1: \documentclass[10pt,preprint]{aastex}
2: %\documentclass[emulateapj,epsfig]{article}
3:
4:
5: %\received{}
6: %\accepted{}
7: %\journalid{}{}
8: %\articleid{}{}
9: \slugcomment{To appear in the Astrophysical Journal, Part 1, Nov 10, 2003}
10: %\lefthead{Hanson}
11: %\righthead{Galactic Super Star Clusters}
12:
13:
14: %\voffset0.7in
15: %\hoffset-.2in
16: \begin{document}
17:
18: \title{A Study of Cyg OB2: Pointing the Way Towards Finding\\
19: Our Galaxy's Super Star Clusters}
20: \author{M.M. Hanson$^1$}
21: \affil{Department of Physics, The University of Cincinnati, Cincinnati,
22: OH 45221-0011}
23: \altaffiltext{1}{Visiting astronomer, Steward Observatory, University of Arizona}
24:
25: \begin{abstract}
26: New optical MK classification spectra have been obtained for
27: 14 OB star candidates identified by Comer\'{o}n et al.\
28: (2002) and presumed to be possible members of
29: the Cyg OB2 cluster as recently described by Kn\"{o}dlseder
30: (2000). All 14 candidate OB stars observed
31: are indeed early-type stars, strongly suggesting the
32: remaining 31 candidates by Comer\'{o}n et al.\ are also
33: early-type stars. A thorough investigation of the properties of
34: these new
35: candidate members compared to the properties of the Cyg OB2
36: cluster star have been completed, using traditional as well
37: as newly revised effective temperature scales for O stars.
38: The cooler O-star, effective temperature scale
39: of Martins et al. (2002) gives a very close distance for the
40: cluster (DM = 10.4). However, even using traditional effective
41: temperature
42: scales, Cyg OB2 appears to be slightly closer (DM = 10.8) than
43: previous studies determined (DM = 11.2; Massey \& Thompson 1992),
44: when the very young age of the stellar cluster ($\sim$ 2 x 10$^6$ yrs)
45: is taken into account in fitting the late-O and early-B dwarfs
46: to model isochrones.
47: Of the 14 new OB stars observed for this study, as many
48: as half appear to be significantly older than the
49: previously studied optical cluster, making
50: their membership in Cyg OB2 suspect. So, while some of the
51: newly identified OB stars may represent a more extended halo
52: of the Cyg OB2 cluster, the survey of Comer\'{o}n et al.\
53: also picked up a large fraction of non-members. Presently,
54: estimates of the very high mass of this cluster ($M_{cl}
55: \approx 10^4 M_{\odot}$ and over 100 O stars) first made by
56: Kn\"{o}dlseder (2000) remain higher than this study can support.
57: Despite this, the recognition of Cyg OB2 being a more
58: massive and extensive star cluster than previously realized
59: using 2MASS images, along with the recently recognized
60: candidate super
61: star cluster Westerlund~1 only a few kpc away (Clark \&
62: Negueruela 2002), reminds us that we are
63: woefully under-informed about the massive cluster
64: population in our Galaxy. Extrapolations of the locally
65: derived cluster luminosity function indicate 10s to perhaps
66: 100 of these very massive open clusters ($M_{cl} \approx 10^4
67: M_{\odot}$, $M_V \approx -11$) should exist within our
68: galaxy. Radio surveys will not detect
69: these massive clusters if they are more than a few million
70: years old. Our best hope for remedying this shortfall is
71: through deep infrared searches and follow up near-infrared
72: spectroscopic observations, as was used by Comeron et al.\
73: (2002) to identify
74: candidate members of the Cyg OB2 association.
75: \end{abstract}
76:
77: \keywords{stars : early type -- Galaxy : open clusters and
78: associations, stellar content -- individual : Cyg OB2 --
79: infrared : stars}
80:
81: \section{INTRODUCTION}
82:
83: An accurate census of our Galaxy's massive star content is
84: essential to understanding its evolution and structure
85: (see Russeil 2003).
86: Yet until very recently, few open clusters were known
87: to exist with extinction greater than A$_V > 5$.
88: Moving our imaging and spectroscopic studies to
89: the near-infrared, specifically around 1 to 2~$\mu$m,
90: has begun to uncover significant massive star populations
91: previously unknown or poorly understood.
92: Because of its conspicuous
93: nature, the galactic center region was among the first to
94: be deeply studied in the near-infrared. These studies identified
95: at least three spectacular and very massive clusters (Morris \&
96: Serabyn 1996; Figer et al.\ 1999; Eckart, Ott \& Genzel 1999)
97: with unique stellar and cluster properties. Arguments
98: about the unusual nature of the galactic center environment have
99: been used to explain the galactic center cluster characteristics
100: (Morris \& Serabyn 1996), but how unique are the galactic center
101: clusters to other massive clusters lurking through out our galaxy?
102: Might similar very massive clusters lie in other regions of our
103: galaxy?
104:
105: The recently completed 2MASS survey (Skrutskie et al.\ 1997)
106: represents a critical
107: first step in uncovering massive open clusters
108: in our galaxy. Not only are numerous new clusters being found
109: through near-infrared surveys (Dutra \& Bica 2000; 2001;
110: Ivanov et al. 2002), but even well known clusters are
111: being found to be more extensive and massive than previously
112: thought. Such is the case for
113: Cygnus OB2, a massive star cluster less than 2 kpc away.
114: Originally discovered and labeled IV Cygni by Munch \& Morgan (1953),
115: Cyg OB2 has been the focus of numerous optical studies, dating
116: back nearly 50 years (Johnson \& Morgan 1954; Morgan et al.\ 1954;
117: Schulte 1956, 1958). The most thorough investigation of Cyg~OB2
118: was completed by Massey \& Thompson (1991, hereafter MT91),
119: identifying 120 possible massive star members based on UBV photometry,
120: and giving optical spectral classifications for just over 70 OB stars in
121: the field. Recently, Kn\"{o}dlseder (2000), using 2MASS infrared
122: observations, has reevaluated the stellar content of Cyg OB2, arguing
123: for a more massive and extensive cluster than previously recognized.
124: Kn\"{o}dlseder (2000) proposed Cyg OB2 to be the most massive
125: stellar association known in our Galaxy, and referred to it as
126: a ``super star cluster'', containing over 100 O stars
127: (Kn\"{o}dlseder et al.\ 2002).
128:
129: The Kn\"{o}dlseder (2000) study of Cyg OB2 was completed
130: using near-infrared imaging photometry alone.
131: Though far more time consuming, and not possible for very heavily
132: reddened sight-lines, optical spectroscopic studies, such as that
133: completed by MT91 on Cyg OB2, give us the needed
134: information to confidently characterize a cluster's mass, age,
135: initial mass function (IMF), and to accurately study the individual members.
136: As a follow up to the Kn\"{o}dlseder (2000) study, Comer\'{o}n et
137: al.~(2002, here after CPR2002) presented a low resolution $H$- and $K$-band
138: spectroscopic survey of 2MASS identified sources towards Cyg OB2.
139: While these spectra were not of the
140: quality required for near-infrared spectral classification (see Hanson,
141: Howarth and Conti 1997), they were sufficient to quickly confirm
142: which stars already showing ``blue'' near-infrared colors lacked
143: discernible molecular bands, consistent with them being early-type
144: stars. They identified 77 early-type candidates.
145: Less than half, just 31 stars, had been previously classified with
146: optical spectra to have been OB stars, leaving 46 new candidate OB
147: stars towards Cyg OB2.
148:
149: In this study, we have obtained classification-quality blue spectra
150: for 14 of the 46 OB star candidates identified by CPR2002 to
151: determine their MK classifications. A 15th star from the CPR2003 list
152: was observed, but it turned out to have been previously studied
153: spectroscopically. The primary goal of this study
154: is to determine if the OB candidates identified using 2MASS colors and
155: low-resolution near-infrared survey spectra are indeed OB stars.
156: In this way, the observations presented here serve as a useful test
157: of this newly explored, but clearly important, near-infrared method
158: of identifying OB star populations through out our galaxy and behind
159: large line-of-sight extinction. Observations are presented in \S 2,
160: and the new spectra are presented in \S 3.
161:
162:
163: A second goal of this paper, which is more difficult than its first,
164: is to determine if the newly found OB stars are associated with the
165: optically studied Cyg OB2 cluster of MT91. This will require a thorough
166: evaluation of the cluster's characteristics to determine the likelihood
167: of membership for any newly found OB stars in the region. A review
168: of the Cyg OB2 cluster characteristics and the characteristics of the
169: newly found OB stars is given in \S 4. A final discussion of the
170: search for possible super star clusters within our Galaxy is presented
171: in \S 5. Concluding remarks are found in \S 6.
172:
173:
174:
175: \section{OBSERVATIONS AND REDUCTIONS}
176:
177: Observations were made the nights of 6, 7, 8 July, 2002 on the University
178: of Arizona's 2.3m Bok Telescope, located on Kitt Peak, outside of Tucson, AZ.
179: The Boller and Chivens (B\&C) spectrograph was employed and operated with an
180: 832 g/mm grating and a Schott 8612 order separating filter. A full-width,
181: half-maximum resolution of FWHM $\approx$ 2.0 \AA\ (about 2.8 pixels) was
182: achieved for a resolution of R $\approx$ 2200 over the spectral range
183: from 3960 to 4800 \AA.
184:
185: The B\&C spectrograph uses a long slit (4$'$). All observations were
186: made using a slit width of 2.5$''$. An average bias as well as sky emission
187: lines were removed by subtracting a median averaged image of several
188: unique slit positions. Pixel-to-pixel gain variations on the CCD detector
189: were removed using observations of an illuminated reflective spot
190: inside the dome. Observations of a Helium-Argon lamp taken
191: periodically through the night provided the wavelength calibrations.
192: Integration times ranged from 16 minutes (Cyg OB2 A46) to as long as
193: an hour (Cyg OB2 A20). The signal-to-noise ratio in the line
194: free continuum exceeds 50 for nearly all spectra. A few spectra
195: drop just below this value at the shortest wavelengths where the
196: CCD response is waning.
197:
198:
199:
200: \begin{deluxetable}{lcccccc}
201: \tablewidth{0pt}
202: \tablecaption{Cygnus OB2 New MK Classification Spectra}
203: \tablehead{
204: \colhead{Star} &
205: \colhead{$\alpha$(2000)} &
206: \colhead{$\delta$(2000)} &
207: \colhead{m$_{B}$} &
208: \colhead{Exp Time} &
209: \colhead{S/N} &
210: \colhead{SpType}
211: }
212: \startdata
213: A20 & 20 33 02.9 & 40 47 25 & $\sim$14.5 & 60 min. & 50 & O8 II((f)) \\
214: A23 & 20 30 39.7 & 41 08 48 & $\sim$14.0 & 24 min. & 80 & B0.7 Ib \\
215: A27 & 20 34 44.7 & 40 51 46 & $\sim$13.0 & 24 min. & 100 & B0 Ia \\
216: A29 & 20 34 56.0 & 40 38 18 & $\sim$14.0 & 24 min. & 80 & O9.7 Iab \\
217: A32 & 20 32 30.3 & 40 34 33 & $\sim$14.0 & 44 min. & 60 & O9.5 IV \\
218: A34 & 20 31 36.9 & 42 01 21 & $\sim$13.0 & 24 min. & 100 & B0.7 Ib \\
219: A36 & 20 34 58.7 & 41 36 17 & $\sim$13.0 & 30 min. & 80 & B0 Ib(n) sb2? \\
220: A37 & 20 36 04.5 & 40 56 13 & $\sim$14.5 & 30 min. & 50 & O5 V((f)) \\
221: A39 & 20 32 27.3 & 40 55 18 & $\sim$14.0 & 30 min. & 60 & B2 V \\
222: A41 & 20 31 08.3 & 42 02 42 & $\sim$13.0 & 30 min. & 100 & O9.7 II \\
223: A42 & 20 29 57.0 & 41 09 53 & $\sim$14.5 & 50 min. & 60 & B0 V \\
224: A43 & 20 32 38.5 & 41 25 13 & 12.03 & 24 min. & 100 & O8 V(n)\tablenotemark{a} \\
225: A44 & 20 31 46.0 & 40 43 24 & $\sim$13.5 & 30 min. & 60 & B0.5 IV \\
226: A45 & 20 29 46.6 & 41 05 08 & $\sim$13.0 & 30 min. & 80 & B0.5 V(n) sb2? \\
227: A46 & 20 31 00.1 & 40 49 49 & $\sim$12.0 & 16 min. & 80 & O7 V((f)) \\
228: \enddata
229: \tablenotetext{a}{SpType from Hutchings (1981) B0; Massey \& Thompson (1991, MT91) O7.5 V.}
230: \end{deluxetable}
231:
232:
233: \subsection{Target Selection}
234:
235: A list of the stars observed and their positions is given in Table 1.
236: Stars were selected from the list of candidate early-type members given in
237: CPR2002. Interstellar extinction towards this sample is
238: exceedingly high in the blue-optical, ranging from A$_B$ = 6 to more than
239: 10 magnitudes. Estimated $B$ magnitudes were calculated from Table 1 in
240: CPR2002, assuming a standard extinction law (Rieke \& Lebofsky
241: 1985). Observations were made of all candidate early-type stars
242: with estimated B magnitudes brighter than $B \approx 14.5$. No stars
243: were observed from the list of Br$\gamma$ emission stars in CPR2002
244: (their Table 2).
245:
246: Optical spectra of the quality presented here were not previously
247: available for any of the target stars with the exception of one
248: star, Cyg OB2~A43. This star was among the first OB stars found in
249: Cyg OB2 based on photometric colors and was identified by Schulte
250: as star 16 (Schulte 1956). Hutchings (1981) and MT91 obtained
251: spectroscopic measures of Cyg OB2~A43,
252: assigning MK spectral types of B0 and O7.5 V, respectively.
253:
254:
255:
256: \begin{figure}
257: \epsscale{0.8}
258: \plotone{f1.ps}
259: \caption{MK Classification spectra of newly confirmed dwarf OB stars toward Cyg OB2.
260: Interstellar features due to diffuse interstellar bands are marked with a vertical
261: dashed line.\label{fig1}}
262: \end{figure}
263:
264:
265: \begin{figure}
266: \epsscale{0.8}
267: \plotone{f2.ps}
268: \caption{OB supergiant stars toward Cyg OB2, see Fig.\ 1.\label{fig2}}
269: \end{figure}
270:
271:
272:
273:
274: \section{THE SPECTRA}
275:
276: The new optical spectra are shown in Figures 1 \& 2. A very quick examination
277: shows that all stars from the sample show HeI 4471 \AA\, indicating
278: they are early-type stars, as predicted by CPR2002.
279: MK spectral types, determined based on visual comparisons with
280: the CCD spectral atlas of Walborn \& Fitzpatrick (1990), are given
281: in Table 1. There is no quantitative method for estimating the
282: uncertainty in spectral types, however, all 15 stars were independently
283: typed by three persons, myself, Phil Massey and Nolan Walborn. There
284: was excellent agreement within a spectral class for all stars (except
285: A39 and A44, which I typed differently). The adopted classifications,
286: independently and consistently given by both Massey and Walborn,
287: are listed in Table 1.
288:
289:
290: Because of the large line-of-sight extinction, very strong interstellar
291: features are seen in our spectra. The commonly detected diffuse interstellar
292: band (DIB) at 4428 \AA\ is shown to be exceedingly strong in all the
293: spectra. A composite DIB spectrum, which was created from all the
294: Cyg OB2 star spectra after prominent stellar photospheric features
295: where removed, is shown in Figure 3. Using the line lists provided by
296: Jenniskens \& Desert (1994), additional DIB features are clearly
297: detected at 4501.80, 4726.59/4727.06 blend, 4761.67/4762.57 blend (one
298: narrow and one quite wide) and 4780.09 \AA. Jenniskens \&
299: Desert (1994) also list a probable, very broad DIB centered at
300: 4595.0 \AA. We may be detecting such a feature, however the line
301: does lie in a region where weak Si III stellar features are plentiful
302: and might not have been adequately removed in the composite spectrum
303: of Fig.\ 3. Line centers, equivalent
304: width and FWHM measures have been made from fitting the spectrum in
305: Fig.\ 3.
306: The fit is shown over-plotted in Fig.\ 3 and the results from the fit
307: are given in Table 2.
308:
309:
310:
311:
312:
313: \begin{figure}
314: \epsscale{0.8}
315: \plotone{f3.ps}
316: \caption{An average spectrum of fifteen Cyg OB2 stars with prominent stellar
317: features removed to show the main diffuse interstellar band (DIB) features.
318: A fit of the DIB features used to create inputs to Table 2 is shown over-plotted
319: with the spectra.\label{fig3}}
320: \end{figure}
321:
322: \begin{deluxetable}{lcc}
323: %\tablewidth{160pt}
324: \tablewidth{0pt}
325: \tablecaption{Composite DIB Features}
326: \tablehead{
327: \colhead{Central} &
328: \colhead{EQW\tablenotemark{a}} &
329: \colhead{FWHM\tablenotemark{b}} \\
330: \colhead{$\lambda$, \AA} &
331: \colhead{\AA} &
332: \colhead{\AA}
333: }
334: \startdata
335: 4428.2 & 4.8 & 18 \\
336: 4501.2 & 0.2 & 3.4 \\
337: 4597.5 & 0.6 & 40 \\
338: 4726.5 & 0.3 & 4.7 \\
339: 4762.0 & 0.2 & 4.3 \\
340: 4762.3 & 1.0 & 35 \\
341: 4779.4 & 0.07 & 2.8 \\
342: \enddata
343: \tablenotetext{a}{EQW: Equivalent width of feature, measured in Angstroms.}
344: \tablenotetext{b}{FWHM: Full
345: width, half-maximum line depth, measured in Angstroms.}
346: \end{deluxetable}
347:
348:
349:
350: \section{THE NEW OB STARS: ARE THEY MEMBERS OF CYG OB2?}
351:
352: Kn\"{o}dlseder (2000) makes a convincing argument, based on 2MASS stellar
353: counts, that a significant fraction of the Cyg OB2 star cluster has been
354: missed in optical surveys. Follow up work by CPR2002 appears to support
355: this interpretation. The amazing success of the CPR2002 study to have
356: identified OB stars in the field would suggest that all 46 candidate OB
357: stars found are indeed OB stars. However, a second equally important goal
358: of this paper is to determine if the OB stars uncovered by CPR2002 are
359: members of the optical OB cluster Cyg OB2 studied
360: by MT91. In this section we will compare the characteristics of the newly
361: identified OB stars with the characteristics of the previously identified
362: OB stars of Cyg OB2 to determine this.
363:
364: \subsection{Location of the new stars}
365:
366: In Figure 4 is shown the CPR2003 survey area,
367: covering a 1.5$^{\circ}$ x 2$^{\circ}$ field of view. The center of the field
368: is aligned with the center of the star distribution seen for
369: Cyg OB2 by Kn\"{o}delseder (2000) using 2MASS, $\alpha = 20^{h} 33^{m} 10^{s}, \
370: \delta=
371: +41^{\circ} 15.7'$. The circles correspond
372: to the 120 stars previously identified to be members of the Cyg OB2
373: cluster by MT91 based on UBV colors. Asterisks in Figure 4 locate
374: a total of 62 new massive star candidates identified by CPR2002.
375: The asterisk sources include 45 candidate early-type members
376: (CPR 2002 A43 is left out since it was previously identified as
377: Schulte 16), 14 stars with Br~$\gamma$ emission (6 of
378: their 20 Br~$\gamma$ sources were previously identified), and 3 new
379: sources showing CO in emission.
380:
381: What is first immediately obvious in this image is that most of
382: the new sources found by CPR2002 lie well away from the
383: previously identified optical cluster stars. Only near the south
384: and west side of the optical field are new candidate OB members seen to
385: lie in close proximity to previously identified cluster members. This study
386: has confirmed 14 new OB stars in the field, all 14 of which are labeled with
387: CPR2002 identifications in Fig\ 4 (A20, A23, etc.). Of the 14 new OB stars
388: confirmed in this study, all lie a significant distance from the previously
389: identified optical cluster of MT91. The reason most of these OB stars were
390: missed by MT91 was simply because the field over which MT91 searched did not
391: extend to these angular distances.
392:
393:
394:
395:
396:
397:
398:
399: \begin{figure}
400: \epsscale{0.8}
401: \plotone{f4.ps}
402: \caption{A 1.5$^{\circ}$ by 2$^{\circ}$ optical field from
403: the Digitized Sky Survey centered on Cyg
404: OB2 as defined by Kn\"{o}delseder (2000) using the 2MASS survey, (J2000)
405: 20$^h$33$^m$10$^s$ $+$41$^{\circ}$15.7$'$. OB stars originally found
406: by Massey \& Thompson (1991, MT91) are shown as circles. Asterisks indicate
407: the location of new OB candidates recently uncovered by CPR2002.
408: Stars from the CPR2002 survey which are confirmed in this
409: study to be OB stars are marked with their star identifications as
410: listed in Table 1. The legend bar in the lower right measures 10$'$.
411: \label{fig4}}
412: \end{figure}
413:
414:
415: \begin{deluxetable}{llccccr}
416: \tablewidth{0pt}
417: \tablecaption{The Dwarf Sample}
418: \tablehead{
419: \colhead{Star} &
420: \colhead{SpTy} &
421: \colhead{T$_{eff}$} &
422: \colhead{Av} &
423: \colhead{m$_V$} &
424: \colhead{M$_V$} &
425: \colhead{V-M$_V$}
426: }
427: \startdata
428: 059 & O8.5V & 37239 & 5.21 & 11.18 & -4.42 & 10.38 \\
429: 070 & O9V & 35974 & 6.96 & 12.99 & -4.19 & 10.22 \\
430: 145 & O9.5V & 34673 & 4.23 & 11.62 & -3.98 & 11.37 \\
431: 187 & B0.5V & 28183 & 5.40 & 13.24 & -2.69 & 10.53 \\
432: 215 & B1V & 25409 & 4.03 & 12.97 & -2.15 & 11.10 \\
433: 227 & O9V & 35974 & 4.73 & 11.47 & -4.19 & 10.93 \\
434: 250 & B1V & 25409 & 4.08 & 12.88 & -2.15 & 10.96 \\
435: 258 & O8V & 38459 & 4.49 & 11.10 & -4.74 & 11.35 \\
436: 259 & B0.5V & 28183 & 3.83 & 11.42 & -2.69 & 10.27 \\
437: 292 & B1V & 25409 & 5.40 & 13.08 & -2.15 & 9.83 \\
438: 299 & O8V & 38459 & 4.61 & 10.84 & -4.74 & 10.97 \\
439: 317 & O8V & 38459 & 4.73 & 10.66 & -4.74 & 10.67 \\
440: 339 & O8.5V & 37239 & 4.94 & 11.60 & -4.42 & 11.08 \\
441: 376 & O8V & 38459 & 4.96 & 11.91 & -4.74 & 11.69 \\
442: 378 & B0V & 31622 & 7.14 & 13.49 & -3.38 & 9.73 \\
443: 390 & O8V & 38459 & 6.72 & 12.95 & -4.74 & 10.97 \\
444: 421 & O9.5V & 34673 & 6.83 & 12.86 & -3.98 & 10.00 \\
445: 429 & B0V & 31622 & 5.64 & 12.98 & -3.38 & 10.71 \\
446: 455 & O8V & 38459 & 6.38 & 12.92 & -4.74 & 11.27 \\
447: 467 & B1V & 25409 & 5.04 & 13.43 & -2.15 & 10.55 \\
448: 470 & O9.5V & 34673 & 5.19 & 12.50 & -3.98 & 11.29 \\
449: 473 & O8.5V & 37239 & 5.29 & 12.02 & -4.42 & 11.14 \\
450: 480 & O7.5V & 39810 & 5.82 & 11.88 & -5.06 & 11.12 \\
451: 485 & O8V & 38459 & 5.58 & 12.06 & -4.74 & 11.21 \\
452: 507 & O8.5V & 37239 & 5.61 & 12.70 & -4.42 & 11.51 \\
453: 515 & B1V & 25409 & 6.97 & 14.66 & -2.15 & 9.85 \\
454: 531 & O8.5V & 37239 & 5.57 & 11.58 & -4.41 & 10.42 \\
455: 534 & O7.5V & 39810 & 6.54 & 13.00 & -5.06 & 11.51 \\
456: 555 & O8V & 38459 & 6.57 & 12.51 & -4.74 & 10.68 \\
457: 588 & B0V & 31622 & 6.02 & 12.40 & -3.38 & 9.76 \\
458: 605 & B0.5V & 28183 & 4.23 & 11.78 & -2.69 & 10.24 \\
459: 692 & B0V & 31622 & 5.69 & 13.61 & -3.38 & 11.30 \\
460: 696 & O9.5V & 34673 & 5.85 & 12.32 & -3.98 & 10.45 \\
461: 716 & O9V & 35974 & 6.10 & 13.50 & -4.19 & 11.59 \\
462: 736 & O9V & 35974 & 5.46 & 12.79 & -4.19 & 11.52 \\
463: & & & \multicolumn{4}{r}{Average Vo - Mv =} 10.80 \\
464: \enddata
465: \end{deluxetable}
466:
467:
468: \subsection{Extinction towards Cyg OB2}
469:
470: From the full sample of 71 Cyg OB2 stars with spectral types published in MT91,
471: a subset of 35 stars, listed in Table 3, was selected to perform specific
472: calibrations of the cluster characteristics. This subset of stars was selected
473: to have optical spectral types between O7.5 and B1 inclusive, all are
474: normal dwarf stars, and all have both UBV values as well as 2MASS JHK$_S$
475: colors. None of the new OB stars presented in this paper were included.
476: Using the intrinsic UBV color terms given by FitzGerald (1970; and recently
477: used by Slesnick, Hillenbrand, \& Massey 2002)
478: and JHK$_S$ color terms given by Lejenue \& Schaerer
479: (2001) for dwarf stars, the extinction characteristics were determined
480: toward each star independently based on the stars observed UBVJHK$_S$
481: colors. This has lead to a well constrained average extinction law
482: towards the cluster as illustrated in Figures 5, 6 and 7. The extinction
483: law created from the fits shown
484: in Figs.\ 5, 6, and 7 are given in Table 4. The optical
485: extinction law was previously measured by MT91 and Torres-Dodgen et al.\ (1991).
486: As seen in these previous studies, the extinction was measured to
487: be consistent with a total to selective extinction ratio, $R_V$, of 3.0.
488: The extinction law through the near-infrared matches that given for the
489: general interstellar medium by Rieke \& Lebofsky (1985), except for the K band.
490: The centroid of the 2MASS K$_S$ band is at 2.16 $\mu$m, while Rieke \& Lebofsky used
491: Johnson K-band filters centered at 2.22 $\mu$m. This leads to a slightly larger
492: extinction ratio in the 2MASS K$_S$ band ($A_{K_s}/A_V = 0.125$) relative to the
493: Rieke \& Lebofsky (1985) K band ($A_K/A_V = 0.112$).
494:
495:
496: \begin{deluxetable}{ccc}
497: %\tablewidth{200pt}
498: \tablewidth{0pt}
499: \tablecaption{Insterstellar Extinction Law toward Cyg OB2}
500: \tablehead{
501: \colhead{$\lambda$} &
502: \colhead{$A_{\lambda}/A_V$} &
503: \colhead{RL85}
504: }
505: \startdata
506: U (0.36 $\mu$m) & 1.600 & 1.531 \\
507: B (0.44 $\mu$m) & 1.333 & 1.324 \\
508: V (0.55 $\mu$m) & 1.000 & 1.000 \\
509: J (1.24 $\mu$m) & 0.282 & 0.282 \\
510: H (1.66 $\mu$m) & 0.175 & 0.175 \\
511: K (2.16 $\mu$m) & 0.125$^1$ & 0.112$^2$ \\
512: \enddata
513: \tablenotetext{1}{K-band centroid at 2.16 $\mu$m.}
514: \tablenotetext{1}{K-band centroid at 2.22 $\mu$m.}
515: \end{deluxetable}
516:
517:
518:
519: \begin{figure}
520: \epsscale{0.8}
521: \plotone{f5.ps}
522: \caption{UBV color excesses for 35 OB dwarf stars in Cyg OB2. See \S4.2 \label{fig5}}
523: \end{figure}
524:
525: \begin{figure}
526: \epsscale{0.8}
527: \plotone{f6.ps}
528: \caption{JHK$_S$ color excesses for 35 OB dwarf stars in Cyg OB2. \label{fig6}}
529: \end{figure}
530:
531:
532: \begin{figure}
533: \epsscale{0.8}
534: \plotone{f7.ps}
535: \caption{BVHK$_S$ color excesses for 35 OB dwarf stars in Cyg OB2. \label{fig7}}
536: \end{figure}
537:
538:
539:
540:
541: With a well established extinction law derived for the cluster, A$_V$ values for every
542: star with optical spectral types in our full sample of 85 (71 previously identified by
543: MT91, plus the 14 new OB stars confirmed in this study) could be calculated from either
544: UBV, JHK$_S$ colors, or both depending on what was available. Extinction values for all
545: stars in this study (inclusive of the 35 dwarf stars also listed in Table 4) are given
546: in Table 5. Of the
547: 85 stars in our complete sample extinction varied from
548: A$_V$ = 3.83 up to A$_V$ = 10.70 for the infamous B5 supergiant, IV Cyg \#12.
549:
550:
551: It is worth noting that
552: the average extinction towards the 14 CPR2002 OB stars was not significantly greater
553: than that found towards the OB stars previously identified in the optical by MT91
554: Also, the new OB stars did not show an extinction law any different from that
555: determined for the general Cyg OB2 cluster, beyond a few intrinsically unusual
556: stars listed below. The former indicates that high extinction
557: was not the reason these 14 stars were missed in the optical surveys. However, the
558: remaining 31 stars in the CPR2002 sample of candidate OB stars {\sl do} have significantly
559: higher extinction than those observed for this study. This may have contributed
560: to their being missed in the optical study of MT91.
561:
562:
563:
564:
565: \subsubsection{Sources with unusual photometric properties}
566:
567: Once the extinction law had been derived for the sight line, it was easy to
568: spot outliers in the numerous color-color diagrams created for the entire
569: sample. Listed below are six sources for which their photometry showed
570: unusual values. With the exception of these six stars, the extinction law towards
571: all the stars in the sample, those found by MT91 and those uncovered by
572: the CRP2002 survey, have essentially identical extinction properties.
573:
574: {\bf MT91 390.} This star was identified by Pigulski \& Kolaczkowki (1998)
575: as having an irregular variability in the I-band, but otherwise looks fine
576: in our colors. This star was classified as an O8V star by MT91.
577:
578: {\bf MT91 556} Schulte 18. This star was classified as B1Ib by MT91, however it
579: shows some peculiarities in its UBV colors. Schulte 18 had been identified by
580: Pigulski \& Kolaczkowki (1998) as having an irregular and rather large variability
581: in the I-band. Unfortunately, the spectrum obtained by MT91 is not shown in
582: their paper.
