Dohlus, M.

Paper	Title	Page
TUPPP23	Numerical Minimization of Longitudinal Emittance in Linac Structures	124
	S. Lange, M. Clemens, L. O. Fichte Helmut-Schmidt-University, Hamburg M. Dohlus, T. Limberg DESY, Hamburg
	Relativistic electron bunches in linear colliders are characterized by 6D phase spaces. In most linear accelerators, the longitudinal phase space distribution does not interact significantly with the transverse distributions. This assumption allows the use of a 2D design model of the longitudinal phase space. The design of linear colliders is typically based on manipulations in the longitudinal phase space. The two dimensional single bunch tracking code LiTrack (Bane/Emma 2005) allows to simulate bunch-compression up to 3rd order and RF acceleration with wake fields. This code is implemented in Matlab with a graphic user interface front end. In order to improve the ability to simulate a two-stage bunch compression system, which consist of a RF accelerating section, a higher harmonic RF section and a dipole magnet chicane, an extension to the LiTrack code is proposed. An analytical model of this two-stage bunch compression system is defined using the energy and the momentum derivatives up to 3rd order of the system. As a consequence, the energy of the system can now be specified directly, for the simulation criteria the peak current and the symmetry of the charge distributions and be specified via parameters. This extended model allows the definition of bunches with an arbitrary energy, phase space correlation, longitudinal emittance, charge distribution and resulting peak current. A minimal longitudinal emittance is generally considered as a quality factor of the bunch, where the bunch energy, peak current and a symmetric charge distribution are represented as constraints. Under these conditions, a constrained optimization problem is defined to minimize the longitudinal emittance with a predetermined bunch-energy and peak-current with respect to the charge distribution symmetry. For the solution of this problem, LiTrack is extended with a optimization solver based on a SQP formulation to find an optimal bunch corresponding to the newly introduced constraints.
WEMPMP01	Computational Needs for XFELS	164
	M. Dohlus DESY, Hamburg
	X-ray Free Electron Lasers (FEL) make use of the principle of Self-Amplified-Spontaneous-Emission (SASE) where electron bunches interact in an undulator with their own co-propagating radiation. They do not require optical resonators and their frequency is therefore not limited by material properties as the reflectivity of mirrors. The performance of X-ray SASE FELs depends exponentially on the beam quality of the electron bunch. Therefore effects in the beamline before the undulator are as important as particle-field interactions of the FEL-SASE process. Critical components are the low emittance electron source, accelerating sections, the bunch compression system and the undulator. Due to the high peak currents and small beam dimensions space charge (SC) effects have to be considered up to energies in the GeV range. Coherent synchrotron radiation (CSR) drives not only the FEL but is also emitted in dispersive sections as bunch compressors. SC, CSR, and wake fields affect significantly longitudinal beam parameters (peak current, correlated and uncorrelated energy spread) and the transverse emittance. Start-to-end simulations use a sequence of various tracking codes (with or without SC, CSR and wake fields) and FEL programs. Usually the particle or phase space information has to be carefully converted for each transition from one tool to another. Parameter studies need many simulations of the complete system or a part of it and beyond that, calculations with several random seeds are necessary to consider the stochastic nature of SASE-FEL process.
	Slides