Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.stsci.edu/science/preprints/prep1563/prep1563.pdf Äàòà èçìåíåíèÿ: Wed Nov 27 00:32:36 2002 Äàòà èíäåêñèðîâàíèÿ: Sat Dec 22 14:59:21 2007 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: m 104

Massive Stars in the Arches Cluster

12

Donald F. Figer3,4 , Francisco Na jarro5 , Diane Gilmore3 , Mark Morris6 , Sungso o S. Kim6 , Eugene Serabyn7 , Ian S. McLean6 , Andrea M. Gilbert8 , James R. Graham8 , James E. Larkin6 , N. A. Levenson4 , Harry I. Teplitz9, 10 figer@stsci.edu ABSTRACT We present and use new spectra and narrow-band images, along with previously published broad-band images, of stars in the Arches cluster to extract photometry, astrometry, equivalent width, and velo city information. The data are interpreted with a wind/atmosphere co de to determine stellar temperatures, luminosities, mass-loss rates, and abundances. We have doubled the number of known emission-line stars, and we have also made the first spectroscopic identification of the main sequence for any population in the Galactic Center. We conclude that the most massive stars are bona-fide Wolf-Rayet (WR) stars and are some of the most massive stars known, having Minitial >100 M , and pro di gious winds, M >10-5 M yr-1 , that are enriched with helium and nitrogen; with these identifications, the Arches cluster contains about 5% of all known WR stars in the Galaxy. We find an upper limit to the velo city dispersion of 22 km s-1 , implying an upper limit to the cluster mass of 7(104 ) M within a radius of 0.23 pc; we also estimate the bulk helio centric velo city of the cluster
3 4 5 6

Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218; figer@stsci.edu Department of Physics and Astronomy, Johns Hopkins University, Baltimore, MD 21218 Instituto de Estructura de la Materia, CSIC, Serrano 121, 29006 Madrid, Spain

Department of Physics and Astronomy, University of California, Los Angeles, Division of Astronomy, Los Angeles, CA, 90095-1562
7 8

Caltech, 320-47, Pasadena, CA 91125; eserabyn@huey.jpl.nasa.gov

Department of Astronomy, University of California, Berkeley, 601 Campbell Hall, Berkeley, CA, 947203411 Laboratory for Astronomy and Solar Physics, Code 681, Goddard Space Flight Center, Greenbelt MD 20771
10 9

NOAO Research Asso ciate


­2­ to be vcluster, +95 km s-1 . Taken together, these results suggest that the Arches cluster was formed in a short, but massive, burst of star formation about 2.5±0.5 Myr ago, from a molecular cloud which is no longer present. The cluster happens to be approaching and ionizing the surface of a background molecular cloud, thus pro ducing the Thermal Arched Filaments. We estimate that the cluster pro duces 4(1051 ) ionizing photons s-1 , more than enough to account for the observed thermal radio flux from the nearby cloud, 3(1049 ) ionizing photons s-1 . Commensurately, it pro duces 107.8 L in total luminosity, providing the heating source for the nearby molecular cloud, Lcloud 107 L . These interactions between a cluster of hot stars and a wayward molecular cloud are similar to those seen in the "Quintuplet/Sickle" region. The small spread of formation times for the known young clusters in the Galactic Center, and the relative lack of intermediate-age stars (age =107.0 to 107.3 yrs), suggest that the Galactic Center has recently been host to a burst of star formation. Finally, we have made new identifications of near-infrared sources that are counterparts to recently identified x-ray and radio sources. Subject headings: Galaxy: center -- techniques: spectroscopic -- infrared: stars

1.

Intro duction

The Arches cluster is an extraordinarily massive and dense young cluster of stars near the Galactic Center. First discovered about 10 years ago as a compact collection of a dozen or so emission-line stars (Cotera et al. 1992; Nagata et al. 1995; Figer 1995; Cotera 1995; Cotera et al. 1996), the cluster contains thousands of stars, including at least 160 O stars (Serabyn et al. 1998; Figer et al. 1999a). Figer et al. (1999a) used HST/NICMOS observations to estimate a total cluster mass ( 104 M ) and radius (0.2 pc) to arrive at an average mass density of 3(105 ) M pc-3 in stars, suggesting that the Arches cluster is the densest, and one of the most massive, young clusters in the Galaxy. They further used these data to estimate
Based on observations with the NASA/ESA Hubble Space Telescope, obtained at the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc. under NASA contract No. NAS5-26555. Data presented herein were obtained at the W.M. Keck Observatory, which is operated as a scientific partnership among the California Institute of Technology, the University of California and the National Aeronautics and Space Administration. The Observatory was made possible by the generous financial support of the W.M. Keck Foundation.
2 1


­3­ an initial mass function (IMF) which been found for the solar neighborho o d (Scalo 1998). They also estimated an mix of spectral types, which makes mo dels. is very flat ( -0.6±0.1) with respect to what has (Salpeter 1955, -1.35) and other Galactic clusters age of 2±1 Myr, based on the magnitudes, colors, and the cluster ideal for testing massive stellar-evolution

Given its extraordinary nature, the Arches cluster has been a target for many new observations. Stolte et al. (2002) recently verified a flat IMF slope for the Arches cluster, finding = -0.8, using both adaptive optics imaging with the Gemini North telescope and the HST/NICMOS data presented in Figer et al. (1999a). Blum et al. (2001) used adaptive optics imaging at the CFHT and HST/NICMOS data (also presented in this paper) to identify several new emission-line stars and estimate an age for the cluster of 2-4.5 Myr. Lang, Goss, & Ro dr´ z (2001a) detected eight radio sources, seven of which have thermal igue spectral indices and stellar counterparts, within 10 of the center of the cluster. They suggest that the stellar winds from the counterparts pro duce the radio emission via free-free emission, consistent with earlier indications from near-infrared narrow-band imaging (Nagata et al. 1995) and spectroscopy (Cotera et al. 1996). In a related study, Lang, Goss, & Morris (2001b) argued that the hot stars in the Arches cluster are responsible for ionizing the surface of a nearby molecular cloud to pro duce the arches filaments, as originally suggested by Cotera et al. (1996) and Serabyn et al. (1998), but in contrast to earlier suggestions (Morris & Yusef-Zadeh 1989; Davidson et al. 1994; Colgan et al. 1996). Yusef-Zadeh et al. (2001) used the Chandra telescope to detect three x-ray components that they asso ciate with the cluster, claiming that hot (107 K) x-ray emitting gas is pro duced by an interaction between material expelled by the massive stellar winds and the lo cal interstellar medium. The Arches cluster has also been the target of several theoretical studies regarding dynamical evolution of compact young clusters. Kim, Morris, & Lee (1999) used FokkerPlanck mo dels and Kim et al. (2000) used N-bo dy mo dels to simulate the Arches cluster, assuming the presence of the gravitational field of the Galactic Center. They found that such a cluster will disperse through two-bo dy interactions over a 10 Myr timescale. PortegiesZwart et al. (2001) performed a similar study and found a similar result, although they note the possibility that the Arches cluster is lo cated in front of the plane containing the Galactic Center. Finally, Gerhard (2001) considered the possibility that compact clusters formed outside the central parsec will plunge into the Galactic center as a result of dynamical friction, eventually becoming similar in appearance to the young cluster currently residing there; Kim & Morris (2002) further consider this possibility. In this paper, we use new and existing observations to determine the stellar properties of the most massive stars in the Arches cluster. We present astrometry and photometry of


­4­ stars with estimated initial masses greater than 20 M (the theoretical minimum mass of O stars), based upon HST/NICMOS narrow-band and broad-band imaging. We also present K-band high-resolution spectra of the emission-line stars, based upon Keck/NIRSPEC observations. We couple these data with previously-reported radio and x-ray data to infer stellar wind/atmosphere properties using a mo deling co de. Finally, we compare our results to those reported in recent observational and theoretical papers.

2.

Observations

A log of our observations obtained using HST and KECK is given in Table 1.

2.1.

HST

The HST data were obtained as part of GO-7364 (PI Figer), a program designed to measure the IMF's in the Arches and Quintuplet clusters (Figer et al. 1999a), determine the star-formation history of the Galactic Center (Serabyn et al. 2002), and determine the nature of the "Pistol Star" (Figer et al. 1999b). Broad-band images were obtained using HST/NICMOS on UT 1997 September 13/14, in a 2â2 mosaic pattern in the NIC2 aperture (19. 2 on a side). Four nearby fields, separated from the center of the mosaic by 59 in a symmetric cross-pattern, were imaged in order to sample the background population. All fields were imaged in F110W (center = 1.10 µm), F160W (center = 1.60 µm), and F205W (center = 2.05 µm). The STEP256 sequence was used in the MULTIACCUM read mo de with 11 reads, giving an exposure time of 256 seconds per image. The plate scale was 0. 076 pixel-1 (x) by 0. 075 pixel-1 (y), in detector co ordinates. The mosaic was centered on RA 17h 45m 50s 35, DEC -28 49 21. 82 (J2000), and . the pattern orientation was -134 6. The spectacular F205W image is shown in Figure 1a, . after first being pro cessed with the standard STScI pipeline pro cedures. The narrow-band images were obtained at roughly the same time as the broad-band images. One image was obtained in each of the F187N (center = 1.87 µm) and F190N (center = 1.90 µm) filters in the NIC2 aperture. The filters widths are 0.0194 µm for F187N and 0.0177 µm for F190N. We used the same exposure parameters for these images as those used for the broad-band images. The difference image, F187N minus F190N, is shown in Figure 2.


­5­ 2.2. Keck

The spectroscopic observations were obtained on July 4, 1999, using NIRSPEC, the facility near-infrared spectrometer, on the Keck II telescope (McLean et al. 1998, 2002), in high resolution mo de, covering K-band wavelengths (1.98 µm to 2.28 µm). The long slit (24 ) was positioned in a north-south orientation on the sky, and a slit scan covering a 24 â14 rectangular region was made by offsetting the telescope by a fraction of a slit width to the west between successive exposures. The slit-viewing camera (SCAM) was used to obtain images simultaneously with the spectroscopic exposures, making it easy to determine the slit orientation on the sky when the spectra were obtained. From SCAM images, we estimate seeing (FWHM) of 0. 6. The plate scales for both spectrometer and SCAM were taken from Figer et al. (2000a). We chose to use the 3-pixel-wide slit (0. 43) in order to match the FWHM of the seeing disk. The corresponding resolving power was R23,300 (=/FWHM ), as measured from unresolved arc lamp lines. The NIRSPEC cross-disperser and the NIRSPEC-6 filter were used to image six echelle orders onto the 10242-pixel InSb detector. The approximate spectral range covered in these orders is listed in Table 2. Coverage includes He I (2.058 µm), He I (2.112/113 µm), Br /He I (2.166 µm), He II (2.189 µm), N III (2.24/25 µm), and the CO bandhead, starting at 2.294 µm and extending to longer wavelengths beyond the range of the observations. Quintuplet Star #3 (hereafter "Q3"), which is featureless in this spectral region (Figer et al. 1998, Figure 1), was observed as a telluric standard (Moneti et al. 1994). Arc lamps containing Ar, Ne, Kr, and Xe, were observed to set the wavelength scale. In addition, a continuum lamp was observed through a vacuum gap etalon filter in order to pro duce an accurate wavelength scale between arc lamp lines and sky lines (predominantly from OH). A field relatively devoid of stars (RA 17h 44m 49s 8, DEC -28 54 6. 8 , J2000) was observed . to provide a dark current plus bias plus background image; this image was subtracted from each target image. A quartz-tungsten-halogen (QTH) lamp was observed to provide a "flat" image which was divided into the background-subtracted target images.

3.

Data Reduction Photometry

3.1.

The NICMOS data were reduced as described in Figer et al. (1999a) using STScI pipeline routines, calnica and calnicb, and the most up-to-date reference files. Star-finding, PSFbuilding, and PSF-fitting pro cedures were performed using the DAOPHOT package (Stetson


­6­ et al. 1987) within the Image Reduction and Analysis Facility (IRAF)10 . For the narrowband photometry, PSF standard stars were identified in the field and used in ALLSTAR. We used these stars to construct a mo del PSF with a radius of 15 pixels (1.125 ). This mo del was then fitted stars found throughout the image using DAOFIND. Aperture corrections were estimated by comparing the magnitudes of the PSF stars with those from an aperture of radius 7.5 pixels (0.563 ), and then adding -2.5log(1.159) in order to extrapolate to an infinite aperture (M. Rieke, priv. comm.). Table 1 gives the net aperture corrections to correct the aperture from a 3 pixel radius to infinity.

3.2.

Source Identification and Astrometry

We culled the list pro duced by the pro cess above by excluding stars with AK <2.8 or AK >4.2, or equivalently, stars with mF160W -mF205W <1.4 or mF160W -mF205W >2.1. These choices are motivated by the fact that the ma jority of stars in the Arches cluster have values within these limits, as can be seen in Figure 4 of Figer et al. (1999a); stars with values outside of these limits are likely to be foreground or background stars. The resultant star list is shown in Table 3. Ob jects are sorted according to inferred absolute K-band magnitudes, in order of decreasing brightness. K-band absolute magnitudes were calculated using AK = EH-K /(AH /AK - 1), where EH-K = (H - K ) - (H - K )0 , and A -1.53 (Rieke et al. 1989). We estimated intrinsic colors by convolving filter profiles with our best-matched mo del spectra. In cases where spectra were not available, i.e. for faint stars, AK = 3.1, a value that is supported by Figer et al. (1999a). This absolute magnitude was then translated into an initial mass according to the pro cedure in Figer et al. (1999a), except that we assumed solar metallicity, a decision supported by our quantitative spectroscopic analysis described later. Alternate identifications were taken from the following: Nagata et al. (1995), Cotera et al. (1996), Lang, Goss, & Ro dr´ z (2001a), and Blum et al. (2001). We had to allow as igue much as a 1 offset in some cases so that the correct ob jects coincided in the various data sets. The position offsets in the table are in right ascension (RA) and declination (Dec) and with respect to the star with the brightest inferred absolute magnitude at K. These offsets were calculated by applying the anamorphic plate scale at the time of observation and rotating x- and y- pixel offsets in the fo cal plane in to RA and Dec offsets in the sky. The stars are plotted and numbered in Figure 1b. Note that the masses in this table apply only in the case that the stars satisfy our mo del assumptions. This is not the case for some of
IRAF is distributed by the National Optical Astronomy Observatories, which are operated by the Asso ciation of Universities for Research in Astronomy, Inc., under cooperative agreement with the National Science Foundation.
10


­7­ the stars, especially the faint ones, given that they are likely to be field stars in the Galactic Center, but otherwise unasso ciated with the Arches cluster.

3.3.

Narrow-band Imaging

The narrow-band filters cover wavelength regions that include several potentially relevant atomic transitions. There are potential contributions to the total observed flux through the F187N and F190N filters from 8 and 2 transitions (H I Paschen-, He I, and He II), respectively. It is clear from the broad feature near 2.166 µm that He I lines are strong in the spectra of the emission-line stars, while the relatively weak 2.189 µm line indicates that the He II lines falling in the F187N filter are minor contributors to FF187N . We estimate that the flux in the two lines falling in the F190N filter is negligible.

3.4.

Narrow-band equivalent-widths

The equivalent-widths in Table 3 were computed according to the following equation: EW1.
87

µm =

(FF187N - FF190N ) â , FF190N

(1)

where the fluxes are in W cm-2 µm-1 and is the FWHM of the F187N filter; this equation assumes that the emission line(s) lie completely within the filter bandwidth, and that there is no contamination of FF190N by emission or absorption lines. We increased FF187N to account for the difference in reddening for the two wavelengths, assuming the extinction law of Rieke et al. (1989); this correction is approximately 8.0% and depends slightly on the estimated extinction. In addition, we reduced FF187N by 9.4% to account for the fact that the F187N filter has a shorter wavelength than the F190N filter and thus FF187N will be greater by this amount than FF190N for normal stars due to increasing flux toward shorter wavelengths on the Rayleigh-Jeans tail of the flux distributions at these wavelengths. Apparently, these two effects nearly cancel each other. The difference image (F187N-F190N) in Figure 2 contains about two dozen emissionline stars with significant flux excesses in the F187N filter, as listed in Table 3. The narrowband equivalent-widths increase with apparent brightness, as seen in Figure 3. In addition, the colors are redder as a function of increasing brightness (Figer et al. 1999a, Figure 4). Both effects are consistent with the notion that the winds are radiatively driven, so that higher luminosity stars will have stronger winds, and thus stronger emission lines, with commensurately stronger free-free emission which has a relatively flat (red) spectrum. In


­8­ addition, the emission lines themselves contribute to a redder appearance. These effects are borne out in our mo dels which show that a zero-age O star will have mF160W -mF205W =-0.07, while a late-type nitrogen-rich WR (WNL) star will have mF160W -mF205W =+0.07 from the continuum alone, and mF160W -mF205W =+0.13 from the emission-lines and continuum. So, a WNL star will have a color that is about +0.20 magnitudes redder than a zero-age O star. This agrees very well with the color trend seen in Figure 4, after one removes the stars with mF160W -mF205W >1.85; these stars are sub ject to extraordinary reddening, consistent with their lo cation far from the cluster center and the behavior of reddening as a function of increasing distance from the center of the cluster (Stolte et al. 2002).

3.5.

Spectroscopy

The spectra were reduced using IDL and IRAF routines. All target and calibration images were bias and background subtracted, flat fielded, and corrected for bad pixels. The target images were then transformed onto a rectified grid of data points spanning linear scales in the spatial and wavelength directions using the lo cations of stellar continuum sources and wavelength fiducials extracted from arc line images and continuum lamp plus etalon images. The rectified images were then used as inputs to the aperture extraction pro cedure. Finally, all 1D spectra were coadded and divided by the spectrum of a telluric standard to pro duce final spectra. The following gives a detailed description of this data reduction pro cedure. A bias plus background image was pro duced from several images of a dark area of sky, observed with the same instrument and detector parameters as the targets. Because the sky level was changing while these images were being obtained, we scaled the sky components before combining them with a median filter. The combined image was scaled in order to match the varying sky emission level in the target images and then added to the bias image formed by taking the median of several bias images. The resultant image was then subtracted from target images, thus subtracting properly scaled bias structure and background. This operation also removes dark current, although the subtraction will be perfect only in the case that the scaling factor for the varying sky level is exactly one. In other cases, a small residual in dark current will remain, although the amount of this residual will typically be less than 1 count (5 e- ). Target images were divided by a normalized flat image, with bias structure and dark current first removed. We then removed deviant pixels from these images with a two-pass pro cedure. First, the median in a 5-by-5 pixel box surrounding each pixel was subtracted from every pixel in order to form an image with the low-spatial-frequency information removed. If the absolute value of a pixel in this difference image was larger than five times


­9­ the deviation in the nearby pixels, then its value was replaced by the median data value in the box. The deviation in nearby pixels is defined as the median of the absolute values of those pixels in the difference image. On the second pass, isolated bad pixels were flagged and replaced if both the following were true: they were higher or lower than both immediate neighbors in the dispersion direction, and their value deviated by more than 10 times the square ro ot of the average of those two neighbors. Isolated bad pixel values are replaced by the average values of their neighbors. We rectified the target images by mapping a set of continuum traces and wavelength fiducials in the reduced images onto a set of grid points. This dewarping in both spatial and spectral directions is done simultaneously and requires knowledge of the wavelengths and positions of several spectral and spatial features; this information was extracted using arc, sky, or etalon lines. Typically, we identified 15 to 20 lines in each echelle order for this purpose. Two stellar continuum spectra and two flat-field edges traced across the length of the dispersion direction were used to define spatial warping. Rectification was done separately for each echelle order. Unfortunately, there is a lack of naturally o ccurring, and regularly spaced, wavelength fiducials (absorption or emission lines) pro duced by the night sky or arc lamps. Because of this, we had to use a three-stage pro cess for determining the relationship between column number and wavelength: rectify the images using the arc and sky lines, measure the etalon line wavelengths in the rectified version of the etalon image and obtain a solution to the etalon equation, and use the analytically determined wavelengths to pro duce a better rectification matrix. This approach gave results that were repeatable (to within an rms of 1 km s-1 ) (Figer et al. 2002). This three-stage pro cess is described in detail below. In the first stage, we chose an order containing many (15 to 20) arc and sky lines so that wavelengths in the rectified images would be fairly well determined. The last few significant figures of some of the arc-line wavelengths, and all of the OH lines, out to seven significant figures, were derived from lists available from the National Institute of Standards and Technology (NIST) Atomic Spectra Database11 . The pro cess for fitting spectral lines was as follows. Each arc or sky line was divided into 10 to 20 sections along the length of the line. Rows from each section were averaged together, and the lo cation of the peak was found by centroiding. Points along the line obtained from the centroiding were typically fitted with a 3rd-order polynomial. Outliers were clipped, and a new fit performed. We used a similar approach to fit the stellar and flat-field edge spatial traces.
11

http://physics.nist.gov/PhysRefData/ASD1/nist-atomic-spectra.html


­ 10 ­ We then used the arrays of co efficients from spatial and spectral fits to pro duce a mapping between points along the spectral features in the warped frame, and points in the dewarped frame. A two-dimensional second or third order polynomial was then fitted to these points, resulting in a list of transformation co efficients, the "rectification matrix." In the second stage, after dewarping the etalon image with the rectification matrix pro duced from the arc and sky lines, we measured the wavelength of each etalon line using SPLOT in IRAF. This provided preliminary estimated wavelengths for each etalon line. Exact etalon wavelengths were given by the etalon equation. Solutions to the etalon equation were found using the constraint that features must have integer order numbers that decrease sequentially toward longer wavelengths. The estimated wavelengths of the 14 to 16 lines were used to pro duce a series of etalon equations that we simultaneously solved by finding the thickness and order numbers giving the least overall variance from the measured wavelengths. Once these parameters were determined, exact wavelengths could be calculated for each etalon line. In stage three, the exact etalon wavelengths were used to determine a new rectification matrix by tracing these features in the rectified etalon image. Together with the same spatial information used to pro duce the first stage rectification matrix, a new rectification matrix was pro duced. The improved quality of the etalon lines allowed for a higher order (fourth or fifth order) polynomial fit to the etalon mapping between warped and dewarped points. The new matrices were applied to the appropriate spectral orders of target images. Three images of the telluric standard ("Q3") were taken using the same setup as was used to obtain the target images. We moved the telescope along the slit length direction between exposures so that the spectra were imaged onto different rows of the detector. We reduced these images using the sames pro cedures used for reducing the target images. Finally, APALL (IRAF) was used to extract spectra in manually chosen apertures. The resultant 1D spectra were then coadded in the case that a single ob ject was observed in multiple slit positionings. Before coadding, we shifted all spectra to a common slit position by cross-correlating the telluric absorption features and shifting the spectra. The coadded spectra of all the stars in Table 3 for which we have spectra are shown in Figure 5; note that these spectra have been smo othed using an 11-pixel (44 km s-1 ) square boxcar function for display purposes. The fluxes were not reddening-corrected, so the spectral shape (roughly flat) is indicative of a hot star observed through a large amount of extinction.


­ 11 ­ 4. Results

We use the data in this paper to form a census of stellar types for massive stars in the cluster, estimate physical properties of the stars, determine the dynamical state of the cluster members, and assess the impact of the cluster on its environment.

4.1.

Spectral Types

The spectra for the most massive stars in our data set are relatively similar, although the features have smaller equivalent-widths for fainter stars. The brightest stars generally have spectra with a weak feature near 2.058 µm (He I), weak emission near 2.104 µm (N III) and 2.112/113 µm (He I), broad and strong emission near 2.166 µm (He I, H I), a weak line near 2.189 µm in P-Cygni profile (He II) in some cases, and weak lines at 2.115/2.24/2.25 µm (N III), where the primary contributors to the emission line fluxes are due to transitions of species listed in parentheses. There are indications of absorption at 2.058 µm in most of the first ten stars, and in some cases, it is strongly in a P-cygni profile (#3 and #8). In both these cases, the absorption appears to have a "double-bottom." This line and the 2.112/113 µm blend are narrow in the spectra of #10, #13, #15, and #29. From these spectra alone, we might assign spectral types of WNL (WN7-WN9) (Figer, McLean, & Na jarro 1997) or O If+ (Hanson et al. 1996) for the brightest stars, just as have been assigned in Nagata et al. (1995) and Cotera et al. (1996); the degeneracy in the classification of stars of these spectral types was noted by Conti et al. (1995). Figure 6 compares an average spectrum of the Arches stars with those of WNL stars (Figer, McLean, & Na jarro 1997) and O If+ stars (Hanson et al. 1996). We can see that the Arches stars have N III emission at 2.104 µm, 2.24 µm, and 2.25 µm, just like that seen in the spectra of WN7-8 stars, while the spectra of the O If+ stars do not have those lines in emission (or they are very weak). Figer, McLean, & Na jarro (1997) argued that N III lines in the K-band might be used to distinguish Wolf-Rayet (WR) from O If+ stars on both observational and theoretical grounds. In addition, Figer, McLean, & Na jarro (1997) showed that WN stars separate by subtype as a function of the relative line strengths in certain lines of their Kband spectra, i.e. W2.189 µm /W2.166 µm or W2.189 µm /W2.11 µm . The relevant values for spectra of the most massive Arches stars are consistent with classifications of WNL. In addition, the equivalent-widths measured for the emission lines are in the range of those measured for WNL stars in Figer, McLean, & Na jarro (1997), but not for O If+ stars. From all these measures, we conclude that the Arches stars are WN9 types, with an uncertainty of ± one subtype. Later in this paper, we use wind/atmosphere mo dels to show that the estimated nitrogen, carbon, and helium abundances verify WNL classifications.


­ 12 ­ One ob jection to the WNL classification might be that it is improbable that all the WR stars in the cluster be in the exact same evolutionary phase, i.e. WNL. Actually, such a situation is predicted by Meynet (1995). Indeed, our new observations show that there are no WR stars in the cluster other than the WNL types, given that our narrow-band observations would easily detect the strong emission lines from all WR stars, such as the WNE star in the Quintuplet cluster (Figer, McLean, & Morris 1995), or the carbon-rich WR (WC) stars seen near the Pistol Star (Figer et al. 1999b). The lack of WC stars suggests an age less than 3.5 Myr (Meynet 1995), consistent with the WNL classification for brightest stars in the Arches cluster. We have identified two groups among the most massive stars with spectral types distinctly different than the WN9 spectral type. The group of stars from #10 to #30 are O If+ types, and fainter stars are O main sequence stars.

4.2.

Identification of the Main Sequence in the Galactic Center

The fainter stars in Table 3 (ID>30) appear to have little to no excess emission at 1.87 µm, consistent with the fact that their spectra appear to be relatively flat. Indeed, some even show absorption at 2.058 µm, 2.112 µm, and 2.166 µm, of a few angstroms equivalent width. In the case of star #68, the spectrum (Br absorption), apparent magnitude, and extinction, suggest a luminosity and temperature consistent with an initial mass of about 60 M , and its presence on the main sequence, given an age of 2.5 Myr and the Geneva mo dels. The star is likely to be a late-O giant or supergiant, albeit still burning hydrogen in its core, and thus its classification on the main sequence; note that it is to o bright to be a dwarf. There are several other similar stars that have relatively featureless spectra, consistent with the spectra of O main sequence stars. This result is a significant spectroscopic verification of the claim in Figer et al. (1999a) that the main sequence is clearly visible in the broadband NICMOS data; also, note that Serabyn, Shupe, & Figer (1999) identified likely O stars with spectra that appeared to be featureless at the resolution of their observations. While a possible identification of the main sequence in the Galactic Center has previously been claimed (Eckart et al. 1999; Figer et al. 2000a), the result in this paper is the first spectroscopic identification of single stars that are unambiguously on the main sequence, as observed by their narrow absorption lines.


­ 13 ­ 4.3. Wind/atmosphere Mo delling

To mo del the massive Arches stars and estimate their physical parameters, we have used the iterative, non-LTE line blanketing metho d presented by Hillier & Miller (1998). The co de solves the radiative transfer equation in the co-moving frame for the expanding atmospheres of early-type stars in spherical geometry, sub ject to the constraints of statistical and radiative equilibrium. Steady state is assumed, and the density structure is set by the mass-loss rate and the velo city field via the equation of continuity. We allow for the presence of clumping via a clumping law characterized by a volume filling factor f(r ), so that the "smo oth" mass loss rate, M S , is related to the "clumped" mass-loss rate, M C , through M S =M C /f1/2 . The velo city law (Hillier 1989) is characterized by an isothermal effective scale height in the inner atmosphere, and becomes a law in the wind. The mo del is then prescribed by the stellar radius, R , the stellar luminosity, L , the mass-loss rate M , the velo city field, v (r ), the volume filling factor, f, and the abundances of the elements considered. Hillier & Miller (1998, 1999) present a detailed discussion of the co de. For the present analysis we have assumed the atmosphere to be composed of H, He, C, N, Mg, Si and Fe. We created a grid of mo dels within the parameter domain of interest. It was bounded by 4.4

­ 14 ­ spectra with mo dified f (f=1, #10b) and terminal velo city (v =1600 km s-1 , #10c). Note that there is no simple M C 1 /f1 1/2 = M C 2 /f2 1/2 scaling (Herald, Hillier, & Schulte-Ladbeck 2001) nor M 1 /v 1 =M 2 /v 2 scaling between the mo dels, and that other stellar parameters require readjustment. Table 4 reveals that ob jects #8 and #10 have very similar luminosities, temperatures, and ionizing-photon rates. However, their wind densities (M ) and abundances reveal different evolutionary phases for these two ob jects. Ob ject #8 fits well with a WNL evolutionary stage. Its wind density and helium abundance are very similar to those derived by Bohannan & Crowther (1999) for WN9h stars. On the other hand, ob ject #10 can be placed into a O If+ phase as its wind density is roughly an order of magnitude lower than that of ob ject #8 and the derived He abundance and upper limits for nitrogen enrichment indicate an earlier evolutionary phase. Indeed, its K-band spectrum is nearly identical to that of the O8 If+ star HD151804 (Hanson et al. 1996). Interestingly, both Arches ob jects have luminosities about one magnitude larger than their counterparts in Bohannan & Crowther (1999).

4.4.

Ionizing Flux

Containing so many massive stars, the Arches cluster pro duces a large ionizing flux. We estimate the total ionizing flux emitted by the cluster to be 4(1051 ) photons s-1 , based on our wind/atmosphere mo del fits for the two emission-line stars in Table 4. To estimate the total flux, we multiplied the estimated ionizing flux from #8 by 10, that from #10 by 20, and determined those of the remaining stars in Table 3 by applying equation 3 in Crowther & Dessart (1998); these factors reflect our choice of #8 and #10 as representatives of the first 30 stars in the table. The ionizing flux estimate is a bit higher than that in Serabyn et al. (1998), a few 1051 photons s-1 , after scaling that number for the fact that the cluster contains 50% more O stars than originally thought (Figer et al. 1999a). This amount of ionizing flux is consistent with the Arches cluster being the ionizing source for the Thermal Arches Filaments (Lang, Goss, & Morris 2001b).

4.5.

Luminosity

Using the same pro cess as applied to estimate the cluster ionizing flux, we estimate a total cluster luminosity of 107.8 L , or one of the most luminous clusters in the Galaxy. About 40% of the total luminosity is contributed by the 30 brightest stars in Table 3.


­ 15 ­ 4.6. Age

As described in Figer et al. (1999b), the absolute magnitudes and mix of spectral types are consistent with a cluster age of 2±1 Myr. This was estimated using the colors and magnitudes of the stars, i.e. the colors give the extinction value, and the apparent magnitudes lead to absolute K-band magnitudes, which are then compared to iso chrones from the Geneva mo dels (Meynet et al. 1994). It is possible for older stars to attain magnitudes as bright as the brightest stars in the cluster, but only at relatively co ol temperatures, i.e. Teff < 25kK. A new age constraint from the data in this paper is given by the absence of WC or WNE stars in the cluster. This observational constraint, when combined with the mo dels of Meynet (1995), gives age < 3.0 Myr for the least limiting case (2âM , Z=0.040) and age < 2.5 Myr for the most limiting case (1âM , Z=0.040). In addition, the presence of WNL stars requires age > 1.5 Myr from these mo dels. Finally, we note the lack of relatively co ol (B-type) supergiant emission-line stars, such as those found in the central parsec (Krabbe et al. 1995) and the Quintuplet cluster (Figer et al. 1999b). The lack of such stars in the Arches cluster implies age < 4.0 Myr. Finally, the detailed mo del for star #8 suggests an age of 2.5 Myr, at least for that star. We combine all this evidence to suggest that the cluster age is 2.5 ±0.5 Myr, where the error is dominated by our lack of information concerning metallicity.

