Astron. Astrophys. 336, 587-603 (1998)

6. H line variations in 1996 and 1997

The H line profiles of HD 51066 are relatively stable from year to year as demonstrated in Fig. 13 though small variations occured on a time scale comparable to our data taking rate of one spectrum per night. The H profile always appears as an absorption line with a filled-in line core and very narrow wings quite typical for late-G/K giants (see also Fig. 5b). Due to the existence of significant chromospheric Ca II H&K emission we interpret the H core filling also to be due to chromospheric emission.

Fig. 13. Time series of residual H profiles in 1996 and 1997 (the 1996 data has been offset by +0.15 for better visibility). The thick line in the 1997 graph is part of a non-residual spectrum. Note that in 1997 the one spectrum at HJD 2,450,548.618 (=0:p846) significantly deviates from the rest. Sharp features marked with are residual water lines.

We have subtracted velocity-shifted and broadened standard star spectra, that were obtained with the same instrumental configuration as HD 51066, in order to remove the non-magnetic part of the spectrum. The resulting difference spectra are furtheron called residual spectra. The best match to the photospheric lines was found when using  Gem (K0IIIb). A fit with o Vir (G8IIIa) results in (photospheric) line cores consistently too shallow by 10% with respect to HD 51066, while 37 Com (G9IIICH-2) has lines consistently deeper by 10%, as does 16 Vir (K0.5III). We had no M-K class III-II reference spectrum to compare with. In Fig. 13 are plotted the residuals after the  Gem subtraction for 1996 January (top) and 1997 April (bottom). In general, all residual spectra show, also if using other standard stars, a blue-shifted emission superimposed on a residual absorption feature. The latter originates most likely from systematic mismatches in the H line wings above a normalized intensity of 0.9 of the continuum, i.e. the upper 1/3 of the absorption profile shown in Fig. 5b. We suspect that this is an artifact due to the difference in luminosity class between our standard(s) and HD 51066. The residual core emission, however, is real and appears blueshifted by 207 km s^-1 with respect to the line center. Its presence implies that at least this part is formed in active surface regions with a net outflow velocity of approximately 20 km s^-1 . Single Gaussian fits to the emission profiles reveal a variable and possibly phase-dependent H -core equivalent width whose minima and maxima are likely anti-correlated with the visual light curve (Fig. 14). The average equivalent width of the residual emission is 270 mÅ compared to 1.40-1.70 Å of the non-residual profile.

Fig. 14. a H -core equivalent width versus rotational phase for 1997 April. Note that the internal error of a single measure is of the order of the symbol size. b The combined V-band light curve from three consecutive rotation cycles in 1997 March-April.

While the residual profiles in Fig. 13 appear relatively stable, the one spectrum at JD 2,450,548.618 (=0:p846) is particularly deviant. It shows a broad absorption wing on the blue side but not on the red side. Since this is only seen in one out of 23 spectra further analysis is impossible but we note that a time-dependent phenomenon like an erupting coronal prominence could produce the observed absorption at velocities larger than the projected rotational velocity . Examples of similar H profiles for, e.g., the active K-dwarf AB Dor were reviewed by Collier Cameron (1996).

© European Southern Observatory (ESO) 1998

Online publication: July 20, 1998
helpdesk.link@springer.de