Bunch Compression

Principle of longitudinal electron bunch compression

Principle of longitudinal electron bunch compression. The bottom row shows an accelerating cavity and the four dipole magnets of the magnetic chicane. The top figures show the bunch shape at various stages and the correlation between the internal position ζ of an electron inside the bunch (and the charge density ρ, respectively) and its relative energy deviation = (W - Wr)/Wr where Wr is the energy of a reference particle at the centre of the bunch: (a) before the cavity, (b) behind the cavity, (c) behind the magnetic chicane. In the RF cavity the particles are accelerated on the falling slope of the RF wave. Thereby the trailing electrons receive a larger energy gain than the leading ones. In the magnetic chicane the electrons at the tail move on a shorter orbit than those at the head and catch up. The ideal linear energy–position correlation (chirp) is indicated by the black curves, the real nonlinear chirp results in the blue curves and the blue-shaded areas.

High peak currents of several 1000 A are needed in extreme-ultraviolet and X-ray free-electron lasers. These cannot be produced directly in the electron gun. Therefore moderately long bunches with a peak­ current of about 50 A are created in the source, quickly accelerated to higher energy and then compressed in length by two orders of magnitude.


The electrons in the linac have speeds very close to the speed of light, and the velocity differences are far too small for a trailing electron to catch up with a leading electron if the particles move on a straight line. This possibility is opened if the particles are passed through a magnetic chicane. Longitudinal bunch compression is achieved in two steps. First an energy slope is imprinted on the bunch by acceleration on the falling slope of the RF wave. Consequently the particles at the head of the bunch receive a smaller energy gain than those at the tail. Afterwards the particles pass through two magnetic chicanes where the trailing electrons of larger energy travel a shorter distance than the leading ones of smaller energy and thus are enabled to catch up.


To realize the energy slope, the RF phase in the first accelerator module is adjusted in such a way that the particles are accelerated on the slope of the RF wave. Due to the cosine shape of the RF wave, adding a nonlinear term to the position–energy relationship inside the bunch, and due to coherent synchrotron radiation effects in the magnetic chicanes, the final bunches do not possess the ideal narrow shape but consist of a leading spike with a width of less than 100 femtoseconds and a tail extending over several picoseconds. The leading spike contains 10 to 20 percent of the total bunch charge and reaches a peak current in excess of 1000 A which is needed in the high-gain FEL process. In the long tail, the local current is too small to expect any significant FEL gain. Work is in progress to supplement the 1.3-GHz cavities with a 3.9-GHz cavity that will linearize the RF wave. With this third-harmonic cavity in operation it should be possible to squeeze almost the entire bunch charge into a narrow pulse.