Thesis: Eindhoven University of Technology, ISBN 90-386-1739-9

The General Particle Tracer code
Design, implementation and application

Bas van der Geer, Marieke de Loos

kaft.bmp (10KB)
Download PDF version (6MB)


Charged particle beams are important tools for scientific, industrial and medical applications. The design and understanding of new charged particle accelerators rely on numerical simulations to predict beam behavior. To aid in the design of these machines, we developed the General Particle Tracer (GPT) code. Because particle tracking strikes the best balance between accuracy and simulation speed for a variety of applications, we based GPT on this principle. The GPT code is a general-purpose code and is currently being used in over twenty institutes worldwide for applications varying from standard beam line design to muon colliders to photo-copiers.

With the GPT code we strive to surpass the generally accepted Parmela code by using transparent physics, better algorithms and a modern programming style. GPT solves the 3D equations of motion of sample particles in time-dependent electromagnetic fields from first principles. The self-fields of the beam, known as space-charge, are also calculated. The code contains a high-order tracking algorithm with variable accuracy to minimize simulation time, a large number of modules to represent beam line components and various space-charge models. Furthermore GPT can be adapted to specific needs, an essential feature in a research environment.

Apart from developing GPT, we have used the code for the design of two novel and challenging electron beam experiments. The project at Eindhoven University of Technology aims at producing ultra-short radiation pulses generated by relativistic electron bunches. The first goal is to produce 100 fs, high quality bunches at an energy of 10 MeV. To prevent space-charge explosion at low energies, a novel acceleration system was designed. An electron bunch is pre-accelerated in a 1 GV/m acceleration field followed by a state-of-the-art rf booster. Our GPT simulations show that the design produces bunches with a cutting-edge density in the 6D phase space, paving the way towards the long-term goal of a table-top (X)UV laser.

The second design is the beam energy recovery system of the 'Rijnhuizen' Free Electron Maser (FEM). The FEM is a tunable high power millimeter-wave source in the range 130 to 260 GHz. To increase the overall efficiency from a few percent to about 50%, a recovery system consisting of an electrostatic decelerator and a multi-stage collector has been designed. Although the concept of such a system is not new, it has never been demonstrated with 1 MW output power and a circulating power of 24 MW. The critical design issue is to reduce beam loss and return current to only a fraction of a percent to reach the target efficiency and to avoid machine damage. The proposed design narrowly meets the targets according to our GPT simulations.