The chemical analyses of the program stars were carried out relative to
the comparison stars 
(G8IIIab) and Procyon (F5IV-V).
was chosen as a primary comparison standard, because the
chemical composition of this star has been analyzed in detail relative to
the Sun (see, for example, Cayrel & Cayrel 1963; Tomkin
& Lambert 1986). The atmospheric composition of 
corresponds to the solar one within the probable error of determination and
no systematic enrichment in heavy-element abundance can be detected. We
adopted the following atmospheric parameters (
, 
, 
), and model metal abundance [A] for 
: 5130 K,
3.0, 
, 0.0. These parameters agree in general with those obtained by
Tomkin & Lambert (1986) and McWilliam (1990).
For Procyon, we have adopted 
, 
, 
, which is the best value according to the compilation of
Steffen (1985).
The final adopted atmospheric parameters (effective temperature,
surface gravity, microturbulence velocity) and model metal abundances
for the program stars are listed in Table 2 (click here). Some of the stars have
previously published parameter determinations. We compared our results
with those obtained by Edvardsson et al. (1993) for
, and by McWilliam (1990) for 6 Per, 5 Tau,
6 Boo, 
, HR 6388, HR 6868, and HR 7180. We obtained an
excellent agreement between our results and results obtained by
Edvardsson et al. (1993) for 
, 6068 K, 
, 
. The comparison between our results and those of McWilliam
(1990) shows a good agreement for temperature (
for 16 Per, 5 Tau, 
, HR 6388, HR 6860, HR 7180), for
gravity (
for 5 Tau, 6 Boo,
), and for microturbulent velocity (
for 5 Tau, 6 Boo, HR 6388, HR 6860, HR 7180). But for some
stars there are disagreements between our results and those obtained by
McWilliam (1990) for gravity (6 Per, HD 168532, HR 7180),
turbulent velocity (6 Per), and effective temperature (6 Boo). Our
gravity values for these three stars are lower than those derived by
McWilliam (1990), who estimated gravity using the relation
between temperature, mass, luminosity, and gravity.
We checked our spectroscopic gravities for dwarfs with those derived using
temperature-mass-luminosity relation and good trigonometric parallaxes (
for 
, 
for HR 3579, 
for 
). There is an excellent agreement for 
, but for HR 3579 and
spectroscopic gravities are somewhat lower than those obtained
from parallaxes. The atmospheric parameters of binaries are found to be typical of
single main-sequence or giant stars, respectively.
The mean differential abundances of the light and iron group elements (Na, Mg,
Si, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu), obtained using neutral or ionized
lines, and the mean abundances of the s-process elements (Y, Zr, Ba, La, Ce, Nd)
for all program stars are given in Table 3 (click here), along with the standard deviations
of averaged abundances estimated from individual lines, and the number of lines
used in the analysis. In Table 4 (click here) the mean abundances of separate elements for
all stars are given. In the literature there are previous abundance
determinations available for some of the stars. We have found a generally good
agreement (Fig. 8 (click here)) for the light and iron-peak elements, as well as for barium
when we compare our results for 
to the results of
Edvardsson et al. (1993). In the following we shall comment on
some details of the abundance distributions.
Table 3: The averaged differential abundances of light and iron group
elements, obtained from neutral (XI) and ionized (XII) lines, and of
s-process elements for the program stars along with the standard deviation,
and the number of lines. The last column indicate the Ba abundance scaled to
the metallicity
Table 4: The mean differential abundances [X] of separate elements for
all program stars, along with the standard deviation of abundances
estimated from individual lines, and the number of lines used in the
analysis
Figure 8: The comparison of abundance determinations of this paper
and others for 
Carbon abundances have been derived for 11 of the program stars using the
neutral carbon line at 
or at 
.
Carbon appears to be enhanced only for HR 6860 and, possibly, slightly for HR
7180. We have found no indication of carbon processing from any other program
star. The oxygen abundance has been derived only for 7 stars using the forbidden
oxygen line at 
. The analysis shows that oxygen may be
slightly (
) overabundant in the atmospheres of six stars.
No significant deviations from the solar abundance pattern can be found
among these elements. The mean abundance for 17 single lined binaries
relative to the standards is 
with typical standard
deviation 
. Only the abundance of Mn shows a larger
dispersion from star to star in the sample: 
.
The seven elements heavier than iron are traditionally identified with
synthesis by the neutron capture s-process. The main sample of
spectroscopic binaries includes seventeen stars (
, 6 Per, 5
Tau, HR 3579, HR 5053, 6 Boo, 
, HR 6388, DR Dra, HD 166478,
HR 6860, HD 170737, FN Aql, HR 7180, HD 179558, HD 181602, HD 204934).
Detailed differential abundance analysis relative to the standard
stars shows that no one from these barium star like binaries have a
significant enhancement (
) of averaged
s-process elements scaled to iron group elements (see Tables 2 (click here) and 3 (click here)).
Taken together the mean abundance of these elements is solar. Only
giant 5 Tau (K0II-III, 
) and supergiant FN Aql (F8Ib,
) show a mild enhancement (
) of
s-process elements, however, overabundance of heavy elements in these
stars are small, different from those of certain BaII stars with
similar orbital periods (Zacs 1994). Although direct
evidence for WD companion has been confirmed only in the case of DR Dra, at
least some of the analyzed systems contain a WD companion, because according
to Jorissen & Boffin (1992) unevolved systems never populate
the right-bottom corner on the (
) plane (Fig. 9 (click here)). Part of the
analyzed single-lined primaries fulfill three constraints: 1)
, (2) 
, and (3) 
, where 
is the maximum eccentricity observed at
period P among barium systems (Jorissen & Boffin 1992).
Constraint (1) ensure that the considered systems followed the same binary
evolution as barium stars. Constraint (2) allows to get rid of systems
with massive main sequence companions, while systems not fulfilling
constraint (3) are likely unevolved and therefore contain a main
sequence rather than a WD companion. Thus the detailed abundance
analyses show that the presence of WD companions in barium stars-like
binaries is not sufficient to produce a strong barium star. This
conclusion confirmed especially the chemical composition of DR Dra
(K0III, 
) with a directly confirmed WD companion
from IUE spectra (Fekel & Simon 1985).
Figure 9: Diagram (
) for single-lined spectroscopic
binaries with barium star like orbital elements. Filled circles
(
): barium stars from Jorissen & Boffin (1992); open
circles (
): giants showing normal barium abundance; crossed
circles (
): giants showing mild barium enhancement;
squares (
): dwarfs
At the same time five of the analyzed giants (5 Tau, DR Dra, HD 166478, HD
168532, HD 176524) show a mild enhancement (
) of
the s-process element barium (see Table 3 (click here)). Although the barium abundance is
based only on three BaII lines, a careful analysis shows that this effect
has not a methodical nature, because:
1) The Ba II lines have a medium intensity (
);
2) The standard deviation of barium abundances obtained from three BaII
lines does not exceed usually 0.1 dex (0.08 for 5 Tau, 0.03 for DR Dra, 0.06
for HR 6860, and 0.09 for HR 7180);
3) Non-LTE effects in BaII lines are small (Mashonkina & Zacs
1996) and our differential line by line analysis relative to the similar
standard canceled possible slight (
) non-LTE deviations;
4) Some of the analyzed BaII star like giants (with similar atmospheric
structure) show a normal barium abundance.
This conclusion support the location of the barium enhanced binaries on the
(e, 
) plane (see Fig. 9 (click here)). As can be seen, barium enhanced giants
occupy a place on the diagram similar to BaII stars. In the right-bottom
corner only dwarfs and one giant HR 5053 (K0III, 
) from
our program stars have a normal barium abundance. However, the unseen
companion of HR 5053 is a G-dwarf (f(M)=0.47, Batten et al. 1989).
We included this star in the observations program to test the possible
influence (due to tidal mixing, as has been sometimes suggested) of a main
sequence companion on the atmospheric structure (chemical composition) of
the primary (giant) star in BaII star like systems. Since the atmosphere of
HR 5053 does not show any peculiarities of chemical composition we concluded
that the main sequence companion does not have a significant influence on
the internal structure of the primary star. On the other hand, on the (e,
) plane outside the region occupied by barium star we do not find
any peculiar (barium enhanced) primary in our sample of barium stars like
binaries. Therefore, we come to the conclusion that a mild barium
enhancement in five giants is, presumably, a result of mass transfer from
the companion during its late phases of evolution. Thus it seems likely that
all red giant spectroscopic binaries (primaries) in barium star like systems
with WD secondaries have chemical peculiarities of barium (s-process
elements). The degree of chemical peculiarities of the primary depends,
apparently, on the efficiency of mass transfer in a specific binary system.
We certainly do not apply this conclusion to the supergiant FN Aql
because:
1) The BaII lines in the spectra of this star are substantially stronger
than for giants (269 mÅ for BaII line at 
);
2) The atmospheric structure of supergiants is significantly different from
the standard giant, therefore, differential analysis might not cancel
possible (greater) non-LTE effects.
Our observational program includes also two mild barium stars HD 119185 (Ba
1.0) and HD 130255 (Ba 1.0) from Jorissen's (1994) list, that did
not show radial velocity (RV) variations. We obtained 
for HD 130255 and 
for HD 119185. The significant
enhancement of s-process elements is shown by both stars. However, we would
like to note that due to the higher dispersion of the chemical abundances
obtained from different lines, the error of the derived atmospheric
parameters and abundances for HD 130255 is higher than for other program
stars. The s-process enhancement of RV non-variable barium stars confirms a
previous conclusion (Zacs 1994) that these barium stars,
probably, have either velocity variations below the limit of detection (very long
orbital period), or high inclined orbital planes.
Acknowledgements
We are indebted to referee Prof. E. Böhm-Vitense for her constructive comments on the paper. This research was supported in part by the ESO C&EE Programme (Grant No. B-01-012) and by the Russian Foundation of Fundamental researches (RFFI grant No. 95-02-04276).
