4 The globular cluster NGC 6139

NGC 6139 (GCl B1624-387, ESO 331-SC4) is located at $\alpha_{1950}=16^{\rm h}$ 24$^{\rm m}$17.4$^{\rm s}$,$\delta_{1950}=-38^\circ$ $44'16\hbox{$^{\prime\prime}$}$ ($l=342.37^\circ$, $b=6.94^\circ$). The integrated Q39 photometry of Zinn (1980) provided E(B-V)=0.78 and ${\rm [Fe/H]}=-1.67$, and the integrated DDO photometry of Bica & Pastoriza (1983) gave E(B-V)=0.68 and ${\rm [Fe/H]}=-1.28$. An integrated visible spectrum of NGC 6139 was presented by Bica & Alloin (1986) and the line strengths were consistent with a metallicity ${\rm [Fe/H]}=-1.5$ and the continuum distribution with E(B-V)=0.70. A photographic B, V CMD of NGC 6139 by Samus et al. (1996) provided E(B-V)=0.87 and they estimated a lower metallicity ${\rm [Fe/H]}=-2.0$. Hazen (1991) studied the RR Lyrae in the cluster and derived $V_{\rm HB}=17.8$,and by Oosterhof group classification she argued that the cluster could be very metal-poor. Recently Zinn & Barnes (1998) studied this cluster by means of CCD V, I photometry and derived E(B-V)=0.76, ${\rm [Fe/H]}=-1.71$ and a distance $d_{\odot}\approx 10$ kpc. In the present study we compare our CMD data with those of Zinn & Barnes (1998) and we derive the parameters with somewhat different methods, in a way consistent with our previous studies.

4.1 Colour-magnitude diagrams

We show in Fig. 5a the V, I CMD for a $1000\times 1000$ pixel extraction ($6.5'\times6.5'$) centered on NGC 6139. The cluster sequences are dominant, with a blue HB and a steep RGB. The spread arises from differential reddening and some crowding effects. The blue disk MS is also present. In Fig. 5b we show a more central extraction ($250\times250$ pixels, or $1.6'\times1.6'$) where the cluster sequences are well-defined, because differential reddening and field contamination are minimised. In the latter CMD only uncrowded stars are given. This CMD is suitable for measuring the cluster properties (Sect. 4.2).

Likewise Terzan 9 (Sect. 3) the best template for the giant branch distribution is that of M 30 (${\rm [Fe/H]}=-2.13$), although NGC 6752 (${\rm [Fe/H]}=-1.54$) cannot be ruled out within uncertainties. A compromise is ${\rm [Fe/H]}=-1.80$, in agreement with the low metallicity trends found in recent stellar photometric studies (Hazen 1991; Samus et al. 1996; Zinn & Barnes 1998). Previous integrated studies of this cluster probably included some contamination by metal-rich bulge stars.

  

Figure 5: V vs. (V-I) CMDs for NGC 6139: a) $1000\times 1000$ pixel extraction ($6.5'\times6.5'$); b) $250\times250$ pixel extraction ($1.6'\times1.6'$) with selection for uncrowded stars

4.2 Reddening and distance

The brighter parts of the blue HB of NGC 6139 are located at $V=17.98\pm0.10$.By taking M 30 as reference, we derive $\Delta (V-I)=1.00$, corresponding to $\Delta (B-V)=0.76$. Assuming E(B-V)=0.06 for M 30 we get E(B-V)=0.82 for NGC 6139, which corresponds to A_V=2.62 (R=3.2).

Adopting for NGC 6139 M_V=0.66, suitable for the cluster metallicity, we get the absolute distance modulus $(m-M)_0=14.7\pm0.15$. This corresponds to a distance of $d_{\odot}=8.7\pm0.6$ kpc for NGC 6139.

Assuming instead NGC 6752 as template one would get E(B-V)=0.70 and a distance of $d_{\odot}=10.4\pm0.6$ kpc.

Giving more weight to the M 30 comparison, we adopt $E(B-V)=0.77\pm0.06$ and $d_{\odot}=9.4\pm1$ kpc for NGC 6139. These values are comparable to those derived by Zinn & Barnes (1998).

Assuming a distance of the Sun to the Galactic center of $R_{\odot}=8.0$ kpc as above, we get X=0.89, Y=-2.85 and Z=1.14. The cluster would be located at $R_{\rm GC}=3.2$ kpc from the Galactic center.