1: \documentclass{aa}
2: %\documentclass[referee]{aa}
3: %
4: \usepackage{graphicx}
5: \usepackage{txfonts}
6: %
7: \def\tmap{\textsc{\bfseries tmap}}
8: \def\pro2{\textsc{\bfseries pro2}}
9: \def\lte2{\textsc{\bfseries lte2}}
10: \def\atoms2{\textsc{\bfseries atoms2}}
11: \def\setf2{\textsc{\bfseries setf2}}
12: \def\line1prof{\mbox{\textsc{\bfseries line1}\raisebox{-0.5ex}{\bfseries--}\textsc{\bfseries prof}}}
13: \def\ironic{\textrm {\bfseries IrOnIc}}
14: \def\etal{{et\,al.}\ }
15: \def\elf{PG\,1159$-$035}
16: \newcommand{\Teff}{$T\mathrm{\hspace*{-0.4ex}_{eff}}$\,}
17: \newcommand{\logg}{$\log\,g$\hspace*{0.5ex}}
18: \newcommand{\Msol}{M\hspace*{0.2ex}$_\odot$}
19: \newcommand{\Lsol}{L\hspace*{0.2ex}$_\odot$}
20: \newcommand{\gppr}{\stackrel{>}{\scriptstyle \sim}}
21: \newcommand{\gappr}{\raisebox{-0.4ex}{$\gppr $}}
22: \newcommand{\lppr}{\stackrel{<}{\scriptstyle \sim}}
23: \newcommand{\lappr}{\raisebox{-0.4ex}{$\lppr $}}
24: %\newcommand{\Ion}[2]{\mbox{#1\,{\scriptsize #2}}}
25: %
26: \begin{document}
27: %
28: \title{Discovery of photospheric argon in very hot central stars of planetary
29: nebulae and white dwarfs\thanks{Based on observations made with the NASA-CNES-CSA Far
30: Ultraviolet Spectroscopic Explorer. FUSE is operated for NASA by the Johns
31: Hopkins University under NASA contract NAS5-32985.}}
32:
33: \author{K\@. Werner\inst{1}
34: \and T\@. Rauch\inst{1}
35: \and J.~W\@. Kruk\inst{2}}
36:
37: \institute{Institut f\"ur Astronomie und Astrophysik, Universit\"at T\"ubingen, Sand 1, 72076 T\"ubingen, Germany
38: \and Department of Physics and Astronomy, Johns Hopkins University, Baltimore, MD 21218, USA}
39:
40: \offprints{K\@. Werner\\ \email{werner@astro.uni-tuebingen.de}}
41:
42:
43: \date{Received; accepted}
44:
45: \authorrunning{K. Werner, T. Rauch, J.W. Kruk}
46: \titlerunning{Discovery of argon in hot central stars and white dwarfs}
47:
48: \abstract
49: %Context
50: {We report the first discovery of argon in hot evolved stars and white
51: dwarfs. We have identified the \ion{Ar}{vii}~1063.55~\AA\ line in some of the
52: hottest known (\Teff=\,95\,000$-$110\,000~K) central stars of planetary nebulae
53: and (pre-) white dwarfs of various spectral type.}
54: %Aim
55: {We determine the argon abundance and compare it to theoretical predictions from
56: stellar evolution theory as well as from diffusion calculations.}
57: %Methods
58: {We analyze high-resolution spectra taken with the \emph{Far Ultraviolet
59: Spectroscopic Explorer}. We use non-LTE line-blanketed model atmospheres and
60: perform line-formation calculations to compute synthetic argon line profiles.}
61: %Results
62: {We find a solar argon abundance in the H-rich central star NGC~1360 and in the
63: H-deficient PG1159 star PG\,1424+535. This confirms stellar evolution modeling
64: that predicts that the argon abundance remains almost unaffected by
65: nucleosynthesis. For the DAO-type central star NGC~7293 and the hot DA white
66: dwarfs PG\,0948+534 and RE\,J1738+669 we find argon abundances that are up to
67: three
68: orders of magnitude smaller than predictions of calculations assuming
69: equilibrium of radiative levitation and gravitational settling. For the hot DO
70: white dwarf PG\,1034+001 the theoretical overprediction amounts to one dex.}
71: %Conclusions
72: {Our results confirm predictions from stellar nucleosynthesis calculations for
73: the argon abundance in AGB stars. The argon abundance found in hot white
74: dwarfs, however, is another drastic example that the current state of
75: equilibrium theory for trace elements fails to explain the observations
76: quantitatively.}
77:
78: \keywords{Stars: abundances --
79: Stars: atmospheres --
80: Stars: evolution --
81: Stars: AGB and post-AGB --
82: White dwarfs}
83:
84: \maketitle
85: %
86: %________________________________________________________________
87:
88: \begin{figure*}[tbp]
89: \includegraphics[width=0.85\textwidth]{fig_ion.ps}
90: \caption[]{Ionization fraction of argon as a function of atmospheric
91: depth. Left: Model with \Teff=\,110\,000~K, \logg=\,7.0 (for
92: PG\,1424+535). Right: Model with \Teff=\,140\,000~K, \logg=\,7.0 (for
93: PG\,1159-035). \ion{Ar}{vii} is dominant in the line-forming regions of the
94: cooler model but is strongly depopulated in the hotter model, explaining the
95: dependency of the \ion{Ar}{vii}~1063.55~\AA\ line strength from \Teff.
96: }
97: \label{fig_ion}
98: \end{figure*}
99:
100: \begin{figure*}[tbp]
101: \includegraphics[width=0.85\textwidth]{fig_ar8.ps}
102: \caption[]{The strongest \ion{Ar}{viii} lines
103: are exhibited by models with \Teff$\gappr$\,140\,000~K. They are located in a
104: region where two unidentified and two \ion{Ne}{viii} photospheric lines are detected in the
105: spectrum of K\,1-16. Unfortunately, the position of the
106: \ion{Ar}{viii} lines is uncertain by $\pm 1.2$~\AA, preventing a unique
107: identification. The parameters for the displayed model are those given for
108: K\,1-16 in Tab.\,\ref{objects_tab} except for the Ar abundance. It is was set
109: to $\log {\rm Ar}=-4.1$ (mass fraction). }
110: \label{fig_ar8}
111: \end{figure*}
112:
113: \begin{figure}[tbp]
114: \includegraphics[width=0.95\columnwidth]{fig_all.ps}
115: \caption[]{
116: Identification of the \ion{Ar}{vii}~1063.55~\AA\ line. It is not detectable in
117: the hottest objects (PG\,1159-035 and K\,1-16). Overplotted are synthetic line
118: profiles with $\log$~Ar (mass fraction), \Teff/kK, and \logg\ as labeled. Unidentified
119: photospheric lines at 1062.45 and 1063.8~\AA\ are seen in several objects
120: (marked by ``?'' in the top spectrum). All other lines are from interstellar
121: \ion{Fe}{ii} and H$_2$; some of them are labeled.}
122: \label{fig_all}
123: \end{figure}
124:
125: \section{Introduction}
126: \label{intro}
127:
128: The determination of elemental abundances in the extremely hot
129: hydrogen-deficient post-AGB stars of spectral type PG1159 is of particular
130: interest, because it is believed that these stars exhibit intershell matter as a
131: consequence of a late helium-shell flash. This allows one to directly study the
132: outcome of AGB star nucleosynthesis taking place in the region between the H-
133: and He-burning shells. While some elements are in line with expectations from
134: stellar evolution calculations, other elements clearly point at shortcomings in
135: the theory. For example, the extreme iron and sulfur depletion found in several
136: PG1159 stars is still not understood. On the other hand, the observed extreme
137: overabundances of fluorine confirms that AGB stellar modeling of complicated
138: nucleosynthesis processes is basically correct in other respects.
139:
140: The detection of metals in PG1159 stars (and stars of other spectral types in
141: the same region of the HRD) is problematic because of their extremely high
142: effective temperatures (\Teff=\,75\,000$-$200\,000~K, Werner \& Herwig 2006). All
143: species are highly ionised so that spectral lines of elements heavier than C, N,
144: and O are, typically, from ionization stages \ion{}{v$-$viii}. The vast majority
145: of these transitions is at wavelengths below the hydrogen Lyman edge and, hence,
146: essentially inaccessible. However, a few exceptions are known and particularly
147: the \emph{Far Ultraviolet Spectroscopic Explorer} (FUSE) is playing a principal
148: role in the discovery of such highly ionised species. Two examples are the
149: \ion{Ne}{vii}~973.3~\AA\ and \ion{F}{vi}~1139.5~\AA\ lines which were never
150: identified before in any stellar atmosphere (Werner \etal 2004, 2005).
151:
152: The argon abundance determination in stellar atmospheres is a challenge. Similarly
153: to neon, the solar abundance of this element cannot be determined directly
154: because of the lack of suitable spectral lines. The currently adopted solar
155: argon abundance (Table~\ref{objects_tab}) has been estimated from solar coronal
156: emission lines and solar energetic particles (Asplund \etal 2005). To our best
157: knowledge, the only stellar photospheric argon abundance determination was
158: performed for B stars which exhibit weak \ion{Ar}{ii} lines in the optical band
159: (the derived mean Ar abundance is 0.3~dex oversolar; Keenan \etal 1990, Holmgren
160: \etal 1990).
161:
162: Taresch \etal (1997) were the first who considered argon in model atmosphere
163: analyses for even hotter stars. From their models tailored to the analysis of UV
164: and FUV spectra of the O3~If supergiant HD\,93129A (\Teff=\,52\,000~K, \logg=\,3.95)
165: they pointed out that the \ion{Ar}{vi}/\ion{Ar}{vii} ionisation balance could be
166: used as an additional independent temperature indicator. However, the blend of
167: potential Ar lines exhibited by their models with strong interstellar H$_2$
168: prevented the successful use of this tool in their case. To our knowledge, this
169: investigation of highly ionised argon was never addressed again in subsequent
170: spectroscopic work. It was this investigation that inspired us to look for argon
171: in hot post-AGB stars.
172:
173: In this paper we report the discovery of an absorption line of highly ionised
174: argon, namely \ion{Ar}{vii}~1063.55~\AA, in a number of hot post-AGB stars and
175: white dwarfs with \Teff$\approx$\,100\,000~K. We describe in some detail the argon
176: model atom and the line identification in Sect.\,\ref{analysis}. We present the
177: abundance analysis in Sect.\,\ref{results} and conclude in
178: Sect.\,\ref{conclusions}.
179:
180: \begin{table*}
181: \caption{List of objects and measured argon abundances. For the two DA white
182: dwarfs \Teff\ and \logg\ are rather uncertain. We list the Ar abundances that
183: we derive for the different sets of published \Teff/\logg\ values. In the last
184: column we note the technique used to derive these sets (LTE or NLTE models, and
185: fits to hydrogen Balmer or Lyman lines).
186: \label{objects_tab}}
187: \begin{tabular}{l l r c c c c l }
188: \hline
189: \hline
190: \noalign{\smallskip}
191: object & spectral type & \Teff & \logg & \multicolumn{3}{c}{\underline{\quad logarithmic argon abundance\quad}} &reference for \Teff, \logg\\
192: & & [K] & (cgs) & \underline{mass fraction}& \multicolumn{2}{c}{\underline{\quad number ratio\quad}} &\\
193: & & & & Ar & Ar/H & Ar/He & \\
194: \noalign{\smallskip}
195: \hline
196: \noalign{\smallskip}
197: NGC 1360 &O(H) CSPN& 97\,000 & 5.3 &$-4.1$&$-5.6$& & Traulsen \etal 2005 \\
198: NGC 7293 & DAO CSPN& 105\,000& 7.0 &$-6.0$&$-7.5$& & Napiwotzki 1999 \\
199: \noalign{\smallskip}
200: PG\,0948+534 & DA & 110\,000& 7.6 &$-4.1$&$-5.7$& & Barstow \etal 2003a, (NLTE Balmer) \\
201: & & 126\,000& 7.3 &$-3.5$&$-5.1$& & Liebert \& Bergeron 1995, (LTE Balmer)\\
202: \noalign{\smallskip}
203: RE\,J1738+669 & DA (CSPN?)& 76\,000 & 7.9 &$-4.5$&$-6.1$& & Barstow \etal 2003b, (NLTE Lyman)\\
204: & & 67\,000 & 7.8 &$-3.7$&$-5.3$& & Barstow \etal 2003a, (NLTE Balmer)\\
205: & & 95\,000 & 7.9 &$-5.5$&$-7.1$& & Finley \etal 1997, (LTE Balmer)\\
206: \noalign{\smallskip}
207: PG\,1034+001 & DO (CSPN?)& 100\,000& 7.5 &$-2.8$& &$-3.8$ & Werner \etal 1995 \\
208: PG\,1424+535 & PG1159 & 110\,000& 7.0 &$-4.5$& & & Dreizler \& Heber 1998 \\
209: PG\,1159-035 & PG1159 & 140\,000& 7.0 &$\leq-5.0$ & & & Werner \etal 1991 \\
210: K\,1-16 & PG1159 CSPN&140\,000&6.4&$\leq-5.1$ & & & Kruk \& Werner 1998 \\
211: PG\,1144+005 & PG1159 & 150\,000& 6.5 &$\leq-4.1$ & & & Werner \& Heber 1991 \\
212: PG\,1520+525 & PG1159 CSPN&150\,000&7.5&$\leq-4.1$ & & & Dreizler \& Heber 1998 \\
213: \noalign{\smallskip}
214: \hline
215: \noalign{\smallskip}
216: The Sun & & & &$-4.6$&$-5.8$&$-4.8$ & Asplund \etal 2005\\
217: \noalign{\smallskip}
218: \hline
219: \end{tabular}
220: \end{table*}
221:
222: \section{Argon abundance analysis}
223: \label{analysis}
224:
225: FUSE observations and data reduction for most of our program stars were
226: described in our previous work as indicated in the Introduction. Because of the
227: higher gravity of our program stars (2$-$4~dex) compared to the supergiant
228: HD\,93129A one can expect that the \ion{Ar}{vi}/\ion{Ar}{vii} lines are
229: strongest at effective temperatures significantly higher than 50\,000~K. This is
230: because the ionization balance is shifted to lower ionisation stages with
231: increasing gravity (and hence increasing gas pressure) which is compensated by
232: increasing temperature. Our models show that the \ion{Ar}{vii} line is strongest
233: at \Teff$\approx$\,100\,000~K. In Fig.\ref{fig_ion} we display the depth
234: dependency of the Ar ionization fractions in the atmospheric models of two
235: PG1159 stars with different effective temperatures (110\,000 and 140\,000~K). The
236: ionization balance in the hotter model is shifted away significantly to
237: ionization stages higher than \ion{Ar}{vii}.
238:
239: We have identified the \ion{Ar}{vii}~1063.55~\AA\ line and performed an Ar
240: abundance determination in six stars (see Sect.~\ref{results}). The following
241: objects do not show this line: PG\,1159-035, K\,1-16, PG\,1144+005, and
242: PG\,1520+525. This allows us to derive a meaningful upper abundance limit for these
243: PG1159 stars.
244:
245: In the spectra of the PG1159 stars NGC~246 and H1504+65 there is no H$_2$ blend
246: at the \ion{Ar}{vii}~1063.55~\AA\ line position, however, \Teff\ and \logg are
247: such that the argon line has disappeared. The following objects have a too
248: strong H$_2$ line blend: The PG1159 stars PG\,1707+427, RX\,J2117+3412,
249: HS\,2324+397, NGC~7094, Abell~43, Longmore~4, the hot DO white dwarf
250: KPD\,0005+5106, the hot DA white dwarf PG\,0038+199, the hydrogen-rich central
251: stars NGC~6853, LSS~1362, NGC~1535, and the hot helium-rich stars (i.e.,
252: spectral type O(He)) K\,1-27, HS\,2209+8229, HS\,1522+6615, and LoTr~4.
253:
254: We have designed an argon model atom for NLTE line-formation calculations. These
255: are performed using and keeping fixed the physical structure (temperature,
256: densities) of line-blanketed NLTE model atmospheres which are described in
257: detail in Werner \etal (2004). In short, they are plane-parallel and in
258: hydrostatic and radiative equilibrium. Model parameters and references to the
259: previous analyses are given in Tab.\,\ref{objects_tab}. The models are composed
260: of H, He, C, O, and Ne. For the H-rich central star NGC~1360 we calculated
261: models with a solar composition. We do not adopt the oversolar helium abundance
262: suggested by Traulsen \etal (2005) because we regard this result as
263: uncertain. For the DAO central star NGC~7293 the elemental abundances are solar,
264: too, except for helium which is reduced to ${\rm He/H}=0.03$ by number. For the DAs we
265: computed H-dominated models setting He and metal abundances to 1\% solar. For
266: the DO PG\,1034+001 we fixed the C and O abundances to the values derived by
267: Werner \etal (1995) and set the unknown Ne abundance to $\log({\rm Ne/He})=-8$ (by
268: number). For each star we performed calculations with different argon content in
269: order to estimate the observed abundance and the error of the analysis.
270:
271: The model atom considers ionization stages \ion{Ar}{v$-$ix}, represented by 1,
272: 15, 20, 41, and 1 NLTE levels, respectively, plus a number of LTE levels. In the
273: ions \ion{Ar}{vi$-$viii} we include 21, 36, and 199 line transitions,
274: respectively. Due to convergence problems which have their origin in a
275: particular \ion{Ar}{vi} EUV line and are not understood, the models for objects
276: with \Teff$\leq$\,110\,000~K (except for PG\,1034+001) do not include \ion{Ar}{vi}
277: lines. This is not expected to have an influence on the calculated
278: \ion{Ar}{vii-viii} line profiles. Atomic data were taken from the
279: NIST\footnote{http://physics.nist.gov/ PhysRefData/ASD/index.html} and Opacity
280: (Seaton \etal 1994) and IRON (Hummer \etal 1993) Projects databases
281: (TIPTOPbase\footnote{ http://vizier.u-strasbg.fr/topbase/}). For all lines we
282: assumed quadratic Stark broadening for the profile calculation, except for
283: \ion{Ar}{viii} lines for which linear Stark broadening is appropriate. Let us
284: comment in some detail on the spectral lines relevant for the present work.
285:
286: The strategic \ion{Ar}{vii}~1063.55~\AA\ line is the singlet transition
287: 3p\,$^{1}$P$^{\rm o}$--3p$^2$\,$^1$D. The same transition in the sulfur
288: isoelectronic configuration gives rise to the \ion{S}{v}~1501.76~\AA\ line that
289: is well known in stellar UV spectroscopy. The exact wavelength position of the
290: \ion{Ar}{vii} line coincides in the
291: Kelly\footnote{http://cfa-www.harvard.edu/amdata/ampdata/kelly/kelly.html}
292: (1987) and {\sc Chianti}\footnote{http://wwwsolar.nrl.navy.mil/chianti.html}
293: (Young \etal 2003) databases and is in accordance with the energy levels
294: provided by Bashkin \& Stoner (1975). The Kentucky
295: database\footnote{http://www.pa.uky.edu/$^\sim$peter/atomic/} gives
296: $\lambda=1063.63$~\AA\ which disagrees with the other data sources and with the
297: line position observed in our FUSE spectra. The $f$-value for this line is
298: reasonably well known. There is only a small difference in the value quoted by
299: Taresch \etal (1997) and the {\sc Chianti} database ($f=0.113$) and the value
300: given in the Kentucky database ($f=0.130$). For our calculations we have adopted
301: $f=0.113$. We have checked with model calculations for the presence of another
302: \ion{Ar}{vii} line listed in Taresch \etal (1997) at 982.0~\AA. It is the
303: strongest component of an intercombination multiplet ($\lambda\lambda~981.97,
304: 999.47, 1066.50$~\AA) that arises from the same lower level as the 1063.55~\AA\
305: line but with an upper level in the triplet system. The $f$-value of the
306: 982.0~\AA\ line, however, is a factor of $\approx$\,100 smaller than that of the
307: 1063.55~\AA\ line and as a consequence is undetectable. A number of other
308: \ion{Ar}{vii} lines appear in the Taresch \etal (1997) list
309: ($\lambda\lambda~1060.1, 1060.4, 1067.4, 1136.7$~\AA), for which no level
310: energies are publicly available. They are not detected in our program stars
311: probably because they arise from too highly excited levels.
312:
313: Concerning the \ion{Ar}{vi} ion, Taresch \etal (1997) list three lines of the
314: intercombination resonance multiplet 3p$^2$\,P$^{\rm o}$--3p$^2$\,$^4$P
315: ($\lambda\lambda~998.4, 1000.2, 1013.3$~\AA). The $f$-values of all multiplet
316: components, however, are $6.6\cdot 10^{-5}$ and smaller. The resulting lines are
317: undetectable in our model spectra as well as in the FUSE spectra of our
318: stars. Four other lines listed ($\lambda\lambda~1008.8, 1013.3, 1164.5,
319: 1169.4$~\AA) are very highly excited and not detectable in our program stars.
320:
321: A look at Fig.\,\ref{fig_ion} suggests that stars with \Teff$\gappr 140\,000$~K
322: could exhibit lines even from \ion{Ar}{viii}. This is a hydrogen-like ion having
323: line transitions between levels with principal quantum numbers
324: n=5$\longrightarrow$6 in the FUSE spectral range. The two lines with the largest
325: $gf$-values are 5f$\longrightarrow$6g and 5g$\longrightarrow$6h located at
326: 1154.9~\AA\ and 1164.1~\AA, respectively. The latter is stronger and it
327: is indeed clearly seen in the hot star models although it is strongly broadened
328: by linear Stark effect (Fig.\,\ref{fig_ar8}). Unfortunately the level energies
329: are not known accurately enough so that the line position is uncertain within
330: $\pm 1.2$~\AA\ (Kentucky database). Within that uncertainty range of the
331: 5f$\longrightarrow$6g line a hitherto unidentified line at 1155.7~\AA\ can be
332: seen in the spectra of the hottest PG1159 stars, among them is K\,1-16
333: (Fig.\,\ref{fig_ar8}), however, the identification with \ion{Ar}{viii} is
334: unclear. The 5g$\longrightarrow$6h line is located in the blue wing of a strong
335: and broad \ion{Ne}{viii} line at 1164.7~\AA\ and cannot be identified
336: clearly. The use of these \ion{Ar}{viii} lines as a spectral diagnostic is
337: questionable in any case because for some levels that probably are important for
338: a proper NLTE model atom the energies are completely lacking.
339:
340: The two \ion{Ne}{viii} lines shown in Fig.\,\ref{fig_ar8} are new
341: identifications. Exploratory modeling suggests that they are of photospheric
342: origin. We are preparing detailed NLTE line formation calculations and will
343: present results in future work.
344:
345: \section{Results}
346: \label{results}
347:
348: We present profile fits to the \ion{Ar}{vii}~1063.55~\AA\ line in
349: Fig.\,\ref{fig_all}. The results are given in Table~\ref{objects_tab}. In the
350: following we comment on the individual objects in their order of appearance in
351: Fig.\,\ref{fig_all} and Table~\ref{objects_tab}.
352:
353: For the white dwarfs we compare our results to the theoretical argon abundance
354: predictions of Chayer \etal (1995a,b). They are based on the assumption that
355: trace elements in H or He dominated white dwarf atmospheres are kept from
356: gravitational settling by radiative levitation. These authors present detailed
357: predictions for the Ar abundance in DA and DO white dwarfs as a function of
358: effective temperature and surface gravity.
359:
360: \paragraph{NGC~1360} The central star of this planetary nebula is of spectral
361: type O(H), i.e., its optical spectrum is dominated by the hydrogen Balmer
362: lines. Hoare \etal (1996) derived \Teff=\,110\,000~K and \logg=\,5.6. A more recent
363: analysis using metal lines in UV spectra indicates slightly different values,
364: namely \Teff=\,97\,000~K and \logg=\,5.3 (Traulsen \etal 2005). We find a marginal
365: Ar overabundance, $\log({\rm Ar/H})=-5.6$ (by number), that is within the error
366: limit identical to the solar abundance ($-5.8$).
367:
368: \paragraph{NGC~7293} This planetary nebula nucleus is classified as a DAO white
369: dwarf, i.e., it is hydrogen-rich with an admixture of helium. The atmospheric
370: parameters were derived by Napiwotzki (1999) using optical spectra:
371: \Teff=\,105\,000~K, \logg=\,7.0, ${\rm He/H}=0.03$. With $\log({\rm Ar/H})=-7.5$ (by number)
372: we find a strong Ar \emph{underabundance} of a factor of $\approx 50$. We can
373: compare this result with the prediction of levitation theory. From Fig.\,2 in
374: Chayer \etal (1995b) we read $\log({\rm Ar/H})=-3.5$ for DA white dwarfs, i.e.,
375: an extreme \emph{overabundance} (factor $\approx 200$). The fact that the star
376: has some residual helium in the atmosphere is probably unimportant for this
377: prediction (note that for DO white dwarfs the predicted Ar abundance is even
378: higher, by about 1~dex). In essence, the predicted Ar abundance is four orders
379: of magnitude higher than the observed one. Traulsen \etal (2005) have claimed
380: higher \Teff\ and lower \logg\ from UV spectra (120\,000~K, 6.5) which both have
381: the tendency to weaken the \ion{Ar}{vii} line and, hence, a 0.5~dex higher Ar
382: abundance is derived in this case which is still much below the theoretical
383: prediction.
384:
385: \paragraph{PG\,0948+534} This is one of the hottest known DA white dwarfs. From
386: Balmer line NLTE fits Barstow \etal (2003a) derived \Teff=\,110\,000~K, \logg=\,7.6. Our
387: best fit model has $\log({\rm Ar/H})=-5.7$ (by number). The Chayer \etal (1995b)
388: prediction for these parameters is much higher, $\log({\rm Ar/H})=-3.9$. (A
389: slight extrapolation was needed because their calculations do not consider
390: \Teff$>$\,100\,000~K.) In their LTE Balmer line analysis Liebert \& Bergeron (1995) find \Teff\
391: even higher and \logg\ lower (\Teff=\,126\,000~K, \logg=\,7.3) which both has the
392: tendency to make the \ion{Ar}{vii} line weaker. We have computed a small model
393: set with their \logg\ and \Teff=\,130\,000~K and different Ar abundances. The best
394: fit is now achieved at $\log({\rm Ar/H})=-5.1$ which makes the difference with the
395: predicted abundance smaller but it is still significant.
396:
397: \paragraph{RE\,J1738+669} This is another extremely hot DA white dwarf. The
398: Balmer line LTE analysis by Finley \etal (1997) resulted in \Teff=\,95\,000~K,
399: \logg=\,7.9. We derive $\log({\rm Ar/H})=-7.1$ (by number). The respective Chayer
400: \etal (1995b) prediction is $\log({\rm Ar/H})=-4.3$ which is again a significant
401: overestimation. The \Teff/\logg\ values for this DA are, however, rather
402: uncertain. Barstow \etal (2003a,b) have performed a separate analysis of the
403: Balmer and Lyman lines and arrived at significantly lower \Teff\ but with
404: discrepant values: 67\,000/7.8 and 76\,000/7.9, respectively. For these two
405: parameter sets we derive $\log({\rm Ar/H})=-5.3$ and $\log({\rm Ar/H})=-6.1$
406: which can be compared to the predicted diffusion abundances for the respective
407: \Teff/\logg\ sets: $-4.8$ and $-4.5$. This means an overestimation of 0.5 and
408: 1.5 dex. Thus, irrespective of the uncertain temperature we find that theory
409: overpredicts the Ar abundance. Only if one accepts the lowest \Teff\ value then
410: the Ar abundance is close to the predicted amount.
411:
412: \paragraph{PG\,1034+001} This DO is among the hottest known helium-rich white
413: dwarfs (\Teff=\,100\,000~K, \logg=\,7.5). It exhibits the strongest \ion{Ar}{vii}
414: line in all stars we examined. We derive $\log({\rm Ar/He})=-3.8$ (by
415: number). From Fig.\,16 in Chayer \etal (1995a) we find that the theoretical
416: prediction is about one dex higher: $\log({\rm Ar/He})=-2.9$.
417:
418: \paragraph{PG\,1424+535} This is the only PG1159 star in which we could identify
419: the \ion{Ar}{vii} line. It is rather strong and we derive a mass fraction of
420: $\log {\rm Ar}=-4.5$. This is close to the solar value ($-4.6$).
421:
422: \paragraph{PG\,1159-035} The prototype of the PG1159 class does not show clear
423: evidence for the \ion{Ar}{vii} line. We derive an upper limit of $\log {\rm
424: Ar}=-4.5$ (mass fraction) which is almost equal to the solar abundance value
425: ($-4.6$).
426:
427: \paragraph{K\,1-16, PG\,1144+005, PG\,1520+525} These PG1159 stars have
428: effective temperatures similar to the prototype. Hence the non-detection of
429: argon allows to derive the solar abundance value as an approximate upper limit,
430: too.
431:
432: From our experience with the profile fitting we estimate the error of our
433: abundance determinations to 0.5~dex.
434:
435: \section{Conclusions}
436: \label{conclusions}
437:
438: \subsection{PG1159 stars and the H-rich central star NGC~1360}
439:
440: Nucleosynthesis calculations for AGB stars predict that the argon abundance in
441: the intershell region remains almost unchanged. In a particular model
442: calculation for a 2~M$_\odot$ AGB star the most abundant argon isotopes
443: $^{36}$Ar and $^{38}$Ar are being destroyed and produced by neutron captures,
444: respectively, resulting in a net reduction of the total Ar abundance of 0.2~dex
445: after 30 thermal pulses (Gallino priv. comm.). This is in line with results from
446: argon isotope analyses of meteoritic SiC grains which were formed in the winds
447: from AGB carbon stars (Gallino \etal 1990). All five noble gases detected in
448: such grains have elemental and isotopic abundances which are very similar to
449: results of nucleosynthesis calculations for the intershell of these stars (Lewis
450: \etal 1990).
451:
452: The solar abundance of argon that we detected in the PG1159 star PG\,1424+535
453: and in the H-rich central star NGC~1360 further confirm these
454: results. PG\,1424+535 displays intershell matter and proves that the Ar
455: abundance remained unchanged during the previous AGB evolution. NGC~1360 could
456: show deviations from the solar Ar abundance if third dredge-up of strongly
457: Ar-depleted/enhanced matter during the previous AGB evolution would have
458: occurred. This is not observed. The solar upper Ar abundance limits determined
459: for the three other PG1159 stars in our sample rule out an extraordinary argon
460: production.
461:
462: \subsection{DA, DAO, and DO white dwarfs}
463:
464: Chayer \etal (1995a,b) have presented detailed predictions about trace element
465: abundances based on diffusion/levitation equilibrium theory. They found,
466: however, that equilibrium theory is unable to reproduce the observed trace
467: element abundances quantitatively for most species. This failure has become
468: increasingly obvious during the last years when more and more observational
469: results became available.
470:
471: Vennes \etal (2005) have for the first time discovered trans-iron group elements
472: in hot DAs (germanium and probably tin). The derived abundances are nearly solar
473: (i.e., of the order $10^{-8}-10^{-9}$ relative to H, by number) and this made
474: these authors ask: ``With the detection of such an unlikely candidate [Ge], what
475: of abundant elements such as Ne, Mg, Ar, and Ca? These elements would be
476: detectable in high-dispersion extreme-ultraviolet spectra. Until their
477: abundances have been determined, the abundance pattern in hot DA white dwarfs
478: remains incomplete.'' Our results on argon are one step towards this goal. They
479: also show that far-ultraviolet data supplied by FUSE offer the only way to this
480: goal as long as high-resolution EUV spectroscopy is unavailable. A significant
481: number of still unidentified photospheric lines in the FUSE spectra of hot white
482: dwarfs (some examples are presented in Figs.\,\ref{fig_ar8} and \ref{fig_all})
483: may hold the key to even more hitherto unidentified elements in white
484: dwarfs. The argon abundances derived in the present paper are another example
485: that equilibrium theory is unable to reproduce the observed trace element
486: abundances quantitatively.
487:
488: To conclude, we recall that not only PG1159 photospheric metal abundances are
489: affected by s-process nucleosynthesis that took place during the AGB phase. This
490: may also be the case for white dwarfs although diffusion has certainly altered
491: the abundance pattern. This is suggested by the recent discovery of strongly
492: oversolar abundances of trans-iron group elements (up to iodine) in cool
493: (\Teff$\approx$\,50\,000~K) DO white dwarfs (Chayer \etal 2005). In contrast, the
494: one dex oversolar argon abundance that we found in the hot DO
495: PG\,1034+001 cannot be the result of nucleosynthesis but must be the result of
496: selective radiative levitation.
497:
498: \begin{acknowledgements}
499: We would like to thank Roberto Gallino for sending his results on
500: nucleosynthesis calculations, and Maria Lugaro for pointing out the argon
501: analysis of meteoritic SiC grains. T.R. is supported by the German Ministry of
502: Economy and Technology through DESY under grant 05\,AC6VTB. J.W.K. is supported
503: by the FUSE project, funded by NASA contract NAS5-32985.
504: \end{acknowledgements}
505:
506: \begin{thebibliography}{}
507:
508: \bibitem{Asp05} Asplund, M., Grevesse, N., \& Sauval, A.~J. 2005, in Cosmic
509: Abundances as Records of Stellar Evolution and Nucleosynthesis, eds.\ T.~G. Barnes, F.~N. Bash III, ASP Conference Series, 336, 25
510:
511: \bibitem{Ba03a} Barstow, M.~A., Good, S.~A., Holberg, J.~B., \etal 2003a, MNRAS, 341, 870
512:
513: \bibitem{Ba03b} Barstow, M.~A., Good, S.~A., Burleigh, M.~R., \etal 2003b, MNRAS, 344, 562
514:
515: \bibitem{BS75} Bashkin, S., \& Stoner, J.~O. Jr. 1975, Atomic Energy Levels \& Grotrian Diagrams, Vol.\ 2, Amsterdam, North Holland
516:
517: \bibitem{Cha95a} Chayer, P., Fontaine, G., \& Wesemael, F. 1995a, ApJS, 99, 189
518:
519: \bibitem{Cha95b} Chayer, P., Vennes, S., Pradhan, A.~K., \etal 1995b, ApJ, 454, 429
520:
521: \bibitem{Cha05} Chayer, P., Vennes, S., Dupuis, J., \& Kruk, J.~W. 2005, ApJ, 630, L169
522:
523: \bibitem{dh98} Dreizler, S., \& Heber, U. 1998, A\&A, 334, 618
524:
525: \bibitem{Fin97} Finley, D.~S., Koester, D., \& Basri, G. 1997, ApJ, 488, 375
526:
527: \bibitem{Gal90} Gallino, R., Busso, M., Picchio, G., \& Raiteri, C.~M. 1990, Nature, 348, 298
528:
529: \bibitem{Ho96} Hoare, M. G., Drake, J. J., Werner, K., \& Dreizler, S. 1996, MNRAS, 283, 830
530:
531: \bibitem{Hol90} Holmgren, D.~E., Brown, P.~J.~F., Dufton, P.~L., \& Keenan, F.~P. 1990, ApJ, 364, 657
532:
533: \bibitem{hea1993} Hummer, D.~G., Berrington, K.~A., Eissner, W., \etal 1993, A\&A, 279, 298
534:
535: \bibitem{Kee90} Keenan, F.~P., Bates, B., Dufton, P.~L., Holmgren, D.~E., \& Gilheany, S. 1990, ApJ, 348, 322
536:
537: \bibitem{Ke87} Kelly, R.~L. 1987, Journal of Physical and Chemical Reference Data, Vol. 16, Suppl. No. 1
538:
539: \bibitem{KW98} Kruk, J.~W., \& Werner, K. 1998, ApJ, 502, 858
540:
541: \bibitem{Lew90} Lewis, R.~S., Amari, S., \& Anders, E. 1990, Nature, 348, 293
542:
543: \bibitem{Li95} Liebert, J., \& Bergeron, P. 1995, in White Dwarfs, eds. D. Koester, K. Werner, Lecture Notes in Physics, 443, 12
544:
545: \bibitem{Na99} Napiwotzki, R. 1999, A\&A, 350, 101
546:
547: \bibitem{Sea94} Seaton, M.~J., Yan, Y., Mihalas, D., \& Pradhan, A.~K. 1994, MNRAS, 266, 805
548:
549: \bibitem{tar97} Taresch, G., Kudritzki, R.~P., Hurwitz, M., \etal 1997, A\&A, 321, 531
550:
551: \bibitem{tra05} Traulsen, I., Hoffmann, A.~I.~D., Werner, K., \etal 2005,
552: in White Dwarfs, eds.\ D. Koester, S.~Moehler, ASP Conference Series, 334, 325
553:
554: \bibitem{Ven05} Vennes, S., Chayer, P., \& Dupuis, J. 2005, ApJ, 622, L121
555:
556: \bibitem{wh91} Werner, K., \& Heber, U. 1991, A\&A, 247, 476
557:
558: \bibitem{wh06} Werner, K., \& Herwig, F., 2006, PASP, 118, 183
559:
560: \bibitem{w91} Werner, K., Heber, U., \& Hunger, K. 1991, A\&A, 244, 437
561:
562: \bibitem{wdw95} Werner, K., Dreizler, S., \& Wolff, B. 1995, A\&A, 298, 567
563:
564: \bibitem{w04} Werner, K., Rauch, T., Reiff, E., Kruk, J.~W., \& Napiwotzki, R. 2004, \aap, 427, 685
565:
566: \bibitem{w05} Werner, K., Rauch, T., \& Kruk, J.~W.\ 2005, \aap, 433, 641
567:
568: \bibitem{Yo03} Young, P.~R., Del Zanna, G., Landi, E., Dere, K.~P., Mason, H.~E., \& Landini, M. 2003, ApJS, 144, 135
569:
570: \end{thebibliography}
571:
572: \end{document}
573: