------------------------------------------------------------------------ From: Ian Glass isg@da.saao.ac.za To: gcnews@aoc.nrao.edu Subject: newly accepted paper for GCNEWS %\documentstyle[epsf,referee]{mn} \documentstyle[epsf]{mn} \begin{document} \title{Luminous variables in the Quintuplet cluster} \author{I.S. Glass$^1$, S. Matsumoto$^2$, B.S. Carter$^3$ and K. Sekiguchi$^4$\\ $^1$South African Astronomical Observatory, PO Box 9, Observatory 7935, South Africa \\$^2$Institute of Astronomy, University of Tokyo, Mitaka, Tokyo 181, Japan\\ $^3$Carter Observatory, PO Box 2909, Wellington, New Zealand \\ $^4$Subaru Telescope, NAOJ, 650 North A'ohoku Place, Hilo, HI 96720, USA} \date{Originally sent to MNRAS 23 Oct 98; revised version sent 24 Dec 98; accepted 13 Jan 99} \maketitle \begin{abstract} We report observations of variability amongst the stars of the ``Quintuplet" cluster located about 30 pc in projection from the centre of the Galaxy. Two of the five cocoon star members, which may be protostars or peculiar Wolf-Rayets, are seen to vary slowly with moderate amplitude (0.4--0.5 mag). The bright star within the ``Pistol" H{\sc ii} region, suspected of being a Luminous Blue Variable (LBV), has in fact been found to show variability, confirming its tentative classification. A second nearby presumed LBV also varies. One of the apparent Quintuplet members is likely to be a Mira variable along the same line of sight. \end{abstract} \begin{keywords} Galaxy:center, Galaxy:open clusters, Stars:variables:others, Stars:supergiants \end{keywords} \section{Introduction} Infrared imaging studies of the inner bulge region have shown that there is a significant increase of the relative numbers of luminous stars towards the Centre of the Galaxy (Catchpole, Whitelock and Glass, 1990; Philipp et al., 1998). Spectroscopy and imaging of very red IRAS sources near the Centre (Moneti, Glass and Moorwood, 1992, 1994) revealed the existence of a number of compact H{\sc ii} regions and other luminous objects. In particular, the ``Quintuplet" cluster\,(AFGL 2004) (Okuda et al., 1990; Glass, Moneti and Moorwood, 1990, GMM) is, like the Galactic Centre cluster itself, rich in luminous stars, including emission-line objects and an M supergiant. Figer, McLean and Morris (1999, FMM) show that its massive stellar population includes many OB supergiants as well as WC and WN stars, indicative of star formation activity within the last 10$^6$--10$^7$ yr. A further cluster near the Centre, G0.121+0.017, possesses numerous hot luminous stars but does not have cool objects similar to those found in the other two (Nagata et al., 1995; Cotera et al., 1996; Serabyn, Shupe and Figer, 1998). Several members of the Galactic Centre cluster are thought to be Luminous Blue Variables (LBVs), a rare category of hot stars conspicuous in the Milky Way and some nearby external galaxies (e.g. Libonate et al., 1995). Near the Quintuplet, a bright H$\alpha$ and HeI emission-line star was found by Moneti, Glass and Moorwood (1994), who referred to it as the ``Serendipitous Source". This object has been classified, largely on spectroscopic grounds (including variable emission-line intensities) as a Luminous Blue Variable by Figer et al.\ (1998) and renamed the ``Pistol Star" after its surrounding radio ``Pistol" nebula. It may be the source of Ly continuum radiation responsible for the ionization of the Pistol radio continuum feature, though Figer et al.\ (1998) and FMM (1999) argue in favour of the hotter stars of the main cluster. At long wavelengths, ISOCAM images show that it is the centre of a bubble-like structure partly coincident with the Pistol radio nebula (A. Moneti, private communication). Its initial mass is estimated at $\sim$ 200 $M_{\odot}$ and its luminosity at $\geq$ 4 $\times$ 10$^6$ $L_{\odot}$ (Figer et al., 1998), also appropriate to its tentative classification as an LBV. A second LBV candidate in the neighbourhood of the Quintuplet has been identified more recently by FMM (1999). \section{Observations} We have been conducting a $K$-band (2.2\,$\mu$m) survey of the inner Bulge of the Milky Way with the intention of identifying luminous long-period variables. This region is obscured by interstellar dust at visible wavelengths. Typically, four observations of 25 fields were made each Galactic Centre season during the years 1994-97 (inclusive). The ``PANIC" camera that was used incorporates a large-format PtSi detector (effectively 1040 $\times$ 520 pixels; see Glass, Sekiguchi and Nakada, 1995). When used with the 0.75\,m telescope at Sutherland it covers a field of 5 $\times$ 5 arcmin$^2$. Taken together, the 25 fields form a survey of extent 24 $\times$ 24 arcmin$^2$, centered on the Galactic Centre. The field which includes the ``Quintuplet" is no.\ 13 of the 25. It is centered at 17h 45m 25.5s, --28$^{\circ}$ 50$'$ 50$''$ (2000). Stellar images from the individual exposures were extracted using DoPHOT (Schechter, Mateo \& Saha, 1993). The coordinates of each star were afterwards placed on a uniform system, which required translation and rotation of the coordinate systems of the individual images. Positional calibration was obtained by identifying visible stars that also appeared on the $K$ images from the astrometric catalogue of Monet et al.\ (1996). The photometry was adjusted to a uniform instrumental system by choosing (iteratively) a number of stars that remained constant throughout the series of exposures and performing a least-squares fit. At this point, variables could be identified by plotting the photometry for each star individually and examining the data by eye. This tedious procedure, rather than a simple statistical calculation, was necessitated by occasional spurious points which contaminate the light curves. When a potential variable was located, the data were Fourier analysed to determine its period and amplitude. The zero-point of the photometry was set by reference to two stars\footnote{GC13-2: $\alpha$ = 17h 46m 25.6s, $\delta$ = --28$^{\circ}$ 49$'$ 44$''$ (2000), $K$ = 7.65; GC13-3: $\alpha$ = 17h 46m 22.9s, $\delta$ = --28$^{\circ}$ 51$'$ 06$''$ (2000), $K$ = 8.19.} elsewhere in field 13 which have been observed with an aperture photometer (SAAO Mk {\sc iii}). The resulting Quintuplet photometry for uncrowded stars is in good agreement with GMM (1990) and should be accurate to 0.05 mag. It may be mentioned here that the effective wavelength of the PANIC $K$ filter is about 0.02\,$\mu$m shorter than that of the standard SAAO system. Figure 1 shows the the standard deviations of the photometry of the brightest 400 stars from sub-field 13. The spurious points were removed prior to this computation. The minimum standard deviation, achieved on the brighter stars, is about 0.04 mag. The points enclosed by circles are reckoned to be genuine variables. Other points with high standard deviations arise from crowding, which can readily be identified by placing an overlay of suspected variables on top of an image. A small number of deviant points are associated with bad data that occurred along the edges of the array, especially in corners and in the first season of observation. The minimum detectable amplitude clearly gets worse as fainter stars are considered. This is a result partly of photon statistics and partly of the very crowded nature of the Galactic Centre fields. Table 1 shows the positions, average $K$ magnitudes and standard deviations of the stars discussed in this paper, located towards the centre of the Quintuplet and in the nearby field. The stars are identified either by their Quintuplet membership numbers as given by GMM (1990) or by their DoPHOT running numbers. The positions should have errors less than 1 arcsec in each coordinate. Figure 2 shows a plan of the Quintuplet, with identifications. \begin{figure} \epsfxsize=8.4cm \epsffile[28 372 539 786]{fig1.ps} \caption{Statistics of 400 of the brightest stars in the complete 5 $\times$ 5 arcmin$^2$ field no.\ 13. Circled stars are near-certain variables. Other stars may show high standard deviations due to crowding and (occasionally) edge effects. Quintuplet stars 2, 3 and 5 are labelled, as are the two Luminous Blue Variables.} \end{figure} \begin{table} \centering \caption{Positions, average $K$ mags and standard deviations of $K$ mags of probable Quintuplet cluster members and some nearby variables.} \begin{tabular}{ll@{\hspace{1mm}}l@{\hspace{1mm}}ll@{\hspace{1mm}}l@{\hspace{1mm}}lll} Name & \multicolumn{3}{c}{R.A.} &\multicolumn{3}{c}{Dec} & Avg. & S.D.\\ (alternative) & \multicolumn{6}{c}{2000} & $K$ & of $K$ \\ Q1 (243) & 17 &46 &14.2 &-28 &49 &37 &7.61 &.04 \\ Q2 (231) & 17 &46 &14.7 &-28 &49 &41 &6.28 &.17 \\ Q3 (211) & 17 &46 &15.9 &-28 &49 &46 &6.91 &.16 \\ Q4 (251) & 17 &46 &14.8 &-28 &49 &35 &7.66 &.04 \\ Q5 LPV (270N) & 17 &46 &15.2 &-28 &49 &28 &8.62 &.49 \\ Q6 (250) & 17 &46 &15.4 &-28 &49 &35 &9.12 &.08 \\ Q7 (192) & 17 &46 &16.6 &-28 &49 &50 &7.57 &.05 \\ Q8 (240) & 17 &46 &16.0 &-28 &49 &38 &9.01 &.07 \\ Q9 (258) & 17 &46 &14.4 &-28 &49 &32 &8.98 &.06 \\ Q10 (241) & 17 &46 &15.2 &-28 &49 &38 &8.83 &.07 \\ Q11 (235) & 17 &46 &15.2 &-28 &49 &41 &10.31 &.26 \\ Q12 (278) & 17 &46 &15.2 &-28 &49 &35 &9.84 &.20 \\ Q14 (307) & 17 &46 &15.5 &-28 &49 &21 &9.12 &.04 \\ Q15 (270S) & 17 &46 &15.1 &-28 &49 &30 &9.49 &.11 \\ D004 Serend (134)& 17 &46 &15.3 &-28 &50 &04 &7.38 &.15 \\ D006 LBV (362) & 17 &46 &18.0 &-28 &49 &03 &7.43 &.26 \\ D018 LPV & 17 &46 &14.4 &-28 &50 &03 &8.65 &.27 \\ D020 LPV & 17 &46 &17.4 &-28 &50 &14 &8.34 &.18 \\ D200 LPV & 17 &46 &14.9 &-28 &48 &43 &10.32 &.56 \\ D215 (256) & 17 &46 &16.6 &-28 &49 &32 &10.21 &.06 \\ D230 LPV & 17 &46 &15.6 &-28 &50 &24 &10.19 &.29 \\ D256 & 17 &46 &15.1 &-28 &49 &12 &10.48 &.09 \\ D271 (301) & 17 &46 &16.1 &-28 &49 &22 &10.54 &.13 \\ D278 & 17 &46 &14.1 &-28 &49 &17 &10.52 &.08 \\ D288 & 17 &46 &15.1 &-28 &49 &41 &10.37 &.23 \\ D296 LPV? & 17 &46 &18.9 &-28 &49 &41 &10.36 &.23 \\ D307 & 17 &46 &15.7 &-28 &49 &18 &10.83 &.15 \\ D309 (242) & 17 &46 &14.5 &-28 &49 &37 &10.95 &.37 \\ D322 (269) & 17 &46 &15.5 &-28 &49 &30 &10.65 &.16 \\ D334 & 17 &46 &15.9 &-28 &49 &40 &11.06 &.32 \\ D3604 & 17 &46 &16.8 &-28 &49 &52 &10.25 &.12 \\ D3605 & 17 &46 &16.1 &-28 &49 &48 &11.08 &.28 \\ D3606 & 17 &46 &14.5 &-28 &49 &41 &10.68 &.40 \\ \end{tabular} \vspace{1mm} \noindent Note: The alternative numbers are those used by FMM (1999). The cross-identifications are based on a table of coordinates kindly provided by D.F. Figer (private communication). \end{table} \begin{figure} \epsfxsize=8.4cm \epsffile[28 192 539 786]{fig2.ps} \caption{Identification of Quintuplet stars and neighbouring variables. The circled stars have numbers from GMM (1990). The D numbers are DoPHOT running numbers from the current work.} \end{figure} \begin{figure} \epsfxsize=7.5cm \epsffile[28 -9 457 786]{quin2.ps} \caption{$K$ light curves of variable stars discovered in the neighbourhood of the Quintuplet. The stars are identified by the DoPHOT numbers. Each vertical tick mark is 0.5 mag. Note that star D4 is the Serendipitous/Pistol star and star D6 is the other suspected luminous blue variable (LBV). Most of the other stars shown are large-amplitude AGB variables. Error bars, based on fig.\ 1, are shown for D200.} \end{figure} Figure 3 shows light curves of variables identified in the outer neighbourhood of the Quintuplet cluster. Nos D4 and D6 are the LBVs. The other variables, except D296, are plausibly identified as being of long period. $K$-light curves of 26 closer-in stars which probably belong to the Quintuplet cluster are shown in Fig.\ 4. Many stars (Quintuplet nos 1, 4, 6, 7, 8?, 9, 10, 12?, 14, 15?; D215, D256, D271, D278 and D3604) do not seem to vary at our level of detectability. Only upper limits to the variability of the fainter stars can be inferred from the figure because of crowding. \section{Discussion} \subsection{Cocoon Variables} From Fig.\ 4, Quintuplet stars 2 and 3 show peak-to-peak variations at $K$ of 0.46 $\pm$ 0.06 and 0.44 $\pm$ 0.06 mag respectively. These are members of the eponymous quintuplet of cocoon-like objects. Crowding and the effects of variable seeing conditions are unlikely to cause false variations in spite of the proximity of D3606 to Q2 and D3605 to Q3, because these objects are typically more than 4 mag fainter at $K$ than their bright neighbours. In addition, the variations appear to be smooth, unlike the case when crowding has occurred. \begin{figure*} \begin{minipage}{17.5cm} \epsfxsize=17.5cm \epsffile[28 195 539 786]{quin1.ps} \caption{$K$ light curves of probable members of the Quintuplet. Each vertical tick mark is 0.5 mag. A bright constant star will have a standard deviation of about 0.04 mag in this field. The scatter in many panels can be due to crowding or faintness: this figure should only be used to establish upper limits to the variability of the sources. Only stars Q2, Q3 and Q5 are likely to be real variables. Q2 and Q3 are members of the eponymous quintuplet and are cocoon stars. Q5 appears to be a Mira variable and is probably not a cluster member.} \end{minipage} \end{figure*} Although early workers (Okuda et al., 1989; GMM, 1990) favoured the idea that the cocoon objects are protostars, others (Figer, Morris and McLean, 1996) believe that they are late WC (Wolf-Rayet) types with extreme dust shells. Unfortunately, they show no intrinsic spectral features except the silicate dust absorption bands. Because many of the less conspicuous (in the infrared) cluster stars are OB supergiants or have Wolf-Rayet (WN or WC) characteristics, it is natural to suggest that the cocoon stars are massive objects, somehow related to them. Their dust shells have temperatures in the range 700--900\,K, but their nature remains ambiguous since the $J$-band spectral features, expected in late-type WC stars, have not been observed (FMM, 1999). The discovery, reported here, that two of these objects are variable, apparently quite smoothly and with a time-scale of order 1--2 years, may help to constrain their nature. It has been suggested by Bernasconi and Maeder (1996) that the apparent lack of very massive ($\leq$ 120 $M_{\odot}$) O-stars close to the ZAMS may be due to their having a much prolonged accretion phase with a duration of 2--2.5 Myr, so that considerable nuclear evolution occurs before they become visible. The duration of this obscured phase is comparable to the ages of the other Quintuplet cluster members as given by FMM (1999). It is therefore suggested that these self-obscured O-stars may be an alternative way to explain the eponymous Quintuplet cocoon stars. Quite possibly the variations that have been noted in two of them arise from instabilites similar to those that characterize the luminous blue variables, themselves representive of a later stage of massive-star evolution. \subsection{Luminous Blue Variables} The Luminous Blue Variables (LBVs), also known as Hubble-Sandage Variables, form a small group of stars which have been noted in the Galaxy, the LMC, M31, M33 and elsewhere. Their properties have been summarized by Humphreys and Davidson (1994). Variations of several types may occur: giant eruptions ($\geq$ 2 mags), eruptions (1--2 mags) on time scales of 10--40 years, oscillations of $\sim$ 0.5 mag on time scales of months to a few years and microvariations of $\leq$ 0.1mag. Their bolometric luminosities lie in the range --8.8 to $>$--11. Fig.\ 3 shows that both the spectroscopically identified LBV candidates near the Quintuplet stars are clearly variable. This result provides confirmation of their classification as luminous blue variables.The Serendipitous source has a peak-to-peak $K$-amplitude of 0.47 $\pm$ 0.06 mag and star 6 has a peak-to-peak amplitude of 0.92 $\pm$ 0.06 mag. Neither object appears to have a regular period on the time-scales covered. The locations of the two possible LBVs are somewhat distant from the densest part of the Quintuplet cluster (see Fig.\ 2; ``serend" and ``D6"), suggesting that they may be amongst the older members of the association as a whole. However, isolated WC, WR and OB stars are also found at comparable distances from the centre of the cluster (Figer et al., 1998). \subsection{Long-Period Variables} Six stars in the area covered by Fig.\ 1 appear to be long-period variables. Four of these, nos Q5 (680d), D18 (466d), D200 (654d) and D230 (259d) have regular light curves, while the other two, nos D20 (415d) and D296 (304d), are less regular. There is a possibility that the apparent variability of no.\ 296 is simply due to errors of measurement. The apparent Quintuplet member Q5 shows variability at $K$ with amplitude $\sim$1.5 mag and a period of 680d. Its spectrum is given as ``late" (FMM, 1999). If, as seems most likely, it is a Mira variable, it cannot be a cluster member for reasons of age, and must be a chance superposition. No.\ 200 is within a few arcsec of the position of the OH/IR star OH 0.178-0.055 (Sjouwerman et al., 1998), given as 17h 46m 15.309s, -28$^{\circ}$ 48$'$ 41.36$''$ (2000). However, other OH/IR stars detected in this survey have positions in closer agreement with the OH work than this one. Long-period variables are common in the 24 $\times$ 24 arcmin$^2$ field of the present survey. Several hundred have been identified so far, occurring at a rate of about 1 arcmin$^{-2}$ in field 13. (The first results of our survey were published by Glass et al., 1996, and a more recent discussion will be found in Glass et al., 1999). It is not suggested that the long-period variables are associated with the cluster. \begin{thebibliography}{} \bibitem[]{} Bernasconi P.A., Maeder A., 1996 A\&A, 307, 829 \bibitem[]{} Catchpole R.M., Whitelock P.A., Glass I.S., 1990 MNRAS, 247, 479 \bibitem[]{}Cotera A.S., Erickson E.F., Colgan S.W., Simpson J.P., Allen D.A., Burton M.G., 1996 ApJ, 461, 750 \bibitem[]{} Figer D.F., Najarro F., Morris M., McLean I.S., Geballe T.R., Ghez A., Langer N., 1998 ApJ, 506, 384 \bibitem[]{} Figer D.F., McLean I.S., Morris M., 1999 (FMM), ApJ, in press \bibitem[]{} Figer D.F., Morris M., McLean I.S., 1996, In {\it Proc. 4th ESO/CTIO Workshop: The Galactic Center}, R Gredel, PASP Conf. Ser. {\bf 102}, p.\ 263. \bibitem[]{} Glass I.S., Matsumoto S., Ono T., Sekiguchi K., 1996 In {\it Proc. 4th ESO/CTIO Workshop: The Galactic Center}, R Gredel, PASP Conf. Ser. {\bf 102}, p.\ 312. \bibitem[]{} Glass I.S., Matsumoto S., Carter B.S., Sekiguchi K., 1999 In {\it IAU Symp.\ 191, Asymptotic Giant Branch Stars}, to be published. \bibitem[]{} Glass I.S. Moneti A., Moorwood A.F.M., 1990 (GMM), MNRAS, 242, 55p (see also 244, 767) \bibitem[]{} Glass I.S., Sekiguchi K., Nakada Y., 1995 In {\it New Developments in Array Technology and Applications}, A.G. Davis Philip et al., Kluwer, Dordrecht, p.\ 109 \bibitem[]{} Humphreys R.M., Davidson K., 1994 PASP, 106, 1025 \bibitem[]{} Libonate S., Pipher J.L, Forrest W.J., Ashby M.L.N., 1995 ApJ, 439, 202 \bibitem[]{} Monet D. et al., 1996 {\it USNO-A1.0: A Catalog of Astrometric Standards}, US Naval Observatory, Washington, DC (CD-ROM set). \bibitem[]{} Moneti A., Glass I.S., Moorwood A.F.M., 1992 MNRAS, 258, 705 \bibitem[]{} Moneti A., Glass I.S., Moorwood A.F.M., 1994 MNRAS, 268, 194 \bibitem[]{} Nagata T., Woodward C.E., Shure M., Kobayashi N., 1995 AJ, 109, 1676 \bibitem[]{} Okuda H., Shibai H., Nakagawa T., Matsuhara H., Kobayashi Y., Kaifu N., Nagata T., Gatley I., Geballe T.R., 1990 ApJ, 351 89 \bibitem[]{}Philipp S., Zylka R., Mezger P.G., Duschl W.J., Herbst T., and Tuffs R.J., 1998 A\&A, in press. \bibitem[]{} Schechter P.L., Mateo M., Saha A., 1993. PASP, 105, 1342 \bibitem[]{} Serabyn E., Shupe D., Figer D.F., 1998. Nature, 394, 448 \bibitem[]{} Sjouwerman L.O, van Langevelde H.J., Winnberg A., Habing H.J., 1998. A\&AS, 128, 35 \end{thebibliography} \end{document}