Central energy equipartition in multi-mass models of globular clusters

Abstract

In the construction of multi-mass King-Michie models of globular clusters, an approximated central energy equipartition between stars of different masses is usually imposed by scaling the velocity parameter of each mass class inversely with the stellar mass, as if the distribution function were isothermal. In this paper, this 'isothermal approximation' (IA) has been checked and its consequences on the model parameters studied by a comparison with models including central energy equipartition correctly. It is found that, under the IA, the 'temperatures' of a pair of components can differ to a non-negligible amount for low concentration distributions. It is also found that, in general, this approximation leads to a significantly reduced mass segregation in comparison with that given under the exact energy equipartition at the centre. As a representative example, an isotropic 3-component model fitting a given projected surface brightness and line-of-sight velocity dispersion profiles is discussed. In this example, the IA gives a cluster envelope much more concentrated (central dimensionless potential W=3.3) than under the true equipartition (W=0.059), as well as a higher logarithmic mass function slope. As a consequence, the inferred total mass (and then the global mass-to-light ratio) results a factor 1.4 times lower than the correct value and the amount of mass in heavy dark remnants is 3.3 times smaller. Under energy equipartition, the fate of stars having a mass below a certain limit is to escape from the system. This limit is derived as a function of the mass and W of the giants and turn-off stars component.