Electron Beam Collimation for Slice Diagnostics and Generation of Femtosecond Soft X-Ray Pulses from a Free Electron Laser

We present the experimental results of femtosecond slicing an ultra-relativistic, high brightness electron beam with a collimator [1]. We demonstrate that the collimation process preserves the slice beam quality, in agreement with our theoretical expectations, and that the collimation is compatible with the operation of a linear accelerator. Thus, it turns out to be a more compact and cheaper solution for electron slice diagnostics than commonly used radiofrequency deflecting cavities and having minimal impact on the machine design. The collimated beam can also be used for the generation of stable femtosecond soft x-ray pulses of tunable duration from a free electron laser. ELECTRON SLICING WITH A COLLIMATOR The feasibility and operability of the electron beam collimation at ultra-relativistic energies for slice diagnostics was investigated in the radiofrequency linear accelerator (RF linac) of FERMI@Elettra FEL [2, 3]. The collimator is a horizontal scraper made of two identical, cylindrical and individually movable rods of copper. The rod diameter is 13 mm wide. Collimation for slice diagnostics was applied to an initial 350 pC, 5 ps FWHM long beam. The magnetic chicane and the upstream linac were set in order to define the bunch length compression by a factor 5.5. All the relevant beam and machine parameters adopted in the experiment are listed in Table 1. The geometric beam size in the middle of BC1 was 160    m so, following the prescription in [4], Eq.1, the scraper blades were inserted into the vacuum chamber to define a half-aperture at least three times bigger, namely 0.5 mm wide. The chromatic beam size x = 2.6 mm, that is the product of the energy dispersion function and the fractional energy spread rms, was much larger than the geometric one, so we can use Eq.1 to evaluate the duration of the collimated beam that is 70   col t fs FWHM. The charge of such a beam is expected to be pC t t C Q Q i col i col 27     , where Qi and ti are, respectively, the initial total charge and the initial bunch duration (FWHM) and C is the compression factor. The scraper aperture was consecutively translated to select 12 longitudinal slices of the bunch. Each slice was accelerated and transported to the linac end, in the socalled TLS region. The slice optical parameters were measured both in the BC1 and in the TLS region with the quadrupole scan technique [5, 6], in dedicated diagnostic stations. COLLIMATOR’S GEOMETRIC TRANSVERSE WAKEFIELD The impact of the scraper transverse wakefield on the emittance of the collimated beam was analytically estimated, with the limitation that the model starts failing when the particles travel at a distance comparable to the collimator half-aperture. Following [7], the FERMI scraper behaves like a flat, long collimator and the kick factor  must be computed in the diffractive regime. For a half-aperture 0.5 mm wide, we have  = 72 V/pC/mm. This is a rather large value but its effect on the emittance is mitigated by the low charge traversing the scraper and a proper optics setting at the collimator location. In the case of a 50pC beam charge traveling 0.4 mm far from the scraper axis, the collimator’s kick is [7] , 8 . 4 rad E hQ      where Q is the bunch charge, h is the bunch centroid distance from the collimator axis,  is the kick factor in the plane of interest and E is the beam mean energy. Following [8], the normalized emittance growth can be estimated with: