Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 318, 198-203 (1997)

Previous Section Next Section Title Page Table of Contents

6. Discussion

We have discovered dramatic changes in the H [FORMULA] emission line object HD 76534. While the line was strongly in emission on the first day of our observing run, it faded to the point of disappearance two days later, but went strongly into emission within the following two hours, making it the fastest transition of the H [FORMULA] line from absorption into emission observed in any Be-type star to date.

Under the assumption that there are no changes in polarization across the H [FORMULA] line, we have used the polarization spectra as timeseries spectra, enabling us to follow the evolution of the line at short time intervals. In the last series of spectra (Run 3), we find that the line strength increased linearly with time. During these 40 minutes, the line profile, the radial velocity and the FWHM remained constant within the errors.

In the above we presented two mechanisms that could be responsible for the sudden onset of H [FORMULA] emission in HD 76534.

The mass outburst mechanism was suggested by Hanuschik et al. (1993) who studied the Be star µ Cen (B2IV-Ve). This object belongs to a small group of Be stars which have been described as showing sudden changes in the H [FORMULA] line. µ Cen has undergone several outbursts over the last years (Peters, 1986; Hanuschik et al. 1993), with rise and fade times of the emission that are an order of magnitude longer than observed in HD 76534. Hanuschik et al. (1993) proposed that the H [FORMULA] emission originates in a circumstellar disk that is built up within a couple of days, and then falls back onto the star as the radiation pressure is insufficient to make the gas reach velocities larger than the local escape velocity. The main point in favour of this hypothesis for µ Cen is that during the rise of the emission strong V/R variations are present, only after several rotation periods is the disk circularized, yielding a stable V/R ratio. In addition, they show that the peak-to-peak separation of the H [FORMULA] line decreases during the rise and increases during the fading, implying that a rotating disk is respectively growing and shrinking.

Neither effect is seen in the case of HD 76534. The medium resolution observations in Run 3 do not show V/R variability, nor kinematical variations in the line. The peak-to-peak velocities remain constant within the observational errors. Of further note is the very symmetric line with V [FORMULA] R. These results imply that the disk was already stable and circularized at the moment the emission was observed. This is not what one would expect based on simple considerations of time scales involved. The estimated rotation period of HD 76534 ((2 [FORMULA] R [FORMULA])/v  sini [FORMULA] 3.5 d/sini) is long when compared to the observed changes. The material, even if it were co-rotating with the star, can never have achieved a circularization of its orbit within a mere two hours.

This apparent contradiction is lifted by the new hypothesis that we bring forward, namely that of a variable photosphere, where the Lyman continuum changes on short timescales. If the circumstellar material was already present before the increased number of ionizing photons became detectable by the sudden onset of H [FORMULA] emission, there is no need to invoke a build-up of the disk, along with its expected line profile variability. Although the variable photosphere hypothesis seems, at first sight, unlikely, it should be noted that not much is known about the Lyman continua of B stars. As the case of [FORMULA] CMa shows, relatively strong changes in the EUV continuum would go almost undetected in the optical. Given the right circumstances, the ionization of circumstellar material can then be a sensitive tracer of such variability. A likely implication of this hypothesis is that the variability would occur with a short (semi-)regular period. Our observations do not cover a large enough time span to establish any periodicity however.

The only previous spectra of HD 76534 were published by Thé et al. (1985) and Praderie et al. (1991). Their spectra were taken one year apart, and showed 'indistinguishable' (cf. Praderie et al. 1991) line profiles. It is difficult to derive much from this fact. One may conclude that the line was stable in emission for a period of more than one year, implying that the observed changes in this paper are a rare event. Alternatively, it is useful to notice that the exposure times used by the above authors were 2 and 2.5 hours respectively, so that any shorter term variations in either the line profile or EW would have been washed out. An indirect argument against the idea that HD 76534 had a stable emission line profile for a long time, is the fact that we observed variability at all. The spectra presented in this paper are the first to have been taken with integration times which are shorter than the timescale of the variability detected, and, unless we were so fortunate to observe a rare moment in the long term spectral behaviour of HD 76534, it is rather more likely that these variations occur very frequently.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1997

Online publication: July 8, 1998