2 Sample selection and observations

Observations were carried out in two 6-nights runs (June and September 2000); due to poor weather conditions in the September run, we were not able to complete the acquisition of the spectra in 47 Tuc. Hence the present paper deals only with data for the two other clusters. In the same runs we also acquired data for 25 field stars with good parallaxes from the Hipparcos catalogue and metallicity similar to that of the GCs.

For each cluster, we selected two groups of stars: (i) dwarfs about 0.2 mag brighter than the turn-off (5 stars in NGC 6397 and 9 in NGC 6752); and (ii) stars at the base of the red giant branch (3 stars in NGC 6397 and 9 stars in NGC 6752), hereinafter called subgiants. Positions of the observed stars in the C-M diagrams of the two clusters (from Grundahl et al. 1999) are given in Fig. 1. A posteriori, radial velocities and chemical composition show that all stars are bona fide members of the clusters. Stars with likely strong and weak CN bands were selected from Strömgren photometry: the c₁index seems in fact to be higher for CN-richer stars (see Grundahl et al. 1999). Given this choice, it should be clear that stars are not a random sample of stars in the different magnitude bins, but rather they were selected in order to include quite extreme cases of CN-poor and CN-rich stars. For comparison, we give in Fig. 2 also the location in a similar C-M diagram of the field stars we observed.

The observations were carried out using the dichroic beamsplitter #2. This allowed us to observe simultaneously two spectral ranges with the two arms of the spectrograph. Slightly different setups were used in the two runs: the blue spectra cover the spectral range from about 3500 Å to about 4700 Å, while the red spectra cover the range from 5700 Å to about 8700 Å. This very broad spectral range allows observations of a large number of spectral features. The resolution was set by the slit width: in most cases this was 1 arcsec (yielding a resolution of 43000 at order centers); however, this value was sometimes changed according to the variable seeing conditions, from 0.7 arcsec in the best cases, down to 1.2 arcsec in the worst ones. The slit length was always set at 8 arcsec, allowing accurate sky subtraction (necessary, since observations were carried out in full moon). Most of the objects are very faint ( $V\;\lower.6ex\hbox{$\sim$ }\kern-7.75pt\raise.65ex\hbox{$>$ }\;17$), so that long exposures were required even with an 8 m telescope like Kueyen, and a very efficient spectrograph like UVES. Typical exposure times were $\sim$1 hr for subgiants, and about 4 hr (split in 3 to 4 segments) for TO stars. The S/N (per pixel at 6700 Å; there are about 5 pixels per resolution element) of the resulting spectra is $\sim$100 in the best cases, and down to about 20 in the worst ones. Relevant data are presented in Tables 1-3.

 

Table 1: Parameters for field stars

Star	Mv	B-V	b-y	[Fe/H]	$\log g$	$T_{\rm eff}$	$T_{\rm eff}$	$T_{\rm eff}$
HD/BD						H$_\alpha$	Best	Alonso et al.
10607	4.03	0.56	0.372	-0.95	4.00	5780	5734
29907	6.01	0.63	0.452	-1.47	4.82	5636	5406
31128	5.05	0.49	0.358	-1.44	4.63	6279	5969
108177	4.87	0.43	0.330	-1.70	4.44	5884	6063	6067
116064	4.74	0.45	0.349	-1.87	4.35	5770	5923
120559	5.93	0.66	0.424	-0.90	4.51	5207	5378
121004	5.12	0.62	0.424	-0.75	4.32	5401	5548
126681	5.71	0.60	0.400	-1.13	4.55	5444	5556	5541
132475	3.62	0.56	0.401	-1.61	3.79	5596	5520	5788
134169	3.82	0.565	0.370	-0.81	4.06	6106	5825	5874
134439	6.72	0.77	0.487	-1.31	4.78	5151	5052	4974
134440	7.05	0.85	0.527	-1.28	4.69	4777	4777	4746
140283	3.34	0.49	0.380	-2.46	3.67	5560	5645	5691
145417	6.81	0.82	0.509	-1.23	4.78	5096	4922
159482	4.92	0.58	0.382	-0.83	4.32	5546	5656
166913	4.22	0.45	0.326	-1.59	4.08	5921	6020
181743	4.96	0.45	0.351	-1.81	4.42	5749	5927
188510	5.59	0.60	0.416	-1.44	4.50	5412	5485	5564
189558	3.59	0.575	0.386	-1.15	3.87	5829	5670	5663
193901	5.44	0.55	0.377	-1.03	4.62	5848	5780	5750
194598	4.60	0.49	0.348	-1.15	4.42	6064	5980	6017
204155	4.07	0.57	0.378	-0.70	4.06	5907	5751
205650	5.39	0.51	0.375	-1.12	4.52	5628	5782
-35 0360	6.31	0.76	0.472	-1.08	4.52	4957	5086	4980
+05 3640	6.50	0.74	0.474	-1.15	4.62	4987	5063

 

Table 2: Data for NGC 6397 stars (identifications from Grundahl et al. 1999); Al abundances from resonance lines

Star	S/N	n.	[Fe/H]	rms	EW	[O/Fe]	[O/Fe]	EW	EW	[Na/Fe]	[Mg/Fe]	[Al/Fe]
		lines			7771	LTE	nLTE	8183	8194	nLTE		nLTE
Dwarfs
1543	91	23	-2.02	0.12	14.7	0.32	0.16	..	38.8	0.21	0.06	0.20
1622	82	21	-2.02	0.10	13.3	0.27	0.11	28.5	40.7	0.28	0.10	0.26
1905	92	22	-2.06	0.11	13.4	0.27	0.11	17.1	29.4	0.02	0.12	0.25
201432	97	22	-2.00	0.11	12.8	0.25	0.08	17.0	35.6	0.08	0.06	0.10
202765	59	18	-2.02	0.14	16.1	0.38	0.21	20.7	28.5	0.06	-0.06	0.19
Subgiants
669	91	27	-2.01	0.11	6.0	0.34	0.26	62.9	86.8	0.48	0.23	0.46
793	105	31	-2.04	0.12	<6.0	<0.34	<0.26	39.9	72.1	0.21	0.28	0.31
206810	85	23	-2.10	0.10	8.0	0.48	0.48	49.5	68.3	0.25	0.10	0.28

 

Table 3: Data for NGC 6752 stars. Stars are ordered according to increasing Na abundance. Star identifications are from Grundahl et al. (1999). Note: 1) Al abundances from resonance lines; (2) Al abundances from high excitation 8772.9/73.9 doublet

Star	S/N	n.	[Fe/H]	rms	EW	[O/Fe]	[O/Fe]	EW	EW	[Na/Fe]	[Mg/Fe]	[Al/Fe]
		lines			7771	LTE	nLTE	8183	8194	nLTE		nLTE
Dwarfs
4428	49	22	-1.52	0.14	35.7	0.35	0.24	24.4	57.1	-0.35	0.04	-0.48 (1)
4383	49	23	-1.42	0.19	46.0	0.60	0.49	33.6	61.5	-0.23	0.01	-0.11 (1)
202316	44	21	-1.56	0.20	28.7	0.31	0.20	42.6	69.0	-0.09	-0.06	-0.45 (1)
4341	42	23	-1.57	0.26	21.6	0.30	0.21	63.4	83.7	0.18	0.06	-0.76 (1)
4458	33	17	-1.52	0.27	19.4	0.04	-0.06	71.9	82.6	0.24	-0.09	-0.63 (1)
4661	21	16	-1.24	0.17	..	..	..	75.4	84.3	0.28	0.15	-0.09 (1)
5048	44	25	-1.43	0.12	11.0	-0.29	-0.37	78.2	93.7	0.37	-0.24	0.19 (1)
4907	66	28	-1.41	0.21	12.0	-0.25	-0.34	102.9	103.0	0.61	-0.09	0.14 (1)
200613	25	18	-1.24	0.18	..	..	..	103.5	110.1	0.64	0.06	0.60 (1)
Subgiants
1406	48	44	-1.39	0.16	18.0	0.40	0.34	82.1	105.7	0.02	0.20	0.10 (2)
1665	49	41	-1.34	0.15	..	..	..	86.7	116.6	0.10	0.10	-0.13 (2)
1445	35	33	-1.53	0.13	..	..	..	95.0	..	0.19	0.06	0.33 (2)
1563	44	38	-1.42	0.17	17.4	0.38	0.33	97.0	131.0	0.25	0.11	0.17 (2)
1400	70	41	-1.44	0.14	15.3	0.33	0.28	103.4	..	0.29	0.10	0.34 (2)
1461	25	32	-1.38	0.24	..	..	..	94.9	144.4	0.29	0.10	-0.06 (2)
202063	29	38	-1.49	0.27	..	..	..	120.4	143.8	0.44	0.13	0.33 (2)
1481	55	46	-1.39	0.15	..	..	..	122.5	166.2	0.54	-0.07	0.69 (2)
1460	38	32	-1.37	0.15	<14.0	<0.25	<0.19	137.4	..	0.65	-0.19	0.86 (2)

Figure 2: Colour-magnitude diagram for the field stars; MV is computed using Hipparcos parallaxes

Figure 3: Comparison between EWs measured on the spectra of two dwarfs in NGC 6397 (stars 1543 and 1622), and two dwarfs in NGC 6752 (stars 4907 and 4428). The rms obtained from these comparisons are 3.4 mÅ for the stars in NGC 6397, and 4.8 mÅ for the stars in NGC 6752 (assuming equal error for both stars)