Astron. Astrophys. 338, 1073-1079 (1998)

3. Analysis and results

Table 3 summarizes our Pleiades abundance results. In col. 3 the temperatures derived from the uvby, photometry are given. The observed indices are the homogeneous means given in the catalogue of Hauck & Mermilliod (1980), they are essentially those of Crawford & Barnes (1969).

Table 3. Abundances of the Praesepe stars

The microturbulent velocity, , is obtained as a pure fitting parameter to obtain equal Fe abundances from lines of different equivalent widths. The determination is good for (single-lined), sharp-lined stars, whose many lines are measured easily and accurately: 4 stars in Table 2. Their microturbulences ( = 4.5 to 5 km/s) are larger by 1.5 km/s than those found in Coupry & Burkhart (1992). In this series large microturbulences were also found for the Pleiades normal A stars; no observational reason can be brought forward since for CMi our equivalent width scale (and so the deduced) is in good agreement with those of Coupry & Burkhart (1992) and Steffen (1985). As no determination was possible for the 4 SB2 stars and HD 72942, we a priori choose a of 4.5 km/s. Microturbulence values affect only abundances determined from strong lines, that is, for S with every star (except the hot star, HD 73666, whose all lines are very weak) and Fe when the only measured line is that at 6678 Å.

The abundance results (Table 3) are far more affected by double-lined spectroscopic binarity: the abundance analyses are compulsorily hampered by assumptions about the poorly-known parameters of the system (temperatures, masses . . .). The results quoted for CncA are minimum values: the dilution effect is not taken into account. Our process for the 3 other systems is equivalent to assume both components to be identical and the results correspond to a "mean" star. We note that the same choice is done in all previous spectroscopic and/or abundance studies.

The abundance results of Table 3 are shown in Fig. 2: the determinations concerning SB2 stars are distinguished from others owing to their sui generis uncertainties. The determinations are better established for the 5 single-lined and sharp-lined stars. On one hand, it is the hot star 40 Cnc with few lines in its spectrum and consequently few elements studied. On the other hand, they are HD 73045, HD 73174, HD 73709, and HD 73730.

Fig. 2. Abundances of Li, Al, Si, S, Fe, Ni, and Eu (on the scale of log N(H) = 12.00) as a function of effective temperature for the Praesepe A stars. The filled squares (triangles if SB2) denote Am stars, that is, underabundant Ca. The filled diamond denotes a hot Am star

These four Am stars exhibit abundance patterns remarkably close to each other : the standard deviation of each abundance mean is almost 0.2 dex or less; they form a very homogeneous group. Since no normal A star could have been studied in Praesepe owing to too large projected rotational velocities, the abundance results are to be compared with those in the Sun and/or cooler Praesepe stars. Compared with the Sun:

• Al, Si, and S are marginally overabundant (+0.2, +0.2, and +0.1 dex).

• Fe and Ni are overabundant (+0.4 and +0.6 dex).

• Eu is largely overabundant (+1.0 dex).

If we turn to the accurate results for F Praesepe stars (Friel & Boesgaard 1992), the mean of [Fe/H], equal to log - log , has been found to be +0.04 0.04, i.e., a mean Fe abundance slightly larger than that of the Sun with small intrinsic dispersion. So in their atmospheres, Am Praesepe stars are overabundant in Fe compared with F Praesepe stars (+0.35 dex).

For lithium, we turn to a study in Praesepe F and G star dwarfs by Soderblom et al. (1993) which extends that of Boesgaard & Budge (1988). Fig. 3, a partial reproduction of their Fig. 4a such as found in Soderblom et al. (1995), shows the Li-temperature profile of all our observed stars with theirs, less than 5400 K and/or upper limits excluded. For late-F to early G dwarfs in the range 6350 to 5950 K, the relationship between Li abundance and temperature is not tight and one-to-one, the most deviant stars not being systematically known binaries. The cluster "Li peak", on the cool side of the Li dip, is, so, poorly defined and comparing the Li content in Am stars with it is questionable. On the hot side of the Li dip, for the three normal F dwarfs near 6700 and 6800 K, the mean of log N(Li) is 3.04; these hot F stars have the same Li content as the Am stars, wether SB2 or not. There is, however, one exception: the Am star, HD 73709, undoubtedly is more abundant than all the others (by a factor of 2).

Fig. 3. The Li temperature profile of A stars with F and G stars (Boesgaard & Budge 1988; Soderblom et al. 1995) in Praesepe. The Am stars are plotted with filled squares, triangles if SB2

The hot star, 40 Cnc, has abundances in S, Fe, and Eu very similar to those of the four cooler stars, even if Fe and Eu abundances are something higher. These results seem too few possibly to promote to a better understanding of blue stragglers in open clusters. We note the significant overabundance in Ca (+0.45 dex compared with the Sun). On the other hand, K-line and metallic-line types only differ by two temperature classes, which are highly reliable following Gray & Garrison (1987). This would mean an underabundance in Ca compared with other metallic elements. We have found a weak underabundance Ca/Fe (-0.15 dex). Thus, there is indeed no contradiction.

For the four SB2 stars, HD 73618, HD 73711, HD 73731, HD 73818, and the broad-lined star, HD 72942, only (Li), S, and Fe could be typically studied. The abundances have been found remarkably near those better established for the four single-lined stars. Both components of HD 73618 show similar Am spectra, suggesting identical stars; the abundances of the "mean" star consistently are those of each component. They have been found the same as those of the three single stars with the same temperature (8100- 8000 K).

This paper series, that of Boesgaard and co-workers, and that of Soderblom and co-workers, separately are homogeneous. Through similar discussions to those stated in Paper I and not redone here, especially concerning temperature scales, our various comparisons are found relevant and significant.

We note the very good agreement with results by Hui-Bon-Hoa and co-workers (1997, 1998): for five stars in common, HD 73045, HD 73174, HD 73618, HD 73709, and HD 73730, our Fe mean is equal to 7.92 with = 0.09 and theirs equal to 7.93 with = 0.13. The methods are similar, the temperature scales identical, and the spectral intervals different. Yet, one star in common, HD 72942, was excluded from the comparison: the effective temperatures derived from the uvby, photometry with the same code (Moon 1985; Moon & Dworetsky 1985) differ by about 200 K. This disagreement is explained by the difficult and ambiguous assignment of the star to a group.

© European Southern Observatory (ESO) 1998

Online publication: September 17, 1998
helpdesk.link@springer.de