1999A&A...349..151M

The long-term visual spectrophotometric (SPh) behaviour of Be stars as a function of fundamental stellar parameters is studied. Some previous SPh results obtained by other authors are confirmed. Moreover, a tendency for temperature and aspect angle dependency of SPh variations is found. From the characteristics of visual SPh behaviour in Be stars we derive constraints for models of regions in circumstellar envelopes where the visual continuum spectrum is formed: (i) The SPh emission and absorption phases should not imply preferential aspect angles, as they can both appear whatever the stellar inclination. This phenomenon cannot be always accounted for by strongly flattened circumstellar envelopes; (ii) Radii of the visible continuum forming regions cannot be larger than a few R_*; (iii) Electron densities of these regions should not exceed N_e∼10¹³cm^–3; (iv) Electron temperature of circumstellar layers producing the SPh emission phases compare with the stellar Balmer continuum radiation temperature and it is much lower in those producing the SPh absorption phases. Three scenarios were studied to produce the observed characteristics of emissions in the V magnitude and in the second component of Balmer discontinuity (ΔD) during the SPh emission phases: (a) expansion of a massive circumstellar shell that preserves circumstellar envelope flattening; (b) expansion of a circumstellar shell which increases the global flattening, so that a disc-like structure is formed; (c) continuous mass ejection that increases the storage of mass in a constant volume with a given flattening. Mechanisms (a) and (b) produce a double valued (ΔV,ΔD) SPh relation, while (c) produces a single valued relation. Only mechanisms (a) and (c) can easily produce the observed amounts of emission ΔV and ΔD without violating the modeling constraints from (i) to (iv) imposed by observations. The model SPh slopes of (ΔV,ΔD) show the global sin i and T_eff observed dependencies. The scenarios used to describe the double valued (ΔV,ΔD) suggest another possible way how to build up circumstellar envelopes around Be stars.