nv\ doublet is much weaker but is clearly observed in both lines between
330: $\sim$$-110$ and $\sim$$+20$ \km. Although there is little information
331: on the $b$-values, the component-centroids are coincident with those
332: of \siiv\ and \civ. The PHVC is partially contaminated in \nv\ $\lambda$1238
333: and fully contaminated in \nv\ $\lambda$1242. The column density reported in
334: Table~\ref{t1} for the PHVC is estimated from the direct integration of the profiles
335: (using the apparent optical depth method, see below) where the narrow and weak absorption at
336: $+80$ \km\ was removed.
337:
338:
339: Finally, the \ovi\ $\lambda$$\lambda$1031, 1037 profiles are certainly the most
340: complicated to model because they are so strong and because both lines are
341: partially corrupted by other intervening absorbers.
342: In regions where the profiles are contaminated (blue part of the spectra
343: in Fig.~\ref{spec1}), we use only one transition of the doublet
344: and we assume that the used transition is not contaminated by intervening
345: Ly$\alpha$ forest absorbers. The velocity-centroids match quite well the velocities of the other high ions
346: for $v<+50$ \km. At $v>0$ \km, \ovi\ $\lambda$1031 is blended with a strong absorber
347: and it is not clear if \ovi\ $\lambda$1037 could be partially blended as well
348: since the velocity structure is different than that of \civ. We report in Table~\ref{t1}
349: a two-component fit to the PHVC, but owing to the uncertainty of possible blending, we
350: will adopt the result from the apparent optical depth method (see below). However,
351: it is worth noting that the $+109$ \km\ component is too narrow to be a Lyman intervening feature
352: \citep[see, e.g.,][]{kirkman97}.
353:
354:
355:
356:
357: \begin{table}
358: \begin{minipage}{8 truecm}
359: \caption{\large Summary of profile fitting measurements for the highly ionized species
360: \label{t1}}
361: \begin{tabular}{lcccc}
362: \hline
363: Ion & $v$ & $b$ & $\log N$ & Note \\
364: & (\km) & (\km) & (cm$^{-2}$) &
365: \\
366: \hline
367: \siiv\ & $-100.3 \pm 0.4 $ & $ 6.1 \pm 0.7$ & $13.00 \pm 0.03$ & \\
368: & $ -85.1 \pm 0.6 $ & $ 3.1 \pm 2.7$ & $12.55 \pm 0.07$ & \\
369: & $ -63.6 \pm 0.3 $ & $ 14.8 \pm 0.4$ & $13.56 \pm 0.01$ & \\
370: & $ -21.2 \pm 0.5 $ & $ 2.8 \pm 1.2$ & $13.23 \pm 0.15$ & \\
371: & $ -10.6 \pm 0.7 $ & $ 14.1 \pm 0.6$ & $13.64 \pm 0.02$ & \\
372: & $ +22.1 \pm 0.7 $ & $ 15.5 \pm 1.1$ & $13.68 \pm 0.03$ & \\
373: & $ +39.9 \pm 0.6 $ & $ 5.1 \pm 1.8$ & $12.77 \pm 0.13$ & \\
374: & $ +98.8 $ & \nodata & $<11.85 $ & $a$ \\
375: \hline
376: \civ\ & $-102.3 \pm 0.5 $ & $ 8.2 \pm 0.4$ & $(>) 14.25 \pm 0.04$ & $b$ \\
377: & $ -84.8 \pm 0.5 $ & $ 4.9 \pm 1.7$ & $(>) 14.16 \pm 0.17$ & $b$ \\
378: & $ -64.7 \pm 0.6 $ & $ 12.5 \pm 0.6$ & $(>) 14.13 \pm 0.02$ & $b$ \\
379: & $ -21.2 $ & $ 2.6 :$ & $(>) 17.27 :$ & $b,c$ \\
380: & $ -10.6 $ & $ 7.7 \pm 3.8: $ & $(>) 14.86 \pm 0.53:$ & $b,c$ \\
381: & $ +16.1 \pm 0.7 $ & $ 11.7 \pm 1.9$ & $14.16 \pm 0.04$ & \\
382: & $ +36.8 \pm 1.4 $ & $ 8.6 \pm 1.2$ & $13.50 \pm 0.11$ & \\
383: & $ +99.1 \pm 0.2 $ & $ 12.9 \pm 0.5$ & $13.68 \pm 0.01$ & \\
384: \hline
385: \nv\ & $-101.5 \pm 1.7 $ & $ 13.0 \pm 3.4$ & $12.94 \pm 0.09$ & \\
386: & $ -82.5 \pm 1.1 $ & \nodata & $12.48 \pm 0.41$ & \\
387: & $ -65.3 \pm 6.6 $ & \nodata & $12.85 \pm 0.21$ & \\
388: & $ -22.0 \pm 0.8 $ & \nodata & $12.42 \pm 0.30$ & \\
389: & $ -11.6 \pm 1.1 $ & $ 15.9 \pm 1.9$ & $13.21 \pm 0.05$ & \\
390: & $ +22.1 $ & \nodata & $<12.55 $ & $d$ \\
391: & $ +39.9 $ & \nodata & $<12.21 $ & $d$ \\
392: & $ +98.5 $ & \nodata & $(<)12.80: $ & $e$ \\
393: \hline
394: \ovi\ & $-99.2 \pm 1.4 $ & $ 9.5 \pm 1.1$ & $14.08 \pm 0.08$ & \\
395: & $-81.5 \pm 0.9 $ & $ 4.6 : $ & $(>) 15.28 : $ & $b$ \\
396: & $-58.4 \pm 0.6 $ & $ 5.4 : $ & $(>) 15.67 : $ & $b$\\
397: & $-25.4 \pm24.5 $ & $ 19.2 : $ & $(>)14.13 : $ & $b$ \\
398: & $ -8.8 \pm 0.8 $ & $ 6.8 : $ & $(>)16.22 : $ & $b$ \\
399: & $+16.4 \pm 4.7 $ & $ 13.2 \pm 6.9$ & $14.00 \pm 0.22$ & \\
400: & $+39.4 \pm 1.0 $ & $ 4.3 : $ & $13.19 \pm 0.14$ & \\
401: & $+92.5 \pm 2.3 $ & $ 29.5 \pm 2.7$ & $\le 13.99 \pm 0.04:$ & $f$ \\
402: & $+109.3 \pm 0.4 $ & $ 6.0 \pm 1.1$ & $\le 13.74 \pm 0.05:$ & $f$ \\
403: \hline
404: \end{tabular}
405: Note: The colons after a value or a 1$\sigma$ error indicate that the measurements are likely more
406: uncertain than
407: the formal errors derived by the profile fitting.
408: $a$: The 3$\sigma$ error is estimated using \siiv\ $\lambda$1393 based on the centroid
409: and width of the \civ\ profile at this velocity.
410: $b$: The ``$(>)$'' indicates that the absorption reaches zero flux in both lines and
411: therefore those
412: are likely lower limits and in all cases remain very uncertain.
413: $c$: For these components, we fix the velocity
414: centroids to those of \siiv.
415: $d$: The 3$\sigma$ error is estimated using \nv\ $\lambda$1238 based on the centroid
416: and width of the \civ\ profile at this velocity (see \S~\ref{sec-fit} for more details).
417: $e$: The "$(<)$" symbol indicates that \nv\ $\lambda$1238 is partially contaminated;
418: \nv\ $\lambda$1242 is completely contaminated and cannot be used.
419: $f$: It is possible that \ovi\ $\lambda$1037 is partially contaminated by an intervening
420: system, which is emphasized by the ``$\le$'' sign (see \S~\ref{sec-fit} for more details).
421: \end{minipage}
422: \end{table}
423:
424:
425: \begin{figure*}
426: \includegraphics[width = 15.5truecm]{f2.eps}
427: \caption{Normalized profiles of several neutral and ionized species observed in the
428: spectra of J1211+0422 at $z = 2.37656$. The vertical dashed lines show the separation
429: between the 4 main components: negative high-velocity component ($-120 \la v \la -40$ \km),
430: main DLA component ($-40\la v \la +35$ \km), the $+39$ \km\ component,
431: and positive high-velocity component only observed in \civ\ and \ovi\ ($+60 \la v \la +120$ \km, see Fig.~\ref{spec1}).
432: The vertical dotted lines display the centroid-velocities of \siiv\ and \civ\ obtained from
433: profile fitting. Unrelated lines are marked with crosses. Note that the $y$-scale changes from panel to panel.
434: \label{spec2}}
435: \end{figure*}
436:
437:
438: The $b$-value derived from the profile fitting can be used to
439: estimate the temperatures of the high ions for each component.
440: In Table~\ref{t2}, we list the temperature of the gas derived from the observed
441: broadening of the absorption \siiv\ and/or \civ\ lines. If $b($\civ$)>b($\siiv) then $T$ can be
442: determined from $b = \sqrt{2 kT/(Am_{\rm H}) + b^2_{\rm nt}}$, where $b_{\rm nt}$ is the non-thermal
443: component and the other symbols have their usual meaning.
444: If $b($\civ$)<b($\siiv) only an upper limit on the temperature can be derived
445: because we cannot estimate the non-thermal broadening. We assume here
446: that \civ\ and \siiv\ reside in the same gas (we will discuss in \S\S\ref{sec-prop}, \ref{sec-ion}
447: that this may not be always the case). We did not use \nv\
448: or \ovi, as their $b$-values are generally more uncertain and
449: these ions possibly exist in different phases (i.e. at different temperatures) than \civ\ or \siiv.
450: Indeed the excitation potentials are 113.9 eV, 77.5 eV, 47.9 eV, and 33.5 eV
451: for \ovi, \nv, \civ, and \siiv, respectively. Therefore, \ovi\ and \nv\ are
452: a priori better diagnostics of shock-heated gas than \civ\ and \siiv, while photoionization
453: is likely more important for \civ\ and \siiv.
454: Half of the component sample may imply temperatures of a few times $10^5$ K gas
455: if thermal broadening dominates. The other half implies temperatures much cooler,
456: a few times $10^4$ K or less.
457:
458: \begin{table}
459: \begin{center}
460: \begin{minipage}{5.8 truecm}
461: \caption{Temperature$^a$ of the highly ionized \civ\ and \siiv\ gas \label{t2}}
462: \begin{tabular}{llccc}
463: \hline
464: Component & Name & $T$ ($10^4$ K) \\
465: \hline
466: $-100$ \km\ & NHVC & $3.9 \pm 0.4$ \\
467: $ -85$ \km\ & NHVC & $< 1.6$ \\
468: $ -64$ \km\ & NHVC & $< 11$ \\
469: $ -21$ \km\ & LVC & $< 1.3$ \\
470: $ -11$ \km\ & LVC & $<34 $ \\
471: $ +22$ \km\ & LVC & $< 10$ \\
472: $ +39$ \km\ & $ +39$ \km\ & $<4.4 $ \\
473: $+100$ \km\ & PHVC & $< 12$ \\
474: \hline
475: \end{tabular}
476: Note: $a$: Temperature estimated from the broadening of
477: \siiv\ and/or \civ\ ($b^2 = 0.129^2 \, T/A + b^2_{\rm nt}$,
478: where $b$ and $b_{\rm nt}$ are in \km).
479: \end{minipage}
480: \end{center}
481: \end{table}
482:
483: We also explore the high-ions profiles with a method that does not rely on knowing the velocity
484: distribution in these profiles. The apparent optical depth (AOD) method
485: (see Savage \& Sembach 1991) provides such a method.
486: In this method, the absorption profiles are converted into
487: apparent column densities per unit velocity $N_a(v) = 3.768\times
488: 10^{14} \ln[F_c/F_{\rm obs}(v)]/(f\lambda)$, where $F_c$ is the
489: continuum flux, $F_{\rm obs}(v)$ is the observed flux as a function of
490: velocity, $f$ is the oscillator strength of the absorption and
491: $\lambda$ is in \AA. This method also allows us to directly check
492: the saturation in the profiles by comparing the line of the same species
493: with different $f\lambda$, and in case the absorption
494: is not saturated, $N_a = \int_{v_-}^{v_+} N_a(v) dv$ provides
495: a direct estimate of the true column density. Another advantage of this
496: method is that it is straightforward to check for any variation in
497: the ratio of $N_a(v)$ with velocities for various species.
498:
499: In Fig.~\ref{aodp}, we show the apparent column density profiles for
500: each doublet. There is a good overall agreement between the $N_a(v)$
501: profiles of \siiv\ $\lambda$1393 and $\lambda$1402, except in
502: the cores of the various components, where \siiv\ $\lambda$1393 appears
503: stronger; but this contamination is small since
504: since $N_a = \int_{-120}^{-40} N_a(v) dv = 10^{13.72 \pm 0.01}$
505: and $10^{13.66 \pm 0.02}$ cm$^{-2}$ and $N_a = \int_{-40}^{+60} N_a(v) dv =
506: 10^{14.06 \pm 0.03}$
507: and $10^{14.02 \pm 0.02}$ cm$^{-2}$ for \siiv\ $\lambda$1393 and $\lambda$1402,
508: respectively. This shows that, despite the
509: absorption being quite strong, the profiles of \siiv\ have essentially
510: no unresolved saturation. The \nv\ doublet is weak and fully resolved between
511: $-120$ and $+30$ \km. For $v >+40$ \km, \nv\ $\lambda$1242
512: is lost in another intervening absorber. For \civ,
513: only parts of the profiles are resolved. For \ovi, we further see
514: that we need to rely on only one line in several portions of the absorption.
515:
516: In Fig.~\ref{aodhvc}, we zoom in on the $N_a(v)$ profiles near the PHVC.
517: The $N_a(v)$ profiles of the \civ\ doublet agree very well in the wing
518: of the profiles. There is some discrepancy between $90$ and 107
519: \km, but not in a systematic manner, and within the noise level
520: both lines give a similar column density since
521: $N_a = \int_{75}^{140} N_a(v) dv = 10^{13.68 \pm 0.01}$ and $10^{13.66 \pm 0.02}$
522: cm$^{-2}$ for \civ\ $\lambda$1548 and $\lambda$1550, respectively (note
523: that the AOD result is fully consistent with the column density derived from the
524: one component fit). This figure also shows clearly
525: the difference between \ovi\ and \civ\ with an increase in the \ovi/\civ\
526: ratio at the edge of the \civ\ profiles where the \ovi\ absorption is still
527: important. It is not clear if those differences
528: at $+60\la v\la +85$ \km\ and $+105 \la v \la +120$ \km\ are real since
529: \ovi\ $\lambda$1037 may be partially blended.
530:
531: Finally, in Fig.~\ref{aodr}, we compare the column density ratios
532: of the high ions obtained from the AOD and profile fitting. We use
533: \siiv\ as the reference ion. For the AOD, we only plot part of
534: the profiles that do not reach zero flux and we use weak lines
535: of the doublet for \civ\ and \siiv, the strong line of the \nv\
536: doublet, and non-contaminated part of the \ovi\ doublet.
537: In view of the complexity of the profiles, there is
538: a good overall agreement between the two methods (when the lines
539: are not saturated), giving us confidence in the results of our fits
540: outside the regions where the profiles are completely saturated.
541: However, we emphasize that the column densities reported in Table~\ref{t1} at $-21$
542: and $-9$ \km\ for \civ\ and \ovi, respectively, are totally uncertain.
543: In the remaining of the text, we will only use the lower limits reported in
544: Table~\ref{t4a} that are based on AOD estimates.
545:
546:
547: \begin{figure}
548: \includegraphics[width = 7.8truecm]{f3.eps}
549: \caption{Apparent column density profiles of the highly ionized species at $z = 2.37656$. Portions of
550: profiles that appear blue show some contamination from another absorption feature.
551: For velocities where the doublet AOD profiles match each other within the errors,
552: the lines are resolved. Missing data occur when the flux reaches zero.
553: \label{aodp}}
554: \end{figure}
555:
556:
557: \begin{figure}
558: \includegraphics[width = 7.8truecm]{f4.eps}
559: \caption{Profile and ratios of the apparent column densities of \civ\ and \ovi\
560: for the positive high-velocity component with respect to $z = 2.37656$. At $v\la -85$ \km\ and $v\ga +105$ \km,
561: \ovi\ might be contaminated. However, at $v\ga +105$ \km, the feature
562: is too narrow to be a Lyman line.
