On the quantitative modelling of core and tails of multiple scattering by Gaussian mixtures

We present a Gaussian mixture model of the projected multiple scattering angle over a large range of target thicknesses. The mixture model has two components describing the core and the tails of the distribution, respectively. It can be used as a fast algorithm for simulation purposes and with advanced methods of track reconstruction capable of coping with non-Gaussian noise. The exact multiple scattering distributions are obtained in a direct manner by repeated convolution of the single scattering distribution. The results are found to be compatible with the Molière approximation and with the standard formula of Highland, and give additional easy-to-use information about the tails of the distribution. Finally, the relative importance of hadronic elastic scattering as compared to multiple Coulomb scattering is evaluated.