1 Introduction

Hg-Mn stars are nonmagnetic late B type stars that display unusually strong lines of many heavy ions (see e.g. Preston 1971; Heacox 1979; Dworetsky 1980; Dworetsky et al. 1984; Adelman 1987; Wahlgren et al. 1995). E.g. the mercury abundances in this type of stars are between 4000 and 40000 times larger than solar ones (see e.g. Aller & Ross 1967; Sargent 1964). In A and B type stars the electron-impact broadening is the main pressure broadening mechanism (e.g. Dimitrijevic 1989). Considering that the resonant lines of ionized heavy elements (z>30) are located in the ultraviolet spectral region, the abundance analysis of these elements has become possible due to satellite observations by high resolution spectrographs as e.g. International Ultraviolet Explorer (IUE) satellite (R=12000) or Goddard High Resolution Spectrograph (GHSR) installed at Hubble Space Telescope. The number of heavy ion lines observations with the higher photometric quality and spectral resolution is growing up. Consequently, experimental and theoretical spectroscopic data for modelling of these lines are required.

In order to investigate the Hg-Mn star atmospheres as well as other types of hot stars, the Stark broadening parameters for heavy ion lines are needed. The most sophisticated theoretical method for the calculation of a Stark broadened line profile is of course the quantum mechanical close coupling approach. However, due to its complexity and numerical difficulties, only a small number of such calculations exist. In a lot of cases, such as e.g. the transitions between more excited energy levels, the more sophisticated quantum mechanical approach is very difficult or even practically impossible to use and the semiclassical approach (e.g. Sahal-Bréchot 1969a,b) remains the most efficient method for Stark broadening calculations. But for radiators with complex spectra, heavy elements or multiply charged ions, even the semiclassical method is often not applicable in an adequate way due to the lack of reliable data on atomic energy levels, or due to the complexity of the spectrum, problems of level mixing or problems with the adequate identification of atomic energy levels. Simpler approaches are as well of interest when line broadening data for a large number of lines are required (e.g. opacity calculations), and the high precision of every particular result is not so important, but only a good average accuracy is sufficient.

Here we present the electron-impact broadening parameters for six Au I lines and eight Au II transitions as a function of temperature, calculated by using the semiclassical (Sahal-Bréchot 1969a,b) and modified semiempirical approach (Dimitrijevic & Konjevic 1980; Dimitrijevic & Krsljanin 1986; Popovic & Dimitrijevic 1996), respectively.

The Au II lines are observed in Hg-Mn and other CP stars (Wahlgren et al. 1995; Fuhrmann 1988; Adelman 1994). The investigation of gold in Hg-Mn stars shows that the abundances, obtained from the Au II $\lambda$174.0476 nm line, are between 4000 (for $\chi$Lup) and 20000 (for k Cnc) times larger than solar ones (Wahlgren et al. 1995).

Since a sufficiently complete set of reliable atomic data needed for the application of the full semiclassical - perturbation approach (Sahal-Bréchot 1969a,b) in an adequate way does not exist, the modified semiempirical method (Dimitrijevic & Konjevic 1980; Dimitrijevic & Krsljanin 1986; Popovic & Dimitrijevic 1996) has been used for Stark broadening calculation of Au II lines, whereas for Au I lines the more sophisticated semiclasical perturbation approach (Sahal-Bréchot 1969a,b) has been applied.

Figure 1: Thermal Doppler and Stark widths for Au II ( $\lambda =174.048$ nm) line as functions of optical depth for an A type star ( $T_{\rm eff}=10000$ K, $\log g=4$)