Speaker: Bruno Muratori
Abstract: Laser wakefield acceleration, together with other types of novel acceleration techniques, has seen considerable progress of late. Together with this progress comes a question, which has only recently started to be addressed, of how to transport and utilise such beams. This is a challenge because of the high initial divergence of these beams. There are several approaches to this problem and we concentrate on one in this paper and look at the implications of it in some detail. The approach we take is to have an FFAG channel made of permanent possibly Halbach magnets. In particular, we look at the implications of quadrupole fringe fields on the aperture of these permanent magnets. It has recently been shown that the pole shape of the quadrupoles influences the fringe field and, further, can have a particularly large effect the more the beam goes off the transverse axis. This could lead to a further restriction in aperture, above the one due to the divergence of the beam exiting the plasma. Several properties of this channel as well as possible geometric rearrangements of it are also discussed.
9 replies on “An FFAG Channel for the Transport of Laser Wakefield Accelerated Beams”
For those unfamiliar with GPT, please see
https://pure.tue.nl/ws/files/1982877/200111445.pdf
Postscript. What is difference between F0D0 transport channel and FFAG quadrupole channel?
I’m looking at the plots for MAD and GPT and I’m not seeing the level of differences you’re talking about. I’m comparing slide 5 and slide 14. With the beta functions shown on slide 5, the beam size would reach about 0.7 mm, as you stated on slide 14. The GPT beam size appears to reach about 1.6 mm; larger rather than smaller, and not orders or magnitude different. I can’t see the paraxial approximation being the issue: the initial RMS divergence is only 1.6 mrad, and I wouldn’t expect it to get to more than around twice that. Are the GPT quadrupoles modeled as having soft edges? That could be enough to lead to the level of differences that I think I see there.
Have you looked at phase space plots for any indicators of nonlinearity appearing (indicating effects of the paraxial approximation or nonlinearities from edge effects in GPT)?
Yes, you are correct. At the moment there is no difference between what I presented and an ordinary FODO channel. My student is currently working on an upgrade to this that would show why I chose this name, but we are not there yet. Maybe she can give a talk at next year’s workshop …
Yes, thank you for the comments. I have to admit that I was a little surprised by these differences … I blame my own stupidity, together with doing things at the last minute … In principle, the quadrupoles do not have any fringe fields but I need to check this with the GPT authors. They may have implemented a default which I have not unserstood properly or am unaware of.
Thanks for the comment. You are correct and this will be looked at next but has not been done yet.
The other thing is, if there is any space charge at the exit of the plasma channel, that would make the beam larger than linear optics would predict.
If you assume zero space charge and perfectly linear transport, in theory you can do anything, but in practice the beams don’t “like” transitioning suddenly from very low to very high betas, you either get space-charge-type effects that are mismatched, or it just becomes very sensitive to errors.
Thanks for the comment. I did consider a case with space charge at 100 MeV on sldes 19 & 20. You are correct that the focus is no longer as precise and the beam is larger. At 1 GeV and for the bunch length considered, space charge makes so little difference that I did not include it. The point is that, this is the only way such a beam can be controlled, from the literature so far. So, through a combination of collimation, optics and such a channel, it should be possible to select out a part of the bunch exiting from the plasma that has longitudinal properties that can be manipulated to yield the desired output.
Thanks again for the comment. I thought everyone knew about GPT, sorry.
I now realised that the link you gave is a link to Bas & Marieke’s PhD thesis. Perhaps a better link would be:
http://www.pulsar.nl/gpt/