1: %\documentstyle[preprint2]{opt/local/aastex/aastex-5.0.2/aastex}
2: %\usepackage{/opt/local/aastex/aastex-5.0.2/aastexug}
3: %\documentstyle[12pt,aasms4]{article}
4: %\documentstyle[11pt,aaspp4]{article}
5: %\documentclass{emulateapj}
6: \documentclass[12pt,preprint]{aastex}
7:
8: \lefthead{}
9: \righthead{Neon Abundances in B-Stars of the Orion Association: Solving the Solar
10: Model Problem?}
11:
12: \begin{document}
13:
14: \title{Neon Abundances in B-Stars of the Orion Association: Solving the Solar
15: model Problem?}
16:
17: \author{Katia Cunha}
18: \affil{National Optical Astronomy Observatory\footnote{On leave from
19: Observat\'orio Nacional - MCT; Rio de Janeiro, Brazil},
20: P.O. Box 26732, Tucson, AZ 85726 USA; kcunha@noao.edu}
21:
22: \author{Ivan Hubeny}
23: \affil{Steward Observatory, University of Arizona, Tucson, AZ 85712, USA;
24: hubeny@aegis.as.arizona.edu}
25:
26: \author{Thierry Lanz}
27: \affil{Department of Astronomy, University of Maryland, College Park, MD 20742 USA;
28: tlanz@umd.edu}
29:
30: \begin{abstract}
31: We report on non-LTE Ne abundances for a sample of B-type stellar members of the Orion
32: Association. The abundances were derived by means of non-LTE fully
33: metal-blanketed model atmospheres and extensive model atoms with updated
34: atomic data. We find that these young stars have a very homogeneous
35: abundance of A(Ne) = 8.27 $\pm$ 0.05. This abundance is higher by $\sim$ 0.4 dex
36: than currently adopted solar value, A(Ne)=7.84, which is derived from
37: lines produced in the corona and active regions. The general agreement
38: between the abundances of C, N, and O derived for B stars with the solar
39: abundances of these elements derived from 3-D hydrodynamical models
40: atmospheres strongly suggests that the abundance patterns of the light
41: elements in the Sun and B stars are broadly similar. If this hypothesis is
42: true, then the Ne abundance derived here is the same within the uncertainties as
43: the value required to reconcile solar models with helioseismological
44: observations.
45: \end{abstract}
46:
47: \keywords{stars: early-type, abundances}
48:
49: \section{INTRODUCTION}
50: One important recent result from studies of stellar atmospheres and
51: chemical compositions in stars is the downward revision in the abundances of
52: carbon, nitrogen and oxygen in the solar photosphere, which was obtained from
53: the adoption of time-dependant, hydrodynamical and 3-Dimensional
54: model atmosphere calculations (Asplund 2005). These more realistic model atmosphere
55: calculations indicate that the solar abundances should be lower by
56: roughly 0.2-0.3 dex when compared to abundances derived from hydrostatic 1-D calculations.
57: It soon became apparent that the significantly lower abundances in the Sun resulted in severe
58: inconsistencies between the solar models and measurements
59: from helioseismology.
60:
61: Different possibilities were investigated in order
62: to try to reconcile the solar models with the seismological observations, such as
63: updating opacities (Bahcall et al. 2004), changing diffusion rates (Guzik et al. 2005), or
64: significantly changing
65: the adopted solar abundances of key elements, such as neon. Allowing for
66: a larger neon abundance, in particular, was justified because the neon abundance in the Sun
67: can be considered to be more uncertain given that it is not measured from lines formed in
68: the solar photosphere.
69: Antia \& Basu (2005) constructed envelope models of the Sun, allowed for different
70: abundance mixtures and focused on the density profile, which is determined from helioseismology.
71: More complete calculations were presented in
72: Bahcall, Basu \& Serenelli (2005) who constructed solar models,
73: consisting of the atmospheres plus the interior, and concluded that an adopted
74: neon abundance A(Ne)= 8.29 $\pm$ 0.05, would suffice in order to bring the solar models
75: and seismological observations into an acceptable agreement.
76: Independently, measurements of neon abundances in
77: a sample of chromospherically active cool stars
78: by Drake \& Testa (2005) indirectly supported high Ne abundances
79: in solar type stars.
80:
81: It has been a long standing puzzle that the C, N and O abundances obtained for B-stars,
82: which are young, were typically lower than the at-the-time generally accepted solar abundances
83: from Anders \& Grevesse (1989). These were puzzling results because from our simplest
84: understanding of how the Galaxy chemically evolves, it is not expected that
85: young stars in the solar vicinity would be less enriched than the Sun, which is much older.
86: Moreover, the abundances obtained from Galactic H II regions were also lower than
87: the accepted solar and in rough agreement with the B-star results.
88: These inconsistencies between the abundances of young stars and H II regions, on one
89: hand, and the Sun on the other, are reconciled nowadays with the revised
90: solar abundances from 3-D models.
91: In this context, it is therefore important and timely to derive accurate
92: neon abundances in the atmospheres of early-type stars.
93: In this study, we report on non-LTE (NLTE) neon abundance calculations for
94: a sample of B stars members of the Orion association. The Ne abundances
95: in young stars
96: can independently shed light on the issue related to the reference
97: Ne abundance in the Galaxy.
98:
99: \section{OBSERVATIONS}
100:
101: The target stars are OB main-sequence members of the different stellar subgroups of
102: the Orion association and drawn from the sample analyzed by
103: Cunha \& Lambert (1992, 1994). Eleven stars were observed with the
104: 2.1m telescope at the McDonald Observatory at high resolution (R=55,000)
105: using the Sandiford echelle. The spectra were obtained on Oct 27, 1994
106: and these have 26 echelle orders covering the total spectral range between 5390 and 6680 \AA.
107: The spectra were reduced with IRAF data package following standard procedures.
108:
109: \section{Non-LTE Abundance Calculations}
110:
111: The stellar parameters for the sample stars were derived in Cunha \& Lambert
112: (1992).
113: Non-LTE model atmospheres were computed using the TLUSTY code (Hubeny 1988,
114: Hubeny \& Lanz 1995).
115: The model calculations assumed a constant microturbulence of 2 km/s.
116: Preliminary models for this study were taken from an extensive grid of
117: NLTE line-blanketed model atmospheres of B stars (Lanz \& Hubeny, in prep.).
118: The final models were computed for the actual effective temperatures
119: and surface gravities of our program stars, and adopting an extended Ne model
120: atom. The BSTAR model grid is analogous to our OSTAR2002 grid (Lanz \& Hubeny
121: 2003), the only difference being the addition of lower ionization stages of the
122: most important species. Concretely, the following ions were considered explicitly
123: in the BSTAR grid models: H I--II, He I--III, C I--V, N I--VI, O I--VI, Ne I--V,
124: Mg II--III, Al II--IV, Si II--V, S II--VI, and Fe II--VI.
125:
126: The Ne model atom constructed consists of 79 levels of Ne I, 138 levels of Ne II,
127: 38 levels of Ne III, 12 levels of Ne IV, plus ground state of Ne V.
128: The energies of the levels were taken from the Opacity Project database
129: TOPBASE (Cunto et al. 1993), updated by the more accurate experimental
130: level energies from the Atomic and Spectroscopic Database at NIST
131: (Martin et al. 1999) whenever available.
132: The $gf$-values were taken from the same sources.
133: However, since the $LS$-coupling, on which the Opacity Project
134: calculations are based, is rather inaccurate for Ne I, we have used results and
135: the procedure suggested by Seaton (1998) to transform the level energies (and
136: designations) from $LS$-coupling to the more appropriate $jK$-coupling.
137: Also, we used a model atom that treats explicitly the fine structure of multiplets.
138: The photoionization cross-sections were taken from TOPBASE, and the collisional
139: excitation rates were considered using the Van Regemorter formula, and for
140: collisional ionization using the Seaton formula -- for a synopsis of expressions,
141: see Hubeny (1988).
142:
143: The detailed synthetic spectra were computed using the {\it Interactive Data
144: Language} (IDL) interface SYNPLOT (Hubeny, unpublished) to the spectrum
145: synthesis program SYNSPEC (Hubeny, Lanz, \& Jeffery 1995).
146: The abundances were obtained from the best fits
147: between observed and synthetic spectra of 8 Ne I transitions:
148: $\lambda$6506.5, 6402.2, 6383.0, 6334.5, 6266.5, 6163.5, 6143.1 and 6096.2 \AA.
149: We have computed NLTE line profiles (using NLTE atomic level populations
150: of all atoms and ions computed by TLUSTY). When computing detailed
151: synthetic spectra we found that the best fits to the observations were
152: obtained for a microturbulent
153: velocity around 5 km/s, although the abundance results were quite insensitive to
154: the microturbulence parameter.
