4. General considerations on the behaviour of He I 6678
At first sight the He I 6678 line profiles reveal at least three distinct components (Fig. 2): a broad photospheric-like absorption line profile, a weak central shell feature, and weak blue (V) and red (R) emissions on the outer edges. However the profiles were highly variable over the run, chiefly in their blue wing, and more complex than those previously observed in August-September 1989 at OHP and clearly showing mid-term and short-term time-scale variations. This behaviour may be the consequence of a quasi-permanent change in the physical and geometrical parameters of layers above the photosphere as a result of stellar activity at the time of this multi-site campaign.
Differently, the mean profile in August-September 1989 principally displayed a broad symmetrical absorption which could, at first sight, be identified as a rotationally distorted photospheric profile; note that this broad absorption was disturbed in its core by a sharp shell feature due to Fe II 6677 and in outer wings by two very weak, somewhat broad, double emission components.
On profiles obtained from August 30 to September 8, 1993 the weak central shell was not due to Fe II alone, because its radial velocity is not consistent with measurements on other Fe II shell lines observed in the vicinity of H. As will be emphasized below, the He I line was strongly disturbed in 1993 by additional broad and narrow absorption features. Moreover the outer weak V and R emissions also varied in width, mean position and intensity. In the following sections it will be stressed that lpv and quantities such as equivalent width (EW) and radial velocity of the centroid of the global absorption component (RV), as well as V and R weak variable emissions, are affected by a mid-term quasi-oscillation that we could not determine accurately, since it is similar to the duration of the observational run. However this variation is fairly well detected in data from all sites and is estimated to be about 9 days, at least.
Since it was not possible to separate the contributions due to the photosphere distorted by stellar rotation and presumably by nrp and by motions in the outer stellar layers and in the circumstellar envelope respectively in each individual profile, we attempted to obtain a qualitative description of the behaviour of both regions. It will be useful in the following to investigate the origin of the detected variations.
4.1. Comparison between 1993 and 1989 data
First, we attempted to fit the broad absorption of the mean He I 6678 profile obtained in September 1989 with a synthetic spectrum of a rotationally distorted photosphere. We adopted K and log g = 3.75, values derived from the BCD system for this star. Then we used the code of static NLTE models of stellar atmospheres provided by TLUSTY (Hubeny 1988) and calculated the profile of He I 6678 line with the code SYNSPEC (Hubeny et al. 1994), with = 340 km s-1 and km s-1. We observed considerable discrepancy between the observed and computed line profiles of He I 6678 for solar abundance , the observed equivalent width of this line being always larger by about 30%. Smith et al. (1994) showed that the equivalent width of the He I 6678 line in pulsating B stars and Be stars without strong emission is larger than that of typical dwarfs, while there is good agreement among them for the He I 4922 line. According to Fig. 1 in Smith et al. (1994) the equivalent width of He I 6678 in EW Lac in 1989 is typical of Be stars without strong emission. However, we are not able to derive a reliable He I 6678 photospheric line profile which properly takes into account local effects (temperature, gravity) at the stellar surface due to rapid rotation. Consequently we decided to adopt as a reference the mean profile obtained in 1989, which could be well fitted with a Gaussian curve with an rms of 0.003 measured on the difference between the two curves.
To compare 1993 and 1989 data, Table 4 lists the average value and the maximum deviation of several quantities (equivalent width EW, full width at half maximum FWHM, radial velocity of the centroid RV, central intensity of the core). These quantities were measured on the absorption part of the He I 6678 line profile using IRAF task "Splot". The integration limits were not fixed. Note the lower value of FWHM in 1993 as compared to 1989 (325 km s-1 compared with 404 km s-1), as well as the amplitude of its mid-term variation () in 1993. The high EW but low FWHM values in 1993 have incited us to introduce the concept of a pseudo-photosphere, as other authors have already done (see Sect. 7.2). However, the term "pseudo-photosphere" will be used only as a line profile description tool, which in fact is devoid of any preconceived idea as to the line formation region or any line formation model or mechanism. In our description of line profiles, the pseudo-photosphere identifies a spectroscopic characteristic which may actually be produced by a particular stellar atmospheric region, such as an extended and/or moving post photospheric zone. At the line profile description level, it is important to stress that the mixing of short-term and mid-term variations reveals strong activity from August 30 to September 8 1993. As He I 6678 is particularly sensitive to local line formation conditions and to non-LTE effects, small variations in opacity will induce changes in line profiles. We shall see that the mid-term variation of He I affects the whole profile. It was probably due to sporadic expulsion of matter from the star which enlarged the photosphere, giving rise to an effect of lower gravity, and to filling up the inner layers of the envelope as shown in Sect. 4.2.
Table 4. He I 6678 line profile parameters in 1989 and in 1993
We examined the quotient of the mean profiles of He I obtained at the same site (OHP) in 1993 and 1989 (Fig. 2). This quotient demonstrates the additional pseudo-photospheric and circumstellar contributions in the 1993 spectra. It is strongly blue-winged, with double V and R emission components, and shows some similarity to Fe II shell lines 6456 and 6516 present in the region of H.
The Fe II lines are slightly blue-winged and have km s-1. They are also flanked by weak V and R emission components, centered at -132 and +144 km s-1 respectively, and possibly oscillating in intensity, with their V/R ratio always 1; extreme emission wings extend to about km s-1.
4.2. The complex variable pseudo-photosphere in 1993
To display the pseudo-photosphere and the circumstellar contribution, quite clearly present in the 1993 data, we divided each individual profile by the 1989 Gaussian fit. The same procedure was also applied to each nightly mean profile from each site to reveal the behaviour of the circumstellar contribution over the run more clearly. This procedure is valid because the mid-term variations in He I 6678 which are induced by changes in physical and geometrical parameters in the pseudo-photosphere/CS inner layers are about 2.5% of the stellar continuum, whereas the expected distortions produced in line profiles by low degree modes of nrp are only about 0.5% (Townsend 1997a).
The resulting daily quotients underscore a complex composite absorption, highly variable in its blue part and flanked by weak outer V and R emissions as seen in Fig. 3 which displays the 9-day gradual fading of a broad component. This very complex absorption can be understood not as a typical profile produced only by an outward moving extended atmosphere, but as the result of the superposition of different contributions (see Fig. 4). Though the composite profiles cannot be resolved unambiguously into different components originating from well-identified regions above the photosphere, an attempt was made to interpret the complex profiles as due to a broad pseudo-photospheric profile, several narrow absorptions superimposed on the former, and to weak V and R outer emissions. Under the assumption just made about the structure of the composite profiles, the following components could be identified (see Fig. 4):
While the daily behaviour of composite features formed in the pseudo-photosphere can be described simply, this is not possible for the individual profiles which are strongly affected by short-term variations induced by local changes in the stellar disc (nrp and/or inhomogeneities). However, the rapid variability in the strength of blue-shifted absorption near -200 km s-1 correlates rather well with the amplitude of short-term variability in the radial velocity of the line centroid at hjd 2449232 and 2449237.8. Moreover, some individual composite circumstellar profiles at both OHP and KPNO show near -150 km s-1 a sharp, discrete and short-lived emission which disturbs the slow blue motion of the broad absorption (a) on hourly time-scales. It could be produced by shocks due to ejecta colliding with each other and/or with the circumstellar medium, as well as to the interaction of a supersonic stellar wind with the discrete ejecta.
© European Southern Observatory (ESO) 2000
Online publication: October 30, 2000