K. U. Leuven, Instituut voor Sterrenkunde
Because of the small pulsation velocity amplitude of delta Scuti stars (less than 10 km/s for the majority of stars), only relatively little attention has been paid to observe these stars spectroscopically. Moreover, since their pulsation period is very short (typically of the order of 2.5 h), high spectral and high temporal resolution are mutually exclusive for an acceptable signal-to-noise ratio (S/N). In order to compare our new data with the results of previous studies, we considered the d Scuti star 14 Aur (A9IV, pulsation period P ~ 0.0881 d, 2K amplitude between 2 and 7.5 km/s ), which has quite often been observed and for which pulsation modes have been deduced (Fitch & Wisniewski 1979, Smith, 1982).
Thanks to the new spectrograph Elodie, both high spectral and high temporal (exposure time around 500s) resolution spectra have been obtained for the first time. Because of the required accuracy, each star spectrum was measured simultaneously with a thorium spectrum, and a correlation mask was used. This should ensure a precision of 50 m/s. The mask is built from an F0 star, a spectral type which is not the same as the target star and can therefore lead to a noise increase. We assume that this has no appreciable influence on the results. The observations have been obtained on December 20, 1994 during nearly 5 hours. The radial velocity curve of 14 Aur is represented on Fig. 1.
Figure 1. Heliocentric radial velocity [km/s] versus the Heliocentric Julian Day (+2449700.) of the observations obtained with the correlation mode of Elodie.
Figure 2. Correlation profiles in the heliocentric frame obtained for 14 Aur. The number given on the right of each spectrum is the Heliocentric Julian Day (+2449707.).
As can be noticed, the heliocentric radial velocity (HRV) curve is extremely well-defined, with an amplitude between 1.5 and 2.5 km/s. The increase of the average HRV is due to the fact that 14 Aur is a member of a multiple system (at least 4 components) with a close companion of short orbital period (Porb ~ 3.79 d, our data covering less than 7% of it) having a large amplitude (K ~ 23 km/s). The origin of the variable amplitude is still a matter of debate: is it due to a quasi-periodicity or to a multiperiodic behaviour? From the results obtained here with Elodie, we clearly have an excellent opportunity to solve this problem by observing this star during a week for instance. The velocity curve is obtained from a Gaussian fit of the correlation profiles. The latter are represented on Fig. 2
Since each profile is the result of the correlation of about 3,000 lines, the S/N is very high, allowing a good study of the structure. Some profiles clearly show a bump (for instance around HJD 2449707.469), a signature of atmospheric motions. Fitch & Wisniewski (1979) have shown that tidal effects with the close companion cause departures from spherical symmetry, leading to a kind of Zeeman splitting of a non-radial mode. They found an l = 1 mode from a period analysis based on photometric observations in the case of 14 Aur. From this period analysis, Smith (1982) modelled his line profile variations using an l = 2 mode. Thanks to the new moment method (see e.g. Aerts et al., 1992; Mathias et al., 1994), an objective mode identification can be undertaken on our high quality profiles with no phase-smearing.
The aim of this kind of study is double: * First, from a larger data set, try to see whether the motion has a strict periodicity (i.e., all the possibly involved periods have to be determined) and use the result to perform a model identification in these stars. In this way, we can test the origin of the modes (the kappa-mechanism in the He II zone) or/and, in the case of 14 Aur (and possibly in other stars as well) we can determine whether or not tidal effects play a primary role. * Second, from building masks corresponding to the spectral type of the star, define sub-masks related to lines having the same forming region. From the resulting correlation profiles, we would be able to detect a possible Van Hoof effect (Mathias & Gillet, 1993) and this could lead to new insights concerning the atmospheric structure and the dynamics. In this way, we hope to build a realistic model of the moving atmosphere.
In conclusion, we would like to stress the very promising results obtained with Elodie. The only criticism that can be made is that, due to the included reduction process, the temporal resolution is limited to about 8 minutes between two consecutive spectra. We recommend to define an option skiping this reduction process, so that it can be done during daytime. If this had been the case for our observations of 14 Aur, we would have got twice our data set, which is of great importance for the accuracy of mode identification methods and for the detection of a Van Hoof effect.
Aerts, C., De Pauw, M., Waelkens, C., 1992, A&A 266, 294
Fitch W.S., Wisniewsk W.Z., 1979, ApJ 213, 808
Mathias P., Gillet D., 1993, A&A 278, 511
Mathias P., Aerts C., De Pauw M., Gillet D., Waelkens C., 1994, A&A 286,109
Smith M.A., 1982, ApJ 254, 242