Replacing 35 Year Old Linac Technologies

1
Replacing 35 Year Old Linac Technologies

• Elliott McCrory
• Based on work by Don Young for Proton Driver 1

2
Basic Idea

• Replace all 35 year old components
• 402 MHz, like SNS
• Reuse 805 MHz systems
• Can get 480 GeV in the same space
• Some of it bought from AccSys Technologies
• Potential collaborations?
• Fermilabs NTF (Arlene Lennox)
• Fermi Institute of Hadron Therapy
• 50M from Hastert?!
• National Taiwan University Hospital, Hsinchu
  Biomedical Science Park
• Taipei, Taiwan
• Gold Platedcan work backwards from this proposal

3
http//www.accsys.com
4
From The AccSys Web Site

• LINSTAR systems are currently in use at several
  facilities around the world 
• A Model PL-2i has been operating at Loma Linda
  University Medical Center since 1990 as the
  injector to the high energy proton synchrotron
  cancer treatment facility. The synchrotron system
  operates 24 hours a day, 6 days a week and has
  been described in a number of technical
  publications. Complete proton therapy systems
  based on the Loma Linda installation are
  commercially available from Optivus Technology,
  Inc.
• Two systems are now in operation in Japan a
  Model PL-3i is the injector to the proton
  synchrotron cancer treatment facility at the
  Shizouka Cancer Center, which was built by
  Mitsubishi Electric Company, and a Model PL-7i is
  the injector for the proton synchrotron cancer
  treatment facility installed by Hitachi, Ltd. at
  the Tsukuba Medical Center.
• A Model PL-7i was operated at the Indiana
  University Cyclotron Facility from 1997 to 2002
  as the injector for the CIS compact synchrotron
  which is in turn the injector for the high energy
  physics cooler ring. This physics system was
  decommissioned. The injector for their cyclotron
  was converted from doing physics research to
  proton cancer therapy and the Model PL-7i linac
  is being upgraded for use as a neutron source for
  research. However, a Model PL-1 RFQ linac is now
  in operation at the facility.

5
Linac Replacement

• 402 MHz front end
• New ion source RFQ
• Two RFQs with double-alpha injection
• New DTL to 90 MeV
• New Transition section
• Gets the match RIGHT!
• Can design to very low losses
• One new 805 MHz Tank to 116 MeV
• Probably room for one or two more 805 MHz tanks
• Sliding existing tanks up one or two slots
• Can get 480 MeV
• Cost Estimate

6
Schematic Layout
Solid NEW
Hatched Existing
RGDTL
DTL Tank 2
DTL Tank 5
DTL Tank 3
DTL Tank 4
RFQ 2
RFQ 1
89 MeV

TransB
TransV
New Module 0
Existing Module 1
To Booster at 480 MeV
Existing Module 7
New Module 8
New Module 9
7
Suggested Parameters
8
Benefits

• Replaces all 35-year-old Linac equipment
• SOA technology
• Increase beam brightness
• Probably 2 to 4 X smaller transverse emittance
• Existing technical staff Days are numbered
• New technical staff can own this new machine
• New machinery and new people might last another
  30 years
• 402 MHz Klystrons exist
• And our 805 MHz klystrons last forever
• Significantly shorter than existing LEL
• Could add more modules at the end to give an
  energy boost
• Higher injection energy into Booster is better
• Reduced losses/activation throughout the Linac
• Transition section is our predominant loss now
• Lower emittance would mean lower 400 MeV-line
  losses

9
Higher Energy Linac Benefits

• The space charge limit for constant aperture at
  Booster injection scales like ß2?3

10
Pet Project (1996)
RFQs
11
Double a Magnet Del Larson
12
Trace-3D Run of PET MEBT
Longitudinal
RFQ 1
RFQ 2
13
From Larsons 96 MEBT paper

• 6.1. Longitudinal Control.
• some control over the longitudinal phase
  space can be exercised by using the trim quads
  located in the cross over arm. The MEBT has been
  designed to have a high dispersion in this
  region, and quadrupole variations in this region
  therefore have a large effect on the dispersion
  function. The dispersion is introduced in the
  first alpha magnet, and the magnet and optics
  have been nominally designed so that the
  dispersion function has zero gradient in the
  middle of the cross over arm. With zero
  dispersion gradient and transverse waists in the
  center of the cross over arm, the beam will have
  symmetric optics about this point, leading to
  zero dispersion and zero dispersion gradient at
  the end of the second alpha magnet. While the
  first goal of the trim quads is to ensure a
  symmetric dispersion function, so as to eliminate
  dispersion at the end of the MEBT, a second
  function can be to control the position of the
  longitudinal waist. Since small trim quad
  variations have a large effect on the dispersion
  and a small effect on the transverse beam optics,
  the longitudinal control exercised by the trim
  quads is largely independent of the transverse
  optics.
• 6.2. Transverse Control.
• Once dispersion has been minimized (by arranging
  for symmetric optics within and between the
  dipoles) the quadrupoles placed downstream of the
  dipoles will have no effect on the longitudinal
  optics and can therefore be used solely for
  transverse optics matching. Since there are three
  such quads, one can in principle exercise control
  over three linear combinations of the four
  independent transverse phase space variables.

14
Rough Cost Estimate
15
Conclusions

• Gold Platedcan work backwards from this proposal