1:
2: \documentclass[usenatbib]{mn2e}
3: %\documentclass[usenatbib,referee]{mn2e}
4: \usepackage{epsfig}
5: %
6: % Version of 21 June 2000; small typo corrected 4 August 2000
7: %
8: \def\nei{\hbox{Ne\,{\sc i}}}
9: \newcommand{\vsini}{\mbox{$v\sin\!i$}}
10: \newcommand{\logg}{\mbox{$\log g$}}
11: \newcommand{\teff}{\mbox{$T_{\mathrm{eff}}$}}
12:
13: \begin{document}
14:
15: \title[HgMn stars]{Neon abundances in normal late-B and mercury-manganese
16: stars}
17:
18: \author[M. M. Dworetsky and J. Budaj]
19: {M. M. Dworetsky$^1$\thanks{Guest Observer, Lick
20: Observatory} and J. Budaj$^{1,2}$\thanks{E-mail: (MMD)
21: mmd@star.ucl.ac.uk (JB) budaj@ta3.sk}\\
22: $^1$Department of Physics and Astronomy, University College London, Gower
23: Street, London, WC1E 6BT \\
24: $^2$Astronomical Institute of the Slovak
25: Academy of Sciences, 05960 Tatransk\'{a} Lomnica, Slovak Republic}
26:
27:
28: \maketitle
29:
30: \begin{abstract}
31:
32: We make new Non-LTE calculations to deduce abundances of neon from
33: visible-region \'{e}chelle spectra of selected \nei\ lines in 7 normal
34: stars and 20 HgMn stars. We find that the best strong blend-free Ne line
35: which can be used at the lower end of the \teff\ range is $\lambda$6402,
36: although several other potentially useful \nei\ lines are found in the red
37: region of the spectra of these stars. The mean neon abundance in the
38: normal stars ($\log A =8.10$) is in excellent agreement with the standard
39: abundance of neon (8.08). However, in HgMn stars, neon is almost
40: universally underabundant, ranging from marginal deficits of 0.1-0.3 dex
41: to underabundances of an order of magnitude or more. In many cases, the
42: lines are so weak that only upper limits can be established. The most
43: extreme example found is $\upsilon$ Her with an underabundance of at least
44: 1.5 dex. These underabundances are qualitatively expected from radiative
45: acceleration calculations, which show that Ne has a very small radiative
46: acceleration in the photosphere and is expected to undergo gravitational
47: settling if mixing processes are sufficiently weak, and there is no strong
48: stellar wind. According to the theoretical predictions of
49: \citet{landolvau}, the low Ne abundances place an important constraint on
50: the intensity of such stellar winds, which must be less than
51: $10^{-14}M_{\sun}$\,yr$^{-1}$ if they are non-turbulent.
52:
53: \end{abstract}
54:
55: \begin{keywords}
56: stars:abundances -- stars:chemically peculiar -- line:profiles\\
57: \end{keywords}
58:
59: \section{Introduction}
60: \label{section:intro}
61: HgMn stars are a subclass of chemically peculiar star occupying the
62: spectral region corresponding approximately to MK types B9--B6
63: (10\,500--16\,000\,K). Due to low helium abundances, their spectral
64: classes are generally placed by observers in the A0--B8 range.
65: Observationally, they are characterized by extremely low rotational
66: velocities, weak or non-detectable magnetic fields and photometric
67: variability, and atmospheric deficiencies of light elements (e.g., He, Al,
68: N) coupled with enhancements of the heavy elements (e.g., Hg, Mn, Pt, Sr,
69: Ga). In addition, the heavy elements also have non-terrestrial isotopic
70: abundances (\citealt{kcs97}; \citealt*{bohl}). The currently favoured
71: mechanism for explaining these anomalies is the radiative diffusion
72: hypothesis \citep{m70}. This work has been advanced in the form of a
73: parameter-free model \citep{m86}.
74:
75: The quiescent atmospheres of these stars makes them one of the best
76: natural laboratories for studying the competing processes of gravitational
77: settling and radiative levitation \citep{vau82}. In the absence of
78: disrupting mechanisms such as convection, rotationally-induced meridional
79: currents, high microturbulence and magnetic fields, certain rare elements
80: can reach a factor of 10$^5$ enhancement over their standard abundance.
81: Because of the strength and sharpness of normally exotic spectroscopic
82: lines, HgMn stars are also useful for constraining fundamental atomic data
83: \citep{l95}.
84:
85: Although there have been many studies of individual HgMn stars and of the
86: abundances of many elements across a sample of HgMn stars, we have been
87: unable to find any papers mentioning the abundance of Ne in HgMn stars,
88: with the recent exception of a paper by \citet{adel00}, in which LTE
89: calculations established that \nei\ line strengths in $\kappa$ Cnc implied
90: overabundance of 0.64 dex, while an underabundance was found in HR\,7245.
91: The He-wk star 3~Cen~A also seemed to be overabundant in Ne relative to
92: the Sun (we will look again at these results in Section~\ref{disc}).
93: However, Ne is known to exhibit strong non-LTE effects in B stars,
94: according to the original investigations by \citet{auer73}, who showed
95: that Non-LTE (NLTE) methods yield \nei\ lines that are nearly double the
96: strength expected from LTE calculations at \teff\ = 15\,000\,K. Thus the
97: overabundances reported may well be due to neglect of NLTE considerations.
98:
99: The lack of Ne observations in late B stars is slightly surprising,
100: because Ne is an important and interesting element. Its standard abundance
101: (\citealt{ag89}; \citealt*{grev96}), which was deduced from the solar
102: wind, nebular spectroscopy, and stellar observations and analyses such as
103: those of \citeauthor{auer73}, is comparable with that of C, N and O. It is
104: interesting also because its atomic structure resembles that of He with
105: very high first ionization potential (about 22\,eV). Consequently, all its
106: resonance lines and the ground state photoionization continuum are in the
107: Lyman continuum where the stellar energy flux in the photosphere of late B
108: stars is low. One would then expect that radiative acceleration on Ne may
109: not be enough to balance gravity and Ne should sink. Indeed, theoretical
110: calculations by \citet*{landolvau}, who considered a nonturbulent mass
111: loss or stellar wind in the stellar envelopes, predict that there will be:
112: (1) neon underabundances if the mass loss rate is less than
113: $10^{-14}M_{\sun}$\,yr$^{-1}$; (2) neon overabundances for mass loss in
114: the range $10^{-14}-10^{-12}M_{\sun}$\,yr$^{-1}$; or (3) normal neon
115: abundances for mass loss over $10^{-12}\,M_{\sun}$\,yr$^{-1}$. Also, our
116: own calculations of radiative accelerations in the atmospheres
117: \citep{budw00} predict a pattern of general photospheric underabundances
118: for Ne.
119:
120: In this paper, we present an abundance analysis of visible \nei\ lines
121: based on a full NLTE treatment of the strength of $\lambda$6402. It is the
122: strongest unblended line of \nei\ in the visible spectrum. We also
123: demonstrate that, to a close order of magnitude, several other \nei\
124: lines tend to have similar NLTE enhancements and
125: can be used if spectra showing $\lambda$6402 are not available. We find
126: that for most HgMn stars, neon turns out to be underabundant, suggesting
127: that the first scenario of \citet{landolvau} is the most likely one.
128:
129: \section{Observations}
130: \label{section:obs}
131:
132: Our stellar sample is based upon that of \citet{sd93}, who analysed IUE
133: data on the UV resonance lines of iron-peak elements in 26 HgMn, 4
134: superficially normal and 10 normal stars. We observed definite detections
135: or determined upper limits for \nei\ in 21 of the HgMn stars in the
136: \citet{sd93} sample and 11 of the normal and superficially normal group.
137: Some of the other stars in the two samples were lacking data in the red
138: region or were cooler than 10\,000\,K and we did not expect to observe any
139: \nei\ lines. Physical parameters of the stars in this study are given in
140: Table~\ref{table:stars}.
141:
142: \begin{table*}
143: \caption{Programme stars: basic data and adopted atmospheric parameters.}
144: \begin{tabular}{llllrllllr}
145: \hline
146: Star &HD &Spectral type &V$_r$ &\teff &\logg &$\xi$ &Ref.
147: &\vsini &Ref. \\
148: & & &(km\,s$^-$$^1$) &(K) &(dex\,cm\,s$^-$$^2$) &(km\,s$^-$$^1$) &
149: &(km\,s$^-$$^1$) \\ \hline \multicolumn{10}{c}{\it {Normal and
150: superficially normal stars}} \\ $\pi$\,Cet &17081 &B7\,V &+15\,SB &13\,250
151: &3.80 &0.0 &[1] &25 &[3] \\ 134\,Tau &38899 &B9\,IV &+18\,V &10\,850 &4.10
152: &1.6 &[1] &30 &[3] \\ $\tau$\,Her &147394 &B5\,IV&$-$14\,V? &15\,000 &3.95
153: &0.0 &[3] &32 &[3] \\ $\zeta$\,Dra &155763 &B6\,III&$-$17\,V? &12\,900
154: &3.90&2.5: & &34 &[3] \\ $\alpha$\,Lyr &172167 &A0\,Va &$-$14\,V &9\,450
155: &4.00 &2.0 &[4] &24 &[3] \\ HR\,7098 &174567 &A0\,Vs &$-$\ 3 &10\,200
156: &3.55 &1.0 &[3] &11 &[3] \\ 21\,Aql &179761 &B8\,II-III &$-$\ 5\,V
157: &13\,000 &3.50 &0.2 &[1] &17 &[3] \\ HR\,7338 &181470 &A0\,III &$-$14\,SBO
158: &10\,250 &3.75 &0.5 &[3] &\ 3 &[3] \\ $\nu$\,Cap &193432 &B9.5\,V &$-$\
159: 2\,V? &10\,300 &3.90 &1.6 &[1] &27 &[3] \\ HR\,7878 &196426 &B8\,IIIp
160: &$-$23 &13\,050 &3.85 &1.0: & &\ 6 &[9] \\ 21\,Peg &209459 &B9.5\,V &$-$\
161: 0 &10\,450 &3.50 &0.5 &[3] &\ 4 &[3] \\ \multicolumn{10}{c}{\it {HgMn
162: stars}} \\ 87\,Psc &7374 &B8\,III &$-$16\,V &13\,150 &4.00 &1.5 &[3] &21.0
163: &[5] \\ 53\,Tau &27295 &B9\,IV &+12\,SBO &12\,000 &4.25 &0.0 &[2] &\ 6.5
164: &[5] \\ $\mu$\,Lep &33904 &B9\,IIIpHgMn &+28 &12\,800 &3.85 &0.0 &[2]
165: &15.5 &[5] \\ HR\,1800 &35548 &B9\,pHgSi &$-$\ 9\,V? &11\,050 &3.80 &0.5
166: &[3] &\ 3.0 &[5] \\ 33\,Gem &49606 &B7\,III &+13 &14\,400 &3.85 &0.5: &
167: &22.0 &[5] \\ HR\,2676 &53929 &B9.5\,III &+\ 6\,V? &14\,050 &3.60 &1.0: &
168: &25.0 &[5] \\ HR\,2844 &58661 &B9\,pHgMn &+21\,V &13\,460 &3.80 &0.5: &
169: &27.0 &[5] \\ $\nu$\,Cnc &77350 &A0\,pSi &$-$15\,SBO &10\,400 &3.60 &0.1
170: &[6] &13 &[6] \\ $\kappa$\,Cnc &78316 &B8\,IIIpMn &+24\,SB1O &13\,500
171: &3.80 &0.0 &[2] &\ 7 &[5] \\ HR\,4072 &89822 &A0\,pSiSr:Hg: &$-$\ 0\,SB2O
172: &10\,500 &3.95 &1.0 &[7] &\ 3.2 &[7] \\ $\chi$\,Lup &141556 &B9\,IV &+\
173: 5\,SB2O &10\,750 &4.00 &0.0 &[11] &\ 2.0 &[7] \\ $\iota$\,CrB &143807
174: &A0\,p:Hg: &$-$19\,SB &11\,000 &4.00 &0.2 &[6] &\ 1.0 &[7] \\
175: $\upsilon$\,Her & 144206 &B9\,III &+\ 3 &12\,000 &3.80 &0.6 &[1] &\ 9.0
176: &[5] \\ $\phi$\,Her &145389 &B9\,p:Mn: &$-$16\,SB1O &11\,650 &4.00 &0.4
177: &[8] &10.1 &[5] \\ HR\,6997 &172044 &B8\,II-IIIpHg &$-26$\,SBO &14\,500
178: &3.90 &1.5 &[3] & 36.0 &[5] \\ 112\,Her &174933 &B9\,II-IIIpHg
179: &$-$20\,SB2O &13\,100 &4.10 &0.0: & &\ 5.5 &[12] \\ HR\,7143 &175640
180: &B9\,III &$-$26\,V? & 12\,100 &4.00 &1.0 &[3] &\ 2.0 &[5] \\ HR\,7361
181: &182308 &B9\,IVpHgMn &$-$20 V? &13\,650 &3.55 &0.0 &[3] &\ 8.2 &[5] \\
182: 46\,Aql &186122 &B9\,IIIpHgMn &$-$32 &13\,000 &3.65 &0.0 &[3] &\ 3.0 &[5]
183: \\ HR\,7664 &190229 &B9\,pHgMn &$-$22\,SB1 &13\,200 &3.60 &0.8 &[8] &\ 8.0
184: &[5] \\ HR\,7775 &193452 &A0\,III &$-$18$^a$ &10\,800 &3.95 &0.0 &[3] &\
185: 0.8 &[10] \\ \hline \end{tabular} \begin{minipage}{15.1cm} $^a$Hoffleit \&
186: Warren (1991) cite HR\,7775 as SB1O although this is a confusion with
187: $\beta$\,Cap (HR\,7776). \\ \ \\ Notes: Spectral types and radial velocity
188: data are from \citet{bsc91}. Values of \teff\ and \logg\ are from
189: \citet{sd93}, or [12] in the case of 112\,Her. Values of `V' and `V?',
190: respectively, indicate known or suspected radial velocity variables; `SB':
191: spectroscopic binary (`SB1' and `SB2', respectively, denote single- and
192: double-lined systems); `O': published orbit (see \citealt*{batten89}).
193: Microturbulence parameters $\xi$ appended by a colon (`:') are approximate
194: and were derived solely from UV Fe\,II lines by \citet{sd93}. \\
195: \\
196: References: [1] \citet{adfuhr85}, [2] \citet{adel88a}; [3]
197: \citet{kcs92}; [4] \citet{gigas}; [5] \citet*{djs98};
198: [6] \citet{ad89}; [7] \citet{harman}; [8] \citet{adel88b}; [9]
199: \citet{cow80}; [10] \citet*{bohl}; [11] \citet*{wahl94};
200: [12] \citet*{r96}.
201:
202: \end{minipage}
203:
204: \label{table:stars}
205: \end{table*}
206:
207: All observations were obtained with the Hamilton \'{E}chelle Spectrograph
208: \citep[HES;][]{vogt} at Lick Observatory, fed by the 0.6-m Coud\'{e}
209: Auxilliary Telescope (CAT), during four runs in 1994--1997. Further
210: details of the instrument can be found in \citet{mischb}. Shortly before
211: our observations in 1994, some of the HES optical components were
212: replaced, improving the resolution and instrumental profile, and making it
213: possible to use the full field of the 2048 $\times$ 2048 CCDs to maximum
214: advantage. We used both the unthinned phosphor-coated Orbit CCD (Dewar
215: 13) and from July 1995 the thinned Ford CCD (Dewar 6), depending on
216: availability as the latter was shared with the multi-object spectrograph
217: on the 3-m telescope. The spectral range for the observations was
218: 3800--9000\,\AA. Typical signal-to-noise (S/N) per pixel in the centres of
219: orders ranged from 75 to 250. The Orbit CCD is cosmetically very clean,
220: with very few bad pixels or columns, while the thinned Ford CCD contains
221: several column defects but offers a much higher detector quantum
222: efficiency in the blue. We used the Ford CCD whenever it was available.
223: With the slit settings used, the combination of spectrographs and CCDs
224: gave resolutions $R\!\approx\!46\,500$. Flat fields were made using the
225: polar axis quartz lamp and wavelength calibrations were obtained with a
226: Th--Ar comparison.
227:
228: The \'{e}chelle spectra were extracted and calibrated using standard {\sc
229: iraf} extraction packages \citep{churchill,valdes}, running on UCL's
230: Starlink node. Previous measurements \citep{csa98} showed that there were
231: no measurable effects of parasitic light (residual scattered light) in
232: the line profiles provided that general scattered light in the adjacent
233: interorder spaces was taken as the subtracted background. In practice the
234: residual scattered light was less than approximately 1 percent; we have
235: therefore made no corrections for it. Allen's method is based on a direct
236: comparison of the solar spectrum (as reflected from the roof of the CAT
237: coelostat) observed with the HES, with the Kitt Peak Solar Flux Atlas
238: \citep{kpno}. As the latter was obtained using a Fourier Transform
239: Spectrometer, it has no measurable parasitic light. The Kitt Peak
240: spectrum is convolved with a suitable instrumental profile to match the
241: HES data; both spectra must be normalized at the same points for a valid
242: comparison. The ratio of summed equivalent widths of various features
243: with good adjacent continuum points, in many different spectral orders,
244: provides the measure of the amount of parasitic light.
245:
246: \section{Abundance determination}
247: \label{section:abunds}
248:
249: \subsection{Stellar parameters and stellar atmospheres}
250: \label{sub:params}
251:
252: Effective temperatures and surface gravities of programme stars are
253: summarized in Table~\ref{table:stars}. In general, the parameters adopted
254: follow our previous work \citep*{sd93,djs98,jda99}. Seven stars are noted
255: as double-lined spectroscopic binaries in Table~\ref{table:ewidths}; one
256: of these (HR\,1800) is better described as a close visual binary in which
257: we can see evidence of the secondary spectrum as rotationally-broadened
258: features. The parameters and light ratios quoted in all seven cases are
259: those adopted for the primary star. Suitable light ratios for other
260: wavelength regions were found by the use of \citet{cd13} model atmosphere
261: fluxes. The light ratio estimated this way for $\lambda$6402 is given in
262: Table~\ref{table:binary}, where it is representative of the values
263: throughout the range $\lambda\lambda$5800-6700. The adopted light ratio
264: for the visual binary HR\,1800 ($\rho = 0.243$ arcsec) is 2.45, based on
265: $\Delta H_p = 0.96$ mag from {\it The Hipparcos Catalogue}
266: \citep{ESA1997}. Another star, 33 Gem, is suspected of being double-lined
267: but there is not yet any information on the orbit or light ratio
268: \citep{hl93}; we treat it as a single star or `average component.' We note
269: that \citet*{ad96} also treated 33 Gem as a single star, noting that the
270: question of binarity could not be conclusively resolved with their data.
271:
272: \begin{table*}
273: \caption{Binary stars: adopted stellar data and light ratios}
274: \begin{tabular}{lcccccr}
275: \hline
276: Star &$\lambda$ &L$_A$/L$_B$ &$\teff_A/\logg_A$ &$\teff_B/\logg_B$
277: &L$_A$/L$_B$ &Ref.\\
278: &(\AA) & &(K)/(cgs) &(K)/(cgs) &6402\AA & \\
279: \hline
280: HR\,7338 &4481 &3.16 &10\,250/3.8 &8\,500/4.0 &2.72 & [1] \\
281: HR\,1800 &$H_p$ &2.45 &11\,050/3.8 &9\,500/4.0 &2.34 & [2] \\
282: $\kappa$\,Cnc &5480 &11.5 &13\,200/3.7 &8\,500/4.0 &10.70 & [3]
283: \\
284: HR\,4072 &4520 &5.45 &10\,650/3.8 &8\,800/4.2 &5.01 & [4] \\
285: $\chi$\,Lup &4520 &3.65 &10\,650/3.9 &9\,200/4.2 &3.35 & [4] \\
286: $\iota$\,CrB &4520 &2.70 &11\,000/4.0 &9\,000/4.3 &2.46 & [4] \\
287: 112\,Her &4520 &6.20 &13\,100/4.1 &8\,500/4.2 &5.20 & [5] \\
288: \hline
289: \end{tabular}
290:
291: \begin{minipage}{10.4cm}
292: Note: The entry for HR\,1800 is the ratio quoted for the broadband $H_p$
293: filter, which we assume to be the light ratio at H$\beta$.
294: References: [1] \citet{petrie}; [2] \citet{ESA1997}; [3] \citet{r98}; [4]
295: \citet{harman} and \citet*{jda99}; [5] \citet{r96}.
296: \end{minipage}
297: \label{table:binary}
298: \end{table*}
299:
300: \subsection{Atomic data for the LTE approximation}
301: \label{sub:atomic}
302:
303: As all the Ne lines in the stars observed are either weak or, except for
304: the hottest stars such as $\tau$ Her, not strongly saturated, the main
305: atomic parameter of critical importance for LTE calculations is the
306: oscillator strength, given as $\log gf$ in Table~\ref{table:eqwid}. We
307: take our oscillator strengths from the calculations of \citet{mjs}, who
308: showed that his calculations were in excellent agreement (within 10 per
309: cent) with other recent theoretical and laboratory data such as that of
310: \citet{hartmetz} for the 3s-3p transitions of interest in this work, and
311: also in excellent agreement with the critically evaluated $gf$-values as
312: given by \citet{auer73}.
313:
314: For the radiative damping, we assumed the classical damping constant
315: $\Gamma_{\rm R} = 2.223\times10^{7}/\lambda^2$ s$^{-1}$ [$\lambda$ in
316: $\mu$m]. This is a good approximation (within a factor of 2) for these
317: lines as the typical lifetime of the upper levels is about 20 ns, and the
318: abundances are not sensitive to the adopted values in any event. Van der
319: Waals contributions to line broadening are also expected to be very small;
320: a suitable approximation by \citet{W67} was used. For Stark broadening,
321: we adopted the recent experimental results of \citet*{delval}, and used an
322: estimate of the temperature scaling factor proportional to $T^{0.4}$ to
323: convert their $w_m$ at 18\,000\,K to values at 12\,000\,K, by multiplying
324: by an average factor of 0.85 \citep{griem74}. One line, $\lambda$5852,
325: was not included in their list and we adopted the simple approximation
326: given in CD23 data (\citealt{cd23}). In general, the measured values
327: which we used are about 3 times the values in CD23 for lines in common. We
328: carried out worst-case sensitivity tests by varying the
329: \citeauthor{delval} Stark broadening by a factor of 2 for the strongest
330: lines in $\tau$ Her; the largest effect on derived abundances was less
331: than 0.01 dex.
332:
333: \subsection{Equivalent widths and LTE results}
334: \label{sub:ew_LTEabunds}
335:
336: Estimated abundances for several identified \nei\ lines were determined
337: using the exact curve-of-growth technique in the LTE approximation. We
338: measured the equivalent widths, $W_{\lambda}$, of Ne absorption lines in
339: the programme spectra by numerical integration in the {\sc dipso v3.5}
340: package \citep{dipso} and compared them to the calculated values for each
341: line, which were generated by our spectrum-synthesis code {\sc uclsyn}
342: \citep{sd88,kcs92}. The necessary atmospheric parameters given in
343: Table~\ref{table:stars} -- \teff, \logg, and microturbulence ($\xi$) --
344: were taken from \citet{sd93}, except for 112\,Her where we used the values
345: given by \citet*{r96}. In the cases of the seven binaries with double
346: spectra, we adopted the light ratios cited in Section~\ref{sub:params} in
347: order to correct for dilution effects. The equivalent widths (corrected
348: for binarity where necessary) and LTE abundances for several lines are
349: given in Table~\ref{table:ewidths}. We used the 2\,km\,s$^{-1}$ grid of
350: \citet{cd13} models, interpolating to produce a model at the chosen \teff\
351: and \logg\ of each star.
352:
353: We searched a list of \nei\ lines from \citet*{wsg66} in the range
354: $\lambda\lambda5800-6800$ and narrowed the list to include only those
355: which were fairly strong, without evident blending problems, and not
356: situated at the ends of \'{e}chelle orders where the spectra are noisiest.
357: In a few cases the quality of the spectra justify quoting equivalent
358: widths to the nearest 0.1\,m\AA. For each star a mean LTE abundance,
359: weighted by equivalent width, was calculated on the scale $\log
360: N(\rm{H})=12.00$. To investigate the consistency of the results from the
361: selected lines, the deviations of each line from the mean of all the
362: lines, $\log (A/A_{\star})$, were calculated for each star where a
363: meaningful average could be computed. These are summarized in
364: Table~\ref{table:ewidths} where the values of $\overline{\log
365: (A/A_{\star})}$ represent the mean deviations from the overall LTE
366: abundance for each line. The small mean deviations imply that the results
367: for each line are broadly consistent with one another and the relative
368: $gf$-values of \citet{mjs}. However, one line ($\lambda$6402) is
369: considerably stronger than all the others, and is well-suited for
370: abundance determinations in the largest number of stars, especially those
371: at the low-\teff\ end of the sequence and with abundances apparently below
372: the standard value. In the remainder of this paper, we shall consider only
373: this line, although future investigators may wish to consider some of the
374: other lines further.
375:
376: \begin{table*}
377:
378: \caption{Ne I equivalent widths (m\AA) and LTE abundances for normal and
379: HgMn programme stars on the scale $\log N ({\rm H}) = 12$}
380:
381: \label{table:eqwid}
382: \begin{tabular}{lcccccccccccccc}
383:
384: \hline
385:
386: &\multicolumn{2}{c}{$\lambda$5852.49}&
387: \multicolumn{2}{c}{$\lambda$6096.16}&
388: \multicolumn{2}{c}{$\lambda$6266.50}&
389: \multicolumn{2}{c}{$\lambda$6382.99}&
390: \multicolumn{2}{c}{$\lambda$6402.25}&
391: \multicolumn{2}{c}{$\lambda$6598.95}&
392: \multicolumn{2}{c}{$\lambda$6717.04}\\
393:
394: Star &$W_\lambda$ & $\log A$ &$W_\lambda$ & $\log A$ &$W_\lambda$ & $\log A$
395: & $W_\lambda$ &$\log A$ & $W_\lambda$&$\log A$ &$W_\lambda$ & $\log A$
396: & $W_\lambda$ &$\log A$ \\
397: \hline
398: &\multicolumn{13}{c}{\it{Normal and superficially normal stars}}&\\
399: $\pi$\,Cet & 8&8.13&10&8.06&15&8.43&18&8.44&39&8.54&--&--&11&8.37\\
400: 134\,Tau &--&-- &--&-- &--&-- &--&-- &$\leq15$&$\leq8.78$
401: &--&-- &--&-- \\
402: $\tau$\,Her &22&8.52&20&8.25&--&-- &26&8.41&59&8.67&--&-- &30&8.76\\
403: $\zeta$\,Dra &13:&8.49: &20:&8.59: &--&-- &24:&8.73:&30&8.32 &--&--
404: &--&-- \\
405: $\alpha$\,Lyr &--&-- &--&-- &--&-- &--&-- &$<10 $&$<9.18$&--&--&--\\
406: HR\,7098 &--&-- &--&-- &--&-- &--&-- &1.8:&7.55:&--&--&--&-- \\
407: 21\,Aql &10&8.21&12&8.12& 9&8.07&12&8.12&35&8.38&--&-- &13&8.43\\
408: HR\,7338$^a$ &--&--&--&--&--&--&--&--&4.0&8.05&--&--&--&--\\
409: $\nu$\,Cap &--&--&--&--&--&--&--&--&$\leq10$&$\leq8.69$&--&--&--&--
410: \\
411: HR\,7878 & 8&8.18&14&8.33& 8&8.10&16&8.42&30&8.34& 4&7.90&10&8.38\\
412: 21\,Peg&5&8.65&4&8.37&$\leq4$&$\leq8.52$&$\leq2$&$\leq8.03$&9&8.36&
413: $\leq2$&$\leq8.40$&$\leq2$&$\leq8.35$\\[1mm]
414: &\multicolumn{13}{c}{\it{HgMn stars}}&\\
415: 87\,Psc &--&--&--&--&--&--&--&--&$\leq8$&$\leq7.41$&--&--&--&--\\
416: 53\,Tau &--&--&--&--&--&--&--&--&$\leq5$&$\leq7.63$&--&--&--&--\\
417: $\mu$\,Lep &--&--&--&--&--&--&--&--&$\leq10$&$\leq7.60$&--&--&--&--\\
418: HR\,1800$^a$ &--&--&--&--&--&--&--&--&$\leq7$&$\leq8.02$
419: &--&--&--&--\\
420: 33\,Gem &19&8.47&27&8.52&12&8.08&14&8.04&38&8.22&15&8.34&17&8.41\\
421: HR\,2676 &--&--&--&--&--&--&--&--&25&7.83&--&--&--&--\\
422: HR\,2844 &--&--&--&--&--&--&--&--&$\leq16$&$\leq7.71$&--&--&--&--\\
423: $\nu$\,Cnc &--&--&--&--&--&--&--&--&$\leq5$&$\leq8.05$&--&--&--&--\\
424: $\kappa$\,Cnc$^a$
425: &21&8.69&22&8.53&22&8.65&21&8.49&38&8.44&12&8.38&20&8.71\\
426: HR\,4072$^a$ &--&--&--&--&--&--&--&--&4.1&8.04&--&--&--&--\\
427: $\chi$\,Lup$^a$ &--&--&--&--&--&--&--&--&4.1&7.95&--&--&--&--\\
428: $\iota$\,CrB$^a$&--&--&--&--&--&--&--&--&3.7&7.76&--&--&--&--\\
429: $\upsilon$\,Her
430: &--&--&--&--&--&--&--&--&$\leq1.0$&$\leq6.63$&--&--&--&--\\
431: $\phi$\,Her &--&--&--&--&--&--&--&--&2.0&7.17&--&--&--&--\\
432: HR\,6997 &15&8.30&19&8.24&20:&8.38:&20&8.25&37&8.15&--&--&--&--\\
433: 112\,Her$^a$ &2.4&7.64&--&--&1.8&7.43&--&--&7.1&7.42&$\leq1.2$
434: &$\leq7.43$&$\leq3.6$&$\leq7.91$\\
435: HR\,7143 &--&--&$\leq1:$&$\leq7.3:$&$\leq1:$&$\leq7.4:$
436: &$\leq2:$&$\leq7.6:$&$\leq3.0$&$\leq7.21$&$\leq1:$&$\leq7.6:$&--&--\\
437: HR\,7361 &18&8.47&23&8.44&18&8.38&20&8.33&40&8.35&13&8.30&--&--\\
438: 46\,Aql &$\leq2$&$\leq7.4$&2:&7.2:&3:&7.4:&3:&7.4:
439: &9&7.41&2:&7.5:&$\leq2$&$\leq7.4$\\
440: HR\,7664 &8&8.07&13&8.16&11&8.17&12&8.10&29&8.16&8&8.13&10&8.24\\
441: HR\,7775 &$\leq3$&$\leq8.42$&$\leq2$&$\leq8.05$&$\leq2$&$\leq8.22$
442: &6&8.72&--&--&$\leq1$&$\leq8.16$&$\leq4$&$\leq8.79$\\
443: \\
444: $\overline{\log (A/A_{\star})}$&& -0.02&&
445: -0.05&&-0.05&&-0.04&&-0.02&&-0.13&&+0.10\\\\
446: $\log gf$ &&-0.49&&-0.31&&-0.37&&-0.24&&+0.33&&-0.35&&-0.35\\
447:
448: \hline
449: \end{tabular}
450: \begin{minipage}{18cm}
451: $^a$Binaries with two spectra. The $W_\lambda$ values are corrected for
452: dilution effects as described in the text. Colons (`:') indicate
453: uncertain values.
454: \end{minipage}
455: \label{table:ewidths}
456: \end{table*}
457:
458: \subsection{A weak blending line?}
459: \label{blend}
460:
461: Although we have chosen our list of Ne lines to be as blend-free as
462: possible, there is a predicted weak blending feature in the CD23 list
463: \citep{cd23} adjacent to the important line \nei\ $\lambda$6402.246:
464: Fe\,{\sc ii} at $\lambda$6402.397. In programme stars
465: (Table~\ref{table:stars}) with approximately standard or lower Fe
466: abundances, this line would have a typical strength of about 0.5~m\AA, too
467: small to affect our results in any significant way. However, in the few
468: stars with enhanced Fe abundances such as 112\,Her \citep[][$\log A({\rm
469: Fe})=8.40$]{sd93}, the strongest iron-rich case, the possible blending
470: effect would have raised the apparent abundance of Ne by 0.3\,dex. It
471: should be noted that for sharp-lined stars, the displacement of the blend
472: is enough to make its existence apparent. The existence of this line with
473: the $gf$-value listed remains to be confirmed; future work directed at
474: refining the neon analysis should address the question of its actual
475: strength with better observations.
476:
477: \section{NLTE calculations and abundances}
478: \label{section:NLTE}
479: \subsection{Calculations and \nei\ model atom}
480: \label{sub:calcsNLTE}
481:
482: The first full NLTE calculations of neon line strengths for \nei\ were
483: made by \citet{auer73} using NLTE model atmospheres. Unfortunately for
484: our purposes, their analysis was restricted to the hotter stars with
485: $\teff>15000$\,K. Recently, their calculations were revisited and
486: extended to cooler temperatures by \citet{sigut}. \citeauthor{sigut} used
487: the $T-\tau$ relations, particle densities, and electron number densities
488: from \citeauthor{cd13} LTE line-blanketed models and solved the restricted
489: NLTE problem, i.e. only the equations of radiative tranfer and statistical
490: equlibrium for Ne. His grid of equivalent widths also has rather large
491: steps for our purposes: 2000\,K, 0.5\,dex, 0.5\,dex in temperature,
492: gravity and Ne abundance, respectively.
493:
494: In this section we examine in detail the temperature, gravity and
495: Ne-abundance region where all our HgMn stars are found, solve the full
496: NLTE problem using NLTE model atmospheres, and find a convenient way to
497: represent the NLTE effects in the \nei\ $\lambda$6402 line so that
498: straightforward interpolation via LTE models can be done. For the
499: calculation of NLTE atmosphere models and level populations we used the
500: {\sc {tlusty195}} code described in more detail in \citet{hub88}, and in
501: \citet{hula92,hula95}. Here H\,{\sc{i}} and \nei\ were treated as explicit
502: ions, which means that their level populations were calculated in NLTE and
503: their opacity was considered. Other elements like He, C, N, and O were
504: allowed to contribute to the particle and electron number density in LTE.
505: Synthetic spectra and equivalent widths were then calculated with the {\sc
506: {synspec42}} code \citep*{hulajef95}. In the following, if not stated
507: otherwise, `in LTE' means `in LTE considering the NLTE model of the
508: atmosphere'.
509:
510: It is not possible to list here all the input parameters entering the NLTE
511: calculations and above mentioned codes. We will mention only those which
512: are most crucial for this particular problem or are different from those
513: which could be generated interactively, e.g. by the very useful interface
514: tool {\sc modion} \citep{varosi}, part of the {\sc tlusty} package for
515: creating the model of the atom from the TOPbase data. We provide a copy of
516: our input model for {\sc tlusty}, which can be downloaded by anyone who
517: wishes to repeat the calculations (Dworetsky \& Budaj 2000).
518:
519: \begin{table}
520: \caption{\nei\ energy levels considered. Column 1: the Paschen
521: level designation, col. 2: the $nlpqr$ notation of Seaton (1998),
522: col. 3: ionization energy in cm$^{-1}$, col. 4: statistical weight of the
523: level.}
524: \begin{center}
525: \begin{tabular}{llll}
526: \hline
527: Paschen & Seaton &Energy & g \\
528: \hline
529: $2p_6$ 1S &2p & 174192.4 & 1. \\
530: $1s_5$ &3s332& 40148.6 & 5. \\
531: $1s_4$ &3s331& 39731.2 & 3. \\
532: $1s_3$ &3s110& 39371.8 & 1. \\
533: $1s_2$ &3s111& 38301.7 & 3. \\
534: $2p_{10}$ &3p311& 25932.7 & 3. \\
535: $2p_9$ &3p353& 24533.4 & 7. \\
536: $2p_8$ &3p352& 24366.2 & 5. \\
537: $2p_7$ &3p331& 24068.8 & 3. \\
538: $2p_6$ &3p332& 23874.6 & 5. \\
539: $2p_5$ &3p131& 23418.3 & 3. \\
540: $2p_4$ &3p132& 23331.9 & 5. \\
541: $2p_3$ &3p310& 23273.0 & 1. \\
542: $2p_2$ &3p111& 23152.0 & 3. \\
543: $2p_1$ &3p110& 21219.7 & 1. \\
544: $2s_5$ &4s332& 15589.3 & 5. \\
545: $2s_4$ &4s331& 15394.4 & 3. \\
546: $2s_3$ &4s110& 14810.5 & 1. \\
547: $2s_2$ &4s111& 14655.8 & 3. \\
548: $3d_6$ &3d310& 12680.8 & 1. \\
549: $3d_5$ &3d311& 12666.3 & 3. \\
550: $3d_4'$ &3d374& 12600.1 & 9. \\
551: $3d_4$ &3d373& 12598.3 & 7. \\
552: $3d_3$ &3d332& 12583.2 & 5. \\
553: $3d_2$ &3d331& 12553.8 & 3. \\
554: $3d_1''$ &3d352& 12490.8 & 5. \\
555: $3d_1'$ &3d353& 12489.0 & 7. \\
556: $3s_1''''$ &3d152& 11781.8 & 5. \\
557: $3s_1'''$ &3d153& 11780.3 & 7. \\
558: $3s_1''$ &3d132& 11770.5 & 5. \\
559: $3s_1'$ &3d131& 11754.8 & 3. \\
560: \hline
561: \end{tabular}
562: \end{center}
563: \label{table:levels}
564: \end{table}
565:
566: The spectrum of \nei\ is that of an inert gas where the LS-coupling breaks
567: down and terms and multiplets do not provide an appropriate description of
568: the atom, so that at least the lower terms and multiplets must be split
569: into individual levels and transitions. We considered explicitly the first
570: 31 levels of \nei\ as \citet{auer73}, plus continuum, with each of the
571: levels treated separately (Table \ref{table:levels}). Paschen designations
572: and experimental energies for the levels were taken from \citet{moore}.
573: Photoionization cross-sections for the terms were taken from the TOPbase
574: data base \citep{cunto}, as calculated by \citet{hibb}. We fit individual
575: photoionization cross-sections to about 10-15 points using the {\sc
576: modion} code. It was assumed that the photoionization cross-section is
577: the same for the term as for the level but it was scaled to the particular
578: level threshold. We used the calculated oscillator strengths of
579: \citet{mjs} as before. For collisional excitation rates we used the van
580: Regemorter formula as in \citet{auer73}. For collisional ionization we
581: used equation (5.79) of \citet{mihalas78} with $\bar{g}_{i}=0.1$. With
582: this input data we calculated the \nei\ $\lambda$6402 equivalent width for
583: the same abundance ($10^{-4}$), $f$-value (0.431), microturbulence $\xi$
584: (4\,km\,s$^{-1}$) and similar H-He NLTE models ($\teff=15000, 20000, \log
585: g=4$) as \citeauthor{auer73} to check and compare our calculations. (For
586: the hotter model, two terms of Ne\,{\sc ii} plus continuum were also
587: considered explicitly.) The results are listed in
588: Table~\ref{table:auer_compare} and are in very close agreement. To compare
589: our results with those by \citet{sigut}, and to check the calculations for
590: lower temperatures, we solved the similar {\em restricted} NLTE problem
591: using \citet{cd13} CD13 LTE line blanketed models (computed with $\xi=2$
592: km\,s$^{-1}$) with our \nei\ atom model plus Hubeny's H\,{\sc{i}} atom
593: model (9 explicit levels plus continuum) and the same Ne abundance ($1.12
594: \times 10^{-4}$), $f$-value (0.428), and $\xi$ (5\,km\,s$^{-1}$) as
595: \citeauthor{sigut}. The results are compared in
596: Table~\ref{table:sigut_compare} and are also in very good agreement.\\
597:
598: \begin{table}
599: \caption{Equivalent widths of \nei\ $\lambda$6402 (m\AA) for two
600: models, comparing this work (D+B) with \citet{auer73}.}
601: \begin{center}
602: \begin{tabular}{ccccc}
603: \hline
604: \teff\ & \multicolumn {2}{c}{15000} & \multicolumn{2}{c}{20000} \\
605: \hline
606: <E & NLTE & LTE & NLTE \\
607: \hline
608: D+B & 30 & 46 & 39 & 74\\
609: A+M & 28 & 45 & 40 & 79\\
610: \hline
611: \end{tabular}
612: \end{center}
613: \label{table:auer_compare}
614: \end{table}
615:
616: \begin{table}
617: \caption{Equivalent widths of \nei\ $\lambda$6402 (m\AA) for two
618: \citeauthor{cd13} models with $\teff=12000\, \&\, 17000$\,K, $\log g=4$,
619: $\xi=5$\,km\,s$^{-1}$, comparing this work (D+B) with \citet{sigut}.}
620: \begin{center}
621: \begin{tabular}{ccccc}
622: \hline
623: \teff\ & \multicolumn{2}{c}{12000} & \multicolumn{2}{c}{17000} \\
624: \hline
625: & NLTE & NLTE/LTE & NLTE & NLTE/LTE \\
626: \hline
627: D+B & 19 & 1.31 & 89 & 1.81\\
628: Sigut & 18 & - & 86 & 1.80\\
629: \hline
630: \end{tabular}
631: \end{center}
632: \label{table:sigut_compare}
633: \end{table}
634:
635: We found that although microturbulence can affect the
636: equivalent widths due to desaturation effects, for the observed stars
637: ($0 \leq\xi\leq 2.5$) it has negligible effect on
638: the atmosphere model, \nei\ level populations, and LTE/NLTE equivalent
639: width ratio, $R = W_{\lambda({\rm LTE})}/W_{\lambda({\rm NLTE})}$. For
640: this reason, our task can be considerably simplified as one can calculate
641: a grid of models and LTE/NLTE ratios for only one $\xi=0$. Also, the NLTE
642: effect of varying the Ne abundance is quite small. While decreasing the Ne
643: abundance from the standard value (8.08) by 1.0 dex reduces the equivalent
644: width considerably, it raises the corresponding LTE/NLTE ratio $R$ by only
645: $0.03\pm0.01$. Consequently, the main results can be gathered into
646: Table~\ref{table:ratios} listing $R(\teff, \logg)$ for standard Ne
647: abundance and zero microturbulence. The ratio $R$ ranges from around
648: $0.6-0.7$ at the high \teff\ end of the HgMn domain to nearly 1.0 at the
649: cool end of the sequence (where the lines of neon disappear and can no
650: longer be studied). \citet{dwbu2000} provide a short Fortran77 code to
651: interpolate in the grid (including the small effects of abundance) for
652: anyone who wishes to undertake their own interpolations for $R$.
653:
654: \begin{table}
655: \caption{LTE/NLTE equivalent width ratio $R$ (\nei\ $\lambda$6402).}
656: \begin{center}
657: \begin{tabular}{crllll}
658: \hline
659: \teff &\logg = & 3.50 & 3.75 & 4.00 & 4.25 \\
660: \hline
661: 11000 & & 0.78 & 0.81 & 0.84 & 0.87\\
662: 12000 & & 0.72 & 0.75 & 0.79 & 0.82\\
663: 13000 & & 0.67 & 0.70 & 0.74 & 0.77\\
664: 14000 & & 0.63 & 0.66 & 0.69 & 0.73\\
665: 15000 & & 0.59 & 0.62 & 0.65 & 0.69\\
666: \hline
667: \end{tabular}
668: \end{center}
669: \label{table:ratios}
670: \end{table}
671:
672: \citet{hub81} pointed out that apart from hydrogen, C\,{\sc{i}} and
673: Si\,{\sc{i}} are the other most important ions to be included explicitly
674: in the NLTE calculations in early A stars. We have checked that for this
675: particular problem, including these elements along with He may again
676: slightly affect the equivalent width, but it does not affect $R$
677: significantly. The line strength results were somewhat sensitive to the
678: collisional excitation. A more detailed analysis of the current
679: precision of neon NLTE calculations can be found in \citet{sigut}.
680:
681: \begin{table}
682:
683: \caption{Non-LTE abundances from \nei\ $\lambda$6402 for normal and HgMn
684: programme stars on the scale $\log A ({\rm H}) = 12.00$. Programme stars
685: with high abundance upper limits or highly uncertain measurements have
686: been omitted.}
687:
688: \begin{tabular}{lccccc}
689:
690: \hline
691:
692: Star &\teff&$\logg$&$W_\lambda$&$W_{\rm LTE}/W_{\rm NLTE}$&$\log A$ \\
693: \hline
694: \multicolumn{6}{c}{\it{Normal and superficially normal stars}}\\
695: $\pi$\,Cet & 13250 & 3.80 & 39 & 0.69 & 8.18 \\
696: $\tau$\,Her & 15000 & 3.95 & 59 & 0.64 & 8.15 \\
697: $\zeta$\,Dra & 12900 & 3.90 & 30 & 0.73 & 8.07 \\
698: 21\,Aql & 13000 & 3.50 & 35 & 0.67 & 8.01 \\
699: HR\,7338 & 10250 & 3.75 & 4.0& 0.86 & 8.04 \\
700: HR\,7878 & 13050 & 3.85 & 30 & 0.71 & 8.04 \\
701: 21\,Peg & 10450 & 3.50 & 9 & 0.82 & 8.22 \\[1mm]
702: \multicolumn{6}{c}{\it{HgMn stars}}\\
703: 87\,Psc & 13150 & 4.00 & $\leq8$ &0.76 & $\leq7.27$ \\
704: 53\,Tau & 12000 & 4.25 & $\leq5$ &0.83 & $\leq7.53$ \\
705: $\mu$\,Lep & 12800 & 3.85 & $\leq10$&0.75 & $\leq7.43$ \\
706: HR\,1800 & 11050 & 3.80 & $\leq7$ &0.83 & $\leq7.90$ \\
707: 33\,Gem & 14400 & 3.85 & 38 & 0.67 & 7.87 \\
708: HR\,2676 & 14050 & 3.60 & 25 & 0.66 & 7.54 \\
709: HR\,2844 & 13460 & 3.80 &$\leq16$&0.72&$\leq7.50$ \\
710: $\nu$\,Cnc & 10400 & 3.60 &$\leq5$ &0.84&$\leq7.94$ \\
711: $\kappa$\,Cnc & 13500 & 3.80 & 38 & 0.69 & 8.08 \\
712: HR\,4072 & 10500 & 3.95 & 4.1& 0.87 & 7.96 \\
713: $\chi$\,Lup & 10750 & 4.00 & 4.1& 0.87 & 7.87 \\
714: $\iota$\,CrB & 11000 & 4.00 & 3.7& 0.85 & 7.67 \\
715: $\upsilon$\,Her & 12000 & 3.80 &$\leq1.0$&0.80&$\leq6.53$ \\
716: $\phi$\,Her & 11650 & 4.00 & 2.0 & 0.83 & 7.08 \\
717: HR\,6997 & 14500 & 3.90 & 37 & 0.67 & 7.82 \\
718: 112\,Her & 13100 & 4.10 & 7.1 & 0.77 & 7.28 \\
719: HR\,7143 & 12100 & 4.00 &$\leq3$&0.81&$\leq7.10$ \\
720: HR\,7361 & 13650 & 3.55 & 40 & 0.65 & 7.94 \\
721: 46\,Aql & 13000 & 3.65 & 9 & 0.72 & 7.23 \\
722: HR\,7664 & 13200 & 3.60 & 29 & 0.68 & 7.85 \\
723:
724: \hline
725: \end{tabular}
726:
727: \label{table:nlte_abunds}
728:
729: \end{table}
730:
731: \subsection{Neon abundances from Non-LTE calculations}
732: \label{nlte_abund}
733:
734: One may obtain a `corrected' LTE equivalent width from $R W_\lambda({\rm
735: obs})$ and analyse the corrected width by using a standard LTE approach
736: including the appropriate microturbulence in a fully line-blanketed case.
737: We used {\sc uclsyn} to calculate the abundance of Ne from equivalent
738: widths given in Table~\ref{table:ewidths} after scaling by $R$ from
739: Table~\ref{table:ratios}. The microturbulence parameters $\xi$ were taken
740: from Table~\ref{table:stars}. The results are shown in
741: Table~\ref{table:nlte_abunds} and plotted as a function of \teff\ in
742: Figure~\ref{fig:nlte_abunds}.
743:
744: \begin{figure}
745: \begin{center}
746: \epsfig{file=neonplot.ps,width=185pt,clip=,angle=270,silent=}
747: \caption{Abundances of Ne in normal stars (open circles) and HgMn stars
748: (filled circles). Upper limits for abundances are indicated by arrows.
749: The standard abundance for Ne on the scale $\log A({\rm H})=12.00$ is from
750: \citet{grev96}.}
751: \label{fig:nlte_abunds}
752: \end{center}
753: \end{figure}
754:
755: Our results yield for normal and superficially-normal stars a mean
756: abundance $\log A ({\rm Ne}) = 8.10\pm0.03$ relative to 12.00 for H.
757: This is in excellent agreement with the standard abundance of Ne, 8.08,
758: given by \citet{grev96}, which is essentially identical to the value
759: (8.09) in the earlier compilation of \citet{ag89}. The standard abundance
760: is partly based on local galactic values (stars and nebulae), and on
761: application of a well-determined correction to solar wind and solar
762: energetic particle values, as well as spectroscopy of solar prominences.
763: We take this agreement as confirmation that our ratio method for LTE/NLTE
764: equivalent width scaling works well in the \teff\ range of HgMn stars.
765:
766: The HgMn stars are, with only one exception ($\kappa$\,Cnc), deficient in
767: Ne, although the deficits in a few cases are only marginal ($0.1-0.2$
768: dex). In many cases we are able only to establish upper limits for Ne
769: abundances. The most extreme case is $\upsilon$\,Her, for which we have
770: particularly good spectra and were able to establish an upper limit 1.5
771: dex below the solar abundance. There is no case in our sample where Ne
772: has an abundance greater than the standard value. The results of Adelman
773: \& Pintado (2000) can also be analysed using our method.
774:
775: \section{Discussion}
776: \label{disc}
777:
778: The error found for the average abundance for the normal stars is based on
779: the scatter in the results. For comparison, following procedures adopted
780: in previous papers in this series on HgMn stars (e.g.,
781: \citealt{sd93,jda99}) we propagate uncertainties in adopted estimates of
782: the errors on each parameter as follows: $\pm$0.25\,dex in \logg,
783: $\pm$250\,K in \teff, $\pm$0.5\,km\,s$^{-1}$ in the microturbulence
784: ($\xi$), and $\pm$5 percent in $W_{\lambda}$. Propagating these errors
785: through the `corrected' LTE analysis used above for the \nei $\lambda$6402
786: line, using a model atmosphere at \teff\ = 13\,000\,K, \logg\ = 4.0,
787: $\xi$=1, leads to the following representative errors in the derived Ne
788: abundances: $\pm$0.10\,dex (\logg); $\pm$0.08\,dex (\teff); $\pm$0.01\,dex
789: ($\xi$); $\pm$0.04\,dex ($W_{\lambda}$) at standard abundance
790: ($W_{\lambda}=27$\,m\AA), and $\pm$0.10\,dex (\logg); $\pm$0.07\,dex
791: (\teff); $\pm$0.01\,dex ($\xi$); $\pm$0.03\,dex ($W_{\lambda}$) with neon
792: underabundant by 0.5 dex ($W_{\lambda}=13$\,m\AA). These are very similar
793: ranges; the combined expected error for one measurement is $\pm0.13$\,dex.
794: The s.d. for the normal stars is $\pm0.08$\,dex. This difference may
795: reflect overestimates in some of the above factors (especially $\Delta\log
796: g$) that comprise the estimated errors, but the two estimates are not in
797: serious disagreement.
798:
799: We have implicitly conducted all our analyses under the assumption of a
800: homogeneous depth distribution of neon in the photospheres of the HgMn
801: stars. It now seems well-founded to conclude that in many HgMn stars the
802: neon atoms may not be distributed with a constant fraction vs. optical
803: depth, due to gravitational settling, but our results offer no method of
804: distinguishing clearly between a uniform depletion in the line forming
805: region and an inhomogeneous distribution in which the total number of
806: absorbers is about the same. Given the scatter in abundance from star to
807: star, line strengths alone will be inadequate to prove the point one way
808: or the other.
809:
810: One should consider the question of whether or not we may lump together
811: the normal and superficially normal stars. The `superficially normal'
812: stars listed in Tables~\ref{table:stars},~\ref{table:ewidths} and
813: \ref{table:nlte_abunds} were originally described as such by
814: \citet{cow80}, owing to their relatively sharp lines, but subsequent
815: investigations by \citet{sd90,sd93} and \citet{kcs93,kcs94,kcs96,kcs97}
816: have shown that HR\,7338, HR\,7878, and 21\,Peg have abundances which are
817: not distinguishable from normal stars for C, N, Cr$-$Ni, Mg, Al, Si, Cu,
818: Zn, Ga and Hg. \citeauthor{cow80} thought that HR\,7878 and HR\,7338 were
819: normal stars with no trace of peculiarity, although he suspected that
820: 21\,Peg might be related to early Am stars due to a weaker than expected
821: Sc~{\sc{ii}} line, and \citet{sad80} found that both 21\,Peg and HR\,7338
822: may have `hot Am' characteristics, such as mild Ba and Y enhancements. One
823: of these, 21\,Peg, is listed as Hg peculiar class by \citet{renson}
824: without a reference cited, but Smith's (\citeyear{kcs97}) study of
825: Hg~{\sc{ii}} lines showed that the Hg abundance is effectively
826: indistinguishable from that found in normal stars. \cite{land98} also
827: describes it as a normal star. In what follows we assume that 21\,Peg is
828: a normal B9.5\,V star and that HR\,7338 is also normal. In this work we
829: therefore feel justified in including these three stars with the normal
830: stars in Table~\ref{table:nlte_abunds} and Fig.~\ref{fig:nlte_abunds}.
831:
832: The equivalent width measures of \citet{adel00} for 3\,Cen\,A,
833: $\kappa$\,Cnc, and HR\,7245 can be used with the results of
834: Section~\ref{section:NLTE} to derive approximate NLTE abundances for neon.
835: For the first two of these stars, those authors give only the equivalent
836: width of $\lambda$5852.49 (although their table headings say 5842.49,
837: which is evidently a misprint). We assume, based on the discussion of
838: Section~\ref{sub:ew_LTEabunds}, that the correction factor $R$ may be
839: taken to be about the same as for $\lambda$6402, and we further assume
840: that we can extrapolate the correction factor to \teff\ = 17\,500 for
841: 3\,Cen\,A, for which we assume $R=0.59$. This well-known peculiar He-weak
842: star has some characteristics similar to HgMn stars. We find a
843: near-standard Ne abundance of 8.17, while $\kappa$\,Cnc has a slight
844: overabundance (8.32), and the HgMn star HR\,7245, which has a measured
845: equivalent width of 8 m\AA, has a low abundance (7.32) similar to that of
846: 112\,Her. Given that our assumptions above could be subject to some
847: uncertainty, at this stage we would not wish to conclude much more than
848: that the abundance of neon seems consistent with the standard value in
849: 3\,Cen\,A and $\kappa$\,Cnc, but in the case of HR\,7245 we are probably
850: on firm ground in assigning a very low abundance of neon.
851:
852: We explored briefly the question of whether the Ne abundances in HgMn
853: stars depend on atmospheric parameters. It seems that largest anomalies
854: (underabundances) are generally observed in the middle of the temperature
855: range of HgMn stars ($11500<\teff<13000$\,K; see
856: Figure~\ref{fig:nlte_abunds}). No apparent correlation with the surface
857: gravity can be seen in Table~\ref{table:nlte_abunds}. It is not possible
858: to draw any conclusions about dependences on rotational velocity as we
859: have chosen to work with a selected sample of HgMn stars with fairly small
860: \vsini\ in order to ensure accurate abundance determinations.
861:
862: \section{Conclusions}
863:
864: We have measured the equivalent widths (or upper limits) of several \nei\
865: lines in the spectra of 11 normal late B stars and 21 HgMn stars in the
866: same \teff\ range. These lines were selected after a search for lines
867: which were well placed in \'{e}chelle orders in the HES and which appeared
868: not to have any significant blending features in the sample of stars
869: studied. When analysed using LTE methods in fully line-blanketed
870: atmospheres, there is a steady increase with \teff\ in the apparent
871: abundance above the standard value of $\log A({\rm Ne})=8.08$. It is
872: apparent from previous studies that this is due to NLTE effects. We note
873: that the strongest line of Ne in the red region, $\lambda$6402, and the 6
874: other lines studied, generally give concordant LTE abundances, suggesting
875: that they are affected by NLTE effects by roughly the same amount. These
876: lines may be of use in future investigations, provided further
877: observations and NLTE calculations are made.
878:
879: We undertook a detailed NLTE analysis of $\lambda$6402 by use of a full
880: analysis including NLTE for H {\sc i} and \nei. We confirm earlier
881: studies by obtaining very similar results when similar inputs are used,
882: and find that the ratio of the NLTE and LTE equivalent widths calculated
883: for a given NLTE model atmosphere is a slowly varying function of \teff\
884: and surface gravity. This ratio only slightly depends on the actual
885: abundance of Ne, and is also very insensitive to the microturbulence, so
886: it becomes possible to interpolate, in a table of the ratio $R$, the
887: `correction factor' by which an observed equivalent width must be scaled
888: in order to produce the NLTE abundance from a much easier LTE analysis.
889:
890: The normal stars in our sample yield a mean logarithmic Ne abundance of
891: 8.10, well within our formal mean error of $\pm0.03$ of the standard
892: value, 8.08, given by \citet{grev96}. This gives us additional confidence
893: that our models, and the ratio method of using LTE calculations as an
894: interpolation device, work satisfactorally for late B stars. The
895: smallness of the scatter (s.d. $\pm0.08$) for the individual stars
896: suggests that the error budget in Section~\ref{disc} is rather
897: conservative.
898:
899: It is clear from our results for the HgMn stars that the abundances of Ne
900: range from standard abundance or slightly below, to extreme deficiencies
901: of an order of magnitude or more. In several cases we have obtained only
902: upper limits, and additional observations of very high quality would be
903: needed in order to attempt to detect the weak Ne lines in these stars.
904: There is a tendency for the Ne abundance to be smallest in the middle of
905: the HgMn effective temperature range, but there is no dependence on
906: surface gravity. There is not a single confirmed case in which Ne has an
907: enhanced abundance, which is strong evidence for the absence in HgMn stars
908: of nonturbulent stellar winds (i.e. hydrogen mass loss rate must be $<
909: 10^{-14}M_{\sun}$\,yr$^{-1}$) that might compete with radiative atomic
910: diffusion and produce accumulations of light elements in the photosphere,
911: as suggested by \citet{landolvau}. That such winds might exist was studied
912: by \citet{babmich}, \citet{babel92} and \citet{krtkub}.
913:
914: \section{Acknowledgements}
915:
916: J. Budaj gratefully acknowledges the support of The Royal Society for a
917: Royal Society/NATO Fellowship (ref 98B) and further support from VEGA
918: Grant No. 7107 from the Slovak Academy of Sciences. MMD is grateful to the
919: Director of the University of California Observatories, Prof.\ J.S.
920: Miller, for allocating time for his Guest Observer programme. We
921: gratefully acknowledge the work of C. M. Jomaron and Mr. Christopher Jones
922: for their assistance in this work. Thanks also to the technical and
923: service staff at Lick Observatory, and especially to Tony Misch, for their
924: efforts on our behalf. NSO/Kitt Peak FTS data used for the measurement of
925: parasitic light were produced by NSF/NOAO. We are grateful to I. Hubeny
926: and J. Krti\v{c}ka for numerous discussions. Research on chemically
927: peculiar stars at UCL is supported by PPARC grant GR/K58500. C. Jones was
928: supported by an Undergraduate Research Bursary from the Nuffield
929: Foundation (NUB-URB98). This work was also supported by the PPARC PATT
930: Rolling Grant GR/K60107 to UCL, for travel to telescopes.
931:
932: \begin{thebibliography}{}
933:
934: \bibitem[\protect\citeauthoryear{Adelman}{1988a}]{adel88a}
935: Adelman S.J., 1988a, MNRAS, 235, 749
936:
937: \bibitem[\protect\citeauthoryear{Adelman}{1988b}]{adel88b}
938: Adelman S.J., 1988b, MNRAS, 235, 763
939:
940: \bibitem[\protect\citeauthoryear{Adelman}{1989}]{ad89}
941: Adelman S.J., 1989, MNRAS, 239, 487
942:
943: \bibitem[\protect\citeauthoryear{Adelman \& Fuhr}{1985}]{adfuhr85}
944: Adelman S.J., Fuhr J.R., 1985, A\&A, 152, 434
945:
946: \bibitem[\protect\citeauthoryear{Adelman \& Pintado}{2000}]{adel00}
947: Adelman S.J., Pintado O.I., 2000, A\&A, 354, 899
948:
949: \bibitem[\protect\citeauthoryear{Adelman, Philip \& Adelman}{Adelman et
950: al.}{1996}]{ad96}
951: Adelman S.J., Philip A.G.D., Adelman C.J., 1996, MNRAS, 282, 953
952:
953: \bibitem[\protect\citeauthoryear{Allen}{1998}]{csa98}
954: Allen C.S., 1998, PhD thesis, University of London
955: \newline URL=http://www.ulo.ucl.ac.uk/ulo\_comms\\/80/index.html
956:
957: \bibitem[\protect\citeauthoryear{Anders \& Grevesse}{1989}]{ag89}
958: Anders E., Grevesse N., 1989, Geochim. Cosmochim. Acta, 53, 197
959:
960: \bibitem[\protect\citeauthoryear{Auer \& Mihalas}{1973}]{auer73}
961: Auer L.H., Mihalas D., 1973, ApJ, 184, 15
962:
963: \bibitem[\protect\citeauthoryear{Babel}{1992}]{babel92} Babel J., 1992,
964: A\&A, 258, 449
965:
966: \bibitem[\protect\citeauthoryear{Babel \& Michaud}{1991}]{babmich} Babel
967: J., Michaud G., 1991, ApJ, 366, 560
968:
969: \bibitem[\protect\citeauthoryear{Batten, Fletcher \& MacCarthy}{Batten et
970: al.}{1989}]{batten89}
971: Batten A.H., Fletcher J.M., MacCarthy D.G., 1989, Publs. DAO, 17, 1
972:
973: \bibitem[\protect\citeauthoryear{Bohlender, Dworetsky \&
974: Jomaron}{Bohlender et al.}{1998}]{bohl}
975: Bohlender D.A., Dworetsky M.M., Jomaron C.M., 1998, ApJ, 504, 533
976:
977: \bibitem[\protect\citeauthoryear{Budaj \& Dworetsky}{2000}]{budw00}
978: Budaj J., Dworetsky M. M., 2000, in preparation
979:
980: \bibitem[\protect\citeauthoryear{Churchill}{1995}]{churchill}
981: Churchill C.W., 1995, Lick Obs. Tech. Rep. No. 74
982:
983: \bibitem[\protect\citeauthoryear{Cowley}{1980}]{cow80}
984: Cowley C.R., 1980, PASP, 92, 159
985:
986: \bibitem[\protect\citeauthoryear{Cunto et al.}{1993}]{cunto}
987: Cunto W.C., Mendoza C., Ochsenbein F., Zeippen C.J., 1993, A\&A, 275, L5
988:
989: \bibitem[\protect\citeauthoryear{del Val, Aparicio \& Mar}{del Val et
990: al.}{1999}]{delval}
991: del Val J.A., Aparicio J.A., Mar A., 1999, ApJ, 513, 535
992:
993: \bibitem[\protect\citeauthoryear{Dworetsky \& Budaj}{2000}]{dwbu2000}
994: Dworetsky M. M., Budaj J., 2000, Comm. Univ. London Observatory. No. 81
995: \newline URL=http://www.ulo.ucl.ac.uk/ulo\_comms\\/81/index.html
996:
997: \bibitem[\protect\citeauthoryear{Dworetsky, Jomaron \& Smith}{Dworetsky et
998: al.}{1998}]{djs98} Dworetsky M.M., Jomaron C.M., Smith C.A., 1998, A\&A,
999: 333, 665
1000:
1001: \bibitem[\protect\citeauthoryear{ESA}{1997}]{ESA1997}
1002: ESA (European Space Agency), 1997, The Hipparcos and Tycho Catalogues, ESA
1003: SP-1200
1004:
1005: \bibitem[\protect\citeauthoryear{Gigas}{1986}]{gigas}
1006: Gigas D., 1986, A\&A, 165, 170
1007:
1008: \bibitem[\protect\citeauthoryear{Grevesse, Noels \& Sauval}{Grevesse et
1009: al.}{1996}]{grev96}
1010: Grevesse N., Noels A., Sauval A. J., 1996, in Holt S. \& Sonneborn G.,
1011: eds, Cosmic Abundances, PASPC No. 99, p. 117
1012:
1013: \bibitem[\protect\citeauthoryear{Griem}{1974}]{griem74} Griem H. R., 1974,
1014: Spectral Line Broadening by Plasmas, Academic Press, New York, Appendix
1015: IVa
1016:
1017: \bibitem[\protect\citeauthoryear{Harman}{1997}]{harman}
1018: Harman D.J., 1997, MSci Project Report, University College London
1019:
1020: \bibitem[\protect\citeauthoryear{Hartmetz \& Schmoranzer}{1984}]{hartmetz}
1021: Hartmetz P., Schmoranzer H., 1984, Z. Phys. A, 317, 1
1022:
1023: \bibitem[\protect\citeauthoryear{Hibbert \& Scott}{1994}]{hibb}
1024: Hibbert A., Scott M.P., 1994, J. Phys. B 27, 1315
1025:
1026: \bibitem[\protect\citeauthoryear{Hoffleit \& Warren}{1991}]{bsc91}
1027: Hoffleit D., Warren Jr. W.H., 1991, Bright Star Catalogue, 5th Rev. Ed.,
1028: ftp://cdsarc.u-strasbg.fr/cats/V/50
1029:
1030: \bibitem[\protect\citeauthoryear{Howarth et al.}{1998}]{dipso} Howarth
1031: I.D., Murray J., Mills D., Berry D.S., 1998, Starlink User Note 50,
1032: Rutherford Appleton Lab./CCLRC
1033:
1034: \bibitem[\protect\citeauthoryear{Hubeny}{1981}]{hub81}
1035: Hubeny I. 1981, A\&A, 98, 96
1036:
1037: \bibitem[\protect\citeauthoryear{Hubeny}{1988}]{hub88}
1038: Hubeny I., 1988, Comput. Phys. Comm., 52, 103
1039:
1040: \bibitem[\protect\citeauthoryear{Hubeny \& Lanz}{1992}]{hula92}
1041: Hubeny I., Lanz T., 1992, A\&A, 262, 501
1042:
1043: \bibitem[\protect\citeauthoryear{Hubeny \& Lanz}{1995}]{hula95}
1044: Hubeny I., Lanz T. 1995, ApJ, 439, 875
1045:
1046: \bibitem[\protect\citeauthoryear{Hubeny, Lanz \& Jeffery}{Hubeny et
1047: al.}{1995}]{hulajef95}
1048: Hubeny I., Lanz T., Jeffery C.S., 1995, Tlusty \& Synspec - A User's Guide
1049:
1050: \bibitem[\protect\citeauthoryear{Hubrig \& Launhardt}{1993}]{hl93}
1051: Hubrig S., Launhardt R., 1993, in Dworetsky M.M., Castelli F., Faraggianna
1052: R., eds, Peculiar versus Normal Phenomena in A-Type and Related Stars.
1053: PASPC No. 44, p. 350
1054:
1055: \bibitem[\protect\citeauthoryear{Jomaron, Dworetsky \& Allen}{Jomaron et
1056: al.}{1999}]{jda99} Jomaron C.M., Dworetsky M.M., Allen C.S., 1999, MNRAS,
1057: 303, 555
1058:
1059: \bibitem[\protect\citeauthoryear{Krti\v{c}ka \& Kub\'{a}t}{2000}]{krtkub}
1060: Krti\v{c}ka J., Kub\'{a}t J., 2000, A\&A, in press
1061:
1062: \bibitem[\protect\citeauthoryear{Kurucz}{1990}]{cd23}
1063: Kurucz R.L., 1990, Trans. IAU, XXB, 168 (CD-ROM 23)
1064:
1065: \bibitem[\protect\citeauthoryear{Kurucz}{1993}]{cd13}
1066: Kurucz R.L., 1993, ATLAS9 Stellar Atmosphere Programs and 2 km s$^{-1}$
1067: Grid (CD-ROM 13)
1068:
1069: \bibitem[\protect\citeauthoryear{Kurucz et al.}{1984}]{kpno}
1070: Kurucz R.L., Furenlid I., Brault J., Testerman L., 1984, Solar Flux
1071: Atlas from 296 nm to 1300 nm (NSO Atlas No 1)
1072:
1073: \bibitem[\protect\citeauthoryear{Landstreet}{1998}]{land98}
1074: Landstreet J. D., 1998, A\&A, 338, 1041
1075:
1076: \bibitem[\protect\citeauthoryear{Landstreet, Dolez \& Vauclair}{Landstreet
1077: et al.}{1998}]{landolvau}
1078: Landstreet J. D., Dolez N., Vauclair S., 1998, A\&A, 333, 977
1079:
1080: \bibitem[\protect\citeauthoryear{Lanz}{1995}]{l95}
1081: Lanz T., 1995, In: Adelman S.J., Wiese W.L., eds,
1082: Astrophysical Applications of Powerful New Databases, PASPC No. 78, p. 423
1083:
1084: \bibitem[\protect\citeauthoryear{Michaud}{1970}]{m70}
1085: Michaud G., 1970, ApJ, 160, 641
1086:
1087: \bibitem[\protect\citeauthoryear{Michaud}{1986}]{m86}
1088: Michaud G., 1986, In Cowley C.R., Dworetsky M.M., M\'{e}gessier C.,
1089: eds, Proc. IAU Coll. 90, Upper Main Sequence Stars with Anomalous
1090: Abundances, Reidel, Dordrecht, p. 459
1091:
1092: \bibitem[\protect\citeauthoryear{Mihalas}{1978}]{mihalas78}
1093: Mihalas D., 1978, Stellar Atmospheres, 2nd Ed., W.H. Freeman \& Co.,
1094: San Francisco
1095:
1096: \bibitem[\protect\citeauthoryear{Misch}{1997}]{mischb}
1097: Misch A., 1997, User's Guide to the Hamil\-ton Echelle
1098: Spec\-tro\-me\-ter\\
1099: URL=http://www.ucolick.org/\verb+~+tony/instruments\\/hamspec/hamspec\_index.html
1100:
1101: \bibitem[\protect\citeauthoryear{Moore}{1949}]{moore}
1102: Moore C.E., 1949, Atomic Energy Levels, Vol. I, NBS Circular No. 467,
1103: Washington, D.C.
1104:
1105: \bibitem[\protect\citeauthoryear{Petrie}{1950}]{petrie} Petrie R.M., 1950,
1106: Publs. DAO, 8, 319
1107:
1108: \bibitem[\protect\citeauthoryear{Renson, Gerbaldi \& Catalano}{Renson et
1109: al.}{1991}]{renson}
1110: Renson P., Gerbaldi M., Catalano F. A., 1991, AAS, 89, 429
1111:
1112: \bibitem[\protect\citeauthoryear{Ryabchikova, Zakhorova \&
1113: Adelman}{Ryabchikova et al.}{1996}]{r96} Ryabchikova T.A., Zakhorova L.A.,
1114: Adelman S.A., 1996, MNRAS, 283, 1115
1115:
1116: \bibitem[\protect\citeauthoryear{Ryabchikova et al.}{1998}]{r98}
1117: Ryabchikova T., Kotchoukhov O., Galazutdinov F., Musaev
1118: F., Adelman S.J., 1998, Contrib. Astron. Obs. Skalnat\'{e} Pleso, 27, 258
1119:
1120: \bibitem[\protect\citeauthoryear{Sadakane}{1980}]{sad80}
1121: Sadakane K., 1980, PASP, 93, 587
1122:
1123: \bibitem[\protect\citeauthoryear{Seaton}{1998}]{mjs}
1124: Seaton M.J., 1998, J Phys B, 31, 5315
1125:
1126: \bibitem[\protect\citeauthoryear{Sigut}{1999}]{sigut}
1127: Sigut T. A. A., 1999, ApJ, 519, 303
1128:
1129: \bibitem[\protect\citeauthoryear{Smith}{1992}]{kcs92}
1130: Smith K.C., 1992, PhD thesis, University of London
1131:
1132: \bibitem[\protect\citeauthoryear{Smith}{1993}]{kcs93}
1133: Smith K.C., 1993, A\&A, 276, 393
1134:
1135: \bibitem[\protect\citeauthoryear{Smith}{1994}]{kcs94}
1136: Smith K.C., 1994, A\&A, 291, 521
1137:
1138: \bibitem[\protect\citeauthoryear{Smith}{1996}]{kcs96}
1139: Smith K.C., 1996, A\&A, 305, 902
1140:
1141: \bibitem[\protect\citeauthoryear{Smith}{1997}]{kcs97}
1142: Smith K.C., 1997, A\&A, 319, 928
1143:
1144: \bibitem[\protect\citeauthoryear{Smith \& Dworetsky}{1988}]{sd88}
1145: Smith K.C., Dworetsky M.M., 1988, In: Adelman S.J., Lanz T., eds,
1146: Elemental Abundance Analyses. Institut d'Astronomie de l'Univ. de
1147: Lausanne, Switzerland, p. 32
1148:
1149: \bibitem[\protect\citeauthoryear{Smith \& Dworetsky}{1990}]{sd90}
1150: Smith K. C., Dworetsky M. M., 1990, in Rolfe, E. J., Evolution in
1151: Astrophysics, ESA SP-310, p. 279
1152:
1153: \bibitem[\protect\citeauthoryear{Smith \& Dworetsky}{1993}]{sd93}
1154: Smith K.C., Dworetsky M.M., 1993, A\&A, 274, 335
1155:
1156: \bibitem[\protect\citeauthoryear{Valdes}{1990}]{valdes}
1157: Valdes F., 1990, The IRAF APEXTRACT Package,
1158: \newline URL=ftp://iraf.tuc.noao.edu/iraf/docs/apex.ps
1159:
1160: \bibitem[\protect\citeauthoryear{Varosi et al.}{1995}]{varosi}
1161: Varosi F., Lanz T., Dekoter A., Hubeny I., Heap S., 1995, MODION (NASA
1162: Goddard SFC)
1163: \newline URL=ftp://idlastro.gsfc.nasa.gov/pub/contrib\\/varosi/modion/
1164:
1165: \bibitem[\protect\citeauthoryear{Vauclair \& Vauclair}{1982}]{vau82}
1166: Vauclair S., Vauclair G., 1982, ARA\&A, 20, 37
1167:
1168: \bibitem[\protect\citeauthoryear{Vogt}{1987}]{vogt}
1169: Vogt S., 1987, PASP, 99, 1214
1170:
1171: \bibitem[\protect\citeauthoryear{Wahlgren, Adelman \& Robinson}{Wahlgren
1172: et al.}{1994}]{wahl94}
1173: Wahlgren G.M., Adelman S.J., Robinson R.D., 1994, Ap.J., 434, 349
1174:
1175: \bibitem[\protect\citeauthoryear{Warner}{1967}]{W67}
1176: Warner B., 1967, MNRAS, 136, 381
1177:
1178: \bibitem[\protect\citeauthoryear{Wiese, Smith \& Glennon}{Wiese et
1179: al.}{1966}]{wsg66}
1180: Wiese W.L., Smith M.W., Glennon B.M., 1966, Atomic Transition
1181: Probabilities, Vol. I, NSRDS-NBS 4
1182:
1183: \end{thebibliography}
1184:
1185:
1186: \end{document}
1187:
1188: