2 The observations

The observations were collected with the 2.15 m telescope at the Complejo Astronómico El Leoncito (CASLEO, San Juan, Argentina) in two runs in June 1994 and May 1995. In the first run we employed the Universidade Federal do Rio Grande do Sul (UFRGS) CCD camera attached to the CASLEO Boller & Chivens Spectrograph. The detector is a GEC6803 chip with $578\times385$ pixels of size $22\mbox{$\,\mu{\rm m}$}\times22\mbox{$\,\mu{\rm m}$}$; one pixel corresponds to 1$.\!\!^{\prime\prime}$32 on the sky. The slit was oriented E-W except for one exposure of Terzan1 (Table 1). We used a grating of 300 grooves mm^-1 producing an average dispersion, in the observed region, of $\approx240\rm\mbox{\,\AA}\,mm^{-1}$ or 5.4Åpixel^-1. The spectral coverage was $\approx6400-9500$Å. A series of exposures (in general 15min each) were employed for the objects (Table 1). The long slit, corresponding to 3$.\mkern-4mu^\prime$3 on the sky, allowed us to sample regions of background sky. The seeing during the nights was typically $\approx2\hbox{$.\!\!^{\prime\prime}$}5$. The slit width was 4$.\!\!^{\prime\prime}$5 resulting in a resolution (FWHM) of $\approx16$Å as deduced from the comparison lamp lines. Dome and sky flat fields were taken and employed in the reductions. The standard stars CD-32$^{\circ}$9927 and LTT7987 (Stone & Baldwin 1983), and HD160233 (Gutiérrez-Moreno et al. 1988), were observed for flux calibrations. HD160233 was also used for the telluric absorption corrections in the near-infrared.

 

Table 1: Logbook of observations

 

• The observations refer to the near-infrared range, except when indicated visible in Col. (1); for Terzan1, 4 exposures were obtained with an E-W slit at different positions in the N-S direction, and 1 at PA $=58\mbox{$^{\circ}$}$.

For the second run we used a CCD camera with a Tektronics chip of $1024\times1024$ pixels (of size $24\mbox{$\,\mu{\rm m}$}\times24\mbox{$\,\mu{\rm m}$}$) attached to a REOSC spectrograph. Similar scanning procedures and gratings as in the first run were adopted. The total field along the slit was 4$.\mkern-4mu^\prime$7. The range observed corresponds to 5800 - 9200Å with a dispersion of 3.36Åpixel^-1. The slit width was 4$.\!\!^{\prime\prime}$2 and the comparison lamp lines provided an average resolution of 17Å. Two objects (UKS2 and ESO93-SC08) have also been observed in the visible range (3500 - 7000Å) with a dispersion of $140\rm\mbox{\,\AA}\,mm^{-1}$ or 3.46Åpixel^-1 leading a resolution of 14Å. The standard stars were LTT4364, EG274 and LTT7379 as well as HD160233.

  

Figure 1: Observed spectra of bulge objects. Spectra are in $F_{\lambda}$ units normalised at $\approx\lambda7500$Å. Constants have been added to the spectra, except for the bottom one

  

Figure 2: Same as Fig. 1 for the remaining clusters of the bulge sample

Reductions were carried out with the IRAF package following standard procedures at the Instituto de Física, UFRGS (Porto Alegre) and the Córdoba Observatory. The spectra were extracted along the slit according to the cluster size and available flux. For larger objects, we scanned the slit along the N-S direction in order to better sample the cluster stellar population. The resulting spatial coverages are given in Table 1. Wavelength calibrations were carried out with a He-Ne-Ar lamp (He-Ar in the second run) with exposures following that of the object or standard star. The spectra were flux calibrated, and telluric absorption bands were corrected using the hot standard stars following the procedures given by BA87.

The cluster sample is listed in Table 1. We emphasize that our observations of Terzan12 refer to the cluster originally designated as such in Terzan (1971) with coordinates $\alpha_{1950}=18^{\rm h}~09^{\rm m}~14.0^{\rm s}$ and $\delta_{1950}=-22\mbox{$^{\circ}$}~45\mbox{$^{\prime}$}~18\mbox{$^{\prime\prime}$}$. We also included in the sample the globular cluster candidate TBJ3 (Terzan & Bernard 1978).

  

Figure 3: Same as Fig. 1 for the clusters projected on the disk

For comparison purposes we included two well-known bulge clusters NGC6624 and NGC6528. The logbook of observations is provided in Table 1, by Cols.: (1) - cluster designations, (2) - date of observation, (3) - number of exposures and duration in seconds, and (4) - slit width (or scanning length) and extraction along the slit (values in arcseconds).

Terzan1 and Terzan4 present bright superimposed stars. The different slit positions were checked for the signature of foreground stars by analysing the spectra extracted from individual peaks. In the case of Terzan1 it was possible to avoid the pixels corresponding to a contaminating foreground late type star which resulted much bluer (thus definitely not as reddened as the cluster) than the remaining cluster extractions, and consequently, it could not be a giant member. In the case of Terzan4 two foreground stars have been identified and have not been included in the final cluster spectrum.

For HP1 and NGC6717 we considered two spatial extractions, a larger one as indicated in Table 1, and a more central extraction for each cluster as discussed respectively in Sects. 4.5 and 4.20.

Figures 1 and 2 show the resulting observed spectra of the bulge sample, and Fig. 3 those projected on the disk. Notice the extremely steep spectra in the upper part of Fig. 1, denoting dramatic dust absorption effects even at these considerably long wavelengths. They correspond to the most reddened globular clusters. The disk objects UKS2 and ESO93-SC08 are discussed in Sect. 5.4.