563: \label{aodhvc}}
564: \end{figure}
565:
566:
567: \begin{figure}
568: \includegraphics[width = 7.8truecm]{f5.eps}
569: \caption{Ratio of column densities at $z = 2.37656$. Filled circles are from the profile fitting method,
570: while histograms are from the apparent optical depth method.
571: \label{aodr}}
572: \end{figure}
573:
574:
575: \section{The low-velocity multiphase gas}\label{sec-prop}
576:
577: \subsection{Overview}
578: We define the LVC as being the absorption observed
579: between about $-40$ and $+35$ \km\ where most of the neutral
580: gas is found. The \oi\ profile in Fig.~\ref{spec2} traces the
581: velocity distribution of neutral gas, noting that
582: \oi\ is one of the best tracers of neutral hydrogen
583: since its ionization potential and charge exchange reactions with hydrogen ensure that
584: \hi\ and \oi\ are strongly coupled. The LVC has
585: a complex multiphase structure of neutral gas (\oi, \nni),
586: ionized gas (\nii, \aliii, \feiii), and highly ionized gas (\ovi, \nv).
587: The LVC differentiates itself from the other components at
588: higher absolute velocities by the fact that it is only between
589: $-40$ and $+35$ \km\ that strong atomic absorption (\oi, \nni) is observed;
590: in contrast the components at higher absolute velocities are nearly fully ionized as we
591: show in \S\ref{sec-ion}. The presence of neutral gas allows us to directly
592: estimate the metallicity of the gas probed (see \S\ref{ssec-metal}).
593: In Table~\ref{t4}, we summarize our column density and abundance estimates for
594: the LVC where for the species other than \hi, the column densities were estimated
595: via the AOD method (see \S\ref{sec-fit}). The \hi\ column density is derived from
596: a simultaneous fit to the damped wings of Ly$\alpha$ (from SDSS) and Ly$\beta$
597: \citep[following][]{herbert06}, and we
598: show the fit to those lines in Fig.~\ref{fithi}.
599:
600: For the high ions, the LVC has at least 3 components derived from the fits.
601: Inspection to the \sii\ profile also reveals that the main absorption can be broken-up
602: in 3 components with velocities centered at $\sim -10, 0, +15$
603: \km. Such structure cannot be seen in \oi\ because the profile is completely saturated,
604: but the velocity structure of \nni\ is consistent with \sii.
605: \oi\ reveals another component at $+22$ \km, which is also seen in the high ions.
606: In Fig.~\ref{spec2}, we show that except for the component at 0 \km\, the velocity-
607: centroids of the high ions match quite well those of the low ions, suggesting some kinematic
608: relation between the weakly and highly ionized phases. The zero velocity component is not included in our
609: fits to the high ions, but as we discussed in \S\ref{sec-fit}, the fit for the
610: high ions may not be unique and can certainly accommodate a component at $\sim 0$ \km.
611:
612: \subsection{Absolute and relative abundances}\label{ssec-metal}
613: To derive the metallicity, it is usually best to use neutral or singly-ionized species that
614: follow the neutral hydrogen to avoid large ionization correction since only $N($\hi$)$ can be directly estimated.
615: The \oi\ absorption is so strong that only a firm lower limit can be placed on the oxygen abundance
616: of the DLA (see Table~\ref{t4}). Because of the nucleosynthetic evolution, nitrogen is often deficient with respect
617: to the other elements \citep[][and see below]{lu98,prochaska02a,henry00,henry06,petitjean08}, and this is observed in
618: this DLA since $[$N$/\alpha] = -0.88 \pm 0.07$, where the $\alpha$-element is
619: here Si (see Table~\ref{t4} and also \S\ref{ssec-ion}). Nitrogen therefore does not provide
620: a good metallicity indicator.
621:
622: In the next section, we show that the singly-ionized
623: species (excluding \nii) predominantly arise in the neutral gas, and
624: therefore these ions can be directly compared to \hi. Sulfur appears not to be
625: depleted into dust and silicon is usually very mildly depleted in DLAs, and $[{\rm Si/S}] = +0.02 \pm 0.06$
626: in this DLA suggests as well no dust depletion of Si. Using these
627: two elements, we derive the metallicity of the DLA, $[{\rm Z/H}] = -1.41 \pm 0.08$,
628: based on the $\alpha$-elements S and Si.
629:
630:
631:
632: \begin{table}
633: \begin{minipage}{8.2 truecm}
634: \caption{\large DLA (LVC) column densities and abundances \label{t4}}
635: \begin{tabular}{lcc}
636: \hline
637: Species X & $\log N($X$)$ & $[$X$/$H$]^a$ \\
638: \hline
639: \hi\ $\lambda$$\lambda$1025,1215 & $20.80 \pm 0.10$ & \nodata \\
640: \ci\ $\lambda$1560 & $< 12.70 $ & ($<-4.49$) \\
641: \cii\ $\lambda$1036 & $> 14.70 $ & $>-2.49 $ \\
642: \cii*\ $\lambda$1335 & $< 12.53 $ & ($<-4.66$) \\
643: \civ\ $\lambda$1550 & $> 14.50 $ & $(>-2.69) $ \\
644: \nni\ $\lambda$1134 & $14.30 \pm 0.05$ & $-2.28 \pm 0.11$ \\
645: \nii\ $\lambda$1083 & $\le14.37 \pm 0.10$ & $(<)-2.21 \pm 0.15$ \\
646: %\nni\ $+$\,\nii\ & $(<)14.64 \pm 0.11$ & $(<)-1.94 \pm 0.15$ \\
647: \oi\ $\lambda$1039 & $> 15.32 $ & $>-2.14 $ \\
648: \mgii\ $\lambda$1239 & $14.87\,^{+0.12}_{-0.17}$& $-1.46\,^{+0.19}_{-0.21} $ \\
649: \alii\ $\lambda$1670 & $> 13.42 $ & $>-1.80 $ \\
650: \aliii\ $\lambda$$\lambda$1854,1862 &$12.69 \pm 0.03$ & ($-2.56 \pm 0.11$) \\
651: \siii\ $\lambda$1808 & $14.91 \pm 0.04$ & $-1.40 \pm 0.11$ \\
652: \siiii\ $\lambda$1206 & $> 13.21 $ & ($>-3.10$) \\
653: \siiv\ $\lambda$1402 & $13.96 \pm 0.04$ & ($-2.30 \pm 0.10$) \\
654: \si\ $\lambda$1425 & $< 12.38 $ & ($<-3.57$) \\
655: \sii\ $\lambda$$\lambda$1250,1259 & $14.53 \pm 0.04$ & $-1.42 \pm 0.11$ \\
656: \ari\ $\lambda$1066 & $< 13.10 $ & $<-1.88$ \\
657: \feii\ $\lambda$$\lambda$1063,1144,1608 & $14.62 \pm 0.03$ & $-1.58 \pm 0.10$ \\
658: \feiii\ $\lambda$1122 & $13.97 \pm 0.03$ & ($-2.28 \pm 0.11$) \\
659: \Niii\ $\lambda$$\lambda$1370,1709,1741 & $13.36 \pm 0.08$ & $-1.64 \pm 0.13$ \\
660: \hline
661: \end{tabular}
662: Note: Errors are 1$\sigma$. Upper limits (``$<$'') are 3$\sigma$ estimated over $[-30,+30]$ \km, full
663: velocity-width of the weak absorption (e.g., \siii\ $\lambda$1808). $a$: Throughout the
664: text we use
665: the following notation ${\rm [X/H]} = \log N({\rm X}^i)/N({\rm H}^0) - \log({\rm
666: X/H})_\odot$ (solar abundances are from Asplund et al. 2006); if ${\rm X}^i$ is
667: a dominant ionization stage in neutral gas, $[$X$/$H$]$ indicates the abundance of a given species X.
668: Estimates within parentheses do not represent a species in a dominant ionization stage,
669: and therefore not the overall abundance of the gas (see text for more
670: details). ``$\le$'' indicates that \nii\ is likely contaminated.
671: \end{minipage}
672: \end{table}
673:
674:
675: \begin{figure}
676: \includegraphics[width = 6.5 truecm, angle = 90 ]{f6.eps}
677: \caption{Simultaneous fit to the damped Ly$\alpha$ (SDSS spectrum) and Ly$\beta$
678: (Keck HIRES) lines at $z = 2.37656$. The dashed lines show the adopted continua.
679: The bump in the continuum of Ly$\beta$ (bottom panel) is due to the blaze
680: function of the echelle spectrum. The solid green line show the solution
681: $\log N($\hi$) = 20.80$ with the shaded region showing the $\pm 1 \sigma$ solutions ($\pm 0.10$ dex).
682: Note that the wavelength scales are different in the top and bottom panels.
683: \label{fithi}}
684: \end{figure}
685:
686: Nucleosynthesis and dust depletion both play a role in the gas-phase abundances of the elements in
687: DLAs \citep[e.g.,][]{lu96,pettini99a,prochaska02}, but disentangling these effects is
688: extremely difficult and requires more elements than available here. However,
689: the low $[$N$/\alpha]$ ratio implies that nucleosynthesis is important
690: since N is not depleted into dust and \nni\ is not affected by ionization in this case.
691: This low $[$N$/\alpha]$ ratio is often explained as a time delay
692: between the production and injection of N into the gas compared with $\alpha$-elements.
693: The $\alpha$-elements (including O, Si, S) are mainly produced by
694: short-lived massive stars ($M>10M_\odot$) through explosions of Type II supernovae (SNe),
695: while N in low metallicity gas is believed to be principally formed in intermediate-mass stars
696: ($3 M_\odot<M< 8 M_\odot$) and injected into the interstellar gas through stellar mass loss. The same
697: low and intermediate-mass stars also produce the bulk of the Fe-group elements through
698: Type Ia SNe. Therefore the $\alpha$-element to Fe-peak element ratio should be
699: systematically enhanced in low metallicity environments. This is observed
700: in low metallicity stars in our Galaxy, where $[{\rm \alpha/Fe}] \approx +0.2$ \citep[e.g.][]{nissen07}.
701: In the present DLA, $[{\rm \alpha/Fe}] = +0.23 \pm 0.05$ and $[{\rm \alpha/Ni}] = +0.24 \pm 0.09$ (where
702: $\alpha=$\,\siii, but as noted above the use of \sii\ would not have changed
703: the results). These values are consistent with others in DLAs,
704: although on the low side. The $[{\rm \alpha/Fe}]$ enhancement coupled
705: with $[{\rm Si/S}]\approx 0$ and $[$N$/\alpha] \approx -0.9$ is consistent
706: with a Type II SN enrichment \citep{prochaska02,henry07}, i.e. the relative
707: abundances in this DLA could be explained by nucleosynthesis alone.
708:
709:
710:
711: \begin{figure}
712: \includegraphics[width = 9truecm]{f7.eps}
713: \caption{{\em Top panel}: Predicted column densities for the Cloudy photoionization model of
714: the LVC assuming a Haardt-Madau (galaxies+QSOs) spectrum. The various lines show the models for
715: each atoms or ions. Relative solar abundances are assumed. The yellow region shows a solution
716: that fits the observations within about $\pm 0.2$ dex (see also Table~\ref{t4a}). This model does, however,
717: not produce enough highly ionized species.
718: {\em Bottom panel}: Variation of the neutral fraction and the path length ($L \equiv N({\rm H})/n_{\rm H}$).
719: For the solution that fits the observations, the fraction of neutral gas is 80--88\% and the linear scale is
720: 2--6 kpc.
721: \label{cloudy-lvc}}
722: \end{figure}
723:
724:
725:
726: On the other hand, in our Galaxy, elements become less and less depleted as the gas changes
727: from cold to warm phases \citep[see, e.g.,][]{savage96}.
728: Assuming now that nucleosynthetic effects are negligible,
729: the low depletion of Fe and Ni with respect to those in other DLAs {\em with similar metallicities}
730: suggests that the neutral gas is likely to be mostly warm. As we show
731: below the Cloudy simulation and the absence of \cii*\ absorption also support that the
732: neutral gas is warm. An intrinsic solar $[{\rm \alpha/Fe}]$ at $[{\rm Z/H}] \approx -1.4$
733: may imply that the star formation occured slowly, allowing a large Fe pollution from
734: type Ia SNe \citep{matteucci01}, but this contradicts the extremely low $[{\rm N/\alpha}]$
735: ratio. More likely, the regions where star formation occurs are very localized and are not directly probed by
736: the J1211+0422 sightline. Hence the gas may be indeed warm with little dust along the line of sight.
737: Both effects from nucleosynthesis and dust depletion may be present, but it is not possible
738: to disentangle these subtle effects at the 20\% (0.1 dex) level (without mentioning possible
739: uncertainties in the atomic parameters or ionization
740: correction, see \S\ref{ssec-ion}). The recent non-LTE abundances of S and Fe of
741: Galactic halo stars with $[{\rm Z/H}] \approx -1.4$ seem to imply that $0.1 <[{\rm \alpha/Fe}] <0.2$
742: \citep{nissen07}, allowing both some Fe and Ni depletion and nucleosynthesis effects to occur
743: if the DLA nucleosynthetic history follows that of Galactic halo stars. In \S\ref{sec-cool},
744: we assume that the Fe depletion is $[{\rm Fe/Si}] = -0.2$ dex \citep[see discussion in][]{wolfe03,wolfe04}, which still
745: allows for an $\alpha$-element enhancement over Fe of $\sim$0.1 dex
746: attributable to nucleosynthesis at the 1$\sigma$ level.
747:
748:
749:
750: \begin{table}
751: \begin{minipage}{8.5 truecm}
752: \begin{center}
753: \caption{Column densities for the LVC from the Cloudy simulation
754: using solar relative abundances and $[{\rm Z/H}] = -1.4$ (Fig.~\ref{cloudy-lvc}). \label{t4a}}
755: \begin{tabular}{lccc}
756: \hline
757: Species & $\log N_{\rm obs}$ & $\log N_{\rm model}$ & $\log N_{\rm model}$ \\
758: & & $\log U = -3.1$ & $\log U = -2.9$ \\
759: \hline
760: \ci\ & $< 12.70 $ & 11.9 & 11.7 \\
761: \cii\ & $> 14.70 $ & 15.8 & 15.8 \\
762: \cii*\ & $< 12.53 $ & 12.5 & 12.4 \\
763: \civ\ & $> 14.50 $ & 12.9 & 13.4 \\
764: \nni\ & $14.30 \pm 0.05$ & 15.2 & 15.2 \\
765: \nii\ & $<14.37 $ & 14.2 & 14.2 \\
766: \nv\ & $13.28 \pm 0.05 $ & 10.6 & 11.2 \\
767: \oi\ & $> 15.32 $ & 16.1 & 16.1 \\
768: \ovi\ & $> 14.86 $ & 8.8 & 9.7 \\
769: \mgii\ & $14.87\,^{+0.12}_{-0.17}$& 14.9 & 14.9 \\
770: \alii\ & $> 13.42 $ & 13.8 & 13.8 \\
771: \aliii\ &$12.69 \pm 0.03$ & 12.8 & 12.9 \\
772: \siii\ & $14.91 \pm 0.04$ & 14.9 & 14.9 \\
773: \siiii\ & $> 13.21 $ & 13.8 & 14.1 \\
774: \siiv\ & $13.96 \pm 0.04$ & 12.8 & 13.2 \\
775: \sii\ & $14.53 \pm 0.04$ & 14.4 & 14.4 \\
776: \ari\ & $< 13.10 $ & 13.1 & 13.0 \\
777: \feii\ & $14.67 \pm 0.03$ & 14.9 & 14.9 \\
778: \feiii\ & $13.97 \pm 0.03$ & 13.8 & 13.9 \\
779: \Niii\ & $13.36 \pm 0.08$ & 13.6 & 13.6 \\
780: \hline
781: \end{tabular}
782: \end{center}
783: \end{minipage}
784: \end{table}
785:
786:
787:
788: \subsection{Ionization}\label{ssec-ion}
789: The comparison of the apparent column densities of \siiv\ and \siii\ shows that the gas traced by
790: singly ionized species is far more important between about $-15$ and $+20$ \km.
791: In contrast the ratios of doubly-ionized
792: species (\aliii\ and \feiii) to \siiv\ show only little variations between
793: $-40$ to $+40$ \km, suggesting some connection between these ions and that ionization
794: may be important. The \nv/\siiv\ ratio also steadily increases
795: from $+20$ to $-40$ \km, implying that the gas becomes more highly ionized at more
796: negative velocities. Although the \ovi\ absorption
797: is too strong to directly use for the column density ratios, the \ovi\ absorption is
798: also far stronger at $-40 \la v \la 0$ \km\ than at $0 \la v \la +30$ \km\
799: (see Figs.~\ref{spec1} and \ref{aodp}).
800:
801: We first investigate whether photoionization can explain the observed properties
802: of the absorbers. We used the photoionization code
803: Cloudy version C07.02 \citep{ferland98}
804: with the standard assumptions there has been enough time
805: for thermal and ionization balance to prevail.
806: We model the column densities of the different ions through a slab illuminated (on both
807: sides) by the Haardt \& Madau (2008, in prep.) UV background ionizing radiation field
808: from quasars {\em and}\ galaxies and the cosmic background radiation appropriate
809: for the redshift $z = 2.377$ and maintained in this case at constant pressure
810: (condition not set for the Cloudy simulations of the higher-velocity components).
811: The Cloudy simulations also assume a priori solar relative heavy element abundances
812: from \citet{asplund06}.
813: We then vary the ionization parameter, $U = n_\gamma/n_{\rm H} =$\,H ionizing photon density/total
814: hydrogen number density [neutral\,+\,ionized], to search for models that
815: are consistent with the constraints set by the column densities and $b$-values.
816:
817: We show in Fig.~\ref{cloudy-lvc} the result of such a Cloudy simulation
818: for the LVC where $[{\rm Z/H}] = -1.4$
819: (see \S\ref{ssec-metal}) and where the \hi\ column density matches
820: the measured value. We use the total column densities rather than modeling
821: each component because i) the ratios of \feiii/\siiv\ and \aliii/\siiv\ show
822: little variation over the LVC velocity range, ii)
823: the total column densities are likely less uncertain than the column
824: densities of the individual components since they do not rely on
825: a particular kinematic model, and iii) the \hi\ column density in the
826: individual components is unknown. The yellow region shows the range of possible
827: solutions ($-3.15 \le \log U \le -2.90$) that fits the column densities of
828: the neutral, singly- and doubly-ionized species within $\pm 0.1$--0.2 dex.
829: For $\log U = -3.1$ and $\log U = -2.9$, column densities for the various species
830: are reported in Table~\ref{t4a}, showing an overall agreement between the model
831: and the observations, except for \nni. The low abundance of N can be understood in terms
832: of nucleosynthetic evolution (see above): using \nni, we find $[$N$/\alpha] \simeq -0.9$
833: (\nii\ is contaminated by an unknown amount and is anyway important only in ionized gas).
834: From Table~\ref{t4a}, we also note
835: some discrepancies with \feii\ and \niii\ (although it is within $\sim 0.2$ dex),
836: which can be interpreted as these species being in part depleted into dust or as a
837: nucleosynthetic effect (see \S\ref{ssec-metal}). However, if in the ionized phase Fe is depleted into
838: dust or deficient relative to $\alpha$-elements as a result of nucleosynthesis,
839: the discrepancy between the model and observations would increase for \feiii, possibly
840: pointing to the need of another ionizing source (see below).
841:
842: The overall good agreement between the model and the observation is quite
843: remarkable in view of the complexity of the profiles and the possible
844: effects of dust and nucleosynthesis.
845: We note that although the agreement would be better for \sii\ if $\log U < -3.1$,
846: smaller ionization parameters are inadequate because too much \ari\ would be produced.
847: $[$\ari$/\alpha] < -0.48$ (where here $\alpha$ is \siii) favors the present model
848: rather than a softer stellar spectrum. Indeed, since Ar is not depleted into dust and
849: is an $\alpha$-element, the observed \ari\ deficiency is due to photoionization \citep{sofia98,prochaska02}.
850: In our own Galaxy, partially photoionized gas leads to a deficiency of
851: the neutral form of Ar \citep{sofia98,jenkins00,lehner03}, but it is generally
852: mild compared to the high redshift systems \citep{vladilo03}.
853: A harder spectrum (i.e. a UV background from QSOs only) would, however,
854: produce a more discrepant model, e.g., $N($\sii$)$ and $N($\feiii$)$ would
855: be more difficult to match simultaneously.
856:
857: This model implies that the singly-ionized species (\cii, \siii, \mgii, \feii, \Niii, and \sii,
858: although the latter may be more subject to ionization correction as illustrated in Fig.~\ref{cloudy-lvc})
859: principally trace the neutral gas. The ionization fraction (\hii/H) is
860: about 12--20\% (not taken into account the high ions, see below).
861: The ``average'' linear size, $L\equiv N({\rm H})/n_{\rm H}$, of the layer of gas in this model is
862: $\sim$2--6 kpc, which seems acceptable for a protogalaxy at $z\sim 2.4$
863: (note that for absorbers with multiple components, $L$ given by Cloudy
864: should be treated as an upper limit).
865: We noted a close kinematic correspondence between
866: the ionized and neutral gas is consistent with the picture of an external
867: UV source that ionizes the external layer of the neutral gas.
868:
869: The Cloudy photoionization simulation, however, fails to produce enough
870: \siiv, \civ, \nv\ and \ovi\ (by orders of magnitude for \civ, \nv\ and \ovi\, see Table~\ref{t4a}),
871: which implies that these ions must be produced
872: by other means, such as collisional ionization processes. Such a conclusion was
873: also drawn by \citet{fox07} in another DLA. We note if
874: $\log U \simeq -2.35$, the modeled and observed column densities of \feiii\ and \siiv\
875: would be similar. This solution would also contribute to a significant amount of \civ, but
876: would require Al to be deficient by $-0.4$ dex compared to Si and Fe. Sulfur would be
877: underproduced by about 0.2 dex.
878: Such a solution would also require a very large path length
879: ($\sim$40 kpc) and would nevertheless fail to produce enough \ovi\ and \nv\ columns.
880: Therefore this solution is less compelling than the $\log U \sim -3$ solution.
881:
882: Collisional equilibrium (CIE) and non-equilibrium collisional ionization (NECI) may also be important
883: sources of ionization, particularly for producing high ions \citep{sutherland93}.
884: Recently, \citet{gnat07} have produced new calculations of CIE and time-dependent
885: radiative cooling gas as a function of metallicity using updated atomic data.
886: In galactic environments, highly ionized species may also be produced via
887: shocks \citep{dopita96}, or in interfaces between cool
888: ($T\la 10^4$ K) and hot ($T\ga 10^6$ K) gas, e.g., in conductive heating models
889: \citep{boehringer87,borkowski90} or turbulent mixing models
890: \citep{slavin93}. In view of the calculations of \citet{gnat07}, metallicity
891: can be an important ingredient; most of the models above assume
892: solar metallicity gas.
893:
894: To constrain the collisional ionization model, the high-ion column density ratios
895: combined with the line-widths of the high-ion components may a priori help us
896: to discern between these various models \citep[e.g.,][]{spitzer96}. Unfortunately,
897: the results from the profile fitting are somewhat uncertain for the LVC since the
898: absorption is so strong (see \S\ref{sec-fit}). From the \siiv\ and \nv\ profiles,
899: there are some suggestions of a mixture of hot ($T \la 3 \times 10^5$ K in
900: the $-11$ components) and cool ($T\la 10^4$ K in the $-21$ \km\ component) gas
901: (see Table~\ref{t2}, \S\ref{sec-fit}, and Fig.~\ref{spec1}). As we argue above,
902: there is also room for an extra component at $\sim 0$ \km, at least in the \siiv\ and
903: \civ\ profiles, and hence the components at $-11$ and $+22$ \km\ could
904: be narrower. The $N_a(v)$ profiles of \nv\ and \ovi\ show that most of the absorption
905: is at $-20 \la v \la +10$ \km\ and in particular the absorption peaks at
906: $-20$ and $-11$ \km (see Fig.~\ref{aodp}). This contrasts from the \siiv\ absorption,
907: where the $N_a(v)$ profiles peak at $\sim$$-15$ but also at $+22$ \km. The $N_a(v)$ profile
908: of \civ\ appears to be somewhat intermediate between \siiv\ on one hand and \nv\
909: and \ovi\ on the other hand. The $N_a(v)$ profiles of the high ions
910: therefore suggest that at $0 < v \la +40$ \km, the highly ionized (in particular \siiv) gas may
911: be more subject to photoionization, but at $-40 \la v \la 0$ \km, collisional ionization becomes
912: more important. However, we emphasize that both mechanims likely occur at any $v$.
913: For the component at $+16$ \km, assuming that a large fraction
914: of \siiv\ is photoionized,\footnote{The velocity centroid of \siiv\ is $+22$ \km,
915: slightly different from the other high ions, consistent with the idea that the origins of these ions
916: may be different.} the ratios of \civ/\ovi\ and \nv/\ovi\ (taken into account
917: the intrinsic deficiency of N) in the models of \citet{gnat07} do not fit the observed
918: ratios, but models such as conductive interfaces or shock ionizations can produce the observed
919: high-ion ratios. For the
920: other components, the high-ion ratios are too poorly determined to draw any firm conclusions
921: about their origin(s), although we note that the line width of the component at $-22$ \km\ implies a very low
922: temperature for \civ\ and \siiv\ (see Table~\ref{t2}) and probably for \nv\ (see Fig.~\ref{spec1}),
923: suggesting an additional photoionization source (possibly from stars) or some NECI process
924: with efficient cooling.
925:
926: According to our profile fitting, there is also a close kinematic relationship
927: between the high ions and the low ions and atoms.
928: A simple explanation for this may involve interface layers between a hot
929: ($T>10^6$ K) and a cooler ($T\la 10^4$ K) gas. Such interfaces are believed to be
930: common in our much evolved Galaxy. However, the column densities of the high ions (in particular \ovi)
931: in model interfaces are much smaller than observed in individual interfaces \citep{savage06}
932: and modeled \citep{borkowski90,slavin93},
933: requiring (too) many interfaces (and this discrepancy is probably worse in
934: the DLA, which has a much lower metallicity than the Galaxy). Other interstellar models of high-ion production
935: (e.g., radiatively cooling clouds, bubbles blown by supernovae or stellar winds) predict
936: large high-ion column densities. However, as discussed by \citet{tripp08},
937: the physics of the interfaces may be oversimplified and not so well understood, especially
938: in the context of high redshift, low metallicity systems.
939:
940: From the above, we conclude that
941: the kinematic and ionization properties of this DLA are consistent with
942: a highly ionized halo that co-rotates with a neutral proto-galactic disk.
943: This picture is quite similar to our own Galaxy, which is surrounded by
944: a hot highly ionized and a warm ionized halo, revealed by the same high
945: and intermediate ions as those at high redshifts \citep[e.g.,][]{zsargo03,savage03,howk03}.
946: However, since several components are observed, it is also possible that we are
947: simply seeing several sheets of neutral gas separated by weakly and highly ionized gas.
948: In galaxies, a major source for heating and ionizing the interstellar medium is
949: massive stars. It is well known that these stars inject large amounts of energy in the interstellar
950: medium over large scales via their supernovae and stellar winds. This environment may cool and/or
951: produce interfaces between cool ($T\la 10^4$ K) and hot ($T>10^6$ K) gas.
952: We have shown that the UV background plays an important role in ionizing and heating the present DLA,
953: but it cannot explain by itself the origin of the high ions, especially \ovi, \nv, but also \civ,
954: and even part of \siiv. Below we argue and show that there is room
955: for stars to be present in this DLA-protogalaxy, hence stellar feedback is a possible
956: origin for the high-ion absorption.
957:
958: \subsection{Nature of the neutral gas and star formation}\label{sec-cool}
959: Above we show that the UV background plays an important role in ionizing and heating
960: the present DLA. There is, however, some leeway for the existence of other
961: (internal) sources of ionization and heating; e.g., those might help to better model
962: the column densities of \feiii\ and \siiv. We further investigate the
963: nature of the neutral gas in this DLA and the implication for the star formation
964: rate (SFR) using the limit on \cii*\ absorption and the
965: technique described by \citet{wolfe03} amd \citet{wolfe04}. In this technique,
966: we allow the heating and cooling to be balanced in a two phase medium
967: (cold and warm neutral medium, hereafter CNM and WNM). The heating
968: is mostly due to grain photoelectron emission, assuming a Fe depletion
969: $[{\rm Fe/Si}] = -0.2$.
970: The cooling in the CNM and WNM is dominated by \cii\ at 158 $\mu$m and Ly$\alpha$ emisions,
971: respectively \citep[for more details, see][]{wolfe03}.
972: The star formation per unit area can be directly related to the photoelectric
973: heating rate. The predicted \cii\ 158 $\mu$m cooling rate is compared with that derived
974: from the strength of the \cii*\ absorption
975: \citep[e.g.,][]{lehner04}:
976: \begin{equation}\label{ecool1}
977: l_c = 2.89 \times 10^{-20}\frac{N({\rm C}^{+*})}{N({\rm H}^{0})}\,
978: \,{\rm erg}\, {\rm s}^{-1}\, ({\rm H\,atom})^{-1}\,.
979: \end{equation}
980: With the column densities of \cii*\ and \hi\ presented in Table~\ref{t5},
981: we find that for LVC, $\log l_c < -27.8$ (3$\sigma$).
982: The upper limits of Wolfe et al. are 2$\sigma$, i.e.
983: for our DLA, $\log l_c < -28.0$ (2$\sigma$).\footnote{An integration of
984: \cii*\ profile over the $[-20,+20]$ \km\ velocity interval would give a 2.5$\sigma$ detection
985: ($W_\lambda = 5.4 \pm 2.2 $ m\AA, $\log N = 12.45\,^{+0.14}_{-0.20}$), which
986: would imply $l_c = -27.9 \,^{+0.2}_{-0.3}$, consistent with the 2$\sigma$ limit.
987: We favor the upper limit because the velocity range is not justified in
988: view of the other absorption profiles, but it is possible that the value of $l_c$
989: is not much smaller than the quoted 3$\sigma$ limit.} This is the lowest known limit
990: on the cooling rate so far found in a DLA \cite[see the up-to-date sample of][]{wolfe08}.
991: Extremely low \cii\ radiative cooling rates can be found either in cold neutral
992: gas where the SFR is low or in warm neutral
993: gas \citep{wolfire03,wolfe03,lehner04}. Below we show that this is the only known DLA
994: where the neutral gas is unambiguously warm and where star formation may occur.
995:
996:
997: \begin{figure}
998: \includegraphics[width = 8.5truecm]{f8.eps}
999: \caption{The panels show a sequence of equilibrium curves for gas subjected to internal heating rates generated by
1000: a grid of SFRs per unit area, $\log \Sigma_{\rm SFR}= -4.0$, $-3.5$, $-3.0$, ....0.0 (above the bottom
1001: dotted red curve to top curves)
1002: in units of M$_\odot$\,yr$^{-1}$\,kpc$^{-2}$.
1003: Extragalactic background radiation is included and the bottom curve shows the result for background radiation alone.
1004: In panels (a) and (b) we show the pressure and the cooling rate, respectively, against the density of
1005: the gas. In (b) the red dash-dotted curve is the locus of all equilibrium CNM solutions in which
1006: $P_{\rm eq} = (P_{\rm max} P_{\rm min})^{0.5}$. The green dash-dotted curve is the same for the WNM solutions.
1007: The solid horizontal line shows the 3$\sigma$ limit derived for the DLA at $z\simeq 2.377$
1008: toward J1211+0422.
