Astron. Astrophys. 347, 473-477 (1999)

3. Discussion and conclusions

The peculiarities that we have described in the previous section show that HD 154791 has not a "standard M spectrum", as it was reported in literature.

All data indicate that we are dealing with an evolved M giant but the comparison of the spectra taken in different epochs suggests that a number of features used for detailed spectral and luminosity classification are variable and are thus useless for this purpose. Lines emission at an energy level higher than in usual Me stars is evident in our spectra. We cannot say if these emissions were absent at the epochs of the other measurements or if they were firstly noticed by us because of our instrumental set-up and S/N ratio.

In any case, the spectral variability, the unusual ratio between Ca II and Ca I lines, as well as the clear presence of the Balmer lines, that are usually barely visible in the M stars, can be the effect of a strong coronal activity. On the other hand, the emission line profile that we found in HD 154791 is different from that of a typical Me star and strongly reminds that of many Be stars (see, e.g., Hanuschik, 1996). The HeI emissions, detected by us in optical range and by Brown et al. (1990) in the near IR, are also difficult to explain in a coronal activity framework. At the same time, this object cannot be classified as a symbiotic star due to the lack of strong and narrow emission features and of forbidden lines.

These facts, as well as the high kT of the X-ray emission during the outbursts, strongly support the presence of a compact companion; on the other hand, the lack of detectable dynamical effects on the optical star suggests that its orbit must be quite wide.

We can thus suggest the following scenario: the SN explosion of the 4U1700+24 precursor happened in a wide, intermediate mass, binary system and left HD 154791 substantially unaffected. This star then followed its natural evolutionary track, without strong interaction with its compact companion, and is now in the red giant phase. But the dense, slow wind typical of this phase is now generating a nebula around the optical star. This nebula can explain the anomalous reddening found by us and the faint IR excess discussed by Schaefer (1986). The neutron star is thus now orbiting in a relatively denser medium, and a continuous, low level accretion is now present. The related X-ray emission is probably well below the detectability threshold of the present day instrumentation, but it is high enough to warm the external layers of the M star, generating a pseudo-coronal effect, as we see in the optical spectra of HD 154791. Inhomogeneities, quite common in M stars wind bubbles, and/or orbital eccentricity give origin to a variable energy output from the neutron star and are detected as spectral variability of the optical star. Time to time, a denser wind blob cross the orbit of the neutron star and a transient high level X-ray outburst rise up, without any periodicity and with random intensity and N_H, depending from the blob density.

In this hypothesis, the neutron star is very old and no momentum transfer from the optical star was present until recent times: thus, its rotation can be very slow, the X-ray pulsed fraction quite small and therefore very difficult to detect during the outbursts. However, the neutron star is now loosing potential energy, due to the interaction with the HD 154791 wind: we can thus foreseen that its orbit is lowering and that this process will increase the X-ray luminosity of the system and the neutron star angular velocity. The higher energy input will encrease the mass loss of the primary star and, after some time, the system will probably look similar to GX 1+4 (M6III-M3III, symbiotic, see, e.g. Chakrabarty & Roche, 1997). During this phase, HD 154791 will loose part of its external layers and will then appear smaller and hotter, as a late dwarf G or K. Thus this process could generate a normal low mass X-ray binary in a time that is difficult to evaluate but that must not be too long. Actually, the uniqueness of the HD 154791/4U1700+24 system strongly suggests that this evolutionary phase should not be too long.

Our scenario explains also the lack of spectral activity during the outburst pointed out by Tomasella et al. (1997): in our hypothesis, this activity is not due to an accretion disk but to the X-Rays reprocessing in the external layers of the HD 154791 photosphere. Since we believe that the neutron star is still quite far from its optical companion, such activity will appear some time after the outburst. Unfortunately, no continuous monitoring following the end of the X-ray outbursts are reported to date in literature: if they will be performed in future, they will prove or discard our scenario and, in case of positive detection, will allow the direct measurement of the orbital radius.

© European Southern Observatory (ESO) 1999

Online publication: June 30, 1999
helpdesk.link@springer.de