1:
2: %% Beginning of file 'sample.tex'
3: %%
4: %% Modified 03 Jan 01
5: %%
6: %% This is a sample manuscript marked up using the
7: %% AASTeX v5.x LaTeX 2e macros.
8:
9: %% The first piece of markup in an AASTeX v5.x document
10: %% is the \documentclass command. LaTeX will ignore
11: %% any data that comes before this command.
12:
13: %% The command below calls the preprint style
14: %% which will produce a one-column, single-spaced document.
15: %% Examples of commands for other substyles follow. Use
16: %% whichever is most appropriate for your purposes.
17:
18: %\documentclass[12pt,preprint]{aastex}
19:
20: %% manuscript produces a one-column, double-spaced document:
21:
22: \documentclass[manuscript]{aastex}
23:
24: %% preprint2 produces a double-column, single-spaced document:
25:
26: %\documentclass[preprint2]{aastex}
27:
28: %% If you want to create your own macros, you can do so
29: %% using \newcommand. Your macros should appear before
30: %% the \begin{document} command.
31: %%
32: %% If you are submitting to a journal that translates manuscripts
33: %% into SGML, you need to follow certain guidelines when preparing
34: %% your macros. See the AASTeX v5.x Author Guide
35: %% for information.
36:
37: %\newcommand{\vdag}{(v)^\dagger}
38: %\newcommand{\myemail}{skywalker@galaxy.far.far.away}
39:
40: %% You can insert a short comment on the title page using the command below.
41:
42: \slugcomment{Accepted for publication in The Astrophysical Journal}
43:
44: %% If you wish, you may supply running head information, although
45: %% this information may be modified by the editorial offices.
46: %% The left head contains a list of authors,
47: %% usually a maximum of three (otherwise use et al.). The right
48: %% head is a modified title of up to roughly 44 characters. Running heads
49: %% will not print in the manuscript style.
50:
51: \shorttitle{The Origin of O{\sc vi} Absorbers}
52: \shortauthors{Heckman et al.}
53:
54: %% This is the end of the preamble. Indicate the beginning of the
55: %% paper itself with \begin{document}.
56:
57: \begin{document}
58:
59: %% LaTeX will automatically break titles if they run longer than
60: %% one line. However, you may use \\ to force a line break if
61: %% you desire.
62:
63: \title{On the Physical Origin of O{\sc vi} Absorption-Line Systems}
64:
65: %% Use \author, \affil, and the \and command to format
66: %% author and affiliation information.
67: %% Note that \email has replaced the old \authoremail command
68: %% from AASTeX v4.0. You can use \email to mark an email address
69: %% anywhere in the paper, not just in the front matter.
70: %% As in the title, you can use \\ to force line breaks.
71:
72: \author{T.M. Heckman\altaffilmark{1}, C.A. Norman\altaffilmark{2},
73: D.K. Strickland\altaffilmark{3},
74: \& K. R. Sembach\altaffilmark{1}}
75: \affil{Department of Physics \& Astronomy, Johns Hopkins University,
76: Baltimore, MD 21218}
77:
78: %% Notice that each of these authors has alternate affiliations, which
79: %% are identified by the \altaffilmark after each name. Specify alternate
80: %% affiliation information with \altaffiltext, with one command per each
81: %% affiliation.
82:
83: \altaffiltext{1}{Guest Investigators on the NASA-CNES-CSA Far
84: Ultraviolet
85: Spectroscopic Explorer. FUSE is operated for NASA by the Johns Hopkins
86: University under NASA contract NAS5-32985.}
87: \altaffiltext{2}{Space Telescope Science Institute, Baltimore, MD 21218}
88: \altaffiltext{3}{Chandra Fellow}
89:
90: %% Mark off your abstract in the ``abstract'' environment. In the manuscript
91: %% style, abstract will output a Received/Accepted line after the
92: %% title and affiliation information. No date will appear since the author
93: %% does not have this information. The dates will be filled in by the
94: %% editorial office after submission.
95:
96: \begin{abstract}
97: We present a unified analysis of the O{\sc vi} absorption-lines
98: seen in the disk and halo of the Milky Way, high velocity clouds,
99: the Magellanic Clouds, starburst galaxies, and the intergalactic medium.
100: We show that these disparate systems
101: define a simple relationship between the O{\sc vi}
102: column density and absorption-line
103: width that is independent
104: of the Oxygen abundance over the range O/H $\sim$ 10\% to twice solar.
105: We show that this relation is exactly that
106: predicted theoretically as a radiatively cooling flow of hot gas passes
107: through the coronal temperature regime - independent of its
108: density or metallicity (for O/H $\gtrsim$
109: 0.1 solar). Since most of the
110: intergalactic O{\sc vi} clouds obey this relation, we infer that they can not
111: have metallicities less than a few percent solar.
112: In order to be able to cool radiatively in less than a Hubble time,
113: the intergalactic clouds must be smaller than $\sim$1 Mpc in size.
114: The implied global
115: heating rate of the warm/hot IGM is consistent with available sources.
116: We show that the cooling column densities for the
117: O{\sc iv}, O{\sc v}, Ne{\sc v}, and Ne{\sc vi} ions are comparable
118: to those seen in O{\sc vi}. This is also true for
119: the Li-like ions Ne{\sc viii}, Mg{\sc x}, and Si{\sc xii} (if
120: the gas is cooling from $T \gtrsim 10^6$ K).
121: All these ions have strong resonance lines in the
122: extreme-ultraviolet spectral range, and would be accessible to $FUSE$
123: at $z \gtrsim$ 0.2 to 0.8.
124: We also show that
125: the Li-like ions can be used
126: to probe radiatively cooling gas at temperatures an order-of-magnitude
127: higher than where their ionic fraction peaks.
128: We calculate that the H-like (He-like)
129: O, Ne, Mg, Si, and S
130: ions have cooling columns of $\sim10^{17}$ (few$\times10^{16}$)
131: cm$^{-2}$. The
132: properties of the O{\sc vii}, O{\sc viii}, and Ne{\sc ix}
133: X-ray absorption-lines
134: towards PKS 2155-304 may be consistent with a scenario of radiatively
135: cooling gas in the Galactic disk or halo.
136: \end{abstract}
137:
138: %% Keywords should appear after the \end{abstract} command. The uncommented
139: %% example has been keyed in ApJ style. See the instructions to authors
140: %% for the journal to which you are submitting your paper to determine
141: %% what keyword punctuation is appropriate.
142:
143: \keywords{intergalactic medium - galaxies: ISM - galaxies: halos
144: - Galaxy: halo - ISM: general}
145:
146:
147: %% From the front matter, we move on to the body of the paper.
148: %% In the first two sections, notice the use of the natbib \citep
149: %% and \citet commands to identify citations. The citations are
150: %% tied to the reference list via symbolic KEYs. The KEY corresponds
151: %% to the KEY in the \bibitem in the reference list below. We have
152: %% chosen the first three characters of the first author's name plus
153: %% the last two numeral of the year of publication as our KEY for
154: %% each reference.
155:
156: \section{Introduction}
157:
158: Diffuse gas in the temperature range of $\sim 10^5$ to $10^6$ K
159: (sometimes called ``coronal gas'')
160: is of considerable interest in many astrophysical
161: contexts. The peak of the radiative cooling curve for plasmas
162: lies at the low temperature
163: end of this range for metallicities greater than a few percent solar
164: (e.g., Sutherland \& Dopita 1993), and since the cooling curve
165: declines with increasing temperature in the coronal regime, this gas will be
166: thermally unstable.
167:
168: In the interstellar medium (ISM)
169: of our own galaxy, coronal gas is created as a consequence
170: of mechanical heating
171: associated with massive stars - supernovae and stellar winds
172: (e.g., Cox \& Smith 1974; Weaver et al. 1977).
173: It is believed that coronal gas traces the interface between hotter
174: (X-ray-emitting) gas and cooler clouds. The rate of mass exchange
175: between the hot and cool phases and the related issue of the
176: fraction of the volume of the disk occupied by the hot gas
177: are crucial components of models of the global structure of the ISM
178: (e.g. McKee \& Ostriker 1977).
179:
180: Since the sound speed in coronal gas exceeds the typical
181: velocity dispersion in the disk of a star-forming galaxy,
182: coronal gas created in the disk can - in the absence of rapid
183: radiative cooling - flow upward to large
184: vertical scale-heights and occupy the lower halo of the galaxy
185: (Shapiro \& Field 1976; Bregman 1980; Norman \& Ikeuchi 1989).
186: On larger scales, since the virial temperature for typical galaxies and
187: small galaxy groups lies
188: near the upper limit of the coronal regime, this gas may be an important
189: component of the large-scale gaseous halos of galaxies and of the
190: intra-group medium (e.g., Norman \& Silk 1979;
191: Verner, Tytler, \& Barthel 1994; Mulchaey et al. 1996).
192:
193: The most significant repository of coronal gas is the intergalactic
194: medium (IGM). Gas and stars inside galaxies
195: account for only 10 to 20\% of the baryonic content of the universe
196: (e.g., Fukugita, Hogan, \& Peebles 1998). Numerical cosmological simulations
197: imply that the bulk of the ``missing baryons'' lies outside galaxies, groups,
198: or clusters (e.g., Cen \& Ostriker 1999; Dav\'e et al. 2001). They also imply
199: that shocks generated by the formation of large-scale structure
200: (and possibly by galactic winds) have heated a substantial fraction
201: of these baryons to temperatures in the coronal regime (the Warm-Hot
202: Intergalactic Medium, or WHIM).
203:
204: The $Far~Ultraviolet~Spectroscopic~Explorer$ mission ($FUSE$ - Moos et al.
205: 2000)
206: has opened the coronal-phase in the ISM and IGM in the low-z universe
207: to direct observation via the
208: O{\sc vi}$\lambda$$\lambda$1032,1038 doublet. Oxygen
209: is the most cosmically-abundant metal. For
210: collisional ionization equilibrium (CIE), O{\sc vi} is a significant species
211: for $T \sim (2- 5)\times 10^5$ K.
212: It is also a byproduct
213: (and hence an important tracer) of hotter radiatively-cooling gas
214: (e.g., Edgar \&
215: Chevalier 1986).
216:
217: To date, $FUSE$ has detected the O{\sc vi}
218: line in absorption in the disk
219: (Jenkins, Bowen, \& Sembach 2001) and halo (Savage et al. 2000)
220: of the Milky Way,
221: in the periphery of high velocity clouds (Sembach et al. 2000), in the
222: interstellar medium of the Large and Small Magellanic Clouds
223: (Hoopes et al. 2002; Howk et al. 2002), in several starburst galaxies
224: (Heckman et al. 2001a,b; Martin et al. 2002), and in the IGM
225: (Oegerle et al. 2000; Sembach et al. 2001; Savage et al. 2002).
226:
227: In each case, the authors have discussed the origin and physical state
228: of the absorbing gas in the particular astrophysical context of
229: their investigation. Our goal in the present paper is to consider
230: the class of O{\sc vi} absorbers more generally. In particular, we will show
231: below that {\it all} of the above classes of O{\sc vi} absorbers obey the
232: simple relationship between the O{\sc vi} column density and line
233: width that is predicted as radiatively cooling hot gas passes
234: through the coronal regime.
235:
236: \section{Data Analysis \& Results}
237:
238: \subsection{Data Analysis}
239:
240: In this investigation we make use of existing O{\sc vi} data in the
241: literature and in a few cases supplement this with additional data that
242: are currently being analyzed. The primary quantities pertinent to our
243: investigation are the column density of O{\sc vi} and the velocity width
244: (line profile second moment) of the O{\sc vi} absorption in different
245: astrophysical regions.
246: We consider data that have been published for the following
247: locations:
248: Galactic disk (Jenkins 1978; $Copernicus$ data), Galactic halo and
249: high velocity clouds (Savage et al. 2000, 2002; Sembach et al. 2000, 2001,
250: 2002; Richter et al. 2001a, 2001b; Heckman et al. 2001a,b), LMC (Howk et al.
251: 2002),
252: SMC (Hoopes et al. 2002), starbursts (Heckman et al. 2001a, 2001b), and
253: the IGM ([PG0953+415: Tripp \& Savage 2000 -- STIS; Savage et al. 2002];
254: [3C\,273: Sembach et al. 2001]; [H\,1821+643: Tripp, Savage, \& Jenkins
255: 2000 -- STIS; Oegerle et al. 2000]).
256:
257: Unless specified otherwise, these results are based
258: upon $FUSE$ data. In most cases, information is available for only the
259: O{\sc vi}$\lambda$1032 line since the 1038\,\AA\ line is blended
260: with other ISM absorption features along the sight lines. Column densities
261: were obtained by integrating the apparent optical depths of the 1032\,\AA\
262: lines, and the line widths were calculated as b$/\sqrt{2} =
263: \sigma = \sqrt{\int{(v-\bar{v})^2 \tau_a(v)dv}/\int{\tau_a(v)dv}}$ over
264: the same velocity ranges used to calculate the column densities.
265:
266: We retrieved the $FUSE$ data for I\,Zw\,18, Mrk~54, and NGC~7673
267: from the $FUSE$ archive to supplement the starburst
268: sample obtained by Heckman et al.
269: (2001a, 2001b). We treated the spectra in a manner similar to that of the
270: data in our previous studies. For I\,Zw\,18, we find $\log N_{OVI} < 13.6~
271: (3\sigma)$ in the $\pm100$ km~s$^{-1}$ velocity range centered on the systemic
272: velocity $v_\odot = +750$ km~s$^{-1}$. This limit is roughly an order of
273: magnitude smaller than the O{\sc vi} columns found for other starbursts.
274: For Mrk~54 and NGC~7673 we find $\log N_{OVI}$ = 14.9 and 15.0 and
275: b = 150 and 180 km~s$^{-1}$ respectively.
276:
277: \subsection{Results \& Possible Selection Effects}
278:
279: Our principal result is shown in Figure 1, where we have plotted the
280: O{\sc vi} column density
281: ($N_{OVI}$) and line widths (b-values) for the various
282: systems listed above.
283:
284: Taken together, the various O{\sc vi} systems define a clear
285: relationship that spans over two-orders-of magnitude in
286: column density and over one order-of-magnitude in line width. The
287: relation is a tight one, with a scatter of only $\sim\pm$0.1 dex in log b
288: at a given value of $\log N_{OVI}$. Above log b $\sim$ 1.6, the
289: column density increases linearly with line width, while at lower
290: line widths the relation rolls over and steepens.
291: Different types of O{\sc vi} systems preferentially
292: populate different regions in parameter space, but with
293: considerable overlap. The Galactic disk,
294: high velocity cloud, and IGM systems tend to
295: have the smallest column densities and line widths,
296: and the starbursts have the largest. The Magellanic Cloud and Milky Way halo
297: systems are intermediate, overlapping well with the low-b, low-column
298: starburst systems.
299: We will discuss the physical interpretation of this correlation in detail
300: in section 3 below. We will show there that the observed correlation
301: is well fit by a simple model {\it with essentially no adjustable parameters.}
302:
303: Before proceeding to this discussion, it is important to establish
304: that the correlation in Figure 1 is not due to an observational selection
305: effect. More specifically, for lines in the linear portion of the
306: curve-of-growth, the ratio of the column density and line width
307: is just proportional to the optical depth of the absorber. Broad lines
308: associated with gas of small column density will correspond to small
309: optical depths. Such shallow lines might elude detection, leading
310: to an absence of detected lines in the lower right corner of Figure 1.
311: The dotted line in the lower right corner of Figure~1 indicates the expected
312: behavior for a line
313: with $N/\sigma = 7.5\times10^{11}$ cm$^{-2}$ km~s$^{-1}$, the approximate
314: $2\sigma$ detection threshold for the $FUSE$ O{\sc vi} $\lambda1032$
315: data considered.
316:
317: We believe this selection effect is not responsible for the correlation
318: in Figure 1 for several reasons. First, it would not account for the absence
319: of points in the upper left corner of Figure 1. Lines produced by
320: gas with a large column density and small line width
321: would be optically-thick and black at line center. These lines
322: would be easy to detect and recognize, if such gas were commonplace.
323:
324: Second, O{\sc vi}
325: absorption is {\it already} seen by $FUSE$ in essentially all the sight lines
326: through the Galactic disk and halo, near high velocity clouds,
327: in the Magellanic Clouds,
328: and in starbursts (except when the relevant wavelength
329: range was contaminated by other strong lines). Thus, only in the case
330: of the intergalactic clouds (whose redshifts are not known
331: {\it a priori}) could selection against weak, broad lines come into play.
332:
333: In fact, larger samples of O{\sc vi}
334: absorption systems in the Galactic halo (Savage et al. 2002) and in
335: high velocity clouds (Sembach et al. 2002) further strengthen the relationship
336: shown in the figure. These larger samples are particularly significant because
337: they are obtained using the light of $\sim100$ quasars and active galactic
338: nuclei as background sources.
339: These objects have very smooth ultraviolet continua,
340: thus minimizing continuum placement uncertainties for any intervening O{\sc vi}
341: absorption features. Numerous broad features exist within those samples, but
342: they typically follow the trend shown in Figure 1.
343: The best example of a broad, weak
344: absorption wing is given by Sembach et al. (2001) for the Galactic O{\sc vi}
345: absorption in the direction of 3C\,273 at velocities
346: $v_{LSR} > 100$ km\,s$^{-1}$.
347: Such absorption wings are relatively uncommon and fall well above the $FUSE$
348: detection threshold indicated in the figure.
349:
350: To further illustrate this, we
351: show in Figure 2 spectra of a broad (log~b = 2.25) redshifted O{\sc vi}
352: synthetic absorption-line added to a typical $FUSE$ spectrum. In the
353: the three cases plotted, the line has a central residual
354: intensity of 85\% (corresponding to the
355: $FUSE$ detection threshold in Figure 1), 28\% (roughly corresponding to the
356: solid line passing through the data points in Figure 1), and
357: 64\% (intermediate between the previous two). The signal-to-noise in the
358: the continuum is representative of the data plotted in Figure 1.
359: It seems clear that the lack of broad, weak O{\sc vi} lines in Figure 1
360: can not be purely due to detectability issues.
361:
362: \section{Discussion}
363:
364: \subsection{The Physics of Radiatively Cooling Gas}
365:
366: Consider the generic situation in which
367: gas cools radiatively from an initial temperature $T$ and density $n$
368: (e.g., Edgar \& Chevalier 1986).
369: The total column density of cooling gas
370: is given by $N_{cool} = \dot{N} t_{cool}$, where
371: $t_{cool}$ is the cooling time and
372: $\dot{N}$ is the rate of cooling per unit area (cm$^{-2}$ s$^{-1}$). The ratio
373: $\dot{N}/n$ has dimensions of velocity - an expression
374: of mass conservation in the cooling gas.
375: Thus, the total cooling column density can be
376: written as $N_{cool} = n t_{cool} v_{cool}$. In the case of radiative cooling
377: of gas with temperature $T$ and metallicity $Z$,
378: $t_{cool} = 3 kT/n \Lambda (T,Z)$, where
379: $\Lambda$$(T,Z)$
380: is the cooling function. Thus $N_{cool} = 3 kT v_{cool}/\Lambda (T,Z)$.
381:
382: The characteristic velocity $v_{cool}$ in the above expression has
383: slightly different physical meanings in different contexts. For example,
384: in the case of gas cooling behind a shock the appropriate velocity
385: is the postshock flow velocity in the O{\sc vi} zone, which is related to
386: the cooling length $l_{cool} \sim v_{cool} t_{cool}$
387: (hence $N_{cool} = n l_{cool}$). In gas collapsing in a cooling instability
388: the characteristic size of the cooling region is $l_{cool} \sim c_s t_{cool}$,
389: where $c_s$ is the isothermal sound speed $(P/\rho)^{1/2}$. Thus,
390: the appropriate velocity to use in this case is the sound speed.
391:
392: The expression $N_{cool} = 3 kT v_{cool}/\Lambda (T,Z)$
393: shows explicitly that $N_{cool}$ is
394: independent of density for radiative cooling.
395: Moreover, at coronal temperatures radiative
396: cooling is dominated by resonance lines due to metals
397: (primarily Oxygen for log $T$ = 5.0 to 5.6,
398: Neon for log $T$ = 5.6 to 5.9, and Iron at higher temperatures). Thus,
399: $N_{cool} \propto \Lambda^{-1}.
400: \propto (Z)^{-1}$.
401: \footnote{The cooling functions computed by Sutherland \& Dopita (1993)
402: show that for coronal gas in CIE,
403: $\Lambda \propto Z^{-1}$ is valid for Oxygen and Neon abundances
404: $Z_{O,Ne} >$ few \% solar. For the
405: non-equilibrium cases they consider, the proportionality holds
406: for $Z_{O,Ne} >$ 10\% solar. Note that their
407: models are parameterized by the Fe abundance, and they
408: assume that [O/Fe] = [Ne/Fe] = +0.5 dex
409: for [Fe/H] $\leq$ -1.0.}
410: Taking the ionic fraction of O{\sc vi} as $f_{OVI}(T)$, we
411: have $N_{OVI} = 3 kT v_{cool} (O/H)_{\odot} Z f_{OVI}/\Lambda$.
412: This column should properly
413: be calculated as a path-integral through the cooling column of gas. However,
414: since $f_{OVI}$ is appreciable over a rather narrow range in temperature
415: centered at $\sim10^{5.5}$ K ($\equiv T_{OVI}$) for CIE, $N_{OVI}$
416: can be approximated by evaluating the above expression for $T = T_{OVI}$.
417: This yields $N_{OVI} = 3 \times 10^{14}$ cm$^{-2}$ for
418: $v_{cool} = 10^2$ km s$^{-1}$.
419: \footnote{We note that the above estimate assumes that the gas is optically
420: thin to its cooling radiation. In CIE at $T = 3 \times 10^5$ K the most
421: important coolants are (in rank order) the \ion{O}{5}$\lambda$629,
422: \ion{O}{5}$\lambda$1218, and \ion{O}{6}$\lambda$$\lambda$1032,1038 lines which
423: together
424: carry 80\% of the total cooling (using the MEKAL plasma code). Based
425: on the observed \ion{O}{6} columns and the implied \ion{O}{5} columns
426: for cooling gas (Table 1), we find that the effective cooling rate would be
427: reduced by
428: 0.0 to 0.2 dex for the systems in Figure 1.}
429:
430: This simple estimate is confirmed by more detailed
431: calculations of radiatively cooling gas.
432: Edgar \& Chevalier (1986) have computed O{\sc vi} columns for the generic
433: case in which hot gas cools from $T = 10^6$ K.
434: They consider isobaric, isochoric, and intermediate cases
435: and find $N_{OVI} \sim 4 \times 10^{14} (v_{cool}/10^2$ km s$^{-1})$ cm$^{-2}$.
436: Calculations of collisionally-ionized gas behind a high-speed
437: radiative shock (Dopita \& Sutherland 1996)
438: give values for $N_{OVI} = (2-3) \times 10^{14}$
439: cm$^{-2}$ for shock speeds ranging from 200 to 500 km s$^{-1}$
440: (postshock flow speeds of $v =$ 50 to 125 km s$^{-1}$).
441: \footnote{For $v_{shock} >$ 300 km s$^{-1}$, higher O{\sc vi}
442: columns can be produced
443: via photoionization of
444: preshock gas by EUV radiation from the shocked gas, if this
445: gas is optically-thick to the radiation.}
446:
447: Begelman \& Fabian (1990) have considered ``turbulent
448: mixing layers'' that arise at the interface between a flow of
449: hot gas past cool gas. The mixing of hot and cool gas creates
450: gas at intermediate temperatures which then cools radiatively. Their model
451: implies $N_{cool} = 5 \eta kT v_t/\Lambda(T)$, where
452: $v_t$ is the turbulent velocity and $\eta$ is a dimensionless efficiency
453: factor which they argue is of order unity.
454: Their equation (4)
455: leads to a predicted O{\sc vi} column of $\sim 3\times 10^{14} \eta$
456: cm$^{-2}$ for a turbulent velocity of 100 km~s$^{-1}$ and a mixing layer
457: temperature of $10^{5.5}$ K (=$T_{OVI}$), where we have
458: used $\Lambda$$(T,Z)$ and $f_{OVI}(T)$
459: for CIE from Sutherland \& Dopita (1993).
460:
461: We can also cast results for conductively heated gas in the same terms.
462: For the case of saturated conduction (Cowie \& McKee 1977), the heat flux
463: is given by $q_{sat} = 5 \phi c_{s} P$
464: where $c_{s}$ is the isothermal
465: sound speed, $P$ is the gas pressure,
466: and $\phi$ is a dimensionless
467: constant of order unity. If we balance this heat flux
468: by radiative cooling we get $q_{sat} = n^2 \Lambda D$,
469: where $D$ is the thickness
470: of the conductively-heated interface.
471: Combining these relations, we have
472: $N_{cool} = nD = 5 \phi c_s P/n \Lambda = 10 \phi c_s kT/\Lambda$. Apart from
473: constants of order unity, we recover the same expression as that
474: above for $N_{cool}$ (where $v_{cool} \sim c_s$).
475: \footnote {In the case of classical (unsaturated) conduction,
476: similar arguments give $N_{cool} = [\kappa(T) T/\Lambda(T,Z)]^{1/2}$, where
477: $\kappa$ is the thermal conductivity (Cowie \& McKee 1977). While $N_{cool}$
478: is still independent of $n$, $N_{OVI}$
479: is no longer independent of metallicity, but will scale as $Z^{1/2}$.
480: For solar metallicity
481: and CIE at $T = 10^{5.5} K$
482: this predicts $N_{OVI} \sim 3 \times 10^{13}$ cm$^{-2}$.}
483:
484: A robust conclusion emerges from the above.
485: {\it The characteristic O{\sc vi} column density in radiatively
486: cooling coronal gas is essentially independent of density, metallicity,
487: and the heating mechanism. It
488: depends only on the value for $v_{cool}$, the characteristic cooling
489: flow speed.}
490:
491: \subsection{Comparison to the Data}
492:
493: In order to compare the predictions of the model described above
494: to the results in Figure 1 we need to relate the observed line
495: width $b_{obs}$ to the cooling flow velocity $v_{cool}$.
496:
497: In the absence of a flow, the O{\sc vi} lines will have an irreducable
498: intrinsic width due purely to thermal broadening: $b_{therm} = (2kT/m_O)^{1/2}
499: = 10 (32)$ km s$^{-1}$ for $T = 10^5~(10^6)$ K.
500: The observed line
501: broadening is due to the combination of the thermal broadening
502: and that produced by the cooling
503: flow: b$_{obs} = \sqrt{b_{thermal}^2+b_{cool}^2}$.
504: We identify $b_{cool}$ with the cooling flow velocity $v_{cool}$ that sets
505: the O{\sc vi} column density. We have therefore computed the relation
506: between $N_{OVI}$ and b$_{obs}$ expected
507: for radiatively cooling gas with
508: b$_{thermal}$ corresponding to $T_{OVI} = 10^5$ and $10^6$ K
509: (chosen as representative bounds for the collisionally ionized gas).
510:
511: The fit of this relationship to the data is excellent. It is important
512: to emphasize that this model has essentially no tunable parameters
513: (apart from $T_{OVI}$, whose range of allowed values is robustly set by
514: simple physics). In particular, this model naturally explains the linear
515: proportionality between b$_{obs}$ and $N_{OVI}$ for large b$_{obs}$
516: (b$_{obs} >>$ b$_{therm}$), the value of the ratio $N_{OVI}$/b$_{obs}$
517: in the ``linear regime'',
518: and the ``turnover'' in the relationship
519: at small b$_{obs}$ (when b$_{obs} \gtrsim $ b$_{therm}$).
520: \footnote
521: {A multiplicity of velocity components at different velocities may also
522: affect the observed line widths. If identical components have the same
523: central velocity, they would follow the linear relationship shown in the
524: figure (neglecting thermal broadening). If, however, the components are
525: separated by an amount $\Delta$$v$, there will be an additional broadening
526: that will cause deviations from this linear relationship. We have shown
527: the curves for a set of three identical components having b = 17 km~s$^{-1}$
528: and symmetrical separations of
529: $\Delta$$v$ = $\pm$ 0, 10, 20, and 30 km~s$^{-1}$ about the central component
530: (thin solid lines). Larger values of $\Delta$$v$ move points further to the
531: right in this plot.}
532:
533: A crucial prediction of the radiatively cooling scenario
534: is that the relationship
535: between O{\sc vi} column density and line width should be independent
536: of the metallicity of the cooling gas, provided that the metallicity
537: is greater than $\sim$ 0.1 solar. This prediction is directly confirmed
538: by Figure 1, which includes O{\sc vi} systems that span a metallicity range from
539: $\sim$10\% solar (SMC sightlines and the starburst NGC1705) through
540: $\sim$solar (Milky Way disk and halo and the starburst NGC3310) to
541: several times solar (the M83 starburst nucleus).
542:
543: Our only probe of the
544: regime where metallicity effects should become significant is
545: the extremely metal-poor starburst I~Zw~18 (O/H = 2\%).
546: Interestingly, we have only
547: an upper limit to the O{\sc vi} absorption in this case:
548: $N_{OVI}$/b$_{obs} \leq
549: 5 \times 10^{11}$ cm$^{-2}$ per km s$^{-1}$. This is about an
550: order-of-magnitude below the average ratio seen for the other starbursts
551: in Figure 1.
552: Since it is likely that the metallicity
553: in the bulk of the ISM of I~Zw~18 is even lower than the 2\% solar
554: value in the HII regions (e.g. Thuan et al. 2002), the absence of detectable
555: O{\sc vi} absorption in this galaxy is consistent with our model.
556:
557: The correlation in Figure 1 is clearly a sequence along which the dynamical
558: state of the O{\sc vi} absorbers changes significantly.
559: Dividing the measured b-values by the speed-of-sound at $T_{OVI}$
560: allows us to roughly characterize the
561: Mach number of the flow ($\cal M$). This provides a simple
562: (but physically instructive) summary of the observational data.
563: We see that the Galactic disk, high velocity cloud, and intergalactic O{\sc vi}
564: systems are quiescent, highly subsonic flows (possibly
565: arising in conductive interfaces
566: between hotter and cooler gas). The Milky Way halo, Magellanic Cloud,
567: and several of the starburst systems appear to be trans-sonic flows
568: ($\cal M \sim$ 1). This is consistent with models for
569: fountain/chimney flows
570: (e.g. Shapiro \& Field
571: 1976; Norman \& Ikeuchi 1989), or dynamically
572: young galactic outflows in the blow-out phase (e.g. Heckman et al. 2001a).
573: Finally, the most extreme starburst systems (those in NGC 3310, NGC 7673,
574: and Mrk 54) show supersonic flows with Mach numbers of several. In these
575: cases the O{\sc vi} absorption is significantly blueshifted relative to the
576: galaxy systemic velocity, and we are probably observing
577: well-developed galactic winds (e.g. Heckman 2001).
578:
579: \section{Implications}
580:
581: \subsection{The Nature of the Intergalactic O{\sc vi} Clouds}
582:
583: We have argued above that the striking correlation in Figure 1 arises naturally
584: from the simple physics of radiatively cooling hot gas.
585: In principle O{\sc vi} can also be produced by
586: the photoionization of relatively cool gas by a suitably hard ionizing source.
587: The only physically plausible such source in the present context is
588: the metagalactic ionizing background due to QSOs. This mechanism can
589: be ruled out for most of the classes of O{\sc vi} absorbers in
590: Figure 1 (e.g., Sembach et al. 2000; Savage et al. 2000) since
591: the feeble intensity of the metagalactic background implies that implausibly
592: large and low-density clouds are required to produce the observed O{\sc vi}
593: columns. A local photoionizing source is implausible in these cases, since
594: normal stars produce
595: insignificant amounts EUV radiation with $h\nu \geq$ 114 eV
596: ($\chi$$_{OVI}$).
597:
598: In principle, either collisional ionization or
599: photoionization by the metagalactic background
600: could produce a significant
601: fraction of the {\it intergalactic} O{\sc vi} clouds (Tripp \& Savage
602: 2000; Tripp et al. 2001;
603: Cen et al. 2001). We would argue that
604: Figure 1 strongly suggests that most intergalactic O{\sc vi} clouds
605: (like the other
606: O{\sc vi} absorption-line systems) are collisionally-ionized,
607: radiatively-cooling gas. In fact, six out of the eight IGM clouds
608: in Figure 1 lie within the same range in parameter space as the
609: high velocity cloud and Milky Way disk systems.
610: {\it There is no plausible reason why photoionization should naturally
611: place the intergalactic clouds directly on top of the correlation
612: between O{\sc vi} column density and line width defined by collisionally
613: ionized coronal gas in the other systems.}
614: \footnote{We note that two of the IGM points do lie well above the trend of the
615: rest of the points shown in Figure~1. These are the $z=0.06807$ and
616: $z=0.14232$ absorbers toward PG\,0953+415. The absorption produced by
617: \ion{O}{6} and other ions in both absorbers can be described
618: self-consistently by photoionization models of dilute gas if absorption
619: arises over path lengths of $\sim100$ kpc (see Savage et al. 2002).}
620:
621: The total ionized gas column density in a photoionized cloud
622: depends only on the ionization parameter ($U$, the ratio of photon
623: to electron density) and the cloud's optical depth in the Lyman continuum:
624: $N_{+} \sim (1 - e^{-\tau})Uc/\alpha$, where $\alpha$ is the recombination
625: coefficient. The predicted O{\sc vi} column is just
626: $N_{OVI} = 8.5\times10^{-4} Z f_{OVI} N_{+}$. In the intergalactic clouds
627: the relative strength of the O{\sc vi} line requires $U \gtrsim 10^{-1}$ and
628: $f_{OVI}$ to be near its maximum value (of order $10^{-1}$)
629: (Tripp \& Savage 2000). The predicted O{\sc vi} column
630: is $N_{OVI} \sim 10^{18} Z (1 - e^{-\tau})$ cm$^{-2}$ for
631: $U \sim f_{OVI} \sim 0.1$. Thus, photoionization
632: does not lead to any ``natural'' value
633: for $N_{OVI}$, nor does it imply any simple proportionality between $N_{OVI}$
634: and line width (Figure 1).
635:
636: For our hypothesis of radiative cooling of hot gas to be correct,
637: it follows first that the intergalactic O{\sc vi}
638: clouds must have metallicities greater than a few \% solar. This
639: is consistent with theoretical models for the WHIM in
640: which it has been chemically-enriched
641: by the outflow of metals from galaxies
642: (Aguirre et al. 2001; Cen et al. 2001).
643:
644: Our hypothesis also requires the clouds to have
645: radiative cooling times less than the Hubble time. This leads directly to
646: a constraint on the sizes of the O{\sc vi} clouds.
647: The cooling time for gas in CIE
648: at $T_{OVI} = 10^{5.5}$ K is
649: $t_{cool} = 4.7 \times 10^{3} n^{-1} Z^{-1}$ years (Sutherland
650: \& Dopita 1993). The O{\sc vi}
651: column for region with a size $D_{OVI}$ is given by
652: $N_{OVI} \sim 8.5\times10^{-4} Z n D_{OVI} f_{OVI}$. Combining these
653: relations and evaluating $f_{OVI}$ at $T_{OVI}$, we have $t_{cool} =
654: 2 \times 10^{10} (D_{OVI}/$Mpc) years. Thus, independent of the metallicity,
655: the requirement $t_{cool} \leq t_{Hubble}$ requires $D_{OVI} \lesssim$ 1~Mpc.
656: We note parenthetically that the ratio of the radiative cooling
657: and sound-crossing time will be of-order unity in these clouds, so
658: the cooling would be intermediate between the isobaric and isochoric cases
659: considered by Edgar \& Chevalier (1986) and in Table 1.
660:
661: The cooling constraint can also be cast in terms of density.
662: For $H_0$ = 70 km s$^{-1}$ Mpc$^{-1}$ and $\Omega_B$ = 0.04,
663: $t_{cool} \leq t_{Hubble}$ requires that the O{\sc vi} clouds correspond
664: to mean overdensities $\rho/\bar{\rho} \geq 1.3 Z^{-1}$. Thus, for
665: $Z$ = 0.03 to 0.3 in the O{\sc vi} clouds (Aguirre et al. 2001; Cen et al.
666: 2001)
667: only moderate overdensities are needed. This agrees with the computations
668: reported by Cen et al. (2001) in which the bulk of the
669: hot O{\sc vi} absorbing clouds have overdensities of 20 to 40.
670:
671: If the O{\sc vi} clouds are cooling on cosmologically-short timescales,
672: they require a suitable heating source.
673: In the numerical simulations
674: by Cen et al. (2001) and Dav\'e et al. (2001) this heating
675: is provided by shocks and compression associated
676: with the formation of large-scale structure.
677: At least on the basis of simple energetic considerations,
678: supernova-driven galactic winds could also heat this material.
679: Tripp, Savage, \& Jenkins (2000) have estimated that the global mass
680: density of intergalactic O{\sc vi} clouds is
681: $\Omega_{OVI} = 4\times10^{-4} Z^{-1}$ or
682: $\rho_{OVI} = 6\times10^7 Z^{-1} M_{\odot}$ Mpc$^{-3}$. For a temperature
683: $T_{OVI} = 10^{5.5}$ K, the corresponding thermal energy density
684: is $1.4\times10^{55} Z^{-1}$ ergs Mpc$^{-3}$. We can compare that
685: to the time-integrated energy density due to supernovae. The galaxy
686: luminosity density measured by Blanton et al. (2001) implies
687: a mean stellar mass density of $6\times10^8 M_{\odot}$ Mpc$^{-3}$
688: (Fukugita, Hogan, \& Peebles 1999). Taking a yield of one supernova
689: per $10^2 M_{\odot}$ implies that supernovae have supplied
690: $6\times10^{57}$ ergs Mpc$^{-3}$. For $Z$ = 0.03 to 0.3, the
691: total supernova energy is $\sim$ one to two orders-of-magnitude
692: larger than the thermal
693: energy of the O{\sc vi} clouds. Cen et al.(2001) estimate that
694: the O{\sc vi} clouds comprise of order 10\% of the WHIM. Given the apparently
695: high efficiency with which starburst-driven winds transport supernova
696: kinetic energy (Heckman 2001), we conclude they could be
697: a significant heating source in the intergalactic O{\sc vi} clouds.
698:
699: \subsection{Predictions for other Coronal Ions}
700:
701: Our hypothesis that the observed O{\sc vi} originates in radiatively cooling
702: hot gas makes an interesting prediction about the detection
703: of both the hotter gas that is the precursor to the O{\sc vi}
704: and the cooler gas that is its byproduct. The columns in
705: the hotter gas are substantial, essentially because the cooling
706: times are long. We have $N_{cool} = 3 kT v_{cool}/\Lambda (T)$. Assuming that
707: $v_{cool} \propto c_s$, and approximating $\Lambda (T) \propto T^{-0.8}$ from
708: $\log T =$ 5.5 to 7.0 (Sutherland \& Dopita 1993), yields $N_{cool} \propto
709: T^{2.3}$.
710:
711: Following Edgar \& Chevalier (1986), we have used the cooling functions in
712: Sutherland \& Dopita (1993) to explicitly calculate the columns
713: for the most important ionic species that peak in fractional
714: abundance at $\log~T =$ 5.0 to 6.5 for
715: gas cooling radiatively
716: from an initial temperature $T = 5 \times 10^6$ K. We have considered
717: gas that cools at constant pressure (isobaric) and constant density
718: (isochoric). We further assume that radiative cooling by dust grains is not
719: significant (e.g. Draine 1981), that the
720: gas is optically thin to its cooling radiation, and that the relative
721: abundances of all the heavy elements have solar values. The results are
722: listed in Table 1.
723: \footnote{In this regime, greater care in discussing the metallicity dependence
724: of the cooling columns is required.
725: For the temperature range where the O{\sc iv}, O{\sc v}, and
726: O{\sc vi} ions are most abundant, Oxygen is the most important coolant.
727: Similarly, at temperatures where the Ne{\sc viii} ion is most abundant,
728: Neon is the most important coolant (Sutherland \& Dopita 1993).
729: Thus the relevant cooling time scales inversely with the O or Ne abundance,
730: implying that the ionic cooling columns are independent of the
731: abundance.
732: In contrast, ions like
733: Mg{\sc x}, Si{\sc xii}, O{\sc viii}, Ne{\sc ix}, and Ne{\sc x} reach peak
734: abundances at temperatures
735: where Fe is the most important coolant. Thus our estimates in Table 1
736: for cooling columns need to be multiplied by the ratio of the
737: actual elemental adundance relative to solar.
738: In the IGM, alpha-element/Fe abundance ratios
739: of up to twice solar are plausible (e.g.
740: Gibson, Loewenstein, \& Mushotzky 1997).}
741:
742: First, consider the Li-like species
743: Ne{\sc viii}, Mg{\sc x}, and Si{\sc xii}
744: which have strong ultraviolet
745: resonance transitions analogous to the O{\sc vi} doublet
746: (Ne{\sc viii}$\lambda$$\lambda$770,780;
747: Mg{\sc x}$\lambda$$\lambda$ 610,625;
748: Si{\sc xii}$\lambda$$\lambda$499,521). These transitions could be observed
749: with $FUSE$ in intergalactic clouds with suitable (moderate) redshifts,
750: and trace ions that reach their peak
751: abundances at $\log~T$ = 5.85 (Ne{\sc viii}) to 6.35 (Si{\sc xii}).
752: In both the isobaric and isochoric calculations, these ions all have
753: cooling columns comparable to O{\sc vi}. Since these transitions
754: have similar oscillator strengths, they will have similar
755: optical depths and equivalent widths,
756: and should be only slightly more difficult
757: to detect than the O{\sc vi} line. See Table 1 for details.
758: Note that the cooling columns for the Li-like species that trace cooler gas
759: (C{\sc iv} and N{\sc v})
760: are 10 to 50 times smaller than O{\sc vi}.
761:
762: Next, consider the other abundant ionic species with strong resonance
763: lines in the EUV band. Notable examples are O{\sc iv}$\lambda$788,
764: O{\sc v}$\lambda$630, Ne{\sc v}$\lambda$568, and
765: Ne{\sc vi}$\lambda$559. Again Table 1 shows that these species
766: have cooling columns comparable to O{\sc vi} (particularly for the
767: ioschoric models) and the implied
768: optical depths and equivalent widths are comparable
769: to the O{\sc vi}$\lambda$1032 line.
770:
771: Finally, consider the Hydrogen-like and Helium-like species of
772: Oxygen and Neon. The cooling columns are very high ($\sim 10^{17}$
773: cm$^{-2}$) because of the steep temperature dependence noted above.
774: These ions have their resonance lines
775: at high energies, and
776: X-ray absorption studies are required to observe them.
777:
778: To provide a graphical summary for a much wider array of
779: astrophyically-significant ions, we show in Figure 3 the cooling columns
780: estimated using the approximation
781: $N_{X(i)} = 3 k T_{X(i)}$ $\cal M$ $c_s (X/H)_{\odot} Z f_i /\Lambda$. This was
782: evaluated
783: at the temperature where ionic species $i$ of element $X$ reaches
784: it peak abundance ($T_{X(i)}$), assuming $v_{cool} = c_s$
785: at $T_{X(i)}$ (i.e. $\cal M$ = 1).
786: The values plotted are approximate
787: only, and are superceded by the detailed calculations described above
788: for the specific subset of ions in Table 1. The strong temperature
789: dependence of the cooling column is readily apparent in this figure.
790:
791: \subsection{Application to the PKS 2155-304 Absorber}
792:
793: Recent $Chandra$ and $FUSE$ data on the sight-line to the AGN
794: PKS 2155-304 enable us to directly test the above predictions.
795: Nicastro et al. (2002) have used the Low-Energy Transmission
796: Grating Spectrograph on $Chandra$ to detect
797: X-ray absorption-lines due to H-like and He-like Oxygen and He-like Neon,
798: and they have compared these data to the $FUSE$ detection of
799: O{\sc vi} absorption (Savage et al. 2000; Sembach et al. 2000).
800:
801: The $FUSE$ data have high enough spectral resolution to show that
802: the O{\sc vi} absorption arises in two components (a narrower
803: component centered at $v_{LSR}$ = +80 km s$^{-1}$ and a broader component
804: centered on $v_{LSR}$ = -92 km s$^{-1}$).
805: Nicastro et al. show that the relative column densities
806: of the O{\sc vi}, O{\sc vii}, O{\sc viii}, and Ne{\sc ix}
807: ions are inconsistent with single-temperature gas in collisional
808: ionization equilibrium. They find that a model of highly dilute gas
809: photoionized by the metagalactic background provides a satisfactory
810: fit to the observed columns provided that the Ne/O abundance ratio
811: is 2 to 3 times solar. This model requires that the absorber have a very low
812: density ($\sim 10^{-5}$ cm$^{-3}$) and correspondingly large size
813: (several Mpc). Thus, even though its small radial velocity suggests
814: that the absorber is part of the disk or halo of the Milky Way, its
815: immense implied size leads Nicastro et al. (2002) to propose that
816: it is a local intergalactic filament (part of the ``Warm-Hot Ionized
817: Medium'').
818:
819: Here we propose an alternative interpretation. We suggest that the absorber
820: arises in gas that is cooling radiatively from an
821: initial temperature high enough to provide O{\sc viii} absorption, but
822: not Ne{\sc x} (e.g. $2 \times 10^6 K\leq T \leq 5 \times 10^6 K$).
823: This model has an important physical difference from the collisional
824: ionization equilibrium models ruled out by Nicastro et al.: while
825: they considered gas at a single temperature, the radiatively cooling
826: model contains gas that spans a broad temperature range.
827:
828: We can use the calculations described in the previous section to compare
829: to the Nicastro et al. data.
830: The results are listed in Table 2. There is a rough agreement, but the
831: models predict larger columns in the O{\sc vii}, O{\sc viii}, and Ne{\sc ix}
832: ions than derived from the X-ray data. Optical depth effects may play a role.
833: For cooling gas, the predicted optical depths in these X-ray lines are
834: roughly 3 to 10 times higher than in the O{\sc vi}$\lambda$1032 line
835: (see Table 1).
836: If this is the case, then the X-ray lines in PKS 2155-304 could be
837: moderately optically-thick, and the
838: column densities reported in Nicastro et al. (2002) should then
839: be taken as lower limits. This would improve the agreement between
840: the models and the data. In fact,
841: the curve-of-growth analysis reported in Nicastro et al.
842: shows that optical depth effects will be non-negligible in the
843: O{\sc vii} and Ne{\sc ix} lines if the Doppler
844: b-parameter is similar to that of the
845: narrow or broad O{\sc vi} absorbers (65 or 95 km s$^{-1}$
846: respectively).
847:
848: If our model is correct, it has an attractive feature. Unlike the
849: photoionization model, it does not require
850: the absorber to have a very low
851: density and very large size-scale. This gas could well
852: reside in the disk or halo of the Milky Way (as might be expected
853: based on its small LSR velocity).
854:
855: \section{Conclusions}
856:
857: Coronal-phase ($T \sim 10^5$ to $10^6$ K) gas is important in many
858: astrophysical contexts, but has been difficult to observe. The
859: O{\sc vi}$\lambda$$\lambda$1032,1038 doublet is a direct probe of this phase,
860: and has been recently detected in absorption by $FUSE$
861: in the disk and
862: halo of the Galaxy, high velocity clouds, the Magellanic Clouds, starburst
863: galaxies, and the IGM.
864:
865: In this paper, we have presented a unified analysis of all these data.
866: We have shown that these disparate systems
867: define a simple, relatively tight relationship between the O{\sc vi}
868: column density ($N_{OVI}$) and absorption-line
869: width (b). The relationship is linear for broad lines (b $\geq$ 40 km s$^{-1}$)
870: but rolls over and steepens for narrower lines. The different
871: classes of O{\sc vi} systems tend to populate distinct regions of
872: the relationship, with Galactic disk clouds and starburst outflows
873: at the extrema. We have shown that the relationship is not due to
874: selection effects.
875: We have also demonstrated it is independent
876: of the Oxygen abundance over the range O/H = one-tenth to twice solar
877: but breaks down for the extremely metal-poor starburst galaxy
878: I~Zw~18 (O/H $\lesssim$ 2\% solar).
879:
880: We have argued that the observed relation between O{\sc vi} column density
881: and line-width is exactly that
882: predicted theoretically as radiatively cooling hot gas passes
883: through the coronal temperature regime. The predicted column density
884: is independent of the gas
885: density or metallicity (as long as the latter is $\gtrsim$
886: 0.1 solar, so that the radiative cooling is dominated by Oxygen).
887: The O{\sc vi} column density depends only the characteristic flow
888: speed in the cooling gas (e.g. Edgar \& Chevalier 1986).
889: {\it This model (which has essentially no adjustable parameters)
890: naturally produces both the form and absolute
891: normalization of the observed column-density vs. line-width relation.}
892: It unifies systems spanning a broad range in size and dynamical state
893: ranging from very small (very large) quiescent clouds in the disk of the
894: Milky Way
895: (IGM), to meso-scale transonic flows in the halo of the Milky Way and Magellanic
896: Clouds, to supersonic global winds in some starbursts.
897:
898: We infer that the intergalactic O{\sc vi} clouds can not have
899: metallicities less than a few percent solar, or they would lie
900: too far off the $N_{OVI}$ {\it vs.} b relation.
901: In order to be able to cool radiatively in less than a Hubble time,
902: the clouds must be smaller than $\sim$1 Mpc in size.
903: We have shown that if the O{\sc vi} clouds are collisionally-ionized,
904: the implied global
905: heating rate is consistent with available sources
906: (structure formation and/or galactic winds).
907: We have also predicted that column densities
908: comparable to those seen in O{\sc vi} ($\sim10^{14}$
909: cm$^{-2}$) will be
910: detected
911: in the other Li-like ions Ne{\sc viii}, Mg{\sc x}, and Si{\sc xii}
912: (if the intergalactic gas is cooling from $T \gtrsim 10^6$ K),
913: and in the O{\sc iv}, O{\sc v}, Ne{\sc v}, and Ne{\sc vi} ions as well.
914:
915: More generally, we have shown that the steep increase in radiative
916: cooling time with temperature means that the Li-like ions can be used
917: to probe radiatively cooling gas at temperatures up to an order-of-magnitude
918: higher than the value where the ionic fraction peaks (e.g. at up to
919: a few $\times 10^6$ K for O{\sc vi}).
920: We have also calculated radiative
921: cooling column densities for all abundant ionic species whose peak
922: fractional abundance occurs at $T \geq 10^5$ K. The H-like (He-like)
923: O, Ne, Mg, Si, and S
924: ions have characteristic cooling columns of $\sim10^{17}$ (few$\times10^{16}$)
925: cm$^{-2}$, but must be observed with X-ray telescopes.
926:
927: Finally, we have shown that the
928: properties of the O{\sc vii}, O{\sc viii}, and Ne{\sc ix} absorption-lines
929: recently detected with $Chandra$ (Nicastro et al. 2002)
930: may be consistent with a scenario of radiatively
931: cooling gas in the Galactic disk or halo.
932:
933: %% If you wish to include an acknowledgments section in your paper,
934: %% separate it off from the body of the text using the \acknowledgments
935: %% command.
936:
937: %% Included in this acknowledgments section are examples of the
938: %% AASTeX hypertext markup commands. Use \url without the optional [HREF]
939: %% argument when you want to print the url directly in the text. Otherwise,
940: %% use either \url or \anchor, with the HREF as the first argument and the
941: %% text to be printed in the second.
942:
943: \acknowledgments
944:
945: The authors thank Robert Benjamin, Ed Jenkins, and Blair Savage for
946: useful suggestions. We also thank Alessandra Aloisi, Charles Hoopes,
947: Chris Howk, and Crystal Martin for their assistance in the analysis of some
948: of the $FUSE$ data shown in this paper. We thank an anonymous referee
949: for several suggestions that improved the paper.
950: This work was supported in part by NASA FUSE GI grants NAG5-9012
951: and NAG5-10302.
952: TMH and KRS also
953: acknowledge support from NASA Long-Term Space Astrophysics grants
954: NAG5-6400 and NAG5-3485 respectively. DKS is
955: supported by NASA through Chandra Postdoctoral Fellowship Award PF0-10012,
956: issued by the Chandra X-ray Observatory Center, which is operated by the
957: Smithsonian Astrophysical Observatory for and on behalf of NASA under
958: contract NAS8-39073.
959:
960: %% Appendix material should be preceded with a single \appendix command.
961: %% There should be a \section command for each appendix. Mark appendix
962: %% subsections with the same markup you use in the main body of the paper.
963:
964: %% Each Appendix (indicated with \section) will be lettered A, B, C, etc.
965: %% The equation counter will reset when it encounters the \appendix
966: %% command and will number appendix equations (A1), (A2), etc.
967:
968: \appendix
969:
970: \section{The Temperature Dependence of Cooling Column for Li-like Ions}
971:
972: In Figure 4 we show the fractional abundance of O{\sc vi} ions
973: as a function of temperature,
974: assuming collisional ionization equilibrium (taken from Sutherland
975: \& Dopita 1993). The asymmetric ``tail'' on the high-temperature
976: side of the peak at $logT \sim$ 5.5 is a consequence of the broad temperature
977: range over which O{\sc vii} (from which O{\sc vi} is produced
978: by recombination) is the dominant ion.
979:
980: In Figure 5 we plot the temperature dependence of the value of the
981: O{\sc vi} cooling column density
982: predicted from $N_{OVI} = (3 kT {\cal M} c_s/\Lambda(T,Z)
983: f_{OVI}(T) O/H$ (taking the flow's Mach number $\cal M$ = 1). The maximum
984: value of the column density provides a reasonable approximation
985: to the total column density derived from a more
986: detailed calculation,
987: where the cooling of a parcel of gas is followed in detail as a
988: function of time.
989:
990: The secondary peak at $logT \approx$ 6.4, is a consequence of the
991: O{\sc vi} fractional abundance remaining significant up to this temperature
992: (Figure 4).
993: The total gas cooling column scales with the cooling time (which increases
994: as $\sim T^{1.8}$) and with the flow speed (which increases
995: with $T^{0.5}$ for constant $\cal M$). These effects compensate for the
996: decrease in the
997: ionic fraction of O{\sc vi} with increasing temperature.
998:
999: Figure 5 therefore shows that for appropriately high initial temperatures,
1000: a significant fraction
1001: of the O{\sc vi} column density in radiatively cooling gas
1002: can arise in gas at temperatures up to an order-of-magnitude higher
1003: than simple consideration of collisional ionization equilibrium would suggest.
1004: For the case of a flow with constant Mach number, the fractional contribution
1005: to $N_{cool}$ from gas hotter than $10^6$ K is about 45\%.
1006:
1007: Note that above $log~T \sim$ 6.0 the primary coolant is Fe, and our
1008: calculations assume a solar O/Fe ratio. For the more general case,
1009: the predicted cooling columns at $T \geq 10^6$ K should be multiplied by
1010: the actual O/Fe relative to the solar value.
1011:
1012: Similar arguments apply to the other Li-like species discussed above
1013: (e.g. the Mg{\sc x} ion could be used to probe radiatively cooling gas
1014: at temperatures up to almost $10^7$ K).
1015:
1016: %% The reference list follows the main body and any appendices.
1017: %% Use LaTeX's thebibliography environment to mark up your reference list.
1018: %% Note \begin{thebibliography} is followed by an empty set of
1019: %% curly braces. If you forget this, LaTeX will generate the error
1020: %% "Perhaps a missing \item?".
1021: %%
1022: %% thebibliography produces citations in the text using \bibitem-\cite
1023: %% cross-referencing. Each reference is preceded by a
1024: %% \bibitem command that defines in curly braces the KEY that corresponds
1025: %% to the KEY in the \cite commands (see the first section above).
1026: %% Make sure that you provide a unique KEY for every \bibitem or else the
1027: %% paper will not LaTeX. The square brackets should contain
1028: %% the citation text that LaTeX will insert in
1029: %% place of the \cite commands.
1030:
1031: %% We have used macros to produce journal name abbreviations.
1032: %% AASTeX provides a number of these for the more frequently-cited journals.
1033: %% See the Author Guide for a list of them.
1034:
1035: %% Note that the style of the \bibitem labels (in []) is slightly
1036: %% different from previous examples. The natbib system solves a host
1037: %% of citation expression problems, but it is necessary to clearly
1038: %% delimit the year from the author name used in the citation.
1039: %% See the natbib documentation for more details and options.
1040:
1041: \begin{thebibliography}{}
1042: \bibitem[]{}Aguirre, A., Hernquist, L., Schaye, J., Katz, N., Weinberg, D.,
1043: \& Gardner, J. 2001, ApJ, 561, 521
1044: \bibitem[]{}Begelman, M. \& Fabian, A. 1990, \mnras, 244, 26
1045: \bibitem[]{}Bregman, J. 1980, \apj, 236, 577
1046: \bibitem[]{}Blanton, M., et al. 2001, \aj, 121, 2358
1047: \bibitem[]{}Cen,R., \& Ostriker, J. 1999, \apj, 514, 1
1048: \bibitem[]{}Cen,R., Tripp, T., Ostriker, J., \& Jenkins, E. 2001, \apj,
1049: 559, L5
1050: \bibitem[]{}Cowie, L. \& McKee, C. 1977, \apj, 211, 135
1051: \bibitem[]{}Cox, D., \& Smith, B. 1974, \apj, 189, L105
1052: \bibitem[]{}Dav\'e, R., Cen, R., Ostriker, J., Bryan, G., Hernquist, L.,
1053: Katz, N., Weinberg, D., Norman, M., \& O'Shea, B. 2001, \apj, 552, 473
1054: \bibitem[]{}Dopita, M., \& Sutherland, R. 1996, \apjs, 102, 161
1055: \bibitem[]{}Draine, B. 1981, \apj, 245, 880
1056: \bibitem[]{}Edgar \& Chevalier, R. 1986, \apj, 310, L27
1057: \bibitem[]{}Fukugita, M., Hogan, C., \& Peebles, J. 1998, \apj, 503, 518
1058: \bibitem[]{}Heckman, T. 2001, in
1059: ``Extragalactic Gas at Low Redshift'', ed. J. Mulchaey and J. Stocke,
1060: ASP Conf. Series, in press (astro-ph/0107438)
1061: \bibitem[]{}Heckman, T., Sembach, K., Meurer, G., Leitherer, Calzetti, D.,
1062: \& Martin, C. 2001a, \apj, 558, 56
1063: Calzetti, D., \& Leitherer, C. 2001a, \apj, 554, 1021
1064: \bibitem[]{}Heckman, T., Sembach, K., Meurer, G., Strickland, D., Martin, C.,
1065: Calzetti, D., \& Leitherer, C. 2001b, \apj, 554, 1021
1066: \bibitem[]{}Hoopes, C.G., Sembach, K.R., Howk, J.C., Savage, B.D., \&
1067: Fullerton, A.W. 2002, \apj, in press (astro-ph/0201351)
1068: \bibitem[]{}Howk, J.C., Sembach, K.R., Savage, B.D., Massa, D., Friedman, S.D.,
1069: \& Fullerton, A.W. 2002, \apj, 569, 214
1070: \bibitem[]{}Jenkins, E.B. 1978, \apj, 219, 845
1071: \bibitem[]{}Jenkins, E.B., Bowen, D.V., \& Sembach, K.R. 2001, in the
1072: Proceedings of the XVIIth IAP Colloquium: Gaseous Matter in Galactic and
1073: Intergalactic Space (astro-ph/0109363)
1074: \bibitem[]{}Martin, C., Heckman, T., Sembach, K., Meurer, G.,
1075: Calzetti, D., \& Leitherer, C. 2002, in preparation
1076: \bibitem[]{}McKee, C., \& Ostriker, J. 1977, \apj, 218, 148
1077: \bibitem[]{}Moos, H.W. et al. 2000, \apj, 538, L1
1078: \bibitem[]{}Mulchaey, J., Mushotzky, R., Burstein, D., \& Davis, D. 1996,
1079: \apj, 456, L5
1080: \bibitem[]{}Nicastro, F., Zezas, A., Drake, J., Elvis, M., Fiore, F.,
1081: Fruscione, A., Marengo, M., Mathur, S., \& Bianchi, S. 2002,
1082: astro-ph/02011058
1083: \bibitem[]{}Norman, C., \& Ikeuchi, S. 1989, \apj, 345, 372
1084: \bibitem[]{}Norman, C., \& Silk, J. 1979, \apj, 233, L1
1085: \bibitem[]{}Oegerle, W.R., Tripp, T.M., Sembach, K.R., Jenkins, E.B., Bowen,
1086: D.V., Cowie, L.L., Green, R.F., Kruk, J.W., Savage, B.D., Shull, J.M.,
1087: \& York, D.G. 2000, \apj, 538, L23
1088: \bibitem[]{}Richter, P., Sembach, K.R., Wakker, B.P., \& Savage, B.D. 2001a,
1089: \apj, 562, L181
1090: \bibitem[]{}Richter, P., Sembach, K.R., Wakker, B.P., \& Savage, B.D.,
1091: Tripp, T.M., Murphy, E.M., Kalberla, P.M.W., \& Jenkins, E.B. 2001b,
1092: \apj, 559, 318
1093: \bibitem[]{}Savage, B., Sembach, K., Jenkins, E.,
1094: Shull, J. M., York, D., Sonneborn, G.,
1095: Moos, H. W., Friedman, S., Green, J.,
1096: Oegerle, W., Blair, W., Kruk, J., \&
1097: Murphy, E. 2000, \apj, 538, L27
1098: \bibitem[]{}Savage, B., Sembach, K., Tripp, T.M., \& Richter, P. 2002, \apj,
1099: 564, 631
1100: \bibitem[]{}Savage, B. et al. 2002, in preparation
1101: \bibitem[]{}Sembach, K.R., Howk, J.C., Savage, B.D., Shull, J.M., \&
1102: Oegerle, W.R. 2001, \apj, 561, 573
1103: \bibitem[]{}Sembach, K., Savage, B., Shull, J. M.,
1104: Jenkins, E., Murphy, E., York, D., Ake, T.,
1105: Blair, W., Davidsen, A., Friedman, S.,
1106: Gibson, B., Kruk, J., Moos, H. W.,
1107: Oegerle, W., Sahnow, D., Sonneborn, G. 2000, \apj, 538, L31
1108: \bibitem[]{}Sembach, K.R. et al. 2002, in preparation
1109: \bibitem[]{}Shapiro, P., \& Field, G. 1976, \apj, 205, 762
1110: \bibitem[]{}Spitzer, L. 1978, Physical Processes in the Interstellar Medium,
1111: Wiley \& Sons
1112: \bibitem[]{}Sutherland, M., \& Dopita, M. 1993, \apjs, 88, 253
1113: \bibitem[]{}Thuan, T.X., Lecavelier des Etangs, A., \& Izotov, Y.I. 2002,
1114: \apj, 565, 941
1115: \bibitem[]{}Tripp, T., \& Savage, B. 2000, \apj, 542, 42
1116: \bibitem[]{}Tripp, T., Savage, B., \& Jenkins, E., 2000, \apj, 534, L1
1117: \bibitem[]{}Tripp, T., Giroux, M., Stocke, J., Tumlinson, J., \&
1118: Oegerle, W. 2001, \apj, 563, 724
1119: \bibitem[]{}Verner, D., Tytler, D., \& Barthel, P. 1994, \apj, 430, 186
1120: \bibitem[]{}Weaver, R., McCray, R., Castor, J., Shapiro, P., \& Moore, R.
1121: 1977, \apj, 218, 377
1122: \end{thebibliography}
1123:
1124: \clearpage
1125:
1126: %% Use the figure environment and \plotone or \plottwo to include
1127: %% figures and captions in your electronic submission.
1128:
1129:
1130: \newpage
1131: \clearpage
1132:
1133: \clearpage
1134: \begin{figure}[ht!]
1135: %\special{psfile=f1.eps hoffset=-40 voffset=-520 vscale=85 hscale=85}
1136: %\vspace{5.5in}
1137: \plotone{f1.eps}
1138: \caption{Column density versus line width for \ion{O}{6} absorption
1139: systems observed in the Galactic disk and halo, high velocity clouds, the
1140: Magellanic Clouds (LMC, SMC), starburst galaxies, and the intergalactic
1141: medium (IGM). Typical errors on these data points are 0.05 to 0.10 dex
1142: for $\log N_{OVI}$ and 0.05 to 0.08 dex for $\log \sigma$.
1143: The relation predicted for radiatively cooling gas is shown as the
1144: two long-dashed lines for an assumed temperature $T_{OVI} = 10^5$ K
1145: and $10^6$ K respectively. The short-dashed line in the lower right
1146: corner represents the typical detection limit in the $FUSE$ spectra.
1147: The solid lines show the effects of the blending of multiple narrow
1148: components with velocity separations as indicated. See text for details.
1149: \label{fig1}}
1150: \end{figure}
1151:
1152: \begin{figure*}
1153: \plotone{f2.eps}
1154: \caption{Spectra showing the addition of three broad (log b = 2.25) redshifted
1155: O{\sc vi}$\lambda$1031.9 absorption-lines to a $FUSE$ spectrum.
1156: From top to bottom these
1157: have residual intensities at line center of 85\%, 64\%, and 28\%.
1158: The strongest line corresponds to the typical broad O{\sc vi}
1159: lines (log b $>$ 2.0) in Figure 1. The weakest line corresponds to
1160: our adopted $FUSE$ detection threshold (dashed diagonal line in Figure 1).
1161: See text for details.}
1162: \label{fig2}
1163: \end{figure*}
1164:
1165: \begin{figure*}[!hb]
1166: \plotone{f3.eps}
1167: \caption{Approximate cooling column densities for various ions of abundant
1168: elements, plotted at the temperature at which each ion has its maximum
1169: fractional abundance assuming collisional ionization equilibrium.
1170: The cooling columns are proportional to the Mach number of the flow,
1171: and are plotted for $\cal M$ = 1.
1172: To maximize the clarity of the figure, the data are presented in two
1173: separate panels, and only some of the ions are explicitly identified.
1174: Oxygen cooling columns are plotted in both panels to aid comparison
1175: between the two panels. This figure is illustrative only, and more accurate
1176: cooling columns for selected ions based on detailed calculations
1177: are listed in Table 1. See text for details.
1178: \label{fig3}}
1179: \end{figure*}
1180:
1181: \begin{figure*}
1182: \plotone{f4.eps}
1183: \caption{Fractional abundance of the O{\sc vi} ion as a function of
1184: temperature, assuming collisional ionization equilibrium (Sutherland \&
1185: Dopita 1993).
1186: \label{fig4}}
1187: \end{figure*}
1188:
1189: \begin{figure*}
1190: \plotone{f5.eps}
1191: \caption{Predicted
1192: column density for the O{\sc vi} ion as a function of
1193: temperature for gas cooling radiatively from $T \geq 10^7$ K at
1194: a velocity $v_{cool} = c_s$ ($\cal M$ = 1). The
1195: secondary peak at high temperature is primarily a consequence
1196: of the increase in radiative cooling time with temperature,
1197: together with the moderate O{\sc vi} ionic fraction at
1198: temperatures up to a few million degrees (Figure 4).
1199: See text for details.
1200: \label{fig5}}
1201: \end{figure*}
1202:
1203: %% If you are not including electronic art with your submission, you may
1204: %% mark up your captions using the \figcaption command. See the
1205: %% User Guide for details.
1206: %%
1207: %% No more than seven \figcaption commands are allowed per page,
1208: %% so if you have more than seven captions, insert a \clearpage
1209: %% after every seventh one.
1210:
1211:
1212: %% Tables should be submitted one per page, so put a \clearpage before
1213: %% each one.
1214:
1215: %% Two options are available to the author for producing tables: the
1216: %% deluxetable environment provided by the AASTeX package or the LaTeX
1217: %% table environment. Use of deluxetable is preferred.
1218: %%
1219: \clearpage
1220:
1221: \begin{deluxetable}{lccccccc}
1222: \tablecolumns{0}
1223: \tablewidth{0pt}
1224: \tablecaption{Selected Cooling Columns and Absorption-Line Strengths}
1225: \tablehead{Ion & $log~T_{max}$ & $log N_{ib}$ & $log N_{ic}$ & Line & $log~N_{ib} f \lambda$ & $log~N_{ic} f \lambda$ & $\tau$$_{ic}$}
1226: \startdata
1227: C{\sc iv} & 5.00 & 12.5 & 13.9 & 1548.2 & 7.0 & 8.4 & 0.25\\
1228: N{\sc v} & 5.25 & 12.6 & 13.6 & 1238.8 & 6.9 & 7.9 & 0.08\\
1229: O{\sc iv} & 5.20 & 13.6 & 15.0 & 787.7 & 7.5 & 8.9 & 0.8\\
1230: O{\sc v} & 5.40 & 13.6 & 14.8 & 629.7 & 8.1 & 9.3 & 2.0\\
1231: O{\sc vi} & 5.45 & 14.2 & 15.0 & 1031.9 & 8.4 & 9.1 & 1.3\\
1232: O{\sc vii} & 5.90 & 16.4 & 16.9 & 21.60 & 9.6 & 10.1 & 13\\
1233: O{\sc viii} & 6.35 & 16.8 & 17.0 & 18.97 & 9.5 & 9.7 & 5\\
1234: Ne{\sc v} & 5.45 & 13.4 & 14.6 & 568.4 & 7.2 & 8.4 & 0.25\\
1235: Ne{\sc vi} & 5.65 & 13.7 & 14.7 & 558.6 & 7.6 & 8.6 & 0.4\\
1236: Ne{\sc viii} & 5.85 & 14.3 & 14.7 & 770.4 & 8.2 & 8.6 & 0.4\\
1237: Ne{\sc ix} & 6.20 & 16.3 & 16.5 & 13.45 & 9.3 & 9.5 & 3\\
1238: Mg{\sc x} & 6.05 & 14.5 & 14.8 & 609.8 & 8.2 & 8.5 & 0.3\\
1239: Si{\sc xii} & 6.35 & 14.8 & 15.0 & 499.5 & 8.3 & 8.5 & 0.3\\
1240: \enddata
1241: \tablecomments{The temperature at which each ion reaches its peak
1242: abundance in CIE is given in column 2. The calculations of cooling columns
1243: are based on the cooling flow behind a 600 km s$^{-1}$ shock (postshock
1244: temperature $5 \times 10^6$ K) using the Sutherland \& Dopita (1993)
1245: non-equilibrium zero-field cooling function and solar abundances. They assume
1246: that the cooling proceeds under isobaric conditions
1247: (column 3) or isochoric conditions (column 4). For the specific
1248: resonance absorption-lines whose wavelength is listed in column 5
1249: (in \AA), the implied line strengths
1250: for the isobaric and isochoric models are given in columns 6 and 7
1251: respectively. These are given in terms of the product of the column
1252: density, oscillator strength (htpp://physics.nist.gov/), and
1253: wavelength (in cm). The optical depth at line center
1254: is just $\tau = 0.015 N f \lambda/ b$ (where the Doppler parameter
1255: b is in cm s$^{-1}$). The implied optical depths at line-center for
1256: the isochoric model (b = $v_{shock}/4$ = 150 km s$^{-1}$) are given in column 8.
1257: The equivalent widths in the optically-thin
1258: limit are $W/\lambda = 8.85 \times 10^{-13} N f \lambda$ (e.g Spitzer 1978).}
1259: \end{deluxetable}
1260:
1261:
1262: \clearpage
1263:
1264: \begin{deluxetable}{lcccc}
1265: \tablecolumns{0}
1266: \tablewidth{0pt}
1267: \tablecaption{Radiative Cooling Models for the PKS 2155-304 Absorber}
1268: \tablehead{Columns & Data (both) & Data (single) & Isobaric & Isochoric}
1269: \startdata
1270: log~$N_{OVI}$ & 14.4$\pm$0.1 & 14.0$\pm$0.1 & 14.2 & 15.0\\
1271: log[$N_{OVII}/N_{OVI}$] & $\geq$1.2$\pm$0.3 & $\geq$1.6$\pm$0.3 & 2.2 & 1.9\\
1272: log[$N_{OVIII}/N_{OVI}$] & $\geq$1.3$\pm$0.4 & $\geq$1.7$\pm$0.4 & 2.6 & 2.0\\
1273: log[$N_{NeIX}/N_{OVI}$] & $\geq$1.4$\pm$0.3 & $\geq$1.8$\pm$0.3 & 2.1 & 1.5\\
1274: \enddata
1275: \tablecomments{The measured values for the column densities are listed
1276: twice. In column 2 we assume both the broad and narrow
1277: O{\sc vi} components seen in the $FUSE$ spectrum are associated with the
1278: X-ray absorber.
1279: In column 3, we assume that only the broad O{\sc vi} component
1280: is associated with the X-ray absorber (as proposed by
1281: Nicastro et al. 2002). In columns 2 and 3 the values for $N_{OVII}$,
1282: $N_{OVIII}$, and $N_{NeIX}$ are treated as lower limits to allow for
1283: the possibility that the X-ray absorption-lines are not optically-thin.
1284: The values in columns 4
1285: and 5 for isobaric and isochoric cooling are taken from the same
1286: calculations
1287: reported in Table 1.}
1288: \end{deluxetable}
1289:
1290: %% The following command ends your manuscript. LaTeX will ignore any text
1291: %% that appears after it.
1292:
1293: \end{document}
1294: