What causes B stars to become Be stars is not yet understood. The proposed mechanisms include the wind-compressed disk model of Bjorkman & Cassinelli (1993), non-radial pulsations, magnetic activity, and binarity.
Also the evolutionary state of Be stars is under debate. Recent studies suggest that Be star fractions are similar for luminosity classes III to V and that conditions during the formation of B stars determine their becoming Be stars soon thereafter (Zorec & Briot 1997).
Among field B stars approximately 18% are Be stars, while the percentage in young clusters can be significantly higher (Grebel 1997). Systematic studies of Be stars in young clusters, where stars share a common origin, are coeval, and of the same metallicity, help to better constrain properties and origins of the Be phenomenon. Similarly, knowledge of the Be star content may be essential in deriving the correct IMF for massive stars.
Thus, in young clusters with Be stars the luminosity function may seem to contain too massive stars, leading to an artificially top-heavy IMF.
We have conducted Monte Carlo simulations to evaluate this effect using B and Be star properties as parameterized by Zorec & Briot (1997) and models from the Geneva group. We ran 30 simulations (each containing 20 000 stars) for different metallicities (Z=0.004, 0.008, 0.02), ages (13, 20, 32 Myr), and Be star fractions (10, 20, 30, 40, 50, 60%).
We always used a Salpeter IMF with a slope of 2.3. Be stars were simulated at random by accounting for their overluminosity, total fraction, and subfraction per magnitude and spectral subclass. Afterwards, the IMF was determined from the luminosity function derived from the resulting color-magnitude diagram. Results are shown in the tables below.
Be fraction | IMF without / with Be stars | |
---|---|---|
Z=0.008, log(t)=7.1, 30 × 20 000 stars | ||
20% | -2.32±0.04 | -2.16±0.04 |
50% | -2.33±0.03 | -2.08±0.04 |
Z=0.008, log(t)=7.3, 30 × 20 000 stars | ||
10% | -2.32±0.04 | -2.25±0.04 |
20% | -2.33±0.04 | -2.24±0.04 |
30% | -2.33±0.04 | -2.21±0.04 |
40% | -2.32±0.03 | -2.17±0.04 |
50% | -2.33±0.03 | -2.14±0.04 |
60% | -2.33±0.04 | -2.11±0.04 |
Z=0.008, log(t)=7.5, 30 × 20 000 stars | ||
20% | -2.32±0.05 | -2.30±0.04 |
50% | -2.32±0.04 | -2.14±0.04 |
Z=0.020, log(t)=7.3, 30 × 20 000 stars | ||
20% | -2.33±0.04 | -2.15±0.03 |
50% | -2.33±0.04 | -2.04±0.03 |
Be star fractions may vary considerably from cluster to cluster (e.g., Grebel 1997). Our simulations demonstrate also that the IMF is affected only slightly for young clusters with Be star fractions of up to ∼30%.
IMF determinations based on ground-based observations are often affected by relatively large errors (∼±0.2), especially when dealing with fairly sparse clusters. In these cases the observational errors will mask effects from Be stars (e.g., Dieball & Grebel 1998). Corrections become more important for accurate HST analyses.
First version: | 18th | March, | 1998 |
Last update: | 29th | September, | 1998 |