Minutes of the MUCOOL meeting, Feb. 12 1999, Fermilab.
(Please report any errors to kaplan@fnal.gov.)
I. Collaboration meeting
---------------------
Steve began the meeting by reminding us of the St. Croix collaboration
meeting at the end of May, for which the agenda is being drafted.
Suggestions are welcome. Rick Fernow is in charge of the Cooling Theory
agenda and Steve is in charge of the MUCOOL agenda. The plan is to have
all plenary sessions this time, instead of parallel sessions as at
previous meetings.
II. RF update
---------
Al reports that Lab G is being cleared out. The magnets are to be moved
out starting probably Monday and will probably take about 2 weeks, during
which time we are to stay out for safety reasons. The PET apparatus has
been packed and is ready to be moved to storage. We expect occupancy the
1st or second week of March.
There is a new design for the windowless cavity with 8 cells (vs. the old
design with 7 cells, including a 1/2-cell at either end). Paul is
simulating the old version (modified with a "cheat" for some reason) and
reports that preliminary results seem too good to be true! He plans to
discuss this further at Monday's Cooling/Optics meeting.
There ensued some discussion of what conventions Al and Paul are using and
whether they are using the same ones. Al states that the accelerating
field on-axis is 30 MV/m = 0.78 E_p cos(32 degrees), where 0.78 is the
transit-time factor and E_p is the peak field on-axis of 60 MV/m,
correspondong to a peak surface field of 90 MV/m. (These numbers are
almost consistent - checking the arithmetic I get 40 MV/m for a peak field
of 60 MV/m on-axis.)
Muzaffer asked whether 90 MV/m would be too much. Jim responded that 120
MV/m is feasible. It helps that this is a pulsed system.
III. Cooling/Optics meeting
----------------------
Paul briefly summarized Dave Carey's talk at the last Cooling/Optics
meeting, for which see Paul's minutes (which he should put in a public
area of our website). Dave plans to work out an analytical transfer
matrix for cavities in a solenoid (and cavities in solenoid fringe fields)
that can be put into Transport.
IV. Pion production and capture
---------------------------
Jim summarized his design for production and capture of pions using a long
shaped solenoid. For a field of 8 T, operating at 460 MeV/c with a 40
MeV/c momentum bite, he finds he can capture 2.8e6 muons/bunch. The
radius is <10 cm and the length is about 25m at a cost of about 0.378
M$/m. The RF design is yet to be worked out. Steve and Jim disagree on
the utility of this approach, since Steve wants a short bunch (~50 ps,
to simulate the beam expected at the input to the 15T cooling section) and
Jim's channel produces a long one (filling the 805-MHz bucket).
V. Wedge simulation
----------------
Rick reported 3 items:
1. In Nashville Bob requested modifications to ICOOL so he could simulate
Jim's matching scenario.
2. Juan checked ICOOL against a transport-matrix calculation and concluded
that the latter calculation is too simple to describe the effects of a
bent solenoid.
3. Rick is revisiting the problems of his previous calculation: poor
matching and dispersion not entirely along the y axis.
VI. Analytical bent-solenoid calculation
------------------------------------
Jim reports that Scott Berg is developing Hamiltonians for bent solenoids.
VII. Lithium-Lens Review
-------------------
Norbert reported on the recent Lithium-Lens Review at Fermilab.
Silvestrov has built at Novosibirsk a 15-cm-long by 2-cm-diameter
liquid-lithium lens and successfully operated it up to 10T surface field
at up to 630 kA (his supply can go beyond 1 MA but he didn't want to
push too hard at first). For now it has stainless-steel end plugs
instead of beryllium. There will need to be titanium foils added (of
thickness 0.1 to 0.5 mm, to be determined) to protect the beryllium from
embrittlement before a test with beryllium can be done.
This summer the plan is to test at 13T for at least 1 million pulses.
Delivery to Fermilab is projected in early- to mid-2000 for use in the
pbar source.
A new agreement is being drafted for a 1-m-long by 1-cm-diameter lens with
15 or 20T surface field. This will cost about $1.4M and awaits funding.
It would require a lithium flow of 6 m/s so Norbert expects some
vibration.
VIII. Liquid-lithium-lens cooling simulation
--------------------------------------
Panagiotis discussed various options that might be considered, including
cooling at 287 MeV/c with 400-MHz RF and cooling at 473 MeV/c at 805 MHz.
He has simulated two different sections for the latter option, both based
on Valeri's approximate calculation. They correspond to Valery's 1st and
13th stages, but without the bent solenoids and wedges. Each section
consists of two lenses separated by shaped solenoids and RF.
The simulated 1st stage works reasonably well, with a 6D-emittance
reduction factor of about 2/3 and total loss about 5% (including decays).
This is in good agreement with Valeri's calculation. However, the beam
envelope in the RF is too big (16-cm diameter) for the RF iris. Valeri
commented that the solenoid field needs to be 4 times higher (2 times for
the matching solenoids). (Panagiotis assumed 2.5T and 3.3T for the
upstream and downstream matching solenoids and 0.78T for the central
solenoid.) The bucket develops a long tail in dt, assuming the usual
"alpha" shape in dp vs. dt. Aside from the tail, the main distribution
narrows in dt (by perhaps a factor 2 - hard to tell quantitatively from
scatter plots) from entrance to exit.
The simulated 13th stage has problems. There are mismatches between the
1st lens and the solenoids and between the solenoids and the 2nd lens,
leading to emittance blowup and large (30%) losses. Valeri pointed out
that his approximate calculation had no fringe fields - it appears
Panagiotis kept Valeri's peak field but introduced fringe fields, thereby
changing the field integral, to which performance is extremely sensitive.
Evidently Panagiotis needs to increase the peak solenoid fields to
maintain the same field integral as in Valeri's calculation. He and
Valeri will get together and make sure they understand what each
did/should do.
Panagiotis ponted out that we are not really simulating Bob's "baseline."
He referred to the famous figure from Bob's "toy calculation" of emittance
and energy vs. distance that shows the longitudinal emittance blowing up
at the end. Dave Neuffer explained that Bob used "antiwedges" at the end
to decrease the transverse emittance at the expense of longitudinal
emittance, based on the desire for extremely small transverse emittance
for a high-energy collider - this may not be appropriate for a First Muon
Collider. We agreed that we are not really simulating the "baseline" -
but we think what we are doing is sensible. In part we are motivated by
cooling (both with alternating solenoids, given the shorter lattice and
Paul/Alvin's new multiple-scattering calculations, and with lithium lenses
in Valeri's calculation) seeming to work better now than Bob assumed.