1: %\documentclass{emulateapj}
2: \documentclass[12pt,preprint]{aastex}
3:
4: %\slugcomment{\today}
5:
6: \lefthead{}
7: \righthead{Argon Abundances in Orion Association B Stars}
8:
9: \begin{document}
10:
11: \title{Argon Abundances in the Solar Neighborhood: \\
12: Non-LTE Analysis of Orion Association B-type Stars\altaffilmark{1}}
13:
14: \author{Thierry Lanz}
15: \affil{Department of Astronomy, University of Maryland, College Park MD 20742, USA;
16: lanz@astro.umd.edu}
17:
18: \author{Katia Cunha\altaffilmark{2}}
19: \affil{National Optical Astronomy Observatory,
20: Casilla 604, La Serena Chile; kcunha@noao.edu}
21:
22: \author{Jon Holtzman}
23: \affil{New Mexico State University, Las Cruces NM 88003, USA;
24: holtz@nmsu.edu}
25:
26: \author{Ivan Hubeny}
27: \affil{Steward Observatory, University of Arizona, Tucson AZ 85712, USA;
28: hubeny@aegis.as.arizona.edu}
29:
30: \begin{abstract}
31: Argon abundances have been derived for a sample of B main-sequence stars in the Orion association.
32: The abundance calculations are based on NLTE metal line-blanketed model atmospheres calculated
33: with the NLTE code TLUSTY and an updated and complete argon model atom.
34: We derive an average argon abundance for this young population of
35: A(Ar)~=~6.66 $\pm$~0.06. While our result is in excellent agreement with a recent analysis
36: of the Orion nebula, it is significantly higher than the currently recommended solar
37: value which is based on abundance measurements in the solar corona. Moreover, the derived argon
38: abundances in the Orion B stars agree very well with a measurement from a solar impulsive flare during which
39: unmodified solar photospheric material was brought to flare conditions. We therefore
40: argue that the argon abundances obtained independently for both the Orion B stars and the Orion nebula
41: are representative of the disk abundance value in the solar neighborhood. The lower
42: coronal abundance may reflect a depletion related to the FIP effect. We propose
43: a new reference value for the abundance of argon in the solar neighborhood,
44: A(Ar)~=~6.63 $\pm$~0.10, corresponding to Ar/O~=~0.009 $\pm$~0.002.
45:
46: %We propose
47: %new reference values for the abundance of neon and argon in the solar neighborhood,
48: %A(Ne)~=~8.09 $\pm$~0.06, and A(Ar)~=~6.63 $\pm$~0.10, corresponding to Ne/O~=~0.26 and
49: %Ar/O~=~0.009.
50:
51: %Thierry, I am not sure if we should keep this info about the neon abundance in the abstract because
52: %we are not really dealing with neon in this paper. I think that having this
53: %in the conclusions would be good but not in the abstract (Just a suggestion!
54: %however, happy if prefer to.
55:
56: \end{abstract}
57:
58:
59: \keywords{astrochemistry --- stars: abundances, early-type --- Sun: abundances}
60:
61:
62: \altaffiltext{1}{Based on observations obtained with the Apache Point Observatory
63: 3.5-meter telescope, which is owned and operated by the Astrophysical
64: Research Consortium.}
65:
66: \altaffiltext{2}{On leave from Observat\'orio Nacional - MCT; Rio de Janeiro, Brazil}
67:
68: \setcounter{footnote}{2}
69:
70: % -----------------------------------------------------------------------------
71:
72: \section{INTRODUCTION}
73:
74: Establishing the abundance of noble gases in the Sun has been a notoriously
75: complex task, with significant uncertainties still remaining in some cases.
76: These difficulties primarily ensue from the atomic structure of noble gases,
77: in particular their high first ionization potential and the large excitation energies
78: of the excited levels in neutral atoms. Noble gases are therefore in the neutral
79: ground state to a great extent in the solar photosphere and, consequently,
80: no spectral lines of these elements have been identified in the photospheric
81: solar spectrum. Well-established model stellar atmosphere and quantitative
82: spectroscopic techniques therefore cannot be applied to study noble gases in
83: the Sun. Furthermore, these species are volatile. Meteoric studies thus yield
84: only a lower limit of the actual abundance of noble gases in the solar system and,
85: therefore, are of limited usefulness for inferring the solar argon abundance.
86:
87: While the solar helium abundance can be determined with a great degree of accuracy
88: from helioseismology, abundance studies of heavier noble gases depend on
89: the analysis of coronal lines and {\sl in-situ\/} measurements of energetic particles
90: from the solar wind. These techniques have greater uncertainty than photospheric
91: abundance studies which have been applied to many other chemical elements.
92: Moreover, measurements are made relative to a reference element, and
93: they actually yield abundance ratios such as the Ar/O ratio.
94: In a recent compilation of solar abundance works, \cite{AGS05} recommend a solar
95: argon abundance, A(Ar)$_{\odot} = 6.18$.\footnote{Abundances are quoted in number densities
96: throughout the paper, in the standard scale where A(H)~=~12.0.} This value
97: significantly revises downward the earlier standard value
98: (A(Ar)$_{\odot} = 6.55$, \cite{SUN89}; A(Ar)$_{\odot} = 6.50$, \cite{lodders08}),
99: mostly because of the downward revision of the solar oxygen abundance
100: \citep{allende01, asplund04}.
101:
102: Because of the shortcomings of solar studies, it appears that abundance measurements
103: in the solar neighborhood may shed more light on the argon abundance at near-solar metallicity.
104: Indeed, this approach is suggested by the similar case of neon, for which
105: recent analyses in B-type stars \citep{cunha06},
106: in the Orion nebula \citep{esteban04}, and in the interstellar medium (ISM) toward
107: the Crab nebula \citep{kaastra07}, provide very consistent results, Ne/O~=~0.26$\pm$0.01,
108: and therefore indicate that the currently adopted solar value, (Ne/O)$_\odot$=0.15,
109: most likely underestimates the actual neon abundance in the solar photosphere and in the solar
110: neighborhood. Defining a reliable set of standard chemical
111: abundances is an essential undertaking, because abundance studies are a common tool used
112: to study physical processes which are revealed by changes in the chemical composition
113: of various astrophysical media, for instance including chemical anomalies in the
114: solar corona, in the ISM (e.g., depletion into grains, ISM ionization),
115: or in stellar photospheres (e.g., mixing of nucleosynthesis products).
116:
117: The optical spectra of early-type stars reveal a number of unblended, weak
118: \ion{Ar}{2} lines. Despite appearing to be the most fitting targets from which to derive
119: the argon abundance in the solar neighborhood, very limited abundance work on argon
120: in early-type stars has been carried out to this day.
121: %%%Trundle et al. (2001) analyzed argon, along with other elements, in a peculiar metal-rich star.
122: %%%THierry, see if you think it is ok to add a reference Trundle et al (2001) work here or somewhere else?
123: %%%She sent us non-published eqws for Ar II lines from this paper which helped us in the beginning
124: %%%of our analysis.
125: \cite{keenan90} analyzed two \ion{Ar}{2} lines in the blue spectrum of 5 bright B-type stars
126: in the field, assuming
127: Local Thermodynamic Equilibrium (LTE), and they derived an argon abundance, A(Ar)$ = 6.49\pm0.1$.
128: \cite{holmgren90} supplemented this study with an investigation of the importance of departures from LTE,
129: which were found to be small. The NLTE result (A(Ar)$ = 6.50\pm0.05$) thus does not
130: differ significantly from the LTE analysis; these NLTE calculations, however, were
131: based upon non-blanketed model atmospheres.
132:
133: In this paper we report on the results of a new NLTE analysis of a sample of
134: 10 main-sequence B-type stars which are members of the young Orion association.
135: The argon abundances are derived from NLTE line-blanketed model atmosphere
136: calculations. We argue that
137: the argon abundances obtained for this young population
138: may well represent the actual argon content of the solar neighborhood and that
139: the Orion B stars can be used as good proxies to define the argon abundance in the
140: solar photosphere.
141: Sect.~2 briefly presents the new observational material,
142: while the NLTE calculations are detailed in Sect.~3. We discuss our
143: results in Sect.~4, comparing them to other determinations of the argon abundance
144: in different sites. In our conclusions (\S5), we propose a new
145: reference value for the argon abundance in the solar neighborhood.
146:
147: % -----------------------------------------------------------------------------
148:
149: \section{OBSERVATIONS AND DATA REDUCTION}
150:
151: The sample under study consists of 10 main-sequence early B-type stars, which
152: have been previously analyzed in our recent work devoted to neon \citep{cunha06}.
153: New spectra were obtained on the nights of 2007 January 29/30, February
154: 5/6, and March 2/3, using the ARC echelle spectrograph on the 3.5~m
155: telescope at the Apache Point Observatory. The ARC echelle covers the
156: entire visible spectrum on a single 2048x2048 detector with a resolving power
157: of $R\approx$ 35,000. Exposure times for the different targets are listed in
158: Table~\ref{tab:exp}. Because of partly cloudy conditions, some stars were
159: observed repeatedly in order to achieve a signal-to-noise ratio of at least 100 in
160: the co-added spectra.
161:
162: The data were reduced and the spectra extracted following fairly standard
163: procedures. Bias frames and flat field exposures were combined; flat
164: frames included frames with a blue filter to insure a high signal-to-noise ratio
165: across the entire spectrum.
166: Individual orders were extracted from both the flats and
167: the object frames, after subtraction of a relatively small scattered light
168: component. Some care was taken with the extraction parameters because of
169: the relatively narrow width of the orders in the spectrograph. The
170: extracted spectra were flattened using the one dimensional extracted
171: flats. Wavelength calibration was performed with Th-Ar exposures,
172: and applied to the object spectra. Noise spectra were obtained
173: using the known gain and readout noise of the detector. The extracted spectra
174: were coadded when multiple observations were available. Finally,
175: the individual orders were combined, using the noise spectra to properly
176: weight the different orders in regions where there was spectral overlap.
177: Sample spectra of all target stars in the region around $\lambda$ 4429 \AA\
178: are shown in Fig.~\ref{fig1}.
179:
180: % ------------------------------------------------------------------------------------
181:
182: \section{NON-LTE ABUNDANCE CALCULATIONS}
183:
184: The NLTE calculations have been performed with the NLTE model stellar atmosphere
185: code TLUSTY \citep{TLUSTY88, TLUSTY95}.\footnote{See
186: http://nova.astro.umd.edu.} The TLUSTY model atmospheres assume
187: hydrostatic and radiative equilibria, and incorporate all essential sources of
188: opacity while allowing for departures from LTE. A large grid of NLTE fully
189: line-blanketed model atmospheres of early B-type stars has been made available
190: recently (BSTAR2006; Lanz~\& Hubeny 2007). The present calculations follow
191: the BSTAR2006 models, but for larger Ne~{\sc i-ii}\ model atoms \citep{cunha06}
192: and for adding extensive Ar~{\sc i-iv}\ model atoms (see \S\ref{modelAtom})
193: in the NLTE calculations. The NLTE argon calculations thus benefit from an
194: extensive and detailed NLTE treatment of all background opacity sources.
195:
196: We have adopted for the new models the same stellar parameters of the target
197: stars as \cite{cunha06}, see Table~\ref{tab:resul}. The stellar parameters
198: were derived by \cite{cunha92}, who provided a detailed discussion about
199: uncertainties. We examine their implications for the argon abundances
200: in \S\ref{abundances}. Like in BSTAR2006, a constant microturbulent velocity
201: of 2 km s$^{-1}$ was assumed in the NLTE model atmosphere calculations.
202:
203: \setcounter{footnote}{3}
204:
205: NLTE synthetic spectra were then calculated with the spectrum synthesis code
206: SYNSPEC\footnotemark\ using the NLTE level populations for all levels
207: included in the TLUSTY models. We have adopted in this step a microturbulent velocity
208: of 5 km s$^{-1}$ that is typical for early B-type main-sequence stars. We
209: stress, however, that the \ion{Ar}{2} lines analyzed in all 10 stars
210: are relatively weak and, therefore, the results are quite insensitive
211: to the microturbulence parameter (see \S\ref{abundances}).
212:
213: \subsection{Argon model atom}
214: \label{modelAtom}
215:
216: The argon model atom consists of 71 levels of \ion{Ar}{1}, 54 levels of \ion{Ar}{2},
217: 44 levels of \ion{Ar}{3}, 36 levels of \ion{Ar}{4}, plus the ground state of
218: \ion{Ar}{5}. Because the level structure of \ion{Ar}{1} is better described with
219: intermediate $jK$-coupling (similarly to \ion{Ne}{1}), the \ion{Ar}{1} model
220: atom was constructed to account for fine structure. It includes the lowest
221: 71 fine-structure levels (up to the $6s$ and $6p$ levels,
222: $E < 121,500$\,cm$^{-1}$). Experimental energy level data and line oscillator
223: strengths were extracted from the Atomic and Spectroscopic Database at
224: NIST \citep{NIST}.\footnote{See http://physics.nist.gov/PhysRefData/ASD/index.html.}
225: The photoionization cross-sections were taken from the Opacity Project (OP) database
226: Topbase \citep{topbase}.\footnote{See http://vizier.u-strasbg.fr/topbase.}
227: Since OP calculations assume $LS$-coupling, we followed a procedure analogous
228: to the case of \ion{Ne}{1} \citep{seaton98} for relating fine-structure levels
229: to $LS$-multiplet levels and thus for assigning the appropriate photoionization
230: cross-section to each level. The collisional excitation rates were considered using
231: the \cite{vanrege62} formula, including a modification for neutral atoms (as also
232: used by \cite{auer73}); for collisional ionization the Seaton formula was used
233: (for a synopsis of expressions, see \cite{TLUSTY88}).
234:
235: For the Ar~{\sc ii-iv}\ ions, we closely followed the approach described in
236: \cite{OS02} regarding the treatment and inclusion of atomic data in model atmosphere
237: calculations. In particular, the bulk of the data are taken from Topbase. The level
238: energies were updated with the more accurate experimental values whenever available
239: in the NIST ASD database. For a detailed description of the process of setting up
240: the atomic data, the reader should refer to \S4 of \cite{OS02}. In particular, we
241: include in each model ions all the energy levels which are in Topbase and below
242: the respective ionization limit, that is, typically up to $n=10$.
243: The \ion{Ar}{2} model atom includes individually the lowest 42 levels
244: up to $4d~ ^2$D ($E < 193,000$\,cm$^{-1}$). The 206 higher levels are grouped
245: into 12 superlevels, 6 in the doublet and 6 in the quartet systems.
246: The \ion{Ar}{3} model atom includes the lowest 27 levels
247: up to $3d~ ^3$P$^o$ ($E < 215,000$\,cm$^{-1}$). The 349 higher levels are grouped
248: into 17 superlevels, 6 in the singlet, 6 in the triplet and 5 in the quintet
249: systems. The \ion{Ar}{4} model atom includes the lowest 22 levels
250: up to $4s~ ^2$D ($E < 270,000$\,cm$^{-1}$), and the first 3 sextet levels.
251: The 322 higher levels are grouped
252: into 10 superlevels, 5 in the doublet and 5 in the quartet systems. Five
253: excited sextet levels, below the ionization limit, are grouped into an
254: additional superlevel.
255:
256: \subsection{Abundances from selected Argon lines}
257: \label{abundances}
258:
259: The optical spectra of early B-type stars contain a number of unblended \ion{Ar}{2} lines.
260: The sample of 11 \ion{Ar}{2} lines analyzed in this study is presented in Table~\ref{tab:line}.
261: The \ion{Ar}{2} $gf$-values for these transitions were extracted from \cite{bennett65}
262: via the NIST ASD database.
263: The accuracy assessment of these oscillator strengths by NIST is very high, with
264: 7 values estimated to be as accurate as 3\%\, while the 4 other lines have
265: estimated uncertainties smaller than 10\%. The linelist used in computation of the synthetic spectra
266: around each of the \ion{Ar}{2} transitions is an update of the \cite{CD23} linelist and
267: is available from the TLUSTY website.\footnotemark[4]
268:
269: Argon abundances were derived from best fits between synthetic and observed spectra.
270: Sample fits for 6 \ion{Ar}{2} lines ($\lambda \lambda$ 4426; 4430; 4735; 4764; 4806 and 5062 \AA)
271: in the star HD35299 are shown in Fig.~\ref{fig2}.
272: The best-fit synthetic profiles for each line
273: were calculated with the argon abundances indicated in the different panels of the figure.
274: The sensitivity of the fits to changes in the argon abundance by $\pm$ 0.1 dex is
275: illustrated for the \ion{Ar}{2}$\lambda$4426 line in the top left panel of Fig.~\ref{fig2}.
276: The final argon abundances for the target stars are given in Table~\ref{tab:resul}. They
277: are the averages of the abundances derived from the individual \ion{Ar}{2} lines.
278: The corresponding standard deviations represent the line-to-line scatter, which may be
279: attributed in part to the remaining uncertainties on the $f-$values (up to 0.05~dex).
280: For completeness, we also show in Table~\ref{tab:resul} the oxygen abundances
281: obtained previously for the target stars, because coronal studies provide the abundance ratio Ar/O.
282: The oxygen abundances are extracted from \cite{cunha94}; see \cite{cunha06} for a discussion of
283: consistency checks between these oxygen abundances and TLUSTY. Abundances are listed
284: as number densities in the standard logarithmic scale, where the hydrogen abundance is A(H)=12.
285:
286: The uncertainties in the derived argon abundances can be evaluated by investigating abundance
287: changes resulting from model atmospheres computed with modified stellar parameters. The error
288: budget in the argon abundances is dominated by the uncertainties in $T_{\rm eff}$. The adopted
289: effective temperatures are derived from Str\"omgren photometry; \cite{cunha94} carefully assessed
290: the uncertainties to $\delta T_{\rm eff}$= $\pm$3$\%$ at most. We verified that the adopted
291: $T_{\rm eff}$ values and uncertainties are consistent with temperature-sensitive spectroscopic
292: diagnostics such as the \ion{Si}{3}$\lambda\lambda$4553, 4568, 4575 lines. These uncertainties
293: translate to an error on the derived abundances, $\delta$A(Ar)= $\pm$0.08 dex. On the other hand,
294: the uncertainty on gravity, $\delta$log g= $\pm$0.1 dex, has a very limited effect on the abundances,
295: $\delta$A(Ar)= $\mp$0.01 dex.
296: %% The abundance
297: %% differences caused by changing the effective temperature by $\delta T_{\rm eff}$= $\pm$3$\%$ and the surface
298: %% gravity by $\delta$log g= $\pm$0.1 dex (typical uncertainties in the adopted stellar parameters;
299: %% \cite{cunha92, cunha94}) correspond to $\delta$A(Ar)= $\pm$0.08 dex and $\mp$0.01 dex, respectively.
300: The errors due to uncertainties in the adopted microturbulent velocities are small given that
301: the sample \ion{Ar}{2} lines are all reasonably weak. A change in microturbulence of 1.5~km/s
302: results in $\delta$A(Ar)= $\pm$0.01 dex. The total error from uncertainties in these stellar parameters
303: is $\pm$0.08 dex, using a quadratic error summation.
304:
305: % -----------------------------------------------------------------------------
306:
307: \section{DISCUSSION}
308:
309: The argon abundance cannot be measured in the photospheres of the Sun and
310: of other cool stars. We discuss and compare argon abundances derived in
311: different sites, including the solar corona, {\sl in-situ\/} measurements in the
312: solar system, the interstellar medium (ISM), \ion{H}{2} regions and planetary
313: nebulae, along with our analysis of early-type stars. The most
314: representative results are discussed in the next sections. This comparison provides
315: for a reliable determination of the abundance of argon in the solar
316: neighborhood, which can then serve as a reference point for further studies.
317: A similar overview of argon abundance determinations has been recently presented by
318: \cite{lodders08}.
319:
320: \subsection{Argon in early-type stars and departures from LTE}
321:
322: The weighted (as well as straight) average of the NLTE abundances measured
323: in 10 early B stars of the Orion association, and listed in Table~\ref{tab:resul},
324: is $<$A(Ar)$> = 6.66 \pm 0.06$. All the \ion{Ar}{2} lines give a consistent result,
325: as indicated by the line-to-line scatter for each star which is always smaller than
326: 0.1~dex, and the abundances span a very tight range
327: ($6.59\leq$ A(Ar) $\leq 6.76$). The error budget from uncertainties in stellar parameters
328: is $\pm$0.08~dex. Furthermore, the available
329: $f$-values are of high accuracy, generally better than 3\%. We therefore
330: believe that the accuracy of the derived argon abundance is likely of the order
331: of 0.1~dex or better.
332:
333: \cite{keenan90} carried out an early, but very limited, study of argon in 5 nearby field
334: B-type stars. They observed two \ion{Ar}{2} lines ($\lambda\lambda$4590, 4658).
335: One star was excluded from the sample because only \ion{Ar}{2}$\lambda$4590 was marginally
336: detected. Argon abundances were derived from equivalent widths and LTE model atmospheres.
337: Results from a second star (HR~1350) were also excluded because the derived abundance
338: differed by 0.25~dex relative to the 3 other stars, which was attributed to larger
339: uncertainties in the stellar parameters. They derived an abundance, A(Ar)$ = 6.49 \pm 0.1$,
340: based on 4 measurements only. Their value is 0.17~dex lower and barely
341: consistent with our results.
342: However, if we retain the abundance derived for HR~1350 that they excluded, and if we correct
343: for small systematic differences in the $f$-values (0.024 and 0.046~dex for
344: $\lambda\lambda$4590, 4658, respectively), we find a better agreement
345: (A(Ar)$ = 6.54 \pm 0.15$). The larger scatter seems to be a more likely estimate of
346: the actual uncertainties of their results.
347:
348: \cite{holmgren90} supplemented Keenan et~al.'s study and evaluated the importance
349: of departures from LTE. They applied the NLTE line formation codes DETAIL and SURFACE
350: \citep{giddings81, butler84} and found that the corrections from LTE are
351: very small. Holmgren et~al. derived an argon abundance, A(Ar)$ = 6.50 \pm 0.05$. However,
352: their NLTE analysis is based on unblanketed NLTE model atmospheres, while Keenan et~al.
353: used metal line-blanketed LTE models. Their comparison therefore amalgamates two
354: different effects, namely direct NLTE effects in argon and differences arising from
355: the changes in atmospheric structure between blanketed and unblanketed models.
356:
357: To ascertain better the extent of departures from LTE in argon, we have repeated our
358: analysis of the target star HD35299 using a Kurucz LTE blanketed model atmosphere and
359: LTE spectrum synthesis. We can then make a direct comparison of two consistent analyses with and without
360: the assumption of LTE, but based on fully line-blanketed model atmospheres in both cases.
361: The LTE abundances obtained for the sample \ion{Ar}{2} lines are not significantly different
362: from the NLTE values: $<\delta$(NLTE - LTE)$>\approx$ 0.02 $\pm$ 0.01~dex, supporting
363: Holmgren et~al.'s earlier claim that departures from LTE are small.
364: LTE argon abundances can also be derived using the same NLTE model atmosphere, but only setting
365: argon populations to their LTE values. This second approach may isolate the non-LTE
366: effects in argon since the same model structure is adopted. In this case, our results
367: show that NLTE effects are also small, the NLTE abundances being larger than the LTE values
368: by $\approx$~0.03~dex. These results therefore support the earlier NLTE study of \cite{holmgren90}.
369: We stress however the importance of performing extensive NLTE calculations
370: when aiming at determining abundances with accuracies as high as 0.1~dex or better. Indeed,
371: even small corrections to LTE may easily introduce systematic errors of the order of
372: the expected accuracy.
373:
374: %%% At physical conditions present in the line-forming region in early B star
375: %%% atmospheres, a large majority of argon is found in the Ar$^{2+}$ ground state.
376: %%% As already pointed out by \cite{keenan90}, small NLTE effects are therefore
377: %%% expected because the \ion{Ar}{2} optical lines are transitions between highly-excited
378: %%% Ar$^+$ levels that are strongly coupled by collisions to the Ar$^{2+}$ ground state.
379:
380: We examine now the reasons why the departures from LTE are small for \ion{Ar}{2}
381: lines while quite significant for \ion{Ne}{1} lines \citep{sigut99, cunha06}. The
382: top panels of Fig.~\ref{fig3} show that the typical NLTE ionization shift toward higher
383: ionization is limited. Because of the high ionization potentials of neon and
384: argon ions, the ionization thresholds are found in the extreme ultraviolet where B stars
385: have a low flux. In this spectral range, the opacity is dominated by hydrogen and helium. The
386: bound-free transition from the ground states of neon and argon are optically thick and
387: in detailed radiative balance in the region of formation of neon and argon optical lines.
388: Consequently, the contribution of radiative photoionization to departures from LTE
389: remains small. Rather than by a global ionization shift, the line strengths are then better described by the
390: departure coefficients, $b_i = n_i / n_i^\star$, where $n_i$ is the population of level $i$
391: and $n_i^\star$ its LTE counterpart. The bottom panels of Fig.~\ref{fig3} show a very
392: different behavior for lines in the $3s - 3p$ \ion{Ne}{1} and in the $4s - 4p$ \ion{Ar}{2}
393: systems to which the analyzed lines belong. In both cases, the optical lines behave mostly
394: like lines in a 2-level atom because the resonance lines are optically thick at the depth
395: of formation of the optical lines. In the neon case, the (red) lines provide a downward channel,
396: overpopulating the lower levels, thus resulting in stronger absorption and yielding thus lower
397: abundances \citep{cunha06}. On the other hand, the Ar$^+$ levels are collisionally coupled
398: to the Ar$^{2+}$ ground state, as indicated by the fact that all Ar$^+$ levels show similar departure
399: coefficients. In first approximation, the ratio of the line source function to the Planck function
400: behaves like $S/B\propto b_u/b_l$. In the line forming region, our calculations typically give
401: $b_u/b_l\approx 0.98$, thus yielding only a small correction to LTE.
402: %%We attribute basically the different behavior of neon and argon levels to the
403: %%lower ionization potential of argon compared to neon, which results in stronger collisional
404: %%coupling to the dominant ion's ground state.
405: Finally, the \ion{Ne}{1} and \ion{Ar}{2}
406: resonance lines are formed in the outermost atmospheric layers,
407: around $\tau_{\rm Ross}\approx 10^{-4}$, where they provide a downward radiative channel,
408: leading to an overpopulation of the ground states and underpopulation of the excited
409: states.
410:
411: Among hot stars, \cite{werner07} have recently reported the identification of
412: \ion{Ar}{7}~$\lambda$1063\ in very hot central stars of planetary nebulae and
413: (pre-)white dwarfs. Because of large remaining uncertainties on stellar parameters,
414: and because of processes such as nucleosynthesis and elemental diffusion occurring
415: in late evolutionary stages, their derived
416: argon abundances cannot be used as a reference value. Several stars
417: however show an argon abundance close to the value that we derived for B-type stars,
418: hence supporting the idea that there is no large production of argon in the
419: previous AGB evolution and that argon remains essentially unchanged in these stages.
420:
421: \subsection{Nebular emission abundances}
422:
423: \cite{esteban04} derived emission line abundances for several elements in the Orion nebula.
424: In particular, they obtained A(Ar)= 6.62 $\pm$ 0.05 when adopting their preferred value for
425: the temperature fluctuation parameter, $t^{2}$=0.02. The argon abundance in the Orion nebula
426: derived earlier by \cite{peimbert77} also ranges between A(Ar)=6.6 (with $t^{2}$=0)
427: and A(Ar)=6.7 (for $t^{2}$=0.35). There is therefore a good indication of a consensus
428: regarding the nebular argon abundance in Orion. Furthermore, the nebular analysis is
429: fully consistent with our results from B stars in the Orion association (A(Ar)= 6.66 $\pm$ 0.06),
430: even with the tight uncertainties of the two studies.
431:
432: Argon abundances can also be derived from planetary nebulae (PNe). \cite{stanghellini06}
433: present argon abundances for a large sample of galactic PNe
434: and find an average argon abundance for the round nebulae (representing the lower mass
435: and probably local population) of $<$A(Ar)$>$~=~6.11. These argon abundances rely on optical
436: argon lines and are deemed as more uncertain in their study.
437: \cite{pottasch06} give
438: a summary of PN abundances from measurements recorded with the {\sl Infrared Space
439: Observatory\/}. The Galactic chemical gradient is apparent from their dataset
440: when plotting abundances as a function of galactocentric distances. Therefore, for comparison
441: with the Orion results, we have retained
442: only the PNe which are within 1~kpc from the Sun. The average argon abundance of the 7 PNe
443: satisfying this criterion is A(Ar) = 6.50$^{+0.14}_{-0.22}$. The average value is 0.12~--~0.16~dex
444: lower than the results from Orion. However, the scatter is much larger and we cannot conclude
445: that there is a significant difference. The results of our study therefore support \cite{pottasch06} conclusion
446: that the argon abundance is essentially the same in young objects (B stars, Orion nebula) and
447: in older PNe in the solar neighborhood.
448:
449: \subsection{ISM absorption studies}
450:
451: Because of its high ionization potential, argon is mostly in the neutral state
452: in the ISM. Argon abundances can be derived from column densities
453: measured from the far-ultraviolet resonance lines, \ion{Ar}{1} $\lambda\lambda 1048, 1067$.
454: \cite{sofia98} detected these \ion{Ar}{1} lines towards several early-type stars using
455: the interstellar medium absorption profile spectrograph during the first {\sl ORFEUS-SPAS\/}
456: shuttle mission in 1993 September. They found deficiencies in \ion{Ar}{1}/\ion{H}{1}
457: ranging from 0.2 to 0.6 dex, relative to the Ar/H abundance ratio in the Sun and in B stars
458: \citep{SUN89, keenan90}. They argued that this result does not provide an
459: indication of argon depletion onto dust grains, but it rather reflects selective argon
460: ionization relative to hydrogen because of the larger \ion{Ar}{1} photoionization
461: cross-section.
462:
463: With the launch of the {\sl Far Ultraviolet Spectroscopic Explorer\/}, many
464: more line-of-sights in the ISM have been observed.
465: \cite{lehner03} carried out an analysis of 23 line-of-sights
466: toward white dwarfs and derived argon abundances in the local ISM. They confirmed
467: deficiencies of the order of 0.4~dex, as found earlier by \cite{sofia98}. Moreover,
468: the idea that these deficiencies actually map the local ISM ionization is well supported
469: by an increase of the abundance scatter outside the local Bubble.
470: It is therefore essential to have a
471: good reference value of the local argon abundance for ISM ionization studies, but
472: \ion{Ar}{1} column densities are thus of limited use for accurately determining
473: the reference argon abundance in the solar neighborhood.
474:
475: \subsection{Coronal vs. photospheric abundances}
476:
477: The standard solar argon abundance recommended by \cite{AGS05},
478: A(Ar)= 6.18 $\pm$ 0.08, has been derived
479: from measured abundance ratios in the solar corona and solar energetic particles.
480: However, significant differences in the chemical composition of the solar corona relative
481: to the photosphere were noticed very early on. These differences were found to correlate
482: with the first ionization potential of the studied elements, the so-called ``FIP effect''.
483: While there is still no agreed theoretical explanation of the FIP effect, the current general
484: consensus is that elements with lower FIP ($\la 10$\,eV) are enhanced in the corona
485: by a factor 4 to 6, while elements with higher FIP ($\ga 10$\,eV) have essentially
486: photospheric abundances \citep{favata03}. However, most coronal studies based
487: for instance on \ion{Ar}{14}~$\lambda 4412$ \citep[e.g., ][]{young97}
488: or on helium-like argon lines observed in X-ray flares \citep{veck81, phillips03}
489: indicate an argon abundance, typically in a range between $6.2\la$~A(Ar)~$\la 6.4$,
490: that is lower than in Orion stars and nebula. The Orion results thus support the idea
491: that the solar corona exhibits significant underabundances of the two elements
492: with a high FIP ($\ga 15$\,eV). However, we cannot define the origin of these underabundances.
493: The FIP effect might either cover
494: a whole range, from overabundances for low FIP elements, to photospheric abundances
495: at intermediate FIP and underabundances for elements with the higest FIP, or
496: different physical processes might be in play and the behavior of low FIP and high FIP
497: elements might be unrelated.
498:
499: Since the neon and argon emission lines observed in the coronal and
500: chromospheric spectra of the Sun are potentially affected by the FIP effect,
501: \cite{feldman03} argued that it would be much preferable to observe and analyze
502: unmodified photospheric plasma. During a low-altitude impulsive flare,
503: photospheric plasma raised to flare temperature was observed by {\sl Skylab\/} \citep{feldman90}.
504: From the high measured plasma density and a Mg/O abundance ratio similar to
505: the photospheric value, Feldman~\& Widing established the origin of the emission.
506: They determined the Ar/Mg and Ne/Mg abundance ratios, yielding A(Ar) = 6.57 $\pm$ 0.12 and
507: A(Ne) = 8.08 $\pm$ 0.10,
508: with A(Mg) = 7.53 from \cite{AGS05}. These values are in excellent agreement
509: with the argon and neon abundances measured in B stars and in the Orion nebula.
510:
511: Argon abundances can also be measured in the corona of other cool stars where the FIP effect
512: might not be as prevalent as in the Sun. \cite{maggio07} recently reported on a large study
513: of X-ray bright pre-main sequence (PMS) stars in Orion. They derived an average argon abundance of
514: A(Ar) = 6.51$^{+0.26}_{-0.17}$. The larger scatter implies that this result remains consistent
515: with the abundance measured in Orion B-type stars and, therefore,
516: does not support an underabundance of argon in the corona of Orion PMS stars similar to the
517: underabundance found in the solar corona.
518:
519:
520: \subsection{{\sl In-situ\/} measurements in the Solar system}
521:
522: Argon abundances have also been measured from energetic particles in the solar
523: wind, from lunar soils, and meteorites. \cite{reames98} contended that the most
524: comprehensive measurements of element abundances in the solar corona comes from
525: measurements of energetic particles that have been ejected during coronal mass
526: ejections. Because of their origin, we expect however that these abundances to show the
527: FIP effect. \cite{reames98} reported an abundance ratio Ar/O = 0.0033~$\pm$~0.0002.
528: Using \cite{AGS05} oxygen abundance, this ratio translates to A(Ar) = 6.18 $\pm$ 0.03.
529: \cite{cerutti74} reported a similarly low Ar/O ratio in lunar soils, yielding
530: A(Ar) = 6.26 $\pm$ 0.1. Finally, C1 chondrites reveal strong depletion because of
531: the volatile nature of noble gases. For instance, \cite{AGS05} list an argon abundance
532: in meteorites that is 6 to 7 orders of magnitude lower than all other measurements.
533:
534: \subsection{Abundances in the Solar neighborhood}
535:
536: Table~\ref{tab:sites} summarizes the argon abundances measured in different sites
537: in the Solar system and in the solar neighborhood, using a broad variety of methods.
538: Excluding meteorites, argon abundances span a range of 0.6~dex (a factor of 4),
539: 6.1~$\la$~A(Ar)~$\la 6.7$. On one hand, solar corona and {\sl in-situ} measurements
540: in the Solar system yield a low value (6.2~$\la$~A(Ar)~$\la 6.4$, or A(Ar) = 6.18 as
541: adopted by Asplund et~al.). On the other hand, Orion B-type stars and nebula provide
542: a tight agreement with emerging photospheric material during an impulse flare,
543: 6.57~$\la$~A(Ar)~$\la 6.66$. What is therefore the best value to represent the argon
544: abundance in the Solar system? Can we use Orion
545: stars and nebula as proxies to the Sun in order to determine standard abundances
546: in the solar neighborhood?
547:
548: Table~\ref{tab:solab} lists the abundances of light elements in the Sun, Orion B stars,
549: and Orion nebula, showing an excellent agreement for all elements if we adopt
550: \cite{feldman03} values for neon and argon in the solar photosphere. This table has
551: been constructed from representative studies that we view as current state-of-the-art works
552: in different objects, and without intending to review comprehensively all abundance studies
553: in the solar neighborhood (in particular for C, N, and O). We aim at putting
554: the present study in a broader context and discuss its relevance to the Ar solar abundance question.
555: First, we need to emphasize
556: that the agreement between the nebular and the photospheric analyses for all five
557: elements (C, N, O, Ne, and Ar) cannot be fortuitous. It rather demonstrates the degree
558: of reliability that can be achieved today in abundance analyses. We believe that we may
559: thus reasonably claim that the accuracy of these abundances is 0.1~dex or better.
560: Second, the long-standing puzzle regarding the C, N, O abundances in the Sun and in
561: young OB stars appears now to have been resolved. In \cite{SUN89} reference work, the CNO abundances
562: were higher in the Sun than in OB stars, in an apparent contradiction with
563: the older age of the Sun. Recent studies
564: have now led to revise downward the solar abundances, and \cite{AGS05} recommend
565: C, N, O solar abundances that are very close to the values derived for Orion stars and
566: nebula. Therefore, these two arguments provide a robust support to the idea
567: that the composition of Orion's young material (in stars and in the nebula) is
568: very similar to the solar composition; hence, Orion's studies may be good proxies
569: to determine the composition in the solar neighborhood, in particular for neon and
570: argon too.
571:
572: Let us assume on the other hand that the coronal abundances are representative of
573: the standard composition in the Solar system. This assumption thus implies that
574: the chemical enrichment over the last 5~Gyr would have been a factor of 2 for neon,
575: and a factor of 2 to 3 for argon, while C, N, and O remained virtually unchanged.
576: Analyses of PNe do not support such an enrichment of neon and argon over the last
577: few billion years. Within 1~kpc of the Sun, 7 PNe have mean abundances of
578: $<$A(O)$>$~=~8.66, $<$A(Ne)$>$~=~8.12 and $<$A(Ar)$>$~=~6.50 \citep{pottasch06},
579: in good agreement with Orion's results (see Table~\ref{tab:solab}). The 7 PNe
580: reveal however an enrichment in carbon and nitrogen resulting from nucleosynthesis
581: in the late evolution of their parent stars.
582: We therefore conclude that neon and argon do not reveal a strong chemical enrichment
583: in the solar neighborhood during the last few billion years, and that the
584: depletion of neon and argon in the solar corona most likely arise from
585: yet little-understood processes in the corona, such as the FIP effect.
586:
587:
588: % -----------------------------------------------------------------------------
589:
590: \section{CONCLUSIONS}
591:
592: We have completed a new analysis of \ion{Ar}{2} lines in the blue spectrum
593: of 10 B main-sequence stars in the Orion association. The analysis involved
594: a fully consistent NLTE treatment based on our NLTE line blanketed model
595: atmospheres. We have derived a mean argon abundance which is in excellent agreement
596: with recent results for the Orion nebula \citep{esteban04}. However, the derived
597: abundance is significantly higher (0.45~dex) than the standard solar value recommended in
598: the recent compilation of \cite{AGS05}. Based on a review of available argon
599: abundance determinations in various sites in the solar neighborhood, we argue
600: that the Orion's results much more likely represent the abundance
601: of argon (and neon) in the solar neighborhood, while the lower coronal abundances
602: reflect a depletion due to physical processes in the corona such as the FIP effect.
603:
604: Therefore, we propose to adopt the following values as new reference abundances in
605: the neighborhood of the Sun in the Galactic disk:
606: \begin{center}
607: A(Ne)~=~8.09~$\pm$~0.06 \\
608: A(Ar)~=~6.63~$\pm$~0.10 \\
609: \end{center}
610: The argon abundance is in excellent agreement with Lodders (2008) recent work,
611: in which she adopts for the solar system (protosolar) abundance, A(Ar)~=~6.57~$\pm$~0.10.
612: These abundances yield the abundance ratios Ne/O~=~0.26 and Ar/O~=~0.009.
613: We believe that these values should provide valuable reference points
614: for various fields of investigations, from the solar corona to the ionization of
615: the interstellar medium.
616:
617: \acknowledgements
618: We thank Martin Asplund for inquiring about the argon abundances in B stars,
619: Carrie Trundle for sending equivalent widths of \ion{Ar}{2} lines measured in the
620: star HD~135485 and Letizia Stanghellini for discussions.
621: The work reported here is supported in part by the National Science
622: Foundation through AST03-07534, AST03-07532, AST06-46790 and NASA through NAG5-9213.
623:
624:
625: % ------------------------------------------------------------------------------------
626:
627: \begin{thebibliography}
628:
629: \bibitem[Allende Prieto et~al.(2001)]{allende01} Allende Prieto, C., Lambert, D.~L.,
630: \& Asplund, M. 2001, \apj, 556, L63
631:
632: \bibitem[Anders~\& Grevesse(1989)]{SUN89} Anders, E. \& Grevesse, N. 1989,
633: Geochimica et Cosmochimica Acta, 53, 197
634:
635: \bibitem[Asplund et~al.(2004)]{asplund04} Asplund, M., Grevesse, N., Sauval, A. J.,
636: Allende Prieto, C., \& Kiselman, D. 2004, \aap, 417, 751
637:
638: \bibitem[Asplund et~al.(2005)]{AGS05} Asplund, M., Grevesse, N., \& Sauval, A. J.
639: 2005, in {\sl Cosmic Abundances as Records of Stellar Evolution and Nucleosynthesis\/},
640: Ed. T.~G. Barnes~III~\& F.~N. Bash, ASP Conf. Ser., 336, 25
641:
642: \bibitem[Auer~\& Mihalas(1973)]{auer73} Auer, L. H., \& Mihalas, D. 1973,
643: \apj, 184, 151
644:
645: \bibitem[Bennett et~al.(1965)]{bennett65} Bennett Jr., W.~R., Kindlmann, P.~J.,
646: \& Mercer, G.~N. 1965, Appl. Opt., Suppl.~2 of Chemical Lasers, 34.
647:
648: \bibitem[Butler(1984)]{butler84} Butler, K. 1984, Ph.D. thesis, University of London
649:
650: \bibitem[Cerutti(1974)]{cerutti74} Cerutti, H. 1974, Ph.D. thesis, University of Bern
651:
652: \bibitem[Christensen-Dalsgaard(1998)]{christensen98} Christensen-Dalsgaard, J. 1998,
653: \ssr, 85, 19
654:
655: \bibitem[Cunha~\& Lambert(1992)]{cunha92} Cunha, K., \& Lambert, D. L. 1992,
656: \apj, 399, 586
657:
658: \bibitem[Cunha~\& Lambert(1994)]{cunha94} Cunha, K., \& Lambert, D. L. 1994,
659: \apj, 426, 170
660:
661: \bibitem[Cunha et~al.(2006)]{cunha06} Cunha, K., Hubeny, I., \& Lanz, T.
662: 2006, \apj, 647, L143
663:
664: \bibitem[Cunto et~al.(1993)]{topbase} Cunto, W., Mendoza, C., Ochsenbein, F.,
665: \& Zeippen, C. J. 1993, \aap, 275, L5
666:
667: \bibitem[Esteban et~al.(2004)]{esteban04} Esteban, C., Peimbert, M., Garc\'ia-Rojas, J.,
668: Ruiz, M.~T., Peimbert, A., \& Rodr\'iguez, M. 2004, \mnras, 355, 229
669:
670: \bibitem[Favata~\& Micela(2003)]{favata03} Favata, F., \& Micela, G. 2003,
671: \ssr, 108, 577
672:
673: \bibitem[Feldman~\& Widing(1990)]{feldman90} Feldman, U., \& Widing, K.~G. 1990,
674: \apj, 363, 292
675:
676: \bibitem[Feldman~\& Widing(2003)]{feldman03} Feldman, U., \& Widing, K.~G. 2003,
677: \ssr, 107, 665
678:
679: \bibitem[Giddings(1981)]{giddings81} Giddings, J.~R. 1981, Ph.D. thesis, University of London
680:
681: %% \bibitem[Girard et~al.(2007)]{girard07} Girard, P., K\"oppen, J., \&
682: %% Acker, A. 2007, \aap, 463, 265
683:
684: \bibitem[Holmgren et~al.(1990)]{holmgren90} Holmgren, D.~E., Brown, P.~J.~F.,
685: Dufton, P.~L., \& Keenan, F.~P. 1990, \apj, 364, 657
686:
687: \bibitem[Hubeny(1988)]{TLUSTY88} Hubeny, I. 1988, Comput. Phys. Commun., 52, 103
688:
689: \bibitem[Hubeny \& Lanz(1995)]{TLUSTY95} Hubeny, I., \& Lanz, T. 1995,
690: \apj, 439, 875
691:
692: \bibitem[Kaastra et~al.(2007)]{kaastra07} Kaastra, J.~S., de~Vries, C.~P.,
693: Constantini, E., \& den~Herder, J.~W.~A. 2007, \aap\ (submitted)
694:
695: \bibitem[Keenan et~al.(1990)]{keenan90} Keenan, F.~P., Bates, B.,
696: Dufton, P.~L., Holmgren, D.~E., \& Gilheany, S. 1990, \apj, 348, 322
697:
698: \bibitem[Kurucz(1994)]{CD23} Kurucz, R. L. 1994, Atomic Line Data,
699: Kurucz CD-ROM 23 (Cambridge, Mass: SAO)
700:
701: \bibitem[Lanz \& Hubeny(2003)]{OS02} Lanz, T., \& Hubeny, I. 2003,
702: \apjs, 146, 417
703:
704: \bibitem[Lanz \& Hubeny(2007)]{BS06} Lanz, T., \& Hubeny, I. 2007,
705: \apjs, 169, 83
706:
707: \bibitem[Lehner et~al.(2003)]{lehner03} Lehner, N., Jenkins, E. B., Gry, C.,
708: Moos, H.~W., Chayer, P., \& Lacour, S. 1998, \apj, 595, 858
709:
710: %%\bibitem[Lodders(2003)]{lodders03} Lodders, K. 2003, \apj, 591, 1220
711:
712: \bibitem[Lodders(2008)]{lodders08} Lodders, K. 2008, \apj\ (in press)
713:
714: \bibitem[Maggio et~al.(2007)]{maggio07} Maggio, A., Flaccomio, E., Favata, F.,
715: Micela, G., Sciortino, S., Feigelson, E.~D., \& Getman, K.~V. 2007, \apj, 660, 1462
716:
717: \bibitem[Martin et~al.(1999)]{NIST} Martin, W. C., Sugar, J., Musgrove, A.,
718: Wiese, W. L., \& Fuhr, J. R. 1999, NIST Atomic Spectra Database
719: (Gaithersburg: NIST)
720:
721: \bibitem[Peimbert~\& Torres-Peimbert(1977)]{peimbert77} Peimbert, M., \&
722: Torres-Peimbert, S. 1977, \mnras, 179, 217
723:
724: \bibitem[Phillips et~al.(2003)]{phillips03} Phillips, K. J. H., Sylwester, J.,
725: Sylwester, B., \& Landi, E. 2003, \apj, 589, L113
726:
727: \bibitem[Pottasch~\& Bernard-Salas(2006)]{pottasch06} Pottasch, S.~R., \&
728: Bernard-Salas, J. 2006, \aap, 457, 189
729:
730: \bibitem[Reames(1998)]{reames98} Reames, D. V. 1998, \ssr, 85, 327
731:
732: \bibitem[Seaton(1998)]{seaton98} Seaton, M. J. 1998, J. Phys. B, 31, 5315
733:
734: \bibitem[Sigut(1999)]{sigut99} Sigut, T. A. A. 1999, ApJ, 519, 313
735:
736: \bibitem[Sofia~\& Jenkins(1998)]{sofia98} Sofia, U. J., \& Jenkins, E.~B.
737: 1998, \apj, 499, 951
738:
739: \bibitem[Stanghellini et~al.(2006)]{stanghellini06} Stanghellini, L., Guerrero, M.~A.,
740: Cunha, K., Manchado, A., \& Villaver, E. 2006, \apj, 651, 898
741:
742: \bibitem[Van Regemorter(1962)]{vanrege62} Van Regemorter, H. 1962, \apj, 136, 906
743:
744: \bibitem[Veck~\& Parkinson(1981)]{veck81} Veck, N. J., \& Parkinson, J. H.
745: 1981, \mnras, 197, 41
746:
747: \bibitem[Werner et~al.(2007)]{werner07} Werner, K., Rauch, T., \& Kruk, J. W.
748: 2007, \aap, 466, 317
749:
750: \bibitem[Young et~al.(1997)]{young97} Young, P. R., Mason, H. E., Keenan, F.~P.,
751: \& Widing, K.~G. 1997, \aap, 323, 243
752:
753: \end{thebibliography}
754:
755: % -----------------------------------------------------------------------------
756:
757: \clearpage
758: \begin{deluxetable}{lllll}
759: %\tabletypesize{\small}
760: \tablewidth{0pt}
761: \tablecaption{Observing log. \label{tab:exp}}
762: \tablehead{
763: \colhead{Star ID} & \colhead{Spectral Type} & \colhead{$V$} &
764: \colhead{UT Date} & \colhead{Exposure times [s]}
765: }
766: \startdata
767: HD 35039 & B2 IV-V & 4.70 & 2007-01-30 & 2x150 \\
768: HD 35299 & B1.5 V & 5.68 & 2007-01-30 & 2x300 \\
769: HD 35912 & B2 V & 6.38 & 2007-02-06 & 720 \\
770: \nodata & \nodata & \nodata & 2007-03-03 & 2x700 \\
771: HD 36285 & B2 IV-V & 6.32 & 2007-03-03 & 2x600 \\
772: HD 36351 & B1.5 V & 5.46 & 2007-02-06 & 330 \\
773: \nodata & \nodata & \nodata & 2007-03-03 & 500, 2x700 \\
774: HD 36591 & B1 IV & 5.33 & 2007-02-06 & 180 \\
775: HD 36959 & B1 V & 5.67 & 2007-02-06 & 2x270 \\
776: HD 37209 & B1 V & 5.72 & 2007-02-06 & 360 \\
777: HD 37356 & B2 IV-V & 6.18 & 2007-03-03 & 2x600 \\
778: HD 37744 & B1.5 V & 6.20 & 2007-02-06 & 480
779: \enddata
780: \end{deluxetable}
781:
782: % -----------------------------------------------------------------------------
783:
784: \clearpage
785: \begin{deluxetable}{crrc}
786: %\tabletypesize{\small}
787: \tablewidth{0pt}
788: \tablecaption{Selected \ion{Ar}{2} lines and oscillator strengths. \label{tab:line}}
789: \tablehead{
790: \colhead{Wavelength [\AA]} & \colhead{$E_{\rm exc}$ [cm$^{-1}$]} & \colhead{$\log gf$} &
791: \colhead{Accuracy\tablenotemark{a}}
792: }
793: \startdata
794: 4426.001 & 135\,085.9960Ê& 0.158 & A \\
795: 4430.189 & 135\,601.7336 & -0.174 & A \\
796: 4589.898 & 148\,620.1411 & 0.100 & A \\
797: 4657.901 & 138\,243.6442 & -0.236 & B \\
798: 4726.868 & 138\,243.6442 & -0.103 & A \\
799: 4735.906 & 134\,241.7392 & -0.108 & A \\
800: 4764.865 & 139\,258.3384 & -0.06\phantom{0} & B \\
801: 4806.020 & 134\,241.7392 & 0.210 & A \\
802: 4847.810 & 135\,085.9960 & -0.223 & B \\
803: 4879.864 & 138\,243.6442 & 0.246 & A \\
804: 5062.037 & 135\,601.7336 & -0.465 & B \\
805: \enddata
806: \tablenotetext{a}{Estimated accuracy is better than 3\% (A) or better than 10\% (B).}
807: \end{deluxetable}
808:
809: % -----------------------------------------------------------------------------
810:
811: \clearpage
812:
813: \begin{deluxetable}{lcccc}
814: \tablecaption{Argon and oxygen abundances. \label{tab:resul}}
815: \tablewidth{0pt}
816: \tablehead{
817: \colhead{Star} & \colhead{$T_{\rm eff}$} & \colhead{$\log g$} & \colhead{A(Ar)} & \colhead{A(O)}}
818: \startdata
819: HD 35039 & 20550 & 3.74 & 6.62 $\pm$ 0.06 & 8.60 \\
820: HD 35299 & 24000 & 4.25 & 6.70 $\pm$ 0.04 & 8.57 \\
821: HD 35912 & 19590 & 4.20 & 6.66 $\pm$ 0.05 & 8.70 \\
822: HD 36285 & 21930 & 4.40 & 6.59 $\pm$ 0.04 & 8.80 \\
823: HD 36351 & 21950 & 4.15 & 6.59 $\pm$ 0.05 & 8.76 \\
824: HD 36591 & 26330 & 4.20 & 6.73 $\pm$ 0.09 & 8.60 \\
825: HD 36959 & 24890 & 4.40 & 6.67 $\pm$ 0.07 & 8.76 \\
826: HD 37209 & 24050 & 4.15 & 6.63 $\pm$ 0.04 & 8.83 \\
827: HD 37356 & 22370 & 4.15 & 6.76 $\pm$ 0.04 & 8.67 \\
828: HD 37744 & 24480 & 4.40 & 6.65 $\pm$ 0.05 & 8.63 \\
829: \enddata
830: \end{deluxetable}
831:
832: % -----------------------------------------------------------------------------
833:
834: \clearpage
835:
836: \begin{deluxetable}{lll}
837: \tablecaption{Argon abundances in various sites. \label{tab:sites}}
838: \tablewidth{0pt}
839: \tablehead{
840: \colhead{A(Ar)} & \colhead{Sites, methods} & \colhead{References} }
841: \startdata
842: 6.66 $\pm$ 0.06 & B stars; NLTE & This study \\
843: 6.50 $\pm$ 0.05 & B stars; NLTE & \cite{holmgren90} \\
844: 6.54 $\pm$ 0.15\tablenotemark{a} & B stars; LTE & \cite{keenan90} \\ [2mm]
845: 6.62 $\pm$ 0.05 & Orion nebula, $t^2=0.022$ & \cite{esteban04} \\
846: 6.6~~--~~6.7 & Orion nebula, $t^2=0.0 - 0.35$ & \cite{peimbert77} \\[2mm]
847: 6.50 $^{+0.14}_{-0.22}$ & Planetary nebulae, $d\leq 1$\,kpc & \cite{pottasch06} \\ [2mm]
848: 6.1\phantom{0} $\pm$ 0.2 & ISM, \ion{Ar}{1}~$\lambda\lambda1048, 1067$ & \cite{sofia98, lehner03} \\ [2mm]
849: 6.62 $\pm$ 0.12 & Low-altitude impulsive flare, Ar/Mg & \cite{feldman90, feldman03} \\[2mm]
850: 6.45 $\pm$ 0.03 & Solar corona, X-ray flares, \ion{Ar}{17} & \cite{phillips03} \\
851: 6.23 $\pm$ 0.1 & Solar corona, \ion{Ar}{14}~$\lambda4412$ & \cite{young97} \\
852: 6.38 $^{+0.18}_{-0.30}$ & Solar corona, X-ray flares & \cite{veck81} \\ [2mm]
853: 6.51 $^{+0.26}_{-0.17}$ & PMS stars in Orion, corona, X-rays & \cite{maggio07} \\ [2mm]
854: 6.18 $\pm$ 0.03 & Solar energetic particles, Ar/O & \cite{reames98} \\
855: 6.26 $\pm$ 0.1 & Lunar soils, Ar/O & \cite{cerutti74} \\
856: \enddata
857: \tablenotetext{a}{Corrected as discussed in text.}
858: \end{deluxetable}
859:
860: % -----------------------------------------------------------------------------
861:
862: \clearpage
863:
864: \begin{deluxetable}{llllll }
865: \tablecaption{Abundances in the solar neighborhood. \label{tab:solab}}
866: \tablewidth{0pt}
867: \tablehead{
868: \colhead{Element} & \colhead{Sun} & \colhead{Orion B stars} & \colhead{Orion nebula} &
869: \colhead{PNe ($d <$~1\,kpc)} & \colhead{References} }
870: \startdata
871: He & 10.98 $\pm$ 0.02 & \nodata & 10.988 $\pm$ 0.003 & \nodata & 2,..., 7, ... \\
872: C & 8.39 $\pm$ 0.05 & 8.35 $\pm$ 0.05 & 8.42 $\pm$ 0.02 & 8.65$^{+0.16}_{-0.27}$ & 1, 4, 7, 8 \\
873: N & 7.78 $\pm$ 0.06 & 7.78 $\pm$ 0.07 & 7.73 $\pm$ 0.09 & 8.13$^{+0.14}_{-0.22}$ & 1, 4, 7, 8 \\
874: O & 8.66 $\pm$ 0.05 & 8.70 $\pm$ 0.09 & 8.65 $\pm$ 0.03 & 8.66$^{+0.06}_{-0.07}$ & 1, 4, 7, 8 \\
875: Ne & 8.08 $\pm$ 0.10 & 8.11 $\pm$ 0.04 & 8.16 $\pm$ 0.09 & 8.12$^{+0.14}_{-0.20}$ & 3, 5, 7, 8 \\
876: Ar & 6.57 $\pm$ 0.12 & 6.66 $\pm$ 0.06 & 6.62 $\pm$ 0.05 & 6.50$^{+0.14}_{-0.22}$ & 3, 6, 7, 8
877: \enddata
878: \tablerefs{(1)~\cite{AGS05}; (2)~\cite{christensen98};
879: (3)~\cite{feldman03};
880: (4)~\cite{cunha94}; (5)~\cite{cunha06};
881: (6)~This paper; (7)~\cite{esteban04}; (8)~\cite{pottasch06}. }
882: \end{deluxetable}
883:
884: % -----------------------------------------------------------------------------
885: %%
886: %% FIGURES
887: %%
888: % -----------------------------------------------------------------------------
889:
890: \clearpage
891:
892: \begin{figure}
893: \epsscale{.8}\plotone{f1.eps}
894: \caption{\label{fig1} \ion{Ar}{2}$\lambda\lambda$4426, 4430 lines in the whole stellar sample.
895: The stars in the panels are arranged in order of increasing $T_{\rm eff}$.
896: }
897: \end{figure}
898:
899:
900: \clearpage
901:
902: \begin{figure}
903: \epsscale{.8}\plotone{f2.eps}
904: \caption{\label{fig2} Sample \ion{Ar}{2} lines in the spectrum of the B1.5V star, HD 35299:
905: observed spectrum (blue points) and best-fit NLTE model spectrum (gray line;
906: red line in electronic edition). The effect of changing the argon abundance by 0.1~dex
907: on the predicted line profile is illustrated in the top left panel.
908: }
909: \end{figure}
910:
911:
912: \clearpage
913:
914: \begin{figure}
915: \epsscale{.9}\plotone{f3.eps}
916: \caption{\label{fig3} Top panels: NLTE ionization fractions of neon and argon as function
917: of depth in a model atmosphere, $T_{\rm eff} = 24,000$\,K, $\log g = 4.25$, and solar composition.
918: Dashed lines show LTE ionization fractions for the same physical conditions. Bottom panels:
919: Departure coefficients of Ne and Ar$^+$ levels of the analyzed lines (lower levels: full lines;
920: upper levels: dashed lines); the thick lines show the departure coefficients of the Ne and Ar$^+$
921: ground states.
922: }
923: \end{figure}
924:
925:
926: % -----------------------------------------------------------------------------
927:
928:
929: \end{document}
930: