1: %%
2: %\documentclass[12pt,preprint]{aastex}
3: \documentclass{emulateapj}
4:
5: %% manuscript produces a one-column, double-spaced document:
6: %% \documentclass[manuscript]{aastex}
7:
8: %% preprint2 produces a double-column, single-spaced document:
9: %%\documentclass[preprint2]{aastex}
10:
11: %% Sometimes a paper's abstract is too long to fit on the
12: %% title page in preprint2 mode. When that is the case,
13: %% use the longabstract style option.
14: %% \documentclass[preprint2,longabstract]{aastex}
15:
16: %\usepackage{natbib} % Bibliography package
17: \newcommand{\me}{Zsarg\'{o}}
18: \newcommand{\mdot}{$\dot M$}
19: \newcommand{\zpup}{$\zeta$~Pup}
20: \newcommand{\vdag}{(v)^\dagger}
21: \newcommand{\myemail}{jaz8@pitt.edu}
22:
23: %% You can insert a short comment on the title page using the command below.
24: \slugcomment{To appear in ApJL}
25:
26: \shorttitle{On the importance of interclump medium}
27: \shortauthors{Zsarg\'{o}et al.}%{\me et al.}
28:
29: \begin{document}
30:
31: \title{On the Importance of the Interclump Medium for Superionization:\\
32: \ion{O}{6} Formation in the Wind of \zpup\ }
33:
34: \author{J.~Zsarg\'{o}\altaffilmark{1,2}, D.~J.~Hillier\altaffilmark{2}, J.-C.~Bouret\altaffilmark{3},
35: T.~Lanz\altaffilmark{4}, M.~A.~Leutenegger\altaffilmark{5,6}, D.~Cohen\altaffilmark{7}}
36:
37: \altaffiltext{1}{E-mail:~\myemail}
38:
39: \altaffiltext{2}{Department of Physics and Astronomy, University of Pittsburgh, 3941 O'Hara St.,
40: Pittsburgh, PA 15260, USA}
41:
42: \altaffiltext{3}{Laboratoire d'Astrophysique de Marseille, CNRS-Universit\'{e} de Provence,
43: 13388 Marseille cedex 13, FRANCE}
44:
45: \altaffiltext{4}{Department of Astronomy, University of Maryland, College Park, MD~20742, USA}
46:
47: \altaffiltext{5}{Goddard Space Flight Center, 8800 Greenbelt Road, Greenbelt, Maryland 20771, USA }
48:
49: \altaffiltext{6}{NASA Postdoctoral Fellow}
50:
51: \altaffiltext{7}{Department of Physics and Astronomy, Swarthmore College, 500 College Ave, Swarthmore,
52: PA 19081, USA}
53:
54: \begin{abstract}
55: We have studied superionization and X-ray line formation in the spectra of \zpup\
56: using our new stellar atmosphere code (XCMFGEN) that can be used to simultaneously analyze
57: optical, UV, and X-ray observations.
58: Here, we present results on the formation of the \ion{O}{6}\
59: $\lambda\lambda 1032, 1038$ doublet.
60: Our simulations, supported by simple theoretical calculations, show
61: that clumped wind models that assume void in the interclump space
62: cannot reproduce the observed \ion{O}{6} profiles.
63: However, enough \ion{O}{6}\ can be produced if the voids are filled by a low density gas.
64: The recombination of \ion{O}{6} is very efficient in the dense material but in the tenuous interclump region an observable amount of \ion{O}{6} can be maintained.
65: We also find that different UV resonance lines are sensitive to different density
66: regimes in \zpup\ :
67: \ion{C}{4} is almost exclusively formed within the densest regions, while
68: the majority of \ion{O}{6} resides between clumps.
69: \ion{N}{5} is an intermediate case, with contributions from both the tenuous gas and clumps.
70: \end{abstract}
71:
72: \keywords{stars:early-type -- stars:winds, outflow -- X-rays: stars -- X-rays:individual:\zpup}
73:
74: \section{Introduction}
75:
76: One of the surprising discoveries of the {\it Copernicus} satellite was the strong P-Cygnii
77: profiles of superions, such as \ion{O}{6} and \ion{N}{5}, in the FUV spectra of many O and B stars
78: \citep{sno76}.
79: The only viable explanation for the presence of \ion{O}{6} is Auger ionization by X-rays
80: from \ion{O}{4} \citep{cas79} which is the dominant form of oxygen in many O-type stars.
81: The X-ray emission, necessary for Auger ionization, was later detected by the first X-ray telescopes
82: \citep[e.g.,][]{har79, sew79}.
83:
84: The origin of the stellar X-ray emission was another enigma until the ``wind-shock'' mechanism
85: \citep{luc80} became the accepted explanation.
86: Massive stars posses strong line-driven winds in which the material is accelerated by numerous
87: C, N, O, and Fe transitions \citep[see e.g.,][]{pau90, CAK}.
88: It was known from the conception of the line-driven wind theory that such flows are unstable and prone
89: to the formation of dense clumps and shocks \citep[see e.g.,][]{owo91,luc80}.
90: The large-scale flow energy is converted to heat in the shock fronts producing high temperature plasma.
91: Numerical simulations confirm \citep{fel97, owo88} that at least the soft X-ray emission of early-type stars
92: can be explained by this mechanism.
93:
94: Evidence for density inhomogeneities (or clumped winds) is provided by variability studies of both WR stars \cite[][and references therein]{lep99}, and O stars \cite[e.g.,][]{eve98, lep08}.
95: Further, density inhomogeneities allow the electron-scattering wings of emission lines to be reduced to the observed level while maintaining the strength of emission lines \citep{hil91, ham98, hil99}.
96: More recently, \cite{cro02} and \cite{hil03} found that they could not simultaneously fit the H$\alpha$ and
97: \ion{P}{5} $\lambda\lambda$1120 profiles in normal O supergiants without assuming an inhomogeneous density distribution in the wind.
98: Using a more statistical approach, \cite{mas03} showed that the phosphorus ionization structure was consistent with expectations only if lower than conventional mass-loss rates were used in their analysis
99: of \ion{P}{5} $\lambda\lambda$1120.
100: Additional observational evidence for wind clumping come from {\it Chandra} and {\it XMM-Newton} high-resolution X-ray spectra of O stars.
101: These spectra revealed that X-ray lines suffer less absorption in the wind than predicted by ``smooth" models \citep[e.g.,][and references therein]{kra03b,coh05,wal07}.
102:
103: Superionization has received only limited attention since the work of \citeauthor{mac93} \citeyearpar{mac93, mac94}.
104: The effect was introduced into modern stellar atmosphere codes \citep[e.g., {\it WM-basic},][]{pau01},
105: but we are unaware of any work that has revisited the question in light of
106: the high-resolution X-ray observations, improved X-ray emission calculations and the new results on clumping.
107: With improvements to CMFGEN, we are developing tools and techniques to move towards this goal.
108: As part of this effort we discovered that the interclump medium is crucial to explain the
109: \ion{O}{6} doublet profile in $\zeta$~Pup.
110: In this paper we demonstrate the effect and discuss its implications.
111: %The paper is organized as follows.
112: In \S\ref{sec:obs} we briefly describe our code, the observations we used,
113: and our models. We present and discuss our results in \S\ref{sec:result}--\S\ref{sec:con}.
114:
115: \section{Observations, Tools, and Models} \label{sec:obs}
116:
117: We coadded {\it Copernicus} U1 scans for $\zeta$~Pup
118: to create the observed profile in Figure~\ref{fig1}.
119: All but two of these scans were blocked U1 scans that minimize stray-light as described in \cite{rog73a,rog73b}.
120: Our data reduction \citep[see,][]{zsa08} produced an \ion{O}{6} profile that strongly resemble
121: those in \cite{mac93} and \cite{mor76}.
122:
123: We have used XCMFGEN \citep{zsa08} to solve for the ionization balance and the non-LTE
124: level populations in our stellar models.
125: XCMFGEN is a new version of CMFGEN \citep{hil98} that can perform X-ray emission calculations
126: in addition to the original CMFGEN tasks.
127: We use APEC \citep{smi01} and its accompanying database of X-ray cross sections (APED) to
128: calculate non-LTE level populations and emissivities in the X-ray emitting plasma.
129:
130: \cite{bou05, bou08} used CMFGEN to derive stellar parameters for seven O supergiants.
131: We used their values for \zpup~to construct the models listed in Table~\ref{tab1}.
132: Our improved models specifically addressed formation of \ion{O}{6} and X-ray lines \citep{zsa08}.
133: The X-ray emitting plasma was distributed in the wind and
134: its emissivity was characterized by three plasma temperatures.
135: The parametrization included a volume filling factor ($f_X$) and an initial turn-on
136: radius (R$_0$), as in \cite{owo01}.
137: The volume filling factor was constant beyond R$_0$ and was adjusted until
138: the prescribed L$_X$/L$_{BOL}$ was met (see Table~\ref{tab1}).
139: Note that L$_X$ is the X-ray luminosity attenuated by the wind and not
140: the intrinsic luminosity.
141:
142: To treat clumping we follow the ``volume filling factor'' approach where the material is compressed
143: into a fraction of the available volume with void in between \citep[see e.g.,][]{hil99, ham98}.
144: The ratio of the volume filled with material (clumps) to the total is
145: \begin{equation}
146: f_{cl}(r) = f_{\infty} + (1 - f_{\infty}) \cdot \exp( -v(r)/v_{cl} ) \label{eq:cl}
147: \end{equation}
148: where $f_{\infty}$ and $v_{cl}$ are free parameters.
149: Eq.~\ref{eq:cl} is an ad-hoc formula, motivated by hydrodynamical simulations, and provides a
150: smooth wind at low $v(r)$ velocities ($r \sim R_*$).
151:
152: In our models, $f_X$ and $f_{cl}$ are two independent parameters despite
153: the fact that both are referred to as ``filling factors''.
154: The two should be related since both of them are the results of the wind instability.
155: However, our understanding of the line-driven winds is too poor to formulate this relationship.
156:
157: \begin{table}
158: \caption{Model Parameters (representative for \zpup). \label{tab1}}
159: \begin{tabular}{lcccr}
160: \tableline\tableline
161: & & & \\
162: {\bf Photosphere/Wind} & & & \\
163: R$_*$\tablenotemark{a} & \multicolumn{4}{c}{19~R$_{\odot}$} \\
164: T$_{eff}$ & \multicolumn{4}{c}{39,000~K} \\
165: log~$g$ (cgs) & \multicolumn{4}{c}{3.6} \\
166: L$_*$ & \multicolumn{4}{c}{7.0$\times$10$^5$~L$_{\odot}$} \\
167: v$_{turb}$ & \multicolumn{4}{c}{12~km~s$^{-1}$} \\
168: v~sin~i & \multicolumn{4}{c}{240~km~s$^{-1}$} \\
169: v$_{\infty}$ & \multicolumn{4}{c}{2300~km~s$^{-1}$} \\
170: $\beta$\tablenotemark{b} & \multicolumn{4}{c}{0.9} \\
171: & \multicolumn{2}{c}{\it Smooth Wind} & \multicolumn{2}{c}{\it Clumped Wind} \\
172: $\dot M$ & \multicolumn{2}{c}{7.6$\times$10$^{-6}$~M$_{\odot}$yr$^{-1}$} &
173: \multicolumn{2}{c}{1.7$\times$10$^{-6}$~M$_{\odot}$yr$^{-1}$} \\
174: $f_{\infty}$ & \multicolumn{2}{c}{$\ldots$} & \multicolumn{2}{c}{0.05} \\
175: v$_{cl}$ & \multicolumn{2}{c}{$\ldots$} & \multicolumn{2}{c}{150~km~s$^{-1}$} \\
176: & & & \\
177: {\bf Plasma} & & & \\
178: R$_0$\tablenotemark{c} & \multicolumn{4}{c}{1.5~R$_*$} \\
179: L$_X$/L$_{BOL}$ & \multicolumn{4}{c}{1.5$\times$10$^{-7}$} \\
180: T$_{pl}$~1\tablenotemark{a} & \multicolumn{4}{c}{4.6$\times$10$^6$~K} \\
181: T$_{pl}$~2 & \multicolumn{4}{c}{2.4$\times$10$^6$~K} \\
182: T$_{pl}$~3\tablenotemark{a} & \multicolumn{4}{c}{1.7$\times$10$^6$~K} \\
183: \tableline
184: \end{tabular}
185: %% Any table notes must follow the \end{tabular} command.
186: \tablenotetext{a}{Value is from \cite{hil93}.}
187: \tablenotetext{b}{Power of the CAK velocity law \citep{CAK}.}
188: \tablenotetext{c}{Value is from \cite{kra03b, leu06}.}
189: \tablecomments{Stellar parameters are from \cite{bou08}, unless noted otherwise.
190: Plasma parameters are from \cite{zsa08}, unless noted otherwise. For simplicity, the same
191: emission measure $EM_{X} = \int n_e n_H dV$ was assumed for each $T_{pl}$.}
192: \end{table}
193:
194: \section{Results and Discussion} \label{sec:result}
195:
196: Figure~\ref{fig1} shows our model calculations together with the {\it Copernicus}
197: spectrum.
198: Auger ionization produces strong \ion{O}{6} features in the case of the smooth wind model,
199: albeit a bit weaker than observed.
200: A better fit can be achieved by increasing the oxygen abundance and by assuming larger
201: turbulent velocities near $v_{\infty}$.
202: A detailed analysis of \zpup\ and the simultaneous fit to the X-ray and UV spectra, will be
203: presented in \cite{zsa08}.
204: The most important feature of Fig.~\ref{fig1} is the absence of an \ion{O}{6} profile
205: in the clumped wind model.
206: This cannot be rectified by fiddling with, for example, the oxygen abundance.
207: Adjustments at the order of magnitude level would be necessary to produce a visible \ion{O}{6} profile;
208: this indicates that something fundamental is wrong with the simplified ``clumped'' model.
