Astron. Astrophys. 329, 131-136 (1998)

5. Discussion

As shown above, the three program stars, of spectral type A7-A9 Ve, have almost identical UV-optical energy distributions. The observed IR excesses are also very similar, and can be fitted with an optically-thin dust model, made up of both a cool and a hot component (Malfait et al. 1997). Sylvester et al. (1996) pointed out that high-resolution IR spectra indicate some differences between the dust features that are observed. The differences between these three stars become clearest, however, when the photometric behaviour and the circumstellar spectral lines are studied. It is the opinion of the authors that these differences are not intrinsic, but are mainly due to the circumstellar matter being confined to disks with different orientations with respect to the observer.

The fact that HD 142666 shows irregular brightness variations, and reddens when becoming fainter, can be explained by assuming that the brightness decrease is caused by dense dust clouds revolving around the star. As this cloud intersects the line of sight, obscuration takes place, while the scattering and absorption of the starlight causes the stellar reddening. This process, called variable circumstellar absorption is the most popular explanation for Herbig Ae/Be stars with 'nonperiodic Algol-like minima', for which UX Ori is the prototype (Grinin et al. 1994).

A very basic calculation leads us to the conclusion that a brightness change of 1.2 magnitude can be caused by an opaque (optical depth ) dust cloud acting as a natural coronograph, obscuring the stellar surface. These clouds have typical sizes of several stellar radii (Grinin & Tambovtseva 1995). From the initial reddening it can be deduced that the size a of the dust particles is only slightly larger than that of the ISM, and for a a value of 0.1 micron can be asumed. For silicate grains the effectivity factor is about 0.5 for visual wavelengths (Grinin V.P., private communication). An expression for the optical depth of a cloud is:

with its characteristic size and the mean number density of grains in the cloud. The size of the cloud is estimated to be 10 , or about . From (1) and a mean number density is calculated and finally the mass of the dust in the cloud is estimated as (Grinin V.P.,private communication).

The nature and orbit of these clouds, however, is still a matter of debate (Herbst et al. 1994). In an UV and optical study of UXors by Eaton & Herbst (1995) a color reversal (first reddening when fading, then becoming bluer) is shown for several Herbig Ae/Be and TTS stars. In Fig. 2 of Eaton & Herbst (1995) it can be seen that this reversal only takes place when the star is already obscured by 1.6 magnitude, an amount of obscuration which is never reached in HD 142666. So, the fact that HD 142666 shows reddening while fading, but not a color reversal, as is observed in UX Ori (Grinin et al. 1994) is because the stellar light is not dimmed enough.

The amount of circumstellar reddening that HD 142666 shows at maximum brightness indicates that HD 142666 has a substantial amount of CS material in the line of sight, while the two fairly constant stars HD 139614 and HD 144432 do not show evidence for material in their line of sight, since they do not show appreciable circumstellar reddening.

Regarding the stars HD 139614 and HD 144432, it is probable that these dense clouds (if present) do not appear in their line of sight, as these stars do not show any substantial brightness changes. Absence of such clouds in the line of sight and a negligible amount of CS extinction, together with the observation of an IR excess caused by CS dust, is a clear indication of a disk structure around the star oriented pole-on with respect to the observer.

The hypothesis that the photometric behaviour of HD 142666 is caused by a disk orientated edge-on, and that the non-variability of the stars HD 139614 and HD 144432 suggests a more pole-on orientated disk is consistent with the observations of CS spectral features described in this paper. The optical CS lines HeI 5876 Å in both stars, and NaI D in HD 144432, show emission profiles, while they are seen in absorption in HD 142666. IRSPEC data of the HeI 1.08 m line further confirm this interpretation, HD 142666 being the only star showing pure absorption, while the other two stars show the same (CS) features in emission.

An alternative explanation for the difference in photometric behaviour of the program stars might be a difference in dust distribution: consider a spherical symmetric envelope in which the CS material around HD 142666 is in a clumpy distribution but homogeneous around the two other stars. This model could indeed account for the photometric behaviour of the stars, but not for the observed difference in CS extinction at maximum brightness and the CS spectral lines. The three stars show a very similar IR excess, so the total amount of dust must be roughly equal. However, the CS extinction for HD 142666 at maximum brightness, when there are no clumps in the line of sight, is still much larger than for the other two stars. This crucial difference cannot be explained by a spherically symmetric envelope.

If the small value observed for the of HD 139614 is considered, it is not unlikely that this object is seen pole-on. The other two stars, HD 142666 and HD 144432, have almost the same value, implying that, if this interpretation is to hold, HD 142666 must be rotating more slowly than HD 144432.

© European Southern Observatory (ESO) 1998

Online publication: November 24, 1997
helpdesk.link@springer.de