1: %%**********************************************************************************
2: %% Martins et al., 2007: On the nature of S2 in the Galactic Center
3: %% Analysis made at MPE, Garching.
4: %%**********************************************************************************
5:
6: %\documentclass[manuscript]{aastex}
7: %\documentclass[preprint]{aastex}
8: %\documentclass[preprint2]{aastex}
9: \documentclass{emulateapj}
10: %\documentclass[12pt,preprint]{aastex}
11:
12: \usepackage{natbib}
13: \usepackage[figuresright]{rotating}
14: \usepackage{graphicx}
15: \usepackage{subfigure}
16: \bibliographystyle{apj}
17:
18: %-----------------------------------------------------------------------
19:
20: \newcommand{\hei}{He~{\sc i} 2.112 $\mu m$}
21: \newcommand{\heii}{He~{\sc ii} 2.189 $\mu m$}
22: \newcommand{\brg}{Br$\gamma$}
23: \newcommand{\mum}{\ifmmode \mu m \else $\mu m$\fi}
24:
25: \newcommand{\teff}{\ifmmode T_{\rm eff} \else $T_{\mathrm{eff}}$\fi}
26: \newcommand{\logg}{\ifmmode \log g \else $\log g$\fi}
27: \newcommand{\lL}{\ifmmode \log \frac{L}{L_{\odot}} \else $\log \frac{L}{L_{\odot}}$\fi}
28: \newcommand{\mdot}{$\dot{M}$}
29: \newcommand{\myr}{M$_{\odot}$ yr$^{-1}$}
30: \newcommand{\vsini}{$V$ sin$i$}
31: \newcommand{\vinf}{\ifmmode v_{\infty} else $v_{\infty}$\fi}
32: \newcommand{\vturb}{v$_{turb}$}
33: \newcommand{\vesc}{v$_{esc}$}
34: \newcommand{\kms}{km s$^{-1}$}
35: \newcommand{\msun}{\ifmmode M_{\odot} \else M$_{\odot}$\fi}
36: \newcommand{\zsun}{\ifmmode Z_{\odot} \else Z$_{\odot}$\fi}
37: \newcommand{\lsun}{\ifmmode L_{\odot} \else L$_{\odot}$\fi}
38: \newcommand{\rsun}{\ifmmode R_{\odot} \else R$_{\odot}$\fi}
39:
40: %-----------------------------------------------------------------------
41:
42: %% You can insert a short comment on the title page using the command below.
43:
44: \slugcomment{}
45:
46:
47: \shorttitle{The nature of S2}
48: \shortauthors{Martins et al.}
49:
50:
51: \begin{document}
52:
53:
54: \title{On the nature of the fast moving star S2 in the Galactic Center \footnote{Based on observations collected at the ESO Very Large Telescope (programs 075.B-0547, 076.B-0259, 077.B-0503, 078.B-0520 and 179.B-0261)}}
55:
56:
57: \author{F. Martins \altaffilmark{1}, S. Gillessen \altaffilmark{1}, F. Eisenhauer \altaffilmark{1}, R. Genzel \altaffilmark{1,2}, T. Ott \altaffilmark{1}, S. Trippe \altaffilmark{1}}
58:
59: \email{martins@mpe.mpg.de}
60:
61: %% Notice that each of these authors has alternate affiliations, which
62: %% are identified by the \altaffilmark after each name. Specify alternate
63: %% affiliation information with \altaffiltext, with one command per each
64: %% affiliation.
65:
66: \altaffiltext{1}{Max Planck Institute of Extraterrestrial Physics, Postfach 1312, D-85741, Garching, Germany}
67: \altaffiltext{2}{Department of Physics, University of California, CA 94720, Berkeley, USA}
68:
69:
70: \begin{abstract}
71: We analyze the properties of the star S2 orbiting the supermassive
72: black hole at the center of the Galaxy. A high quality SINFONI H and K band
73: spectrum obtained from coadding 23.5 hours of observation
74: between 2004 and 2007 reveals that S2 is an early B dwarf
75: (B0--2.5V). Using model atmospheres, we constrain its stellar and wind
76: properties. We show that S2 is a genuine massive star, and not the
77: core of a stripped giant star as sometimes speculated to resolve the
78: problem of star formation so close to the supermassive black hole. We
79: give an upper limit on its mass loss rate, and show that it is He
80: enriched, possibly because of the presence of a magnetic field.
81: \end{abstract}
82:
83: \keywords{Stars: early-type --- stars: fundamental parameters --- Galaxy: center}
84:
85:
86: %%#####################################################################
87: %%------------------------------- Introduction ---------------------
88:
89:
90: \section{Introduction}
91: \label{intro}
92:
93: The central parsec of our Galaxy hosts a large population of young
94: massive stars \citep{krabbe95,pgm06}. Their presence is puzzling since
95: according to standard theories, star formation should not happen so
96: close to the supermassive black hole SgrA*: the tidal forces are so
97: large that any molecular cloud should be disrupted before being able
98: to collapse \citep{morris93}. For most of the young stars, the
99: solution seems to be star formation in dense accretion disks as
100: witnessed by the two counter-rotating flat stellar structures around
101: SgrA* \citep{lb03,genzel03,pgm06}. However, this attractive solution
102: does not explain the so-called ``S stars'', the group of objects
103: located within one arcsecond (=0.038 pc) of the black hole. Their
104: orbits are randomly oriented \citep{frank05} and thus do not fit the
105: accretion disk scenario. Spectroscopically identified as B stars
106: \citep{ghez03,frank05}, their true nature as massive stars has been
107: challenged: they have been proposed to be the hot, luminous cores of
108: evolved red giant and/or AGB stars the envelope of which has been
109: stripped by tidal interactions with SgrA* \citep{tal05,kd05}. In that
110: case, the problem of their formation process vanishes since they might
111: very well have formed away from the hostile environment of the black
112: hole before having migrated towards the Galactic Center. Such a
113: process is too slow to explain the presence of short lived massive
114: stars close to SgrA*. Hence, these S stars constitute a ``paradox of
115: youth'' \citep{ghez03}.
116:
117: In this letter, we analyze the properties of the brightest S-star --
118: S2 -- and show that it is a genuine massive star.
119:
120:
121:
122:
123: %%#####################################################################
124: %%----------------------- Observation / models ---------------------
125:
126: \section{Observational data and models}
127: \label{obs_mod}
128:
129: To constrain the physical parameters of S2, a high quality spectrum is
130: needed. Since the observation of the S stars is time consuming, the
131: S/N ratio obtained in one night is usually limited to 10 at
132: maximum. To increase this ratio, we have co-added all the S2 spectra
133: obtained with SINFONI \citep{sinf} in adaptive optics mode (0.0125\arcsec
134: $\times$0.025\arcsec) on the VLT since 2004. This corresponds to a
135: total integration time of 23.5 hours in K band. Each spectrum was
136: carefully extracted by selection of source and background pixels from
137: which nebular contamination was removed. Note however that this is not
138: such a critical problem for S2 which has a large radial velocity so
139: that the position of the stellar absorption core is blueshifted beyond
140: the nebular emission. We have determined the combined spectrum by a
141: cross-correlation technique. First, the position of the \brg\ line was
142: estimated by a simple Gaussian fit to the line. From the such obtained
143: radial velocities a first combined spectrum was calulated. This
144: spectrum was then used to cross-correlate all individual spectra
145: against it. This yielded better estimates for the radial
146: velocities, which in turn were used to get a more accurate combined
147: spectrum. This procedure was iterated until the result did not change
148: anymore. The resulting spectrum shown in Fig.\ \ref{spec_s2} has a S/N
149: ratio of a few tens in the K band and a resolution of $\sim
150: 4000$. The H band spectrum is noisier due to shorter observation time
151: and higher extinction.
152:
153: The main K band lines (\hei\ and \brg) are clearly
154: identified. But we also detect several weak He~{\sc i} lines at 2.149,
155: 2.161 and 2.184 \mum. Similarly, one He~{\sc i} and three H~{\sc i}
156: lines are identified for the first time in the H band \footnote{These
157: lines were seen by \citet{frank05} in the average spectrum of several
158: stars, but not in individual spectra.}. We do not see any trace of
159: \heii. With the current S/N ratio, this line would be detected if it
160: had a depth of about 1\% of the continuum level.
161:
162: For the quantitative analysis of this spectrum, we used the atmosphere
163: code CMFGEN \citep{hm98}. It allows the computation of non-LTE
164: atmosphere models including winds and line blanketing and is appropriate for hot massive stars. A description of the code is given in \citet{hm98}, and we refer to \citet{gc07} for a summary of
165: the procedure used to build the models. To derive the parameters of
166: interest, we have run models with: 19000 $< T_{\rm eff} <$ 30000 K,
167: 3.0 $< \log g <$ 4.75, 0.1 $<$ He/H $<$ 2.0 and 10$^{-8}$ $< \dot{\rm
168: M} <$ 10$^{-5.5}$ \myr. These models and synthetic spectra
169: included, in addition to H and He, C, N, O, Si, S and Fe. We adopted the solar
170: abundances of \citet{gs98}.
171: \begin{figure}
172: \epsscale{1}
173: \plotone{f1.eps}
174: \caption{Spectrum of S2 after addition of all the SINFONI observations since 2004. The main lines are indicated. \label{spec_s2}}
175: \end{figure}
176:
177:
178:
179:
180: %%#####################################################################
181: %%------------------- A genuine B star --------------------
182:
183: \section{The nature of S2: a genuine B star}
184: \label{param}
185:
186: %----------------------------------------------
187: \subsection{Spectroscopic classification}
188:
189:
190: The spectrum shown in Fig.\ \ref{spec_s2} is typical of an early B
191: star \citep[see also][]{ghez03}. Inspection of the atlases of
192: \citet{wh97} and \citet{hanson05} shows that for late O stars, \heii\
193: is expected. We do not detect this line in spite of the good enough
194: S/N ratio. For stars with spectral types later than B3, the He~{\sc i}
195: lines vanish. Since we detect several of them, we can safely argue
196: that S2 has a spectral type between B0 and B2.5. Fig.\ 12 of
197: \citet{hanson05} shows that in supergiants and giants, \brg\ is
198: clearly separated from He~{\sc i} 2.161 \mum, while for dwarfs \brg\
199: is broad enough to merge with the He~{\sc i} line. This latter
200: morphology is similar to what we observe for S2 (see below for a
201: quantitative comparison). Hence, in addition to being
202: spectroscopically identified as an early B star, we can conclude from
203: these qualitative arguments that \textit{S2 is also a dwarf and not a
204: supergiant/giant}.
205:
206:
207: %----------------------------------------------
208: \subsection{A mass estimate}
209:
210: The key question we want to answer here is the exact nature
211: of S2. Spectroscopically, S2 is unambiguously an early B dwarf.
212: However, this does not necessarilly mean that S2 is a young,
213: massive star. It could be the core of an older, evolved star which
214: would have lost its envelope through tidal interaction with SgrA* but
215: would still spectroscopically look like a B star. In that case, the
216: star could have formed far away from SgrA* before being dragged to its
217: proximity and experiencing an envelope stripping. There would be no
218: paradox of youth.
219:
220:
221: Recently, \citet{kd05} and \citet{dray06} modeled this stripping
222: process as well as the subsequent evolution of the remaining core in
223: the vicinity of SgrA*. They showed that, under certain condition on
224: the IMF and the capture rate, the population of S stars could be
225: explained by tidal stripping. The argument is however based on the
226: number of stars predicted at the position of the S stars in the HR
227: diagram. Expressed differently, the conclusion comes from the fact
228: that the modeled stripped stars can reach the luminosities and
229: effective temperatures of B stars. But strictly speaking, this does
230: not exclude that the S stars are genuine massive stars, which would
231: lie at the same position in the HR diagram.
232:
233: For that, the only parameter which can unambiguously be used is the
234: stellar mass. The most massive stars able to experience envelope
235: stripping are AGB stars. According to stellar evolution, such objects
236: are the evolved descendents of main sequence stars with M $<$ 8
237: \msun. The core of such objects is actually much less massive: Table 4
238: of \citet{fc97} shows that their mass is lower than 1 \msun. In the
239: scenario explored by \citet{dray06}, only stars having lost at least
240: 99\% of their envelope could explain the S stars. Thus, their mass
241: should be at most $\sim$ 1 \msun.
242:
243: To estimate the mass of S2, we can rely on gravity and radius: $M = g
244: R^{2}/G$ where $g$ is the gravity, $R$ the radius and $G$ the
245: gravitational constant. Gravity can be derived from the shape of \brg,
246: while radius is straightforwardly obtained from the knowledge of
247: effective temperature and luminosity. Below, we explain how we
248: proceeded to constrain these parameters.
249:
250:
251:
252: \begin{table}
253: \begin{center}
254: \caption{Physical properties derived from our quantitative analysis for different \teff\ appropriate for B0--2.5V stars.\label{tab_res}}
255: \begin{tabular}{l|rrrrr}
256: \tableline
257: \teff\ [K] & 19000 & 22000 & 25000 & 27000 & 30000 \\
258: \hline & & & & \\
259: \lL\ & 4.30 & 4.45 & 4.60 & 4.65 & 4.80 \\
260: \logg\ & 3.80 & 4.00 & 4.00 & 4.00 & 4.25 \\
261: $\log g_{min}$ & 3.55 & 3.72 & 3.77 & 3.89 & 3.86 \\
262: R [\rsun] & 13.1 & 11.6 & 10.7 & 9.7 & 9.4 \\
263: R$_{min}$ [\rsun] & 11.4 & 10.1 & 9.3 & 8.5 & 7.3 \\
264: M$_{min}$ [\msun] & 16.8 & 19.5 & 18.6 & 20.5 & 14.1 \\
265: He/H [\#] & 1.20 & 0.50 & 0.45 & 0.55 & 0.80 \\
266: He/H$_{min}$ [\#] & 0.85 & 0.25 & 0.25 & 0.30 & 0.30 \\
267: \mdot\ [$10^{-7}$\myr] & $<$3 & $<$3 & $<$3 & $<3$ & $<$3 \\
268: \tableline
269: \end{tabular}
270: \end{center}
271: \end{table}
272:
273:
274:
275:
276: \textit{Effective temperature}. \teff\ is usually derived from the
277: ratio of lines from two consecutive ionization states: He~{\sc i} /
278: He~{\sc ii} lines for O stars and Si~{\sc iii} / Si~{\sc iv} for early
279: B stars. Unfortunately, S2 does not show He~{\sc ii} and Si lines in
280: the K band. Hence, we can only rely on semi-quantitative arguments to
281: estimate \teff. We have shown above that S2 is a B0V to B2.5V
282: star. According to the recent studies of \citet{dufton06} and
283: \citet{trundle07}, such stars have 19000 $<$ \teff\ $<$ 30000 K. We
284: have thus run models for \teff\ = 19, 22, 25, 27 and 30 kK.
285:
286:
287:
288: \textit{Luminosity}. To estimate the star's luminosity, for each \teff\, we adjusted \lL\ in our models to match the absolute K band magnitude of S2. This magnitude is defined as
289: %\begin{equation}
290: $MK=mK-A_{K}-DM$
291: %\end{equation}
292: with $mK=14.0$ the observed magnitude of S2 \citep{pgm06},
293: $A_{K}=2.25$ the K band extinction at the position of S2
294: \citep{schoedel07} and $DM=14.50$ the distance modulus for a distance
295: to the Galactic Center of 7.94 kpc \citep{frank03a}. The resulting
296: absolute K magnitude of S2 is -2.75. The range of values we obtain for
297: \lL\ is 4.30--4.80. In practice, the uncertainty of $\pm$0.2 on
298: $A_{K}$ and 0.5 kpc on the distance to the Galactic Center translate
299: into an uncertainty of about 0.12 in \lL.
300:
301:
302: \textit{Radius}. The radius of the star is simply derived from $L = 4\pi \sigma_{B} R^{2} T_{eff}^{4}$ ($\sigma_{B}$ being the Boltzmann constant). The minimum radius allowed by our study (corresponding also to the minimum mass of the star -- see below) is estimated for a luminosity reduced by the uncertainty (0.12 dex). The values of $R_{min}$ are given in Table \ref{tab_res}.
303:
304:
305: \textit{Gravity}. \logg\ was constrained from the shape
306: of \brg. \citet{repolust05} showed that its wings (resp. absorption
307: core) get broader (resp. weaker) when gravity increases. For large
308: \logg, the blue wing merges with the He~{\sc i} 2.161\mum\ line as
309: discussed previously. Fig.\ \ref{s2_logg_t22} shows how the line
310: profile changes when \logg\ increases from 3.0 to 4.5 for \teff\ =
311: 22000 K. The best models correspond to \logg\ $\sim$ 4.0. A
312: quantification of the goodness of the fit in the \brg\ region of the
313: spectrum was performed by means of a $\chi^{2}$ analysis. To minimize
314: contamination by HeI, we used only the range 2.164-2.179\mum. Fig.\
315: \ref{chi2_logg_t22} shows the result of this analysis and confirms
316: that \logg\ = 4.0 is preferred. From this curve, we also derive a 3
317: $\sigma$ lower limit to \logg\ ($\log g_{min}$). Table \ref{tab_res}
318: shows the results for other \teff. The values of \logg\ we obtain are
319: all typical of a dwarf star \citep[\logg $\sim$ 4.0,][]{trundle07},
320: confirming our previous ``spectroscopic'' findings.
321:
322: \begin{figure}
323: \epsscale{1}
324: \plotone{f2.eps}
325: \caption{Determination of \logg\ for the case \teff\ = 22000 K. The black solid line is the observed \brg\ line. The dashed (dotted, dot-dashed) line is a model with \logg\ = 4.0 (3.00, 4.50). All other parameters are kept constant (especially He/H=0.5). The wavelength range used for the \logg\ determination is indicated. \label{s2_logg_t22}}
326: \end{figure}
327:
328:
329: \begin{figure}
330: \epsscale{1}
331: \plotone{f3.eps}
332: \caption{$\chi^{2}$ as a function of \logg\ for the \teff\ = 22000 K model. The best fit model has \logg\ = 4.0, as also seen from Fig.\ \ref{s2_logg_t22}. \label{chi2_logg_t22}}
333: \end{figure}
334:
335: \textit{Rotational velocity}. We derived a projected rotational
336: velocity of 100$\pm$30 \kms. It best accounts for the shape of the
337: \hei\ doublet: for larger \vsini\ the lines merge into a single
338: component, for lower values, they are too seperated. This value is
339: commonly found for early B stars: according to \citet{abt02}, \vsini\
340: = 127$\pm$8 (108$\pm$8) \kms\ for B0--2V (B3--5V) stars.
341:
342:
343:
344: Table \ref{tab_res} summarizes, for each \teff, the values of \logg\
345: and radius. Since we want to test if S2 is the core of an AGB star, we
346: also give the minimum mass of S2. All our estimates are larger than
347: 14.1 \msun. This is larger than the initial mass of AGB stars, and
348: consequently is well above the mass of the core of such a star. From
349: that, one can thus safely conclude that \textit{S2 is not a stripped
350: AGB star, and is really a genuine B star}.
351:
352:
353:
354:
355:
356:
357:
358:
359: %%#####################################################################
360: %%------------------ He/H and Mdot --------------------
361:
362: \section{Mass loss rate and Helium content}
363: \label{param}
364:
365:
366:
367: %----------------------------------------------
368: \subsection{Mass loss rate}
369:
370: Below a certain threshold, \brg\ is insensitive to \mdot. When \mdot\
371: increases above this limit, the \brg\ absorption profile is
372: progressively filled by emission. By determining this threshold, we
373: can set an upper limit on the mass loss rate of S2 (see Fig.\
374: \ref{s2_mdot_t22}). In practice, the quantity we constrain is the wind
375: \textit{density} $\rho = \dot{M}/(4 \pi R^{2} v_{\infty})$ (where $R$
376: is the stellar radius and $v_{\infty}$ the terminal
377: velocity). Deriving a mass loss rates implies we know
378: $v_{\infty}$. Since the lines we observe in the S2 K band spectrum are
379: insensitive to $v_{\infty}$, we have to assume values. Terminal
380: velocities of B dwarfs are difficult to estimate since the winds are
381: usually weak and the spectra do not show P-Cygni lines or emission
382: lines from which it is usually derived. Hence, we have adopted
383: $v_{\infty}$ = 1000 \kms\ which is appropriate for late O dwarfs
384: \citep[e.g.][]{jc03}. If $v_{\infty}$ would be lower than 1000
385: \kms, estimated mass loss rates would be lower. Our assumptions thus
386: lead to conservative upper limits of \mdot\ $\lesssim 3\times 10^{-7}$
387: \myr. This is lower than \mdot\ required by \citet{loeb04} to feed
388: SgrA* without transport of angular momentum.
389:
390:
391: \begin{figure}
392: \epsscale{1}
393: \plotone{f4.eps}
394: \caption{Effect of mass loss on \brg. When \mdot\ increases, \brg\ starts to depart from a pure absorption to an emission profile. Characterizing the value of \mdot\ for which this occurs provides an upper limit on the mass loss rate of S2. \label{s2_mdot_t22}}
395: \end{figure}
396:
397:
398:
399:
400: %----------------------------------------------
401: \subsection{Helium content}
402:
403:
404:
405: The near infrared spectrum of S2 is dominated by H and He lines. We
406: thus could constrain the He/H abundance ratio. Qualitatively, when
407: this ratio increases, all He~{\sc i} lines have stronger absorption
408: profiles, while H~{\sc i} lines get weaker. In practice, we used the
409: He~{\sc i} 2.149 $\mu m$ and He~{\sc i} 2.184 $\mu m$ lines to
410: constrain He/H since these lines are almost insensitive to
411: microturbulence, which is not the case of \hei\ and He~{\sc i} 2.161
412: $\mu m$. Fig.\ \ref{s2_he} shows the changes in the He~{\sc i} line
413: profiles when He/H is varied between 0.1 and 0.5 (model with \teff\ =
414: 22000 K). We find that the He/H ratio varies from about 0.45 at \teff\
415: = 25000 K up to as high as 1.20 at \teff\ = 19000 K. We proceeded as
416: for \logg\ ($\chi^{2}$ analysis) to derive a 3 $\sigma$ lower limit on
417: the value of He/H. All these lower limits are larger than 0.25,
418: confirming that S2 is He enriched.
419:
420:
421: With He/H $>$ 0.25, S2 falls into the category of the so-called ``He
422: rich'' stars, a class of B stars with He/H between 0.3
423: and 10 \citep{smith96}. Interestingly, these peculiar stars have a
424: very narrow distribution of spectral types centered around B2, similar
425: to what we find for S2 (B0--2.5V). The best studied He-rich star is
426: $\sigma$ OriE. The origin of its surface abundance pattern is
427: explained by a combination of specific wind properties and magnetic
428: field. The star should have a wind weak enough for ion decoupling to
429: occur, i.e. the radiative acceleration, essentially due to metals, is
430: not redistributed among the passive plasma (mainly H and He) because
431: of a too low density and reduced Coulomb forces \citep{kk01}. In that
432: case, Helium might accumulate at the surface of the star. But this
433: extra Helium can remain on the surface only if turbulence, and its
434: associated mixing effect, is suppressed. This calls for a magnetic
435: field strong enough to freeze the stellar surface
436: \citep{hg99}. Our upper limit on the mass loss rate of S2 is
437: consistent with a relatively weak wind. In this scenario, chemical
438: inhomogeneities (spots) are expected on the surface as a result of the
439: interplay between the magnetic field geometry and the wind. We cannot
440: test this prediction with our observations since we do not spatially
441: resolve the stellar surface of S2. Similarly, no CNO abnormalities are
442: expected in this model, which we cannot test either due to the absence
443: of CNO lines (observed and expected) in the near-IR spectrum of S2.
444:
445: \begin{figure}
446: \epsscale{1}
447: \plotone{f5.eps}
448: \caption{Determination of He/H. The black solid line is the observed spectrum of S2. The red dashed (blue dotted) line is a model with He/H=0.10 (0.50). \teff\ = 22000 K and \logg = 4.0. The He rich model is required to fit the S2 spectrum. \label{s2_he}}
449: \end{figure}
450:
451:
452: If the wind is strong enough to develop into a normal homogenous
453: outflow, one might still observe a significant He enrichment if the
454: star has a magnetic field and rotates fast enough. Indeed
455: \citet{mm05} predict that a strong (10$^{4}$ G) magnetic field
456: can lead to solid body rotation of the star. This favors the diffusion
457: of species by meridional circulation. As a consequence, a star can
458: experience strong surface He enrichment: Fig.\ 10 of \citet{mm05}
459: shows that a 15 \msun\ star can reach a He mass
460: fraction of 0.3 (around 0.12 in He/H) in 12 Myr. Contrary to the
461: previous scenario, large N and reduced C abundances are also expected.
462:
463: In view of the available models and of the currently known stellar
464: properties, we thus tentatively propose that S2 is a magnetic star.
465:
466:
467:
468:
469:
470:
471:
472:
473:
474:
475:
476:
477:
478:
479: %%#####################################################################
480: \acknowledgments
481:
482: We thank John Hillier for his help with CMFGEN, as well as Corinne
483: Charbonnel and Ana Palacios for interesting discussions on AGB
484: stars. We thank Andrea Ghez and an anonymous referee for useful
485: comments.
486:
487:
488: %{\it Facilities:} \facility{ESO/VLT}.
489:
490:
491:
492: %%#####################################################################
493: %\bibliography{biblio_letter.bib}
494: \begin{thebibliography}{}
495: \bibitem[Abt et al.(2002)]{abt02} Abt, H., et al., 2002, \apj, 573, 389
496: \bibitem[Alexander(2005)]{tal05} Alexander, T, 2005, PhR, 419, 65
497: %\bibitem[Allen et al.(1990)]{ahh90} Allen, D.A., et al., 1990, \mnras, 244, 706
498: \bibitem[Bouret et al.(2003)]{jc03} Bouret, J.~C., et al., 2003, \apj, 595, 1182
499: \bibitem[Davies \& King(2005)]{kd05} Davies, M.B. \& King, A.R., 2005, \apj, 624L, 25
500: \bibitem[Dray et al.(2006)]{dray06} Dray, L.M., et al., 2006, \mnras, 372, 31
501: \bibitem[Dufton et al.(2006)]{dufton06} Dufton, P.L., et al., 2006, \aap, 457, 265
502: \bibitem[Eisenhauer et al.(2003)]{frank03a} Eisenhauer, F., et al., 2003, \apj, 597L, 121
503: \bibitem[Eisenhauer et al.(2003)]{sinf} Eisenhauer, F., et al., 2003, SPIE, 4841, 1548
504: \bibitem[Eisenhauer et al.(2005)]{frank05} Eisenhauer, F., et al., 2005, \apj, 628, 246
505: \bibitem[Figer et al.(1999)]{figer99} Figer, D.F., et al., 1999. \apj, 506, 384
506: \bibitem[Forestini \& Charbonnel(1997)]{fc97} Forestini, M. \& Charbonnel, C., 1997, A\&AS, 123, 241
507: %\bibitem[Genzel et al.(1996)]{genzel96} Genzel, R., et al., 1996, \apj, 472, 153
508: \bibitem[Genzel et al.(2003)]{genzel03} Genzel, R., et al., 2003, \apj, 594, 812
509: %\bibitem[Ghez et al.(1998)]{ghez98} Ghez, A.M., et al., 1998, \apj, 509, 678
510: \bibitem[Ghez et al.(2003)]{ghez03} Ghez, A.M., et al., 2003, \apj, 586L, 127
511: \bibitem[Grevesse \& Sauval(1998)]{gs98} Grevesse, N. \& Sauval, A.J., 1998, SSRv, 85, 161
512: \bibitem[Hanson et al.(2005)]{hanson05} Hanson, M.M., et al., 2005, \apjs, 161, 154
513: \bibitem[Hillier \& Miller(1998)]{hm98} Hillier, D.J., Miller, 1998, \apj, 496, 407
514: \bibitem[Hunger \& Groote(1999)]{hg99} Hunger, K. \& Groote, D., 1999, \aap, 351, 554
515: %\bibitem[Krabbe et al.(1991)]{krabbe91} Krabbe, A., et al., 1991, \apj, 382L, 19
516: \bibitem[Krabbe et al.(1995)]{krabbe95} Krabbe, A., et al., 1995, \apj, 447L, 95
517: \bibitem[Krti$\breve{\rm c}$ka \& Kub\'at(2001)]{kk01} Krti$\breve{\rm c}$ka, J. \& Kub\'at, J., 2001, \aap, 369, 222
518: \bibitem[Levin \& Beloborodov(2003)]{lb03} Levin, Y. \& Beloborodov, A.M. , 2003, \apj, 590L, 33
519: \bibitem[Loeb(2004)]{loeb04} Loeb, A., 2004, \mnras, 350, 725
520: %\bibitem[Lyubimkov et al.(2004)]{lyu04} Lyubimkov, L.S., et al., 2004, \mnras, 351, 745
521: %\bibitem[Marrone et al.(2007)]{dm07} Marrone, D.P., et al., 2007, \apj, 654L, 57
522: \bibitem[Martins et al.(2007)]{gc07} Martins, F. et al., 2007, \aap, 468, 233
523: \bibitem[Maeder \& Meynet(2005)]{mm05} Meynet, G., Maeder, A., 2005, \aap, 440, 1041
524: \bibitem[Morris(1993)]{morris93} Morris, M., 1993, \apj, 408, 496
525: %\bibitem[Najarro et al.(1997)]{paco97} Najarro, F., et al., 1997, \aap, 325, 700
526: \bibitem[Paumard et al.(2006)]{pgm06} Paumard, T., et al., 2006, \apj, 643, 1011
527: \bibitem[Repolust et al.(2004)]{repolust04} Repolust, T. et al., 2004, \aap, 415, 349
528: \bibitem[Repolust et al.(2005)]{repolust05} Repolust, T. et al., 2005, \aap, 440, 261
529: \bibitem[Sch$\ddot{\rm o}$del et al.(2007)]{schoedel07} Sch$\ddot{\rm o}$del, R., et al., 2007, \aap, 469, 125
530: \bibitem[Smith(1996)]{smith96} Smith, K.C., 1996, Ap\&SS, 237, 77
531: \bibitem[Trundle et al.(2007)]{trundle07} Trundle, C., et al., 2007, \aap, 471, 625
532: \bibitem[Wallace \& Hinkle(1997)]{wh97} Wallace, L., Hinkle, K., 1997, ApJS, 111, 445
533: \end{thebibliography}
534:
535:
536: \end{document}
537: