4 Discussion

   
4.1 AS 78

Now we can take into account all the observed and derived properties of AS 78 in order to discuss its possible evolutionary state. With the fundamental parameters derived above, the star is clearly located above the main sequence. Direct comparison with the evolutionary tracks of Schaller et al. 1992 yields a mass for the star of M_* $\sim$ 9 $M_{\odot }$ (Fig. 14). The bright emission-line spectrum and the presence of a strong near-IR excess due to hot circumstellar dust make AS 78 similar to B[e] supergiants (Zickgraf et al. 1986). Its position in the HR diagram is close to that of HD 87643, which shows some features of this type of object (e.g., Oudmaijer et al. 1998). However, most B[e] supergiants are much more luminous (see Fig. 14) and more massive. One can argue that these stars might have lost a significant fraction of their masses through the stellar wind. This matter would appear as a nebula surrounding the stars which is seen in HD 87643, but not in AS 78. Therefore, this interpretation is hardly applicable to the latter.

Let us now explore other options for the nature of AS 78. The location outside the main sequence rules out the possibility that it is a pre-main-sequence Herbig Ae/Be star. The calculations by Palla & Stahler (1993) and their analysis of the distribution of young intermediate-mass stars in the HR diagram show that they would still be hidden inside their parental clouds if they were as far from the zero-age main-sequence as AS 78 is now. Even if they are visible, they should be surrounded by a large amount of protostellar dust, which produces a significant far-IR excess (e.g., Berrilli et al. 1992). The SED of AS 78 clearly shows the lack of such an excess.

AS 78 may be a post-AGB object because of its intermediate luminosity. According to Blöcker (1995), the parameters we derived ( $L_{\rm bol}$ and $T_{\rm eff}$) would correspond to a star with an initial mass $M_{\rm ZAMS}$ = 4 $M_{\odot }$ and a core mass $M_{\rm H}$= 0.7 $M_{\odot }$ nearly 50 years after the beginning of its post-AGB evolution. Usually these intermediate-mass stars evolve through the post-AGB phase to the planetary nebula phase very quickly. During this evolutionary phase, they are usually heavily obscured by the circumstellar envelope which was ejected during the previous AGB phase. In fact, all type I (massive) and most of the type II (less massive) PNe are known to be strong emitters in the far infrared having the peak of emission between 25 and 60 $\mu$m (e.g. Pottasch et al. 1988). Thus, this possibility seems to be very unlikely for AS 78.

The presence of a large number of Fe II lines in the spectrum and the shape of the IR excess make AS 78 similar to the group 2 B[e] stars (26 objects) selected by Allen & Swings (1976). The nature and evolutionary state of most of these stars have not yet been determined (e.g. Lamers et al. 1998). However, a few objects from this group (e.g., GG Car, MWC 342, MWC 84, MWC 623) have been suspected to be binaries (see Miroshnichenko 1998 for a recent review). MWC 84 (CI Cam) and MWC 342 (V1972 Cyg) both show strong Balmer and Fe II emission-line spectra as well as IRAS fluxes rapidly decreasing towards longer wavelengths. MWC 342 displays cyclic variations of its optical brightness, while CI Cam was detected as a strong X-ray transient source in 1998. Nevertheless, the presence of their secondary components can be detected only indirectly. This suggests that these components are rather compact. At the same time, all quoted suspected B[e] binaries display very strong emission-line spectra, which are usually not observed in single non-supergiant stars.

AS 78 also shows such a strong emission-line spectrum. Moreover, as we speculated in Sect. 3.3, the near-IR variations detected in AS 78 might imply that its secondary is surrounded by a dusty disk, which is sometimes occulted by the primary. Location of the dusty disk around the visible component seems to be unrealistic since the latter's gaseous envelope does not show a noticeable flattening (see Sect. 3.3). That fact we have not found any periodicity in the photometric variations of AS 78 might suggest that the binary separation is large and the corresponding orbital period is comparable with the period of our observations. Thus, the binary hypothesis agrees with the information we derive for AS 78 better than any other considered above, but it still requires additional observations to be proven.

Figure 14: Hertzsprung-Russell diagram. The zero-age main-sequence and evolutionary tracks from Schaller et al. (1992) for stars of different initial masses (denoted by the numbers in $M_{\odot }$ near corresponding tracks) are shown by solid lines, the Humphreys-Davidson limit by a dashed line. Positions of LBVs and candidate LBVs collected from literature are shown by filled circles, those of some high-luminosity B[e] supergiants by triangles. The positions of AS 78 and MWC 657 (this work), MWC 342 (Miroshnichenko & Corporon 1999), and HD 87643 (Zorec 1998) are shown by the open circles

   
4.2 MWC 657

Despite the fact that the spectra of MWC 657 and AS 78 are quite similar, it can be seen from the above discussion that their photometric and spectroscopic properties are different. MWC 657 seems to be surrounded by a dense circumstellar gaseous disk, while the emission-line spectrum of AS 78 is more likely due to a mostly spherical stellar wind. The photometric variations of MWC 657 show signs of periodicity, while those of AS 78 are rather gradually changing with time. These differences might imply a different nature and/or evolutionary state for the objects. However, these differences may be due to different contributions of the system components to the overall radiation.

The position of MWC 657 as a single star in the HR diagram is close to that of AS 78 (Fig. 14). Therefore, all the above arguments about the B[e] supergiant, pre-main-sequence Herbig Ae/Be star, and post-AGB scenarios that we applied to AS 78 are valid for MWC 657. Moreover, the observed properties of MWC 657 resemble those of MWC 84 and MWC 342, as in the case of AS 78. The spectral line profiles of MWC 342 (which are very similar to those of MWC 657) and CI Cam suggest that they are formed in dense gaseous disks. The shape of the IR excess of MWC 657 (which is almost identical to those of AS 78, MWC 342, and MWC 84) may be understood if it is formed in the circumstellar disk. Due to these similarities, the binary hypothesis seems to be a very attractive explanation of the behaviour of MWC 657.

Let us assume that the photometric variations are only due to the variable contribution from the secondary component + disk. Then the upper limit of the V-band brightness of the visible component is nearly 13.2 mag. This, in combination with the interstellar E_B-V = 1.1 and D = 3.5 kpc, gives us its absolute visual magnitude $M_{v} \ge -2.8$ mag. The presence of the broad He I absorption lines, which most likely belong to the visible component, indicates that it is a moderately rotating B-type star. The width and strength of the He I 5876 and 6678 Å lines in the 1994 OHP spectrum are consistent with a temperature for the star of $15\, 000$ K and vsin $i \sim$ 100 kms^-1 (see Fig. 7), assuming no disk emission contribution to the profiles. The same BC = -1.3 mag as in Sect. 3.2.2 gives $M_{\rm bol} \ge -$4.1 mag, log $L_{\rm bol}$/ $L_{\odot} \le$ 3.5, and a radius for the star of $R_* \le$ 10 $R_{\odot}$. Assuming the total mass of the system is 10 $M_{\odot }$, the orbital period of 86 days gives a separation between the components of $\sim$ 10¹³ cm. This value is an order of magnitude larger than the deduced radius of the brighter component. Thus, the brighter component does not fill its Roche lobe and is the presumed mass gainer.

Since no spectral lines of the fainter (secondary) component are seen, it is most likely that the disk fills the Roche lobe of this component and is the source of mass transfer in the system. Such a scenario has been developed for massive X-ray binaries by Harmanec (1985), one of which may well be MWC 657. If this is the case, the fainter component of MWC 657 is probably a compact helium star, which means that the system is rather evolved. To get more information about both possible components, one needs high-resolution spectroscopy of MWC 657 at the minimum and maximum brightness phases.