Detailed neutron transport requires significantly more data than i.e. photon transport. One reason is the larger energy range of emission energies of the binding energy of nucleons around 10 MeV down to thermal energies of few meV. Further, the cross sections depend not only on the atomic but also on the mass number of the interacting nuclide. Finally, nuclear reactions with more than two emitted particles can only be described through multiple differential cross sections corresponding to energy and direction. Thus, efficient simulations require a uniform representation of the interaction data, combining detailed information with little memory. Basis are evaluated nuclear interaction data in ENDF-format, i.e. ENDF/B-VII.0, JEFF-3.1.
For the preparation groups the interaction data in an optimised energy grid. Differential cross sections are processed as linearly interpolated density functions. Multidimensional density functions from double differential cross sections are split up in outbound density and conditional density functions. Random samples are drawn via a self-developed algorithm called stochastic trapezium algorithm.
The simulation of elastic and inelastic scattering is carried out in particular detail due to its frequency of occurrence. The primary angular distributions given in centre of mass representation are linearly interpolated within every energy group with respect to the primary energy, and finally converted into secondary energy density. Primary and secondary energy combined with Q-value of the reaction enable the calculation of the scattering angle in the laborator system. Relativistic kinematics are generally applied. When treating thermal neutrons, the motion of the nuclei has to be taken into account. Generally, this is done by using the free gas approximation, but, if available, the so-called S(α,β)-approximation can be used instead. The regards collective binding effects of fluids and solids.
As calculational example, the intercomparison result for the spectral fluency of a 3He-counter set up in an irradiation chamber with Am-Be-source is depicted. In the sketch, red is polyethylene, blue the concrete floor, brown the source and green the helium gas. The spectral fluency is displayed in lethargy representation (log(E)). To be able to better compare the different results, the graphs are shifted against each other. Generally, a good agreement can be observed. The differences in the meV-region can be traced back to the discretised representation of the S(α,β) for polyethylene which are applied by PHITS and MCNP.