CLIC


3. Drive Beam and RF Power Source

3.2 Drive-beam injector

3.2.1 Overview of the system

 

The basic scheme for the drive-beam injector has been further developed and adapted to the new parameter set. The total pulse, at the injector exit, is 92 ms (Fig. 3.3 ) and is composed of 32  ¥  22 subpulses as mentioned in Section 3.1 . The time structure of this pulse is produced after the thermionic gun in a subharmonic buncher, in such a way that the electron bunches of each subpulse occupy alternately even and odd buckets (Fig. 3.3). This is indeed the requisite for subsequent separation of the subpulses, and their recombination in the combiner rings.

Fig. 3.3 : Combined pulses at the injector exit or linac entrance.

Figure 3.4 shows a layout of the CLIC drive-beam injector. The latter is composed of five subsystems:
1) A thermionic gun; 2) A bunching system providing a bunched beam at 10 MeV; 3) An injector linac accelerating the beam up to 50 MeV; 4) A spectrometer line with beam diagnostic and collimation; 5) A matching section to the drive-beam accelerator.

Fig. 3.4 : Layout of the CLIC drive-beam injector.

The thermionic gun

The gun provides a beam at 200 keV. The rise and the fall time of the electron pulse from the gun are assumed to be linear. Their duration ranges between 10 and 50 ns. The power calculated for the gun and for the grid is based on the total charge of about 1300mC. It takes into account the bunching efficiency and the collimation in order to get 750mC at the injector exit. At 200 kV, the gun provides an energy of 253 J. With a repetition rate of 100 Hz, the beam power out of the gun is 25 kW. The average current delivered by the gun is 0.127 A. The duty cycle is 0.0092. It implies an average current of 13.8 A during the pulse. The gun is of the triode type. With the cathode at -200 kV, a pulsed grid allows this peak current. The grid limits the electron current (Ib) by space charge effects [3.2]. The grid voltage is plotted versus the beam current on Fig. 3.5 based on the LEP Injector Linac (LIL) gun characteristics. The CLIC grid voltage is estimated to be around 600 V.

Fig. 3.5 : Grid voltage versus beam current.

A pass tube is fed from a power supply and provides the -200 kV. A function generator drives the grid through optical fibres and the linear power amplifier must have the necessary bandwidth to reach the required grid modulation. A ripple of one per mil, during the 92 ms flat-top, could be obtained with the voltage. The beam emittance was measured out of the LIL gun [3.4] . The measured geometrical emittance containing 85% of the beam is 120 mm.rad (not normalized) for a beam current of 1 A.

The bunching system

The bunching system is composed of two subharmonic bunchers (SHB1 and SHB2) working at 937/2 MHz, followed by one standing-wave pre-buncher (B1) and one travelling-wave buncher (B2), both working at 937 MHz. The bunching efficiency is estimated at 66% (based on LIL) while the collimation efficiency (at 50 MeV) is estimated at 90%. After a drift of 50 cm from the gun exit, the first subharmonic buncher (SHB1) receives a modulation of 10 kV. The second subharmonic buncher (SHB2) is 2 m downstream and works with a modulation of 50 kV. The phase of the SHBs is rapidly switched by 180° every 130 ns, in order to produce the `phase-coded' subpulses. Then the pre-buncher (B1) is 1.25 m downstream and works with a voltage of 100 kV. Finally the buncher B2 continues the bunching process while accelerating the beam up to 10 MeV. It has 12 cells (4 ¥  3 cells) with a phase advance of 2p /3. The magnetic field along the front end keeps the beam sizes to a reasonable value of 10 mm (90% of particles). The front end is roughly 5 m long.

The injector linac

The linac is composed of three damped and detuned accelerating structures (S1 to S3) and works in the fully loaded steady-state mode. It accelerates the beam up to 50 MeV with travelling wave sections at 937 MHz. A long sequence of solenoids along the entire section is installed for focusing. The accelerating structures are 3.4 m long with a loaded accelerating field of 3.9 MV/m. They are similar to those of the drive-beam accelerator. The injector linac provides an energy gain of 40 MeV. Adding the beam energy from the front end, the total energy at the injector linac exit is around 50 MeV. A beam collimation is implemented with losses up to 10% before injection of the beam into the accelerating linac.

The matching section and the spectrometer line

A matching section composed of two quadrupole triplets is implemented at the injector exit (50 MeV). A chicane will allow the beam collimation (3sE) in order to achieve the required beam characteristics. Losses up to 10% before injection of the beam into the drive-beam accelerator could be accepted. A spectrometer line will be used for beam diagnostics and beam performance optimization.

3.2.2 Beam characteristics required at injector exit

The main condition to be fulfilled is that the total number of bunches should be a multiple of 32 since there are two combiner rings giving a multiplication factor of 4 and a delay loop providing a multiplication factor of 2.

The energy spread at the exit of the injector linac is partly correlated due to the short-range wakefields. The uncorrelated energy spread is assumed to be 0.75% and the total energy spread is assumed to be less than 1%. Such a value is expected after beam collimation around 3sE before the injection in the accelerating linac. An r.m.s. value of 4 mm for the single bunch length could be achieved at the injector exit. A large beam size could cause losses in the decelerating structures. Therefore, it is crucial to obtain at 1.2 GeV a beam with an emittance as small as possible. Assuming an emittance blow-up of 50% between the injector and the decelerator, an upper limit of 100mm.rad is required at 50 MeV. Table 3.2 summarizes the beam characteristics required at the injector exit.

Table 3.2 : Beam characteristics required at the injector exit

 

3.2.3 Beam dynamics simulations

Beam dynamics simulations are done with the PARMELA code [3.5] . Preliminary simulations have been performed for the bunching system (up to B2) and with an energy gain of 3 MeV. As initial beam conditions, a given number of particles (between 100 and 500) are generated randomly in a four-dimensional transverse hyperspace with uniform phase and random energy-spread at the gun exit. The normalized r.m.s. emittance is 10mm.rad at this point and 100% of particles are assumed to be confined within six times this emittance. At 200 keV a total emittance of 62m m.rad is used for the simulations. With a radius of 10 mm for the hole in the anode, the horizontal (and vertical) b-value of the ellipse is 1.61m/rad. The r.m.s. beam radius is s =7 mm. A straight ellipse (a  = 0) is assumed in both transverse planes at the gun exit. The longitudinal coordinate z =0 is taken at the anode exit where the magnetic field is still zero. The total charge of 29.2 nC (1.8 ¥ 1011 e-) is distributed over two RF periods of 937 MHz. The transverse focusing is provided by a longitudinal magnetic field satisfying the Brillouin-flow condition. Both axial bunch velocity and bunch current are functions of z . For preliminary simulations, the magnetic field is represented by a fast rise at the beginning, starting from zero and followed by a constant amplitude along the front end (ª 5 m). Fig. 3.6 shows the energy gain from the gun exit.

Figure 3.7 shows the horizontal beam envelopes along the front end. The continuous curve is the r.m.s. envelope and the dotted curves are the 90% and 100% beam envelopes. The entire beam remains inside the 40 mm aperture (radius) of the travelling-wave structures. The buncher's aperture (radius) is 53 mm. The core of the bunch has an extension of 22 degrees while the total bunch (a few particles in the tails) extends over twice this value. The full energy spread is roughly 10% which corresponds to a r.m.s. value of a few per cent. Figure 3.8 shows the bunch length obtained at 3.7 MeV.

Figure 3.9 shows the initial phase versus the current phase. It provides a figure of merit of the bunching efficiency. For this run with 500 particles and an initial phase of 500°, one obtains 300° in the correct bucket. The bunching efficiency in this case is 60%.

The normalized emittance is 245 mm.rad. The results presented here are preliminary. The target values for the injector are not yet completely reached and further optimization is necessary. The input and resulting parameters are given in Table 3.3.

Fig. 3.6 : Energy gain along the front end

Fig.3.7 : Horizontal beam envelopes
(r.m.s., 90% and 100% of particles)

Fig. 3.8 : Bunch length at the front-end exit.

Fig. 3.9 : Bunching efficiency of the injector.

 

 

3.2.4 A photo-injector option

The possibility to use a RF photo-injector as the drive-beam source is under investigation. Figure 3.10 gives a sketch of a possible layout. A CW laser working at 468.5 MHz provides a continuous train. During 92ms the necessary power at 262 nm is generated in order to create the charge of 750 mC by the photocathode. The `even' and `odd' photon pulses are directly produced by electro-optics components. The resulting laser beam illuminates the photocathode of an RF gun powered by a klystron at 937 MHz. It generates an electron beam with a momentum of several MeV at the exit of the photo-injector and with the required sequence of bunches which can then be directly injected into the drive-beam accelerating linac. Such an option would represent several advantages: it would replace the thermionic gun and the bunching system by a single RF gun; the pulse shaping would be much easier and could be adjusted in order to optimize the RF power generation; small emittances would be achievable. However, several issues remain to be addressed: the UV power and stability for the laser and the necessary charge for the photocathodes. An R&D programme has been set up to try to overcome these problems.

Fig. 3.10 : Possible layout for a photo-injector option.

 


| Welcome | Table of Contents | Overall Layouts | Publications | Minutes |

sln December 2000