Comparison between electrostatic PP and PIC simulations on electron bunch expansion

With the great development of parallel computing techniques, the particle-particle (PP) model has been successfully applied in a number of plasma applications. Comparing to particle-mesh (PM) models, for example the widely used particle-in-cell (PIC) method, PP has the advantages of high accuracy in solving Coulomb interactions. In this paper, it is shown that PP is also advantageous to simulate non-neutral plasmas, such as electron bunch expansion in vacuum. The numerical effects of the macro-particle weight and the time step length are investigated for a PP model, accurate and convergent results can be obtained with less effort. On the contrary, PIC needs to simulate the same problem with extremely large effort. It is found that the simulation accuracy does not grow with reduced cell size monotonously, thus no convergence can be easily obtained. In the long run, PIC must apply large enough domain to cover all the expanding particles and avoid non-physical effects caused by imperfect infinite boundary condition, which may result in too heavy computation and make PIC infeasible.