• G. De Ninno
  ELETTRA, Basovizza, Trieste
• W.M. Fawley
  LBNL, Berkeley, California
• W. Graves
  MIT, Middleton, Massachusetts

The FERMI at ELETTRA project at Sincotrone Trieste involves two FEL's, each based upon the principle of a seeded harmonic cascade and using the existing ELETTRA injection linac at 1.2 GeV beam energy. Scheduled to be completed in 2008, FEL-1 will operate in the 40-100 nm wavelength range and will involve one stage of harmonic up-conversion. The second phase, FEL-2, will begin operation two years later in the 10^-40 nm wavelength range and will involve two cascade stages. FEL design assumes wavelength tunability over the full wavelength range and polarization tunability of the output radiation including helical polarization. The design considers focusing properties and segmentation of realizable undulators and available input seed lasers. We discuss how the interplay between various limitations and self-consistent accelerator simulations [1,2] have led to our current design. We present results of simulations using GENESIS and GINGER simulation codes including studies of various shot-to-shot fluctuations and undulator errors. Findings for the expected output radiation in terms of the power, transverse and longitudinal coherence for the short pulse (50-200 fs) and long pulse (~1 ps) modes of operation are reported.

[1] S. Lidia et al. in these proceedings. [2] S. Di Mitri et al. in these proceedings.

• H. Freund
  SAIC, McLean
• M. Abo-Bakr, K. Goldammer, D. Kraemer, B.C. Kuske, A. Meseck
  BESSY GmbH, Berlin
• S. Biedron
  ANL, Argonne, Illinois

The BESSY FEL is based on a seeded cascade of High Gain Harmonic Generation (HGHG) sections followed by an amplifier to produce coherent and stable short wavelength output. Here, we report on comparative design studies carried out using the MEDUSA [1], and GENESIS [2] simulation codes. These two codes have each been used to successfully predict a variety of FEL designs and have agreed well with a number of important experiments. In addition, they were included in a comparative study of FEL simulation [3] that reported substantial agreement between the codes for the specific configurations studied. However, these codes are based on different assumptions. GENESIS treats the particle dynamics using a wiggler-averaged orbit approximation, the transverse electromagnetic field is treated using a field solver, and harmonics are not included. MEDUSA does not use the wiggler-averaged orbit approximation to treat particle dynamics, the transverse fields are treated using a Gaussian modal superposition, and harmonics are included self-consistently. Hence, the comparative study for an HGHG cascade is important. We report the results where the parameters of each stage have been optimized.

[1] H.P. Freund et al., IEEE JQE 36, 275 (2000). [2] S. Reiche, NIMA 429, 243 (1999). [3] S.G. Biedron et al., NIMA 445, 110 (2000).

• G. Lambert, M. Bougeard, W. Boutu, P. Breger, M.-E. Couprie, D. Garzella, H. Merdji, P. Monchicourt, P. Salieres
  CEA/DSM/DRECAM/SPAM, Gif-sur-Yvette
• B. Carre
  CEA/Saclay, Gif-sur-Yvette
• T. Hara, H. Kitamura, T. Shintake
  RIKEN Spring-8 Harima, Hyogo

In order to reach very short wavelengths in FEL, and to have a more compact, fully coherent and tunable source, a particular seeding configuration is foreseen to be tested as a demonstration experiment in 2006 into the SCSS phase 1 facility (Spring-8 Compact Sase Source, Japan). The external source is the 13th harmonic (61.5 nm) of a Ti: Sa laser (25 mJ, 10 Hz, 100 fs) generated in 10 Hz pulsed Xe gas cell. The harmonic generation process provides us with a intense (1 μJ) and ultra-short (50 fs) VUV beam. The design of the experiment implantation is discussed, taken into account the performances of the generation process, the focusing of the selected harmonic into the modulator, and the resistance of the optical components. Besides one should consider the vacuum needs, the geometrical problems and the mechanics for the under UHV mirrors translation. One first chamber is dedicated to the harmonic generation. A second one is used for spectral selection and adaptation of the harmonic in the modulator. Finally theoretical estimates of the performances relying on 1D simulations using PERSEO code and 3D simulations using GENESIS code are also given.