16 August 2006

Emittance Simulations

Axel has calculated the bunch lengthening expected at 1.5-m for the ILC bunch (6.4 nC) as a function of cathode bias and initial bunch length. The conclusion is that 300 kV is needed to increase the frequency of SHB1, 500 kV to eliminate SHB1 altogether, and 700 kV to eliminate both SHBers.

There is another interesting question that can be answered by simulations. How much can the beam normalized transverse emittance be lowered by increasing the bias voltage? This involves at least the following:

1. What is the emittance out of the gun as a function of bias. The emittance increase of the gun is presumably due to the focusing effect of the Pierce lens.

2. Each of the bunchers (two SHBers, S-band buncher) have radial fields that increase the emittance. If we can decrease the required longitudinal bunching fields (by going to a higher bias and then using a shorter laser pulse), we would presumably decrease the emittance. Simulations should be able to tell us how much this decrease will be.

Actually a lot of these issues are discussed in Mary James' thesis: SLAC-319 (1987), especially in Section 2.4, "Radial Dynamics." For the SLC injector, she calculates a normalized rms emittance at 30 MeV of 2.5x10^-4 m compared to a measured value of 1.5x10^-4 m. She also estimates an emittance right out of the gun of 3x10^-5 m, probably also pessimistic. Probably she assumed the bias was 150 kV. Also note that the cathode radius for the SLC gun at the time of James' thesis was 0.7 cm.

Since there have been advances in overcoming the surface charge limit (SCL), it is possible now to also consider reducing the laser radius at the cathode. This will decrease the bunch emittance (varies linearly with the radius). The trade off between such parameters as radius and bias for minimizing the bunch emittance is studied for a cw beam (thus much lower charge per bunch) by Bazarov et al. [The initial study was I.V. Bazarov et al., PAC2003, p. 2062; a more thorough study using "multiobjective evolutionary algorithms" is in I.V. Bazarov and C.K. Sinclair, PRST-AB 8, 034202 (2005).] We could do a simpler study varying only a couple parameters, including bias and radius.

Emittance Compensation with a DC Gun (and other topics)

1. Motivation for low emittance beams. Colliders or low DF pulsed accelerators in general have not usually paid much attention to achieving the lowest possible emittance. For ILC, lower emittance has a number of benefits, especially lower lossses upstream of the DR, which in turn will reduce undesirable radiation effects.

2. A simple goal is to preserve the normalized emittance of the photogun, which is about 10^-5 m. With a bias voltage of 120 kV, the present ILC design uses 2 SHBers and an S-band buncher. The bunchers increase the emittance to about 10^-4 m. At 500(700) kV, 1(both) SHBers could be eliminated. A more reasonable goal of 300 kV [a guess] would allow both SHBers to operate so that the nonlinear terms in the bunching forces will be lower. Will this be low enough to be able to effectively use emittance compensation?

3. B. Yunn (PAC01, p. 2254) has claimed that the emittance compensation scheme used for rf guns will work for dc guns if the bunching process is linear. In his paper, he quotes the results of studies of various cw injector designs with emittance compensation, but doesn't actually give any details of the designs other than to bunch early and then accelerate. He states that in some cases the final emittance is lower than out of the gun [this implies that there is a significant correlated emittance at the gun exit] while reducing the bunch length by a factor of 10! But note that Yunn is always working with cw (i.e., high DF) designs, which implies low charge density and thus relatively weak bunching forces.

Apparently Sinclair expects to use emittance compensation at Cornell with his ultra-HV gun, but again his application is a cw injector.

4. Finally of course there are rf guns, which have the advantage of exiting the gun at high energy and already bunched and also with most of the emittance energy-correlated.

J. Clendenin