583:
584: {\bf MT91 575} CPR2003 B13. This star was previously known to be an emission
585: line star from Merrill \& Burwell (1950). Its spectral type from MT91 is
586: B1.5V. This star has the largest IR excess in the entire sample, as seen in
587: our (H-K$_S$) vs (J-H) plots. The star is listed in HAe/Be star catalog
588: of The et al.\ (1994). The spectrum as shown in MT91 looks normal.
589:
590: {\bf MT91 605.} This star was previously identified to be an emission line star
591: by Merrill \& Burwell (1950). The UBV and JHK$_S$ colors for this star appear
592: normal. The spectral type, B0.5V, comes from MT91. The spectrum of this star,
593: as shown in MT91 also appears normal.
594:
595: {\bf MT91 793} CPR2002 B16. This star shows a strong near-infrared excess in
596: its JHK$_S$ colors. The spectral type from MT91 is uncertain (B1.5III?). The
597: spectrum
598: shown in MT91 has numerous emission lines, both in hydrogen and from other
599: transitions. It seems likely that this is a Be type star, though it is
600: not mentioned as such in the literature.
601:
602: {\bf CPR2003 A34.} This star shows a relatively strong near-infrared excess
603: in the (H-K$_S$) vs (J-H) diagram. From this work, the star has been classified
604: as a B0.7 Ib star.
605:
606:
607:
608:
609: \subsection{Distance to Cyg OB2}
610:
611: The subsample of dwarf late-O and early-B stars given in Table 3 was used to
612: derive the cluster distance. Calculating a star's distance requires knowing
613: its apparent magnitude, line of sight extinction and its absolute magnitude.
614: The first two measures are well constrained for our sample. The greatest uncertainty
615: in obtaining the distance will be in choosing the absolute magnitude for the
616: dwarf star sample.
617:
618: There are numerous tables of absolute magnitude as a function of spectral type
619: available in the literature. These show relatively good agreement over the past 3
620: decades (Walborn 1973; Humphreys \& McElroy 1984; Conti 1988; Vacca et al.\ 1996).
621: Because of the large number of dwarf
622: stars with similar age and distance, the Cyg OB2 data set can be used to
623: test the consistency of the absolute magnitude calibration.
624: First, the distance modulus was determined using the absolute magnitudes of
625: Massey and collaborators, most recently published in Slesnick et al.\ 2002,
626: based on earlier studies by Conti (1988) for the O stars and
627: Humphreys \& McElroy (1984) for the B stars.
628: These values for absolute magnitude are comparable
629: to what has been typically used in the literature, dating back
630: as far as Walborn (1973) and are within a tenth of a magnitude of
631: those quoted in Vacca et al.\ (1996).
632: Figure 8 shows the distance modulus of each
633: star from the subsample (35 dwarf stars) as a function of spectral type
634: (converted to T$_{eff}$). The average of all 35 stars yields a distance
635: modulus of 11.16. This is very similar to what was
636: determined in MT91 (DM = 11.20). Some structure is seen
637: in Fig.\ 8, however. While the O stars (Log T$_{eff} > 4.50$) give
638: a consistent distance, a slight increase is seen in the predicted
639: distance modulus with the cooler stars. If just the O stars are
640: averaged, a slightly lower value of DM = 10.99 is obtained.
641:
642:
643: \begin{figure}
644: \epsscale{0.7}
645: \plotone{f8.ps}
646: \caption{The distance modulus calculated using the spectral type, effective temperature
647: and absolute magnitude relations presented in Slesnick, Hillenbrand \& Massey
648: (2002).\label{fig8}}
649: \end{figure}
650:
651:
652: The distance to Cyg~OB2 was also calculated using the Hipparcos derived
653: absolute magnitudes given by Wegner (2000). The average value for the
654: distance modulus, DM = 10.07, and shown in Fig.\ 9, is considerably
655: closer than originally given by MT91. Could the distance modulus
656: derived from the Hipparcos values be correct? Comparisons to nearly every
657: other study of absolute magnitude as a function of spectral type in
658: OB stars in the last few decades shows the Wegner values for absolute
659: magnitude to be
660: significantly lower, typically by a full magnitude in mid and
661: late O and early B stars. This strongly suggests that the Hipparcos
662: distances have a strong ``near'' bias not fully appreciated. Fig.\ 9 shows
663: a weak but significant dependency of spectral type with average distance,
664: in the same sense as is seen in Fig.\ 8. Such a dependency would be introduced if
665: the age of the stars used in the calibration is not taken into consideration.
666: Because Cyg OB2 is only a few million years old, the cooler dwarf stars
667: would be found exceedingly close to the ZAMS. Typical late-O and
668: early-B dwarfs used for the calibration of $M_V$ exist in clusters
669: less massive and typically older than Cyg OB2. While these older dwarf
670: stars may have the same spectral type and luminosity class, they represent
671: slightly more massive stars with higher intrinsic luminosities than our
672: near-ZAMS B stars. This effect, also discussed by Walborn (2002), is
673: borne out in the increased DM measured as a function of spectral type in
674: both the Slesnick et al.\ (2002)
675: absolute magnitudes (Fig.\ 8) and the Hipparcos values (Fig.\ 9).
676:
677:
678: \begin{figure}
679: \epsscale{0.7}
680: \plotone{f9.ps}
681: \caption{The distance modulus calculated from every star listed in Table 4 using
682: absolute magnitudes determined from Hipparcos (Wegner 2000) with effective
683: temperature scales of Slesnick et al.\ (2002).
684: \label{fig9}}
685: \end{figure}
686:
687:
688: \subsubsection{Effective Temperatures of OB stars}
689:
690: Fig.\ 8 and 9 demonstrate that some consideration of the very young age of
691: the Cyg OB2 cluster needs to be considered in order to properly determine
692: its distance modulus. If Cyg OB2 is to be fit to a very young isochrone,
693: there are just two things that can change in Fig.\ 8 and 9 to make a
694: better fit. Either
695: the absolute magnitudes are correct and the temperature scales needs
696: to be changed or
697: the other way around. This then introduces a second calibration which is
698: equally important to estimating distance modulus: the stellar effective
699: temperature scale.
700: Knowing a star's effective temperature is crucial in estimating its absolute
701: luminosity, particularly in theoretical models. Perhaps the most widely used
702: reference to this relationship is given by Vacca et al.\ (1996)
703: based on calculations using plane parallel, pure hydrogen and helium,
704: non-LTE atmosphere models. However, several
705: independent groups have recently shown that more sophisticated atmospheric models lead
706: to significantly lower effective temperatures in the O-type stars, both in
707: the dwarf stars (Martins et al.\ 2002; Markova et al.\ 2003) and the
708: giant and supergiant O supergiants (Bianchi \& Garcia 2002; Herrero, Puls
709: \& Najarro 2002). The shift in the spectral-type effective-temperature scale
710: is most pronounced in mid-O dwarf stars, and diminishes to near zero offset
711: from the Vacca et al.\ estimates of effective
712: temperatures for early-B stars. The implications of a new, lower effective
713: temperature scale, as discussed in Bianchi \& Garcia (2002), includes a
714: lower luminosity, thus lower mass, when compared to evolutionary tracks.
715: Since the typical goal in studying most OB clusters is to compare the stars
716: to theoretical evolutionary tracks, a lowering of the O star temperature scale
717: will have an immediate effect on results for age, mass and distance to OB
718: clusters.
719:
720:
721:
722: \begin{figure}
723: \epsscale{0.7}
724: \plotone{f10.ps}
725: \epsscale{0.8}
726: \caption{The distance modulus calculated using the effective temperature
727: relations of Martins, Schearer and Hillier (2002) with absolute magnitudes
728: scaled to match a 2 million year isochrone.
729: \label{fig10}}
730: \end{figure}
731:
732: \begin{figure}
733: \epsscale{0.7}
734: \plotone{f11.ps}
735: \caption{The distance modulus calculated from every star listed in Table 4 using
736: the effective temperature scale of Slesnick et al.\ (2002) with the absolute magnitudes
737: forced to match a 2 million year isochrone.
738: \label{fig11}}
739: \end{figure}
740:
741:
742: It is worth noting,
743: the reduced stellar effective temperatures as proposed by recent
744: theoretical arguments appear drastic only when these new values are compared to
745: Vacca et al.\ (1996). As discussed by Massey \& Hunter (1998) and Massey et
746: al. (2000), the Vacca et al.\ values
747: represented a significant upward trend from previous effective temperature
748: scales (Conti 1973; Humphreys \& McElroy 1984; Chlebowski \& Garmany 1991).
749: In fact, the Vacca et al.\ absolute magnitudes, bolometric
750: corrections, and many other calibrations are all highly generous with respect
751: to previous estimates of all these values. Recent studies of young open clusters
752: in the Large Magellanic Cloud (LMC) confirm this. Heydari-Malayeri et al.\ (2003) showed
753: a marked discrepancy between the high luminosities suggested by Vacca et al.
754: (1996) for their cluster stars over those they derived
755: directly from photometry and a known LMC distance. A comparison of the
756: effective temperature and spectral type calibrations from the previous
757: decade (most notably Humphreys \& McElroy 1984) give good
758: agreement, though not entirely coincident, with the dwarf model atmosphere
759: temperatures proposed by Martins et al.\ (2002).
760:
761:
762: Our ultimate goal in studying this cluster is to place the stars
763: onto the HR diagram to determine distance, age, and stellar masses.
764: It is reasonable to assume the best fit would be
765: one where the 35 dwarf stars line up along a single isochrone
766: on the HR diagram, provided the stars were predominately co-eval
767: in formation.
768: Using the new spectral type T$_{eff}$
769: calibration given by Martins et al., we employed the Z = 0.02,
770: 2x mass loss isochrones of LeJeune \& Schaerer (2001) to predict
771: total luminosity as a function of temperature for the dwarf stars.
772: An isochrone of 2 million years was chosen, though the cluster clearly
773: shows an age spread of from 1 to 3 million years (see \S 4.4.1).
774: Using the bolometric
775: corrections of Slesnick et al.\ (2002), this gave a set of absolute V
776: magnitudes as a function of effective temperature. For the early B-stars,
777: a seamless extrapolation from the Martins et al. values was created
778: based on T$_{eff}$ scales given by Humphreys \& McElroy
779: (1984) and Chlewbowski \& Garmany (1991). A third plot, showing the distance
780: modulus as a function of T$_{eff}$ using the absolute magnitudes
781: determined in this way is shown in Figure 10. The temperature dependence
782: seen in Figs.\ 8 and 9 is now entirely gone. Not surprisingly,
783: the average distance, DM = 10.44, is less than previous measures.
784: The reduced stellar temperatures
785: leads to a reduced absolute luminosity for O dwarf stars. Distances
786: predicted using O dwarf stars with such low absolute magnitudes can be
787: tested against other distance measures, such as
788: to the Large Magellanic Cloud. Here, OB stars are routinely observed
789: (Walborn \& Blades 1997; Massey, Waterhouse \& DeGioia-Eastwood 2000,
790: Heydari-Malayeri et al.\ 2003) and there are independent checks on
791: distance to derive the absolute magnitudes.
792:
793:
794: \subsubsection{Absolute Magnitudes for ZAMS OB Stars}
795:
796: If the temperature scale for OB stars as given in Figs.\ 8 and 9 is
797: correct, we can improve the fit of Fig.\ 8 based on Slesnick et al.\
798: (2002) effective
799: temperatures by making small adjustments to the absolute magnitude
800: scale. Again, the Z = 0.02, 2x mass loss isochrones of LeJeune
801: \& Schaerer (2001) where used to provide a smooth predicted total
802: luminosity as a function of temperature for the dwarf sample.
803: An isochrone of 2 million years was chosen. Using the bolometric
804: corrections of Slesnick et al.\ (2002) with the original effective
805: temperature scale of MT91 (also listed in Slesnick, but these are
806: based primarily on Humphreys \& McElroy
807: 1984, and Conti 1988), this provides a set of absolute
808: V magnitudes. This final set of spectral type, effective temperature
809: and absolute V magnitude, $M_V$, used for the dwarf sub-sample is
810: listed in Table 3. The DM value
811: of each star is shown as a function of effective temperature
812: (spectral type) in Fig.\ 11 and is also given in Table 3. The average
813: distance modulus determined in this way was found to be 10.80.
814: Fig.\ 11 shows a weak trend not yet seen. The hotter the star,
815: the greater the predicted distance modulus, quite opposite to what
816: was seen in Fig.\ 8 and 9. One possible explanation is that the
817: cluster is older than 2 million years, and the hotter stars are
818: further from the ZAMS than we've estimated. However, its clear based
819: on the presence of the extremely massive O3 If, and other early O
820: supergiants, that the cluster must indeed be quite young. Another explanation
821: is that the assumed effective temperatures of the mid-O stars is too
822: hot, giving them too great a luminosity and thus placing them at
823: a further distance. It will be possible to show in the next section
824: that explanation number one is ruled out. Explanation number two
825: lends a bit of support to the notion that traditional temperature
826: scales for mid-O stars are a bit too high, even after coming down from
827: the generous temperature scale of Vacca et al.\ (1996).
828:
829:
830: \subsubsection{Final word on the Distance to Cyg OB2}
831:
832: Using the traditional set of absolute magnitudes and stellar
833: effective temperatures for Cyg OB2 yields a distance modulus for the
834: cluster which is consistent with previous studies (MT91, Torres-Dodgen
835: et al.\ 1991). However, weak but significant dependences of the
836: distance calculated with the stellar spectral type are shown in many
837: of our figures (Figs.\ 8, 9, and 11 in particular), indicating possible
838: room for improvement. It is of interest to note that any attempt
839: at better fitting the DM values as a function of spectral type
840: ultimately leads to a closer distance for Cyg OB2. It is perhaps
841: premature to boldly accept the
842: much closer distance for Cyg OB2 suggested by Fig.\ 10 using the
843: Martins et al.\ (2002) derived values for effective temperature.
844: Beyond the possible inconsistencies with other measures of OB
845: distances as mentioned above, there are significant ramifications
846: to applying a new, closer distance to the Cyg OB2 cluster. Perhaps
847: the most important is a close distance reduces the luminosity of the
848: early O supergiants in Cyg OB2. Among the most important and best studied
849: of the Cyg OB2 supergiants are the O3 If$^*$ star, IV Cyg \# 7 ([MT91] 457),
850: and the O5 If stars, IV Cyg \#11, \#8A, and \#8C ([MT91] 734, 465, and 483,
851: respectively). In a recent paper, one in a series of papers studying
852: the Cyg OB2 supergiant stars, Herrero et al.\ (2002) assume a distance
853: of 1700 pc (the MT91 DM value of 11.2) for deriving the luminosity of
854: these stars and in modeling their atmospheres and mass loss. Moving
855: the distance to just over 1200 pc (DM=10.44) reduces the intrinsic luminosity
856: of the O supergiant stars by a factor of two! This low a luminosity is
857: simply not consistent with current wind theory because of the difficulty
858: driving the mass loss seen in the Cyg OB2 supergiant stars if their
859: luminosities were reduced a factor of two. However, a
860: DM = 10.80 is not such a problem. The derived luminosities of the
861: before mentioned supergiants, O3 If$^*$ star, IV Cyg \# 7 ([MT91] 457),
862: and the O5 If stars, IV Cyg \#11, \#8A, and \#8C, (MT[91] 734, 465, and 483,
863: respectively) using a DM = 10.80 is given in Table 5. For these stars,
864: we derived $Log \frac{L}{L_{\odot}}$ = 5.98, 6.06, 6.22, and 5.83,
865: respectively. Herrero et al.\ (2002), derived $Log \frac{L}{L_{\odot}}$ =
866: 5.91, 5.92, 6.19, and 5.66 respectively. Moreover, the luminosities
867: calculated by Herrero et al.\ (2002) for stars [MT91] 632, 217, and
868: 83 of $Log \frac{L}{L_{\odot}}$ = 5.77, 5.41, and 4.85, are a very good
869: match the values given in Table 5 for these stars: 5.80, 5.83, and
870: 4.64, respectively. For the remainder of this work,
871: a distance modulus of 10.80 will be used for the Cyg OB2 cluster.
872:
873:
874: \subsection{The HR Diagram for Cyg OB2}
875:
876:
877: A new HRD for Cyg OB2 using the distance modulus of 10.80 is shown in Figure 12.
878: Here it can be seen how the location of the dwarf stars are well fit to a 2 million
879: year old isochrone, as given by LeJeune \& Schaerer (2001), with a spread in
880: age of between 1 and 3 million years old. The luminosity
881: spread seen in the B1.5 and B2 dwarfs looks to be greater and in the direction of
882: being over-luminous than is seen for the O and earlier B dwarfs. There may be
883: increased contamination from foreground stars at these dwarf star masses which
884: are slightly lower than their hotter dwarfs.
885:
886:
887: \begin{figure}
888: \epsscale{0.8}
889: \plotone{f12.ps}
890: \caption{A revised HR diagram for Cyg OB2 assuming the new distance of 1500 pc
891: (DM=10.8). Newly identified stars from this study are identified with an asterisk
892: following their spectral types. The distance of 1500 pc
893: was determined through main sequence fitting of the previously known (none of those
894: found in this study) O7.5 through B1 dwarfs to a 2
895: million year old isochrone using effective temperatures given by Slesnick
896: et al.\ (2002). See Fig.\ 11.
897: \label{fig12}}
898: \end{figure}
899:
900:
901:
902: \subsubsection{Age of the Cyg OB2 Cluster}
903:
904: As first stated in MT91, a reasonably well-defined main sequence is seen in the
905: HRD of Cyg OB2. However, contamination is present in the form of several
906: evolved giant and supergiant stars (Fig. 12). This indicates contamination
907: from a non-coeval population within the Cyg OB2 sample region. Despite this
908: contamination, we can concentrate on the dwarf stars to estimate the approximate
909: age of the primary constituents of the cluster. The presence the O5 and O5.5
910: dwarf stars indicates an age not greater than 1 to 1.5 million years old.
911: However, the O5 dwarf (A37) comes from the CPR2002 study and has yet to
912: be confirmed
913: as being part of the main Cyg OB2 cluster. Furthermore, the luminosity of
914: the O5.5 V ([MT91] 516), puts it far from the main sequence (or possibly at a
915: nearer distance), making it also suspect as a reliable age measure for the bulk of
916: the cluster. A well behaved population of dwarf stars is found starting
917: at around O6 and O7 and by O7.5 and O8 there exists a particularly numerous
918: sample of dwarf stars. These mid-O dwarfs put a strong upper limit on the age
919: of the cluster of about 3 million years. An age of 2 million years, with a spread
920: of perhaps one million years, represents a reasonable age for the bulk of
921: the Cyg OB2 cluster based on its most massive dwarf stars. Indeed, a
922: young age of not more than a few million years is needed to
923: explain several very high luminosity blue supergiant stars (such as Cyg OB2 \#7,
924: the O3 If) and the possible Wolf-Rayet star members, the positions of which are
925: also shown in Fig.\ 12. The Wolf-Rayet stars where placed on
926: the HRD in Fig.\ 12 based on observations of the stars apparent magnitude
927: and line of sight extinction as measured by
928: Schmutz \& Vacca (1991) for WR145; Dougherty
929: et al.\ (2000) for WR146 and Massey et al.\ (2001) and references therein
930: for WR~144. The temperature of the three WR stars were simply assumed to be 50,000 K.
931: The temperature, apparent magnitude
932: and line of sight extinction of MWC349A was estimated by Hofmann et al.\ (2002).
933: Citing proximity arguments,
934: Massey et al.\ (2001) dismissed the Van der Hucht et al.\ (1981) claim
935: that all three WR stars where members of Cyg OB2. Only WR~144
936: is near the optically distinct cluster as first studied by MT91. However,
937: if a new extended radius for the Cyg OB2 cluster is to be considered, then
938: WR~145, WR~146 and possibly MWC~349A should be re-considered as possible
939: members.
940:
941: \subsubsection{Mixed-aged Stars towards Cyg OB2}
942:
943: Let's consider the possibility of non-coeval members appearing in the HRD shown
944: in Fig.\ 12.
945: Over the temperature range Log T$_{eff}$ = 4.30 to 4.37, four early
946: B stars lie well above the main sequence, A39 (B2 V), 642 (B1 III),
947: 575 (B1.5V), and 793 (B1.5III?) in Fig.\ 12. Two show near infrared
948: excesses (575, 793), consistent with them being Be stars. A39 is likely
949: foreground. An additional 11 giant and supergiant stars, lying mostly
950: in the temperature range Log T$_{eff}$ = 4.40 to 4.53, are undeniably
951: older than the few million years we've assigned to the dwarf stars.
952: Of these eleven older, evolved stars, over half of them
953: originated from the CPR2002 survey (A32, A23, A27, A29, A36, A41).
954: The other five, S03, and MT[91] 632, 601, 642, and 83 where from the original
955: optical survey. The O5 V((f)) star identified by CPR2002 as A37, seems to be
956: two young for the remainder of the Cyg OB2 cluster, indicating an age
957: below 1 million years. Summing things up, 7 of the 14 new stars identified
958: by CPR2002 show ages inconsistent with the remainder of the cluster, and
959: are unlikely members of Cyg OB2. Extending the search
960: for cluster members to larger distances from the central core has lead to
961: an increase in contamination from non-member stars. The IMF values and
962: stellar counts made by MT91 still remain the best estimates for the cluster
963: until the remaining OB candidates identified by CPR2002 are observed.
964: However, it is clear that in an effort to move towards greater completeness
965: of the cluster members, we have also moved towards a greater incidence of
966: contamination. Without reliable measures to establish membership,
967: this will be a chronic problem for all studies attempting deep photometric
968: studies of distant clusters within our galaxy.
969:
970: \begin{deluxetable}{llccccc}
971: \tablewidth{0pt}
972: \tablecaption{Cyg OB2 Stars with Spectral Types}
973: \tablehead{
974: \colhead{Star} &
975: \colhead{SpTy} &
976: \colhead{T$_{eff}$} &
977: \colhead{mv} &
978: \colhead{Av} &
979: \colhead{BC} &
980: \colhead{Log$\frac{L}{L_{\odot}}$}
981: }
982: \startdata
983: 005 & O6V & 43551.2 & 12.93 & 6.19 & -3.73 & 5.13 \\
984: 021 & B2? & 19952.6 & 13.74 & 4.41 & -2.00 & 3.29 \\
985: 059 & O8.5V & 37239.2 & 11.18 & 5.21 & -3.38 & 5.26 \\
986: 070 & O9V & 35975.0 & 12.99 & 6.96 & -3.31 & 5.20 \\
987: 083 & B1I & 26302.7 & 10.64 & 4.03 & -2.67 & 4.64 \\
988: 138 & O8.5I & 34435.0 & 12.26 & 6.73 & -3.25 & 5.36 \\
989: 145 & O9.5V & 34673.7 & 11.62 & 4.23 & -3.23 & 4.62 \\
990: 169 & B1.5V & 22387.2 & 13.90 & 4.51 & -2.27 & 3.37 \\
991: 174 & B1.5V & 22387.2 & 12.55 & 4.46 & -2.27 & 3.89 \\
992: 187 & B0.5V & 28183.8 & 13.24 & 5.40 & -2.84 & 4.22 \\
993: 213 & B0Vp & 31622.8 & 11.95 & 4.20 & -3.14 & 4.38 \\
994: 215 & B1V & 25409.7 & 12.97 & 4.03 & -2.58 & 3.68 \\
995: 217 & O7III & 39902.5 & 10.22 & 4.42 & -3.59 & 5.41 \\
996: 227 & O9V & 35975.0 & 11.47 & 4.73 & -3.31 & 4.92 \\
997: 250 & B1V & 25409.7 & 12.88 & 4.08 & -2.58 & 3.73 \\
998: 258 & O8V & 38459.2 & 11.10 & 4.49 & -3.43 & 5.04 \\
999: 259 & B0.5V & 28183.8 & 11.42 & 3.83 & -2.84 & 4.32 \\
1000: 292 & B1V & 25409.7 & 13.08 & 5.40 & -2.58 & 4.18 \\
1001: 299 & O8V & 38459.2 & 10.84 & 4.61 & -3.43 & 5.19 \\
1002: 300 & B1?V & 25409.7 & 13.05 & 4.24 & -2.58 & 3.73 \\
1003: 304 & B5Ie & 13182.6 & 11.46 & 10.79 & -1.83 & 6.38 \\
1004: 317 & O8V & 38459.2 & 10.66 & 4.73 & -3.43 & 5.31 \\
1005: 339 & O8.5V & 37239.2 & 11.60 & 4.94 & -3.38 & 4.99 \\
1006: 358 & B2? & 19952.6 & 14.81 & 6.50 & -2.00 & 3.69 \\
1007: 376 & O8V & 38459.2 & 11.91 & 4.96 & -3.43 & 4.91 \\
1008: 378 & B0V & 31622.8 & 13.49 & 7.14 & -3.14 & 4.94 \\
1009: 390 & O8V & 38459.2 & 12.95 & 6.72 & -3.43 & 5.20 \\
1010: 395 & B1.5V & 22387.2 & 14.14 & 5.94 & -2.27 & 3.85 \\
1011: 403 & B1.5V & 22387.2 & 12.94 & 5.53 & -2.27 & 4.17 \\
1012: 417 & O4III & 48194.8 & 11.55 & 7.15 & -4.01 & 6.18 \\
1013: 421 & O9.5V & 34673.7 & 12.86 & 6.83 & -3.23 & 5.16 \\
1014: 425 & B0V & 31622.8 & 13.62 & 6.78 & -3.14 & 4.74 \\
1015: 426 & B0V & 31622.8 & 14.05 & 6.60 & -3.14 & 4.50 \\
1016: 429 & B0V & 31622.8 & 12.98 & 5.64 & -3.14 & 4.54 \\
1017: 431 & O5If & 44771.3 & 10.80 & 6.99 & -3.76 & 6.33 \\
1018: 448 & O6V & 43551.2 & 13.61 & 7.57 & -3.73 & 5.41 \\
1019: 455 & O8V & 38459.2 & 12.92 & 6.38 & -3.43 & 5.07 \\
1020: 457 & O3If & 50699.1 & 10.49 & 5.44 & -4.11 & 5.98 \\
1021: 462 & O6.5III & 41304.8 & 10.33 & 5.24 & -3.66 & 5.73 \\
1022: 465 & O5.5Ia & 43351.1 & 9.06 & 5.06 & -3.63 & 6.22 \\
1023: 467 & B1V & 25409.7 & 13.43 & 5.04 & -2.58 & 3.90 \\
1024: 470 & O9.5V & 34673.7 & 12.50 & 5.19 & -3.23 & 4.65 \\
1025: 473 & O8.5V & 37239.2 & 12.02 & 5.29 & -3.38 & 4.96 \\
1026: 480 & O7.5V & 39810.7 & 11.88 & 5.82 & -3.48 & 5.30 \\
1027: 483 & O5If & 44771.3 & 10.08 & 5.01 & -3.76 & 5.83 \\
1028: 485 & O8V & 38459.2 & 12.06 & 5.58 & -3.43 & 5.10 \\
1029: 507 & O8.5V & 37239.2 & 12.70 & 5.61 & -3.38 & 4.81 \\
1030: 515 & B1V & 25409.7 & 14.66 & 6.97 & -2.58 & 4.18 \\
1031: 516 & O5.5V & 44874.5 & 11.84 & 7.63 & -3.84 & 6.18 \\
1032: 531 & O8.5V & 37239.2 & 11.58 & 5.57 & -3.38 & 5.25 \\
1033: 534 & O7.5V & 39810.7 & 13.00 & 6.54 & -3.48 & 5.14 \\
1034: 555 & O8V & 38459.2 & 12.51 & 6.57 & -3.43 & 5.31 \\
1035: 556 & B1Ib & 26302.7 & 11.01 & 6.81 & -2.67 & 5.61 \\
1036: 575 & B1.5V & 22387.2 & 13.41 & 7.58 & -2.27 & 4.79 \\
1037: 588 & B0V & 31622.8 & 12.40 & 6.02 & -3.14 & 4.92 \\
1038: 601 & O9.5III & 32961.0 & 11.06 & 6.01 & -3.16 & 5.50 \\
1039: 605 & B0.5V & 28183.8 & 11.78 & 4.23 & -2.84 & 4.34 \\
1040: 611 & O7Vp & 41020.4 & 12.77 & 5.61 & -3.55 & 4.90 \\
1041: 632 & O9.5I & 31477.5 & 9.88 & 5.70 & -3.00 & 5.80 \\
1042: 642 & B1III & 23550.5 & 11.78 & 5.83 & -2.39 & 4.80 \\
1043: 646 & B1.5?V & 22387.2 & 13.34 & 5.13 & -2.27 & 3.84 \\
1044: 692 & B0V & 31622.8 & 13.61 & 5.69 & -3.14 & 4.31 \\
1045: 696 & O9.5V & 34673.7 & 12.32 & 5.85 & -3.23 & 4.99 \\
1046: 716 & O9V & 35975.0 & 13.50 & 6.10 & -3.31 & 4.65 \\
1047: 734 & O5If & 44771.3 & 10.04 & 5.54 & -3.76 & 6.06 \\
1048: 736 & O9V & 35975.0 & 12.79 & 5.46 & -3.31 & 4.68 \\
1049: 745 & O7V & 41020.4 & 11.91 & 5.30 & -3.55 & 5.12 \\
1050: 771 & O7V & 41020.4 & 12.06 & 7.00 & -3.55 & 5.73 \\
1051: 793 & B1.5III? & 22908.7 & 12.29 & 6.13 & -2.33 & 4.69 \\
1052: S03 & O8.5III & 35727.3 & 10.22 & 6.01 & -3.30 & 5.92 \\
1053: S73 & O8V & 38459.2 & 0.00 & 6.04 & -3.43 & 5.28 \\
1054: A20 & O8II & 37153.5 & 0.00 & 7.77 & -3.37 & 5.88 \\
1055: A23 & B0.7Ib & 26915.3 & 0.00 & 7.28 & -2.73 & 5.55 \\
1056: A27 & B0Ia & 28840.3 & 0.00 & 7.06 & -2.88 & 5.75 \\
1057: A29 & O9.7Iab & 30199.5 & 0.00 & 6.92 & -3.00 & 5.48 \\
1058: A32 & O9.5IV & 34673.7 & 0.00 & 6.73 & -3.23 & 5.43 \\
1059: A34 & B0.7Ib & 26915.3 & 0.00 & 5.96 & -2.73 & 5.21 \\
1060: A36 & B0Ib & 28840.3 & 0.00 & 6.51 & -2.88 & 5.48 \\
1061: A37 & O5V & 46131.8 & 0.00 & 6.77 & -3.95 & 5.51 \\
1062: A39 & B2V & 19952.6 & 0.00 & 5.89 & -2.00 & 4.42 \\
1063: A41 & O9.7II & 31622.8 & 0.00 & 5.97 & -3.14 & 5.29 \\
1064: A42 & B0V & 31622.8 & 0.00 & 5.83 & -3.14 & 4.72 \\
1065: A44 & B0.5IV & 28183.8 & 0.00 & 5.04 & -2.84 & 4.62 \\
1066: A45 & B0.5V & 28183.8 & 0.00 & 4.64 & -2.84 & 4.53 \\
1067: A46 & O7V & 41020.4 & 0.00 & 4.56 & -3.55 & 5.19 \\
1068: \enddata
1069: \end{deluxetable}
1070:
1071:
1072: \section{DISCUSSION}
1073:
1074:
1075:
1076: Comer\'{o}n et al.\ (2002) estimate that they have uncovered between
1077: 90 and 100 O-type stars or closely related objects in the Cyg OB2
1078: association. Indeed, our results suggest the numerous OB candidates
1079: they have identified are likely to be OB stars. The fraction of O-stars
1080: in our sample of 14 newly identified OB stars is lower than they had
1081: predicted for the candidate sample (77\%), though if we include the
1082: B supergiants which {\sl used} to be O stars, we get better agreement
1083: (60\%) when one considers the sample size.
1084:
1085: However, the high fraction of suspected non-members found in this
1086: study of 14 new OB stars would suggest that many if not most of the
1087: candidate OB stars identified by Comer\'{o}n et al.\ (2002) and
1088: still yet to be confirmed will also be non-members.
1089: The very extended spatial distribution of the CPR2002 sample, as
1090: shown in Fig.\ 4, also supports this prediction. The Cyg OB2 cluster
1091: may be more extensive and contain more O stars than previously thought,
1092: however it now seems less likely that
1093: a concentrated effort would yield over 100 O stars
1094: as predicted by Kn\"{o}dlseder (2000). Based on this definition
1095: used by Kn\"{o}dlseder (2000), Cyg OB2 would not be classified
1096: as a super star cluster. Naturally, spectroscopic observations of the
1097: remaining OB candidates listed by Comer\'{o}n et al.\ are needed to
1098: resolve this debate. Moreover, we can not rule out that deeper
1099: near-infrared investigations concentrated on the center of the the
1100: Cyg OB2 cluster region may reveal the underlying massive cluster
1101: the Kn\"{o}dlseder (2000) near-infrared study suggests.
1102:
1103:
1104: \subsection{Does the Milky Way contain Super Star Clusters?}
1105:
1106: The term ``super star cluster'' was first used to describe very
1107: luminous young star clusters in nearby late type galaxies (Hodge 1961).
1108: Early examples include 30 Dor in the Large Magellanic Cloud and NGC 330
1109: in the Small Magellanic Cloud, each harboring tens of thousands of stars.
1110: No such galactic counterparts, outside of the very old globular clusters,
1111: were known to exist in the galaxy. Recently, the term has been used to
1112: describe quite a number of galactic clusters, first the Arches and
1113: Quintuplet clusters near the galactic center, and now Cygnus OB2
1114: (Kn\"{o}dlseder 2000) and Westerlund 1
1115: (Negueruela \& Clark 2003). Based on the proximity of the last two, and
1116: assuming a similar distribution within the solar circle, Kn\"{o}dlseder
1117: et al. (2002) predict there may be as many as 100 similar clusters
1118: in the Galaxy. Interestingly, evidence for such a large population
1119: of very massive clusters, is not unprecedented.
1120: Extrapolating the cluster luminosity function of our
1121: galaxy, van den Bergh \& Lafontaine (1984) predicted a total of $\sim$
1122: 10$^2$ clusters with $M_V$ = -11, the absolute visual magnitude of 30 Dor!
1123: A continued extrapolation of our galaxy's cluster luminosity function
1124: predicts our galaxy to contain one cluster with $M_V$ = -12 (Larsen 2002).
1125:
1126: Van den Bergh \& Lafontaine (1984) found it hard to believe clusters with
1127: such mass could be contained within our galaxy. Perhaps for this reason,
1128: it's been assumed that our galaxy's cluster luminosity function steepened
1129: over the range $-11 < M_V < -8$ (Larsen 2002). However, there are many
1130: factors contributing to massive clusters being hidden from view. Most
1131: important is the fact that the most massive open clusters are also among the
1132: youngest (see Fig.\ 6 of Larsen \& Richtler 1999). This is because
1133: very massive and presumably more extended clusters are preferentially
1134: destroyed over lower mass clusters by dynamical friction in a relatively
1135: short timescales (10$^9$ yrs; de Grijs et al.\ 2003). Recently, Portegies
1136: Zwart et al.\ (2001; 2002) showed the exceedingly fast dynamical evolution
1137: and fatal disruption of massive clusters formed in the inner 200 pc
1138: of the galaxy, predicting their apparent loss of detection in just
1139: 5 million years. They predict the central portion of our galaxy
1140: could easily harbor as many as 50 clusters with properties similar
1141: to the Arches or Quintuplet massive cluster systems. Thus, our
1142: galaxy's most massive clusters are severely affected by
1143: line of sight (due to their close proximity to the galactic plane)
1144: and local extinction (due to their youth) severely hindering our
1145: efforts to find such objects. Given the rather low star formation
1146: rate (SFR) seen at the sun's galactic radius, and the known
1147: correlation between SFR and cluster luminosities
1148: (Larsen 2002 and references therein), one would not expect to
1149: find examples of very massive star clusters locally. An absence of massive
1150: clusters in the well sampled, but small region out at the galactic
1151: radius of our Sun can not be used as evidence for the absence of massive
1152: clusters elsewhere. Finally, the known absence of massive clusters
1153: in M31 while considering they might yet exist in the Milky Way is reasonable
1154: considering M31's earlier Hubble type (Kennicutt \& Chu 1988).
1155:
1156:
1157:
1158: {\subsection{How might galactic Super Star Clusters be found?}
1159:
1160: Historically, radio surveys of the galactic plane have been successful to
1161: identify very young and massive galactic HII regions (Westerhout 1958), and
1162: to trace the structure of our galaxy (Georgelin \& Georgelin 1976). A recent
1163: series of papers by Blum and collaborators (Blum, Damineli, \& Conti 1999;
1164: Blum, Conti, \& Damineli 2000, Blum, Damineli, \& Conti 2001) have further
1165: investigated some of the largest of these radio identified HII regions.
1166: Referred to as Giant HII regions (GHII), these HII regions have Lyman
1167: continuum fluxes more than 10 times that of Orion, or about 10 O stars.
1168: Their near-infrared studies were able to spectroscopically confirm OB stars
1169: in all three GHII regions. However, emission line sources consistent with very
1170: massive young stellar objects were also found to lie among the brightest
1171: sources in each of the three clusters, a clear indication of the very
1172: young age of clusters (not more than perhaps 1 million years old) found
1173: via radio surveys. Curiously, the spectroscopic distances of all
1174: three of the GHII regions studied by Blum and collaborators turned out
1175: to be {\sl nearer} than the radio measurements had estimated. This
1176: significantly reduced their intrinsic Lyman continuum flux, demoting two of
1177: them from being true GHII regions to simply large HII regions. Most important,
1178: their results indicate we are still biased to detecting HII regions on the
1179: near side of the galaxy, even when using radio surveys. Only a few true
1180: GHII regions found via radio surveys are confirmed to lie on the far side
1181: of the galaxy (such as W49A, see Conti \& Blum 2002).
1182:
1183:
1184: Neither Cyg OB2 nor Westerlund 1 are easily detected in the radio. This is
1185: an important point to recognize. Massive clusters older than a few million years
1186: will have been missed in galactic radio surveys. How do we find our Galaxy's
1187: super star clusters?
1188:
1189: Systematic infrared surveys to search for massive clusters with follow up
1190: near-infrared spectroscopic studies of their stellar members may be the
1191: only method. Sifting through deep near-infrared imaging surveys to search
1192: for new massive clusters will be an important direction for galactic studies
1193: in the next decade. Equally important will be determining the completeness
1194: limits for such objects, though it may only be possible within a few kpc from
1195: the Sun. It is important to know if massive clusters found in the disk share
1196: any of the extreme qualities of the
1197: very massive clusters found at the galactic center. To do so means uncovering
1198: and then investigating a significant number of disk super star clusters (if they
1199: exist) to get a statistically meaningful result on the characteristics of
1200: this new class of galactic objects.
1201:
1202: The Cyg OB2 cluster serves as an important first calibration. Because it has
1203: significant extinction, Kn\"{o}dlseder et al. (2002) was forced to investigate,
1204: isolate and characterize the cluster using near-infrared methods. However, the
1205: extinction is not so great that cross-checks with more traditional cluster
1206: studies, namely optical MK classification of many of its members, was
1207: ruled out. Combining near-infrared and optical observations to study Cyg OB2
1208: has given us the best of both methods and resulted in a clearer description of
1209: the cluster and member properties. The 100\% success rate, so far, of
1210: the CPR2002 study to select early-type stars from a dense field using very low
1211: resolution near-infrared measurements alone, gives great confidence in the
1212: ability to apply similar techniques to heavily
1213: reddened, near-infrared clusters now being uncovered. MK classification spectra
1214: are the most desirable, but for stars with A$_V > 8$ or so, the benefits of a
1215: more accurately determined optical spectral type are eclipsed by the need for
1216: a very large telescope to obtain such spectra. For these stars, and this
1217: would include the remaining 31 OB candidate stars listed by CPR2002 which were
1218: too extinguished for this study, stellar classification will need to be made at
1219: near-infrared wavelengths (Hanson et al.\ 1996).
1220:
1221:
1222:
1223:
1224: \section{CONCLUSIONS}
1225:
1226:
1227: New optical MK classification spectra have been obtained for 14 candidate OB stars from
1228: the Cyg OB2 cluster which were sifted and identified using near-infrared photometric
1229: and low resolution spectroscopic measurements. Our optical spectra confirm the
1230: early-type nature of these stars and lends support to Kn\"{o}dlseder's (2000)
1231: result that Cyg OB2 contains more early-type stars than previously thought. However,
1232: about half of the OB stars appear to be non-members based on their evolutionary age.
1233: For this and other reasons, the present O-star count, still under 100,
1234: does not yet allow Cyg OB2 to be classified as a super star cluster,
1235: as defined by Kn\"{o}dlseder et al. (2002).
1236:
1237: Our investigation of the Cyg OB2 cluster characteristics indicates that its distance
1238: may be slightly closer than previous work quoted (DM = 11.2; Massey \& Thompson
1239: 1992) when the very young age, and thus under-luminous nature of these stars (as
1240: compared to the more evolved clusters used in the derivation of dwarf star
1241: absolute magnitudes) is taken into consideration. A revised distance of DM = 10.8
1242: is predicted based on fitting 35 dwarf cluster members with spectral types
1243: between O7.5 and B1 to a 2 x 10$^6$ year isochrone. We also explored the newly
1244: presented effective temperature scale for dwarf O stars of Martins
1245: et al.\ (2002), using a similar fit of the dwarf population to
1246: a 2 x 10$^6$ year isochrone. The new lower temperatures and thus lower absolute
1247: luminosities (particularly for mid-O dwarf stars) gives a very close
1248: value for the distance to Cyg OB2 (DM = 10.4). This
1249: lends some urgency to further observational testing and investigation of the
1250: new lower effective temperature scale for O dwarf stars, as it looks to be
1251: strongly supported by many independent theoretical groups at this time. Observational
1252: studies of additional massive clusters containing a significant sample of OB
1253: dwarf stars would be useful in this regard.
1254:
1255: Our study has examined and confirmed the reliability of near-infrared
1256: broad band colors combined with low resolution survey spectra to accurately
1257: sift OB stars from a dense field population. It would appear evidence for
1258: very massive clusters in our galaxy is mounting, based both on direct observations
1259: of fairly massive clusters near the Sun, as well as extrapolations of our galaxy's
1260: cluster luminosity function. Neither optical nor radio searches are likely
1261: to be successful in
1262: locating these objects. However, numerous near-infrared clusters are currently
1263: being uncovered with 2MASS. For extinction values A$_V > 8$, determination of
1264: spectral types for the stars in these clusters will need to be obtained using
1265: near-infrared classification techniques. Great possibilities lie ahead for
1266: the discovery and precise study of very massive galactic OB clusters, uncovering
1267: very massive, possible, super star clusters currently hidden within the Galaxy.
1268:
1269:
1270:
1271: \acknowledgements
1272:
1273:
1274: We are grateful to Steward Observatory and The University of Arizona for
1275: their support of this program through their generous allowance of
1276: telescope time.
1277: Phil Massey gave the author extensive, critical comments
1278: on the manuscript and, along with Nolan Walborn, performed an independent
1279: check of the MK spectral types for the stars in this study.
1280: The author continues to benefit from enlightening discussions with
1281: Peter Conti. Finally, the referee provided
1282: insightful suggestions that greatly improved the quality of the manuscript.
1283: The Digitized Sky Surveys were produced at the Space Telescope Science
1284: Institute under U.S. Government grant NAG W-2166. Near-infrared photometry
1285: was obtained from The Two Micron All Sky Survey (2MASS), a joint project
1286: of the University of Massachusetts and the Infrared Processing and Analysis
1287: Center/California Institute of Technology, funded by the National Aeronautics
1288: and Space Administration and the National Science Foundation. This manuscript
1289: is based upon work supported by the National Science Foundation under
1290: Grant AST-0094050 to the University of Cincinnati.
1291:
1292:
1293:
1294: \begin{references}
1295:
1296: \reference{} Bianchi, L. \& Garcia, M. 2002, ApJ, 581, 610
1297:
1298: \reference{} Blum, R. D., Damineli, A., \& Conti, P.S. 1999, AJ, 117, 1392
1299:
1300: \reference{} Blum, R. D., Conti, P.S. \& Damineli, A. 2000, AJ, 119, 1860
1301:
1302: \reference{} Blum, R. D., Damineli, A., \& Conti, P.S. 2001, AJ, 121, 3149
1303:
1304: \reference{} Chlebowski, T., \& Garmany, C. D. 1991, ApJ, 368, 241
1305:
1306: \reference{} Comer\'{o}n, F., et al. 2002, A\&A, 389, 874
1307:
1308: \reference{} Conti, P. S. 1988, in O Stars and Wolf-Rayet Stars, ed. P. S. Conti \&
1309: A.B. Underhill (NASA SP-497), 121
1310:
1311: \reference{} Conti, P. S. \& Blum, R. D. 2002, ApJ, 564, 827
1312:
1313: \reference{} de Grijs, R., Basian, N. \& Lamers, H. J. G. L. M. 2003, ApJ, 583, L17
1314:
1315: \reference{} Dougherty, S.M., williams, P.M., \& Pollacco, D.L. 2000, MNRAS, 316, 143
1316:
1317: \reference{} Dutra, C. M., Bica, E. 2000, A\&A, 359, L9
1318:
1319: \reference{} Dutra, C. M., Bica, E. 2001, A\&A, 376, 434
1320:
1321: \reference{} Eckart, A., Ott, T., Genzel, R.\ 1999, A\&A, 352, L22
1322:
1323: \reference{} Figer, D. et al.\ 1999, ApJ, 525, 750
1324:
1325: \reference{} FitzGerald, M. P. 1970, A\&A, 4, 234
1326:
1327: \reference{} Georgelin, Y.M. \& Georgelin, Y.P. 1976, A\&A, 49, 57
1328:
1329: \reference{} Hanson, M. M., Conti, P. S., Rieke, M. J. 1996, ApJS, 107, 281
1330:
1331: \reference{} Hanson, M. M., Howarth, I. D., Conti, P. S. 1997, ApJ, 489, 698
1332:
1333: \reference{} Herrero, A., Puls, J., \& Najarro, F. 2002, A\&A, 396, 949
1334:
1335: \reference{} Heydari-Malayeri, M., Meynadier, F., \& Walborn, N. R. 2003, A\&A, 400, 923
1336:
1337: \reference{} Hodge, P. 1961, ApJ, 133, 413.
1338:
1339: \reference{} Hofmann, K.-H., Balega, Y., Ikhsanov, N.R., Miroshnichenko, A.S., \& Weigelt,
1340: G. 2002, A\&A, 395, 891
1341:
1342: \reference{} Humphreys, R. M. \& McElroy, D. G. 1984, ApJ, 284, 565
1343:
1344: \reference{} Hutchings, J. B. 1981, PASP, 93, 50
1345:
1346: \reference{} Ivanov, V. D., Borissova, J., Pessev, P., Ivanov, G. R., Kurtev, R. 2002,
1347: A\&A, in press
1348:
1349: \reference{} Jenniskens, P. \& Desert, F.-X., 1994, A\&A, 106, 39
1350:
1351: \reference{} Johnson, H. L. \& Morgan, W. W. 1954, ApJ 119, 344
1352:
1353: \reference{} Kennicutt, R. C. \& Chu, Y. 1988, AJ, 95, 720
1354:
1355: \reference{} Kn\"{o}dlseder, J. 2000, A\&A, 360, 539
1356:
1357: \reference{} Kn\"{o}dlseder, J. et al.\ 2002, A\&A, 390, 945
1358:
1359: \reference{} Larsen, S. S. 2002, ApJ, 124, 1393
1360:
1361: \reference{} Larsen, S. S. 1999, A\&A, 345, 59
1362:
1363: \reference{} Markova, N., Puls, J., Repoloust, T., \& Markov, H. 2003, in press
1364:
1365: \reference{} Martins, F., Schearer, D., Hillier, D. J. 2002, A\&A, 382, 999
1366:
1367: \reference{} Massey, P. \& Hunter, D., 1998, ApJ, 493, 180
1368:
1369: \reference{} Massey, P. \& Thompson, A. B., 1991, AJ, 101, 1408
1370:
1371: \reference{} Massey, P., Waterhouse, E., \& DeGioia-Eastwood, K. 2000, AJ, 119, 2214
1372:
1373: \reference{} Massey, P., Degioia-Eastwood, K., \& Waterhouse, E. 2001, AJ, 121, 1050
1374:
1375: \reference{} Morgan, W. W., Meinel, A. B., Johnson, H. M. 1954, ApJ 120, 506
1376:
1377: \reference{} Munch, L. \& Morgan, W. W. 1953, ApJ, 118, 161
1378:
1379: \reference{} Clark, J. S., Negueruela, I. 2002, A\&A, 396, L25
1380:
1381: \reference{} Negueruela, I., Clark, J. S. 2003, in a Massive Star
1382: Odyssey, from Main Sequence to Supernova, Proceedings IAU Symp.\ 212,
1383: eds.\ van der Hucht, K. A., Herrero, A., Esteban, C., in press.
1384:
1385: \reference{} Portegies Zwart, S. F., Makino, J., \& McMillan, S. L. W. 2001, ApJ, 546, L101
1386:
1387: \reference{} Portegies Zwart, S. F., Makino, J., McMillan, S. L. W. \& Hut, P. 2002, ApJ, 565, 265
1388:
1389: \reference{} Rieke, G. H., Lebofsky, M. J.\ 1985, ApJ, 288, 618
1390:
1391: \reference{} Russeil, D.\ 2003, A\&A, 397, 133
1392:
1393: \reference{} Schmutz, W., \& Vacca, W.D. 1991, A\&AS, 89, 259
1394:
1395: \reference{} Schulte, D. H. 1956, ApJ, 124, 530
1396:
1397: \reference{} Schulte, D. H. 1958, ApJ, 128, 41
1398:
1399: \reference{} Slesnick, C. L., Hillenbrand, L. A., Massey, P. 2002, ApJ 576, 880
1400:
1401: \reference{} Skrutskie, M., et al.\ 1997, in The Impact of Large Scale Near-IR
1402: Sky Surveys, ed. Garson, et al. (Dordrecht: Kluwer), 25.
1403:
1404: \reference{} Torres-Dodgen, A. V., Carroll, M., Tapia, M. 1991, MNRAS, 249, 1
1405:
1406: \reference{} van den Bergh, S. \& Lafontaine, A. 1984, AJ, 89, 1822
1407:
1408: \reference{} Walborn, N. R. \& Fitzpatrick, E. L. 1990, PASP, 102, 379
1409:
1410: \reference{} Walborn, N. R. \& Blades, J. C. 1997, ApJS, 112, 457
1411:
1412: \reference{} Walborn, N. R. 2002, AJ, 124, 507
1413:
1414: \reference{} Wegner, W. 2000, MNRAS, 319, 771
1415:
1416: \reference{} Westerhout, G. 1958, BAN, 14 215
1417:
1418: \end{references}
1419:
1420:
1421: \end{document}
1422:
1423:
1424: