Requirements on Electron Beam Quality

We have already addressed the importance of a high peak current for reducing the gain length. In practice this means that the bunches must be longitudinally compressed. An additional bonus of the bunch compression is the possibility to generate extremely short FEL pulses with durations in the 10-femtosecond regime.

The second requirement is a small beam cross section. A measure for the beam diameter is the emittance which is, loosely speaking, the product of radial beam size and divergence. A low emittance means that it is possible to maintain a small beam diameter over a very long distance; this is exceedingly important in the undulators of an X-ray FEL which may be more than 100 m long. During acceleration the emittance shrinks inversely proportionally to particle energy, because the longitudinal momentum component of a particle is increased in the cavities, whereas the transverse component remains invariant. There are, however, perturbing effects (such as “wake fields”) that tend to enlarge the emittance. The amplification process in the high-gain FEL depends critically on a good transverse overlap between the electron and the photon beam. Ideally both beams should have the same cross section to ensure that the energy transfer is optimized all along the undulator axis.

The third requirement is a very low energy spread within the beam. In order to achieve laser saturation the energy spread must be less than half the FEL parameter. The achievement of an extremely low energy spread in the order of 0.01 percent is a serious technical challenge for FELs operating at sub-nanometer wavelengths.

The three requirements on the drive beam of an X-ray FEL – high peak current, very low emittance and very small energy spread – are so demanding that none of them can be fulfilled by an electron storage ring. Only linear accelerators (linacs) are suited to produce the drive beam. In particular, the extremely short electron bunches that are needed to generate femtosecond X-ray pulses can only be produced in linacs equipped with magnetic bunch compressors. It would be impossible to recirculate these ultrashort electron bunches through a storage ring without disrupting the charge distribution.