155:
156: The nature of NLTE effects in Ne I line formation was already discussed
157: by Auer \& Mihalas (1973), Dworetsky \& Budaj (2000), and, in particular,
158: by Sigut (1999). Our models do not offer anything fundamentally different
159: from these studies; only we are using significantly more extended model
160: atoms, more recent atomic data, and fully blanketed model atmospheres.
161: The nature of NLTE effects for above listed optical lines of Ne I was
162: already explained in the earlier studies, namely since the lower levels
163: of the optical lines are connected to the ground state of Ne I by resonance
164: lines that are located in far UV, they are essentially in detailed balance
165: with the ground state. The optical lines thus behave like classical lines
166: in a two-level atom, that is the lower level is somewhat overpopulated, while
167: the upper level is depopulated (because of an imbalance of the number of
168: excitations compared to de-excitations caused by the photon escape through
169: the boundary). The source function is thus lower than the Planck function,
170: and the optical lines are consequently predicted stronger than in LTE.
171: This in turn means that the deduced NLTE abundances are expected to be
172: lower that the LTE ones.
173:
174: Our results are in qualitative agreement with the earlier studies.
175: We compared our calculations with the two previous studies (Sigut 1999 and
176: Dworetsky \& Budaj 2000). For one target star, HD35299,
177: we measured equivalent widths for the Ne I lines and computed Ne
178: abundances for an assumed microturbulence $\xi$=5.0 km/s. A modest systematic difference
179: was obtained for HD35299: $\delta$Ne(This study - Sigut) = 0.1 dex, for an average
180: of all lines (eqws computed by Sigut, priv. comm.)
181: Dworetsky \& Budaj (2000), as discussed in their paper, find good agreement
182: with the results by Sigut (1999). Dworetsky \& Budaj have also used TLUSTY for
183: their study, and made their TLUSTY-compatible Ne I atomic data input
184: available online. We tested their input data, and found that they inadvertently
185: set up parameters for the collisional excitation rates in such a way that they
186: were effectively set to 0 (which means they they in fact maximized possible
187: NLTE effects). When we artificially modified our data by setting
188: collisional excitation rates also to zero, we obtained an excellent agreement
189: with Dworetsky \& Budaj. This also demonstrates that the collisional excitation
190: rates are not critical for basic features of Ne I NLTE line formation.
191:
192:
193:
194: \subsection{Ne Abundance Results and Previously Derived O Abundances}
195:
196: The studied Orion stars span a significant range in effective temperature,
197: from $\sim$ 20000 K to 29000 K (Table~1). In Figure 1 (top panel) we show that
198: there is no significant trend of the Ne abundances with the stars' effective temperature,
199: indicating the absence of important systematic errors in this study.
200: A comparison between the LTE and
201: NLTE abundance trend with $T_{\rm eff}$ is shown in the bottom panel of Figure 1.
202: There is a trend of the derived LTE abundances with $T_{\rm eff}$,
203: which again demonstrates the inadequacy of LTE for abundance determinations,
204: as already shown by earlier investigations.
205: The LTE line profiles were
206: obtained by setting the Ne level populations to their LTE values, while
207: the atmospheric structure (temperature, electron density, etc.), as well
208: as level population of other species, were kept at their NLTE values.
209: We stress that such line profiles are different from truly LTE line
210: profiles (computed for a consistent LTE structure and LTE level populations
211: of all species), but our ``LTE'' line profiles best demonstrate NLTE
212: effects in determining the Ne abundance.
213: The fact that LTE
214: abundances are {\em larger} than NLTE ones is in fact quite comforting in the
215: context of Ne abundance determinations, because otherwise one may be worried
216: whether the deduced high Ne abundance is an artifact of some spurious
217: NLTE effect caused for instance by some inadequacy in atomic data. However,
218: since the NLTE abundance is {\em smaller}, this potential worry can be
219: ruled out.
220:
221: The Ne abundances obtained for the targets (listed in Table 1)
222: are quite homogeneous
223: and show a small scatter that can be completely explained in terms of
224: uncertainties in the abundance determinations.
225: The average Ne abundance for all target stars is A(Ne)=8.27 $\pm$ 0.05.
226: The uncertainties in the Ne I abundances can be estimated from
227: the sum in quadrature of the abundance uncertainties due to errors in the
228: adopted stellar parameters, microturbulence, continuum location, as well as,
229: atomic data. We estimate that our derived Ne abundances are accurate to within
230: roughly 0.1 dex. The data available to us do not allow for Ne II abundances
231: to be derived in this study as a consistency check. (We note that Kilian-Montenbruck
232: et al. (1994) found high LTE Ne II abundances for OB stars.)
233:
234: It is the Ne/O ratio, and not only neon, that is obtained from abundance measurements in the
235: solar corona
236: and this ratio constitutes an important ingredient in the construction of solar models. Before
237: a comparison can be done with the Orion results for B stars, it is important to stress
238: that the oxygen and neon abundances for the Orion targets were not derived homogeneously.
239: The methodology presented in this study to
240: derive neon abundances consisted of a full NLTE treatment, including
241: NLTE line formation and the computation of fully-blanketed model atmospheres in NLTE.
242: For oxygen, however, we adopt the previously published results from Cunha \& Lambert (1994)
243: which were derived in LTE and using Kurucz LTE model atmospheres, and finally corrected by means
244: of the NLTE calculations by Becker \& Butler (1988).
245: It is important then to verify
246: whether the previously derived oxygen abundances for Orion are consistent with
247: results from the more sophisticated calculations presented here for neon.
248:
249: As a consistency check on the published oxygen results, we re-derived the oxygen
250: abundance for one of the Orion stars, HD35299, using TLUSTY/SYNSPEC.
251: We adopted the same published
252: equivalent widths for O II lines and same stellar parameters but used the
253: TLUSTY NLTE model atmospheres calculated for the Ne analysis. We calculated
254: O II abundances versus the microturbulence parameter and derived $\xi$= 5 km/s
255: and an oxygen abundance A(O)= 8.65 $\pm$ 0.05. This abundance compares favorably
256: with the oxygen abundance obtained in Cunha \& Lambert (1992), within the uncertainties.
257: Such agreement justifies the adoption of the published oxygen results in order
258: to investigate Ne/O ratios.
259:
260: \section{DISCUSSION}
261:
262: Solar photospheric abundances can be readily compared to
263: meteoritic C1 chondrite abundances and good agreement is found for most of the elements,
264: or more specifically, for those elements that form rocks (see e.g. Lodders 2003).
265: Noble gases, as well as carbon, nitrogen and oxygen are volatiles and
266: their abundances are therefore significantly depleted in meteorites. For C,N, and O
267: one can rely on abundances measured in the solar photospheres, available
268: from both 1-D and 3-D model atmospheres calculations.
269: For neon, however, the solar abundances are subject to further uncertainties due to
270: the absence of photospheric lines because even the lower excited states
271: of the Ne atom have very high energy. Alternatively, Ne abundances in the Sun are inferred
272: from measurements of Ne/O in the solar coronal gas, solar wind and solar energetic particles.
273: The most recent assessment of the Ne abundance in the Sun is obtained from
274: measurements of Ne/O in the solar corona and from energetic particles is A(Ne)=7.84
275: (Asplund, Grevesse \& Noels 2005). The solar value according to Lodders (2003)
276: is just slightly higher (A(Ne)=7.87).
277:
278: The neon abundances derived here for a sample of early-type stars in the Orion
279: association are found to be quite homogeneous.
280: The average neon abundance for the studied stars (A(Ne)=8.27 $\pm$ 0.05) is higher than the
281: quoted solar value by $\sim$ 0.4 dex.
282: In Figure 2 we show our Ne results versus oxygen abundances.
283: The average oxygen abundance for the sample Orion stars is: A(O)=8.70 $\pm$ 0.09.
284: which is entirely consistent with a single oxygen abundance and agrees with
285: the solar abundance of A(O)=8.66 (Asplund et al. 2005).
286:
287: Recent results from detailed calculations of oxygen abundances in a sample of 3 B0.5V
288: stars in the Orion nebula indicate an average oxygen abundance of
289: A(O)=8.63 $\pm$ 0.03 (Simon-Diaz et al. 2006).
290: The oxygen abundance data presented here is a subsample of the stars
291: analyzed in Cunha \& Lambert (1994). For the full sample with 18 stars this previous study
292: obtained an oxygen
293: abundance spread which was larger: A(O)= 8.72 $\pm$ 0.13, but marginally within the
294: abundance uncertainties.
295: Our target stars in this study, however, span a tighter
296: range in oxygen abundances (A(O)= 8.70 $\pm$ 0.09) and are considered to represent
297: a single oxygen abundance for all purposes.
298: Therefore, the B stars in the Orion association can be
299: represented by a Ne/O ratio of 0.38, which is much higher than the currently
300: adopted solar value of 0.15 (Asplund et al. 2005).
301:
302: The Orion Nebula has been the most extensively studied galactic H II region
303: and is recognized as the standard reference for nebular abundance studies in the Galaxy.
304: Recently, Esteban et al. (2004) conducted a careful emission line study of several elements
305: in the Orion nebula and obtained
306: A(O)=8.65 and A(Ne)=8.05 (or, Ne/O=0.25); these are gas
307: abundances and corrections for any element trapped in grains have not been considered.
308: This nebular result is lower than our average for the Orion stars but higher
309: than the solar value (see Figure 2).
310:
311: \section{Conclusions}
312:
313: Measurements of neon abundances in a variety of objects that can help define the
314: uncertain neon abundance in the Sun are potentially of great importance for solar physics.
315: We find the Ne abundance in B stars members of the Orion association is
316: significantly higher than the solar value by roughly 0.4 dex ($\sim$ 2.5X).
317: We argue that the Ne abundances measured in young OB stars
318: should be a good representation of the solar chemical composition, as is indicated
319: from the good agreement between the abundances in B stars and Sun for other elements
320: such as C, N, and O. The high Ne abundances obtained here
321: come to the rescue of the solar models that require, according to Bahcall, Basu \& Serenelli (2005),
322: an increase in the Ne abundance by $\sim$2.8 $\pm$ 0.4.
323:
324: \acknowledgements
325: We especially thank C. Allende Prieto for discussions and for communicating with T. Sigut.
326: We also thank D. Arnett, L. Stanghellini, S. Wolff and D. Garnett for several discussions.
327: We thank V. V. Smith for helping with the observations.
328: The work reported here is supported in part by the National Science
329: Foundation through AST03-07534, AST03-07532, and NASA through NAG5-9213.
330:
331: \begin{thebibliography}{}
332:
333: \bibitem[]{336} Anders, E. \& Grevesse, N. 1989, Geochimica et Cosmochimica Acta, vol. 53, p. 197
334:
335: \bibitem[]{338} Antia, M., \& Basu, S. 2005, ApJ, 621, L85
336:
337: \bibitem[]{340} Asplund, M. 2005, Ann. Rev. Astron. Astrophys., 43, 481
338:
339: \bibitem[]{342} Asplund, M., Grevesse, N., \& Sauval, A. J.2005 In Cosmic
340: Abundances as Records of Stellar Evolution and Nucleosynthesis, ed. F. N. Bash, \&
341: T. G. Barnes p. 25
342:
343: \bibitem[]{346} Auer, L. H., \& Mihalas, D. 1973, ApJ, 184, 151
344:
345: \bibitem[]{348} Bahcall, J. N., Basu, S., \& Serenelli, A. M. 2005, ApJ, 631, 1281
346:
347: \bibitem[]{350} Bahcall, J. N., Serenelli, A. M., Basu, S. 2005b, ApJ, 621, L85
348:
349: \bibitem[]{352} Becker, S. R. \& Butler, K. 1988, A\&A, 201, 232
350:
351: \bibitem[]{354} Cunha, K. \& Lambert, D.L. 1992, ApJ, 399, 586
352:
353: \bibitem[]{356} Cunha, K. \& Lambert, D.L. 1994, ApJ, 426, 170
354:
355: \bibitem[]{358} Cunto, W., Mendoza, C., Ochsenbein, F., \& Zeipen, C. J. 1993,
356: A\&A, 275, L5
357:
358: \bibitem[]{361} Drake, J. J., \& Testa, P. 2005, Nature, 43, 525
359:
360: \bibitem[]{363} Dworetsky, M. M., \& Budaj, J. 2000, MNRAS, 318, 1264
361:
362: \bibitem[]{365} Guzik, J. A., Watson, L. W., \& Cox, A. N. 2005, ApJ, 627, 1049
363:
364: \bibitem[]{367} Grevesse, N., \& Sauval, A. J. 1998, Space Sci. Rev. 107, 665
365:
366: \bibitem[]{369} Hubeny, I. 1988, Comput. Phys. Commun., 52, 103
367:
368: \bibitem[]{371} Hubeny, I., \& Lanz, T. 1995, ApJ, 439, 875
369:
370: \bibitem[]{373} Hubeny, I., Lanz, T., \& Jeffery, C. S. 1994, SYNSPEC - A User's
371: Guide, in Newsletter on Analysis of Astronomical Spectra No 20, St. Andrews Univ.
372:
373: \bibitem[]{375} Kilian-Montenbruck, J., Gehren, T., \& Nissen, P. E. 1994, A\&A, 291, 757
374:
375: \bibitem[]{376} Lanz, T., \& Hubeny, I. 2003, ApJS, 146, 417
376:
377: \bibitem[]{378} Lester, J. B, Gray, R. O., \& Kurucz, R. L. 1986, ApJS, 61, 509
378:
379: \bibitem[]{380} Martin, W. C., Sugar, J., Musgrove, A., Wiese, W. L., \& Fuhr, J. R.
380: 1999, NIST Atomic Spectra Database
381:
382: \bibitem[]{381} Nieva, M.F., \& Przybilla, N. 2006, ApJ, 639, L39
383:
384: \bibitem[]{383} Seaton, M. J. 1998, J. Phys. B, 31, 5315
385:
386: \bibitem[]{385} Sigut, T. A. A. 1999, ApJ, 519, 313
387:
388: \bibitem[]{387} Simon-Diaz, S., Herrero, A., Esteban, C., \& Najarro, F. 2006, A\&A, 351, 366
389:
390: \end{thebibliography}
391:
392: \clearpage
393:
394: \begin{deluxetable}{lcccc}
395: \tablecaption{Sample Stars and Abundances}
396: \tablewidth{0pt}
397: \tablehead{
398: \colhead{Star} & \colhead{Teff} & \colhead{Log g} & \colhead{A(Ne)} & \colhead{A(O)}}
399: \startdata
400: HD35039 & 20550 & 3.74 & 8.25 & 8.60 \\
401: HD35299 & 24000 & 4.25 & 8.30 & 8.57 \\
402: HD35912 & 19590 & 4.20 & 8.21 & 8.70 \\
403: HD36285 & 21930 & 4.40 & 8.29 & 8.80 \\
404: HD36351 & 21950 & 4.16 & 8.24 & 8.76 \\
405: HD37356 & 22370 & 4.13 & 8.33 & 8.67 \\
406: HD37209 & 24050 & 4.13 & 8.28 & 8.83 \\
407: HD37744 & 24480 & 4.40 & 8.35 & 8.63 \\
408: HD36959 & 24890 & 4.41 & 8.21 & 8.76 \\
409: HD36591$^{a}$ & 26330 & 4.21 & 8.26 & 8.60 \\
410: HD36960 & 28920 & 4.33 & 8.21 & 8.72
411:
412: \enddata
413: \tablecomments{(a): Recent results from IUE flux,
414: 2MASS and Johnson magnitudes (Nieva \& Przybilla 2006)
415: indicate quite good agreement with the adopted
416: stellar parameters.
417: }
418: \end{deluxetable}
419:
420: \clearpage
421:
422: \begin{figure}
423: \epsscale{.8}\plotone{f1.eps}
424: \caption{\label{fig1} Top panel:
425: The Ne derived abundances versus the adopted
426: effective temperatures for the sample Orion stars. The $T_{\rm eff}$ range covered
427: here is relatively large and no trends are found for the derived neon abundances.
428: For comparison, we also indicate the currently adopted solar Ne abundance
429: from Asplund et al. (2005) as the dashed line.
430: Bottom panel:
431: LTE (open circles) and NLTE (filled circles) abundances calculated for
432: the strongest Ne I line $\lambda$ 6402 \AA. This is the only line that
433: is strong enough to be measured in the hottest stars in our sample.
434: The LTE Ne abundances are
435: found to have a trend with effective temperature. This trend
436: is erased with the NLTE calculations.}
437: \end{figure}
438:
439: \begin{figure}
440: \includegraphics[scale=0.3,angle=270]{f2.eps}
441: \caption{\label{fig2} Neon abundances derived for the Orion B stars in this study versus the
442: oxygen abundances from Cunha \& Lambert (1994; filled circles). The average Ne and O
443: abundance for the studied sample is represented by the open square. For comparison
444: we also show the currently adopted solar value (Asplund et al. 2005),
445: as well as the H II region abundance obtained for the
446: Orion nebula by Esteban et al. (2004; filled square), without accounting for any depletion
447: onto grains.}
448:
449: \end{figure}
450:
451: \end{document}
452: