No abstract

No abstract

The machine learning models are constructed to predict the fragment production cross sections in projectile fragmentation (PF) reactions using the Bayesian neural network (BNN) techniques. The massive learning for the BNN models is based on the 6393 fragments from 53 measured projectile fragmentation reactions. A direct BNN model and a physical guiding BNN by FRACS parametrization (BNN + FRACS) model have been constructed to predict the fragment cross section in projectile fragmentation reactions. It is verified that the BNN and BNN + FRACS models can well reproduce the wide range of fragment production in PF reactions with incident energy from 40 MeV/u to 1 GeV/u, reaction systems of projectile nucleus from $^{40}$Ar to $^{208}$Pb and various target nucleus. The high precision of the BNN and BNN + FRACS models makes them applicable in the low production rate of extreme rare isotopes in the future PF reactions with large asymmetry of projectile nucleus in the main new generation of radioactive nuclear beam factories.

A numerical study of the formation of a radiation-induced defect cluster center in neutron-irradiated silicon has been performed by solving a set of semilinear parabolic reaction-diffusion-coupled equations. It is found that most of the primary displacement defects [interstitial (I) and vacancy (V)] will be annihilated by I-V direct recombination in an extremely short time. In particular, the formation of four-vacancy defects is independent of the concentration of sinks and impurities in a sample, and of the energy of the recoil particle relatively. The threshold energy for the formation of a vacancy cluster has also been studied. The results are discussed with the experimental observations.

No abstract