1009: \label{fig-cool}}
1010: \end{figure}
1011:
1012:
1013:
1014: Fig.~\ref{fig-cool} shows the
1015: equilibrium solutions for gas subjected to internal and external sources of heating. The internal source
1016: is photoelectric heating by dust grains absorbing starlight produced locally (and an assumed
1017: cosmic-ray contribution), while the external heating is produced
1018: by the extragalactic background from galaxies and QSOs including UV and
1019: a higher energy component \citep[see][for more details]{wolfe03,wolfe04}.
1020: We assume no enhancement in $\alpha$-element abundances in estimating the
1021: dust content \citep[see][]{wolfe04}.
1022: The red dash-dottedcurve is the locus of all equilibrium CNM solutions
1023: and the green dash-dotted curve is the same for the WNM solutions.
1024: The solid horizontal line shows our 3$\sigma$ limit on $l_c$, which rejects a CNM solution.
1025: Therefore the neutral gas is warm. The WNM solution also suggests that star formation may
1026: occur, but only at a rate $\Sigma_{SFR} < 7.1\times 10^{-3}$ M$_\odot$\,yr$^{-1}$\,kpc$^{-2}$. This limit
1027: is about a factor twice larger than the local rate in the Milky Way.
1028: As we discuss in \S\ref{ssec-metal}, there might be very little dust in this DLA.
1029: We therefore also ran a dust-free model where the only heating is
1030: by X-ray photoionization and cosmic rays. As in the previous model, their heating
1031: rates are proportional to the SFRs. The CNM solution is again ruled out and star formation
1032: is still posible in the WNM solution. A population of DLAs with warm neutral gas and possible
1033: ongoing star formation exists based on our findings.
1034: Hence many of the sightlines where only upper limits on $l_c$ are derived may actually probe WNM gas.
1035:
1036: \section{The fully ionized high-velocity gas}\label{sec-ion}
1037: \subsection{Overview}\label{sec-ion-over}
1038: The strong \oi\ $\lambda$1302 transition is completely absent outside
1039: the velocity range $[-40,+35]$ \km\ (see Fig.~\ref{spec2}). Absorption
1040: from singly-ionized species with large $\lambda f$ values, \siiii,
1041: and highly ionized species is seen at higher absolute velocities.
1042: The NHVC, the $+39$ \km\ component,
1043: and the PHVC (see Table~\ref{t2}) have
1044: $[$O/Si$]\equiv \log N({\rm O^0})/N({\rm Si^+}) - \log ({\rm O/Si})_\odot <-0.8$.
1045: Since Si and O are both $\alpha$-elements (i.e., have similar nucleosynthetic
1046: history), we expect the intrinsic O/Si should be nearly solar.
1047: The strong underabundance of {\em neutral}\ oxygen compared with
1048: singly ionized silicon implies these clouds are strongly ionized \citep[see, e.g.,][]{lehner01}.
1049: In all these components, we also find that the column densities of the highly ionized
1050: species are systematically larger than respective singly-ionized species,
1051: i.e., $N($\civ$) > N($\cii$)$ and $N($\siiv$) > N($\siii$)$, implying not only hydrogen
1052: fully ionized in the gas, but also the metals are highly ionized.
1053:
1054: While these components could certainly arise in the IGM and be unrelated
1055: to the DLA at $z \simeq 2.377$, the close velocity proximity to the DLA
1056: ($|v-v_{\rm LVC}| \la 100$ \km) suggests some association, possibly
1057: via outflows or infalling material on the DLA protogalaxy
1058: or in form of a highly ionized halo about the protogalaxy that does not
1059: share the same overall motion.
1060: Further, ``typical" diffuse IGM conditions would require simultaneously
1061: $\log n_{\rm H} \la -5$ ($\rho/\bar{\rho} \la 1$) and $[{\rm Z/H}] < -2$ \citep[e.g.,][]{simcoe06},
1062: and are unlikely to occur for these absorbers
1063: since the Hubble
1064: broadening $b_{\rm H} \approx 220 L({\rm Mpc})$ (where we assume
1065: a standard cosmology at $z\sim 2.4$) would be larger than the observed line
1066: widths of the individual components.
1067:
1068: Following, \citet{fox07}, the escape velocity of a DLA may be estimated if
1069: the virial velocity, $v_{\rm vir}$, of the DLA-protogalaxy is known and assuming
1070: a spherical geometry: $v_{\rm esc} \approx \sqrt{2} v_{\rm vir}$.
1071: \citet{haehnelt98} found $\Delta v_{\rm neut} \approx 0.5$--$0.75 v_{\rm vir} $ \citep[see also,][]{maller01}.
1072: For the DLA presented here $\Delta v_{\rm neut} \simeq 70 $ \km, hence
1073: none of the high-velocity components could escape the gravitational potential
1074: of the DLA. However, the relation between $v_{\rm vir}$ and $v_{\rm esc}$
1075: is not well known and the velocity width and virial velocity,
1076: while linearly correlated according to cosmological simulations, shows a large scatter around the median value.
1077: Also, in both \citet{haehnelt98} and \citet{maller01} simulations, the probability distribution of
1078: $\Delta v_{\rm neut} / v_{\rm vir} $ has a large scatter from $\Delta v_{\rm neut} / v_{\rm vir} \sim 0.2$
1079: to $\sim$1.5, enabling the high-velocity components to be also signatures of feedback-driven flows
1080: or even an escaping galactic wind from the protogalaxy. It could also be that our line-of-sight passes
1081: through a highly ionized halo of another galaxy \citep[see, e.g., Fig.~1 in][]{maller03} since
1082: clustering of galaxies near DLAs exists at both low and high redshifts
1083: \citep[e.g.,][]{chen03,bouche04,cooke06a,cooke06b,ellison07}.
1084: Yet it is unlikely that the fully ionized components probe the core of other galaxies
1085: (since there is no signature of neutral gas).
1086:
1087: The amount of total hydrogen is unknown in these ionized components,
1088: making it impossible to determine the metallicity. We therefore
1089: assume that all have a metallicity similar to the LVC, i.e., we set a priori
1090: $[{\rm Z/H}] = -1.4$ in our Cloudy simulations. This similarity in the metallicities might arise if
1091: the high-velocity ionized components have their origins in the
1092: protogalaxy or at least were polluted by the DLA itself or by another protogalaxy
1093: with a similar metallicity. Metallicity measurements in DLAs at $z \sim 2.4$ show
1094: a large scatter at this redshift with $[{\rm Z/H}]$ ranging from about $-0.5$ dex to $-
1095: 2.0$ dex, but with a mean around $-1.4$ dex
1096: \citep{prochaska03}. Yet if the high-velocity gas was only slightly polluted
1097: by materials from protogalaxies and is being accreted by the present DLA, a much
1098: lower metallicity may be expected \citep[e.g., see][]{simcoe04,simcoe06}.
1099:
1100: Below we review the properties of each high-velocity components and in particular
1101: explore the possible ionization mechanisms and the connection (if any) between
1102: the various ionized species. Our main aims are to understand (i) if ``simple'' photoionization
1103: and collisional ionization models can reproduce the observed properties (column density
1104: ratios and Doppler broadening) of the ionized gas;
1105: (ii) if \ovi\ and \nv\ {\em necessarily}\ probe {\em hot}\ collisionally ionized gas; (iii) the
1106: origin(s) of the highly ionized high-velocity components.
1107:
1108:
1109: \subsection{The negative high-velocity component}\label{sec-NHVC}
1110: For the NHVC, the velocity centroids at $-85$ and $-64$ \km\ are quite similar for
1111: \cii, \alii, \siiv, and \nv\ (see Figs.~\ref{spec1} and \ref{spec2}),
1112: suggesting some relationship between all these ions. Inspecting the
1113: ratios of the apparent column densities of various ions with respect to \siiv,
1114: we find that, in the velocity range
1115: $[-80,-40]$ \km, there is little or no variation for the low, intermediate,
1116: and high ions with respect to \siiv, suggesting that the ionization conditions do not change
1117: much over this velocity range. There is, however, a drop in $N({X^+})/N($\siiv$)$ at
1118: more negative velocities than $-80$ \km, which also corresponds to an increase of
1119: $N($\nv$)/N($\siiv$)$ and $N($\civ$)/N($\siiv$)$, implying that
1120: gas becomes more highly ionized at more negative velocities (which is the same
1121: pattern seen for the LVC).
1122:
1123: In the components at $-100$ and $-85$ \km, the $b$-values of \siiv\
1124: and \civ\ imply a temperature for the gas bearing these ions
1125: of a few times $10^4$ K. In the $-64$ \km\ component,
1126: higher temperatures may occur, with $T < 1.1\times 10^5$ K.
1127: At $T<7\times 10^4$ K, collisional ionization models produce
1128: a fraction of \cii\ larger than the fraction of \civ\ \citep[see][]{gnat07},
1129: inconsistent with our observations. Thus
1130: photoionization could play a role for the origin of \civ\
1131: and \siiv\ in components at $-100$ and $-85$ \km.
1132: Below we therefore model the ionization of the NHVC by producing first a Cloudy
1133: photoionization simulation that matches the integrated columns of the NHVC, which we
1134: feel are a more reliable indicator of the average properties of this gas
1135: than the individual component columns (see \S\ref{sec-fit}).
1136:
1137:
1138: \begin{table}
1139: \begin{center}
1140: \begin{minipage}{5 truecm}
1141: \caption{Total column densities for the NHVC \label{t5}}
1142: \begin{tabular}{lccc}
1143: \hline
1144: Species & $\log N$ \\
1145: \hline
1146: \oi\ $\lambda$1302 & $<12.94$ \\
1147: \cii\ $\lambda$$\lambda$1036,1334 & $13.69 \pm 0.05$ \\
1148: \civ\ $\lambda$$\lambda$1548,1550 & $>14.66 $ \\
1149: \nii\ $\lambda$1083 & $<12.77$ \\
1150: \nv\ $\lambda$$\lambda$1238,1242 & $13.27 \pm 0.13$ \\
1151: \ovi\ $\lambda$1031 & $>14.80 $ \\
1152: \alii\ $\lambda$1670 & $11.95 \pm 0.05 $ \\
1153: \aliii\ $\lambda$$\lambda$1854,1862 & $12.05 \pm 0.08$ \\
1154: \siii\ $\lambda$$\lambda$1260,1304,1524 & $12.67 \pm 0.08$ \\
1155: \siiii\ $\lambda$1206 & $>13.12$ \\
1156: \siiv\ $\lambda$1393,1402 & $13.70 \pm 0.02$ \\
1157: \feiii\ $\lambda$1122 & $< 13.20 $ \\
1158: \hline
1159: \end{tabular}
1160: Note: Upper limits are 3$\sigma$.
1161: \end{minipage}
1162: \end{center}
1163: \end{table}
1164:
1165:
1166: Fig.~\ref{cloudy-NHVC} shows the result of our Cloudy simulation using
1167: the Haardt \& Madau UV background (see \S\ref{ssec-ion}). In the left panel we
1168: show the predicted and observed column densities against the density. The metallicity
1169: is a priori assumed to be similar to the LVC. $N($\hi$)$ is
1170: chosen so that the model predicts enough column densities for the weakly ionized species.
1171: It is, however, quite possible, that other (NECI)
1172: mechanisms may produce both low and high ionization species (see below).
1173: We do not show $N($\oi$)$ and $N($\siiii$)$ in Fig.~\ref{cloudy-NHVC} because the estimated
1174: limits are not constraining.
1175:
1176: In the panel on the right-hand side of Fig.~\ref{cloudy-NHVC},
1177: we show the column density ratios for multiple
1178: ionization stages for C, N, Si, and Al against the density. In
1179: this representation, nucleosynthetic effects do not play any role since
1180: the same elements in different ionization stages are considered.
1181: We do make the assumption, however, that ions in different ionization
1182: stages are all produced by the same ionization mechanism.
1183: With those assumptions, the best solution is found at
1184: $\log U = -1.95$ where the observed ratios of
1185: \siii/\siiv\ and \alii/\aliii\ and the limits on \nii/\nv\ and \cii/\civ\
1186: agree with the model within about 1$\sigma$.
1187:
1188: \begin{figure*}
1189: \includegraphics[width = 16truecm]{f9.eps}
1190: \caption{{\em Left:} Predicted column densities for the Cloudy photoionization model of
1191: the NHVC assuming a Haardt-Madau (QSOs+galaxies) spectrum.
1192: Relative solar abundances are assumed. The various lines show the models for each
1193: atom or ion (the gray solid and dotted lines for \siiv\ and \siii\ are the Cloudy
1194: solution $-0.25$ dex, see text for more details), while the symbols are the measurements.
1195: Symbols with downward arrows are 3$\sigma$ upper limits
1196: and with upward arrows are lower limits. The vertical dotted line corresponds to the
1197: solution found from the right diagram. The vertical dot-dashed line correspond to
1198: another possible solution (see text for more details).
1199: {\em Right:} Ratio of column densities for the observed and models versus the hydrogen
1200: density. Data are plotted at $\log U = -1.95$ where within about 1$\sigma$ the
1201: observed ratios of \siii/\siiv\ and \alii/\aliii\ and the limits on \nii/\nv\ and \cii/\civ\
1202: agree with the model.
1203: \label{cloudy-NHVC}}
1204: \end{figure*}
1205:
1206:
1207: The vertical dotted line on the left panel of Fig.~\ref{cloudy-NHVC} shows this solution
1208: in the column density versus density plot. \civ, \nv, \nii, \alii, and \aliii\ are consistent
1209: with such a solution (since N may be deficient by $-0.9$ dex, see
1210: \S\ref{sec-prop}, this would require another mechanism to produce \nv, see below),
1211: but both \siii\ and \siiv\ are over-produced by 0.25 dex.
1212: The gray dotted and solid lines show the column densities of
1213: \siii\ and \siiv, respectively, minus 0.25 dex, which fit the data quite well.
1214: Therefore, if $[$Si/C$] = -0.25$ and $[$Al/C$] = 0$, this photoionization
1215: model provides a good description for ions with ionization potential less
1216: than about 50 eV. However, this relative abundance pattern is quite
1217: peculiar: neither depletion into dust nor standard nucleosynthesis yields can provide a simple
1218: explanation for such a relative abundance pattern. If the former plays
1219: a role, Al should also be deficient, while solar or supersolar
1220: $[$Si/C$]$ are usually observed in low metallicity protogalaxies.
1221:
1222: The assumption that all the ionization stages are produced by
1223: the same ionizing spectrum could be wrong. For example, if $\log U = -2.3$
1224: (shown in Fig.~\ref{cloudy-NHVC} by the dot-dashed line),
1225: there is a good agreement between the model and observation for
1226: \siiv, \aliii, and \cii\ within about 1$\sigma$.
1227: We note that this solution matches the \siii/\siiv\ ratio in the velocity
1228: range $[-100,-40]$ \km\ (i.e. the most negative velocity absorption is
1229: ignored where the gas becomes more highly ionized, see above).
1230: \civ\ would be under produced, but \civ\ can be also produced
1231: via collisional ionization (see below). However, with this latter solution,
1232: \alii\ and to lesser extent \siii\ would be overproduced and the \alii/\aliii\
1233: ratio does not fit the observations. The ionization potentials
1234: for \cii, \siii, and \aliii\ are in the range 11.3--24.4 eV, 8.2--16.4 eV, and 5.9--18.8 eV,
1235: respectively: the discrepancy between the model at $\log U = -2.3$ and
1236: observations increases as the minimum ionization potentials become smaller.
1237: It is therefore possible that the model with the current UV background produces too many
1238: photons at low energies. We note that the typical size for $\log U = -2.3$
1239: would be $L \sim 12$ kpc, more reasonable than at $\log U = -1.95$ where $L \sim 90$ kpc.
1240:
1241: It is possible that both the shape and intensity of the UV background
1242: vary in the Universe. In addition, the escape of ionizing flux from protogalaxies
1243: is uncertain and could change both the shape and intensity of the ionizing
1244: radiation. The impact of an ``internal" ionizing source could be similar.
1245: If the DLA is close to an AGN or a QSO, its ionizing radiation
1246: would have an important impact on the shape and intensity of the UV background.
1247: We therefore first investigated if a softer ionizing source
1248: would produce a better match to the observations, at least for the low
1249: ionized species. We considered ionization by a late O-type
1250: spectrum ($T_{\rm eff} = 35000$ K) with an atmospheric metallicity $[$Z/H$]=-1.4$
1251: and an atmosphere from the TLUSTY models \citep{hubeny95}. We find that a model
1252: with $\log N($\hi$) =17$ fits fairly well the observed column densities of
1253: the singly ionized species \cii, \siii, \alii, and \nii\ (assuming
1254: that N is deficient by at least $-0.7$ dex).
1255: On the other hand, column densities of the higher ionized gas (including \aliii)
1256: cannot be reproduced by this model, falling short by several orders of magnitude.
1257: While such a model cannot be sufficient on its own, it is possible that the true ionizing
1258: spectrum is a combination of soft and hard UV spectra. We therefore considered
1259: a Haardt \& Madau UV background from QSOs only, i.e. a UV background harder than
1260: considered previously. A solution with $\log U = -1.35$ ($\log n_{\rm H} = -3.5$)
1261: can match the column densities of \siiv\ and \nv\ (if nitrogen is deficient by $-1.7$ dex)
1262: and the lower limits on \civ\ and \ovi, while producing negligible \siii\ and \alii.
1263: But it fails to produce enough \aliii.
1264: The QSOs only background also produces a solution at $\log U = -2.05$ ($\log n_{\rm H} = -2.8$)
1265: that fits the \siii, \siiv, and \civ\ column densities. This model
1266: requires the relative abundance of Al to be deficient
1267: by $-0.6$ dex to match \aliii\ and, and in any case does not produce
1268: enough \civ, \nv, and \ovi.
1269:
1270: We have therefore not found a satisfactory Cloudy photoionization model with a single
1271: source of ionization unless the NHVC has
1272: peculiar abundances or the UV background model
1273: that includes galaxies and QSOs overestimates the contribution from low energy photons.
1274: The latter model and explanation ($\log U \simeq -2.3$) appear plausible because
1275: it better fits the \siii/\siiv\ ratio over the velocities $-90 \la v \la -40$ \km.
1276: We note that if $[{\rm Z/H}]$ increases in this model,\footnote{To
1277: first order, reducing/increasing the metallicity
1278: increases/decreases the hydrogen column density, but the effect is not solely
1279: linear and the shape of the ionization fractions as a function
1280: of temperature or density can change with $[{\rm Z/H}]$. We also note that
1281: if we adopted $\log N($\hi$) =17 \pm 0.5$ dex (with matching metallicities),
1282: $\log U$ would only change by about $\pm 0.2$ dex.} the discrepancy between
1283: model and observations for the singly-ionized species decreases, but modeled \siii\ and \alii\ column densities
1284: are still not in agreement with the observations even at $[{\rm Z/H}] = -0.5$, which is the current
1285: maximum metallicity observed in a DLA at $z\sim 2.4$. A hard spectrum would seem unlikely to produce the
1286: observed amount of \nv\ and \ovi\ (and some \civ) because the Hubble flow would
1287: broaden the absorption profiles too much. Therefore, other ionization mechanisms,
1288: including collisional ionization, may play a role in this gas.
1289:
1290: The observed high-ion ratios, broadenings of the high ions, and
1291: $N($\civ$)\gg N($\cii$)$ and $N($\siiv$)\gg N($\siii$)$ are
1292: incompatible with CIE or NECI models. Other models involving shocks or interfaces
1293: also require temperatures for the highly ionized gas of a few times $10^5$ K.
1294: Hence a combination of simple photoionization and collisional ionization models
1295: cannot reproduce all the observed properties, implying that the assumptions in these models
1296: are unlikely to hold for the NHVC, especially for the high ions. Therefore physical conditions
1297: are likely to be more complicated than these simplified scenarios (see \S\ref{sec-comp}).
1298:
1299: \subsection{The component at $+39$ \km}\label{sec-cloud39}
1300: As for the NHVC, the kinematics of the high, intermediate,
1301: low ions in the $+39$ \km\ component appear connected (see, e.g., \cii, \siii, \siiv\ in Fig.~\ref{spec2}).
1302: We modeled this component with Cloudy adopting the same assumptions as described in the
1303: previous section. In Table~\ref{t6}, we summarize the column densities for the various
1304: species. The low ions (\siii, \cii) and high ions (\siiv\ and about 80\% of $N($\civ$)$)
1305: can originate in photoionized gas when $\log n_{\rm H} = -2.0$,
1306: which implies $\log U = -2.7$, $T = 1.8 \times 10^4$ K,
1307: and $\log N({\rm H}) = \log N({\rm H}^+) = 19.4$ (see Fig.~\ref{cloudy-c39}).
1308: The linear size of the absorbing material is about 800 pc. Lower metallicity
1309: cannot be ruled out based on the broadening of component.
1310:
1311: As for the NHVC, the column densities of \ovi\ and \nv\ (and part of \civ)
1312: cannot be reproduced by ionization by the UV background from galaxies and QSOs.
1313: Assuming that the photoionization model is correct, the amount of \civ\
1314: not produced by this model would be $\log N($\civ$) \sim 13.2$. With
1315: this estimate, the ratios of \civ/\ovi\ and \nv/\ovi\ would be compatible
1316: with a CIE model only if N is deficient by at least $-1.2$ dex. However,
1317: this model requires $ T \sim 2 \times 10^5$ K, inconsistent with the $b$-value of \civ\ that
1318: implies $T < 7 \times 10^4$ K. For \ovi, the (uncertain)
1319: $b$-value does not suggest temperatures higher than $10^5$ K. It is not clear if other models
1320: (e.g., conductive interfaces) in low temperature conditions could reproduce the observed ratios.
1321: As for the NHVC, we discuss the properties of this component in the context
1322: of feedback models in \S\ref{sec-comp}.
1323:
1324:
1325: \begin{figure}
1326: \includegraphics[width = 8truecm]{f10.eps}
1327: \caption{Predicted column densities for the Cloudy photoionization model of
1328: $+39$ \km\ component assuming a Haardt-Madau (galaxies+QSOs) spectrum.
1329: Relative solar abundances are assumed.
1330: The various lines show the models for each atoms or ions, while the symbols
1331: are the measurements. Symbols with downward arrows are 3$\sigma$ upper limits,
1332: upward arrows are lower limits. An adequate solution is found for $\log U = -2.7$, which
1333: implies $\log n_{\rm H} = -2.0$, $T = 1.8 \times 10^4$ K, and $\log N({\rm H}) = \log N({\rm
1334: H}^+) = 19.4$; i.e. this cloud is fully ionized. The implied path length is $L = 0.8$ kpc.
1335: The most discrepant values to this model are the high ions
1336: (\ovi, \nv, \civ), which likely probe other ionization mechanism(s)
1337: (see \S\ref{sec-cloud39} for more details).
1338: \label{cloudy-c39}}
1339: \end{figure}
1340:
1341: \begin{table}
1342: \begin{center}
1343: \begin{minipage}{5 truecm}
1344: \caption{Column densities of the $+39$ \km\ component \label{t6}}
1345: \begin{tabular}{lccc}
1346: \hline
1347: Species & $\log N$ \\
1348: \hline
1349: \oi\ $\lambda$1302 & $<12.71$ \\
1350: \ovi\ $\lambda$$\lambda$1031,1037 & $13.19 \pm 0.14$ \\
1351: \cii\ $\lambda$$\lambda$1334 & $13.28 \pm 0.05$ \\
1352: \civ\ $\lambda$$\lambda$1548,1550 & $13.50 \pm 0.11$ \\
1353: \nii\ $\lambda$1083 & $<12.48$ \\
1354: \nv\ $\lambda$$\lambda$1238 & $<12.21$ \\
1355: \aliii\ $\lambda$$\lambda$1854,1862 & $<11.61$ \\
1356: \siii\ $\lambda$1193 & $12.37 \pm 0.05$ \\
1357: \siiv\ $\lambda$1393,1402 & $12.77 \pm 0.13$ \\
1358: \feiii\ $\lambda$1122 & $<12.88 $ \\
1359: \hline
1360: \end{tabular}
1361: Note: Upper limits are 3$\sigma$.
1362: \end{minipage}
1363: \end{center}
1364: \end{table}
1365:
1366:
1367:
1368: \subsection{The positive high-velocity component}\label{sec-hvc}
1369: With only the detection of \civ, and possibly \ovi\ and \nv, the PHVC ($[+60,+140]$ \km)
1370: appears even more highly ionized than the gas probed by the NHVC, although the
1371: column densities of \civ\ are quite similar in both components and the column density of \ovi\
1372: appears smaller in the PHVC compared to the NHVC.
1373: In Table~\ref{t7}, we summarize the measurements for the high ions and the
1374: limits for other strong transitions. Because $N($\ovi$)$ is uncertain, we assume the
1375: $N($\civ$)/N($\ovi$)$ ratio is constant over the profile and use $N($\civ) to estimate
1376: $N($\ovi$)$. This gives the value reported in Table~\ref{t7}. However, for the ionic
1377: ratio used below, we will also take into account the direct measurement $\log N($\ovi$) \le 14.18$
1378: reported in Table~\ref{t1}
1379: (where the lower-equal sign emphasizes that the \ovi\ $\lambda$1037 could be tainted).
1380:
1381: \begin{table}
1382: \begin{center}
1383: \begin{minipage}{5 truecm}
1384: \caption{Total column densities for the PHVC \label{t7}}
1385: \begin{tabular}{lccc}
1386: \hline
1387: Species & $\log N$ \\
1388: \hline
1389: \cii\ $\lambda$1334 & $<12.63$ \\
1390: \civ\ $\lambda$$\lambda$1548,1550 & $13.68 \pm 0.01$ \\
1391: \nii\ $\lambda$1083 & $<12.63$ \\
1392: \nv\ $\lambda$1238 & $\le 12.80:$ \\
1393: \ovi\ $\lambda$1037 & $\le 13.83:$$^a$ \\
1394: \siiii\ $\lambda$1206 & $<11.59$ \\
1395: \siiv\ $\lambda$1393 & $<11.85$ \\
1396: \hline
1397: \end{tabular}
1398: Note: Upper limits are 3$\sigma$. Because \nv\ is partially
1399: contaminated, its measurement remains uncertain. $a$: We report the \ovi\
1400: column density derived from $N($\ovi$)/N($\civ$) \simeq 1.4$ in the velocity
1401: range $[85,110]$ \km\ (see Figure~\ref{aodhvc}) and $N($\civ). The ``$\le$''
1402: sign emphasizes that the line could be partially contaminated by an unrelated absorber.
1403: \end{minipage}
1404: \end{center}
1405: \end{table}
1406:
1407: The upper limit on the temperature, $T < 1.2 \times 10^5$ K (derived from the
1408: profile fitting to the \civ\ doublet, see Table~\ref{t2}), could suggest
1409: that collisional ionization may play a role if thermal motions dominate
1410: the broadening and if the \ovi\ absorption
1411: is largely contaminated. However,
1412: at such relatively low temperatures, one would expect a priori
1413: detection of other ions, in particular \siiv. Models involving
1414: interfaces (see \S\ref{ssec-ion}) are ruled out because
1415: even strong transitions of low and intermediate ionized species
1416: (e.g., \siiii\ $\lambda$1206, \cii\ $\lambda$1334) are not
1417: detected. The absence of these very strong
1418: transitions is rarely observed in highly ionized gas.
1419: The high-ion ratios that can help us to discern between ionization processes
1420: are unfortunately quite uncertain: For \ovi, we have to make
1421: the dubious assumption that the absorption is not contaminated by some metal line
1422: absorption. The \nv\ absorption is partially blended
1423: at 1238 \AA\ and completely contaminated at 1242 \AA; therefore only an upper limit can be derived.
1424: Furthermore, nitrogen is likely to be deficient with respect to the other elements. With these assumptions
1425: and caveats, we find that $N($\civ$)/N($\siiv$) > 68$, $N($\civ$)/N($\ovi$) \ge 0.3$--0.7
1426: (the first value is derived using the total \ovi\ column density
1427: from the profile fitting -- see Table~\ref{t1}, the second value
1428: is derived using the AOD column density ratio in the velocity
1429: range $[+85,+110]$ \km), and $N($\civ$)/N($\nv$)\ga 7.7$
1430: (but note that N is likely deficient relative to C). These high-ion ratios with the additional
1431: constraints on the temperature and the absence of detection
1432: of low ions are not consistent with any models involving collisional
1433: ionization.
1434:
1435:
1436: \begin{figure}
1437: \includegraphics[width = 8truecm]{f11.eps}
1438: \caption{Predicted column densities for the Cloudy photoionization model of
1439: the PHVC assuming a Haardt-Madau (galaxies+QSOs) spectrum.
1440: Relative solar abundances are assumed. The various lines show the models for
1441: each atoms or ions. The horizontal yellow line shows the range of observed
1442: \civ\ column density at 2$\sigma$. Other estimates are reported in Table~\ref{t7}.
1443: The dashed line represents the solution for \nv\ with $[{\rm N/\alpha}] = -0.9$.
1444: (see \S\ref{sec-hvc} for more details).
1445: \label{cloudy-hvc}}
1446: \end{figure}
1447:
1448: We have therefore produced a Cloudy photoionization
1449: simulation with the same assumptions used for the other components,
1450: in particular that the gas has a metallicity [Z/H$]=-1.4$,
1451: similar to the LVC, and the UV background is from galaxies
1452: and QSOs. In Fig.~\ref{cloudy-hvc}, we show the Cloudy model
1453: that matches the \civ\ column density for $-1.1\la \log U \la -0.8$.
1454: Within this range, this model predicts an \ovi\
1455: column density consistent with the range of estimated values, assuming weak
1456: or no blending. The model
1457: produces too much \nv\ when nitrogen has solar relative abundance
1458: with respect to the other metals. However,
1459: the model is consistent with $N($\nv$)$ if nitrogen is deficient
1460: by $-0.9$ dex (dashed line in Fig.~\ref{cloudy-hvc}).
1461: All the other limits reported in Table~\ref{t7} are consistent
1462: with the model. Models with $\log U \ll -1.1 $ and $\log U \gg -0.7 $
1463: are rejected because they would contradict the limits on \siiii\ (and \siiv)
1464: and \ovi, respectively. The path length for this cloud would be quite large
1465: ranging between 50 to 250 kpc. However, if the UV background is dominated by only the QSOs,
1466: suitable solutions can also be found with smaller $U$, and hence reducing the path length.
1467: Therefore, despite the absence of low ions, the PHVC does not appear to probe hot
1468: gas but could represent an extended cool, low density, highly photoionized layer of gas.
1469:
1470: \subsection{Origin(s) of the highly ionized high-velocity gas}\label{sec-comp}
1471: So far we have discussed the highly ionized high-velocity gas in the
1472: context of simple ionization models. We have shown that photoionization by the UV background
1473: is a possible source for producing the low and intermediate ions, and \siiv\ and part
1474: of \civ. But except for the PHVC, \ovi\ and \nv\ cannot be produced
1475: by photoionization by the UV background. Photoionization from
1476: point sources (e.g., from extremely massive stars) is also unlikely as it is not clear that
1477: such stars can produce enough $> 113.9$ eV photons to produce the observed
1478: strong absorption in \ovi. It would also require
1479: the highly ionized gas to be extremely close to the ionizing object,
1480: which is not consistent with the many components observed at both high positive
1481: and negative velocities (we also note that this scenario would have problem
1482: with the excellent match in the kinematics between low and high ions since the required
1483: large $U$ would make the absorption of the low ions weak).
1484: For \ovi\ and \nv\ (and in part \civ),
1485: CIE and NECI models fail to reproduce the
1486: observed properties (ionic ratios and broadening) of the high ions.
1487: There is some evidence that the highly ionized gas can be at low temperatures
1488: based on the $b$-values obtained from the profile fits. The highly ionized gas
1489: must therefore be relatively dense ($n\ga 0.01$ cm$^{-3}$), or have a relatively
1490: high metallicity, or have a combination of both
1491: in order to be able to cool to a few times $10^4$ K. Indeed,
1492: the instantaneous radiative cooling time can be written as \citep[e.g.,][]{gnat07}
1493: \begin{eqnarray*}
1494: t_{\rm cool} & = & 219 \left(\frac{T}{10^6\, K}\right) \left(\frac{n}{10^{-3}\, {\rm cm}^{\-3}}\right)^{-1} \\
1495: & & \times \left(\frac{\Lambda(T,Z)}{1.3\times 10^{-22}\, {\rm erg\,cm^3\,s^{-1}}}\right)^{-1} \, {\rm Myr},
1496: \end{eqnarray*}
1497: %$$
1498: %\tau_{\rm cool} = 219 \left(\frac{T}{10^6\, K}\right) \left(\frac{n}{10^{-3}\, {\rm cm}^{\-3}}\right)^{-1}
1499: % \times \left(\frac{\Lambda(T,Z)}{1.3\times 10^{-22}\, {\rm erg\,cm^3\,s^{-1}}}\right)^{-1} \, {\rm Myr},
1500: %$$
1501: where $\Lambda(T,Z)$ is the cooling function, which is dependent on the temperature and the metallicity.
1502: This expression is valid if the gas is isochoric
1503: \citep[if instead it is isobaric, 219 should be replaced by 365, see][]{gnat07}.
1504: For subsolar metallicity ($Z = 0.01 Z_\odot$), and $n = 10^{-3}$ cm$^{-3}$, it would take more than
1505: the age of the Universe at the absorption epoch ($t_{\rm H}$) for a $10^6$ K gas to cool (if the gas had initially $10^5$ K,
1506: it would take less than 100 Myr to cool). Therefore, a scenario involving accretion of primordial and low density
1507: gas is not viable if the initial temperature of the hot gas is at least
1508: $10^6$ K since $t_{\rm H} < t_{\rm cool}$. Gas at such high temperatures may exist
1509: if it was heated, e.g., by supernovae feedback. As the density and the metallicity increase, the
1510: gas can cool more rapidly.
1511: It is more likely that the highly ionized high-velocity gas is
1512: a signature of either protogalactic winds/outflows or intra-galaxy, enriched material
1513: falling onto the protogalaxy.
1514:
1515: At $z>1$, feedback processes such as galactic winds/outflows
1516: or shock heating by supernova remnants are expected to be more common
1517: because star formation is more active, and these are in fact essential
1518: for galaxy formation and evolution \citep[e.g.,][]{croton06,mashchenko08}.
1519: On the other hand galaxy mergers and clustering of galaxies are
1520: also important at high redshift, favoring the enrichment of gas near
1521: protogalaxies that may fall onto them or the possibility that a sightline intersects unrelated absorbers
1522: to the DLA. \citet{fox07a} favor a scenario
1523: involving galactic winds because they find high-ion absorption at velocities
1524: larger than the escape velocities of the protogalaxies. However, the knowledge
1525: of $v_{\rm esc}$ in DLAs is not well known (see \S\ref{sec-ion-over}), and we believe there is not
1526: yet enough evidence to favor one scenario over the other one. In principle,
1527: one could use cosmological simulations of protogalaxies that incorporate feedback mechanisms
1528: to try to disentangle these scenarios.
1529: However, many aspects of feedback are too complex to reliably model (e.g., metal cooling,
1530: non-equilibrium ionization effects) and current simulations often lack the spatial and mass resolutions
1531: to resolve the supernova environment \citep[e.g.,][]{dalla08}. Further
1532: \citet{fangano07} argue it is even difficult to differentiate between outflowing
1533: and infalling materials in simulations. Nevertheless, we find interesting
1534: that these recent models can produce low temperatures (a few times
1535: $10^4$ K) for the highly ionized gas in the galactic outflows, which is so
1536: difficult to produce in the simplified ionization models used above \citep{oppenheimer06,kawata07,fangano07}.
1537: These models also predict a similarity in the kinematics of
1538: the low, intermediate, and high ions as observed here.
1539:
1540: Local analogs of highly ionized HVCs may also
1541: help to better understand the origin(s) of the highly ionized HVCs at high $z$.
1542: With the {\em Far Ultraviolet Spectroscopic Explorer (FUSE)}\
1543: and the Space Telescope Imaging Spectrograph (STIS) on board of the {\em Hubble Space Telescope},
1544: many highly ionized HVCs (i.e. gas that moves at much different
1545: velocities than gas associated with the Galactic disk) were discovered in the halo of the Milky
1546: Way toward extragalactic sightlines \citep{sembach99,lehner02,sembach03,collins04,collins05,fox06}.
1547: At least 60--70\% of the high-latitude sky is in fact covered by \ovi\ HVCs \citep{sembach03}.
1548: It is interesting to note that both high and low redshift highly ionized
1549: HVCs share some common properties: By definition, they are highly ionized,
1550: but they also often show very complex kinematic and ionization structures, they are often
1551: multiphase and (nearly) fully ionized, and there seems to be some kinematic connection
1552: between the low and high ions.
1553:
1554: As for their high redshift counterparts, the origin of low redshift HVCs remains
1555: largely enigmatic, and both a ``Galactic" origin (where the HVCs are the result of feedback processes
1556: and trace the disk-halo mass exchange, perhaps including the accretion of matter condensing
1557: from an extended corona) and a ``Local Group" origin (where they are part of the local warm-hot
1558: intergalactic medium) have been proposed \citep[e.g.,][]{sembach99,nicastro03,gnat04,fox06}.
1559: Evidence for and against any of these theories are mostly circumstantial,
1560: although recently \citet{zech07} have discovered two highly ionized HVCs toward
1561: the star ZNG\,1 in the globular cluster M\,5, firmly demonstrating that some of
1562: these highly ionized HVCs originate near the Galactic disk and are signatures
1563: of Galactic-feedback (the metallicity being supersolar for these HVCs).
1564: Other highly ionized HVCs associated to galactic
1565: feedback mechanisms are also found near the LMC \citep{lehner07}.
1566: On the other hand, some HVCs are also clearly related to an accretion phenomenon or tidal products
1567: some HVCs are also related to an accretion phenomenon or tidal products
1568: \citep[e.g., Complex C and Magellanic Stream,][]{tripp03,sembach03}.
1569: Therefore, while an IGM
1570: origin cannot be ruled out for some HVCs, several highly ionized HVCs in the
1571: low redshift Universe have been associated with galactic feedback and accretion processes.
1572: At high redshift we noted that the column densities of the high ions are far too
1573: large for these ions to arise in the IGM.
1574: At both high and low redshifts, highly ionized HVCs may therefore be connected to
1575: gas outflowing from the host protogalaxy or galaxy. For example, \citet{simcoe06} presented
1576: evidence of chemically enriched high-velocity \ovi\ and \civ\ absorbers
1577: near galaxies at $z\sim 2.3$ with possible high [Si/C] suggestive of
1578: supernovae type II enrichment, and therefore favoring outflowing matter.
1579: For the present DLA, there is not enough evidence to favor feedback over
1580: accretion.
1581:
1582: While we listed similar properties for the highly ionized HVCs at low and high $z$, they
1583: also have differences. At high $z$, \civ\ (IP: 47.9--64.5 eV) appears to trace
1584: both photoionized and collisionally ionized gas, much like
1585: \siiv\ (IP: 33.5--45.1 eV) at low redshift. Narrow components in \nv\ and \ovi\ absorption
1586: (and \civ\ unexplained by photoionization by the UV background)
1587: are observed at high $z$, but are relatively rare or absent at low redshift
1588: (although part of this could be an instrumental bias since the spectra of the \ovi\ HVCs at $z\sim 0$
1589: have $R \simeq 20,000$ compared to 45,000 for the present observations -- e.g., Savage, Sembach, \& Howk
1590: 2001 reported a narrow \nv\ component with $b = 7 \pm 2$ \km\ in the Milky Way gas).
1591: The high-velocity absorption near DLAs is often much stronger than near
1592: our Galaxy and the high ion ratios appear to vary much more in DLAs (at least in the present DLA)
1593: than in or near the Galaxy. The change in the ionizing background spectral energy distribution
1594: may explain some of these differences for \siiv\ and part of \civ.
1595: However, as we have demonstrated here, many of the highly ionized high-velocity
1596: components (\ovi, \nv, and some \civ) in the DLA cannot be explained by
1597: photoionization alone. Therefore some of the listed differences above between
1598: the highly ionized high-velocity gas at low and high $z$ also suggest an evolution in
1599: the properties of feedback processes or the environment near the (proto)galaxies.
1600:
1601:
1602: \section{Discussion}\label{sec-disc}
1603: Before summarizing our main findings, we argue that this DLA is typical in many regards.
1604: From a standpoint of the metallicity ($[{\rm Z/H}] = -1.41 \pm 0.08$), the main
1605: component of the DLA at $z=2.377$ toward J1211+0422 is actually quite ordinary
1606: \citep[see, e.g.,][for a summary of the metallicity evolution of DLAs]{prochaska03}.
1607: The full velocity width of the neutral absorption of about 70 \km\ is also not
1608: far from the median value of 90 \km\ derived in a large sample of DLAs \citep{prochaska97}.
1609: The $N($\hi$)$ value lies near the median of the DLA \hi\ column density distribution.
1610: The $\alpha$-element enhancement relative to Fe-peak elements and N is also not
1611: out of the ordinary \citep{lu96,prochaska02,petitjean08}. But what may
1612: really distinguish this DLA from others is the absence of cold neutral gas signature. However,
1613: $32/76 = 42$\% of the DLAs studied by \citet{wolfe08} have no detectable \cii*\ absorption,
1614: possibly indicating the neutral gas is warm. None of the upper limits derived on $l_c$ are as stringent
1615: as in the present DLA and improving the signal in the spectra of these DLAs would help to understand
1616: if the population of DLAs with WNM is important.
1617:
1618: We also saw that the high, intermediate, and low ions in
1619: this DLA seem to follow each other quite well kinematically.
1620: \citet{wolfe00a} found
1621: that \civ\ and low ions rarely follow the same kinematic distribution,
1622: \siiv\ and \civ\ appear strongly correlated, and in many cases the
1623: \aliii\ profiles are better correlated with those of the low ions than with those
1624: of \civ\ and \siiv\ \citep[see also][]{lu96}. However, even though
1625: the components of the high-ion profiles are generally disjoint from the
1626: low-ion profiles, \citet{wolfe00a} noted that the high-ion
1627: and low-ion profiles are roughly centered on similar velocities and
1628: that the high-ion profiles have a full-width systematically similar
1629: to or larger than that low-ion profiles, properties similar to the present
1630: DLA \citep[see also][]{fox07a,fox07}. The ionized DLA described
1631: by \citet{prochaska02a} shows
1632: some correlation between the velocity centroids of the low, intermediate, and
1633: high ions.
1634: The absence of clear kinematic alignment between the high, intermediate, and low-ion profiles
1635: in every DLA may, however, not be entirely surprising. In our own Galaxy, there is a significant
1636: variety in the type of associations between the various phases of the interstellar gas,
1637: which are generally attributed to sightline projection effects and the physical processes
1638: that give rise to the high-ions.
1639:
1640: We also found that the amount of ionized gas associated to this DLA is very large.
1641: To gauge the importance of the highly ionized gas relative to the neutral gas
1642: we can estimate the amount of hydrogen in the highly ionized phase from
1643: $N($\hii$) = N($\nv$)/(f_{\rm N\,V} \, ({\rm N/H})_\odot 10^{[{\rm N/H}]})$, where $f_{\rm
1644: N\,V} = N($\nv$)/N({\rm N}) \simeq 0.2$
1645: is the ionization fraction of N at $T \simeq 2 \times 10^5$ K
1646: when using the NECI calculations from \citet{gnat07}.
1647: At $T$ greater or lower than $2 \times 10^5$ K, $f_{\rm N\,V}< 0.2$.
1648: We assume that the N abundance is the same in the neutral
1649: and ionized phase, so that $[{\rm N/H}] = [$\nni/\hi$] \simeq -2.3$.
1650: For the LVC, $\log N($\hi$) = 20.8$, $\log N($\hii$) \sim 20.5$ (highly ionized gas)
1651: and $\log N($\hii$) \sim 20.1$ (photoionized gas).
1652: For the NHVC, $\log N($\hii$) \sim 20.5$ (highly ionized gas)
1653: and $\log N($\hii$) \sim 20.2$ (photoionized gas). Using the lower limits
1654: on $N($\ovi$)$ (Tables~\ref{t4a} and \ref{t5}), we find consistent \hii\ amounts
1655: for the highly ionized gas with $\log N($\hii$) > 20.3$
1656: in the LVC and NHVC (in that case $[{\rm O/H}] = -1.4$ and $f_{\rm O\,VI} \la 0.2$).
1657: The $+39$ \km\ component and PHVC are negligible with
1658: respect to hydrogen amounts in the LVC and NHVC.
1659: Summing the various components, $N_{\rm tot}($\hii$)/N($\hi$) \ga 1.5$,
1660: where $N_{\rm tot}($\hii$)$ includes contribution from both the photoionized gas and
1661: highly ionized gas. As we discuss above there is some uncertainty
1662: in the processes that give rise to \ovi\ and \nv. Nevertheless the ionization
1663: fraction of these species is unlikely to be much more than 0.2 based
1664: on CIE/NECI and photoionization models. The metallicity
1665: correction in the high-velocity components is uncertain because no direct metallicity
1666: estimates can be done. If the metallicity is not similar to the DLA
1667: itself, it is more likely to decrease in these components, although
1668: if there is enough pollution from other nearby protogalaxies an increase
1669: may be possible. The limit $N_{\rm tot}($\hii$)/N($\hi$) \ga 1.5$ is larger than those typically
1670: derived by \citet{fox07} for DLAs with similar total \hi\ column density, but is comparable to
1671: ionized-to-neutral ratios derived in DLAs with lower $N($\hi$)$ and sub-DLAs
1672: \citep{fox07,fox07c}. The use of \nv\ may have allowed us to simply derive a more stringent lower limit.
1673:
1674: Finally, we observe that \citet{fox07} argue that the highly ionized gas
1675: may help to solve (in part) the missing metal ``problem" at high redshift \citep{pettini99} since if the
1676: \ovi-bearing gas has $T\sim 10^6$ K, then $f_{\rm O\,VI} \sim 0.003 $ (according to CIE models), which
1677: would imply that the \ovi\ phase has enough metals to solve the issue.
1678: In our analysis, several components have $T \la 3\times 10^5$ K, consistent
1679: with $f_{\rm O\,VI} \la 0.2$. However, there is also evidence for \ovi-bearing gas with $T<9\times 10^4$ K,
1680: which would also imply $f_{\rm O\,VI} \ll 0.2$ if CIE/NECI applies. But CIE/NECI (and photoionization) provides a
1681: poor description of this gas.
1682: To quantify accurately the importance of the highly ionized phase in DLAs requires first
1683: a better understanding of the high-ion absorption in a large and (relatively) uncontaminated sample
1684: (with the purpose of obtaining a statistical sample where the distribution of $b$ in the {\em individual}\
1685: components could be studied), the ionization fractions of the high ions (requiring models with more realistic physics),
1686: and metallicities in the ionized plasma (which is likely to rely on similar assumptions made in
1687: the present work).
1688:
1689:
1690:
1691:
1692: \section{Summary}\label{sec-sum}
1693: We have presented Keck HIRES observations of the absorbing material associated with a DLA at
1694: $z\simeq 2.377$ along the QSO J1211+0422 ($z_{\rm em} = 2.5412$) sightline with the goal of
1695: exploring the properties of the highly ionized gas, its connection with the neutral gas, its origin(s),
1696: and its possible relationship to feedback processes.
1697: The absorbing material at $z\simeq 2.377$ has little contamination, especially
1698: in key species, such as \siiv, \civ, and \nv, which allows us to undertake detailed
1699: modeling of the high-ion profiles, to assess their relationship with low ions and study
1700: the ionization mechanisms at play.
1701:
1702: The main results of our analysis for the gas that has velocities in the range $[-40,+35]$
1703: \km\ interval where the neutral absorbing material is confined are as follows:
1704:
1705: 1. Where there is neutral absorbing material, intermediate- and high-ion absorption
1706: is also observed. The profiles of the low and high ions have several
1707: components. The profile fits to the high ions show that there is a close kinematic relationship
1708: between the high and low ions.
1709:
1710: 2. The metallicity of this DLA, $[{\rm Z/H}] = -1.41 \pm 0.08$, is similar to
1711: the average metallicity of other DLAs at $z\sim 2.4$. With $[{\rm Si/S}] = +0.02 \pm 0.06$,
1712: there is little evidence for silicate grains. The low ratio
1713: of $[{\rm N/Si}] = -0.88 \pm 0.07$ implies that nucleosynthesis plays
1714: a role in the observed relative abundances. Hence, the enhancement of
1715: $[{\rm Si/Fe}] = +0.23 \pm 0.05$ could be solely due to nucleosynthesis. However,
1716: within the uncertainties, a mixture of dust depletion of Fe and Ni ($[{\rm Si/Ni}]\simeq [{\rm Si/Fe}]$)
1717: and nucleosynthesis is a possibility.
1718:
1719: 3. A Cloudy simulation, where the source
1720: of the photoionization is dominated by the Haardt \& Madau galaxies plus QSOs spectrum,
1721: reproduces well the column densities of the singly- and doubly-ionized species.
1722: Within this simulation, about 85\% of the gas is neutral and the gas is warm ($T\sim 10^4$ K).
1723: However, this model fails completely to produce large enough high-ion columns (especially for \civ, \nv, and \ovi).
1724: Possible models for the production of the high ions include conductive interfaces
1725: or shocks. Including the highly ionized gas, the column of ionized gas in this
1726: velocity range is similar to the amount of the neutral gas, with
1727: $N($\hii$)/N($\hi$) \ga 0.8$.
1728:
1729: 4. We derive the lowest upper limit on \cii*\ in a DLA to date,
1730: which implies a cooling rate $l_c < 10^{-27.8}$ erg\,s$^{-1}$ per H atom at 3$\sigma$.
1731: Using the models of \citet{wolfe03} and assuming a Fe dust depletion of $[{\rm Fe/Si}] = -0.2$,
1732: we show that the neutral gas can only be warm. We also show that the star formation rate is
1733: $< 7.1\times 10^{-3}$ M$_\odot$\,yr$^{-1}$\,kpc$^{-2}$, i.e. we cannot rule out
1734: that star formation may occur at a rate about twice the local rate of the Milky Way.
1735: These conclusions hold even if the environment is essentially free of dust.
1736: DLAs may therefore have warm neutral gas and ongoing star formation at the same time.
1737: Improved upper limits on $N($\cii*$)/N($\hi$)$ in DLAs may help to discern the importance of this population
1738: of warm neutral absorbers.
1739:
1740: 5. This DLA is made up of warm neutral and warm (and possibly hot) ionized gas
1741: that are comoving with similar velocities given the close
1742: kinematic relationship between the low and high ions.
1743: Despite the similar velocities for all the species, the DLA has multiple
1744: phases. A picture where the neutral gas is surrounded by a highly ionized halo is plausible,
1745: but not unique. Because several components are observed, it is also possible that we are seeing
1746: several sheets of neutral gas separated by weakly and highly ionized gas.
1747:
1748: For gas outside of the LVC absorption, i.e., at
1749: $|v| \ga +39$ \km, absorption is only observed in singly-ionized and higher ionized
1750: species (i.e. there is no signature of neutral gas). The amount of gas
1751: in the highly ionized phase is larger than the amount in the weakly ionized phase.
1752: Our main results for the high-velocity gas are as follows:
1753:
1754: 6. In the velocity range $[+80,+120]$ \km, {\em only}\ \civ\ and \ovi\
1755: absorption is detected. This gas has one of the highest known \civ/\siiv\ ratios,
1756: with $N($\civ$)/N($\siiv$)>68$. Other ions with very strong transitions, such as \cii\ and
1757: \siiii, are totally absent. Paradoxically, the observed properties ($N, b$) are better
1758: explained with a cool, photoionized model than a model involving collisional ionization
1759: (in or out of equilibrium).
1760:
1761: 7. For the components at $+39$ \km\ and in the velocity range $[-120, -40]$ \km,
1762: low and high ions are detected and appear to be kinematically
1763: related. A photoionization model using the Haardt \& Madau (galaxies+QSOs) spectrum
1764: can reproduce the column densities of the ionized species in the $+39$ \km\
1765: component, except those of \ovi\ and \nv. For the gas in the velocity range $[-120, -40]$ \km,
1766: Cloudy results are difficult to reconcile with the observations without invoking
1767: peculiar relative abundances (subsolar $[{\rm C/Si}]$ and solar $[{\rm C/Al}]$)
1768: or a different spectral energy distribution for the ionizing spectrum. In any case, photoionization
1769: does not appear to be a major source for \ovi\ and \nv. The broadening of the profiles implies temperatures
1770: that are too low to produce \nv\ and \ovi\ (and in part \civ) in CIE (e.g., at $-100$ \km, $b($\ovi$) = 9.5 \pm 1.1$ \km,
1771: implying $T <10^5$ K; in several components of \civ, $b($\civ$) \approx 4$--10 \km, implying $T\la 1$--$7\times 10^4$ K).
1772: This indicates that cooling must be important and efficient, hence requiring the gas
1773: to be relatively dense ($n_{\rm H} > 10^{-3}$ cm$^{-3}$) and/or rich in metals.
1774:
1775: Combining our results with other works \citep[in particular][]{fox07},
1776: we make some more general conclusions concerning the high ions observed in high redshift DLAs:
1777:
1778: 8. In the present DLA, \siiv\ probes predominantly photoionized gas, while \civ\ has the most
1779: ambiguous origin. Photoionization by a hard spectral source (e.g., QSOs) can produce
1780: significant \civ\ absorption but is generally not sufficient to explain the total observed column.
1781: In our Galaxy, \siiv\ is the ``ambiguous'' high ion. The difference in the spectral energy distribution
1782: of the ionizing background flux between the high and low $z$ Universe
1783: may explain this difference for the most part. In view of other \siiv\ and \civ\ studies in DLAs,
1784: these conclusions are likely to hold for the whole population of DLAs.
1785:
1786: 9. Combining our results with those of \citet{fox07}, DLAs can have strong \ovi\
1787: absorption with in some cases \nv\ absorption. In our DLA and for another DLA in
1788: \citet{fox07}, simple photoionization models fail to reproduce enough columns for these ions.
1789: However, these ions do not necessarily probe hot ($T> 2\times 10^5$ K) gas, i.e. their mere presence
1790: does not necessarily imply the presence of hot gas. Non-thermal, turbulent motions may broaden
1791: the high-ion profiles, and for the present DLA, there is evidence for narrow ($T <10^5$ K)
1792: components not associated with photoionized gas. We also find that
1793: simple CIE and non-equilibrium radiative cooling ionization models fail for these
1794: ions and part of \civ\ when the line-widths are available
1795: (i.e. the models cannot reproduce simultaneously the high-ion ratios and $b$-values).
1796: Both collisional and photoionization models are likely more complex than those presented here.
1797: However, in view of our results and those of \citet{fox07}, we believe it is safe to conclude that
1798: \nv\ and \ovi\ (and part of \civ) associated to DLAs must
1799: exist in a separate phase than the photoionized phase. In that sense,
1800: they appear better tracers of feedback (or accretion) processes in high redshift DLAs than \siiv\ and \civ.
1801:
1802: Future detailed analysis of \ovi\ and \nv\ absorption in a relatively uncontaminated and larger sample
1803: than presently published would be extremely useful, in particular to constrain the $b$-value distribution
1804: of the individual components. \civ\ and \siiv\ should not be dismissed since they
1805: are far less contaminated by the Ly$\alpha$ forest.
1806: A detailed profile fitting on a large sample of \civ\ and \siiv\ absorbers associated to DLAs
1807: in order to extract the individual $b$-values would be very valuable since \siiv\ mostly
1808: probes photoionized gas while \civ\ probes both photoionized and collisionally ionized gas.
1809: This work provides a step in that direction and we
1810: plan to further investigate the nature of the highly ionized gas in DLAs in the future.
1811:
1812:
1813:
1814: \section*{acknowledgements}
1815:
1816: The authors wish to recognize and acknowledge the very significant cultural
1817: role and reverence that the summit of Mauna Kea has always had within the
1818: indigenous Hawaiian community. We are most fortunate to have the opportunity
1819: to conduct observations from this mountain. We also acknowledge the Keck support
1820: staff for their efforts in performing these observations. This research has made use of
1821: the NASA's Astrophysics Data System Abstract Service, the SIMBAD database,
1822: operated at CDS, Strasbourg, France, the NASA/IPAC Extragalactic Database (NED)
1823: operated by the Jet Propulsion Laboratory, California Institute of Technology,
1824: under contract with the NASA, and SDSS data.
1825:
1826:
1827:
1828: \begin{thebibliography}{}
1829: \bibitem[\protect\citeauthoryear{Aguirre et al.}{2001}]{aguirre01}
1830: Aguirre A., Hernquist L., Schaye J., Katz N., Weinberg D.~H., Gardner J., 2001, ApJ,
1831: 561, 521
1832:
1833: \bibitem[\protect\citeauthoryear{Asplund, Grevesse, \& Sauval}{2006}]{asplund06}
1834: Asplund M., Grevesse N., Sauval A.~J., 2006, CoAst, 147, 76
1835:
1836: \bibitem[\protect\citeauthoryear{Bland-Hawthorn \& Cohen}{2003}]{bland03}
1837: Bland-Hawthorn J., Cohen M., 2003, ApJ, 582, 246
1838:
1839: \bibitem[\protect\citeauthoryear{Boehringer \& Hartquist}{1987}]{boehringer87}
1840: Boehringer H., Hartquist T.~W., 1987, MNRAS, 228, 915
1841:
1842: \bibitem[\protect\citeauthoryear{Borkowski, Balbus, \& Fristrom}{1990}]{borkowski90}
1843: Borkowski K.~J., Balbus S.~A., Fristrom C.~C., 1990, ApJ, 355, 501
1844:
1845: \bibitem[\protect\citeauthoryear{Bouch{\'e} \& Lowenthal}{2004}]{bouche04}
1846: Bouch{\'e} N., Lowenthal J.~D., 2004, ApJ, 609, 513
1847:
1848: \bibitem[\protect\citeauthoryear{Chen \& Lanzetta}{2003}]{chen03}
1849: Chen H.-W., Lanzetta K.~M., 2003, ApJ, 597, 706
1850:
1851: \bibitem[\protect\citeauthoryear{Collins, Shull, \& Giroux}{2004}]{collins04}
1852: Collins J.~A., Shull J.~M., Giroux M.~L., 2004, ApJ, 605, 216
1853:
1854: \bibitem[\protect\citeauthoryear{Collins, Shull, \& Giroux}{2005}]{collins05}
1855: Collins J.~A., Shull J.~M., Giroux M.~L., 2005, ApJ, 623, 196
1856:
1857: \bibitem[\protect\citeauthoryear{Cooke et al.}{2006a}]{cooke06a}
1858: Cooke J., Wolfe A.~M., Gawiser E., Prochaska J.~X., 2006a, ApJ, 652, 994
1859:
1860: \bibitem[\protect\citeauthoryear{Cooke et al.}{2006b}]{cooke06b}
1861: Cooke J., Wolfe A.~M., Gawiser E., Prochaska J.~X., 2006b, ApJ, 636, L9
1862:
1863: \bibitem[\protect\citeauthoryear{Croton et al.}{2006}]{croton06}
1864: Croton D.~J., et al. 2006, MNRAS, 365, 11
1865:
1866: \bibitem[\protect\citeauthoryear{Dalla Vecchia \& Schaye}{2008}]{dalla08}
1867: Dalla Vecchia C., Schaye J., 2008, MNRAS, submitted, [arXiv:0801.2770]
1868:
1869: \bibitem[\protect\citeauthoryear{Dopita \& Sutherland}{1996}]{dopita96}
1870: Dopita M.~A., Sutherland R.~S., 1996, ApJS, 102, 161
1871:
1872: \bibitem[\protect\citeauthoryear{Ellison et al.}{2007}]{ellison07}
1873: Ellison S.~L., Hennawi J.~F., Martin C.~L., Sommer-Larsen J., 2007, MNRAS, 378, 801
1874:
1875: \bibitem[\protect\citeauthoryear{Fangano, Ferrara, \& Richter}{2007}]{fangano07}
1876: Fangano A.~P.~M., Ferrara A., Richter P., 2007, MNRAS, 381, 469
1877:
1878: \bibitem[\protect\citeauthoryear{Ferland et al.}{1998}]{ferland98}
1879: Ferland, G.~J., Korista, K.~T., Verner, D.~A., Ferguson, J.~W.,
1880: Kingdon, J.~B., \& Verner, E.~M., 1998, PASP, 110, 761
1881:
1882: \bibitem[\protect\citeauthoryear{Fitzpatrick \& Spitzer}{1997}]{fitzpatrick97}
1883: Fitzpatrick, E.L., Spitzer, L. 1997, ApJ, 475, 623
1884:
1885: \bibitem[\protect\citeauthoryear{Fox et al.}{2007a}]{fox07a}
1886: Fox A.~J., Ledoux C., Petitjean P., Srianand R., 2007a, A\&A, 473, 791
1887:
1888: \bibitem[\protect\citeauthoryear{Fox et al.}{2007b}]{fox07}
1889: Fox A.~J., Petitjean P., Ledoux C., Srianand R., 2007b, A\&A, 465, 171
1890:
1891: \bibitem[\protect\citeauthoryear{Fox et al.}{2007c}]{fox07c}
1892: Fox A.~J., Petitjean P., Ledoux C., Srianand R., 2007c, ApJ, 668, L15
1893:
1894: \bibitem[\protect\citeauthoryear{Fox, Savage, \& Wakker}{2006}]{fox06}
1895: Fox A.~J., Savage B.~D., Wakker B.~P., 2006, ApJS, 165, 229
1896:
1897: \bibitem[\protect\citeauthoryear{Gnat \& Sternberg}{2004}]{gnat04}
1898: Gnat O., Sternberg A., 2004, ApJ, 608, 229
1899:
1900: \bibitem[\protect\citeauthoryear{Gnat \& Sternberg}{2007}]{gnat07}
1901: Gnat O., Sternberg A., 2007, ApJS, 168, 213
1902:
1903: \bibitem[\protect\citeauthoryear{Haehnelt, Steinmetz, \& Rauch}{1998}]{haehnelt98}
1904: Haehnelt M.~G., Steinmetz M., Rauch M., 1998, ApJ, 495, 647
1905:
1906: \bibitem[\protect\citeauthoryear{Heckman et al.}{2000}]{heckman00}
1907: Heckman T.~M., Lehnert M.~D., Strickland D.~K., Armus L., 2000, ApJS, 129, 493
1908:
1909: \bibitem[\protect\citeauthoryear{Henry, Edmunds, K{\"o}ppen}{2000}]{henry00}
1910: Henry R.~B.~C., Edmunds M.~G., K{\"o}ppen J., 2000, ApJ, 541, 660
1911:
1912: \bibitem[\protect\citeauthoryear{Henry, Nava, \& Prochaska}{2006}]{henry06}
1913: Henry R.~B.~C., Nava A., Prochaska J.~X., 2006, ApJ, 647, 984
1914:
1915: \bibitem[\protect\citeauthoryear{Henry \& Prochaska}{2007}]{henry07}
1916: Henry R.~B.~C., Prochaska J.~X., 2007, PASP, 119, 962
1917:
1918: \bibitem[\protect\citeauthoryear{Herbert-Fort et al.}{2006}]{herbert06}
1919: Herbert-Fort S., Prochaska J.~X., Dessauges-Zavadsky M., Ellison S.~L.,
1920: Howk J.~C., Wolfe A.~M., Prochter G.~E., 2006, PASP, 118, 1077
1921:
1922: \bibitem[\protect\citeauthoryear{Howk, Sembach, \& Savage}{2003}]{howk03}
1923: Howk J.~C., Sembach K.~R., Savage B.~D., 2003, ApJ, 586, 249
1924:
1925: \bibitem[\protect\citeauthoryear{Howk, Wolfe, \& Prochaska}{2005}]{howk05}
1926: Howk J.~C., Wolfe A.~M., Prochaska J.~X., 2005, ApJ, 622, L81
1927:
1928: \bibitem[\protect\citeauthoryear{Hubeny \& Lanz}{1995}]{hubeny95}
1929: Hubeny I., Lanz T., 1995, ApJ, 439, 875
1930:
1931: \bibitem[\protect\citeauthoryear{Jenkins et al.}{2000}]{jenkins00}
1932: Jenkins E.~B., et al., 2000, ApJ, 538, L81
1933:
1934: \bibitem[\protect\citeauthoryear{Jenkins \& Tripp}{2006}]{jenkins06}
1935: Jenkins E.~B., Tripp T.~M., 2006, ApJ, 637, 548
1936:
1937: \bibitem[\protect\citeauthoryear{Kawata \& Rauch}{2007}]{kawata07}
1938: Kawata D., Rauch M., 2007, ApJ, 663, 38
1939:
1940: \bibitem[\protect\citeauthoryear{Kirkman \& Tytler}{1997}]{kirkman97}
1941: Kirkman D., Tytler D., 1997, ApJ, 484, 672
1942:
1943: \bibitem[\protect\citeauthoryear{Ledoux, Petitjean, \& Srianand}{2003}]{ledoux03}
1944: Ledoux C., Petitjean P., Srianand R., 2003, MNRAS, 346, 209
1945:
1946: \bibitem[\protect\citeauthoryear{Lehner}{2002}]{lehner02}
1947: Lehner N., 2002, ApJ, 578, 126
1948:
1949: \bibitem[\protect\citeauthoryear{Lehner \& Howk}{2007}]{lehner07}
1950: Lehner N., Howk J.~C., 2007, MNRAS, 377, 687
1951:
1952: \bibitem[\protect\citeauthoryear{Lehner et al.}{2003}]{lehner03}
1953: Lehner N., Jenkins E.~B., Gry C., Moos H.~W., Chayer P., Lacour S., 2003,
1954: ApJ, 595, 858
1955:
1956: \bibitem[\protect\citeauthoryear{Lehner, Keenan, \& Sembach}{2001}]{lehner01}
1957: Lehner N., Keenan F.~P., Sembach K.~R., 2001, MNRAS, 323, 904
1958:
1959: \bibitem[\protect\citeauthoryear{Lehner, Wakker, \& Savage}{2004}]{lehner04}
1960: Lehner N., Wakker B.~P., Savage B.~D., 2004, ApJ, 615, 767
1961:
1962: \bibitem[\protect\citeauthoryear{Lu et al.}{1996}]{lu96}
1963: Lu L., Sargent W.~L.~W., Barlow T.~A., Churchill C.~W., Vogt S.~S., 1996,
1964: ApJS, 107, 475
1965:
1966: \bibitem[\protect\citeauthoryear{Lu, Sargent, \& Barlow}{1998}]{lu98}
1967: Lu L., Sargent W. L. W., Barlow T. A., 1998, AJ, 115, 55
1968:
1969: \bibitem[\protect\citeauthoryear{Maller et al.}{2001}]{maller01}
1970: Maller A.~H., Prochaska J.~X., Somerville R.~S., Primack J.~R., 2001,
1971: MNRAS, 326, 1475
1972:
1973: \bibitem[\protect\citeauthoryear{Maller et al.}{2003}]{maller03}
1974: Maller A.~H., Prochaska J.~X., Somerville R.~S., Primack J.~R., 2003,
1975: MNRAS, 343, 268
1976:
1977: \bibitem[\protect\citeauthoryear{Martin}{2006}]{martin06}
1978: Martin C.~L., 2006, ApJ, 647, 222
1979:
1980: \bibitem[\protect\citeauthoryear{Mashchenko, Wadsley, \& Couchman}{2008}]{mashchenko08}
1981: Mashchenko S., Wadsley J., Couchman H.~M.~P., 2008, Sci, 319, 174
1982:
1983: \bibitem[\protect\citeauthoryear{Matteucci}{2001}]{matteucci01}
1984: Matteucci F., 2001, ASSL, 253
1985:
1986: \bibitem[\protect\citeauthoryear{Meiksin}{2008}]{meiksin08}
1987: Meiksin A.~A., 2008, Reviews of Modern Physics, submitted [arXiv:0711.3358]
1988:
1989: \bibitem[\protect\citeauthoryear{Morton}{2003}]{morton03}
1990: Morton, D. C. 2003, ApJS, 149, 205
1991:
1992: \bibitem[\protect\citeauthoryear{Nicastro et al.}{2003}]{nicastro03}
1993: Nicastro F., et al., 2003, Natur, 421, 719
1994:
1995: \bibitem[\protect\citeauthoryear{Nissen et al.}{2007}]{nissen07}
1996: Nissen P.~E., Akerman C., Asplund M., Fabbian D., Kerber F., Kaufl H.~U., Pettini M., 2007, A\&A, 469, 3
1997:
1998: \bibitem[\protect\citeauthoryear{Noterdaeme et al.}{2008}]{noterdaeme08}
1999: Noterdaeme P., Ledoux C., Petitjean P., Srianand R., 2008, A\&A, 481, 327
2000:
2001: \bibitem[\protect\citeauthoryear{Oppenheimer \& Dav{\'e}}{2006}]{oppenheimer06}
2002: Oppenheimer B.~D., Dav{\'e} R., 2006, MNRAS, 373, 1265
2003:
2004: \bibitem[\protect\citeauthoryear{Petitjean, Ledoux, \& Srianand}{2008}]{petitjean08}
2005: Petitjean P., Ledoux C., Srianand R., 2008, A\&A, 480, 349
2006:
2007: \bibitem[\protect\citeauthoryear{Pettini}{1999}]{pettini99}
2008: Pettini M., 1999, cezh.conf, 233
2009:
2010: \bibitem[\protect\citeauthoryear{Pettini et al.}{1999}]{pettini99a}
2011: Pettini M., Ellison S.~L., Steidel C.~C., Bowen D.~V., 1999, ApJ, 510, 576
2012:
2013: \bibitem[\protect\citeauthoryear{Pettini et al.}{2001}]{pettini01}
2014: Pettini M., Shapley A.~E., Steidel C.~C., Cuby J.-G., Dickinson M.,
2015: Moorwood A.~F.~M., Adelberger K.~L., Giavalisco M., 2001, ApJ, 554, 981
2016:
2017: \bibitem[\protect\citeauthoryear{Prochaska et al.}{2002a}]{prochaska02a}
2018: Prochaska J.~X., Henry R.~B.~C., O'Meara J.~M., Tytler D., Wolfe A.~M., Kirkman D.,
2019: Lubin D., Suzuki N., 2002a, PASP, 114, 933
2020:
2021: \bibitem[\protect\citeauthoryear{Prochaska et al.}{2003}]{prochaska03}
2022: Prochaska J.~X., Gawiser E., Wolfe A.~M., Castro S., Djorkovski S.~G., 2003, ApJ, 595, L9
2023:
2024: \bibitem[\protect\citeauthoryear{Prochaska et al.}{2002b}]{prochaska02}
2025: Prochaska J.~X., Howk J.~C., O'Meara J.~M., Tytler D., Wolfe A.~M., Kirkman D.,
2026: Lubin D., Suzuki N., 2002b, ApJ, 571, 693
2027:
2028: \bibitem[\protect\citeauthoryear{Prochaska \& Wolfe}{1997}]{prochaska97}
2029: Prochaska J.~X., Wolfe A.~M., 1997, ApJ, 487, 73
2030:
2031: \bibitem[\protect\citeauthoryear{Prochaska \& Wolfe}{2002}]{prochaska02}
2032: Prochaska J.~X., Wolfe A.~M., 2002, ApJ, 566, 68
2033:
2034: \bibitem[\protect\citeauthoryear{Rauch et al.}{2008}]{rauch08}
2035: Rauch M., et al., 2008, ApJ, in press [arXiv:0711.1354]
2036:
2037: \bibitem[\protect\citeauthoryear{Savage \& Lehner}{2006}]{savage06}
2038: Savage B.~D., Lehner N., 2006, ApJS, 162, 134
2039:
2040: \bibitem[\protect\citeauthoryear{Savage \& Sembach}{1991}]{savage91}
2041: Savage, B. D., Sembach, K. R. 1991, ApJ, 379, 245
2042:
2043: \bibitem[\protect\citeauthoryear{Savage \& Sembach}{1996}]{savage96}
2044: Savage B.~D., Sembach K.~R., 1996, ARA\&A, 34, 279
2045:
2046: \bibitem[\protect\citeauthoryear{Savage, Sembach, \& Howk}{2001}]{savage01}
2047: Savage B.~D., Sembach K.~R., Howk J.~C., 2001, ApJ, 547, 907
2048:
2049: \bibitem[\protect\citeauthoryear{Savage et al.}{2003}]{savage03}
2050: Savage B.~D., et al., 2003, ApJS, 146, 125
2051:
2052: \bibitem[\protect\citeauthoryear{Sembach et al.}{1999}]{sembach99}
2053: Sembach K.~R., Savage B.~D., Lu L., Murphy E.~M., 1999, ApJ, 515, 108
2054:
2055: \bibitem[\protect\citeauthoryear{Sembach et al.}{2003}]{sembach03}
2056: Sembach K.~R., et al., 2003, ApJS, 146, 165
2057:
2058: \bibitem[\protect\citeauthoryear{Shapley et al.}{2003}]{shapley03}
2059: Shapley A.~E., Steidel C.~C., Pettini M., Adelberger K.~L., 2003, ApJ, 588, 65
2060:
2061: \bibitem[\protect\citeauthoryear{Simcoe, Sargent, \& Rauch}{2004}]{simcoe04}
2062: Simcoe R.~A., Sargent W.~L.~W., Rauch M., 2004, ApJ, 606, 92
2063:
2064: \bibitem[\protect\citeauthoryear{Simcoe et al.}{2006}]{simcoe06}
2065: Simcoe R.~A., Sargent W.~L.~W., Rauch M., Becker G., 2006, ApJ, 637, 648
2066:
2067: \bibitem[\protect\citeauthoryear{Slavin, Shull, \& Begelman}{1993}]{slavin93}
2068: Slavin J.~D., Shull J.~M., Begelman M.~C., 1993, ApJ, 407, 83
2069:
2070: \bibitem[\protect\citeauthoryear{Sofia \& Jenkins}{1998}]{sofia98}
2071: Sofia U.~J., Jenkins E.~B., 1998, ApJ, 499, 951
2072:
2073: \bibitem[\protect\citeauthoryear{Spitzer}{1996}]{spitzer96}
2074: Spitzer L.~J., 1996, ApJ, 458, L29
2075:
2076: \bibitem[\protect\citeauthoryear{Sutherland \& Dopita}{1993}]{sutherland93}
2077: Sutherland, R. S., Dopita, M. A. 1993, ApJS, 88, 253
2078:
2079: \bibitem[\protect\citeauthoryear{Tremonti, Moustakas, \& Diamond-Stanic}{2007}]{tremonti07}
2080: Tremonti C.~A., Moustakas J., Diamond-Stanic A.~M., 2007, ApJ, 663, L77
2081:
2082: \bibitem[\protect\citeauthoryear{Tripp et al.}{2008}]{tripp08}
2083: Tripp T.~M., Sembach K.~R., Bowen D.~V., Savage B.~D., Jenkins E.~B.,
2084: Lehner N., Richter P., 2007, ApJS, in press, [arXiv:0706.1214]
2085:
2086: \bibitem[\protect\citeauthoryear{Tripp et al.}{2003}]{tripp03}
2087: Tripp T.~M., et al., 2003, AJ, 125, 3122
2088:
2089: \bibitem[\protect\citeauthoryear{Vladilo et al.}{2001}]{vladilo01}
2090: Vladilo G., Centuri{\'o}n M., Bonifacio P., Howk J.~C., 2001, ApJ, 557, 1007
2091:
2092: \bibitem[\protect\citeauthoryear{Vladilo et al.}{2003}]{vladilo03}
2093: Vladilo G., Centuri{\'o}n M., D'Odorico V., P{\'e}roux C., 2003, A\&A, 402, 487
2094:
2095: \bibitem[\protect\citeauthoryear{Weiner et al.}{2008}]{weiner08}
2096: Weiner B.~J., Coil A.~L., Newman, J.~A., Cooper M.~C., Prochaska, J.~X., Rubin, K.~H.~R.,
2097: 2008, 4th UC Irvine Center for Cosmology Workshop, submitted.
2098:
2099: \bibitem[\protect\citeauthoryear{White \& Frenk}{1991}]{white91}
2100: White S.~D.~M., Frenk C.~S., 1991, ApJ, 379, 52
2101:
2102: \bibitem[\protect\citeauthoryear{Wolfe \& Chen}{2006}]{wolfe06}
2103: Wolfe A.~M., Chen H.-W., 2006, ApJ, 652, 981
2104:
2105: \bibitem[\protect\citeauthoryear{Wolfe, Gawiser, \& Prochaska}{2003}]{wolfe03b}
2106: Wolfe A.~M., Gawiser E., Prochaska J.~X., 2003, ApJ, 593, 235
2107:
2108: \bibitem[\protect\citeauthoryear{Wolfe, Gawiser, \& Prochaska}{2005}]{wolfe05}
2109: Wolfe A.~M., Gawiser E., Prochaska J.~X., 2005, ARA\&A, 43, 861
2110:
2111: \bibitem[\protect\citeauthoryear{Wolfe et al.}{2004}]{wolfe04}
2112: Wolfe A.~M., Howk J.~C., Gawiser E., Prochaska J.~X., Lopez S., 2004, ApJ, 615, 625
2113:
2114: \bibitem[\protect\citeauthoryear{Wolfe \& Prochaska}{2000a}]{wolfe00a}
2115: Wolfe A.~M., Prochaska J.~X., 2000a, ApJ, 545, 591
2116:
2117: \bibitem[\protect\citeauthoryear{Wolfe \& Prochaska}{2000b}]{wolfe00b}
2118: Wolfe A.~M., Prochaska J.~X., 2000b, ApJ, 545, 603
2119:
2120: \bibitem[\protect\citeauthoryear{Wolfe, Prochaska, \& Gawiser}{2003}]{wolfe03}
2121: Wolfe A.~M., Prochaska J.~X., Gawiser E., 2003, ApJ, 593, 215
2122:
2123: \bibitem[\protect\citeauthoryear{Wolfe et al.}{2008}]{wolfe08}
2124: Wolfe A.~M., Prochaska J.~X., Jorgenson R.~A., Rafelski M., 2008, ApJ, in press (arXiv:0802.3914)
2125:
2126: \bibitem[\protect\citeauthoryear{Wolfire et al.}{1995}]{wolfire95}
2127: Wolfire M.~G., McKee C.~F., Hollenbach D., Tielens A.~G.~G.~M., 1995, ApJ, 453,
2128: 673
2129:
2130: \bibitem[\protect\citeauthoryear{Wolfire et al.}{2003}]{wolfire03}
2131: Wolfire M.~G., McKee C.~F., Hollenbach D., Tielens A.~G.~G.~M., 2003, ApJ, 587,
2132: 278
2133:
2134: \bibitem[\protect\citeauthoryear{Yao \& Wang}{2007}]{yao07}
2135: Yao Y., Wang Q.~D., 2007, ApJ, 666, 242
2136:
2137: \bibitem[\protect\citeauthoryear{Zech et al.}{2008}]{zech07}
2138: Zech W.~F., Lehner, N. Howk J.~C., Dixon, W.~V.~D., Brown, T.~M., 2008, ApJ, 679, 460
2139:
2140: \bibitem[\protect\citeauthoryear{Zsarg{\'o} et al.}{2003}]{zsargo03}
2141: Zsarg{\'o} J., Sembach K.~R., Howk J.~C., Savage B.~D., 2003, ApJ, 586, 1019
2142:
2143:
2144: \end{thebibliography}
2145:
2146:
2147: \bsp
2148:
2149: \label{lastpage}
2150:
2151: \end{document}
2152: