In this work we analyze the ballistic phase of displacement cascades. Monte Carlo simulations performed with realistic cross-sections were used to calculate the probability density function, pL(l), describing the length traveled by a particle between two successive collisions within the binary collision approximation framework. From the analysis of pL(l), this work shows that the trajectories set displays a bifractal behavior. The comparison of different MC simulations clearly shows that this bifractality is only due to the long tail of pL(l). The unusual form of pL(l) implies that atoms displaced in the collision cascade during irradiation follow a Levy flight. Since the Levy law is scale invariant, an anomalous diffusion of atoms at large distances is then expected under irradiation and may explain experimental results.

No abstract

No abstract

No abstract

The object of this paper is to present how perturbation techniques implemented in TRIPOLI-4 can be used to treat the uncertainties on nuclear data. The perturbation scheme used in these calculations is the correlated sampling method. Cross sections uncertainties are processed in multigroup mode via the modules GROUPR and ERROR of NJOY nuclear data processing system. We use the relative error output of the Error module. Perturbated cross sections are then sampled according to the covariance matrix. The perturbation scheme is used to obtain the perturbated score corresponding to these cross sections. This procedure is repeated to obtain an estimation of the dispersion of the response due to the cross section variations. This dispersion is a measure of the effect on the result of uncertainties on the cross sections. This method is demonstrated by application to the Winfrith Iron Benchmark.