"How do massive stars form?" is still a major question. Some authors have considered that massive stars cannot form by mass accretion like intermediate or low-mass stars, because the radiation pressure of the very luminous protostars would repel the infalling matter and prevent the accretion. Thus, various scenarios have been explored, like the coalescence of protostars.
The interesting point is that if accretion is strong enough of the order of 10~3 M0 yr-1, the momentum of the infalling matter may overcome the radiation momentum of massive stars and accretion becomes possible. In addition, the accreting matter, especially if it forms a disk, protects itself against rapid grain destruction and ionization. Massive protostars show bipolar outflows indicating that a part of the infalling matter is re-directed in ejections around the polar axes. It is remarkable that the rates of bipolar ejection are related to the stellar luminosities over 6 orders of magnitude. There is an excellent review by H. Zinnecker and H.W. Yorke  on massive star formation.
22.1 The Various Scenarios for Massive Star Formation
Several competing scenarios for the formation of massive stars have been proposed over recent years. Some observations  favor the accretion scenario, as for lower mass stars. However, there are still uncertainties and the reality may be complex, with situations in dense clusters where both accretions and collisions play a role.
This is the case, studied since the 1960s, of the pre-MS evolution at constant mass, characterized by horizontal blueward evolutionary tracks in the HR diagram, from the Hayashi line to the ZAMS as shown in Fig. 20.2. The timescale is the Kelvin-Helmholtz timescale iKH « GM2/(RL), which corresponds to ~ 1% of the MS
A. Maeder, Physics, Formation and Evolution of Rotating Stars, Astronomy and Astrophysics Library, DOI 10.1007/978-3-540-76949-1-22, "
© Springer-Verlag Berlin Heidelberg 2009
lifetime. For example, for a 30 M0 star, ¿kh = 3 x 104 yr. The evidences of disks make this scenario difficult to support.
However, we note that the constant mass scenario is an asymptotic limit of the accretion scenarios; it corresponds to the case of an extremely intense and fast initial accretion, which then declines steeply. Thus, the case of constant mass, although oversimplified, is an interesting limit (which might not be too far from the reality). As such, it remains a valuable comparison basis. Some timescales of constant mass pre-MS evolution are given in Table 20.1.
Protostars are moving within interstellar clouds which later form a cluster. Thus, collisions of intermediate mass protostars have been suggested as a possible formation mechanism for massive stars . Often in literature, the coalescence scenario has been supported with the argument that the accretion scenario is not possible for massive stars, because the high-radiation field of massive stars may reverse the infall. We stress that the coalescence may well be important, however not for the above negative reason. If the coalescence scenarios applies, it is due to positive reasons, because the collision probability in dense clusters is high. A nice feature of the coalescence scenario is that it considers the internal cluster dynamics as part of the game.
Observations: there are several arguments in favor of this scenario:
- Most massive stars do not form in isolation, but in the central regions of rich young clusters.
- There seems to be a relation between the mass of the most massive star in a cluster and the cluster mean density .
- There is a mass segregation in young clusters, with the most massive stars in the cluster center (even in very young not dynamically relaxed clusters). Also, the intermediate mass stars show some mass segregation. (We note that the process of cloud fragmentation, occurring according to the Jeans criterion, would produce the smallest stellar masses in the cluster center, if the gas is isothermal with a higher density in the cloud center.)
- The frequency of binaries, and in particular the occurrence of many short-period SB2 systems (e.g., with periods < 5 d), is often high among OB stars . This feature is in agreement with the collision theory, which predicts a high frequency of tight binaries among massive stars due to tidal captures and star-disk encounters .
Theory: for stellar collisions, as well as for particle collisions, the average timescale icoll between two successive collisions behave like (1/icoll) ~ an v, where a is the cross-section, n the concentration of stars and v the average velocity. The full expression is 
Was this article helpful?