209:
210: \begin{figure}
211: \includegraphics[angle=270,scale=0.35]{f1.eps}
212: \caption[]{Comparison of the {\it Copernicus} observation
213: of \ion{O}{6} $\lambda\lambda$1035 (rugged line) with those
214: calculated by XCMFGEN for a smooth wind (red), for a clumped wind
215: model (blue), and for the interclump medium (green).
216: See Table~\ref{tab1} and \S\ref{sec:result} for details.}
217: \label{fig1}
218: \end{figure}
219:
220: There is a simple theoretical explanation for the weak \ion{O}{6} P-Cygnii profiles
221: in clumped models.
222: Since the \ion{O}{6} $\lambda\lambda$1032, 1038 doublet forms only in the wind by
223: redistribution of the stellar radiation, the optical depth controls the shape and
224: strength of the profiles \citep[see e.g.,][]{lam87}.
225: In a spherical and accelerating flow it is appropriate to write the optical depth as
226: \begin{equation}
227: \tau_{O~{\rm VI}} \sim \frac{\pi e^2}{m c} \; \frac{f \, n_{O~{\rm VI}} \, \lambda_0 }{\frac{v}{r} +
228: \left( \frac{dv}{dr} - \frac{v}{r} \right) \mu^2}
229: \label{eq:tau}
230: \end{equation}
231: where the atomic parameters have the usual meanings and $\arccos(\mu)$ is
232: the angle between a line-of-sight and the radial direction.
233: The average \ion{O}{6} density can be written as
234: \begin{equation}
235: n_{O~{\rm VI}} = A_O \, n \, q_{O~{\rm VI}} \sim A_O \, \frac{\dot M}{4 \pi r^2 v(r) m_p } \, q_{O~{\rm VI}} \;
236: \label{eq:n_i}
237: \end{equation}
238: by using the oxygen abundance $A_O$, the \ion{O}{6} fraction $q_{O~{\rm VI}}$, and the average
239: particle mass $m_p$.
240: Note, that $n_{O~{\rm VI}}$ is the {\it mean ion density over a Sobolev length}, so
241: \begin{equation}
242: n_{O~{\rm VI}} = n_{{O~{\rm VI}}, \, cl} \, f_{cl} = A_O \, n_{cl} \, f_{cl} \, q_{O~{\rm VI}}
243: \label{eq:n_icl}
244: \end{equation}
245: must be used if the wind material occupies only a fraction $f_{cl}$ of the available volume with
246: $n_{cl}$ number density.
247:
248: The key unknown in the expression for $\tau_{O~{\rm VI}}$ is the \ion{O}{6} fraction.
249: We can estimate $q_{O~{\rm VI}}$ by a simplified rate equation:
250: \begin{equation}
251: \frac{dn_{ O~{\rm VI} }}{dt} = n_{ O~{\rm IV} } \, \alpha (J_X) - n_{O~{\rm VI}} \, n_e \, \gamma = 0 ,
252: \label{eq:SEE}
253: \end{equation}
254: where $\alpha (J_X)$ is the X-ray-flux-dependent rate for Auger ionization, and $\gamma$ is
255: the effective recombination rate coefficient.
256: Solving for the ion fraction,
257: \begin{equation}
258: \frac{n_{O~{\rm VI}}}{n_{O~{\rm IV}}} = \frac{q_{O~{\rm VI}}}{q_{O~{\rm IV}}} = \frac{\alpha
259: (J_X)}{n_e \gamma}.
260: \label{eq:orat}
261: \end{equation}
262: Under the approximation that $q_{O~{\rm IV}} \sim 1$ and $n_e \sim n$, we find that $q_{O~{\rm VI}} =
263: \frac{\alpha (J_X)}{n \gamma}$, or
264: \begin{equation}
265: n_{O~{\rm VI}} \; = \; A_O \, n \, q_{O~{\rm VI}} \; = \; A_O \, \frac{\alpha (J_X)}{\gamma} \; .
266: \label{eq:nosix}
267: \end{equation}
268: The ion density of \ion{O}{6}, therefore, is {\it independent} of the wind density, and thus
269: the mass-loss rate.
270: Substituting this in the expression for the optical depth, we find for a smooth wind
271: \begin{equation}
272: \tau_{O~{\rm VI}, sm} \sim \frac{\pi e^2}{m c} \; \frac{A_O \, \alpha(J_X)}{\gamma}
273: \; \frac{f \, \lambda_0}{ \frac{v}{r} + \left( \frac{dv}{dr} - \frac{v}{r} \right) \mu^2} \; .
274: \label{eq:tauosix}
275: \end{equation}
276: The Sobolev optical depth in a smooth wind is thus also independent of the mass-loss rate.
277:
278: Let us generalize this to a clumped wind.
279: Equations~\ref{eq:orat} and \ref{eq:nosix} still give the \ion{O}{6} fraction and density,
280: but these are the values for the dense clumps, and not means over a Sobolev length.
281: Thus we need to use Eqs.~\ref{eq:tau}, \ref{eq:n_icl}, and \ref{eq:orat}, together with
282: the assumptions of $q_{O~{\rm IV}} \sim 1$ and $n_e \sim n$ to derive
283: \begin{equation}
284: \tau_{O~{\rm VI},cl} \; \sim
285: \; f_{cl} \, \tau_{O~{\rm VI}, sm}
286: \label{eq:tauosixcl}
287: \end{equation}
288: in a clumped wind.
289:
290: It is clear now why clumped models fail to produce observable \ion{O}{6} lines.
291: The volume filling factor for \zpup\ ($f_{cl}\sim$~0.05 at $r>>R_*$)
292: decreases $\tau_{O~{\rm VI},cl}$ well below unity.
293: The results of our XCMFGEN simulations fully support the above calculation.
294: Figure~\ref{fig2} shows the radial optical depth of the important UV resonance lines in our
295: models (Table~\ref{tab1}).
296: The curves show what is predicted by Eq.~\ref{eq:tauosixcl} for small $f_{cl}$;
297: the \ion{O}{6} optical depth is nearly two orders of magnitude lower than in the smooth wind model.
298: The dense clumps, therefore, contribute little or nothing to the \ion{O}{6} profiles.
299:
300: \begin{figure}
301: \includegraphics[angle=270,scale=0.35]{f2.eps}
302: \caption{
303: The radial optical depth as a function of normalized wind velocity for the
304: stronger component of the \ion{N}{5} $\lambda\lambda$1240 (red), \ion{C}{4}
305: $\lambda\lambda$1550 (blue), and \ion{O}{6} $\lambda\lambda$1035 (green) doublets.
306: The solid and dashed lines are for smooth and clumped wind models (Table~\ref{tab1}),
307: respectively.
308: \label{fig2}}
309: \end{figure}
310:
311: There are two other factors that further decrease the \ion{O}{6} optical depth, and hence
312: adversely influence the model \ion{O}{6} profile.
313: First, \cite{bou08} found sub-solar oxygen abundances for all stars in their sample (a factor of
314: five difference for \zpup); this affects only comparisons with earlier calculations.
315: Second, the X-ray flux available for Auger ionization tends to decrease when $f_{cl}$ is lowered;
316: the wind absorption is smaller, hence less X-ray emission is needed to meet the observed
317: L$_X$/L$_{BOL}$.
318:
319: What are the implications of our failure to produce observable \ion{O}{6}
320: lines with clumped winds?
321: Is the wind smoother than suggested by \cite{bou08}?
322: Their filling factor for \zpup~is not extremely low; it is very similar to those found for
323: other O supergiants \citep{hil03,mas03}. Further, increasing $f_{cl}$ to 0.1 or 0.2 would not
324: solve the problem.
325: It is more likely that the use of the traditional ``filling factor'' approach in XCMFGEN causes
326: the failure to produce observable \ion{O}{6} lines.
327: Equations~\ref{eq:tauosix} and \ref{eq:tauosixcl} suggest that the part of the wind with the largest filling
328: factor will be the most important at producing O VI optical depth.
329: Thus we need to consider the influence of the tenuous interclump medium (ICM) on the \ion{O}{6} profile.
330:
331: This is a non-trivial exercise; at least three components (dense clumps, ICM, and
332: the hot plasma) need to be treated simultaneously for a fully self-consistent solution.
333: Fortunately, the radiation field at any radius is almost independent of the ICM.
334: Therefore, we can use $J_{\nu}(R)$ from our clumped model and solve the statistical
335: equilibrium equations in the ICM only.
336: We created such a model from the clumped model of Table~\ref{tab1}.
337: We scaled down the densities by a factor of $f_{cl}^2$ to simulate the ICM at $r>>R_*$, and
338: imported the radiation field from the original clumped model.
339: We also assumed that the new model is smooth, reflecting the high volume filling fraction of
340: the ICM.
341: Effectively, we took the ICM component out of the real (clump+ICM+hot plasma) wind and built a
342: stand-alone model for it.
343: Since we imported the radiation field from the clumped model, the coupling between the ICM and
344: the rest of the wind has been taken into account at least in the first order.
345:
346: The choice of density in the ICM is somewhat arbitrary, but in the absence of reliable hydrodynamical
347: predictions it is still reasonable.
348: There is a density contrast of four hundred between the dense clumps and the ICM at $r>>R_*$.
349: This means that the ICM contributes little to the overall mass-loss rates.
350: If we combined the models for the ICM and clumps the total \mdot\ would be nearly identical
351: (within 5\%) to that of the clumped model.
352:
353: The result of this experiment is displayed in Figs.~\ref{fig1} and \ref{fig3}.
354: It is obvious from both figures that the ICM has substantial \ion{O}{6} optical depth
355: and contributes a strong P-Cygnii profile.
356: Our results also offer a glimpse on the behavior of other lines.
357: The \ion{C}{4} optical depth in the ICM is very low and thus most of the \ion{C}{4}
358: $\lambda\lambda$1550 lines are produced by the clumps.
359: The behavior of \ion{N}{5} $\lambda\lambda$1240 is very interesting.
360: It appears from Fig.~\ref{fig3} that both the ICM and the dense medium contribute.
361: %Other photospheric and wind features are less likely to be influenced by the tenuous interclump medium.
362:
363: \begin{figure}
364: \includegraphics[angle=270,scale=0.35]{f3.eps}
365: \caption{
366: Same as figure~\ref{fig2} but the solid
367: curves are now for the interclump medium.
368: See \S\ref{sec:result} for details. \label{fig3}}
369: \end{figure}
370:
371: Our results warrant an investigation of the physical conditions in the ICM and also its role
372: in the formation of UV resonance lines in different environments (e.g., low and high density winds).
373: For example, what is the density contrast between the ICM and clumps?
374: At very low densities
375: the average charge state of oxygen (and other species) may be drastically different
376: from the one in the dense regions.
377: Also, how closely does the ICM follow the velocity of the clumps?
378: Do \ion{O}{6} and \ion{C}{4} follow a different velocity structure?
379: Further, how are the ICM and hot plasma related?
380:
381: \section{Conclusion} \label{sec:con}
382:
383: In this letter we present our first results on superionization
384: in clumped winds, and showcase the potential of the interclump medium to
385: produce observable features.
386: Clumped wind models that use the classical ``volume filling factor" approach
387: (clumps with voids in between) cannot reproduce the observed \ion{O}{6}
388: profile in $\zeta$~Pup.
389: The recombination of \ion{O}{6} is too efficient and the necessary fractional
390: abundance cannot be sustained in the clumps.
391: However, a tenuous interclump medium can contribute enough \ion{O}{6} to produce an
392: observable \ion{O}{6} profile.
393: Only a small amount of mass is necessary in the interclump medium, so its overall
394: effect on the derived mass-loss rates is negligible.
395:
396: Our result highlights the need for improved treatment of clumping
397: in the winds of massive stars.
398: It is impossible to achieve a simultaneous fit to all
399: UV P-Cygnii profiles with a single component wind model for \zpup.
400: Our simulations suggest that in $\zeta$~Pup, different UV resonance lines
401: probe different density regimes.
402: \ion{C}{4} is formed almost exclusively in the dense
403: material, while \ion{O}{6} likely originates from the interclump medium.
404: \ion{N}{5} is an intermediate case with similar contributions from both components.
405: In cooler O stars, when N$^{2+}$ become the dominant ionization stage, we might expect
406: that \ion{N}{5}\ shows the same behavior as \ion{O}{6}\ in the hotter stars.
407: Obviously other possible effects of the interclump medium, in both O and W-R stars,
408: should be investigated.
409:
410: \acknowledgments
411:
412: This research was supported by STScI grant HST-AR-10693.02 and by SAO grant TM6-7003X.
413: We are also grateful to Dr. Randall Smith for providing us the source code of APEC, and
414: to the Chandra X-ray Center for the use of ATOMDB.
415: Maurice A. Leutenegger acknowledges support from a fellowship administered by Oak Ridge
416: Associated Universities under the NASA Postdoctoral Program.
417: J.-C.~Bouret acknowledges financial support from the French National Research Agency (ANR)
418: through program number ANR-06-BLAN-0105.
419:
420: %\include{ms_bbl}
421:
422: \begin{thebibliography}{38}
423: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
424:
425: \bibitem[{{Bohlin}(1975)}]{boh75}
426: {Bohlin}, R.~C. 1975, \apj, 200, 402
427:
428: \bibitem[{{Bouret} {et~al.}(2005){Bouret}, {Lanz}, \& {Hillier}}]{bou05}
429: {Bouret}, J.-C., {Lanz}, T., \& {Hillier}, D.~J. 2005, \aap, 438, 301
430:
431: \bibitem[{{Bouret} {et~al.}(2008){Bouret}, {Lanz}, {Hillier}, \&
432: {Foellmi}}]{bou08}
433: {Bouret}, J.-C., {Lanz}, T., {Hillier}, D.~J., \& {Foellmi}, C. 2008, \aap, in
434: preparation, ~
435:
436: \bibitem[{{Cassinelli} \& {Olson}(1979)}]{cas79}
437: {Cassinelli}, J.~P. \& {Olson}, G.~L. 1979, \apj, 229, 304
438:
439: \bibitem[{Castor {et~al.}(1975)Castor, Abbott, \& Klein}]{CAK}
440: Castor, J., Abbott, D.~C., \& Klein, R.~I. 1975, \apj, 195, 157
441:
442: \bibitem[{{Cohen} {et~al.}(2006){Cohen}, {Leutenegger}, {Grizzard}, {Reed},
443: {Kramer}, \& {Owocki}}]{coh05}
444: {Cohen}, D.~H., {Leutenegger}, M.~A., {Grizzard}, K.~T., {et~al.} 2006, \mnras,
445: 368, 1905
446:
447: \bibitem[{{Crowther} {et~al.}(2002){Crowther}, {Hillier}, {Evans}, {Fullerton},
448: {De Marco}, \& {Willis}}]{cro02}
449: {Crowther}, P.~A., {Hillier}, D.~J., {Evans}, C.~J., {et~al.} 2002, \apj, 579,
450: 774
451:
452: \bibitem[{{Eversberg} {et~al.}(1998){Eversberg}, {L\'{e}pine}, \&
453: {Moffat}}]{eve98}
454: {Eversberg}, T., {L\'{e}pine}, S., \& {Moffat}, A.~F.~J. 1998, \apj, 494, 799
455:
456: \bibitem[{{Feldmeier} {et~al.}(1997){Feldmeier}, {Puls}, \&
457: {Pauldrach}}]{fel97}
458: {Feldmeier}, A., {Puls}, J., \& {Pauldrach}, A.~W.~A. 1997, \aap, 322, 878
459:
460: \bibitem[{{Hamann} \& {Koesterke}(1998)}]{ham98}
461: {Hamann}, W.-R. \& {Koesterke}, L. 1998, \aap, 335, 1003
462:
463: \bibitem[{{Harnden} {et~al.}(1979){Harnden}, {Branduardi}, {Gorenstein},
464: {Grindlay}, {Rosner}, {Topka}, {Elvis}, {Pye}, \& {Vaiana}}]{har79}
465: {Harnden}, Jr., F.~R., {Branduardi}, G., {Gorenstein}, P., {et~al.} 1979,
466: \apjl, 234, L51
467:
468: \bibitem[{{Hillier}(1991)}]{hil91}
469: {Hillier}, D.~J. 1991, \aap, 247, 455
470:
471: \bibitem[{{Hillier} {et~al.}(1993){Hillier}, {Kudritzki}, {Pauldrach}, {Baade},
472: {Cassinelli}, {Puls}, \& {Schmitt}}]{hil93}
473: {Hillier}, D.~J., {Kudritzki}, R.~P., {Pauldrach}, A.~W., {et~al.} 1993, \aap,
474: 276, 117
475:
476: \bibitem[{{Hillier} {et~al.}(2003){Hillier}, {Lanz}, {Heap}, {Hubeny}, {Smith},
477: {Evans}, {Lennon}, \& {Bouret}}]{hil03}
478: {Hillier}, D.~J., {Lanz}, T., {Heap}, S.~R., {et~al.} 2003, \apj, 588, 1039
479:
480: \bibitem[{{Hillier} \& {Miller}(1998)}]{hil98}
481: {Hillier}, D.~J. \& {Miller}, D.~L. 1998, \apj, 496, 407
482:
483: \bibitem[{{Hillier} \& {Miller}(1999)}]{hil99}
484: {Hillier}, D.~J. \& {Miller}, D.~L. 1999, \apj, 519, 354
485:
486: \bibitem[{{Kramer} {et~al.}(2003){Kramer}, {Cohen}, \& {Owocki}}]{kra03b}
487: {Kramer}, R.~H., {Cohen}, D.~H., \& {Owocki}, S.~P. 2003, \apj, 592, 532
488:
489: \bibitem[{{Lamers} {et~al.}(1987){Lamers}, {Cerruti-Sola}, \&
490: {Perinotto}}]{lam87}
491: {Lamers}, H.~J.~G.~L.~M., {Cerruti-Sola}, M., \& {Perinotto}, M. 1987, \apj,
492: 314, 726
493:
494: \bibitem[{{L{\'e}pine} \& {Moffat}(1999)}]{lep99}
495: {L{\'e}pine}, S. \& {Moffat}, A.~F.~J. 1999, \apj, 514, 909
496:
497: \bibitem[{{L{\'e}pine} \& {Moffat}(2008)}]{lep08}
498: {L{\'e}pine}, S. \& {Moffat}, A.~F.~J. 2008, \aj, 136, 548
499:
500: \bibitem[{{Leutenegger} {et~al.}(2006){Leutenegger}, {Paerels}, {Kahn}, \&
501: {Cohen}}]{leu06}
502: {Leutenegger}, M.~A., {Paerels}, F.~B.~S., {Kahn}, S.~M., \& {Cohen}, D.~H.
503: 2006, \apj, 650, 1096
504:
505: \bibitem[{{Lucy} \& {White}(1980)}]{luc80}
506: {Lucy}, L.~B. \& {White}, R.~L. 1980, \apj, 241, 300
507:
508: \bibitem[{{Macfarlane} {et~al.}(1994){Macfarlane}, {Cohen}, \& {Wang}}]{mac94}
509: {Macfarlane}, J.~J., {Cohen}, D.~H., \& {Wang}, P. 1994, \apj, 437, 351
510:
511: \bibitem[{{MacFarlane} {et~al.}(1993){MacFarlane}, {Waldron}, {Corcoran}, J.,
512: {Wang}, \& {Cassinelli}}]{mac93}
513: {MacFarlane}, J.~J., {Waldron}, W.~L., {Corcoran}, M.~F., {et~al.} 1993, \apj,
514: 419, 813
515:
516: \bibitem[{{Massa} {et~al.}(2003){Massa}, {Fullerton}, {Sonneborn}, \&
517: {Hutchings}}]{mas03}
518: {Massa}, D., {Fullerton}, A.~W., {Sonneborn}, G., \& {Hutchings}, J.~B. 2003,
519: \apj, 586, 996
520:
521: \bibitem[{{Morton}(1976)}]{mor76}
522: {Morton}, D.~C. 1976, \apj, 203, 386
523:
524: \bibitem[{{Owocki} {et~al.}(1988){Owocki}, {Castor}, \& {Rybicki}}]{owo88}
525: {Owocki}, S.~P., {Castor}, J.~I., \& {Rybicki}, G.~B. 1988, \apj, 335, 914
526:
527: \bibitem[{{Owocki} \& {Cohen}(2001)}]{owo01}
528: {Owocki}, S.~P. \& {Cohen}, D.~H. 2001, \apj, 559, 1108
529:
530: \bibitem[{{Owocki} \& {Rybicki}(1991)}]{owo91}
531: {Owocki}, S.~P. \& {Rybicki}, G.~B. 1991, \apj, 368, 261
532:
533: \bibitem[{{Pauldrach} {et~al.}(2001){Pauldrach}, {Hoffmann}, \&
534: {Lennon}}]{pau01}
535: {Pauldrach}, A.~W.~A., {Hoffmann}, T.~L., \& {Lennon}, M. 2001, \aap, 375, 161
536:
537: \bibitem[{{Pauldrach} \& {Puls}(1990)}]{pau90}
538: {Pauldrach}, A.~W.~A. \& {Puls}, J. 1990, \aap, 237, 409
539:
540: \bibitem[{{Rogerson} {et~al.}(1973{\natexlab{a}}){Rogerson}, {Spitzer},
541: {Drake}, {Dressler}, {Jenkins}, {Morton}, \& {York}}]{rog73a}
542: {Rogerson}, J.~B., {Spitzer}, L., {Drake}, J.~F., {et~al.} 1973{\natexlab{a}},
543: \apjl, 181, L97+
544:
545: \bibitem[{{Rogerson} {et~al.}(1973{\natexlab{b}}){Rogerson}, {York}, {Drake},
546: {Jenkins}, {Morton}, \& {Spitzer}}]{rog73b}
547: {Rogerson}, J.~B., {York}, D.~G., {Drake}, J.~F., {et~al.} 1973{\natexlab{b}},
548: \apjl, 181, L110+
549:
550: \bibitem[{{Seward} {et~al.}(1979){Seward}, {Forman}, {Giacconi}, {Griffiths},
551: {Harnden}, {Jones}, \& {Pye}}]{sew79}
552: {Seward}, F.~D., {Forman}, W.~R., {Giacconi}, R., {et~al.} 1979, \apjl, 234,
553: L55
554:
555: \bibitem[{{Smith} {et~al.}(2001){Smith}, {Brickhouse}, {Liedahl}, \&
556: {Raymond}}]{smi01}
557: {Smith}, R.~K., {Brickhouse}, N.~S., {Liedahl}, D.~A., \& {Raymond}, J.~C.
558: 2001, \apjl, 556, L91
559:
560: \bibitem[{{Snow} \& {Morton}(1976)}]{sno76}
561: {Snow}, T.~P. \& {Morton}, D.~C. 1976, \apjs, 32, 429
562:
563: \bibitem[{{Waldron} \& {Cassinelli}(2007)}]{wal07}
564: {Waldron}, W.~L. \& {Cassinelli}, J.~P. 2007, \apj, 668, 456
565:
566: \bibitem[{{Zsarg\'{o}} \& {Hillier}(2008)}]{zsa08}
567: {Zsarg\'{o}}, J. \& {Hillier}, D.~J. 2008, \apj, in preparation
568:
569: \end{thebibliography}
570:
571: \end{document}
572: