A Finite Element, Adaptive Mesh Particle-In-Cell Code

72498-Existing computer codes for modeling charged particles in three-dimensional geometries use finite-difference-analysis tech¿niques and fixed, orthogonal, structured elements. The computational size of the problems of interest is extremely large, thereby severely limiting the types of problems that can be solved and the information that can be gained. A new approach that maintains accuracy while reducing computational cost is needed. This project will develop a finite element, particle-in-cell code with adaptive meshing, which could reduce the number of elements by two to three orders of magnitude while maintaining or improving the simulation accuracy. In Phase I, quasi-static field solvers for both electrostatics and magnetostatics were implemented using the finite element methods with adaptive meshing. The quasi-static solvers permitted time varying boundary conditions, but the electric and magnetic fields remained uncoupled. A particle pusher in Particle-In-Cell (PIC) mode, using the leapfrog integrator with proper nodal charge weighting, also was implemented. Special particle boundaries such as absorbing, reflection and periodic surfaces also were included in the implementation. Phase II will expand and improve the Phase I algorithms to make them more robust and stable. Magnetic field effects will be included in the quasi-static PIC analyses, and the effects of nonuniform warm plasma in an unstructured mesh will be studied. Finally, transient beam optics will be investigated and characterized as a test of the algorithms. Commercial Applications and Other Benefits as described by awardee: The new code should dramatically lower the computational cost of modeling charged particles in accelerators, microwave sources, and plasma devices. It would allow development of components and systems that can¿not be designed with existing codes, thereby providing improved performance and reduced cost.