3. Data analysis
On September 10, 1993, the mean V magnitude of the star was 14.49, whereas on July 1996, it was V . The color index remained constant, being 0.28 on September 1993 and 0.29 on July 1996; the was 0.16 on the night of July 13, 1996.
The search for periodic light variations in the V band has been performed with a Discrete Fourier Transform (DFT) algorithm.
To reduce the noise in the DFT power spectrum, we have shifted the data points to a common magnitude level, computed by using the mean magnitude for each night. This could appear somewhat arbitrary, but is suggested by the fact that we are able to determine the mean luminosity of V1101 Aql in both parts of the observing run since we observed the object for more than one orbital cycle (see below) during each night.
The DFT power spectrum of the V data points (Fig. 1a) shows a series of peaks, the most prominent being at 7 cycles day-1 (=0.144193 days, or ), with a slight prevalence on its one-day alias at 6 cycles day-1 (=0.166809 days, or ). All the peaks belong to the same family of one-day aliases and are therefore produced by the sampling of one single real periodicity of the V lightcurve. We therefore applied the CLEAN algorithm (Roberts et al. 1987) to discriminate the modulation responsible for producing the alias series. The results of this analysis, reported in Fig. 1b, suggest as real period the modulation. As a further check, we subtracted this periodicity with appropriate amplitude and phase to the V data set and computed a new DFT (Fig. 2a): its power spectrum peaks at 7 cycles day-1, but the power of the peak is lower than that of Fig. 1a. We followed the same procedure using the modulation (Fig. 2b): in this case, the DFT power spectrum is the same of Fig. 1a but the peaks had their spectral power doubled. Finally, we applied two least squares best-fit methods to the V data points, i.e. the Sterken's (1977) and the Schöneich-Lange's (1981) algorithms, and both methods indicate as best fit the periodicity at 0.144 days.
The V lightcurves folded with the and are shown in Fig. 3a,b, respectively.
All the images taken at La Silla under photometric conditions (seeing 1".5) seem to reveal around V1101 Aql the presence of an asymmetric nebulosity, best seen in V (Fig. 4a) rather than in B (Fig. 4b). This feature was not detected in the frames obtained on July 1997 frames due to poor seeing conditions.
To better disentangle the faint nebulosity from the background, we have designed, through DAOPHOT II, the average PSF using nearby field stars, and then we have subtracted it to the image of V1101 Aql. The result is presented in Fig. 5a, b: an underlying bow-shaped nebulosity is visible in V (Fig. 5a) and in B (Fig. 5b).
The spectrum secured on June 9, 1996, is shown in Fig. 6. We can notice the Balmer lines in absorption with a small emission core and, perhaps, the presence of a noisy emission of He II at 4686 Å. Fluxes and EW's of these lines are reported in Table 2. is outside the spectral range as well as the absorption lines and bands of the secondary, possibly present at longer wavelengths.
Table 2. Fluxes (in units of 10-15 erg cm-2 s-1) and EW's (in Å) of the lines detected in the spectrum of Fig. 6. If an emission core is present, emission and absorption fluxes and EW's are both reported
The NaD interstellar absorption at 5890 Å is also present. This line might be possibly contamined by the presence of He I 5876 in absorption; a double gaussian fit yields Å, which corresponds, according to the relation by Barbon et al. (1990), to a mag. This value is higher than that ( 0.1 mag) found by Pastukhova & Shugarov (1994) using the ()/() color ratio.
© European Southern Observatory (ESO) 1998
Online publication: February 4, 1998