Nhat Is the Mass-To-Light Ratio of the Old Magellanic Globular Cluster NGC 1835? G
Total Page:16
File Type:pdf, Size:1020Kb
~ra to a common intensity scale (no scal acceptable loss to guarantee the quality in spectral modelling analysis with a sin Ing in intensity was made here), and that of the photometry. The data in Figure 3 gle model profile. But of course, the t,here is no significant detector non were processed with these columns derivation of the appropriate profile re i1nearity present (which would be mani masked away. mains problematical. fest as residual differences at the posi One negative side effect results, how tions of strong Iines). Furthermore, the ever. With two fewer signal columns, the difference of the spectra has an rms fitting algorithm now no longer effective Conclusions noise very near to the calculated photon Iy ignores the single pixel outliers due to shot noise limit, given the quoted 6.8 particle detections. We have had to in Based on our, admittedly incomplete, electron/ADU calibration. clude a routine, therefore, to test for analysis, we feel we can make the From these and other results from the pixels more than 5 times the (a priori following conclusions. rUn it is clear that apparent S/N ratios of known) noise sigma from the fit, and (i) The double density RCA CCO on 300 : 1 are achievable with this detector. throw the worst single one out. Although the CES long camera works very weil, ~igher ratios may be possible with mul not particularly elegant, this strategy has even at 4000 A. tiPle integrations. proven very effective in removing parti (ii) Its lower noise per pixel compared The only serious problem found in our cle detections. to the Reticon, and its registry of each data is shown in Figure 2b. The two spectral channel with multiple pixels, Columns marked there consistently allows particle detections to be discov ered and easily removed. Show residuals (data-minus-fit) depen Focus Effects dent on signal level. We understand (iii) The expected quantum noise limit (Sandro O'Odorico, private comm.) that Our data-minus-fit residual frames are is achieved on single integrations, allow these RCA chips are known to exhibit exquisitely sensitive to focus variations ing S/N ratios of several hundred to be such behaviour - that is, to have occa along the spectra (although the final in obtained. We have not done tests to S~onal column pairs in which part of the tegrated intensities at each point should determine whether S/N ratios of 1,000 Signal in one column seems to end up in not bel. We find that the width of the and greater are possible, by summing the adjacent column, when the signal is spectra does vary, being broadest at the multiple integrations. above some threshold level. We have two ends, but that the effect is so small (iv) Photometrically unacceptable col experimented with trying to fix this prob that we cannot measure it in the widths umns on the detector have been discov lem by applying an interpolated re of individual emission lines in the cali ered, which should be masked out dur transfer of signal after the fact, but could bration spectra. We suppose that a ing analysis if results of the highest qual not Convince ourselves that the results small tilt of the CCO with respect to the ity are to be attained. were always reliable. focal plane of a half degree is sufficient (v) Least squares fitting of templates Gur solution to the problem is to mask to give the magnitude of what we see. for data extraction, and probably also the offending columns of data away, From the residual frames we infer that for wavelength calibrations, seems a ~n? exclude them from the template over about a quarter of the total length good way to determine the length of ~ttlng process. We thereby lose some of the recorded data, the instrumental spectrum over which the instrumental 5% of our signal, whichwe deem an profile is clearly constant enough for use profile is sensibly constant in shape. 'Nhat is the Mass-to-Light Ratio of the Old Magellanic Globular Cluster NGC 1835? G. MEYLAN, ESO p. DUBATH, M. MA YOR, Observatoire de Geneve, Switzerland, and p. MAGAIN, Institut d'Astrophysique de Liege, Belgium 1. Richness of the Southern Sky From the determinations found in the vefocity dispersion, by detecting the h We astronomers are lucky: our Galaxy literature of the individual masses of the very small line broadening present in the as two companion galaxies, the Large richest old clusters, a systematic differ integrated light spectra. and Small Magellanic Clouds, situated ence seems to exist between the globu lars in our Galaxy and in the Magellanic weil above the galactic plane, which 2. Magellanic and Galactic Clouds, Magellanic clusters appearing Contain a huge potential of astrophysi M/LvRatio ?al information. For example, concern fighter than galactic clusters. This differ :ng star clusters, the realm of the globu ence in mass between old rich 2. 1 Mage/lanic globular clusters ar clusters is much richer and more Magellanic and galactic clusters obvi Varied in the Magellanic Clouds than in ously has direct consequences on the The method most often used for ob the Galaxy: rich clusters of all ages are mass-to-light ratio determination of the taining the total mass of Magellanic observed, from the youngest, having considered clusters, reflecting perhaps clusters is related to the systemic rota ages of a few tens 106 yr to the oldest systematic differences in mass function. tion of the Magellanic Clouds. The ob h ' " 1aVlng ages of the order or larger than This was challenged and discussed re served value of the tidal radius r of the 9 t o10 yr. In this paper, only old cently (Meylan 1988 b). A way to resolve cluster is transformed into mass, in a Magellanic and galactic globular clus this controversy consists in obtaining way similar to the case of galactic open ters are considered. good observational values of the central clusters. It is assumed that the clusters 55 are in rotation along circular orbits TABLE 1: The different values of the total mass of NGC 1835 published during these last around the centre of mass of the 15 years. Clouds, the old clusters seeming to form a disk-like system in the LMC. Tidal Year x Mtet M/Lv ra Authors 6 [G units) masses, particularly for the outer clus [10 MG) [rcl ters, may be underestimated, if these 1974 " . 0.045 0.2 ., . Freeman 1974 clusters are in radial rather than circular 1978 " . 0.044 0.12 ... Chun 1978 orbits. 1978 " . 0.062 0.17 . .. Chun 1978 Observational determinations of the 1985 " . 0.073 0.18 . .. Elson & Freeman 1985 tidal radii by star counts in the outer 1985 ... 0.16 0.42 iso Elson & Freeman 1985 parts of Magellanic clusters is pioneer 1988 1.75 0.39 1.30 iso Meylan 1988b work of a difficult nature, since the 1988 1.50 0.28 0.94 30 Meylan 1988b pollution by Magellanic fjeld stars still 1989 1.25 1.03 3.58 iso present study induces uncertainties. Determination of 1989 1.00 0.81 2.83 30 present study tidal radius is a difficult task even for the galactic globular clusters, and the "ob served limiting radius" determinations are rather weak for nearly all of them. from (J) Cen, the (unique) giant cluster of nation processes. Comparison between 6 Using the above method, Elson and the Galaxy (Mtot = 3.9 10 MG), the total galactic and Magellanic M/Lv values Freeman (1985) found for the following masses range from 0.4 to 1.1 106 MG, should be done only between results three old LMC clusters, NGC 1835, whereas all the mass-to-light ratios are coming from the same kind of models NGC 2210, and NGC 2257, total masses located between about 2 and 3. The constrained by the same kind of obser equal to, respectively, Mtot = 7.3, 1.9, above values can be considered as typi vational data. The more elaborated and 3.7 104 MG, with corresponding cal of the masses and mass-to-light King-Michie dynamical models have mean M/Lv ratios equal to 0.18, 0.11 and ratios of the rich globulars of our Galaxy. been applied so far only to galactic 0.56 (fable 1). globular clusters (with the exception of Reasonably good dynamical con NGC 1835), due to the lack of observa 2.3 A difference in M/L bya factor of straints - surface brightness profile and v tional data concerning the Magellanic 10? central value of the velocity dispersion clusters. have been published so far only for one The typical mass of the globular clus This situation is on the verge of Magellanic cluster: NGC 1835. It is only ters in the Clouds (Iess than 105 MG) is change: if it is still difficult to obtain in the case of this cluster that the deter smaller than the typical mass of the velocity dispersion profiles of mination of the velocity dispersion (o(Vr) globular clusters in the Galaxy (greater Magellanic clusters, it appears noW = 5 km S-1, obtained by Eison and Free than 105 MG)' This difference in mass feasible to obtain at least the central man (1985) from a Fourier method ap between rich galactic and Magellanic value of the velocity dispersion, from plied to integrated light spectra) can be clusters obviously has direct conse integrated light spectra. The obtaining converted into mass, by using the core quences on the mass-to-light ratio: M/Lv of such an essential observational con radius and the dimensionless mass de = 0.1-0.5 for those in the Magellanic straint, in the case of the old Magellanic rived from the fit ofthe observed surface Clouds and M/Lv = 2.0-3.0 for those in globular NGC 1835 is presented below, brightness profile to single-mass iso the Galaxy.