4.7.

Velo cities

A velo city determination for the emission-line stars is complicated by several facts. First, the strongest spectral features are blended emission lines, making it impractical to simply compare the measured wavelength centroid of a "line" to the expected vacuum wavelength. Because of this, one must cross-correlate the target spectra with respect to a template spectrum composed of features that accurately represent relative strengths of the blended lines. Second, the emission lines are broad, so that small wavelength shifts will pro duce little change in the cross-correlated power when compared to a template spectrum. Third, there are slight differences in the shapes of the emission lines between spectra of the various stars, so a mo del blend for one spectrum might not faithfully repro duce the features in another spectrum, at least not to the fidelity required for high precision velo city measurements. Because of these difficulties, we approached the velo city estimates using two techniques. We smo othed the spectra using a box-car filter with varying widths between 1 and 31 pixels (4 and 120 km s-1 ), finding little difference in the velo cities as a function of these widths. First, we estimated an absolute velo city for star #8 by cross-correlating its spectrum with that of our mo del spectrum. This metho d gave a velo city of +54 km s-1 (redshifted),


­ 16 ­ in the helio centric frame, for the blend near 2.166 µm. Our estimates using other lines are somewhat less than this value, as low as +20 km s-1 , but those lines are weaker than the blend at 2.166 µm, and thus pro duce larger velo city errors. In the second metho d, we cross-correlated all the spectra against each other. The crosscorrelations were performed separately on three wavelength regions, the first containing the 2.104 µm and 2.115 µm features, the second containing the 2.166 µm and 2.189 µm features, and the third containing the doublet at 2.25 µm. This metho d was used to compute the standard deviation of relative velo cities, allowing us to infer the mass enclosed within some orbital radius that represents an average of the emission-line stars' orbital radii. Given that small differences in intrinsic blend morphology can affect the lo cation of the maximum point of the cross-correlated power, we also repeated this approach using line centroids to compute velo city differences. Using both approaches, we found a standard deviation of 22 km s-1 for a sample containing eight emission-line stars. This value represents an upper limit on the intrinsic dispersion, given that the effects described above would tend to increase the estimated value over the intrinsic velo city dispersion. We also found that the stars were redshifted by +41 km s-1 with respect to star #8. We therefore estimate a helio centric "cluster" velo city of +95±8 km s-1 , where the error is simply the standard deviation divided by the square ro ot of eight; note that this error neglects the systematic effects described above, so the true velo city might differ from the estimate by significantly larger than the quoted error. For a gravitationally bound and spherically symmetric cluster of mass Mcluster , the virial theorem gives Mcluster = 3 2 R/G (Ho and Filippenko 1996), where is the one-dimensional velo city dispersion, R is the appropriate radius, G is the gravitational constant, the velo cities are assumed to be isotropic, and all stars have equal mass. This simple formula can be compared to the more general case where the cluster can be resolved into individual stars (Illingworth 1976). Using R = 0.23 pc for the sample in question, we calculate Mcluster < 7(104 ) M , or about five times greater than what would be expected from direct integration of the mass function in Figer et al. (1999a) over the area sampled by the stars used in the analysis. This high mass limit results from the systematic effects inherent in our radial velo city determinations, as described above.

5.

Discussion

In this section, we compare our measurements to those in previous papers and use measurements at other wavelengths to determine the physical parameters of the observed stars. Finally, we discuss how the Arches cluster interacts with its lo cal environment to


­ 17 ­ create heating and ionization of a nearby cloud.

5.1.

Comparison to Previous Near-infrared Measurements

Table 3 lists over 30 probable emission-line stars, albeit the faintest having relatively weak emission lines; Figure 3a confirms that there are roughly this number of stars with reliable emission-line excesses. This list contains over a factor of two increase in the number of emission-line stars previously identified in the cluster (Blum et al. 2001; Nagata et al. 1995; Cotera et al. 1996). The line and continuum fluxes presented here largely agree with earlier results (Nagata et al. 1995; Cotera et al. 1996). The FF187N and FF190N fluxes reported in this paper are similar to those reported in Blum et al. (2001), after correcting for differences in the assumed zero points, the fact that we corrected for the difference in extinction at the two narrow-band wavelengths, and that we also corrected for the intrinsic shape of the stellar continuum; in addition, our extinction estimates are higher in many cases than those used in Blum et al. (2001). The spectra in this paper are consistent with the narrow-band photometry in Nagata et al. (1995) and Blum et al. (2001) and the spectra in Cotera et al. (1996), although our high-resolution spectroscopy shows that the photometry is significantly affected by blending of absorption and emission features in P-Cygni profiles. We confirm the discovery of a new bright emission-line star (#5, B22, N9) near the southern edge of the cluster, reported in Blum et al. (2001), and note that it is a counterpart of the x-ray source, "AR8," in Lang, Goss, & Ro dr´ z (2001a). We also confirm that star igue #16 (B19) is an emission-line star, as suspected by Blum et al. (2001). Blum et al. (2001) listed some additional candidate emission-line stars. We confirm that the following stars from that list are, indeed, emission-line stars (their designations in parentheses): #15 (B8), #27 (B16), #17 (B29), #10 (B30), #10 (B20), #13 (B31).

5.2.

Comparison to X-ray Flux Measurements

Yusef-Zadeh et al. (2001) reported Chandra X-ray observations of a region including the Arches cluster. They detected three extended sources, one (A1) near the center of the Arches cluster, another (A2) lo cated to the North and West of the center by about 7 and a third weaker source (A3) about 90 â 60 in size underlying the first two. The centroid of source A2 coincides within 1 arcsecond with an emission-line star, #9 in Table 3. The apparent spatial coincidence of the X-ray sources and Arches cluster strongly suggests that


­ 18 ­ the X-ray sources are physically asso ciated with the cluster. Yusef-Zadeh et al. estimate the total X-ray luminosity between 0.2 and 10 keV to be 3.3, 0.8 and 0.16 (1035 ) ergs s-1 for A1, A2 and A3, respectively. They attribute the emission from A1 and A2 to either colliding winds in binary systems or to the winds from single stars interacting with the collective wind from the entire cluster. The coincidence of source A2 with an emission-line source is very interesting in the context of the latter scenario. A3, on the other hand, has roughly the characteristics expected from sho ck-heated gas created by the collisions of the multitude of 1000-km s-1 stellar winds emanating from the stars in the rich, dense cluster (Ozernoy, Genzel, & Usov 1997; Canto, Raga, & Ro dr´ z 2000). Because the X-ray sources are ´ igue extended, it is unlikely that they can be attributed to single X-ray binary systems. However, the rough coincidence of source A1 with the core of the cluster raises the possibility that it may be comprised of many relatively weak stellar X-ray sources, binary or single, residing in the cluster core, and unresolved spatially from each other.

5.3.

Comparison to Radio Flux Measurements

From Table 3 we see that one of the ob jects analysed in this work, #8, has also been detected at 8.5 GHz (Lang, Goss, & Ro dr´ z 2001a). Our derived mass-loss rate is consistent igue with the observed radio flux (0.23 mJy) only if the outer wind regions are unclumped. Such a behavior for the clumping law has been suggested by Nugis, Crowther, & Willis (1998) from analysis of galactic WR stars. They found that the observed infrared to radio fluxes of WR stars are well repro duced by a clumping law where the filling factor is unity close to the stellar surface, increasing to a minimum at 5 to 10 R and returning again to unity in the outer wind where the radio flux forms. Note, however, that the line fluxes of the weaker lines like Br or He I remain unaltered with this new description of the clumping law, but the line fluxes of the strongest lines, such as Pa, formed in the outer wind can be significantly reduced. From Table 4 we see that our mo dels are fully consistent with the observed equivalent-width of Pa. We consider now the possible correlation between line fluxes and radio-continuum flux analogous to the one discussed above for K-band fluxes (see Figure 3a). In principle, we also expect the near-infrared emission line strengths to scale with the free-free emission detected at radio wavelengths (Nugis, Crowther, & Willis 1998; Leitherer, Chapman, & Koribalski 1997). However, we do not find such a correlation, as can be seen in Figures 8a,b. A similar result was obtained by Bieging, Abbott, & Churchwell (1982) for a sample of eight WR stars. We believe this apparent absence of correlation between Pa line-strength and radio flux is caused by both observational and physical effects. The observational effect is related to the


­ 19 ­ fact that the radio measurements are picking up only the tip of the iceberg, i.e., those stars with the densest winds of the cluster. The physical effect is related to the fact that all three components contributing to EW1.87µm (H I, He I, and He II) are very sensitive to changes in temperature in the parameter domain appropriate to these ob jects. Further, both the line and continuum fluxes depend strongly not only on the mass-loss rate but also on the shape of the velo city field and the clumping law. Therefore, such strong dependence of the Pa line flux on several stellar parameters intro duces a large scatter in the expected line-strength vs radio-flux relationship. The radio fluxes of the most massive Arches stars are comparable to those of WNL stars, but not to those of O If+ stars. The WN8 star WR105 (van der Hucht 2001) would emit 0.14 mJy at the distance of the Arches cluster, comfortably within the range of fluxes measured for the Arches stars. Similar values are reported for WR stars in Bieging, Abbott, & Churchwell (1982). On the other hand, HD 16691 (O4 If+ ) emits 0.3 mJy at 4.9 GHz, according to Wendker (1995), implying an expected flux of 1.7 µJy at the distance of the Arches cluster, assuming that the star has a parallax of 1.7 mas (Perryman et al. 1997). The expected flux is two orders of magnitude below the flux levels of the brightest Arches stars (Lang, Goss, & Ro dr´ z 2001a). No doubt, this difference is due to the relatively low igue mass-loss rate for HD 16691, about 1/20 of that of the bright Arches stars. A similar trend can be seen in Figure 6 where the emission lines in the spectra of HD 16691 are shown to be much weaker than those in the spectra of the Arches emission-line stars. Again, weak winds pro duce weak emission lines and weak free-free emission. Finally, we r They were found (2001a) with the have designated in Table 3. eport several additional radio sources having emission-line star counterparts. by comparing the radio continuum contour plot in Lang, Goss, & Ro dr´ iguez difference image in Figure 2; they are marked by squares in this figure. We the four near-infrared counterparts to these newly identified radio sources

5.4.

Evolutionary Status of the Massive Stars in the Arches Cluster

The emission-line stars appear to contain significant amounts of hydrogen, while also exhibiting considerable helium content. We believe that this can be explained by the most massive stellar mo dels in Meynet et al. (1994). For the brightest 10 or so stars in Table 3, the observations can be fit by these mo dels for Minitial 120 M stars that have evolved to co ol temperatures while retaining hydrogen. In particular, star #8 can be fit by a Minitial 120 M star with solar abundance, standard mass-loss rates, age of 2.4 Myr to 2.5 Myr, and presentday mass of 72 M to 76 M (Schaller, Schaerer, Meynet, & Maeder 1992).


­ 20 ­ 5.5. Relation to the Nearby Molecular Cloud (M0.10+0.03)

It appears that the Arches cluster heats and ionizes the surface of M0.10+0.03, the nearby molecular cloud (Serabyn & Guesten 1987; Brown & Liszt 1984), given that the cluster can easily provide the necessary flux to account for the infrared emission and recombinationline flux from the cloud. The relative helio centric velo city between the cluster stars (+95±8 km s-1 ) and the ionized gas on the surface of the cloud (-20 to -50 km s-1 ) suggests that the physical asso ciation is accidental and that the cluster stars are ionizing the surface of the cloud. This difference in velo city is reminiscient of that observed between the Quintuplet (+130 km s-1 ) (Figer 1995; Figer et al. 1999a) and the Sickle cloud, M0.10+0.03 (+30 km s-1 ) (Lang, Goss, & Wo o d 1997). In both cases, it appears that young clusters happen to lie near molecular clouds whose surfaces are ionized by the photons from the hot stars in the clusters. The following shows that the ionizing flux and energy required to heat the cloud can be provided by the Arches cluster. Note that a differential velo city of 100 km s-1 would pro duce a relative drift of 100 pc in 1 Myr, a distance that would bring a cluster within the vicinity of a few clouds, given the spatial distribution of clouds in the central few hundred parsecs.

5.5.1. Ionization and Heating Even before the discovery of the Arches cluster, Serabyn & Guesten (1987), Genzel et al. (1990), and Mizutani et al. (1994) suggested that the Thermal Arched Filaments are photoionized by nearby hot stars. After the discovery of the cluster, many authors considered the possibility that the cluster is ionizing the cloud. One problem with this idea is the fact that the filaments are very large, and have roughly constant surface brightness and excitation conditions (Erickson et al. 1991; Colgan et al. 1996), indicating that the ionizing source is either evenly distributed over many parsecs or is relatively far away. Given the new ionizing flux estimates in this paper and in (Serabyn et al. 1998), the cluster would pro duce enough flux to account for the filaments, even if 20 pc away, far enough to allow for the even illumination that is observed. Indeed, Timmermann et al. (1996) predicted this result, and Lang, Goss, & Morris (2001b) present a detailed analysis that confirms it. Cotera et al. (1996) give estimates for the total ionizing flux of 2-5(1050 ) photons s-1 , depending on whether one mo dels the spectral energy distributions for the emission-line stars with Kurucz (1979) atmospheres or blackbo dy functions; however, this estimate includes only flux from the dozen or so emission-line stars that were known at the time. Nonetheless, the results of this paper suggest that the cluster ionizes the cloud.


­ 21 ­ The heating in the cloud pro duces an infrared luminosity of 107 L (Morris, Davidson, & Werner 1995). Assuming a covering fraction of 10%, we find that the Arches cluster can deliver about this much luminosity, within a factor of two.

5.5.2. Location along the line of sight Given the Br flux from the filaments measured by Figer (1995), we know that the line emission is extincted by about AK 3. This implies that the filaments are on the near side of the cloud, since such an extinction corresponds only to the typical foreground extinction to the Galactic Center, and precludes any substantial additonal extinction. This information leads to the conclusion that the Arches cluster is on the near side of, and is approaching, the wayward molecular cloud that is moving in opposition to the bulk motion of stars and gas around the Galactic Center (McGinn, Sellgren, Becklin, & Hall 1989), consistent with the geometry described in Lang, Goss, & Morris (2001b).

5.6.

Dynamical Evolution and Uniqueness of the Arches and Quintuplet Clusters

The temporal coincidence of the star formation events that pro duced the massive clusters in the Galactic Center, and the lack of older red supergiants, suggest that the Galactic Center has been host to a recent burst of star formation. Kim, Morris, & Lee (1999) and Kim et al. (2000) predicted that compact young clusters in the Galactic Center would evaporate on short timescales, i.e. a few Myr. Portegies-Zwart et al. (2001) argue that other clusters similar to, yet somewhat older than, the Arches and Quintuplet clusters exist in the central 100 pc. This argument is based upon a dynamical analysis which predicts that such clusters evaporate after 55 Myr, and further that the clusters' pro jected surface number density in stars drops below the limit of detectability in a few Myr. The statement that members of dispersed clusters could have gone undetected is incorrect. Such stars would easily be detectable for their extreme brightness, i.e. there would be hundreds of stars as bright as IRS 7 (in the central parsec) strewn about the central 100 pc for each Arches/Quintuplet-like cluster between the age of 5 and 30 Myr. Given the claim in Portegies-Zwart et al. about the expected number of "hidden" young clusters in the central hundred parsecs, we would expect to see of order ten thousand red supergiants in this region. Only a few are seen, as demonstrated by surveys for such stars (Figer 1995; Cotera 1995). Portegies-Zwart et al. suggest that clusters could be "hiding" near bright stars due to


­ 22 ­ limitations in dynamic range, of array-based detections are multiple coadds are used (e.g. in agreement with Kim et al., but these arguments are specious, since the obviously not limited by the digitization of Figer et al. (1999a) reach a dynamic range of we conclude that the clusters must disperse r dynamic range a single read if over 105 ). Thus apidly

5.7.

Comparison to NGC 3603 and R136 in 30 Dor

The Arches cluster is similar in age and content to NGC 3603 and R136 in 30 Dor, and is surrounded by a giant H II region as is R136. In contrast to these clusters, we do not see WN5h or WN6h stars (Crowther & Dessart 1998), suggesting that the Arches cluster is older. This is consistent with our age determination, as suggested by other means described earlier. While our spectra exhibit no primary diagnostic lines to estimate metallicity, we may use our estimates for the helium and nitrogen abundances in ob ject #8 in conjunction with the evolutionary mo del for 120 M to infer metallicity. For Z(He)=0.7, the mo dels predict the star to have already reached its maximum nitrogen surface mass fraction. Hence, we may compare our derived nitrogen mass fraction, Z(N)=0.016, with the evolutionary mo dels values at different metallicities (Schaller, Schaerer, Meynet, & Maeder 1992). We see that this value is met for solar metallicity. A more detailed analysis of the metallicity of the Arches stars will be presented in Na jarro et al. (2002). We thank Nolan Walborn and Nino Panagia for critical readings of the paper. We thank John Hillier for providing his co de. F. N. acknowledges DGYCIT grants ESP981351 and PANAYA2000-1784. We acknowledge the work of: Maryanne Angliongto, Oddvar Bendiksen, George Brims, Leah Buchholz, John Canfield, Kim Chin, Jonah Hare, Fred Lacayanga, Samuel B. Larson, Tim Liu, Nick Magnone, Gunnar Skulason, Michael Spencer, Jason Weiss and Wo on Wong.


­ 23 ­

Table 1. Log of Observations
Type Filter
a

center µm K-band 1.10 1.60 2.05 1.874 1.900

PHOTFNU/ZP(Vega)b sec. DN-1 ·· · 61E-10 86E-09 64E-09 95E-08 36E-08

Ap. Corr. 3 pix. to inf ··· 1.35 1.67 1.81 1.75 1.76

Integ.

Date

Spectroscopyc Imaging Imaging Imaging Imaging Imaging

NIRSPEC-6 F110W F160W F205W F187N F190N

9. 1. 1. 4. 5.

150 256 256 256 256 256

s. s. s. s. s. s.

13 13 13 13 13

3 July September September September September September

1999 1997 1997 1997 1997 1997

a NIRSPEC-6 has half-p ower p oints of 1.85 µm and 2.62 µm (Figer et al. 2000b). Because the orders are longer than the width of the detector, the spectra are not contiguous in wavelength. b Multiplying PHOTFNU/ZP(Vega) by the observed count rate gives the ratio of the ob ject's flux with resp ect to Vega. Values for F110W, F160W, and F205W are from the HST Data Handbook (Keyes et al. 1997). Values for F187N and F190N are from Marcia Rieke (priv. communication). The zero points in Blum et al. (2001) were also provided by Marcia Rieke; however, the values used in the table above were more recently provided. c All sp ectroscopy images were obtained with the slit p ositioned approximately north-south using the multiple correlated read mode ("Fowler" sampling) with 16 reads at the beginning and end of each integration. The resolution was 23,300, (/FWHM , where FWHM is the full-width at half maximum of unresolved arc lamp lines). The slit size was 0. 43 â 24 .


­ 24 ­

Table 2. Wavelength Coverage in Spectra Echelle Order 33 34 35 36 37 38 min µm 2.281 2.214 2.152 2.092 2.036 1.983 max µm 2.315 2.248 2.184 2.124 2.067 2.013


­ 25 ­

Table 3. Massive Stars in Arches Cluster
ID
a

Desig./Ref. (2)

b

RA (3) 0. -6. 8. 4. 3. 2. 3. 2. 0. -1. -1. 1. -2. 6. 7. 4. -0. 3. -5. 2. 7. 0. 12. -1. -3. 4. 5. 5. 7. 0. 2. 5. 6. 7. 6. -6. 3. 6. 11. 5. 2. 1. 3. 5. 3. 11. 6.

c

DEC (4)

c

(1) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

mF110W mag. (5) 16. 17. 16. 15. 16. 15. 15. 16. 16. 17. 17. 16. 17. 16. 16. 16. 18. 16. 18. 17. 16. 17. 17. 18. 19. 17. 17. 17. 17. 17. 17. 17. 17. 17. 17. 19. 99. 17. 18. 17. 19. 18. 18. 18. 18. 19. 18. 30 84 06 63 69 75 74 31 10 37 02 40 59 38 12 62 13 70 89 49 85 46 56 27 42 70 13 26 23 87 70 59 53 67 29 03 00 51 10 78 53 02 45 25 47 03 01

mF160W mag. (6) 12. 13. 12. 12. 12. 12. 12. 12. 12. 13. 12. 12. 13. 12. 12. 13. 14. 13. 14. 13. 13. 13. 13. 14. 15. 13. 13. 13. 13. 14. 13. 13. 13. 14. 13. 14. 14. 13. 14. 14. 15. 14. 14. 14. 14. 14. 14. 33 39 28 12 80 05 16 54 44 35 72 67 63 84 78 01 05 17 58 88 29 65 82 40 05 98 49 69 81 16 97 96 95 03 84 53 55 85 37 24 47 41 69 47 64 69 46

mF205W mag. (7) 10. 11. 10. 10. 10. 10. 10. 10. 10. 11. 10. 10. 11. 11. 11. 11. 12. 11. 12. 12. 11. 12. 12. 12. 13. 12. 12. 12. 12. 12. 12. 12. 12. 12. 12. 12. 12. 12. 12. 12. 13. 12. 13. 12. 13. 12. 12. 45 18 46 37 86 37 48 76 77 46 92 99 74 22 27 40 15 63 60 16 77 02 19 61 05 34 01 17 26 53 41 42 42 49 37 60 63 38 65 67 53 82 04 88 01 90 90

MK d mag. (8) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 8. 7. 7. 7. 7. 7. 7. 7. 7. 7. 7. 7. 6. 6. 6. 6. 6. 6. 6. 6. 6. 6. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 5. 0 9 7 6 6 6 5 3 3 1 1 0 9 7 5 5 5 2 2 1 1 1 9 8 8 8 7 7 7 6 5 5 4 4 4 4 3 3 3 3 2 2 1 1 1 1 0

Minit M (9)

e

EWf ° A (10) 128.2 152.3 218.2 230.5 199.3 166.3 150.2 206.0 106.7 51.8 ·· · 162.7 18.2 124.9 41.2 97.5 47.1 17.6 0.6 1.6 17.2 11.4 11.4 1.4 10.9 -0.7 22.1 6.7 3.5 1.5 1.6 8.7 23.2 6.3 8.4 ·· · 46.3 38.6 ·· · 10.7 6.6 2.2 -3.7 1.2 6.8 ·· · 5.0

FF190N er g s cm-2 s-1 ° A (11) 3. 1. 2. 3. 2. 3. 3. 2. 2. 1. 1. 9. 1. 1. 1. 6. 1. 4. 5. 1. 7. 6. 4. 2. 6. 8. 6. 6. 4. 5. 5. 5. 5. 6. 00e39e88e18e02e83e25e55e28e28e72e39e59e69e39e69e10e03e89e00e25e40e34e81e10e58e83e49e67e89e52e40e31e10e-

-1

N4 C9 AR3 B28 N1 C13 AR17 B34 N14 C11 AR7 B3 N11 C2 AR5 B17 N9 AR8 B22 N8 C8 AR1 B23 N10 C5 AR4 B21 N7 C6 AR2 B24 N 5 C1 B30 ··· N6 C3 AR16 B25 B31 B12 N12 CB B8 B19 B29 AR9 B20 A R6 ··· B7 B27 B2 ··· ··· B18 B16 B14 B9 ··· ··· B15 B13 B5 B10 ··· N10 C5 AR4 B21 B11 ··· ··· ··· ··· ··· ··· ··· ··· ···

00 75 20 83 29 87 53 46 80 83 03 01 08 24 24 22 89 58 81 90 36 24 50 42 26 60 31 77 08 20 87 53 08 93 51 19 54 51 91 59 22 37 76 80 11 18 49

0. 00 -3. 53 -4. 13 4. 66 -9. 64 -0. 03 2. 73 1. 01 10. 50 -4. 25 14. 41 4. 98 -1. 39 -0. 32 5. 67 1. 59 -4. 90 4. 34 -3. 72 2. 58 2. 65 5. 55 -1. 08 1. 55 -4. 30 -1. 27 2. 74 0. 55 4. 62 3. 66 2. 60 2. 41 2. 36 1. 22 5. 26 14. 87 2. 99 2. 28 -13. 94 3. 93 -4. 66 1. 52 3. 59 -2. 83 -1. 49 -5. 42 -2. 52

>120 >120 >120 >120 >120 >120 >120 >120 >120 >120 >120 >120 116.9 106.3 99.7 99.5 97.6 92.0 92.0 89.1 87.8 87.7 70.3 66.7 66.1 64.9 63.0 61.7 61.0 59.0 58.7 57.6 57.0 56.6 56.1 56.0 55.5 55.3 55.2 54.7 53.5 53.0 51.9 51.9 50.9 50.6 50.2

3.14e5.64e4. 1. 3. 2. 3. 3. 57e85e71e64e59e35e-

3.58e-

015 015 015 015 015 015 015 015 015 015 ··· 015 016 015 015 015 016 015 016 016 015 016 016 016 016 016 016 016 016 016 016 016 016 016 016 ··· 017 016 ··· 016 016 016 016 016 016 ··· 016


­ 26 ­

Table 3--Continued
ID
a

Desig./Ref. (2) ··· ··· ··· ··· ··· ··· ··· ··· ··· B11 ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ···

b

RA (3) 4. -1. -1. 10. 12. 5. 0. 6. 4. 6. -3. 11. 7. -1. 2. 7. 2. -2. 3. 2. 7. 7. -0. -2. 6. 5. 9. 7. 1. 6. 22. -2. 4. 5. 6. -2. 3. 2. 9. 3. -7. 9. 4. -3. 3. 1. 5.

c

DEC (4) -4. 14. -3. -1. -10. -2. 6. 0. 0. 2. 3. 8. 0. 23. 6. -2. 6. 1. 2. 7. 4. 2. -0. 6. 2. -2. -4. 11. 5. 2. -1. 4. 3. -1. 3. 2. 6. 4. -10. 6. 5. 5. -10. 5. 1. -0. -5.

c

(1) 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94

mF110W mag. (5) 18. 18. 19. 19. 18. 18. 18. 18. 18. 18. 18. 18. 18. 18. 18. 18. 18. 18. 18. 18. 17. 18. 18. 19. 18. 18. 18. 19. 18. 18. 19. 19. 18. 18. 18. 19. 18. 18. 18. 18. 20. 18. 19. 19. 18. 19. 19. 73 36 57 82 59 25 77 13 15 01 87 27 32 56 10 21 25 67 32 68 74 10 89 28 21 81 73 83 81 40 03 91 45 74 35 21 70 50 33 71 11 48 37 36 59 18 81

mF160W mag. (6) 15. 14. 15. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 15. 14. 14. 14. 15. 14. 14. 15. 15. 14. 15. 14. 15. 15. 15. 15. 15. 15. 15. 15. 15. 14. 15. 15. 00 59 38 99 67 51 78 57 54 53 90 63 56 69 57 64 69 90 63 94 33 52 94 32 62 96 98 32 98 74 08 41 92 02 93 34 08 03 00 10 55 07 38 52 95 43 91

mF205W mag. (7) 13. 12. 13. 12. 12. 12. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 12. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 14. 28 94 53 94 94 94 02 03 03 04 05 05 02 09 04 15 13 16 11 35 93 06 30 62 13 35 36 36 38 25 44 46 47 48 40 53 54 54 56 60 62 65 68 69 48 81 14

MK d mag. (8) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 5. 5. 5. 5. 5. 5. 5. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 0 0 0 0 0 0 0 9 9 9 9 9 9 9 8 8 8 8 8 7 7 7 7 7 6 6 6 6 6 6 5 5 5 5 5 4 4 4 4 4 3 3 3 3 3 3 3

Minit M (9) 50. 49. 49. 49. 49. 49. 48. 48. 48. 47. 47. 47. 47. 47. 46. 46. 45. 45. 44. 43. 43. 43. 43. 43. 42. 42. 42. 42. 41. 41. 40. 40. 40. 40. 40. 39. 39. 39. 38. 38. 37. 37. 36. 36. 36. 36. 36.

e

EWf ° A (10) -3.2 ·· · 2.1 -7.8 ·· · 6.4 -5.4 -10.4 -11.4 ·· · -8.7 ·· · 7.0 ·· · -0.5 -22.0 0.7 -5.6 -1.4 -4.2 -5.1 -0.4 -9.4 -3.7 32.1 -12.4 ·· · ·· · -14.2 -1.4 ·· · -9.4 -5.2 -10.9 -2.2 -9.4 -13.6 -8.8 ·· · -5.4 ·· · -7.1 -20.6 -6.8 -2.9 0.4 -1.1

FF190N er g s cm-2 s-1 ° A (11)

-1

77 74 65 80 43 96 14 93 35 67 84 54 08 53 20 60 91 34 57 82 82 72 10 63 88 50 63 42 07 71 07 45 33 95 26 62 50 94 53 39 98 14 08 70 39 52 22

20 97 41 62 55 25 95 15 57 45 16 42 58 67 15 92 05 23 34 66 61 44 63 33 49 82 73 51 70 93 95 65 30 48 63 09 48 16 83 18 53 61 21 24 31 02 27

2 9 9 9 9 7 4 1 1 9 8 8 0 0 4 0 9 7 9 8 7 5 2 1 5 3 2 2 9 3 8 5 3 1 1 3 1 1 7 1 8 3 9 8 8 5 5

2.39e-016 ··· 2.01e-016 2.94e-016 ··· 3.53e-016 2.75e-016 3.13e-016 3.33e-016 ··· 2.78e-016 ··· 3.06e-016 ··· 2.90e-016 2.74e-016 2.62e-016 2.61e-016 2.98e-016 2.16e-016 3.53e-016 3.13e-016 2.34e-016 1.72e-016 2.89e-016 2.37e-016 ··· ··· 2.04e-016 2.61e-016 ··· 1.98e-016 2.25e-016 2.00e-016 2.24e-016 1.80e-016 1.95e-016 1.85e-016 ··· 1.74e-016 ··· 1.84e-016 1.47e-016 1.39e-016 2.18e-016 1.57e-016 1.06e-016


­ 27 ­

Table 3--Continued
ID
a

Desig./Ref. (2) ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· B14 ··· ··· ··· ··· ··· ··· ··· B29 ··· ··· ··· ··· ··· ··· ··· ··· ··· ···

b

RA (3) 4. 6. 14. 11. 4. 10. 8. 8. -7. 4. 4. -4. -5. 7. 3. 4. 0. 3. -3. 7. 2. 3. 6. 2. 7. 5. -0. 1. 5. 6. 16. 8. 12. 12. -9. 4. -1. 7. 6. 3. 1. -2. 16. -0. 12. 3. -18.

c

DEC (4)

c

(1) 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141

mF110W mag. (5) 18. 18. 19. 19. 19. 18. 18. 19. 21. 19. 20. 99. 19. 19. 19. 19. 19. 18. 19. 18. 19. 19. 19. 19. 19. 19. 19. 19. 99. 19. 20. 19. 20. 19. 20. 19. 20. 19. 19. 20. 19. 20. 20. 20. 19. 19. 20. 73 47 02 16 42 71 69 19 30 13 07 00 11 29 19 04 47 94 80 97 27 53 14 66 29 82 51 24 00 49 41 12 41 69 61 94 04 61 25 15 91 53 92 61 75 19 93

mF160W mag. (6) 15. 15. 15. 15. 15. 15. 15. 15. 16. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 15. 16. 15. 15. 15. 16. 15. 16. 15. 15. 16. 16. 16. 16. 16. 15. 15. 16. 18 01 33 36 35 21 28 35 45 39 71 47 39 54 44 44 67 40 70 46 32 71 54 82 62 86 66 58 96 66 03 73 98 86 16 97 01 90 90 18 06 21 32 37 96 83 41

mF205W mag. (7) 13. 13. 13. 13. 13. 13. 13. 13. 14. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 13. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 71 54 73 75 76 77 78 78 53 79 81 82 87 89 91 93 94 87 97 99 83 01 06 08 06 09 09 09 11 11 15 16 16 18 20 21 23 27 27 28 29 30 32 35 36 38 39

MK d mag. (8) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 4. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3 3 2 2 2 2 2 2 2 2 2 1 1 1 1 0 0 0 0 0 0 0 9 9 9 9 9 9 9 9 8 8 8 8 8 8 7 7 7 7 7 7 7 6 6 6 6

Minit M (9) 36. 36. 36. 35. 35. 35. 35. 35. 35. 35. 34. 34. 34. 33. 33. 33. 33. 32. 32. 32. 32. 32. 31. 31. 31. 31. 31. 30. 30. 30. 30. 30. 30. 29. 29. 29. 29. 28. 28. 28. 28. 28. 27. 27. 27. 27. 26.

e

EWf ° A (10) -8.5 -5.3 ·· · ·· · -47.3 ·· · -10.0 -9.7 3.0 ·· · ·· · -169.5 ·· · ·· · -10.2 -11.2 ·· · -0.3 -8.0 -6.3 2.9 ·· · -8.0 -8.0 0.2 -14.9 -11.4 -13.6 -16.8 -8.5 ·· · ·· · ·· · ·· · ·· · -12.9 -17.2 ·· · -14.2 ·· · -13.2 -6.4 ·· · ·· · -10.7 -9.9 ·· ·

FF190N er g s cm-2 s-1 ° A (11)

-1

18 10 78 74 01 53 59 09 64 17 41 92 10 22 42 13 65 25 19 15 32 64 51 54 65 68 73 29 58 02 14 80 14 56 61 53 28 04 72 14 35 41 11 63 45 99 83

0. 3. 6. -14. 5. 5. 4. -6. 0. -21. -16. -0. -8. 11. 2. 6. 18. 6. 5. 3. 2. 16. 3. -3. 0. -7. 3. 5. 0. 5. 1. 19. -8. -13. 10. -4. -4. 20. 7. -13. -2. -5. 1. -16. -0. 4. 3.

76 03 19 30 67 15 39 04 18 01 84 47 23 83 06 41 90 40 29 69 72 48 32 21 01 20 12 48 41 50 01 13 76 77 04 74 70 08 79 78 97 08 50 45 65 28 95

4 3 2 8 6 6 4 4 3 3 9 8 1 7 5 2 1 9 7 3 3 1 4 1 1 1 0 9 8 7 1 0 0 8 4 3 1 5 5 4 3 1 9 5 3 2 9

1.67e-016 2.03e-016 ··· ··· 1.44e-016 ··· 1.63e-016 1.51e-016 6.98e-017 ··· ··· 1.41e-016 ··· ··· 1.42e-016 1.25e-016 ··· 1.31e-016 1.14e-016 1.27e-016 1.42e-016 ··· 1.25e-016 1.16e-016 1.13e-016 1.10e-016 1.06e-016 1.17e-016 1.04e-016 8.40e-017 ··· ··· ··· ··· ··· 1.01e-016 8.29e-017 ··· 1.03e-016 ··· 9.86e-017 8.96e-017 ··· ··· 8.77e-017 7.72e-017 ···


­ 28 ­

Table 3--Continued
ID
a

Desig./Ref. (2) ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· AR11 ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ···

b

RA (3) 5. 5. 16. 15. -5. 14. 20. 5. 2. 7. 5. 1. -3. 2. 5. -0. 10. 8. 3. 8. 7. 2. 7. 11. 5. 6. 3. 1. 5. -12. -0. -0. 6. 2. 0. 4. 7. 14. 0. 4. 4. -5. -3. 0. -3. 3. 8.

c

DEC (4)

c

(1) 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188

mF110W mag. (5) 19. 19. 20. 19. 19. 19. 19. 20. 19. 19. 20. 20. 20. 19. 20. 20. 99. 20. 19. 20. 20. 20. 19. 19. 19. 19. 20. 20. 19. 20. 20. 20. 20. 20. 20. 20. 21. 20. 20. 20. 19. 21. 21. 20. 21. 20. 20. 76 69 22 65 96 63 69 39 57 91 25 15 43 74 38 12 00 43 81 86 12 01 75 62 80 34 48 09 87 39 22 13 30 65 34 07 14 64 77 08 92 11 23 71 36 14 15

mF160W mag. (6) 15. 16. 16. 16. 16. 16. 16. 16. 15. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 15. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 16. 97 02 19 05 07 11 07 16 81 09 28 12 34 00 30 33 33 32 11 48 27 17 15 14 20 86 55 32 16 48 31 30 53 56 66 31 69 53 65 43 17 74 90 65 91 42 49

mF205W mag. (7) 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 14. 41 43 43 44 45 46 47 47 31 48 48 51 52 52 54 55 56 57 59 59 59 60 61 65 69 42 70 72 74 76 76 80 81 82 83 83 85 85 89 90 70 93 94 95 95 96 97

MK d mag. (8) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 6 5 5 5 5 5 5 5 5 5 5 5 5 4 4 4 4 4 4 4 4 4 4 3 3 3 3 3 2 2 2 2 2 1 1 1 1 1 1 1 1 0 0 0 0 0 0

Minit M (9) 26. 26. 26. 26. 26. 26. 26. 26. 25. 25. 25. 25. 25. 25. 25. 25. 24. 24. 24. 24. 24. 24. 24. 23. 23. 23. 23. 23. 23. 22. 22. 22. 22. 22. 22. 22. 22. 21. 21. 21. 21. 21. 21. 20. 20. 20. 20.

e

EWf ° A (10) ·· · -8.5 ·· · ·· · -9.7 ·· · ·· · ·· · 12.8 ·· · ·· · -5.7 ·· · -12.2 ·· · -15.2 -5.2 ·· · -12.4 ·· · ·· · 4.2 ·· · ·· · -10.4 9.8 -13.4 -6.3 -12.2 ·· · -4.7 -15.9 -9.2 ·· · -13.4 -7.8 ·· · ·· · ·· · ·· · 7.8 ·· · -11.0 -10.4 -18.3 6.6 -9.4

FF190N er g s cm-2 s-1 ° A (11)

-1

64 99 64 57 16 38 04 54 57 48 71 31 52 13 02 53 98 17 16 35 36 31 91 83 46 22 09 54 83 94 37 74 38 67 30 52 41 40 31 04 29 10 31 29 40 71 89

-2. 5. 2. 2. 7. 2. -5. 21. 3. -4. -13. 2. -10. 3. 20. -1. 0. 7. 6. -17. -16. -0. 14. 7. 3. 5. -7. 2. 1. 10. 5. 2. -5. 19. -1. 5. 15. 1. 23. 15. -0. -19. -6. -4. -5. -0. 0.

17 52 18 69 86 76 66 20 12 87 80 76 44 63 61 42 77 07 95 92 34 64 29 40 19 03 45 50 80 63 24 52 44 47 09 89 81 39 12 06 15 55 67 80 48 54 96

7 5 5 3 2 2 0 0 9 9 8 5 4 4 2 0 9 8 6 6 6 5 4 9 5 4 4 3 1 8 8 5 4 2 1 1 0 9 6 4 3 2 0 9 9 8 7

··· 8.40e-017 ··· ··· 7.53e-017 ··· ··· ··· 8.94e-017 ··· ··· 7.45e-017 ··· 7.38e-017 ··· 7.83e-017 6.58e-017 ··· 7.01e-017 ··· ··· 3.17e-017 ··· ··· 7.22e-017 8.94e-017 6.18e-017 6.15e-017 6.59e-017 ··· 5.87e-017 5.96e-017 5.83e-017 ··· 5.83e-017 5.61e-017 ··· ··· ··· ··· 6.92e-017 ··· 4.90e-017 5.32e-017 4.43e-017 5.01e-017 4.97e-017


­ 29 ­

Table 3--Continued
ID
a

Desig./Ref. (2) ··· ··· ··· ··· ··· ··· ··· ···

b

RA (3) 4. -6. 11. -2. 1. 3. 3. 2.

c

DEC (4) -0. 3. 4. 17. 13. 14. 7. 15.

c

(1) 189 190 191 192 193 194 195 196

mF110W mag. (5) 19. 20. 20. 20. 20. 21. 20. 20. 90 61 03 57 32 41 13 54

mF160W mag. (6) 16. 16. 16. 16. 16. 16. 16. 16. 24 72 51 66 60 96 48 69

mF205W mag. (7) 14. 14. 15. 15. 15. 15. 15. 15. 77 99 00 00 00 01 03 03

MK d mag. (8) - - - - - - - - 3. 3. 3. 3. 3. 3. 2. 2. 0 0 0 0 0 0 9 9

Minit M (9) 20. 20. 20. 20. 20. 20. 20. 20.

e

EWf ° A (10) 4.0 -18.3 ·· · ·· · -4.7 ·· · -8.0 ·· ·

FF190N er g s cm-2 s-1 ° A (11)

-1

40 58 56 78 61 56 55 67

76 00 59 82 20 83 46 02

7 6 5 4 4 3 1 1

5.87e-017 4.81e-017 ··· ··· 4.17e-017 ··· 4.44e-017 ···

a ID numb ers were assigned by sorting the data in order of increasing M the range 1.4 < m160 - m205 < 2.1

K

(decreasing luminosity), and rejecting stars outside

b Designations are taken from the following, in order of preference: (1) Nagata et al. (1995) (N#), (2) Cotera et al. (1996) (C#), (3) Lang et al. (2001) (AR#), and (4) Blum et al. (2001) (B#). Radio sources AR9-17 are newly identified in this paper, and their coordinates have been extracted from Figure 2 of Lang et al. (2001). c

Positions are with respect to RA(J2000) 17h 45m 50.26s DEC(J2000) -28 49 22. 76 and have a relative error of ±0.008 . M
K

d e

assumes (m

205

-K)0 =0, d=8000 pc (Reid 1993), (m

160

-m

205 )0

=-0.05 and A

K

= 1.95 â E(H - K).

Minitial assumes the relation between mass and magnitude for =2.5 Myr from the Geneva models with solar metallicity and enhanced mass-loss rates. Equivalent is calculated as EW = [(FF187N - FF190N )/FF190N ] F187N , where FF187N is corrected for extinction and the intrinsic shape of the spectral energy distribution.
f


­ 30 ­

Table 4. Mo del Parameters star# #8 #10a #10
b

#10

c

log L /L R / R T kK R =2/3 /R Teff kK v km s-1 log M /M yr-1 f mF205W mF160W mF110W EWP ° A

6.26 6.27 6.15 6.24 43.5 46.8 46.8 46.8 32.2 31.1 29.1 30.7 47.5 48.2 48.6 48.2 30.9 30.7 28.5 30.3 1100 1000: 1000: 1600: -4.35 -5.37 -4.79 -5.21 1.25 2.25 1.50 2.25 0.1 0.1 1.0 .1 10.71 11.44 11.45 11.43 12.50 13.33 13.35 13.32 16.27 17.41 17.43 17.40 242. 54.0 63.3 54.9 3.34 3.80 3.75 3.75 Ak a H 0.27 0.42 0.42 0.42 a 0.71 0.56 0.56 0.56 He a C 0.0002: 0.0008: 0.0008: 0.0008: Na 0.016 0.006: 0.006: 0.006: 49.9 49.8 49.5 49.7 log Q(H+ ) + log Q(He ) 48.5 48.4 47.6 48.2

a

Mass fraction

Note. -- Quantities followed by a colon are upper limits.


­ 31 ­ REFERENCES
Bieging, J. H., Abb ott, D. C., & Churchwell, E. B. 1982, ApJ, 263, 207 Blum, R. D., Schaerer, D., Pasquali, A., Heydari-Malayeri, M., Conti, P. S., Schmutz, W. 2001, AJ, accepted Bohannan, B., & Crowther, P.A., 1999, ApJ, 511, 374 Brown, R. L. & Liszt, H. S. 1984, ARA&A, 22, 223 Canto, J., Raga, A. C., & Rodr´ ez, L. F. 2000, ApJ, 536, 896 ´ igu Colgan, S. W. J., Erickson, E. F., Simpson, J. P., Haas, M. R., & Morris, M. 1996, ApJ, 470, 882 Conti, P. S., Hanson, M. M., Morris, P. W., Willis, A. J., & Fossey, S. J. 1995, ApJ, 445, L35 Cotera, A. S., Erickson, E. F., Simpson, J. P., Colgan, S. W. J., Allen, D. A., & Burton, M. G. 1992, American Astronomical Society Meeting, 181, 8702 Cotera, A. S. 1995, Ph.D. Thesis, Stanford University Cotera, A. S., Erickson, E. F., Colgan, S. W. J., Simpson, J. P., Allen, D. A., & Burton, M. G. 1996, ApJ, 461, 750 Crowther, P. A. & Dessart, L. 1998, MNRAS, 296, 622 Davidson, J. A., Morris, M., Harvey, P. M., Lester, D. F., Smith, B., & Werner, M. W. 1994, NATO ASIC Proc. 445: The Nuclei of Normal Galaxies: Lessons from the Galactic Center, 231 Eckart, A., Ott, T., & Genzel, R. 1999, A&A, submitted Erickson, E. F., Colgan, S. W. J., Simpson, J. P., Rubin, R. H., Morris, M., & Haas, M. R. 1991, ApJ, 370, L69 Figer, D. F., Na jarro, F., Morris, M., McLean, I. S., Geballe, T. R., Ghez, A. M., & Langer, N. 1998, ApJ, 506, 384 Figer, D. F. et al. 2000a, ApJ, 533, L49 Figer, D. F., McLean, I. S., Becklin, E. E., Graham, J. R., Larkin, J. E., Levenson, N. A., & Teplitz, H. I. 2000b, Proc. SPIE, 4005, 104 Figer, D. F., McLean, I. S., & Na jarro, F. 1997, ApJ, 486, 420 Figer, D. F., McLean, I. S., & Morris, M. 1995, ApJ, 447, L29 Figer, D. F. 1995, Ph.D. Thesis, University of California, LosAngeles


­ 32 ­
Figer, D. F., Kim, S. S., Morris, M., Serabyn, E., Rich, R. M., & McLean, I. S. 1999a, ApJ, 525, 750 Figer, D. F., Morris, M., Geballe, T. R., Rich, R. M., Serabyn, E., McLean, I. S., Puetter, R. C., & Yahil, A. 1999b, ApJ, 525, 759 Figer, D. F., et al. 2002, in preparation Genzel, R. et al. 1990, ApJ, 356, 160 Gerhard, O. 2001, ApJ, 546, L39 Hanson, M. M., Conti, P. S., & Rieke, M. J. 1996, ApJS, 107, 281 Herald, J.E., Hillier, D. J. & Schulte-Ladb eck, R.E. 2001, ApJ, 548, 932 Hillier, D. J. 1989, ApJ, 347, 392 Hillier, D. J. & Miller, D. L. 1998, ApJ, 496, 407 Hillier, D. J. & Miller, D. L. 1999, ApJ, 519, 354 Ho, L. C. & Filipp enko, A. V. 1996, ApJ, 466, L83 Illingworth, G. 1976, ApJ, 204, 73 Keyes, T., et al. 1997, HST Data Handb ook, Ver. 3.0, Vol. I (Baltimore: STScI) Krabb e, A., et al., ApJ, 447, L95 Kim, S. S., Morris, M., & Lee, H. M. 1999, ApJ, 525, 228 Kim, S. S., Figer, D. F., Lee, H. M., & Morris, M. 2000, ApJ, 545, 301 Kim, S. S., & Morris, M. 2002, in preparation Kurucz, R. L. 1979, ApJS, 40, 1 Lang, C. C., Goss, W. M., & Wood, D. O. S. 1997, ApJ, 481, 1016 Lang, C. C., Goss, W. M., & Rodr´ ez, L. F. 2001a, ApJ, 551, L143 igu Lang, C. C., Goss, W. M., & Morris, M. 2001b, AJ, 121, 2681 Leitherer, C., Chapman, J. M., & Koribalski, B. 1997, ApJ, 481, 898 McGinn, M. T., Sellgren, K., Becklin, E. E., & Hall, D. N. B. 1989, ApJ, 338, 824 McLean et al. 1998, SPIE Vol. 3354, 566


­ 33 ­
McLean et al. 2002, PASP, in preparation Meynet, G., Maeder, A., Schaller, G., Schaerer, D., & Charb onnel, C. 1994, A&A Supp., 103, 97 Meynet, G. 1995, A&A, 298, 767 Mizutani, K. et al. 1994, ApJS, 91, 613 Moneti, A., Glass, I. S. & Moorwood, A. F. M. 1994, MNRAS, 268, 194 Morris, M. & Yusef-Zadeh, F. 1989, ApJ, 343, 703 Morris, M., Davidson, J. A., & Werner, M. W. 1995, ASP Conf. Ser. 73: From Gas to Stars to Dust, 477 Na jarro, F. et al. 2002, in preparation Nagata, T., Woodward, C. E., Shure, M., & Kobayashi, N. 1995, AJ, 109, 1676 Nugis, T., Crowther, P. A., & Willis, A. J. 1998, A&A, 333, 956 Ozernoy, L. M., Genzel, R., & Usov, V. V. 1997, MNRAS, 288, 237 Perryman, M. A. C. et al. 1997, A&A, 323, L49 Portegies-Zwart, S. F., Makino, J., McMillan, S. L. W., & Hut, P. 2001, ApJ, 546, L101 Reid, M. J. 1993, ARA&A, 31, 345 Rieke, G. H., Rieke, M. J., & Paul, A. E. 1989, ApJ, 336, 752 Salp eter, E. E. 1955, ApJ, 121, 161 Scalo, J. 1998, in The Stel lar Initial Mass Function, G. Gilmore and D. Howell (eds.), vol. 142 of 38th Herstmonceux Conference, San Francisco, ASP Conference Series, p. 201 Schaller, G., Schaerer, D., Meynet, G., & Maeder, A. 1992, A&AS, 96, 269 Serabyn, E. & Guesten, R. 1987, A&A, 184, 133 Serabyn, E., Shup e, D., & Figer, D. F. 1998, Nature, 394, 448 Serabyn, E., Shup e, D., & Figer, D. F. 1999, ASP Conf. Ser. 186: The Central Parsecs of the Galaxy, 320 Serabyn, E., Figer, D. F., Kim, S. S., Morris, M., & Rich, R. M. 2002, ApJ, in preparation Stetson, P. 1987, PASP, 99, 191 Stolte, A., Greb el, E. K., Brandner, W., & Figer, D. F. 2002, ° submitted a,


­ 34 ­
Timmermann, R., Genzel, R., Poglitsch, A., Lutz, D., Madden, S. C., Nikola, T., Geis, N., & Townes, C. H. 1996, ApJ, 466, 242 van der Hucht, K. A. 2001, New Astronomy Review, 45, 135 Wendker, H. J. 1995, A&AS, 109, 177 Yusef-Zadeh, F., Cotera, A., Fruscione, A., Lang, C., Law, C., Wang, D., & Wardle, M. 2001, American Astronomical Society Meeting, 198, 8709 Yusef-Zadeh, F., Law, C., Wardle, M., Wang, Q. D., Fruscione, A., Lang, C. C., & Cotera, A. 2002, in preparation

A This preprint was prepared with the AAS L TEX macros v5.0.


­ 35 ­

Fig. 1a.-- (a) F205W image, after pro cessing by calnica and calnicb.


­ 36 ­

Fig. 1b.-- (b) ID's of bright stars.


­ 37 ­

Fig. 2.-- Difference image made by subtracting F190N image from F187N image. Positions of radio sources AR1-8 (circles) are from Lang, Goss, & Ro dr´ z (2001a). Radio sources igue AR9-17 (squares) are newly identified in this paper, and their co ordinates are taken from Figure 2 in Lang, Goss, & Ro dr´ z (2001a). Positions of x-ray sources (diamonds) are igue from Yusef-Zadeh et al. (2002).


­ 38 ­

Fig. 3.-- (a) Plot of EW1. F187N versus FF190N .

87

µm as a function of m

F205W

. (b) Linear plot of excess flux in

Fig. 4.-- Plot of EW1.87 µm as a function of color in m data points where the color error is less than 0.2.

F160W

-m

F205W

. The plot includes


­ 39 ­

Fig. 5a.-- Spectra of selected stars from Table 3. The spectra have not been dereddened, and the flux scale is arbitrary. The wavelength gaps in the spectra are due to incomplete coverage of the cross-dispersed echelle format by the detector. The two sharp absorption features near 2.32 µm are due to imperfect correction for telluric absorption. Other sharp features (a few pixels wide), especially in the left-most order, are similarly due to this imperfect correction, or are due to detector artifacts, c.f. the feature near 2.294 µm in the spectrum of star #2.


­ 40 ­

Fig. 5b.-- Same as Figure 5a.


­ 41 ­

Fig. 5c.-- Same as Figure 5a. The sharp features near 2.294 µm in the spectra of star #10 and #14 are due to detector artifacts.


­ 42 ­

Fig. 5d.-- Same as Figure 5a.


­ 43 ­

Fig. 5e.-- Same as Figure 5a.


­ 44 ­

Fig. 5f.-- Same as Figure 5a.


­ 45 ­

Fig. 5g.-- Same as Figure 5a.


­ 46 ­

Fig. 5h.-- Same as Figure 5a.


­ 47 ­

Fig. 5i.-- Same as Figure 5a.


­ 48 ­

Fig. 6.-- Comparison between average spectrum of Arches emission-line stars, spectra of representative WN7, WN8, and WN9 stars (Figer, McLean, & Na jarro 1997), and spectra of individual O If+ stars (Hanson et al. 1996). All spectra have been smo othed to match the resolution of the WR star spectra (R525). Notice that the WNL and Arches stars have N III features in common at 2.104 µm and 2.25 µm. In addition, the equivalent-widths and relative ratios of equivalent-widths are similar in the WNL and Arches spectra. The spectra are all on the same scale, but shifted by a constant for presentation purposes.


­ 49 ­

Fig. 7a.-- Observed (dots) and mo delled (line) spectrum of star #8.


­ 50 ­

Fig. 7b.-- Observed (dots) and mo delled (line) spectrum of star #10.


­ 51 ­

Fig. 8a.-- Plot of (F

F187N

-F

F190N

) â versus F

8.5 GHz

.

Fig. 8b.-- Plot of (F

F187N

-F

F190N

) â versus F

4.9 GHz

.


­ 52 ­

Fig. 9.-- Histogram of massive star separations from cluster center. Data are taken from Table 3.