2 Physical parameters of the cool component

In the previous work (Jurdana & Kotnik-Karuza 1994) we followed the behaviour of the excitation temperature of neutral metals in the cool giant's photosphere, which approximately equals its effective temperature and determines the physical state of the photospheric layers (Komarov et al. 1974). The observed monotonic decrease of the $T_{\rm exc}$ for each element cannot be due to the red giant's intrinsic variability since the total temperature change in the observed period exceeded 1000 K. This value is remarkably higher compared to 420 K of intrinsic variability for the cool component of CH Cyg (Taranova 1990). This parameter for the same elements was determined for the Sr 30 Her over a ten year interval as well as for the invariable $\beta$ And (Kotnik-Karuza & Jurdana-Sepic 1997). The revised and completed results of the mean values of $T_{\rm exc}$ (Fig. 1a), show remarkably higher oscillations in CH Cyg than in 30 Her and negligible small oscillations of $T_{\rm exc}$ in $\beta$ And.

  

Figure 1: Physical parameters of CH Cyg compared with the corresponding values of the late type giant's of the same spectral type 30 Her and $\beta$ And at different epochs: a) mean values of effective temperature for neutral metals as a function of time, b) mean values of microturbulent velocity for neutral metals as a function of time, c) microturbulent velocity of a Fe I as a function of the excitation potential of the lower state

  

Table 1: Results for excitation temperature and microturbulent velocity with basic features of empirical curves of growth for 30 Her and $\beta$ And

The data for 30 Her and $\beta$ And are presented in Table 1. The corresponding data for CH Cyg are taken from the previous work (Jurdana & Kotnik-Karuza 1994).

The microturbulent velocities have been well defined in the optical region during quiescence since most lines that could be measured with reasonable accuracy in these highly crowded absorption spectra were more or less saturated. In such cases the classical methods make an accurate determination of microturbulent velocities possible under the assumption that the oscillator strengths are sufficiently well known. The m icroturbulent velocities of CH Cyg, taken as mean values over different multiple groups for each of the four neutral metals, show erratic variations of about 6 km s^-1 throughout the investigated time interval (Fig. 1b). These data compared with the results for the variable 30 Her and nonvariable $\beta$ And (Table  1) confirm the general opinion that there is no clear distinction regarding this parameter between variable and non-variable stars, i.e. it is not likely that intrinsic variability would have direct effect upon stellar turbulence (Tsuji 1986). From Fig. 1b it is evident that the turbulent velocities vary from atom to atom for all investigated stars. This confirms the belief that such stratification most likely exists in the largely extended atmospheres of giant stars (Farragiana & Hack 1971).

In the four spectra taken at different times of the quiescent period we measured the dependence of the microturbulent velocity $\xi$ on the excitation potential of the lower state $\chi$ for Fe I which rendered itself most reliable because of the greatest number of measurable lines (no fewer than 40). A surprising evidence of anomalous behaviour of the microturbulent velocity was found in CH Cyg up to February 1988: this parameter was increasing with the multiplet number. In July 1988 the dependence $\xi=f(\chi)$ common for the red giant stars has been recovered (Fig. 1c). Namely, a decrease of microturbulent velocities with increasing excitation potential of the lower state is expected because of their dependence on the height in the photosphere. This could be a consequence of their possible relationship to the granular convection in the red giant's photosphere (Tsuji 1991) and of their origin in the gradient of the velocity field along the line of sight.

It is evident from our observations that the time behaviour of the physical parameters contradicts the picture of a quiescent red giant's photosphere.