5. Abundance analysis
Data for the lines used in this analysis are taken from the tables of Wiese et al. (1966) and correspond closely to the line lists of Sadakane & Okyudo (1989) (hereafter SO) and Takada-Hidai & Takeda (1995) (hereafter THT). The f-values for nitrogen are derived from the Opacity Project data base provided by the CDS Strasbourg (Cunto et al. 1993, Seaton 1992, hereafter OP). For sulphur and iron the transition probabilities are taken from Wiese et al. (1966) and Fuhr et al. (1988), respectively.
Equivalent widths for all detectable lines in the programme stars are given in Table 2 . The nitrogen line is blended with sulphur and the equivalent width given refers to the entire blend. For spectrum synthesis the sulphur abundance for this line is assumed to be the same as that corresponding to the blend. Owing to rotational broadening the blend cannot be resolved into single components - except in the case of the two very sharp-lined stars HR 3383 and HR 4138, where the equivalent width of the components are given in brackets. Otherwise the table gives the equivalent width of the blend.
Unfortunately the spectrum of HR 6548 is heavily distorted by some remaining fringes. The three nitrogen lines , 8703.26 and 8718.84 give only an upper limit of .
In HR 7883 the spectral range above is also heavily distorted by fringes. The line gives only an upper limit of the nitrogen abundance.
The iron line in HR 7773 is hardly detectable, so the iron abundance given is an upper limit.
Table 2. Nitrogen abundances (LTE and NLTE) and non-LTE abundance corrections ( ) derived from individual photospheric lines. Transition probabilities for nitrogen, sulphur and iron are taken from OP, Wiese et al. (1966) and Fuhr et al. (1988) respectively. Equivalent width are given in mÅ
Table 2. continued
As already mentioned by THT the non-LTE abundance corrections ( ) depend markedly on effective temperature; in the parameter range investigated in this work ( 8800 < T ) increases with T . Two other processes should not be neglected. First, a higher gravity generally leads to higher collisional rates and hence reduces the non-LTE effects. Secondly, in the special case of nitrogen the carbon abundance plays an important role in non-LTE calculations (Rentzsch-Holm 1996), because it strongly affects the resonance lines of nitrogen via bound-free continua in the UV. A higher carbon abundance leads to smaller non-LTE abundance corrections.
However, the abundance corrections derived in this work are not representative of truly photospheric N I lines. Owing to the presence of the Paschen lines the nitrogen lines form higher in the atmosphere where departures from LTE are generally larger. A clear trend to larger non-LTE abundance corrections towards the line core of P13 is visible from Table 2 , reflecting the decreasing depth of line formation.
For all program stars the non-LTE abundances from the individual lines reveal a significantly smaller scatter than when LTE is employed.
Where a determination of the equivalent width of the blend is possible, sulphur shows significant overabundances with respect to the Sun.
Non-LTE effects for sulphur are taken into account by interpolating for T and in the tables of THT for the two lines and .
The resulting non-LTE abundance corrections are too small, dex, to remove fully the general overabundance of sulphur. Only in HR 1448 and HR 6070 does agree with the solar abundance within the error limits.
The LTE iron abundance derived from the line is larger in all program stars than the LTE abundance derived in other spectral regions (Lemke 1989). The latter are expected to be more reliable, since they are based on a richer sample of lines.
Non-LTE effects in the infrared are expected to be larger than in the visible owing to the presence of the Paschen lines (see Sect. 5.1 ). For normal A stars, non-LTE abundance corrections of neutral iron are always positive (Rentzsch-Holm 1995). Hence, non-LTE cannot account for the discrepancy between the iron abundances in the two different spectral ranges.
© European Southern Observatory (ESO) 1997