1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[12pt,referee]{aastex}
3:
4: %% preprint2 produces a double-column, single-spaced document:
5:
6: %% \documentclass[preprint2]{aastex}
7:
8: %% Sometimes a paper's abstract is too long to fit on the
9: %% title page in preprint2 mode. When that is the case,
10: %% use the longabstract style option.
11:
12: %% \documentclass[preprint2,longabstract]{aastex}
13:
14: %% You can insert a short comment on the title page using the command below.
15: \slugcomment{To appear in Ap.\ J., \hspace{5mm} v. \today}
16:
17: \shorttitle{The stellar population of RSGC1}
18: \shortauthors{Davies et al.}
19: \usepackage{graphicx}
20:
21: \usepackage{lscape}
22: \usepackage[bf,small]{caption}
23: \newcommand{\vdag}{(v)^\dagger}
24: \newcommand{\myemail}{davies@cis.rit.edu}
25: \newcommand{\microns}{$\mu$m}
26: \def\EW{$W_{\lambda}$}
27: \def\brg{Br\,$\gamma$}
28: \def\hei{He\,{\sc i}}
29: \def\feii{Fe\,{\sc ii}}
30: %\def\deg{$^{\circ}$}
31: \def\ga{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}}
32: \def\la{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}}
33: \def\msunyr{$M$ \mbox{$_{\normalsize\odot}$}\rm{yr}$^{-1}$}
34: \def\msun{$M$\mbox{$_{\normalsize\odot}$}}
35: \def\minit{$M_{\rm init}$}
36: \def\lsun{$L$\mbox{$_{\normalsize\odot}$}}
37: \def\kms{\,km~s$^{-1}$}
38: \def\EW{$W_{\lambda}$}
39: \def\arcsec{$^{\prime \prime}$}
40: \def\arcmin{$^{\prime}$}
41:
42: %% This is the end of the preamble. Indicate the beginning of the
43: %% paper itself with \begin{document}.
44:
45: \begin{document}
46: \title{The cool supergiant population of the massive young \\ star
47: cluster RSGC1} \author{Ben Davies\altaffilmark{1}, Don
48: F.\ Figer\altaffilmark{1}, Casey J.\ Law\altaffilmark{2}, Rolf-Peter
49: Kudritzki\altaffilmark{3}, Francisco Najarro\altaffilmark{4},
50: Artemio Herrero\altaffilmark{5} and John W.\ MacKenty\altaffilmark{6} }
51:
52: %Antonio Marin\altaffilmark{4},
53: %Melvin Hoare\altaffilmark{5}
54:
55: \affil{$^{1}$Chester F.\ Carlson Center for Imaging Science, Rochester
56: Institute of Technology, 54 Lomb Memorial Drive, Rochester NY, 14623,
57: USA}
58:
59: \affil{$^{2}$Astronomical Institute "Anton Pannekoek", University of
60: Amsterdam, Kruislaan 403, 1098 SJ, Amseterdam, The Netherlands}
61:
62: \affil{$^{3}$Institute for Astronomy, University of Hawaii, 2680
63: Woodlawn Drive, Honolulu, HI, 96822, USA}
64:
65: \affil{$^{4}$Instituto de Estructura de la Materia, Consejo Superior
66: de Investigaciones Cientificas, Calle Serrano 121, 28006 Madrid,
67: Spain.}
68:
69: \affil{$^{5}$Instituto de Astrofísica de Canarias, Via L\`{a}ctea S/N,
70: E-38200 La Laguna, Tenerife, Spain}
71:
72: \affil{$^{6}$Space Telescope Science Institute, 3700 San Martin
73: Drive, Baltimore, MD 21218}
74: %\affil{$^{5}$School of Physics \& Astronomy, University of Leeds,
75: % Woodhouse Lane, Leeds LS2 9JT, UK}
76:
77: \begin{abstract}
78: We present new high-resolution near-IR spectroscopy and OH maser
79: observations to investigate the population of cool luminous stars of
80: the young massive Galactic cluster RSGC1. Using the 2.293\micron\
81: CO-bandhead feature, we make high-precision radial velocity
82: measurements of 16 of the 17 candidate Red Supergiants (RSGs)
83: identified by Figer et al. We show that F16 and F17 are foreground
84: stars, while we confirm that the rest are indeed physically-associated
85: RSGs. We determine that Star F15, also associated with the cluster, is
86: a Yellow Hypergiant based on its luminosity and spectroscopic
87: similarity to $\rho$~Cas. Using the cluster's radial velocity, we have
88: derived the kinematic distance to the cluster and revisited the stars'
89: temperatures and luminosities. We find a larger spread of luminosities
90: than in the discovery paper, consistent with a cluster age 30\% older
91: than previously thought (12$\pm$2Myr), and a total initial mass of
92: $(3\pm1) \times 10^{4}$\msun. The spatial coincidence of the OH maser
93: with F13, combined with similar radial velocities, is compelling
94: evidence that the two are related. Combining our results with recent
95: SiO and H$_2$O maser observations, we find that those stars with maser
96: emission are the most luminous in the cluster. From this we suggest
97: that the maser-active phase is associated with the end of the RSG
98: stage, when the luminosity-mass ratios are at their highest.
99: \end{abstract}
100:
101: \keywords{open clusters \& associations, supergiants, stars:evolution,
102: stars:late-type, masers}
103:
104: %_____________________________________________________________________
105: % ********************************************************************
106:
107: \section{Introduction} \label{sec:intro}
108: The Red Supergiants (RSGs) represent a key evolutionary phase in the
109: life-cycle of stars with initial masses of $\sim$8--30\msun\
110: \citep[e.g.][]{Mey-Mae00}. Though comparatively brief, the
111: mass-loss rate in this stage can be many orders of magnitude greater
112: than on the main-sequence \citep[e.g.][]{Repolust04,v-L05}, and the
113: mass lost in the RSG phase can determine the terminal mass of the
114: star, the appearance of the supernova (SN) explosion, and nature of
115: the stellar remnant \citep[e.g.][]{Heger03}.
116:
117: The study of RSGs is hampered by low-number statistics -- until
118: recently only $\sim$200 were known in the Galaxy, and around 40 in the
119: Large Magellanic Cloud (LMC). Further, the majority of these stars are
120: isolated, and hence have a variety of ages, initial masses and
121: metallicities. Ideally, we would like to study large numbers of RSGs
122: {\it coeval clusters}, where we can be confident that the variables of
123: metallicity and initial stellar mass are fixed.
124:
125: Two recent discoveries now present us with the opportunity to study
126: statistically-significant numbers of RSGs in such clusters. In
127: \citet[][hereafter FMR06]{Figer06} the discovery of an apparent
128: cluster of 14 RSGs was presented; while \citet[][hereafter
129: DFK07]{RSGC2paper} report on the remarkable stellar population of a
130: second cluster in the same region, which it was shown contains 26
131: RSGs. In order to use these two clusters as testbeds with which to
132: study the pre-SN evolution of massive stars, we must first
133: quantitatively study the properties of the stars within each.
134:
135: In the case of the second cluster, hereafter RSGC2 (also known as
136: Stephenson~2), DFK07 used high-resolution spectroscopy to obtain
137: accurate radial velocities of many stars in the region of the cluster,
138: and were able to separate proper cluster-members from
139: background/foreground stars. Further, from the radial velocities they
140: made a quantitative discussion of the kinematic distance to the
141: cluster, thus enabling them to determine the stars' luminosities,
142: temperatures, ages and initial masses.
143:
144: The details of the stellar properties of the first cluster, hereafter
145: RSGC1, are less well constrained. In the discovery paper FMR06 showed
146: with low-resolution spectra that the stars were all late-type; but
147: with no radial velocity information, they relied on the stars' similar
148: near-IR colours to argue that the stars were all at the same
149: distance. They determined this distance by associating a nearby OH
150: maser, detected by \citet{Blommaert94}, with the RSGs. However this
151: maser was {\it singly-peaked} -- such masers are formed in the winds
152: of RSGs and hence are typically {\it doubly-peaked}, with separation
153: twice the outflow speed. If the second peak was missed in the original
154: observations, this would lead to an incorrect radial velocity for the
155: cluster, and hence kinematic distance. Consequently, the cluster
156: membership, as well as the properties of the luminous cool stars of
157: RSGC1 are poorly-constrained when compared to RSGC2.
158:
159: Here, we study the cool supergiants of RSGC1 with data of similar
160: quality to that presented in DFK07 for RSGC2. We present high
161: resolution spectroscopy of the CO bandhead feature of 16 of the 17
162: $K$-bright stars in the field identified in FMR06, allowing us to
163: determine accurate radial velocities of the stars, and establish the
164: cluster membership. We also present new observations of the OH maser
165: source OH25.25-0.16, showing that it is indeed doubly-peaked, and that
166: the central radial velocity of the profile is consistent with the
167: average stellar radial velocity. We use this value, in conjunction
168: with contemporary Galactic rotation curve parameters, to reappraise
169: the distance to the cluster, the stars' temperatures and luminosities,
170: and the age and intial mass of the cluster.
171:
172: We begin in Sect.\ \ref{sec:obs} with a description of the
173: observations and data reduction steps, and we describe the results and
174: analysis of the data in Sect.\ \ref{sec:res}. In Sect.\ \ref{sec:disc}
175: we derive the cluster's age and mass, and compare with similar
176: analyses of RSGC2. Finally, we use the stellar population of RSGC1 to
177: investigate the maser-active phase of RSGs.
178:
179: %_____________________________________________________________________
180: % ********************************************************************
181:
182: \section{Observations \& data reduction} \label{sec:obs}
183:
184: \begin{table}[t]
185: \caption{Data for the stars observed. Designations are from FMR06,
186: coordinates and apparent magnitudes are from 2MASS. }
187: \label{tab:obsdata} \centering
188: \begin{tabular}{lccccccccc}
189: \hline \hline
190: ID & $\alpha$ & $\delta$ & m$_{\rm J}$ & m$_{\rm H}$ & m$_{\rm K_{S}}$ \\
191: & \multicolumn{2}{c}{J2000} & & & \\
192: \hline
193: F01 & 18 37 56.29 & -6 52 32.2 & 9.748 & 6.587 & 4.962 \\
194: F02 & 18 37 55.28 & -6 52 48.4 & 9.904 & 6.695 & 5.029 \\
195: F03 & 18 37 59.72 & -6 53 49.4 & 9.954 & 6.921 & 5.333 \\
196: F04 & 18 37 50.90 & -6 53 38.2 & 9.658 & 6.803 & 5.342 \\
197: F05 & 18 37 55.50 & -6 52 12.2 & 10.547 & 7.178 & 5.535 \\
198: F06 & 18 37 57.45 & -6 53 25.3 & 9.866 & 7.038 & 5.613 \\
199: F07 & 18 37 54.31 & -6 52 34.7 & 9.941 & 7.065 & 5.631 \\
200: F08 & 18 37 55.19 & -6 52 10.7 & 10.772 & 7.330 & 5.654 \\
201: F09 & 18 37 57.77 & -6 52 22.2 & 10.262 & 7.240 & 5.670 \\
202: F10 & 18 37 59.53 & -6 53 31.9 & 10.179 & 7.218 & 5.709 \\
203: F11 & 18 37 51.72 & -6 51 49.9 & 10.467 & 7.325 & 5.722 \\
204: F12 & 18 38 3.30 & -6 52 45.1 & 10.143 & 7.238 & 5.864 \\
205: F13 & 18 37 58.90 & -6 52 32.1 & 10.907 & 7.716 & 5.957 \\
206: F14 & 18 37 47.63 & -6 53 2.3 & 10.495 & 7.576 & 6.167 \\
207: F15 & 18 37 57.78 & -6 52 32.0 & 10.651 & 8.070 & 6.682 \\
208: F16 & 18 38 1.30 & -6 52 52.0 & 13.617 & 9.608 & 7.558 \\
209: F17 & 18 37 48.77 & -6 53 7.7 & 12.763 & 10.188 & 9.003 \\
210: \hline
211: \end{tabular}
212: \end{table}
213:
214: \subsection{Radio observations \& data reduction}
215: OH maser observations at 18 cm were carried out with the
216: VLA\footnote{The Very Large Array (VLA) is operated by the National
217: Radio Astronomy Observatory under cooperative agreement with the
218: National Science Foundation.} on 25 May, 2006 in the AB configuration.
219: The 256-channel spectrum had a bandwidth of 1.56 MHz giving a velocity
220: resolution of 1.1 km s$^{-1}$ and was centered at $v_{\rm{LSR}}=100$
221: km s$^{-1}$. The observation consisted of a single pointing with
222: J2000 coordinates of (18h37m52s,-06\degr53\arcmin40\arcsec) and had a
223: field of view of about 27\arcmin. The beam size is
224: 3.5\arcsec$\times$1.8\arcsec. The total integration time was 340
225: minutes, giving an rms error of about 4 mJy beam$^{-1}$ per 4 kHz
226: channel.
227:
228: The data were calibrated using the AIPS package. The flux and
229: bandpass calibrator was 1331+305, while 1822-096 was used as a phase
230: calibrator. The phase calibrator was mildly resolved in this
231: observation, so baselines with uv distances greater than 50 $k\lambda$
232: were removed when solving for a phase solution. Minimal flagging was
233: required to remove interference. The astrometric accuracy of our
234: positions is dominated by the uncertainty in the position of our phase
235: calibration source, which is 0.3\arcsec. The compact source
236: GPSR5 25.266-0.161 has a J2000 position of (18h37m57.9934s,
237: -06\degr53\arcmin30.973\arcsec), consistent with previous measurements
238: with an astrometric accuracy of ~2\arcsec\ \citep{Becker94}.
239:
240: \subsection{High-resolution spectroscopy}
241: \subsubsection{Observations}
242: Observations were taken with {\it NIRSPEC}, the cross-dispersed
243: echelle spectrograph mounted on Keck-II, during the night of 5$^{th}$
244: May 2006. We observed stars F01 -- F17, with the exception of
245: F12 which was missed due to time constraints. F15 was
246: re-observed in a separate observing run on 12$^{th}$ Aug 2006, using
247: the same technical setup as described below. Table \ref{tab:obsdata}
248: lists the coordinates and 2MASS magnitudes of the 17 stars, as
249: identified in FMR06. The stars' locations are illustrated in Fig.\
250: \ref{fig:ohpic}.
251:
252: We used the NIRSPEC-7 filter and 0.576\arcsec$\times$24\arcsec\
253: slit. Setting the dispersion angle to 62.53\degr, and cross-disperser
254: angle to 35.53\degr, this gave us a spectral resolution of
255: $\sim$17,000 in the wavelength range 1.9--2.4\microns. We integrated
256: for 20s in two nodded positions along the slit for each star. We
257: observed the B0V star HD~171305 as a telluric standard. We used a
258: continuum lamp to obtain flat-field frames, while for wavelength
259: calibration, we observed Ar, Ne, Xe and Kr lamps to obtain as many
260: spectral lines as possible in the narrow wavelength range. To sample
261: the gaps between these lines, we also observed the continuum lamp
262: through an etalon filter.
263:
264: \subsubsection{Data reduction}
265: To remove sky emission, dark current and bias level, we subtracted nod
266: pairs of spectra. Fluctuations in pixel-to-pixel sensitivity were
267: corrected for by dividing through by the normalized flat-field frame.
268:
269: The optics of {\it NIRSPEC} produce spectral orders which are warped
270: in both the spatial and dispersion directions, and before the spectra
271: can be extracted this warping must be corrected for in a process known
272: as {\it rectification}. The methodology we use is the same as that in
273: DFK07, and is explained in more detail in \citet{Figer03}. Here we
274: give a brief summary.
275:
276: Spatial rectification is done by adding nod-pair spectra and fitting
277: the spectral traces with a polynomial. Rectification in the dispersion
278: direction is more complicated; it involves obtaining accurate
279: wavelengths of the etalon lines using the arc frames, and assuming
280: that the wavelength of the $n$th order etalon-line $\lambda_{n}$ is
281: governed by the separation of the etalon plates $t$ via the relation
282: $\lambda_{n} = t/2n$ (the {\it etalon equation}).
283:
284: The arc lines were used to get initial estimates of the etalon-line
285: wavelengths, and hence of the etalon plate separation. The etalon-line
286: wavelengths were then recomputed using the etalon equation, and used
287: to re-estimate the wavelengths of the arc lines. The etalon thickness
288: was fine-tuned in an iterative process until the residuals between the
289: measured and predicted arc-line wavelengths across all orders were
290: minimized.
291:
292: After rectification, the spectra were extraced from each frame by summing
293: the pixels across the trace in each channel. Shifts between spectra of
294: up to 4\kms\ ($\la$1 pixel), caused by the star not being quite in the
295: centre of the slit, were corrected for by cross-correlating the
296: atmospheric CO$_{2}$ feature at 2.05\microns\ in each spectrum.
297:
298: The accuracy of the final wavelength solution is determined from the
299: residuals between the observed and predicted arc-line wavelengths in
300: the etalon-fitting process described above, and is better than
301: $\pm$4\kms. The internal error between spectra, from the CO$_{2}$
302: telluric feature, is $\ll$1\kms, and so is dominated by systematics in
303: our analysis process which we estimate to be $\pm$1\kms\ (DFK07).
304:
305:
306: %_____________________________________________________________________
307: % ********************************************************************
308:
309: \section{Results \& analysis} \label{sec:res}
310:
311: \begin{figure}[t]
312: \centering
313: % \includegraphics[width=12cm,bb=18 180 544 580,clip]{spec_G25.25_maser.eps}
314: \includegraphics[width=12cm,bb=18 180 544 580,clip]{fig1.eps}
315: \caption{Spectrum at 1612MHz of the source OH25.25-0.16. Our new
316: observations clearly show the second peak at
317: $\sim$135\kms. Errorbars show $\pm1\sigma$ uncertainty and triangles
318: show $5\sigma$ upper limits in the flux density. }
319: \label{fig:ohspec}
320: \end{figure}
321:
322: \begin{figure}[t]
323: \centering
324: % \includegraphics[width=12cm,bb=40 20 750 750,clip]{maserpos.eps}
325: \includegraphics[width=12cm,bb=40 20 850 750,clip]{fig2.eps}
326: \caption{Contour-plot of the 1612Hz radio image, overlayed on
327: the 2MASS K$_{S}$-band image of the same region. The image is
328: centered on the indicated coordinates (epoch J2000). Contours are
329: drawn at 5, 10, 20 and 40$\sigma$ above the background; the
330: 1$\sigma$ level corresponds to 0.87mJy/beam. The position of the OH
331: maser is indicated, and is coincident with star F13. The other
332: identified radio sources in the field, labelled in the figure, were
333: shown by \citet{T-R06} to be extragalactic in origin. The stars F01
334: to F17, as designated by FMR06, are also labelled.}
335: \label{fig:ohpic}
336: \end{figure}
337:
338: \subsection{The maser source OH25.25-0.16} \label{sec:maser}
339: As mentioned in Sect.\ \ref{sec:intro}, the 1612MHz OH maser forms in
340: the outflows of RSGs, far above the stellar surface. The velocity
341: profiles are therefore typically doubly-peaked, with a separation
342: twice the terminal velocity of the outflow centred on the star's
343: systemic velocity. However, when observed by \citet{Blommaert94},
344: OH25.25-0.16 appeared only as a single peak with a radial velocity
345: $v_{\rm LSR} = 102.2$\kms.
346:
347: Our new observation of the OH maser source is shown in velocity-space
348: in Fig.\ \ref{fig:ohspec}. Here it can be seen that we clearly detect
349: the second peak. The flux-weighted mean velocities of the two peaks
350: are $103.8 \pm 0.1$\kms\ and $138.1 \pm 0.9$\kms, calculated using all
351: channels with emission greater than 5$\sigma$ above the
352: background. The average velocity of the peaks is $120.9 \pm 0.9$\kms,
353: consistent with the average velocities of the SiO masers in the
354: cluster, $120.7 \pm 3.2$\kms, observed by \citet{N-D06}. The implied
355: outflow speed is 17.1$\pm$0.6\kms, a typical outflow speed for RSGs
356: \citep{R-Y98}, and similar to the outflow speed of S~Per
357: \citep[$\sim$16\kms\,][]{Diamond87}, which occupies a similar location
358: in the HR-diagram as the stars of RSGC1 \citep[][see Sect.\
359: \ref{sec:disc}]{G-H76}.
360:
361: In Fig.\ \ref{fig:ohpic} we overlay a contour plot of the 1612MHz
362: emission on the 2MASS $K_{S}$-band image of the cluster. We find the
363: positional centroid of the OH maser to be
364: 18$^{h}$37$^{m}$58.882$^{s}$, -06\degr52\arcmin32.28\arcsec\ (J2000),
365: with a positional uncertainty of 0.3\arcsec. The J2000 position of the
366: maser is consistent with previous measurement by Blommaert et
367: al. (1994), which had a positional accuracy of ~4". The maser is also
368: spatially coincident with star F13, whose 2MASS coordinates are
369: 18$^{h}$37$^{m}$58.908$^{s}$, -06\degr52\arcmin32.11\arcsec\ (J2000),
370: with positional uncertainty 0.06\arcsec.
371:
372: The central radial velocity of the OH maser is also consistent with
373: the SiO observation of F13 by \citet{N-D06}\footnote{This radial
374: velocity was found from the average velocity of the high- and
375: low-velocity edges. The centroid of the peak of this source was found
376: to have a radial velocity of 116.5$\pm$2\kms}, $v_{\rm LSR, F13} =
377: 120.5 \pm 2.0$\kms, and with the CO-bandhead radial velocity
378: measurement of F13 presented in this paper, 125.4$\pm$4\kms\ (see
379: Sect.\ \ref{sec:radvs}). From this evidence, it seems highly likely
380: that the OH maser originates in the outflow of F13.
381:
382: In addition to the OH maser, we also detect the continuum sources
383: GPSR5 (25.266-0.161, 25.252-0.139, 25.237-0.15). These sources were
384: shown conclusively to be extragalactic in origin by \citet{T-R06}.
385:
386:
387: \begin{figure}[p]
388: \centering
389: % \includegraphics[width=16cm,bb=25 30 560 679,clip]{order2_1_dummy.eps}
390: \includegraphics[width=16cm,bb=25 30 560 679,clip]{fig3.eps}
391: \caption{High-resolution spectra of all stars observed in the region
392: of the CO-bandhead feature. The dotted blue line indicates the zero
393: velocity of the blue edge of the CO bandhead. }
394: \label{fig:cobandhead}
395: \end{figure}
396:
397: \begin{table}[p]
398: \begin{center}
399: \caption{Observed data for RSGs. (1): the stellar IDs (from
400: FMR06); (2): the radial velocity of each star ($\pm$4\kms), and
401: the radial velocity measured by \citet{N-D06}, where available
402: ($\pm$2\kms); (3): effective temperature; (4): spectral type,
403: accurate to $\pm$2 subtypes; (5) derived extinction towards each
404: star; (6): absolute $K$-band magnitude; and (7): bolometric
405: luminosity. }
406: \begin{tabular}{lccccccc}
407: \hline \hline
408: (1) & \multicolumn{2}{c}{(2)} & (3) & (4) & (5) & (6) & (7) \\
409: ID & $V_{\rm LSR}$ & (ND06)$^{1}$ & $T_{\rm eff}$~(K) & Spec Type
410: & $A_{K_{S}}$ & $M_{K}$ & log($L_{bol}$/\lsun) \\
411: & \multicolumn{2}{c}{(\kms)} & & & & & \\
412:
413: \hline
414: F01 & 129.5 & (117.7) & 3450$\pm$127 & M5 & 2.58$\pm$0.09 & -11.75$^{+0.34}_{-0.30}$ & 5.42$^{+0.12}_{-0.13}$ \\
415: F02 & 114.2 & (119.7) & 3660$\pm$127 & M2 & 2.83$\pm$0.07 & -11.92$^{+0.34}_{-0.30}$ & 5.56$^{+0.12}_{-0.13}$ \\
416: F03 & 127.2 & -- & 3450$\pm$127 & M5 & 2.46$\pm$0.09 & -11.28$^{+0.34}_{-0.30}$ & 5.24$^{+0.12}_{-0.13}$ \\
417: F04 & 121.2 & (124.3) & 3752$\pm$117 & M1 & 2.46$\pm$0.04 & -11.24$^{+0.32}_{-0.28}$ & 5.32$^{+0.11}_{-0.13}$ \\
418: F05 & 124.8 & -- & 3535$\pm$130 & M4 & 2.77$\pm$0.08 & -11.36$^{+0.34}_{-0.30}$ & 5.29$^{+0.12}_{-0.14}$ \\
419: F06 & 120.7 & -- & 3450$\pm$127 & M5 & 2.19$\pm$0.09 & -10.70$^{+0.34}_{-0.30}$ & 5.00$^{+0.12}_{-0.13}$ \\
420: F07 & 121.6 & -- & 3605$\pm$151 & M3 & 2.33$\pm$0.12 & -10.81$^{+0.36}_{-0.32}$ & 5.10$^{+0.13}_{-0.14}$ \\
421: F08 & 128.2 & -- & 3605$\pm$151 & M3 & 2.84$\pm$0.12 & -11.33$^{+0.36}_{-0.32}$ & 5.30$^{+0.13}_{-0.14}$ \\
422: F09 & 121.6 & -- & 3399$\pm$150 & M6 & 2.44$\pm$0.08 & -10.92$^{+0.33}_{-0.29}$ & 5.07$^{+0.12}_{-0.13}$ \\
423: F10 & 122.0 & -- & 3605$\pm$151 & M3 & 2.45$\pm$0.12 & -10.86$^{+0.36}_{-0.32}$ & 5.12$^{+0.13}_{-0.14}$ \\
424: F11 & 124.1 & -- & 3535$\pm$130 & M4 & 2.63$\pm$0.08 & -11.03$^{+0.34}_{-0.30}$ & 5.16$^{+0.12}_{-0.14}$ \\
425: F13 & 125.4 & (120.5) & 4015$\pm$140 & K2 & 3.19$\pm$0.09 & -11.39$^{+0.34}_{-0.30}$ & 5.45$^{+0.12}_{-0.13}$ \\
426: F14 & 122.0 & -- & 3605$\pm$151 & M3 & 2.29$\pm$0.12 & -10.25$^{+0.36}_{-0.32}$ & 4.87$^{+0.13}_{-0.14}$ \\
427: F15 & 120.8 & -- & 6850$\pm$350 & G0 & 2.65$\pm$0.04 & -10.07$^{+0.40}_{-0.36}$ & 5.36$^{+0.14}_{-0.16}$ \smallskip \\
428: {\it F16 } & {\it 42.6 }$^{2}$ & -- & -- & -- & -- & -- & -- \\
429: {\it F17 } & {\it 33.2 }$^{2}$ & -- & -- & -- & -- & -- & -- \\
430: \hline \label{tab:rsgparams}
431: \end{tabular}
432: \end{center}
433: {$^{1}$\footnotesize We quote the \citet{N-D06} values measured
434: by taking the average of the high- and low-velocity edges of the
435: maser profiles. \\
436: $^{2}$Stars F16 \& F17 are determined to be foreground stars, and due
437: to the uncertainties in reddening and distance we derive no
438: stellar parameters for these stars.}
439: \end{table}
440:
441: \subsection{The high-resolution spectra}
442: The high-resolution observations of the region around the CO-bandhead
443: feature at 2.293\microns\ for the 17 K-bright stars are shown in Fig.\
444: \ref{fig:cobandhead}. All stars observed, with the exception of F15,
445: show the feature strongly in absorption; F15 has it weakly in
446: emission. Quantitative analysis of the high-resolution spectroscopy
447: results are described below.
448:
449: \subsubsection{Stellar radial velocities} \label{sec:radvs}
450: The radial velocities of each star serve two purposes; firstly they
451: allow us to distinguish between genuine cluster stars and foreground
452: stars with similar colours; and secondly they allow us to derive a
453: kinematic distance to the cluster.
454:
455: The differences in radial velocities of the stars can be seen
456: qualitatively in Fig.\ \ref{fig:cobandhead} -- the blue-edges of the
457: CO bandhead in stars F16 and F17 are noticably blue-shifted compared
458: to stars F01 -- F15, which all have very similar radial velocities.
459:
460: In order to accurately quantify the radial velocities of the stars, we
461: implemented the same technique as presented in \citet{Figer03} and
462: DFK07. We cross-correlated the spectra shown in Fig.\
463: \ref{fig:cobandhead} with that of the high-resolution spectrum of Arcturus presented in \citet{W-H96arct}, which had been degraded to
464: the same spectral resolution as our data. For star F15, which has CO
465: in emission, we inverted the spectrum before analysis. We experimented
466: with isolating different spectral ranges during the cross-correlation,
467: to determine the robustness of our measurements. We found that the
468: measured velocities were stable to within $\pm$1\kms, which we take to
469: be the {\it internal} error between individual measurements. The
470: absolute uncertainty on the measurements is limited by the accuracy of
471: the wavelength solution, $\pm$4\kms\ (see Sect.\ \ref{sec:obs}).
472:
473: We find that stars F01 -- F15 all have radial velocities in the range
474: $v_{\rm LSR} \sim$115-125\kms, while stars F16 and F17 have $v_{\rm
475: LSR}$ of 33\kms\ and 43\kms\ respectively. From this, we conclude that
476: the two faintest stars observed are foreground stars, while stars F01
477: -- F15 are physical members of the cluster. As F12 was unobserved, the
478: status of this star is still unclear. We note that the CO absorption
479: strengths of stars F16 and F17 are much lower than those of the RSGs,
480: and instead are more typical of less luminous stars. This is
481: consistent with these stars being foreground objects.
482:
483: The measured radial velocities of all stars observed are listed in
484: Table \ref{tab:rsgparams}. Also listed in Table \ref{tab:rsgparams},
485: where available, are the radial velocities determined from SiO maser
486: emission by \citet{N-D06}. We see that the two measurements of stars
487: F02, F04 and F13 are within $\sim$2$\sigma$, while for F01 it is
488: $\sim$3$\sigma$. As SiO masers are commonly thought to trace the
489: stellar systemic velocity \citep{Jewell91}, we would expect the two
490: velocity measurements to agree well. However, we note that the
491: observations of \citet{N-D06} may have been hampered by their
492: beam-size; their observations of F01, F02 and F13 have at least one
493: other RSG within the beam FWHM. In the case of F04, which is well
494: separated from the other RSGs, the CO bandhead and SiO maser
495: measurements are in excellent agreement. As mentioned in Sect.\
496: \ref{sec:maser}, the measured radial velocity of F13 is in excellent
497: agreement with observations of the OH and SiO maser sources at the
498: same location.
499:
500:
501: \subsubsection{Distance to RSGC1} \label{sec:dist}
502: We take the mean radial velocity of the stars observed at high
503: spectral resolution and compare with the Galactic rotation curve,
504: using the contemporary measurements collated by \citet{K-D07}. The
505: mean radial velocity of the stars from the CO observations is
506: 123.0$\pm$1.0\kms, with an uncertainty determined from Poisson
507: statistics of the measurements. The absolute uncertainty on the radial
508: velocity is therefore dominated by that in the wavelength solution,
509: $\pm$4\kms. This compares well to the average radial velocity found by
510: \citet{N-D06}, $\sim$120$\pm$2\kms, from their SiO maser
511: observations\footnote{These values are the average of their two
512: different methods of measuring the radial velocity of each star from
513: the maser line profile}, and the central velocity of our new 1612MHz
514: OH maser observation, $120.9 \pm 0.9$\kms.
515:
516: In Fig.\ \ref{fig:grotcurve} we compare our radial velocity to the
517: Galactic rotation curve in the direction of $l=25.15$\degr,
518: $b=-0.15$\degr. We use the distance to the Galactic centre $D_{\rm
519: GAL} = 7.5 \pm 0.3$kpc \citep{Eisenhauer05}, and solar rotational
520: velocity $\Theta_{\sun} = 214 \pm 7$\kms\ \citep{F-W97,R-B04}. We use
521: the uncertainties on these values to construct `maximal' and `minimal'
522: rotation curves in Fig.\ \ref{fig:grotcurve}.
523:
524: \begin{figure}[t]
525: \centering
526: % \includegraphics[width=14cm,bb=47 10 605 480]{grotcurve2.eps}
527: \includegraphics[width=14cm,bb=47 10 605 480]{fig4.eps}
528: \caption{Galactic rotation curve in the direction of RSGC1
529: \citep{B-B93}, using the measurements of collated by
530: \citet{K-D07}. }
531: \label{fig:grotcurve}
532: \end{figure}
533:
534: The cluster radial velocity actually extends beyond the
535: asymptotic point of the curve; however it lies well within the two
536: `error' curves, and therefore could simply be due to the uncertainties
537: in $D_{\rm GAL}$ and $\Theta_{\sun}$. To determine the distance to the
538: cluster, we take the average of the two points where the radial
539: velocity intercepts the `maximal' rotation curve. This gives us a
540: kinematic distance to the cluster of 6.60$\pm$0.89\,kpc, slightly
541: larger than the distance quoted in FMR06 when using the radial
542: velocity of the singly-peaked OH maser.
543:
544:
545: \subsubsection{Effective temperatures} \label{sec:temp}
546: To determine the spectral-types of the RSGs, and hence their effective
547: temperatures, we used the same empirical method described in
548: \citet{RSGC2paper}. We compared the equivalent width (\EW) of the CO
549: bandhead absorption with that of template stars, taken from the
550: catalogues of \citet{K-H86}, \citet{W-H96} and \citet{W-H97}. We
551: defined a measurement region of 2.294-2.304\microns, and defined the
552: continuum as the median of the range 2.288-2.293\microns. We estimated
553: the uncertainty by repeating the measurements with slightly adjusted
554: continuum regions, and found that measurements were stable to
555: $\sim$1\AA, or $\sim$5\%. To find the spectral-types of the RSGs, we
556: compared the \EW\ measurements with a linear fit to the \EW\ of the
557: template stars as a function of spectral-type. Using this method, we
558: are able to determine spectral-types to within $\pm$2 subtypes (see
559: DFK07). In converting spectral-type to effective temperature, we used
560: the temperature scale of \citet{Levesque05}. The derived
561: spectral-types and effective temperatures for the RSGs are listed in
562: Table \ref{tab:rsgparams}.
563:
564: For star F15 the method described above breaks down, as this star has
565: CO in emission. The star's radial velocity suggests that it is part of
566: the cluster, and hence a supergiant. It was assigned the spectral-type
567: G6~I in FMR06, and hence deemed to be a `Yellow Hypergiant' (YHG),
568: based on its weak CO-bandhead absorption. Given that this was only a
569: marginal detection of CO absorption here we reappraise the star's
570: spectral type.
571:
572: \begin{figure}[p]
573: \centering
574: % \includegraphics[width=14cm,bb=25 10 560 720]{yhg.eps}
575: \includegraphics[width=14cm,bb=25 10 560 720]{fig5.eps}
576: \caption{The high-resolution K-band spectrum of F15, compared to the
577: spectra of $\gamma$\,Cyg and $\alpha$\,Sge from \citet{W-H97}. }
578: \label{fig:yhg}
579: \end{figure}
580:
581: Figure \ref{fig:yhg} shows the spectra of F15 on the two occasions it
582: was observed. It can be seen that the CO-bandhead, which was present
583: in emission in May, was not detected at all in August. Over the rest
584: of the star's spectrum, no discernable variability is observed.
585:
586: For comparison, Fig.\ \ref{fig:yhg} shows similar template spectra of
587: $\gamma$~Cyg, spectral-type F8~Iab, and $\alpha$~Sge, spectral-type
588: G1~III, which have been resampled to the same spectral resolution as
589: our data \citep[taken from][]{W-H97}. Though admittedly $\alpha$~Sge
590: has a lower luminosity class than that inferred for F15, it serves to
591: give some insight into F15's temperature. At spectral-type G1, the
592: CO-bandhead absorption is still seen, albeit weakly. The star also
593: shows the atomic absorption lines of Al~{\sc i}, Fe~{\sc i}, Mg~{\sc
594: i}, and \brg. However, at this temperature molecular absorption,
595: mostly from CN, can be seen as the undulating `noise'-like features
596: throughout the spectrum. At spectral-type F8, the CN and CO absorption
597: are gone, while the atomic absorption lines remain.
598:
599: From these comparison spectra, we assign a spectral-type to F15 of G0,
600: $\pm$2 subtypes, and hence an effective temperature of $T_{\rm eff} =
601: 6500-7200$K. Indeed, the star is very similar spectroscopically to
602: $\rho$~Cas, one of the archetypal YHGs, which also shows transient CO
603: emission / absorption \citep[see spectra presented in][]{Gorlova06}.
604:
605:
606: \subsubsection{Extinction} \label{sec:extinct}
607: The extinction towards each star is measured by comparing the 2MASS
608: infrared colours of the stars to the intrinsic colours of supergiants
609: with the same spectral-type. For the RSGs we use the observations of
610: Galactic stars with luminosity class Iab from \citet{Elias85}; while
611: for F15 (the YHG) we use the tables of \citet{Koorneef83}. We define
612: the excess between the observed and intrinsic colours as the reddening
613: towards each star. We convert this reddening to an extinction towards
614: each star using the relationship of \citet{R-L85},
615:
616: \begin{equation}
617: A_{K_{S}} = \frac{E_{\lambda - K_{S}}}
618: {(\lambda/\lambda_{K_{S}})^{1.53} - 1}
619: \label{equ:reddening}
620: \end{equation}
621:
622: \noindent where $\lambda$ is the wavelength appropriate for the 2MASS
623: $J$ or $H$ filters. To determine the uncertainty in each extinction
624: measurement we derive extinctions for the upper and lower limits to
625: each star's spectral-type. Where the difference between the derived
626: $A_{K_{S}} (J-K_{S})$ and $A_{K_{S}} (H-K_{S})$ values is outside this
627: uncertainty, we adopt half this difference as the error in the
628: measurement.
629:
630: Using this method, we find a median extinction of $A_{K_{S}} (J-K_{S})
631: = 2.58$, and $A_{K_{S}} (H-K_{S}) = 2.62$. Each have uncertainties of
632: 0.07mags from Poisson statistics, and are therefore in good agreement
633: with one another. We adopt the mean of these measurements, $A_{K_{S}}
634: = 2.60 \pm 0.07$ as the median extinction towards the cluster. We note
635: that the extinction towards the YHG, $A_{\rm K, F15} = 2.65\pm0.02$,
636: is consistent with that derived for the RSGs.
637:
638: The median cluster extinction is slightly lower than the $A_{K_{S}} =
639: 2.74 \pm 0.02$ derived in FMR06. We consider the latest measurement to
640: be the more reliable, due to the extra assumptions used in FMR06:
641: instead of dereddening each star according to the intrinsic colours
642: appropriate for its spectral-type, they dereddened {\it all} stars to
643: the mean colour of M supergiants from \citet{Elias85}. The ranges in
644: colours of M supergiants are $\Delta (J-K) \sim$0.3 and $\Delta (H-K)
645: \sim$0.1, which each correspond to $\Delta A_{K_{S}} \sim$0.2 using
646: Eqn.\ (\ref{equ:reddening}). Hence, if the spectral-types of the RSGs
647: are asymmetrically distributed about the mean spectral-type, this may
648: produce a derived extinction out by as much as 0.2, consistent with
649: the difference between the extinctions derived here and in FMR06.
650:
651:
652:
653: \subsubsection{Luminosities} \label{sec:lum}
654: We take the extinctions toward each star derived in Sect.\
655: \ref{sec:extinct}, in conjunction with the distance to the cluster
656: estimated in Sect.\ \ref{sec:dist} to determine the absolute
657: $K_{S}$-band magnitudes of the stars. To convert these to bolometric
658: luminosities ($L_{\star}$) we interpolate over the contemporary
659: bolometric corrections $BC_{K}$ for RSGs, given in \citet{Levesque05},
660: for the stellar temperatures derived in Sect.\ \ref{sec:temp}.
661:
662: The absolute uncertainty in each $L_{\star}$ determination is $\sqrt(
663: \delta A_{K} + \delta BC_{K} + \delta D_{\rm cl})$. As we can
664: confidently make the approximation that all the stars are all located
665: at the same distance, the uncertainty in distance $\delta D_{\rm
666: cl})$ can be neglected when analysing the luminosity spread of the
667: stars to infer the cluster's age (see Sect.\ \ref{sec:age}). The
668: uncertainties in both $A_{K}$ and $BC_{K}$ are carried forward from
669: the error in $T_{\rm eff}$, and are determined by substituting the
670: upper and lower limits to the stars' temperatures.
671:
672: We list the stars' luminosities in Table \ref{tab:rsgparams}, along
673: with uncertainties which include the error in the cluster
674: distance. The newly-derived values are similar to those quoted in
675: FMR06, typically within $\pm$0.3dex. As stated above, FMR06 used the
676: same extinction towards all stars in the cluster, an approximation
677: which breaks down if there are large variations in the interstellar
678: extinction across the field or if a star has extra circumstellar
679: extinction. Also, our high S/N, high-resolution spectra give a more
680: accurate picture of variations in CO equivalent width, and hence
681: better constrained stellar temperatures -- key in evaluating the
682: stars' bolometric corrections. For these reasons, we conclude the
683: bolometric luminosities derived here to be the more accurate than
684: those quoted in FMR06.
685:
686:
687: \subsubsection{Spectral energy distributions}
688: Using the Galactic plane surveys of MSX and GLIMPSE
689: \citep{Egan01,Benjamin03}, as well as 2MASS, we have collated IR
690: photometry for all stars observed here. For the brighter stars, mid-IR
691: photometry is unavailable due to the stars saturating in the images
692: (e.g. F01 and F02); while fainter stars in crowded regions (e.g. F15)
693: are dwarfed by brighter nearby stars (e.g. F09). We rejected all
694: upper-limit measurements and all detections fainter than 10$\sigma$.
695:
696: In Fig.\ \ref{fig:seds} we plot the spectral energy distribution (SED)
697: of each star. Also plotted in the figure are the stars' de-reddened
698: photometry, which were calculated using the interstellar extinction
699: toward each star in conjuction with the extinction-laws quoted by
700: \citet{Indebetouw05} and \citet{Messineo05}. We have overplotted
701: blackbody curves appropriate for the stellar temperatures and
702: luminosities calculated in Sects.\ \ref{sec:temp} and \ref{sec:lum}.
703:
704: In all cases, the blackbodies provide excellent fits to the near-IR
705: photometry, even in the case of F15 where a less accurate method of
706: temperature estimation was possible. This serves to validate the
707: stellar luminosities and temperatures of the stars derived above
708: (c.f.\ Fig.\ 13 of FMR06, where poorer fits to the IR photometry were
709: obtained.).
710:
711: In all cases where photometry is available, the mid-IR MSX data shows
712: that the stars have considerable excess emission. This is indicative
713: of warm circumstellar dust, a product of the high mass-loss rates of
714: the stars. Also, the SEDs appear to show bumps around 12\microns,
715: which can be understood as silicate emission from the oxygen-rich
716: dust. A detailed study of the circumstellar material around these
717: objects will be the subject of a future paper.
718:
719:
720:
721: \begin{figure}[p]
722: \centering
723: % \includegraphics[width=16cm,bb=10 20 566 566]{sed_1.eps}
724: \includegraphics[width=16cm,bb=10 20 566 566]{fig6a.eps}
725: \caption{Spectral energy distributions of cluster members. Crosses
726: are raw photometry from 2MASS, GLIMPSE \& MSX; filled circles are
727: dereddened according to the stellar extictions derived in Sect.\
728: \ref{sec:extinct}. Over plotted in green is a black-body curve, with
729: $T$ determined from spectral-type of star and luminosity according
730: to dereddened K-band magnitude at a distance of 5.88kpc. }
731: \label{fig:seds}
732: \end{figure}
733: \addtocounter{figure}{-1}
734: \begin{figure}[t]
735: \centering
736: % \includegraphics[width=16cm,bb=10 180 566 566]{sed_2.eps}
737: \includegraphics[width=16cm,bb=10 180 566 566]{fig6b.eps}
738: \caption{ Cont. }
739: \end{figure}
740:
741: %\subsection{The kinematic distance of RSGC1}
742:
743:
744: %_____________________________________________________________________
745: % ********************************************************************
746:
747: \section{Discussion} \label{sec:disc}
748:
749: \subsection{Cluster age} \label{sec:age}
750: \begin{figure}[t]
751: \centering
752: % \includegraphics[width=12cm,bb=15 10 560 450]{rsg1_hrd.eps}
753: \includegraphics[width=12cm,bb=15 10 560 450]{fig7.eps}
754: \caption{H-R diagram, showing the location of the RSGs and one
755: YHG. Overplotted are rotating Geneva isochrones from
756: \citet{Mey-Mae00}, with ages 10, 12 and 14 Myr. The luminosity
757: spread of the RSGs is well-matched by the 12Myr isochrone. Errors on
758: the data-points do not include the uncertainty in cluster distance
759: $\Delta D_{\rm cl}$ -- the magnitude of this error is indicated at the
760: right of the plot. The data-point separated from the rest is the YHG
761: F15. }
762: \label{fig:hrd}
763: \end{figure}
764:
765: In Fig.\ \ref{fig:hrd}, we plot the derived temperatures and
766: luminosities of the cool, luminous stars on a H-R diagram. Also
767: plotted are isochrones taken from the stellar evolutionary models of
768: \citet{Mey-Mae00} which include the effects of rotation, and have
769: initial rotational velocity set to 300\kms. As we can confidently make
770: the approximation that all the stars are at the same distance, we do
771: not include the error in distance on each data-point. The magnitude of
772: the error in $L_{\star}$ when this uncertainty {\it is} included is
773: shown on the right of the plot.
774:
775: The figure shows that the temperatures and luminosity spread of the
776: RSGs are well-matched by the 12Myr isochrone. The 10Myr isochrone
777: cannot reproduce the low-luminosity stars, while the 14Myr isochrone
778: is too faint to fit the high-luminosity stars. When the uncertainty in
779: cluster distance is taken into account ($\Delta D_{\rm cl}$ in Fig.\
780: \ref{fig:hrd}), the 10Myr and 14Myr isochrones appear more
781: reasonable. We experimented with different evolutionary models, namely
782: non-rotating models with varying mass-loss rates and metallicity
783: \citep{Schaller92,Schaerer93,Meynet94}. We found generally that the
784: non-rotating models gave ages that were $\la$2\,Myr younger. We settle
785: on an age estimate for RSGC1 of 12$\pm$2Myr.
786:
787: We note that the YHG F15 does {\it not} lie on the 12Myr
788: isochrone in Fig.\ \ref{fig:hrd}. Fitting the star with the rotating
789: Geneva isochrones, we get an age of 10Myr (pre blue-loop) or 8Myr
790: (post blue-loop). For the specific models used in this analysis, the
791: masses of RSGs in a 12Myr-old cluster do not experience a
792: blue-loop. However, the specific masses of stars which experience
793: blue-loops are extremely sensitive to the input physics, such as
794: rotation \citep[see Fig.~1 of][]{Hirschi04}. Hence, a blue-loop may be
795: introduced for stellar initial masses relevant to RSGC1 simply by
796: changing the rotational speed. Additionally, it is likely that
797: blue-loops would be affected by the inclusion of extra physics (e.g.\
798: magnetic fields). In summary, We do not necessarily interpret F15's
799: location in the H-R diagram as evidence of cluster non-coevality, it
800: could simply be that the input physics of the evolutionary models used
801: in the analysis are not fine-tuned to this cluster.
802:
803: If the RSGs of this cluster do experience some form of blue-loop, then
804: the position of F15 on the H-R diagram is consistent with it evolving
805: {\it away} from the RSGs. A post-RSG nature for this star would make
806: it a member of a very exclusive club -- arguably IRC +10420 is the
807: only object which is widely accepted to be a post-RSG, though a case
808: has also been argued for HD~179821 \citep[see review
809: of][]{Oudmaijer08}. We note that F15 does not exhibit the same
810: considerable IR excess, nor the bright maser emission of IRC
811: +10420. This may be due to F15's lower initial mass -- the Geneva
812: models imply $\sim$18\msun\ for F15, while the larger luminosity of IRC
813: +10420 makes it consistent with a star of initial mass $\sim$40\msun\
814: (see also Sect.\ \ref{sec:masers}).
815:
816: The cluster age we derive is slightly greater than the age of
817: $\la$9Myr derived in FMR06. This previous estimate was determined by
818: comparing the luminosity spread of the RSGs to that predicted by the
819: non-rotating isochones of \citet{Schaller92} as a function of age. The
820: more rigourous investigation of the stars' luminosities in this present
821: paper results in a larger luminosity spread for the RSGs, while the
822: non-rotating models do not reproduce the higher luminosities. Much
823: better agreement is found between the contemporary rotating models and
824: the new luminosity estimates.
825:
826: Finally, we remark that the observed temperatures of the stars
827: are systematically cooler than the isochrones (see also Fig.\
828: \ref{fig:masers}). This could be reconciled by increasing either the
829: relative metal abundances or the stellar rotational velocities. Both
830: lead to slightly increased stellar radii, the former due to the
831: increased opacity of the envelope, the latter due to the lower
832: effective gravities. A super-solar metallicity would certainly be
833: consistent with the Galactic metallicity gradient and the cluster's
834: Galacto-centric distance ($\sim$3\,kpc). However, given the well-known
835: disparities between observations and theory in the field of RSGs, we
836: attach a cautionary note to any conclusions derived from this
837: evidence. While recent progress has been made in uniting theory and
838: observation at solar metallicity \citep{Levesque05}, discrepancies
839: still exist at sub-solar metallicities \citep{Levesque06}. The
840: location in the Galaxy of the Scutum clusters would seem to make
841: non-solar metallicities likely. Accurate abundance measurements of the
842: clusters would make them ideal testbeds for evolutionary models, as
843: well as probes of the Galactic metallicity gradient.
844:
845:
846: \subsection{Cluster mass} \label{sec:mass}
847: To determine the cluster mass we employ the same Monte-Carlo technique
848: used in FMR06 and DFK07. We generate a synthetic cluster of a
849: pre-defined initial mass, containing stars whose masses are randomly
850: drawn from a distribution consistent with a Salpeter initial mass
851: function \citep{Salpeter55}. Then, for a given cluster age, we
852: determine the present-day masses, temperatures and luminosities from
853: the Geneva isochrones used in Sect.\ \ref{sec:age}. We then count the
854: number of RSGs in the cluster, where we define a RSG as a star whose
855: temperature is lower than 4000K and luminosity greater than
856: $10^{4}$\lsun. As this is a random process, we repeat each simulation
857: 1000 times to reduce statistical noise.
858:
859: \begin{figure}[t]
860: \centering
861: % \includegraphics[width=12cm,bb=20 10 570 450]{N_age_model_rot.eps}
862: \includegraphics[width=12cm,bb=20 10 570 450]{fig8.eps}
863: \caption{The number of RSGs as a function of age for a coeval
864: cluster of a given initial mass, calculated using the rotating
865: Geneva isochrones \citep{Mey-Mae00}. The age-range for RSGC1,
866: determined from isochrone fitting, is indicated on the plot, as is
867: the number of RSGs observed in the cluster.}
868: \label{fig:mass_evo}
869: \end{figure}
870:
871: In Fig.\ \ref{fig:mass_evo}, we plot the number of RSGs contained in a
872: synthetic cluster as a function of cluster age, for two initial
873: cluster masses: 20,000\msun\ and 40,000\msun. The plot shows that
874: significant numbers of RSGs begin to be seen after $\sim$7Myr -- in
875: clusters younger than this the post-MS stars are massive enough to
876: evolve directly to the WN phase, skipping the RSG stage. Peaks in the
877: number of RSGs are reached at $\sim$14Myr and $\sim$20Myr, the dip in
878: between is caused by the onset of a blue-loop in the stars' evolution
879: for a narrow range of initial masses. The mean luminosity of the RSGs
880: as a function of time decreases, as the initial masses of the stars in
881: the RSG zone becomes smaller (see Fig.\ \ref{fig:hrd}, and discussion
882: in Sect.\ \ref{sec:clusters}). For a cluster containing 14 RSGs with
883: an inferred age of 12$\pm$2\,Myr (see above), we find that the initial
884: mass of the cluster must be somewhere in between these two, implying
885: an initial mass of RSGC1 of $(3 \pm 1) \times 10^{4}$\msun.
886:
887: Using the velocity dispersion of the RSGs, we can compare this value
888: of the cluster's initial mass to the cluster's {\it dynamical} mass,
889: under the assumption that the cluster is currently in virial
890: equilibrium. The dynamical mass $M_{\rm dyn}$ is derived using the
891: relation,
892:
893: \begin{equation}
894: M_{\rm dyn} = \frac{\eta \sigma^{2}_{v} r_{\rm hp}}{G}
895: \label{equ:dynmass}
896: \end{equation}
897:
898: \noindent where $r_{\rm hp}$ is the half-light radius,
899: $\sigma^{2}_{v}$ is the velocity dispersion, $G$ is the gravitational
900: constant, and $\eta$ is a constant which depends on the stellar
901: distribution with radius, and is typically taken to be $\sim$10
902: \citep[see review in Introduction of][]{Mengel02}.
903:
904: For the velocity dispersion, we find $\sigma^{2}_{v} = 3.7$\kms\ after
905: the internal uncertainty in the wavelength solution ($\pm$1\kms) has
906: been subtracted in quadrature. To find the half-light radius, we plot
907: the cumulative brightness profile of the cluster using stars F01--15
908: as tracers of the cluster's spatial distribution (see Fig.\
909: \ref{fig:halflight}). We used a cluster centre of
910: 18$^{h}$37$^{m}$57.4$^{s}$, -6\degr52\arcmin58.11\arcsec\ (J2000), the
911: approximate mid-point of the cool stars. We experimented with moving
912: the cluster centre by up to 0.5\arcmin, using different photometric
913: bands, and using the luminosities derived in Sect.\ \ref{sec:lum}
914: rather than the raw photometry. From all these methods we found that
915: the cluster half-light radius was stable at $0.8 \pm 0.1$\arcsec. At
916: the distance derived in Sect.\ \ref{sec:dist}, this gives a cluster
917: size $r_{\rm hp} = 1.5 \pm 0.3$pc.
918:
919: Using these values, we find the dynamical mass of RSGC1 to be {$M_{\rm
920: dyn} = (5 \pm 1) ~\eta/10 \times 10^{4}$\msun.} Due to the extra
921: uncertainty in the density parameter $\eta$, we consider this to be an
922: order-of-magnitude estimate only, which compares well to the initial
923: cluster mass derived using evolutionary models.
924:
925:
926: \begin{figure}[t]
927: \centering
928: % \includegraphics[width=12cm,bb=50 15 700 530]{halflight_rsgs.eps}
929: \includegraphics[width=12cm,bb=50 15 700 530]{fig9.eps}
930: \caption{Cumulative apparent magnitude distribution of the cluster,
931: centred on 18$^{h}$37$^{m}$57.4$^{s}$, -6\degr52\arcmin58.11\arcsec\
932: (J2000), using the 15 cool luminous stars as tracers of the
933: cluster's spatial distribution. }
934: \label{fig:halflight}
935: \end{figure}
936:
937:
938: \begin{table}[t]
939: \begin{center}
940: \caption{Properties of the two Scutum RSG clusters. (1): cluster
941: name; (2): total initial cluster mass derived from evolutionary
942: models; (3): cluster mass derived from stellar velocity
943: dispersion; (4): cluster age; (5): cluster diameter; (6): cluster
944: distance; (7): distance from cluster to the Galactic centre; (8):
945: initial mass range of the RSGs within the clusters. RSGC2 values
946: come from DFK07.}
947: \begin{tabular}{l|cc|c|c|cc|c}
948: \hline
949: (1) & (2) & (3) & (4) & (5) & (6) & (7) & (8) \\
950: Cluster & $M_{\rm init}$ (evol) & $M_{\rm dyn}$ & Age & $d$ & $D_{\odot}$
951: & $D_{\rm GC}$ & $M_{\rm init} (RSGs) $\\
952: & \multicolumn{2}{c|}{($\times10^{4}$\msun)} & (Myr) & (pc) &
953: \multicolumn{2}{c|}{(kpc)} & (\msun) \\
954: \hline \hline
955: RSGC1 & 3$\pm$1 & 5$\pm$1 & 12$\pm$2 & 1.5$\pm$0.3 & 6.60$\pm$0.89 & 3.2 & 18$^{+4}_{-2}$ \\
956: RSGC2 & 4$\pm$1 & 6$\pm$4 & 17$\pm$3 & 3.2$^{+1.2}_{-0.7}$ & 5.83$^{+1.91}_{-0.76}$ & 3.5 & 14$\pm$2 \\
957: \hline
958: \end{tabular}
959: \label{tab:clusters}
960: \end{center}
961: \end{table}
962:
963:
964: \subsection{Comparison of the two RSG clusters} \label{sec:clusters}
965: Given that, until recently, the largest number of RSGs in any one
966: cluster was 5, the discovery of the two Scutum RSG clusters lying so
967: close to one another is remarkable. After applying the same analysis
968: techniques to each cluster, we summarize the physical parameters of
969: RSGC1 and RSGC2 in Table \ref{tab:clusters}.
970:
971: The radial velocities of each cluster put them at the tangential point
972: of the Galactic rotation curve, with Galacto-centric distances of
973: 3-4kpc. The clusters are separated from one another by
974: $0.8^{+1.6}_{-0.7}$\,kpc. Their proximity, combined with their similar
975: ages, suggests that they were both formed in a region-wide starburst
976: phase some 10-20Myr ago, and that the chemical abundances of their
977: natal material should be similar.
978:
979: First-order evidence of uniform metallicity between the clusters comes
980: from the median spectral-types of the RSGs, which is M3 in each
981: cluster. The average spectral-type of RSGs has been shown to be
982: dependent on environment, shifting gradually to later types with
983: increasing metallicity -- averages of K5, M1, and M2 were found for
984: the SMC, LMC and Galaxy respectively \citep{Elias85,M-O03}. Proposed
985: physical explanations for this include (a) reduced metallicity leading
986: to lower envelope opacities, increased stellar radii and hence to
987: systematically lower effective temperatures; or (b) the effect of
988: lower metal abundances on the strengths of the diagnostic TiO
989: lines. Regardless, the average spectral-type of M3 for the two Scutum
990: clusters would seem to suggest that they have similar, possibly
991: super-solar metallicities. The chemical abundances of the stars in
992: these clusters will be the subject of a future paper.
993:
994: Both from the analysis of evolutionary models, and the velocity
995: dispersion, we find similar masses for each cluster -- the factor of
996: two difference in the number of RSGs is caused by the difference in
997: cluster ages. This may also explain the slightly larger size of RSGC2:
998: it has been suggested by \citet{B-G06} that a cluster of this initial
999: mass and age may be {\it out} of virial equilibrium. This arises when
1000: the left-over natal material is expelled from the cluster by the first
1001: SNe explosions of the cluster's most massive stars. This leaves the
1002: remaining stars with a velocity dispersion larger than that of a
1003: virialized cluster, resulting in cluster expansion. This may explain
1004: (i) the slightly different sizes of the clusters, (ii) the apparent
1005: lack of any obvious diffuse nebular emission in the GLIMPSE
1006: images. Also, we note that the dynamical masses of each cluster are
1007: slightly higher than the `evolutionary' masses, though the errors on
1008: the dynamical masses, in particular the density parameter $\eta$, make
1009: it difficult to draw any firm conclusions from this.
1010:
1011: The difference in ages between the two clusters implies that the
1012: initial masses of the RSGs in each must be different; in the younger
1013: RSGC1 we are seeing stars with larger initial masses than in the
1014: slightly older RSGC2. To investigate the initial masses of the stars
1015: in each cluster, we again use the evolutionary models of
1016: \citet{Mey-Mae00}. For a given isochrone, we find the minimum and
1017: maximum initial masses and luminosities of all stars with effective
1018: temperatures cooler than 4000K, i.e.\ the temperature range of
1019: the RSGs. In Fig.\ \ref{fig:minmaxmass} we plot the luminosity and
1020: mass ranges of cool stars as a function of time. From the luminosity
1021: ranges of the stars in each cluster, we then determine the initial
1022: masses of the RSGs. For RSGC1 we find $M_{\star, \rm init} =
1023: 18^{+4}_{-2}$\msun, and for RSGC2 $M_{\star, \rm init} = 14 \pm
1024: 2$\msun. This underlines the potential importance of the RSG clusters
1025: to the study of stellar evolution: {\it they allow us to study large
1026: numbers of RSGs at uniform metallicity as a function of initial mass.}
1027:
1028:
1029: \begin{figure}[t]
1030: \centering
1031: % \includegraphics[width=12cm,bb=20 10 700 600]{minmaxmass.eps}
1032: \includegraphics[width=12cm,bb=20 20 542 538]{fig10.eps}
1033: \caption{The minimum and maximum luminosities (top panel) and
1034: initial masses (bottom panel) of stars with effective temperatures
1035: cooler than 4000K in a coeval cluster. Calculated using the rotating
1036: Geneva isochrones of \citet{Mey-Mae00}. }
1037: \label{fig:minmaxmass}
1038: \end{figure}
1039:
1040: \subsection{Masers in RSGs} \label{sec:masers}
1041: One aspect of RSG evolution which the two Scutum RSG clusters allow us
1042: to study is the onset of the maser-active phase. Masers are often
1043: observed in `extreme' cool stars, e.g.\ the RSGs VY~CMa,
1044: NML~Cyg, S~Per \citep[see][ and refs therein]{R-Y98}, which are also
1045: synonymous with large IR excesses. In the standard picture, these
1046: phenomena are caused by episodes of high mass-loss, producing large
1047: amounts of circumstellar material which gives rise to the IR
1048: excess. The masers themselves originate in the outflowing
1049: material. The SiO 43GHz maser is formed low in the wind at higher
1050: temperatures; at larger radii the formation of dust grains leads to a
1051: depletion of SiO. In the SiO maser-forming region the outflow velocity
1052: is low, and hence SiO masers typically have radial velocities similar
1053: to the stellar systemic velocity, $v_{\rm sys}$
1054: \citep{Jewell91}. Higher in the wind the H$_2$O 22GHz maser forms, and
1055: typically has peaks at many velocities between $v_{\rm sys} \pm
1056: v_{\infty}$, where $v_{\infty}$ is the wind's terminal velocity
1057: (typically around 10-30\kms). At larger radii still, H$_2$O is
1058: photo-disassociated to OH, giving rise to the 1612MHz OH maser. Here
1059: the outflow has reached its terminal velocity, giving the line-profile
1060: its typical double-peaked morphology (where the separation of the
1061: peaks is 2$v_{\infty}$ and the centroid is $v_{\rm sys}$). For a more
1062: comprehensive review of masers in luminous cool stars, see
1063: \citet{Habing96}.
1064:
1065: \begin{figure}[t]
1066: \centering
1067: % \includegraphics[width=12cm,bb=10 10 520 500,clip]{hrd_masers.eps}
1068: \includegraphics[width=12cm,bb=10 10 520 500,clip]{fig11.eps}
1069: \caption{H-R diagram of the RSGs in RSGC1, showing those stars which
1070: are maser sources. A 12Myr model isochone is overplotted, from the
1071: rotating Geneva models of \citet{Mey-Mae00}.}
1072: \label{fig:masers}
1073: \end{figure}
1074:
1075: It is unclear as to whether the maser stage is one which all RSGs will
1076: go through, or whether only extreme objects will pass through this
1077: phase. Much work has been done on the detailed physical conditions
1078: under which masers form \citep[see review of][]{Habing96}, however in
1079: a simplified picture one may say that the presence of the different
1080: masers is determined by the wind density, i.e. by mass-loss rate and
1081: by pulsations. For an increasing mass-loss rate, gradually the
1082: critical density will be reached in the formation zones of each
1083: transition; while stellar pulsations will create density-contrasts in
1084: the outflow, conducive to population inversions. That is to say that
1085: we expect to see maser emission from those RSGs which have the
1086: strongest mass-loss rates and which are pulsationally unstable.
1087:
1088: Mass-loss from RSGs is driven by radiation pressure on dust grains in
1089: the outer atmosphere, and hence depends upon the star's effective
1090: temperature (cool enough to allow dust to form) and luminosity (high
1091: radiation pressure) \citep{v-L05}. Hence we may expect to find masers
1092: in high-mass RSGs which have the higher luminosities, while among a
1093: coeval sample of RSGs one may expect to see masers in those objects
1094: furthest along their evolution, where their path on the HR diagram
1095: takes a sharp upturn (see Fig.\ \ref{fig:hrd}). Further, as
1096: masers are linked to stellar pulsations, we may expect to find masers
1097: in unstable stars -- i.e.\ stars with high $L_{\star}/M_{\star}$
1098: ratios that are evolving closer to the Humphreys-Davidson limit at
1099: $L_{\star} \sim 10^{5.7}$\lsun\ \citep{H-D79}. This zone of the H-R
1100: diagram is often linked to the so-called `modified' Eddington limit,
1101: when the contributions of atomic/molecular transitions to the
1102: continuum opacity are included when calculating the Eddington
1103: luminosity.
1104:
1105: To investigate the presence of masers in RSGs, in Fig.\
1106: \ref{fig:masers} we plot a HR diagram of the RSGs in RSGC1, and
1107: indicate those stars which are maser sources. In addition to the OH
1108: maser observations presented here, SiO and H$_2$O maser observations
1109: of RSGC1 were also taken by \citet{N-D06}. They found that stars F01,
1110: F02, F04 and F13 were spatially coincident with SiO emission, while
1111: they concluded that the H$_{2}$O maser emission they found was likely
1112: to come from F13\footnote{The large beamsize of the 22GHz H$_{2}$O
1113: maser, 73\arcsec, overlaps several other stars in the
1114: cluster. However, as these masers are often observed in stars with
1115: maser emission from SiO, and as the central velocity of the H$_{2}$O
1116: maser was consistent with that from the SiO maser emission of F13,
1117: \citet{N-D06} concluded that F13 was likely to be the origin of the
1118: H$_{2}$O maser emission. }.
1119:
1120: The figure shows that it is the most luminous stars of the cluster
1121: which exhibit maser emission. Further, F13, which is the source of
1122: SiO, H$_{2}$O and OH masers, appears to be at the point of evolving
1123: back toward the blue. The figure therefore appears to support the
1124: hypothesis that maser emission activates (or becomes strong enough to
1125: be observed) in the latest stages of RSG evolution, when the star's
1126: mass-loss rate is highest and the star becomes unstable to
1127: pulsations.
1128:
1129: This hypothesis could be tested with a study of the second Scutum RSG
1130: cluster, RSGC2. This cluster also has many RSGs and so we are again
1131: seeing stars in both the earlier and later RSG stages. As mentioned in
1132: Sect.\ \ref{sec:clusters}, the stars in RSGC2 are less luminous and
1133: hence further from the H-D limit. Though they likely have lower
1134: initial masses, they have lower $L_{\star}/M_{\star}$ ratios. Thus, we
1135: may not expect to see maser emission from these stars, especially
1136: those with lower $L_{\star}$. A comprehensive maser study of the RSGC2
1137: region would be able to test at which phase of RSG evolution stars
1138: become maser-active, while non-detections would place lower-limits to
1139: the initial-mass requirements to pass through the maser-active phase.
1140:
1141: Finally, we note that the YHG F15 is not observed to have maser
1142: emission, unlike the prototype post-RSG IRC +10420. Masers are rarely
1143: observed around stars hotter than $\sim$4000, and the presence of
1144: maser emission in IRC +10420's outflow is commonly accepted as
1145: evidence of its rapid evolution away from the RSG phase. That no maser
1146: is observed in F15 may be indicative of a lower wind-density while in
1147: the RSG phase, due to its lower initial mass ($\sim$18\msun\ compared
1148: to $\sim$40\msun\ for IRC +10420).
1149:
1150:
1151: %_____________________________________________________________________
1152: % ********************************************************************
1153:
1154: \section{Conclusions}
1155: We have presented a comprehensive investigation into the physical
1156: properties of the luminous cool stars in the massive cluster
1157: RSGC1. Using high-resolution spectroscopy we have accurately measured
1158: the cluster's radial velocity, derived its kinematic distance, and
1159: reappraised the stars' temperatures and luminosities. We find a larger
1160: luminosity spread than in the discovery paper, which is well fitted by
1161: a cluster age of 12Myr and cluster mass of $(3 \pm 1) \times
1162: 10^4$\msun. The mass is similar to that of the nearby cluster RSGC2,
1163: and we suggest that the difference in the number of RSGs in each is
1164: due to the separation in cluster ages, with RSGC2 being somewhat
1165: older. This implies that the initial masses of the RSGs in each
1166: cluster are different, which we determine to be $\sim$18\msun\ for
1167: RSGC1 and $\sim$14\msun\ for RSGC2. The clusters therefore allow the
1168: study of RSG evolution as a function of initial mass, while
1169: constraining the variable of metallicity. Finally, with new 1612MHz
1170: radio observations we find compelling evidence that the OH maser is
1171: associated with star F13, and collating recent maser observations of
1172: the cluster we argue that the maser-active phase is associated with
1173: stars in the latter stages of RSG evolution.
1174:
1175:
1176:
1177: \acknowledgments Acknowledgments: We would like to thank Maria
1178: Messineo for useful discussions concerning masers in cool stars, and
1179: the anonymous referee for comments and suggestions which improved the
1180: paper. We thank Mark Claussen for discussions regarding the
1181: astrometric accuracy of the VLA observations. The material in this
1182: work is supported by NASA under award NNG~05-GC37G, through the
1183: Long-Term Space Astrophysics program. This research was performed in
1184: the Rochester Imaging Detector Laboratory with support from a NYSTAR
1185: Faculty Development Program grant. The National Radio Astronomy
1186: Observatory is a facility of the National Science Foundation operated
1187: under cooperative agreement by Associated Universities, Inc. Part of
1188: the data presented here were obtained at the W.\ M.\ Keck Observatory,
1189: which is operated as a scientific partnership among the California
1190: Institute of Technology, the University of California, and the
1191: National Aeronautics and Space Administration. The Observatory was
1192: made possible by the generous financial support of the W.\ M.\ Keck
1193: Foundation. This publication makes use of data products from the Two
1194: Micron All Sky Survey, which is a joint project of the University of
1195: Massachusetts and the Infrared Processing and Analysis
1196: Center/California Institute of Technology, funded by the National
1197: Aeronautics and Space Administration and the NSF. This research has
1198: made use of MSX and Spitzer's GLIMPSE survey data, the {\sc simbad}
1199: database, Aladin \& IDL software packages, and the GSFC IDL library.
1200:
1201: \clearpage
1202: \bibliographystyle{/fat/Data/model_paper/aa}
1203: %\bibliography{/fat/Data/model_paper/biblio}
1204: \bibliography{/fat/Data/bibtex/biblio}
1205:
1206:
1207: \end{document}
1208:
1209:
1210:
1211:
1212: \begin{figure}
1213: \centering
1214: \includegraphics[width=16cm]{sp_hist.eps}
1215: \caption{Histogram of spectral types for the RSGC1 stars.}
1216: \end{figure}
1217:
1218: \begin{figure}
1219: \centering
1220: \includegraphics[width=16cm]{v_ew.eps}
1221: \caption{Plot of CO equivalent width versus radial velocity for the
1222: stars listed in Table XXX. Stars in the cross-hatched region have
1223: similar velocities and are likely to be part of the same
1224: cluster. Stars in the top left are probably foreground stars due to
1225: their low \EW\ and much lower velocities. }
1226: \label{fig:v_ew}
1227: \end{figure}
1228:
1229:
1230: \begin{figure}
1231: \centering
1232: \includegraphics[width=16cm]{sed_1.eps}
1233: \caption{Spectral energy distributions of cluster members. Crosses
1234: are raw photometry from 2MASS, GLIMPSE \& MSX; filled circles are
1235: dereddened according to the stellar extictions derived in BLAH. Over
1236: plotted in green is a black-body curve, with $T$ determined from
1237: spectral-type of star and luminosity according to dereddened K-band
1238: magnitude at a distance of 5.88kpc. }
1239: \end{figure}
1240: \addtocounter{figure}{-1}
1241:
1242: \begin{figure}
1243: \centering
1244: \includegraphics[width=16cm]{sed_2.eps}
1245: \caption{ Cont. }
1246: \end{figure}
1247: \addtocounter{figure}{-1}
1248:
1249:
1250: % \begin{landscape}
1251: \begin{table}
1252: % \centering
1253: \begin{tiny}
1254: \begin{tabular}{lcclccclcccc}
1255: \hline \hline
1256: ID\# & RA & DEC & & J & H & K$_{\rm S}$ & & \multicolumn{3}{c}{W$_{\lambda}$ (\AA)} & Type \\
1257: & \multicolumn{2}{c}{(J2000)} & & & & & & Br$\gamma$ & He {\sc i} & CO & \\
1258: \hline
1259: 1 & 18 37 48.50 & -6 54 57.1 & & 13.767 & 13.326 & 12.912 & & 1.55$\pm$0.37 & 0.46$\pm$0.41 & 6.32$\pm$0.64 & K/M \\
1260: 2 & 18 37 48.83 & -6 54 38.1 & & 14.142 & 11.733 & 10.694 & & -0.27$\pm$0.19 & 0.55$\pm$0.06 & 21.77$\pm$0.42 & K/M \\
1261: 3 & 18 37 49.02 & -6 54 30.3 & & 15.256 & 11.366 & 9.517 & & -1.18$\pm$0.24 & 0.38$\pm$0.32 & 33.37$\pm$0.47 & K/M \\
1262: 4 & 18 37 49.44 & -6 54 17.9 & & 14.678 & 12.336 & 11.291 & & 0.67$\pm$0.21 & 0.67$\pm$0.34 & 15.56$\pm$0.65 & K/M \\
1263: 5 & 18 37 49.65 & -6 54 20.9 &$^{1}$ & & & &$^{2}$ & 0.06$\pm$0.20 & 1.26$\pm$0.35 & 17.77$\pm$1.75 & K/M \\
1264: 6 & 18 37 49.80 & -6 54 22.6 & & 14.736 & 13.760 & 13.505 &$^{2}$ & -1.12$\pm$0.39 & -0.55$\pm$0.29 & 16.21$\pm$1.61 & K/M \\
1265: 7 & 18 37 49.95 & -6 54 06.7 & & 15.767 & 13.545 & 12.577 & & 0.19$\pm$0.26 & 0.40$\pm$0.25 & 13.51$\pm$0.51 & K/M \\
1266: 8 & 18 37 50.30 & -6 53 54.7 & & 14.618 & 12.006 & 10.753 & & -0.10$\pm$0.11 & 0.61$\pm$0.30 & 23.24$\pm$0.71 & K/M \\
1267: 9 & 18 37 50.50 & -6 54 23.9 & & 14.425 & 11.323 & 9.855 & & -0.94$\pm$0.31 & 0.46$\pm$0.49 & 33.74$\pm$0.40 & K/M \\
1268: 10 & 18 37 50.90 & -6 53 38.2 & & 9.658 & 6.803 & 5.342 & & -0.53$\pm$0.44 & -0.88$\pm$0.63 & 52.29$\pm$3.55 & K/M \\
1269: 11 & 18 37 50.98 & -6 53 22.4 &$^{1}$ & & & &$^{2}$ & -0.76$\pm$0.16 & 0.55$\pm$0.22 & 24.51$\pm$0.63 & K/M \\
1270: 12 & 18 37 51.12 & -6 54 20.7 & & 14.234 & 11.236 & 10.042 & & -0.49$\pm$0.18 & 0.06$\pm$0.10 & 12.71$\pm$2.87 & K/M \\
1271: 13 & 18 37 51.18 & -6 53 31.8 & & 12.324 & 11.071 & 9.700 & & 0.37$\pm$0.33 & 0.55$\pm$0.22 & 13.98$\pm$0.62 & K/M \\
1272: 14 & 18 37 52.82 & -6 54 48.0 & & 16.741 & 13.786 & 12.407 & & -0.79$\pm$0.41 & 0.80$\pm$0.43 & 23.68$\pm$1.55 & K/M \\
1273: 15 & 18 37 53.13 & -6 53 15.6 & & 15.937 & 13.676 & 12.576 & & 1.20$\pm$0.09 & -0.15$\pm$0.19 & 12.18$\pm$0.71 & K/M \\
1274: 16 & 18 37 53.82 & -6 54 33.3 & & 14.973 & 11.404 & 9.706 & & -0.74$\pm$0.26 & 0.90$\pm$0.40 & 32.23$\pm$0.66 & K/M \\
1275: 17 & 18 37 53.90 & -6 53 37.0 & & 13.237 & 10.399 & 9.051 & & -1.11$\pm$0.29 & 0.69$\pm$0.18 & 29.04$\pm$0.83 & K/M \\
1276: 18 & 18 37 54.03 & -6 51 37.2 & & 15.850 & 12.628 & 11.262 & & 0.04$\pm$0.16 & 0.65$\pm$0.31 & 25.80$\pm$0.48 & K/M \\
1277: 19 & 18 37 54.55 & -6 51 30.8 & & 15.155 & 11.281 & 9.339 & & -1.47$\pm$0.31 & 0.38$\pm$0.35 & 34.62$\pm$0.30 & K/M \\
1278: 20 & 18 37 54.62 & -6 53 42.1 & & 10.726 & 9.224 & 8.525 & & -1.19$\pm$0.25 & 0.73$\pm$0.16 & 20.79$\pm$0.80 & K/M \\
1279: 21 & 18 37 54.64 & -6 53 14.2 & & 15.933 & 13.872 & 12.330 & & 4.39$\pm$0.08 & 0.47$\pm$0.16 & -0.07$\pm$0.19 & A \\
1280: 22 & 18 37 54.75 & -6 52 51.8 &$^{1}$ & & & &$^{2}$ & 0.49$\pm$0.21 & 0.95$\pm$0.40 & 18.20$\pm$0.73 & K/M \\
1281: 23 & 18 37 54.84 & -6 53 33.1 &$^{1}$ & & & &$^{2}$ & 2.27$\pm$0.05 & 0.94$\pm$0.24 & 16.85$\pm$0.72 & K/M \\
1282: 24 & 18 37 54.85 & -6 53 36.1 &$^{1}$ & & & &$^{2}$ & 0.45$\pm$0.17 & 0.40$\pm$0.44 & 20.35$\pm$0.65 & K/M \\
1283: 25 & 18 37 54.90 & -6 53 38.9 &$^{1}$ & & & &$^{2}$ & 0.61$\pm$0.14 & 0.86$\pm$0.09 & 8.67$\pm$1.14 & K/M \\
1284: 26 & 18 37 54.97 & -6 53 13.7 & & 14.364 & 12.317 & 11.257 & & 3.00$\pm$0.04 & 0.75$\pm$0.03 & 0.16$\pm$0.13 & B \\
1285: 27 & 18 37 55.00 & -6 52 53.1 &$^{1}$ & & & &$^{2}$ & 1.44$\pm$0.18 & 0.61$\pm$0.11 & 14.73$\pm$0.59 & K/M \\
1286: 28 & 18 37 55.09 & -6 52 56.5 &$^{1}$ & & & &$^{2}$ & 0.69$\pm$0.26 & 1.17$\pm$0.30 & 16.04$\pm$0.54 & K/M \\
1287: 29 & 18 37 55.21 & -6 52 55.1 &$^{1}$ & & & &$^{2}$ & 2.32$\pm$0.07 & 0.80$\pm$0.02 & 2.68$\pm$0.12 & K/M \\
1288: 30 & 18 37 55.41 & -6 53 13.5 & & 13.553 & 11.266 & 10.203 & & 2.01$\pm$0.23 & 2.10$\pm$0.41 & 23.17$\pm$2.10 & K/M \\
1289: 31 & 18 37 55.46 & -6 54 44.3 & & 8.873 & 7.485 & 6.934 & & -1.01$\pm$0.14 & 0.58$\pm$0.21 & 27.65$\pm$0.72 & K/M \\
1290: 32 & 18 37 55.56 & -6 52 56.8 &$^{1}$ & & & &$^{2}$ & 2.41$\pm$0.03 & 0.81$\pm$0.02 & -0.73$\pm$0.24 & B \\
1291: 33 & 18 37 55.76 & -6 52 59.7 & & 13.763 & 11.761 & 10.687 & & 2.81$\pm$0.06 & 1.53$\pm$0.03 & -0.24$\pm$0.26 & B \\
1292: 34 & 18 37 55.94 & -6 53 33.1 &$^{1}$ & & & &$^{2}$ & -0.29$\pm$0.16 & 0.53$\pm$0.22 & 21.89$\pm$0.62 & K/M \\
1293: 35 & 18 37 56.05 & -6 54 36.6 & & 13.052 & 11.604 & 11.017 & & 0.50$\pm$0.08 & 0.39$\pm$0.15 & 15.47$\pm$0.38 & K/M \\
1294: 36 & 18 37 56.11 & -6 53 23.2 & & 15.108 & 12.523 & 11.348 & & 0.67$\pm$0.26 & 0.37$\pm$0.27 & 19.87$\pm$0.53 & K/M \\
1295: 37 & 18 37 56.42 & -6 53 39.3 & & 14.272 & 11.373 & 10.169 & & 0.22$\pm$0.14 & 0.64$\pm$0.33 & 21.78$\pm$0.56 & K/M \\
1296: 38 & 18 37 56.48 & -6 52 23.8 &$^{1}$ & & & &$^{2}$ & -0.85$\pm$0.22 & 0.84$\pm$0.40 & 27.16$\pm$0.47 & K/M \\
1297: 39 & 18 37 56.48 & -6 52 30.8 &$^{1}$ & & & &$^{2}$ & -1.24$\pm$0.42 & 0.26$\pm$0.54 & 57.09$\pm$1.65 & K/M \\
1298: 40 & 18 37 56.63 & -6 52 32.8 &$^{1}$ & & & &$^{2}$ & 2.91$\pm$0.21 & 0.24$\pm$0.10 & 5.76$\pm$0.75 & A \\
1299: 41 & 18 37 56.96 & -6 53 04.0 &$^{1}$ & 13.805 & 11.195 & 11.016 &$^{2}$ & 2.56$\pm$0.03 & 0.80$\pm$0.04 & -1.80$\pm$0.29 & B \\
1300: 42 & 18 37 57.04 & -6 53 07.5 & & 13.103 & 10.932 & 10.295 & & 3.82$\pm$0.11 & 2.29$\pm$0.06 & 1.50$\pm$0.64 & B \\
1301: %\input{papertab}
1302: \hline
1303: \end{tabular}
1304: \end{tiny}
1305: \end{table}
1306: \addtocounter{table}{-1}
1307:
1308: \begin{table}
1309: % \centering
1310: \begin{tiny}
1311: \begin{tabular}{lcclccclcccc}
1312: \hline \hline
1313: ID\# & RA & DEC & & J & H & K$_{\rm S}$ & & \multicolumn{3}{c}{W$_{\lambda}$ (\AA)} & Type \\
1314: & \multicolumn{2}{c}{(J2000)} & & & & & & Br$\gamma$ & He {\sc i} & CO & \\
1315: \hline
1316: 43 & 18 37 57.26 & -6 53 15.9 &$^{1}$ & 14.827 & 12.189 & 10.687 &$^{2}$ & 3.44$\pm$0.11 & 0.94$\pm$0.03 & 2.32$\pm$0.53 & B \\
1317: 44 & 18 37 57.33 & -6 53 57.7 & & 15.033 & 11.176 & 9.138 & & -0.88$\pm$0.23 & 0.27$\pm$0.34 & 33.06$\pm$0.39 & K/M \\
1318: 45 & 18 37 57.35 & -6 53 09.8 & & 13.088 & 10.917 & 9.814 & & 1.65$\pm$0.07 & 0.96$\pm$0.03 & 1.40$\pm$0.21 & B \\
1319: 46 & 18 37 57.40 & -6 53 36.8 & & 13.752 & 12.045 & 11.035 & & -0.24$\pm$0.55 & 0.67$\pm$0.60 & 25.82$\pm$0.91 & K/M \\
1320: 47 & 18 37 57.50 & -6 53 25.5 & & 9.866 & 7.038 & 5.613 & & -1.50$\pm$0.65 & 0.83$\pm$0.65 & 44.47$\pm$1.46 & K/M \\
1321: 48 & 18 37 57.51 & -6 52 25.1 &$^{1}$ & & & &$^{2}$ & 1.99$\pm$0.22 & 1.91$\pm$0.17 & 9.10$\pm$1.17 & K/M \\
1322: 49 & 18 37 57.66 & -6 53 13.8 & & & & &$^{2}$ & 2.39$\pm$0.05 & 0.72$\pm$0.07 & -3.15$\pm$0.12 & B \\
1323: 50 & 18 37 57.74 & -6 52 10.9 &$^{1}$ & & & &$^{2}$ & 2.98$\pm$0.70 & -2.55$\pm$0.68 & 78.45$\pm$1.08 & K/M \\
1324: 51 & 18 37 57.77 & -6 52 22.2 & & 10.262 & 7.240 & 5.670 & & -0.15$\pm$0.30 & 0.68$\pm$0.43 & 54.48$\pm$0.80 & K/M \\
1325: 52 & 18 37 57.78 & -6 52 32.0 & & 10.651 & 8.070 & 6.682 & & 2.50$\pm$0.59 & -0.07$\pm$0.30 & -2.36$\pm$0.30 & \\
1326: 53 & 18 37 57.79 & -6 52 37.4 &$^{1}$ & & & &$^{2}$ & 6.60$\pm$1.83 & 0.30$\pm$0.36 & -10.22$\pm$2.45 & A \\
1327: 54 & 18 37 57.81 & -6 53 25.8 &$^{1}$ & & & &$^{2}$ & 2.18$\pm$0.15 & 1.17$\pm$0.12 & 0.06$\pm$0.08 & B \\
1328: 55 & 18 37 57.98 & -6 53 06.8 &$^{1}$ & & & &$^{2}$ & 0.80$\pm$0.08 & 0.39$\pm$0.17 & 10.13$\pm$0.57 & K/M \\
1329: 56 & 18 37 57.99 & -6 51 43.5 & & 12.597 & 9.982 & 8.730 & & -1.17$\pm$0.14 & 0.75$\pm$0.23 & 31.59$\pm$0.56 & K/M \\
1330: 57 & 18 37 58.06 & -6 53 05.9 & & 14.547 & 12.789 & 11.950 & & 6.44$\pm$0.41 & -1.50$\pm$0.16 & -1.80$\pm$0.55 & A \\
1331: 58 & 18 37 58.11 & -6 53 05.1 &$^{1}$ & & & &$^{2}$ & 4.83$\pm$0.17 & 0.40$\pm$0.25 & -0.88$\pm$0.45 & A \\
1332: 59 & 18 37 58.22 & -6 52 51.5 & & 15.684 & 12.731 & 11.183 &$^{2}$ & 0.00$\pm$0.09 & 0.79$\pm$0.31 & 23.40$\pm$0.71 & K/M \\
1333: 60 & 18 37 58.31 & -6 53 01.8 & & 15.196 & 13.741 & 11.564 & & 5.18$\pm$0.23 & 1.50$\pm$0.23 & 4.00$\pm$1.84 & B \\
1334: 61 & 18 37 58.39 & -6 52 43.8 & & 14.516 & 11.951 & 10.638 &$^{2}$ & 3.37$\pm$0.10 & 1.31$\pm$0.02 & -1.39$\pm$0.39 & B \\
1335: 62 & 18 37 58.52 & -6 53 03.2 & & 15.660 & 13.500 & 12.247 & & 3.81$\pm$0.04 & 1.05$\pm$0.05 & 0.73$\pm$0.40 & B \\
1336: 63 & 18 37 58.55 & -6 52 57.9 & & 15.065 & 11.863 & 10.289 & & -0.89$\pm$0.03 & 0.31$\pm$0.06 & 25.39$\pm$0.57 & K/M \\
1337: 64 & 18 37 58.65 & -6 52 49.1 & & 16.937 & 14.332 & 13.762 & & 0.57$\pm$0.41 & 1.43$\pm$0.11 & 19.46$\pm$1.14 & K/M \\
1338: 65 & 18 37 58.67 & -6 51 35.2 & & 12.138 & 10.542 & 9.856 & & -0.14$\pm$0.24 & 1.43$\pm$0.30 & 20.04$\pm$0.61 & K/M \\
1339: 66 & 18 37 58.85 & -6 53 23.6 &$^{1}$ & 14.469 & 12.482 & 11.516 &$^{2}$ & 1.98$\pm$0.03 & 0.35$\pm$0.03 & -0.32$\pm$0.29 & B \\
1340: 67 & 18 37 59.10 & -6 53 22.5 &$^{1}$ & & & &$^{2}$ & 5.28$\pm$0.26 & 1.23$\pm$0.22 & 2.40$\pm$0.73 & B \\
1341: 68 & 18 37 59.31 & -6 52 15.3 & & 15.848 & 12.278 & 9.786 & & -4.41$\pm$0.11 & -0.14$\pm$0.05 & -0.23$\pm$0.64 & Be \\
1342: 69 & 18 37 59.31 & -6 52 05.7 & & 15.582 & 14.094 & 13.061 & & 6.35$\pm$0.27 & 2.15$\pm$0.24 & 2.40$\pm$2.36 & B \\
1343: 70 & 18 37 59.98 & -6 53 12.5 & & 15.929 & 12.908 & 11.491 & & 2.36$\pm$0.05 & 0.88$\pm$0.18 & 8.88$\pm$0.79 & K/M \\
1344: 71 & 18 38 00.02 & -6 53 21.1 & & 15.455 & 12.293 & 10.987 & & -0.87$\pm$0.11 & 0.39$\pm$0.11 & 24.53$\pm$0.58 & K/M \\
1345: 72 & 18 38 00.19 & -6 53 30.8 &$^{1}$ & & & &$^{2}$ & 0.03$\pm$0.13 & -0.78$\pm$0.04 & 10.24$\pm$0.36 & K/M \\
1346: 73 & 18 38 00.28 & -6 51 58.1 &$^{1}$ & & & &$^{2}$ & 0.54$\pm$0.46 & 1.90$\pm$0.31 & 15.26$\pm$1.05 & K/M \\
1347: 74 & 18 38 00.58 & -6 51 59.1 & & 15.711 & 12.875 & 11.409 & & 0.15$\pm$0.40 & 0.95$\pm$0.36 & 19.23$\pm$0.90 & K/M \\
1348: 75 & 18 38 00.97 & -6 53 21.6 & & 13.497 & 10.393 & 8.997 & & -1.12$\pm$0.16 & 0.78$\pm$0.27 & 30.73$\pm$0.47 & K/M \\
1349: 76 & 18 38 01.09 & -6 52 00.6 & & 15.597 & 12.983 & 11.460 & & -0.88$\pm$0.28 & 0.07$\pm$0.43 & 28.67$\pm$0.87 & K/M \\
1350: 77 & 18 38 01.18 & -6 53 22.8 &$^{1}$ & & & &$^{2}$ & 3.67$\pm$0.41 & 0.82$\pm$0.22 & 6.66$\pm$0.52 & K/M \\
1351: 78 & 18 38 01.46 & -6 51 43.9 & & 15.539 & 11.495 & 9.224 & & -1.75$\pm$0.36 & -0.35$\pm$0.31 & 31.25$\pm$0.60 & K/M \\
1352: 79 & 18 38 01.74 & -6 53 29.5 & & 15.455 & 12.824 & 11.472 & & 0.30$\pm$0.08 & 1.12$\pm$0.36 & 19.46$\pm$0.89 & K/M \\
1353: 80 & 18 38 02.05 & -6 51 35.4 & & 16.902 & 12.836 & 10.739 & & -1.36$\pm$0.17 & 0.44$\pm$0.16 & 32.06$\pm$0.43 & K/M \\
1354: 81 & 18 38 02.24 & -6 52 05.1 & & 14.742 & 13.407 & 12.870 & & 1.25$\pm$0.08 & 0.85$\pm$0.15 & 8.41$\pm$0.87 & K/M \\
1355: 82 & 18 38 02.82 & -6 51 23.7 & & 16.403 & 14.045 & 12.899 & & 1.80$\pm$0.21 & -0.30$\pm$0.18 & 5.47$\pm$0.29 & K/M \\
1356: 83 & 18 38 59.43 & -6 53 22.1 &$^{1}$ & & & &$^{2}$ & 2.48$\pm$0.29 & 0.39$\pm$0.19 & 6.42$\pm$0.39 & K/M \\
1357: \hline
1358: \end{tabular}
1359: \end{tiny}
1360: \end{table}
1361:
1362: % \end{landscape}
1363:
1364:
1365:
1366: \begin{figure}
1367: \centering
1368: \includegraphics[width=14cm]{RSG1cmd.eps}
1369: \caption{RSGC1 col-mag diagram. A non-rotating 10Myr isochrone is
1370: overplotted. }
1371: \end{figure}
1372:
1373: \begin{figure}
1374: \centering
1375: \includegraphics[width=14cm]{loresfind.eps}
1376: \caption{Finding chart for the lo-res stars. }
1377: \end{figure}
1378:
1379:
1380:
1381:
1382:
1383:
1384:
1385:
1386: \subsection{Low-resolution spectroscopy}
1387: Target objects were identified as stars within 3\arcmin of the cluster
1388: centre (as determined from the RSGs) with similar 2MASS colours as the
1389: RSGs ($H-K_{S} \ga$1), and $K_{S}$-band magnitudes of $13 \la K_{S} \la
1390: 8$, determined from the predicted location of BSGs from the isochrone
1391: fits in FMR06.
1392:
1393: Observations were taken with {\it NIRSPEC}/Keck-II on 12$^{th}$ Aug
1394: 2006, this time with the cross-disperser angle set to 179.7\degr
1395: (i.e.\ {\it no} cross dispersion). We observed through the $K$-filter,
1396: using the 0.570\arcsec$\times$42\arcsec\ slit, and achieved spectral
1397: resolution R=2000 in the wavelength range 1.996-2.382\microns. We
1398: integrated for 100s on each star in two nod positions along the slit.
1399:
1400: Reduction of the data followed the exact same process as for the
1401: high-resolution data, except that rectification in the dispersion
1402: direction was done using arc/sky lines only.
1403:
1404: In practice, the targets selected from the 2MASS images often were
1405: resolved into multiple stars. Also, in addition to the target stars,
1406: many other stars of detectable brightness fell into the slit. We
1407: extracted all spectra from our data for which a S/N of $\ga$50 could
1408: be achieved. We identified the stars we observed using our new
1409: higher-resolution photometry (see below), and numbered them in order
1410: of increasing RA. A list of the stars observed can be found in Table
1411: REF!!!
1412:
1413:
1414:
1415: \subsection{Near-IR photometry}
1416: Antonio needs to write this...
1417:
1418: \subsection{Low-resolution spectra}
1419: In order to distinguish intrinsically `blue', reddened cluster stars
1420: from foreground red giants, we used a systematic process which
1421: measures the equivalent width (\EW) of each spectrum in the regions of
1422: \brg\ (2.1605 -- 2.1705\microns) and the CO bandhead
1423: (2.293-2.313\microns). We defined a relevant continuum measurement for
1424: each spectral feature from the median continuum values in the regions
1425: 2.13--2.15\microns\ and 2.27--2.28\microns\ respectively. If a star
1426: had a CO detection of 8$\sigma$ or greater, it was classified as
1427: spectral-type `K/M', and therefore likely a foreground star. Likewise,
1428: if a \brg\ detection of 8$\sigma$ or greater resulted in a preliminary
1429: `B/A' spectral classification.
1430:
1431: Of the 83 stars observed, 21 had a positive detection of \brg. The
1432: spectra of these stars are shown in Fig.\ \ref{fig:bspectra}. The
1433: position of \brg\ is marked on the plot, as are the \hei\ lines at
1434: XXX, YYY and ZZZ\microns, and the \feii\ line at QQQ\microns. It can
1435: be seen that many of the stars with \brg\ also have \hei\ in
1436: absorption. D67 actually has spectral features in emission, while D44
1437: apparently has \hei\ XXX in emission. We note that the sample is
1438: contaminated by the star F15 ($=$D52 in this numbering system), the
1439: supposed YHG from FMR06. This star also has \brg\ and \hei\ XXX in
1440: absorption, but also has {\bf(Mg??)} in emission -- typical of G-type
1441: stars.
1442:
1443: \begin{figure}[t]
1444: \centering
1445: \includegraphics[width=14cm,clip,bb=110 70 710 580]{ukfinder.eps}
1446: \caption{Finding chart for those stars found to have Br$\gamma$ in
1447: emission/absorption, overlayed on the UKIDSS $H$-band
1448: image. Coordinates for the stars are given in Table REF!!!}
1449: \label{fig:ukfinder}
1450: \end{figure}
1451:
1452: A finding chart for these stars is shown in Fig.\ \ref{fig:ukfinder}.
1453:
1454: \begin{figure}[p]
1455: \centering
1456: \includegraphics[width=16cm,clip,bb=30 50 566 651]{bstars.eps}
1457: \caption{Low-resolution K-band spectra of all stars observed with
1458: Br$\gamma$ in emission/absorption.}
1459: \label{fig:bspectra}
1460: \end{figure}
1461:
1462:
1463:
1464:
1465:
1466:
1467:
1468:
1469:
1470:
1471: