Section Navigation

Velocity bunching experiment at the Neptune Laboratory


In recent years electron beam users have increased their demands for high brightness beams in short sub-ps pulses. These beams find applications in the advanced accelerator community for injection into short wavelength high gradient accelerators, or as plasma wake-field drivers, and in the light source community for short wavelength SASE Free Electron Laser and for Thompson-scattering generation of short X-ray pulses. Recent designs of systems capable of delivering high brightness very short electron beams include the use of conventional photoinjectors in conjunction with magnetic compressors. While the magnetic compression scheme has been proved successful in increasing the beam current, the impact on the phase space has been shown to be quite dramatic [1].

An alternative scheme that could preserve the phase space quality while still shortening the beam to sub-ps bunch length has been recently proposed as an injector for X-ray Free Electron Laser, by Serafini and Ferrario [2]. This scheme, commonly known as "velocity bunching" is an elaboration of the old idea of RF rectilinear compression to the RF photoinjector system. The idea is based on the weak synchrotron motion that the beam undergoes at moderate energies in the RF wave of a linac accelerating structure. The compression happens in a rectilinear section so that the damage suffered by going through bending trajectories is avoided. This compression scheme has been recently observed and studied experimentally in different laboratories. [3-4]. A main ingredient of the Serafini and Ferrario recipe to produce high brightness sub-ps electron beam is to integrate this compression section in the emittance compensation scheme, by keeping the transverse beam size under control through solenoidal magnetic field in the region where the bunch is compressing and the electron density is increasing.

A small variation inside this framework is the thin lens version of velocity bunching. Here the synchrotron motion of the electrons inside the RF structure is very limited. There is almost no phase advance inside the longitudinal lens and all the bunching happens in the drift following the linac. We studied this configuration, also sometimes termed "ballistic bunching", at the Neptune Laboratory [5].

Neptune ballistic bunching experimental setup

At the Neptune laboratory at UCLA a 4 ps rms long laser pulse hits a single crystal copper cathode inside a 1.6 cell RF gun. The photoelectrons generated are then accelerated by the RF fields and go through the emittance compensation solenoid. At this point the beam can be energy chirped inside a 6+2 1/2 cell S-band PWT RF cavity. There is the capability of controlling and measuring independently the phases of the two accelerating structures allowing us to test the ballistic bunching scheme. Downstream of the linac an aluminum foil can be inserted and the transition radiation generated is collected by a parabolic mirror and reflected to a Martin Puplett autocorrelator for pulse length diagnostic. There are also four chicane dipoles along the beamline and two of them can be turned on in a 45 degrees dipole mode. On the 45 degrees beam line there is a quadrupole lens and a Yag screen for emittance measurement via quad scan.

Longitudinal compression results

We measured the pulse length in the frequency domain by Coherent Transition Radiation interferometry. For 210 pC of charge and 70 degrees off crest in the PWT cavity, we obtain the interferogram shown in the figure. The data analysis yields a pulse length of 0.39 ps. The peak current of the bunched beam is in excess of 500 Amps. It is worth noticing the compressed beam is shorter than what was possible to get with magnetic compression for comparable beam charge, confirming the fact that in this case a more linear part of the RF wave is sampled.

Transverse dynamics study

For optimal compression the beam runs through the high gradient structure far from the crest of the RF wave so that the energy spread at the exit of the Linac is very large. For example, for the case in which the focus of the longitudinal lens is 3 m downstream on the beamline, the RF phase was set 70 degrees off crest, resulting in a energy spectrum extending from 5 MeV to 9 MeV. This is a limitation to the determination of the transverse projected emittance because the energy spread translate to an angle spread and will appear to all measurement techniques (that are trace space measurements) as unphysical transverse emittance. On the other hand, the energy is correlated with the longitudinal position of the beam and with a small window of acceptance in energy, a longitudinal slice of the beam can be selected. Experimentally, we used the 45 degrees dispersing bending dipole configuration to select the central beam slice over which a vertical quad scan emittance measurement was performed. We measured the vertical phase space parameters of the electron beam varying the phase of the linac to understand the effect of the compression on the transverse dynamics of the beam.

The experimental data show an increase in emittance even for phases for which the beam is not yet fully compressed. To completely understand this, it requires a deeper look into the dynamics of a chirped beam going through the bending magnet that we used as a slice selector in the measurement.

For the vertical phase space, the bending magnet is in fact just a drift, but if we look at physical beam volume in x-s configuration space, we observe a strong compression if the beam is chirped in energy with particles with higher energy being in the tail of the beam. Note that a bending magnet has a negative R56 so that the compression is anomalous. The sizes of the projections of the beam density onto the curvilinear longitudinal axis s or onto the transverse dimension x and y increase as the beam bends, but the three-dimensional physical volume of the beam gets smaller. The effect of this electronic density spike at the cross-over point is dramatic for the electron beam quality especially at moderately relativistic energies where space charge forces are dominating the dynamics of the beam. For a mathematical treatment of the anomalous compression mechanism based on linear beam matrix formalism we refer to [5].


The Neptune ballistic bunching experiment demonstrated the efficiency of the rectilinear RF compression. A compression ratio in excess of 10 was achieved due to the strong suppression of the effect of the RF-non linearities. Experimental investigation of the transverse phase space quality showed emittance degradation. A deeper look into the beam dynamics shows that the technique used to select a small energy spread slice affects the beam quality in a serious way. A first order linear analysis is performed to quantify the anomalous beam compression of a chirped beam going through bending magnet. Three-dimensional simulations are in reasonable agreement with the experimental data.

The Neptune experiment points out the deleterious effect of crossovers in a space charge dominated beam dynamics, but because of the lack of post-acceleration we can not make conclusive statements on the quality of a beam compressed with a velocity bunching scheme. Future experiments are needed to investigate the full potential of the velocity bunching method for increasing the brightness of photoinjector beams, and the use of the magnetic solenoids to keep the beam phase spaces under control. One important point to be addressed is to investigate the difference between the thin lens version 'ballistic' bunching and the long version of the rectilinear compressor.


[1] S.G. Anderson, et al., ``Horizontal phase-space distortion arising from magnetic pulse compression of an intense relativistic electron-beam.'' Phys. Rev. Lett. 91, 074803 (2003).

[2] L. Serafini, M. Ferrario, ``Velocity Bunching in Photo-Injectors'', AIP Conference Proceedings 581, 87 (2001)

[3] P. Piot, et al., ``Subpicosecond compression by velocity bunching in a photoinjector'', Phys. Rev. STAB, 6, 033503

[4] S.G. Anderson, et al., ``Pulse Compression via velocity bunching with the LLNL Thomson x-ray source photoinjector'' PAC 2003 Proceedings

[5] P. Musumeci, et al., ``Velocity bunching experiment at the Neptune Laboratory'' PAC 2003 Proceedings