Microbunching and Beam Break-up in Duv-fel Accelerator

We present the results of electron beam longitudinal modulation studies in the DUV-FEL accelerator. For bunch length determination we used the “zero-phasing” method, based on a measurement of the chirped electron bunch energy spectra. The measurements revealed a spiky structure in the longitudinal phase space [1]. A model based on space charge effect is considered [2] to explain of the obtained phenomena. The analysis of the energy spectra has shown a sensitivity of the structure to the electron beam peak current, energy and longitudinal non-uniformity of the RF gun drive laser. Analytical calculations have demonstrated a qualitative agreement with experimental observations. Several experiments have been made to compare with theory; measured results are reviewed in this paper. The obtained effect is briefly discussed in relation to high brightness accelerators. INTRODUCTION AND MOTIVATION A detailed description of the DUV-FEL accelerator can be found in our earlier publications [3]. The DUV-FEL includes a photocathode RF gun, illuminated by a short pulse Ti:Sa laser. The following two linac sections accelerate the beam up to an energy of 70 MeV. The second linac tank is dephased, producing a time-energy correlation, which a four-magnet chicane, located downstream, converts into longitudinal bunching. Since the bunch is “undercompressed”, the third linac tank is used to remove residual chirp, and, in combination with the fourth tank, accelerates the bunch up to the nominal energy for FEL operations (177 MeV). For a successful FEL performance a peak current of 300 A is needed, which requires an initial bunch length of 1.5 ps RMS for the 300 pC electron bunch to be reduced to 0.4 ps RMS (compression ratio of 3.75). For the electron bunch length characterization we use the “zero-phasing” method [4], in which no further acceleration is done after the bunch compressor, but an energy chirp is imparted to the beam with the last linac section (with calibrated RF amplitude). A 72° spectrometer dipole transforms the correlated energy spread into a spatial distribution on a downstream monitor. The RMS size of the horizontal projection σx, due to a known amount of chirp ERF · kRF, corresponds to the bunch length σz: