HgMn stars are a subclass of chemically peculiar star occupying the spectral region ~A0/B9-B6 (10500-16000K). Observationally, these stars are characterised by extremely low rotational velocities, weak or non-detectable magnetic fields and photometric variability, and atmospheric deficiencies of light elements (e.g., He, Al, N) coupled with enhancements of the heavy elements (e.g., Hg, Mn, Pt, Sr, Ga). In addition, the heavy elements also have non-terrestrial isotopic abundances (Smith [1997]; Bohlender, Dworetsky & Jomaron [1998]). The currently favoured mechanism for explaining these anomalies is the radiative diffusion hypothesis (Michaud [1970]). This work has been advanced in the form of a parameter-free model (Michaud [1986]).
The quiescent atmospheres of these stars makes them one of the best
natural laboratories for studying the competing processes of gravitational
diffusion and radiative levitation (Vauclair & Vauclair [1982]). In
the absence of disrupting mechanisms such as convection,
rotationally-induced meridional currents, high microturbulence and
magnetic fields, certain elements reach a factor of 10
enhancement
over their cosmic abundance. Because of the strength and sharpness of
normally exotic spectroscopic lines, HgMn stars are also useful for
constraining fundamental atomic data (Lanz [1995]).
The study of UV resonance lines of manganese by Smith & Dworetsky
[1993], hereafter SD93, showed that the Mn abundances were not
only greatly enhanced over the solar value, but were also correlated with
. Their work confirmed and extended previous studies that had drawn
attention to this correlation, in particular Aller [1970],
Heacox [1979] and Adelman [1989]. Their data also
confirmed the existence of a small class of hot, but only mildly-enhanced
Mn stars (Cowley [1980]), with 46 Aql as its exemplar
(Smith [1993]). In the
hottest stars it was also found that there was great difficulty in fitting
the wings of the resonance lines simultaneously with the cores, the latter
being deeper than expected for LTE formation. However, the abundances
were consistent with the detailed calculations of Alecian & Michaud
[1981]. SD93 speculated on the presence of vertical abundance
gradients of Mn within the atmosphere (i.e., stratification), but the
evidence was insufficient to conclude one way or the other.
In this paper, we present an abundance analysis of visible Mn I and Mn II lines using both spectrum synthesis (i.e., line profile fitting methods) and exact curve of growth (equivalent width matching). These analyses, performed under the assumption of Local Thermodynamic Equilibrium (LTE), agree closely with, and confirm in detail, the abundances of Mn found by SD93 from resonance lines measured in IUE spectra. The only differences were for double-lined binary stars where SD93 made no allowance for dilution or blending effects. The use here ( Stellar parameters, etc. ) of model atmospheres with improved blanketing has only a small effect on the results.
Abt [1952] and Booth & Blackwell [1983] demonstrated
that hyperfine structure (hfs) can have severe effects on abundance
calculations of Mn I if it is not properly taken into account. Although
their particular interests were concerned with F and G stars and the solar spectrum, Booth
& Blackwell's estimates for hotter stars indicated that abundance errors
of more than 1 dex could be made by ignoring hfs. In this paper, we
consider hfs models for Mn II lines similar to those discussed for Mn I
by Booth & Blackwell. Adelman [1992] suggested that the
neglect of hfs in his analysis produced spurious microturbulent velocities
deduced from MnII in HgMn stars; we note that he also omitted the
lines 
4206 and 4326 in analyses
of stars with high Mn anomalies.
Although there are no extant laboratory data for MnII in the literature
as of the writing of this paper, we are able to show that some Mn II
lines must have hfs structures 0.06--0.09Å wide, and that it is the
most likely explanation for anomalous strengths of 
4206 and
4326.