4 Spectral analysis

The goal of this study is to establish an HR diagram for nearby A0 stars free of any peculiar object. The comparison between the observed spectra and those computed with the above derived parameters should allow to select the "normal" stars.

A grid of synthetic spectra, covering the range of the previously determined $T_{\mathrm{eff}}$ and logg values and based on Kurucz stellar model atmosphere (Kurucz 1993) has been constructed; solar abundances and a microturbulence value equal to 2 $\mathrm {km\, s^{-1}}$ are adopted. Computations have been made with various rotational broadening values for each synthetic spectrum.

From the grid of the Kurucz fluxes (1993) we computed also the UV fluxes in the four TD1 bands; the comparison between the observed and computed colour indices $m_{\mathrm{UV}-V}$has been systematically done. The choice of these data is due to the fact that the Thompson et al. (1978) Catalogue contains all but 2 stars of our sample, while only 14 of them have been observed by IUE in the low resolution mode and have spectra available from the Final Archive.

In order to compare the optically observed spectra with the computed ones, the needed shift in wavelength to be applied to the observed spectrum is obtained by cross correlation. The templates used are the spectra computed with the appropriate stellar parameters. The cross correlation program differs from the Midas command XCORRELATE in the sense that the correlation index is normalized to 1. The template and the spectrum are rebinned in the velocity space, the rebinning being largely oversampled, the shift is applied and the spectrum rebinned back in wavelength at its original stepsize.

As by-product of this program, the correlation curve contains informations on the quality of the fit between observed and computed spectra.

The broadening parameter (Table 2, Col. 14) refers to the $v\,\sin\,i$ value adopted for the best fit of the observed and computed spectrum and, in the case of spectroscopic binaries, does not have any physical meaning.

The selection of the normal stars follows from the application of tests based on:

$\bullet$ The correlation curve
The height and the shape of the correlation curve are related to similarity of the synthetic spectrum and the observation; in particular, a non symmetric correlation curve indicates a possible binarity.

$\bullet$ Coherence between MD and Geneva parameters
The comparison between the MD and Geneva parameters for stars with E(b-y)=0.00 shows that the largest difference in $T_{\mathrm{eff}}$ is 352 K (HD 125473), significantly larger than the uncertainties expected for the $T_{\mathrm{eff}}$ determination from errors on photometric colour indices. We note that for this star the MD parameters are those from which the best fit with the computed spectrum is obtained.

We remark that for several stars the logg computed by MD from Strömgren photometry is higher than what expected for dwarf A0 stars. We looked at the eleven stars with logg higher than 4.3 (HD 3003, HD 16152, HD 38206, HD 60629, HD 71043, HD 80950, HD 101615, HD 106797, HD 109573, HD 188228, HD 193571). None of them has an H$_\gamma$ profile which fits the spectrum computed with the MD parameters. For all these stars $E(b-y)\leq 0.01$, so that the Geneva parameters are computed as well. The logg computed from the Geneva photometry is systematically lower and the spectrum computed with Geneva parameters from undereddened colours, fits better the H$_\gamma$ profile.

$\bullet$ Observed and computed profiles
The core of the observed H$_\gamma$ is expected to be deeper than that computed for the appropriate $v\,\sin\,i$ value with Kurucz models which do not include NLTE effects; therefore an observed profile shallower than the computed one is a sign of abnormality. In highly rotating stars Mg II 4481 is the only metallic feature with a measurable profile; this feature is expected to be rotationally broadened, as soon as the 0.02 Å separation of the Mg II doublet becomes irrelevant compared to the stellar $v\,\sin\,i$ value.