Astron. Astrophys. 317, L17-L20 (1997)

4. Luminous Blue Variables

Humphreys and Davidson (1994) have published recently a comprehensive review of the properties of LBVs. The LBVs are stars of large bolometric luminosity, L/L , and large mass loss rate M y ^-1. They are thought to be massive stars, , in a post main sequence phase of evolution. Their spectra show broad emission lines of H I, He I, He II, Fe II, Fe [II], ..., with P Cyg profiles sometimes. At minimum the spectrum resembles that of a late O or B star, characteristic of a temperature in the range 1.6 - 3 x K; at maximum the spectrum resembles an A star with a temperature of about 8000 K.

One of their most distinctive characteristics is variability over a wide range of time scales. On time scales of months to years, variability at a level of is common. On time scales of tens of years, variations of occur, with the time from minimum to maximum of order a few months. For both of these levels of variability, L constant. On times of (?) y, eruptions or violent ejections occur. During such eruptions, , and L may increase. The overall behaviour is reminiscent of terrestrial volcanoes: long periods of quiescence interrupted from time to time by minor activity, with violent ejections at much more widely spaced times. For some LBVs ejected material is directly observed around the star. For others excess infrared emission provides evidence for the presence of circumstellar material.

Until recently there have been many suggestions but no agreement as to what is the physical cause or causes of the LBV eruptions. Nevertheless, because LBVs have luminosities, L, close to their Eddington limits, L (Appenzeller 1989), where L and is some appropriate flux averaged opacity, it has generally been suspected that some physical process occurring in layers close to the photosphere, where L/L , may lead to an instability which is responsible for the subsequent variability observed. Recently, Stothers and Chin (1993, 1994, 1995) have discovered a specific destabilizing mechanism, which seems capable of explaining the variations that occur on time scales of decades. In brief Stothers and Chin found that the outer layers of the envelope of a post main sequence supergiant can become almost detached from the remainder of the star due to the large iron opacity at a temperature of 2 x 10 ⁵ K, and that a dynamical instability will arise in these layers due to the effects of a high radiation force in combination with partial ionization of hydrogen and helium. They predicted that a massive star can undergo this type of dynamical instability at two evolutionary phases: just after the end of central hydrogen burning and again before the end of central helium burning. Only stars with initial masses > 60 enter this first phase of instability. Stars with masses > 30 are expected to pass through the second phase of instability. In each case the instability phase is one of enhanced mass loss. During this phase the stellar model remains essentially in the same place in the theoretical HR diagram. The large colour and temperature variations that are observed arise from the enhanced, optically thick wind. Stothers and Chin also note that such behaviour may actually be cyclical, with the mean interval between these variations decreasing with increasing luminosity of the star.

© European Southern Observatory (ESO) 1997

helpdesk.link@springer.de