4 Bulge objects

Since care is necessary to interpret reddened integrated spectra in such crowded regions, we discuss in this section individual clusters comparing the present results with previous ones from the literature, especially with those from CMDs, when available.

Different techniques have been most frequently used to study such clusters: Malkan (1982, hereafter M82) derived reddening from integrated infrared photometry; ZW84 and Zinn (1985, hereafter Z85) based on the same data, estimated metallicities $.\!\!^{\rm h}$. By means of near-infrared integrated spectra AZ88 derived $.\!\!^{\rm h}$ from the CaII triplet, and reddening from the interstellar band at 8621Å. Origlia et al. (1997) derived $.\!\!^{\rm h}$ by means of infrared integrated spectra using a calibration of the 1.62$\,\mu{\rm m}$ feature (mostly CO). Minniti (1995) employed ($J,\ K$) photometry and visible spectroscopy of giants in the clusters. Finally, information from ($V,\ I,\ z$) and ($J,\ K$) CMDs are available for many clusters as discussed below in conjunction with the present spectroscopic results. In particular, ($V,\ I$) CMDs are a very sensitive metallicity discriminator among metal-rich clusters, owing to curvature effects on the GB by blanketing effects (Ortolani et al. 1991).

Taking into account the growing body of evidence that the HB morphology is related to the cluster structure (Buonanno et al. 1997 and references therein), we also indicate for each cluster the HB characteristics from CMDs, and structural analysis results, whenever available, in view of a future test of these ideas for bulge clusters, as improved CMDs will become available.

4.1 NGC6256

In an early photographic study, Alcaino (1983) obtained a B,V CMD with detection limit at the HB level, and estimated E(B-V)=0.80 and a metallicity comparable to that of 47Tuc. However, a deeper CCD photometry (Ortolani et al. 1998) shows a blue-extended HB similar to that of the intermediate metallicity cluster NGC6522 (Barbuy et al. 1994).

The derived spectroscopic $\mbox{$[Z/Z_{\odot}]$}=-1.01$ (Table 3), indicates a lower metallicity than that of 47Tuc. This value might still be an upper limit, owing to possible metal-rich field star contamination. We remind that the cluster NGC6256 is in a post-core collapse phase (Trager et al. 1995, hereafter TKD95). NGC6256 appears to belong to an intermediate metallicity bulge family (IMBF) like NGC6522 itself and NGC6540 (Bica et al. 1994), which are characterised by blue-extended HBs and a post-core collapse structure.

4.2 TBJ3 = TJ13

This object at $\alpha(1950)=17^{\rm h}~15^{\rm m}~06^{\rm s}$ and $\delta(1950)=-27\mbox{$^{\circ}$}~51\mbox{$^{\prime}$}$, is the only remaining globular cluster candidate in the lists of Terzan & Bernard (1978) and Terzan & Ju (1980). Djorgovski et al. (1990) found that TBJ1 = TJ17 with radial velocity $V_R=6270\mbox{$\rm\, km\, s^{-1}$}$ and TJ15 with $V_R=8619\mbox{$\rm\, km\, s^{-1}$}$ are galaxies. They also reported images of these objects together with TBJ2 = TJ16 indicating their galaxy nature. Finally, TJ5 and TJ23 are planetary nebulae (Bica et al. 1995).

The integrated spectrum (Fig. 1) presents no evidence of a redshifted component, indicating that the dominant light source is not a background galaxy. Terzan & Bernard (1978) presented an R plate of TBJ3 and described the object as circular and diffuse with a diameter of $\approx10\mbox{$^{\prime\prime}$}$. An R CCD image taken in May 1994 with the Danish telescope at ESO, La Silla, showed that the central clump is partially resolved into stars, and a faint halo of stars might be present (Ortolani, private communication).

The spectral properties are compatible with those of a metal-rich globular cluster like NGC6528 (Table 2 and Fig. 5). The derived reddening (E(B-V)=0.95) is not as large as those of the very reddened bulge clusters (Table 3). As discussed by e.g. Aguilar (1993), there are several dynamical processes which might lead to the destruction of globular clusters, such as mass loss, tidal shocks, dynamical friction, and disruption of clusters venturing the central 2kpc. In this respect the appearance and spectrum of TBJ3, together with the moderate reddening for this faint object, places TBJ3 as a candidate for a low luminosity post-core collapse fossil of a globular cluster. We note, however, that some bulge stars might be dominating the observed spectrum, and the derived reddening would be, in this case, a lower limit so that TBJ3 might be a more distant and luminous object. Deep CMDs would be necessary to establish the nature of this interesting object.

4.3 Terzan2

The different techniques provided quite similar reddening and metallicity values. The infrared integrated photometry indicated E(B-V)=1.3 and $\hbox{$.\!\!^{\rm h}$}=-0.47$ (M82/ZW84/Z85); AZ88 derived E(B-V)=1.54 and $\hbox{$.\!\!^{\rm h}$}=-0.25$; Kuchinski et al. (1995) employed K, (J-K) CMDs and estimated $\hbox{$.\!\!^{\rm h}$}=-0.25$; Ortolani et al. (1997a) employed V, (V-I) CMD and derived E(B-V)=1.54 and $\hbox{$.\!\!^{\rm h}$}=-0.55$.

The reddening and metallicity derived in Table 3 are consistent with the previous results. This cluster belongs to the metal-rich bulge family (MRBF) and is in a post-core collapse phase (TKD95).

4.4 Terzan4

From integrated infrared photometry, M82/ZW84/Z85 obtained E(B-V)=1.5 and $\hbox{$.\!\!^{\rm h}$}=-0.21$; AZ88 found E(B-V)=1.52 and $\hbox{$.\!\!^{\rm h}$}=-0.94$ from near-infrared spectroscopy; Ortolani et al. (1997) used a deep ESO NTT V, (V-I) CMD (under exceptional seeing conditions 0$.\!\!^{\prime\prime}$35 - 0$.\!\!^{\prime\prime}$55) and detected a blue HB and a general morphology which might place this cluster as metal-poor as $\hbox{$.\!\!^{\rm h}$}=-2.0$. They derived a high reddening E(B-V)=2.35, which suggests that contamination by foreground field stars affected previous determinations.

In the present study we have eliminated pixel rows to avoid two obvious foreground (less reddened) bright stars (Sect. 2). Indeed, the resulting E(B-V) for the cluster (Table 3), is close to that derived from the CMD. However, the metallicity (Table 3) is high as compared to that from the CMD, suggesting that metal-rich bulge stars reddened as much as the cluster affect the present spectrum. We conclude that Terzan4 is not a member of the MRBF.

4.5 HP1

The near-infrared spectroscopy provided E(B-V)=1.44 and $\hbox{$.\!\!^{\rm h}$}=-0.56$ (AZ88). From individualgiants, Minniti (1995) obtained E(B-V)=1.68 and $\hbox{$.\!\!^{\rm h}$}=-0.30$, and Minniti et al. (1995) favoured E(B-V)=1.25 and $\hbox{$.\!\!^{\rm h}$}=-0.5$ from a K, (J-K) CMD. On the other hand, the deep V, (V-I) CMDs in Ortolani et al. (1997b) revealed that HP1 has a blue extended HB and a general morphology of a globular cluster with $\hbox{$.\!\!^{\rm h}$}\approx-1.5$; they obtained a reddening E(B-V)=1.19.

In the present study we analyse two spatial extractions of HP1: that of Table 1 corresponding to 83$^{\prime\prime}$ along the slit, and a core extraction of 10$^{\prime\prime}$ along the slit. The observed and reddening-corrected spectra are shown in Fig. 8, while the equivalent width values are given in Table 2. The core spectrum provides a low metallicity $\mbox{$[Z/Z_{\odot}]$}=-1.09$ (Table 3), while the total spectrum is considerably contaminated by metal-rich stars ($\mbox{$[Z/Z_{\odot}]$}=-0.37$). This radial change explains why high values of metallicity have been obtained in previous studies. The present metallicity of the core is possibly still an upper limit, but taking into account the CMD results of Ortolani et al. (1997b), we confirm that HP1 belongs to the IMBF, likewise NGC6256 (Sect. 4.1) with post-core collapse (TKD95). The present reddening value is consistent with the previous lower values.

  

Figure 8: HP1: different extractions. Top panel, observed spectra; bottom panel, reddening-corrected

Interestingly, metal-rich stars at the same distance as the cluster are seen in the CMD (Ortolani et al. 1997b). They define tighter sequences than would be expected from depth and differential reddening effects for the bulge field, which led Bica et al. (1997) to consider the possibility of capture of such stars by globular clusters in dense bulge fields. The mechanism appears to be efficient enough to change significantly the stellar population content of a globular cluster over a Hubble time. The metallicity contrast produced by this effect would be more pronounced in metal-poor clusters like HP1. It is expected to operate as well in metal-rich bulge family clusters, although in this case the effect would be difficult to detect.

Another interesting mechanism which might create composite CMDs is that of mergers (Catelan 1997). This effect would require that mergers have previously occurred in dwarf galaxies like Sagittarius, because the velocity dispersion of globular clusters in the bulge is exceedingly high. Consequently, one would not expect the very metal-rich globular clusters to take part in a merger event.

4.6 Liller1

From integrated infrared photometry, M82/ZW84/Z85 obtained E(B-V)=2.9 and $\hbox{$.\!\!^{\rm h}$}=-0.21$; AZ88 derived E(B-V)=2.71 and $\hbox{$.\!\!^{\rm h}$}=+0.20$; Origlia et al. (1997) found $\hbox{$.\!\!^{\rm h}$}=-0.29$; Frogel et al. (1995) employed K, (J-K) CMD and derived $\hbox{$.\!\!^{\rm h}$}=+0.25$. Finally, Ortolani et al. (1996) from an I, (I-z) CMD found E(B-V)=3.05 and $.\!\!^{\rm h}$ comparable to that of the inner bulge stellar population.

The values obtained in Table 3 for reddening and metallicity confirm the previous results. The cluster belongs to the MRBF and is possibly among the most metal-rich ones. Deeper studies are necessary for a definitive diagnosis. The cluster is very concentrated and possibly is in a post-core collapse phase (Trager et al. 1993, hereafter TDK93).

4.7 NGC6380

The integrated infrared photometry (M82/ZW84/Z85) provided E(B-V)=1.4 and $\hbox{$.\!\!^{\rm h}$}=-1.0$. On the other hand, Ortolani et al. (1998) using V, (V-I) CMDs found GB and HB morphologies of a metal-rich globular cluster, deriving $\hbox{$.\!\!^{\rm h}$}\approx-0.5$, and a lower reddening value E(B-V)=1.07.

The present results (Table 3) are consistent with the latter study, placing NGC6380 in the MRBF. TDK95 pointed out NGC6380 as a possible post-core collapse cluster, however, Ortolani et al. (1998) did not find any evidence of this effect using ESO NTT images with 1$^{\prime\prime}$ seeing.

4.8 Terzan1

The integrated infrared photometry provided E(B-V)=1.5 and $\hbox{$.\!\!^{\rm h}$}=+0.24$ (M82/ZW84/Z85) while the integrated near-infrared spectroscopy by AZ88 indicated a similar reddening E(B-V)=1.49 and a lower metallicity $\hbox{$.\!\!^{\rm h}$}=-0.71$, albeit still placing the cluster in the metal-rich family. Ortolani et al. (1993a) using ($V,\ I,\ z$) CMDs found evidence of a high metallicity for this cluster, from the strong curvature effects on the GB, and also derived E(B-V)=1.67. Bica et al. (1993) pointed out the occurrence of two relatively bright disk blue main sequence stars close to the cluster core which might explain a dilution of the CaII triplet, and consequently, the lower metallicity value derived by AZ88. From integrated infrared spectroscopy, Origlia et al. (1997) derived a high metallicity, $\hbox{$.\!\!^{\rm h}$}=+0.03$.

The present integrated spectrum shown in Fig. 1 samples a large area, $10\mbox{$^{\prime\prime}$}\times90\mbox{$^{\prime\prime}$}$ (Table 1) and should represent as much as possible the cluster global properties. As explained in Sect. 2, the pixels corresponding to a bright foreground (much less reddened than the cluster) late-type giant have been excluded. The derived reddening is E(B-V)=1.80 (Table 3), and the resulting reddening-corrected spectrum (Fig. 5) clearly shows absorption features characteristic of a very metal-rich cluster. The derived metallicity $\mbox{$[Z/Z_{\odot}]$}=-0.18$ places Terzan1 nearly at the solar value. We cannot rule out some contamination either by blue disk stars which would dilute the CaII triplet, or bulge/disk red stars which would enhance the metallicity, with a reddening comparable to that of the cluster. We conclude that Terzan1 is a member of the MRBF, probably one of the most metal-rich clusters with post-core collapse (TKD95).

4.9 Tonantzintla2

The ($V,\ I$) CMDs in Bica et al. (1996) showed GB and HB morphologies of a metal-rich globular cluster, and they derived E(B-V)=1.26 and $\hbox{$.\!\!^{\rm h}$}=-0.60$.

The spectroscopic reddening and metallicity (Table 3) confirm the latter results, placing Tonantzintla2 as a new member of the MRBF. The cluster is not in a post-core collapse phase (TDK95).

4.10 Palomar6

The integrated infrared photometry (M82/ZW84/Z85) provided E(B-V)=1.4 and $\hbox{$.\!\!^{\rm h}$}=-0.74$; Minniti et al. (1995) derived E(B-V)=1.63 and $\hbox{$.\!\!^{\rm h}$}=+0.20$; Minniti et al. (1995) from K, (J-K) CMD favoured E(B-V)=1.25 and $\hbox{$.\!\!^{\rm h}$}=+0.2$; finally, Ortolani et al. (1995a) using ($V,\ I$) CMDs found GB and HB morphologies of a metal-rich globular cluster, deriving E(B-V)=1.33 and $\hbox{$.\!\!^{\rm h}$}=-0.4$. Although there is some dispersion in the metallicity values, all studies agree that Palomar6 is a MRBF.

The present metallicity value $\mbox{$[Z/Z_{\odot}]$}=-0.09$ (Table 3) is consistent with the average of the previous results, and favours the lower reddening values. Palomar6 is not in a post-core collapse phase (TDK95).

4.11 Djorgovski1

Ortolani et al. (1995a) using ($V,\ I$) CMDs found GB and HB morphologies of a metal-rich globular cluster, deriving E(B-V)=1.71 and $\hbox{$.\!\!^{\rm h}$}=-0.4$.

The present determinations (Table 3) agree well. This cluster belongs to the MRBF. The cluster structure is somewhat loose with a concentration parameter c=1.50, and there is no evidence of post-core collapse (TDK93, see also the image in Ortolani et al. 1995a).

4.12 Terzan5

The integrated infrared photometry (M82/ZW84/Z85) provided E(B-V)=2.1 and $\hbox{$.\!\!^{\rm h}$}=+0.24$; AZ88 derived E(B-V)=1.65 and $\hbox{$.\!\!^{\rm h}$}=-0.28$; Liu et al. (1994) employed a K$^{\prime}$, (J-K$^{\prime}$)CMD, and derived E(B-V)=1.6 and $\hbox{$.\!\!^{\rm h}$}=-0.3$; Origlia et al. (1997) derived $\hbox{$.\!\!^{\rm h}$}=-0.64$; and Ortolani et al. (1996) from ($V,\ I$) CMDs found GB and HB morphologies of a very metal-rich globular cluster, obtaining $\hbox{$.\!\!^{\rm h}$}=0.0$ and E(B-V)=2.49.

The different studies present a considerable range of reddening values, and the present result (E(B-V)=2.60) is consistent with the higher estimates. A range in metallicity is also present in the literature, but all agree in the sense that the cluster is metal-rich and the present value ($\mbox{$[Z/Z_{\odot}]$}=+0.07$) is consistent with the higher estimates. The cluster is MRBF and is not in a post-core collapse phase (TDK95).

4.13 Terzan6

The near-infrared spectroscopy indicated E(B-V)=2.93 and $\hbox{$.\!\!^{\rm h}$}=-0.61$ (AZ88); Barbuy et al. (1997) using ($V,\ I,\ z$) CMDs found GB and HB morphologies of a metal-rich globular cluster, deriving E(B-V)=2.24 and a metallicity intermediate between those of 47Tuc and NGC6528/6553.

The reddening in Table 3 is consistent with that derived from ($V,\ I$) CMDs; the reddening in AZ88 is exceedingly high. The metallicity found in all studies places the cluster in the MRBF. The cluster is in a post-core collapse phase (TKD95).

4.14 UKS1

The integrated infrared photometry (M82/ZW84/Z85) provided E(B-V)=3.1 and $\hbox{$.\!\!^{\rm h}$}=-1.18$;Liu et al. (1994) estimated E(B-V)=2.8 and $\hbox{$.\!\!^{\rm h}$}=-1.2$ from infrared CMDs; Minniti et al. (1995) showed a K, (J-K) CMD detecting the GB; Ortolani et al. (1997c) using I, (I-z) CMDs also detected the GB consistent with a very reddened globular cluster with E(B-V)=3.1.

The present value (Table 3) confirms the high reddening. However, the near-infrared reddening-corrected spectrum (Fig. 4) and absorption features (Table 2) are those of a metal-rich globular cluster. If metal-rich bulge stars contribute to the integrated light of UKS1, they must be nearly at the cluster distance, since there is no change in E(B-V) as deduced from the CMDs. Deeper studies are necessary for a definitive diagnosis of this cluster. The cluster is rather concentrated and may be in a post-core collapse phase (TDK93, see also the CCD image in Ortolani et al. 1997c).

4.15 Terzan9

The integrated infrared photometry (M82/ZW84/Z85) provided E(B-V)=1.7 and $\hbox{$.\!\!^{\rm h}$}=-0.38$; AZ88 derived lower values for reddening and metallicity (E(B-V)=1.25 and $\hbox{$.\!\!^{\rm h}$}=-0.99$); Liu et al. (1994) obtained E(B-V)=1.8 and $\hbox{$.\!\!^{\rm h}$}=-1.0$.

The present reddening (E(B-V)=1.60) is consistent with the highest values found, and the metallicity ($\mbox{$[Z/Z_{\odot}]$}=-1.01$) with the lower estimates. The cluster probably does not belong to the MRBF, and deep CMDs, including the HB morphology, are necessary for a more conclusive result. The cluster is in a post-core collapse phase (TKD95).

4.16 ESO456-SC38

By means of V, (V-I) CMDs, Ortolani et al. (1997c) found evidence of a metal-rich GB. They derived $\hbox{$.\!\!^{\rm h}$}=-0.5$ and E(B-V)=0.89.

The reddening and metallicity in Table 3 are consistent with the ($V,\ I$) CMD results, placing the cluster in the MRBF. The cluster structure is rather loose with a concentration parameter c=1.50, and there is no evidence of post-core collapse (TDK93). See also the image in Ortolani et al. (1997c), where some central bright sources appear to be stars.

4.17 Terzan10

Liu et al. (1994) estimated E(B-V)=2.6 and $\hbox{$.\!\!^{\rm h}$}=-0.7$ from an infrared CMD. Ortolani et al. (1997c) using V, (V-I) CMDs detected the GB (and possibly some evidence of a blue HB) and derived E(B-V)=2.40.

The integrated spectrum indicates an intermediate metallicity ($\mbox{$[Z/Z_{\odot}]$}=-1.35$ and E(B-V)=1.90, Table 3). The reddening value is somewhat lower than those from the CMD studies, which might suggest some contamination by foreground stars (bulge or disk). Taking together the evidence of a blue HB and the present metallicity value, Terzan10 probably belongs to the IMBF, unless contamination by field stars affects the estimate. However, differently of the other IMBF clusters discussed in the present paper, the CCD image of Terzan10 in Ortolani et al. (1997c) shows no evidence of post-core collapse, where the bright sources superimposed appear to be bright stars. Deeper studies are necessary for a definitive diagnosis of this cluster.

4.18 Terzan12

Ortolani et al. (1998) using V, (V-I) CMDs found GB and HB morphologies of a metal-rich globular cluster, deriving E(B-V)=2.06 and $\hbox{$.\!\!^{\rm h}$}\approx-0.5$.

The present values (Table 3) basically confirm these results. The cluster is a new member of the MRBF and is not in a post-core collapse phase (TDK95, therein referred to as Terzan11). The image in Ortolani et al. (1998) supports the latter result.

4.19 Palomar8

A visible integrated spectrum indicated $\hbox{$.\!\!^{\rm h}$}=-0.48$, and from the (B-V) integrated colour, E(B-V)=0.35 (ZW84). Armandroff (1988) using ($V,\ I$) CMDs found a metal-rich morphology compatible with ZW84's metallicity value, and derived E(B-V)=0.30.

The reddening in Table 3 is consistent with the ($V,\ I$) CMDs, and the metallicity of all studies place the cluster in the MRBF. The cluster is not in a post-core collapse phase (TDK95).

4.20 NGC6717

An early photographic CMD by Goranskii (1979) showed a blue HB at the detection limit and he derived E(B-V)=0.17 and [M/H]=-1.3. By means of visible integrated spectroscopy, ZW84 found $\hbox{$.\!\!^{\rm h}$}=-1.32$ and E(B-V)=0.22. Recently a deep V, (B-V) CMD of NGC6717 was presented by Brocato et al. (1996), where they adopted E(B-V)=0.22 (from ZW84) and derived $\hbox{$.\!\!^{\rm h}$}=-1.26$. Their CMD also shows a blue extended HB morphology for this post-core collapse cluster (TDK95), similarly to NGC6522 (Barbuy et al. 1994). In the present sample, NGC6717 is the only cluster in common with the large sample of globular clusters by Rutledge et al. (1997) who employed CaII triplet of individual giants as metallicity indicator. They derived $\hbox{$.\!\!^{\rm h}$}=-1.33$ in ZW84's metallicity scale and $\hbox{$.\!\!^{\rm h}$}=-1.09$ in Carretta & Gratton's (1997) scale.

In addition to the total extraction (Table 1) we analyse a core extraction corresponding to 10$^{\prime\prime}$ along the slit. The observed spectra are shown in the upper panel of Fig. 9. The core spectrum has a strong TiO band typical of bulge late-type stars, whereas in the total spectrum this feature is diluted, presumably by the cluster intrinsic stellar population, which is not metal-rich. The equivalent widths (Table 2) and metallicities derived for the total and core spectra (Table 3) are too high compared to the previous values, indicating important contamination. Contrary to HP1, the contamination is stronger in the core. The reddening derived for the total spectrum (Table 3) is similar to that of ZW84.

A simple decontamination experiment can be carried out assuming that the contaminating bulge stars affect both spectra in different amounts. Our approach is to test different proportions for the core spectrum and subtract them from the total spectrum in order to cancel out the TiO band. The match was obtained for a proportion of 60% (flux fraction at $\approx7500\mbox{\,\AA}$), and the resulting decontaminated spectrum is shown in the lower panel of Fig. 9. By measuring Ws of the CaII triplet (Table 2) we derive $\mbox{$[Z/Z_{\odot}]$}=-1.05$ (Table 3), which is now compatible with previous published values. NGC6717 can be conclusively placed in the IMBF.

  

Figure 9: Same as Fig. 8 for the core collapse case of NGC6717. We also show the decontaminated spectrum according to Sect. 4.20

  

Figure 10: Metallicity calibration for the sum of CaII triplet equivalent widths