3 Abundance analysis, and derived abundances of the lighter elements

3.1 Model atmosphere and stellar parameters

The adopted model atmospheres (OSMARCS) were computed with the latest version of the MARCS code, initially developed by Gustafsson et al. (1975) and subsequently updated by Plez et al. (1992), Edvardsson et al. (1993), and Asplund et al. (1997). The current version includes up-to-date continuum and line opacities for atomic and molecular species, treated in opacity sampling with more than 10⁵ sampling points between 910 Å and 20 $\mu$m. Models for CS 31082-001 were computed for a metallicity 1/1000th Solar, with the $\alpha$-elements boosted by 0.4 dex relative to iron.

In our preliminary analysis of the star (Hill et al. 2001), we were using the synthetic spectrum code of Spite (1967 and subsequent improvements in the last thirty years). In the present analysis we have employed a more consistent approach based on the turbospec synthesis code developed by Plez (Plez et al. 1993), which shares routines and input data with OSMARCS. The latest version (Alvarez & Plez 1998) features: Full chemical equilibrium including 92 atomic and over 500 molecular species, Van der Waals collisional broadening by H, He, and H₂ following Anstee & O'Mara (1995), Barklem & O'Mara (1997), and Barklem et al. (1998), and updated continuum opacities, and plane-parallel or spherical geometry. The main differences between the Spite et al. and the Plez codes lie in the continuum opacity, the source function (diffusion is included in the latter), and the collisional broadening calculation.

The effective temperature for the star was computed from multicolor information, using the Alonso et al. (1999) color-temperature transformations. A number of photometric data are available for CS 31082-001: $UBVR_{\rm C}I_{\rm C}$ (subscript C indicating the Cousins system) are from Beers et al. (2002a); the V magnitude and Strömgren photometry are from Twarog et al. (2000); infrared data are available from the DENIS (Fouqué et al. 2000) and 2Mass surveys (Cutri et al. 2000). A summary of these photometric data, and the corresponding derived temperatures, is given in Table 2

   

Table 2: Colors and effective temperature of CS 31082-001.

Index	value	$T_{{\rm eff}}$ (K)	$T_{{\rm eff}}$ (K)
		E(B-V)	E(B-V)
		=0.00	=0.03
V	11.674 $\pm$ 0.009
(B-V)	0.772 $\pm$ 0.015	4822 $\pm$ 120	4903 $\pm$ 120
b-y	0.542 $\pm$ 0.009	4917 $\pm$ 70	4980 $\pm$ 70
$(V-R)_{{\rm C}}$	0.471 $\pm$ 0.015	4842 $\pm$ 150	5027 $\pm$ 150
$(V-I)_{{\rm C}}$	0.957 $\pm$ 0.013	4818 $\pm$ 125	4987 $\pm$ 125
(V-K)	2.232 $\pm$ 0.008	4851 $\pm$ 50	4967 $\pm$ 50

At a Galactic latitude $l = -76\degr$, the observed colors of CS 31082-001 are not expected to be significantly affected by reddening. The Burstein & Heiles (1982) maps suggest negligible reddening; the Schlegel et al. maps (Schlegel et al. 1998) suggest $E(B-V) \approx 0.03$. Table 2 lists effective temperatures derived both for the situation of no reddening, and for an adopted reddening of 0.03 mags. The no-reddening values are in better agreement with the derived excitation temperature from Fe I lines. The final adopted temperature of $T_{{\rm eff}} =$ 4825 K is consistent with that obtained from the excitation equilibrium of the Fe I lines. A gravity of $\log g =$ 1.5 $\pm$ 0.3 dex was assumed in order to satisfy the ionization equilibrium of iron and titanium, and a microturbulence velocity of $\xi=$ 1.8 $\pm$ 0.2 km s^-1 was obtained from the requirement that strong and weak lines of iron yield the same abundance.

In this paper we are mostly concerned with the relative abundances of elements in this star, especially the abundance pattern of the heavy neutron-capture elements. The relative abundances are only very weakly dependent on the adopted stellar parameters; all of the lines of interest respond similarly to small changes in temperature and gravity, hence the pattern of the heavy elements relative to one another is hardly affected. A thorough discussion of errors is provided in Sect. 4.2.

   
3.2 Abundances of light and iron-peak elements

Most of the abundances for the light and the iron-group elements were determined via equivalent width measurement of a selection of unblended lines. Exceptions are Li, C, N, and O, for which synthesis spectra were directly compared to the observed spectrum. The linelist used for all light and iron-peak elements will be published together with the analysis of the complete sample of our Large Program. For the compilation of this linelist, we used the VALD2 compilation of Kupka et al. (1999).

Table 3 lists the mean abundances, dispersion of the single line measurements around the mean ($\sigma$), and the number of lines used to determine the mean abundances of all measured elements from lithium to zinc. Also listed are the abundances relative to iron, [X/Fe], and the total uncertainty on this ratio, $\Delta$[X/Fe], including errors linked both to observations and the choice of stellar parameters (similarly to the $\Delta$[X/Th] reported in Table 6 and explained in Sect. 4.2). Notes on specific elements are:

 

 

Table 3: LTE abundances for lighter elements in CS 31082-001.

El.	Z	$\log \epsilon$	[X/H]	$\sigma$	$N_{{\rm lines}}$	[X/Fe]	$\Delta$[X/Fe]
2 $^{12}{\rm C}/^{13}{\rm C}$	>20
Li I	3	0.85			1		0.11
C	6	5.82	-2.7	0.05		+0.2
N	7	<5.22	<-2.7			<+0.2
O I	8	6.52	-2.31		1	0.59	0.20
Na I	11	3.70	-2.63	0.02	2	0.27	0.13
Mg I	12	5.04	-2.54	0.13	7	0.36	0.16
Al I	13	2.83	-3.64		1	-0.74	0.17
Si I	14	4.89	-2.66		1	0.24	0.10
K I	19	2.87	-2.25	0.08	2	0.65	0.10
Ca I	20	3.87	-2.49	0.11	15	0.41	0.08
Sc II	21	0.28	-2.89	0.07	7	0.01	0.06
Ti I	22	2.37	-2.65	0.09	14	0.25	0.10
Ti II	22	2.43	-2.59	0.14	28	0.31	0.05
Cr I	24	2.43	-3.24	0.11	7	-0.34	0.11
Mn I	25	2.14	-3.25	0.09	6	-0.35	0.10
Fe I	26	4.60	-2.90	0.13	120	0.00
Fe II	26	4.58	-2.92	0.11	18	0.02
Co I	27	2.28	-2.64	0.11	4	0.26	0.12
Ni I	28	3.37	-2.88	0.02	3	0.02	0.14
Zn I	30	1.88	-2.72	0.00	2	0.18	0.09

Lithium
With an equivalent width of more than 15 mÅ, the lithium 6708 Å line is easily detected in this star. The abundance was determined using spectrum synthesis techniques to account for the doublet nature of the line. With an abundance of $\log \epsilon$(Li) = 0.85, CS 31082-001 falls well below the lithium plateau for hot dwarf halo stars, as expected for a red giant. However, not all lithium in this star has been diluted after the first dredge-up, as observed in many metal-poor giants. In fact, in a sample of 19 giants with [Fe/H] $\leq -2.7$ (Depagne et al. 2002), we find that among the eight giants with gravities close to $\log g =$ 1.5, four have detectable lithium. Hence, in this respect, CS 31082-001 is not exceptional.

Carbon and nitrogen
These two elements are detected via the molecular bands of CH and CN. Line lists for ¹²CH, ¹³CH, ¹²C¹⁴N, and ¹³C¹⁴N were included in the synthesis. The CN linelists were prepared in a similar fashion as the TiO linelist of Plez (1998), using data from Cerny et al. (1978), Kotlar et al. (1980), Larsson et al. (1983), Bauschlicher et al. (1988), Ito et al. (1988), Prasad & Bernath (1992), Prasad et al. (1992), and Rehfuss et al. (1992). Programs by Kotlar were used to compute wavenumbers of transitions in the red bands studied by Kotlar et al. (1980). For CH, the LIFBASE program of Luque & Crosley (1999) was used to compute line positions and gf-values. Excitation energies and isotopic shifts (LIFBASE provides only line positions for ¹²CH) were taken from the line list of Jørgensen et al. (1996). By following this procedure, a good fit of CH lines could be obtained, with the exception of a very few lines which we removed from the list.

¹³C isotopic lines could not be detected, hence we only provide a lower limit on the ¹²C/¹³C ratio, based on the non-detection of ¹³CH lines.

Figure 2: Fit of CH lines of the G band in CS 31082-001. Dots: observations, lines: synthetic spectra computed for the abundances indicated in the figure.

Figure 3: The forbidden [O I] 6300 Å line in CS 31082-001. Symbols as in Fig. 2.

The carbon abundance was derived primarily from CH lines in the region 4290-4310 Å (the CH A-X 0-0 bandhead), which is almost free from intervening atomic lines. We derive an abundance $\log \epsilon ({\rm C})= 5.82 \; \pm \;0.05$ (see Fig. 2). With this same abundance, a good fit is obtained in the more crowded regions around 3900 Å and 3150 Å, where the B-X and C-X bandheads occur.

The nitrogen abundance was then derived from CN lines. It was only possible to set an upper limit, since the CN lines are extremely weak. A conservative estimate of the nitrogen abundance is $\log \epsilon (N)= 5.22$, from the 3875-3900 Å B-X 0-0 bandhead. An attempt was also made to use the NH 0-0 bandhead of the A-X system around 3350 Å. Lines were extracted from the Kurucz line database (1993), and the gf-values were scaled with the mean correction derived by comparison of the Kurucz and Meyer & Roth (1991) gf-values for the 2 (0-0) R₁(0) and ^R Q₂₁(0) lines at 3358.0525 Å and 3353.9235 Å. This derived correction, -0.807 in $\log(gf)$, was applied to all of the NH lines. The fit was poor, but if the gf-values are correct, which is a bold assumption, the nitrogen abundance is at most $\log \epsilon (N)= 5.02$. Given the many uncertainties attached to this latter determination, we adopt the more conservative estimate obtained from CN.

Oxygen
The forbidden oxygen line at 630 nm is clearly detected in the three spectra taken in October 2000, where the motion of the Earth relative to the star resulted in a Doppler shift that moved this weak line clear of the neighboring telluric absorption lines (Fig. 3). The measured equivalent width of the [OI] line is 2.7 mÅ; the corresponding abundance is $\log \epsilon$(O) = 6.52, which in turn implies an overabundance of oxygen with respect to iron of [O/Fe] = +0.59 $\pm$ 0.2.

Figure 4: The observed Nb II 3215 Å, Ru I 3799 Å, and Rh I 3434 Å lines in CS 31082-001. Symbols as in Fig. 2.

Figure 5: The observed Pd I 3242 Å, Ag I 3281 Å, and 3383 Å lines in CS 31082-001. Symbols as in Fig. 2.

Thus, CS 31082-001 has an oxygen abundance that is consistent with the mild oxygen enhancement observed for other halo stars, at least when derived from the same forbidden [OI] line (e.g. Barbuy et al. 1988; Sneden et al. 2001; Nissen et al. 2001). The linearly increasing trend of [O/Fe] with decreasing metallicity suggested from measurements of OH lines in the UV of halo turnoff stars (e.g., Israelian et al. 1998; Boesgaard et al. 1998) is not a relevant comparison here, since it is known that there exist systematic differences between the two indicators (UV OH and [OI]), that could arise, for example, from temperature inhomogeneity effects (see Asplund & García Pérez 2001).

All of the $\alpha$-elements in CS 31082-001 are enhanced by 0.35 to 0.6 dex relative to iron ([$\alpha$/Fe] = +0.37 $\pm$ 0.13, where $\alpha$ is the mean of O, Mg, Si, Ca, and Ti), consistent with the observed behavior of other metal-poor halo stars.

Potassium is also observed in CS 31082-001, from the red lines at 7664 Å and 7698 Å; an LTE analysis yields [K/Fe] $_{{\rm LTE}} =$ +0.65. However, Ivanova & Shimanskii (2000) have shown that these transitions suffer from significant NLTE effects. For a star with $T_{{\rm eff}} = 4800$ K, $\log g =$ 1.5, the correction amounts to $\sim$-0.33 dex. Therefore, the true potassium abundance in CS 31082-001 should be around [K/Fe] = +0.3 dex.

Iron group (Cr through Ni)
Among the elements of this group, all are depleted to a similar level as iron, although Co is up by $\sim$0.3 dex, while Cr and Mn are down by $\sim$0.3 dex, a typical behaviour for metal-poor stars (e.g., McWilliam et al. 1995; Ryan et al. 1996). Thus, both in this respect, and from its standard [$\alpha$/Fe] enhancement, CS 31082-001 appears to be a typical very metal-poor halo star, except for its high neutron-capture-element abundances. This behavior is understandable if the light elements are produced in the hydrostatic burning phase of the evolution of massive stars, whereas the r-process elements are produced during the explosive synthesis phase of the SNe.

Zinc
Zinc is present in CS 31082-001 at a level comparable to the iron-group elements, and thus does not share the strong enhancement of neutron-capture elements in this star. This is relevant to the ongoing discussions on the origin of Zn in halo stars (Umeda & Nomoto 2002 and references therein), and the importance of the neutron-capture channel for Zn production. Either Zn does not arise from neutron-capture processes, or these processes are unimportant in r-process conditions.