HIGHINTENSITY HEAVYION DRIVER LINAC FOR THE RIA FACILITY

1
HIGH-INTENSITY HEAVY-ION DRIVER LINAC FOR THE RIA
FACILITY

• P. N. Ostroumov, J.A. Nolen, K.W. Shepard

ICFA-HB2004 Bensheim, Germany, October 18-22,
2004
ICFA-HB2004, October 18-22, 2004
2
Outline

• RIA Facility
• Layout
• Major requirements
• Driver linac hardware development
• Uranium beam dynamics in the Baseline Driver
  Linac
• Beam-based steering algorithm
• Two options of the Driver Linac
• High-statistics beam loss calculations in the
  presence of errors
• Proton beam dynamics
• Existing Baseline design
• Possibility to produce 2 MW proton beam
• Summary/Conclusion

3
Layout of the RIA facility (if sited at ANL)
4
RIA Driver Linac
Multi-ion, multi-charge-state, 1.4 GV Ion Linac
400 kW beams of ALL ions from protons to Uranium
Baseline Design 393 SC cavities of 9 different
types arrayed in three linac sections
5
SC ECR Ion Source VENUS, LBNL Development
Superconducting Structure
Beam Transport
8 pµA of U28 and U29
Conventional Components
See D. Leitner, C. Lynes et al LBNL
6
57.5 MHz CW RFQ
Cold Al Model
Exploded View 57 MHz RFQ. Will be built using
high-T brazing
7
Intermediate-velocity cavities have been
developed at ANL
115 MHz ?0.15 Steering Corrected QWR
345 MHz ?0.4 Double-spoke
172.5 MHz ?0.14 HWR
8
All SC prototype cavities demonstrated excellent
performance
2.0K
4.5K
9
Drift-Tube Cavity Cryomodule Design
10
Elliptical-cell cavities and cryomodules have
been developed at JLAB for SNS, beta0.49 is
being developed by MSU and JLAB (for the
baseline design).
From Claus Rode SNS DOE Review, May 2004
11
Triple-spoke resonator option of the driver linac

• Box cryomodule applied to TSR
• preserves convenience of the top-loading design

12
High power stripper foils and films for the
driver linac

• Determination of optimal foil thickness, Z, and
  degree of straggling required experiments at
  Texas AM with 10 MeV/u uranium, GSI with 85
  MeV/u uranium, and MSU with 80 MeV/u
• Results show that optimal thickness and Z for
  the strippers are 300 micrograms/cm2 and Z4 at
  the lower energy, and 15-20 mg/cm2 and Z6-12 at
  the higher energy

• Windowless Liquid Stripper Loop
• Features
• Once through system
• Compressed gas driven
• Rotatable deflector
• Easy nozzle change

13
Multi-Q beam matching, 180? bend, collimation
High-?
QuadrupoleSextupole
Beam losses (400 kW driver beam) main
collimator 2 kW cleaning
collimators 85 W
High dispersion area
86 MeV/u (after the stripping)
Stripper
Medium-?
14
Linac design has been iterated repeatedly

• We are about to start next iteration
• 200 kV HV platform, redesign the RFQ
• Charge states for uranium 34 and 35
• DT resonators 25 MV/m surface field

15
Front End of the Driver Linac
MHB

• Main features
• Two charge states of heavy-ion beams, Agt180
• Achromatic bend
• Sextupole corrections is required to minimize
  transverse emittance growth
• Beam dynamics is dominated by space charge
• Includes Multi-Harmonic Buncher to form very low
  longitudinal emittance of heavy-ion beams.
  Required for multi-q acceleration.
• Deliver any beam from protons to Uranium to the
  fixed velocity acceptance of the RFQ

16
Uranium beam from the ECR
Q1 Q2 Qavr A Ieff
1 2 1.5 238 0.214
3 7 5.0 238
0.511 8 12 10.0 238
0.583 13 17 15.0 238
0.661 18 22 20.0 238
0.731 23 27 25.0 238
0.749 28 238
0.125 29 238
0.125 30 34 32.0 238
0.408 35 39 37.0 238
0.095
I tot, U 4.200 mA 1 1 1.0
16 0.280 2 2
2.0 16 0.208 3
7 5.0 16 0.204
Itot 0.700 mA
17
LEBT must be designed for high current beams
U27, U28 , U29, O2, U30
Total current 4.9 mA U28U29 125 e?A
125 e?A
Extraction voltage is 100 kV
18
Final optimization is followed by end-to-end
simulations
Uranium beam emittance evolution, 106 particles,
no errors except the stripping foil thickness
fluctuation Vertical Longitudinal
Chicane-1 Chicane-2 Chicane-1
Chicane-2
19
Beam-Based Steering

• Multiple charge states effective transverse
  emittance growth due to misalignments.
• Frequent machine-settings retuning to
  accommodate many different ion species.
• Algorithm is an integral part of the TRACK code.
• Is capable of being implemented in real
  machines.
• Method
• Measure beam positions at BPMs
• Apply known deflections (kicks ) to the
  trajectory
• Measure the new beam positions and calculate the
  differences
• Measure beam responses to induced kicks
• Find that minimizes
• Apply steering
• Compensates static misalignment errors.

20
Correction of multi-q beam position (50 seeds)
Residual deviation of beam centroid along the
linac
21
Beam-Loss Calculations

• Final step of BD design studies
• Simulations on the multi-processor computer
• Up to 500 randomly seeded accelerators with all
  types of errors and misalignments, typically 200
  seeds
• Beam steering is applied
• Wide range of rf errors, thickness fluctuation
  and their combinations have been studied
• Number of tracked particles
• Up to 106, typically 2105 in each seed
• Total number of simulated particles 40 million,
  some cases up to 200 million.

22
Two options of the Driver Linac
Baseline Design 6-cell elliptical cavities 805
MHzEpeak27.5 MV/m
Triple-spoke resonators 345 MHzEpeak 27.5 MV/m
23
Beam losses Fractions Locations

• Two types of losses
• Controlled losses Beam halo particles
  intercepted by collimators placed at designated
  areas of the accelerator (example MTS after a
  stripper).
• Uncontrolled losses Beam particles lost in the
  structures of the accelerator resulting in the
  radio-activation of the equipment.
• Case with only stripper thickness 5 (FWHM)
  fluctuations
• No uncontrolled losses.
• Controlled losses 0.2 in MTS-1 and 0.3
  in MTS-2.
• Case with errors
• 6 combinations of errors have been studies (see
  next table).
• Uncontrolled losses observed in the high-?
  section of the baseline design.
• Controlled losses slight increase but remains in
  the 0.2-0.4 range.

24
Beam emittances, image of 40 million particles
Baseline Triple-spoke
25
Beam losses
26
Summary of beam loss studies

• Most critical sources of error
• RF errors (field phase).
• Fluctuations in stripper thickness.
• The Triple-spoke design is more tolerant of
  errors than the Baseline design.
• Uncontrolled losses observed for the Baseline
  design. To keep the losses below the 1 Watt/m
  limit
• Rms RF errors field lt 0.5 and phase lt 0.5 deg.
• Stripper thickness fluctuation lt 5 FWHM.
• No uncontrolled losses observed for the
  Triple-spoke design even with RF errors (0.7,
  0.7deg, RMS) and thickness fluctuation of 10
  FWHM.

27
Proton beam dynamics

• Main issues
• The driver linac is optimized for the heaviest
  ions. Must be retunable for protons and other
  light ions.
• 400 MeV/u Uranium and 1 GeV protons
• Space charge effects of low-energy pre-bunched
  beams upstream of the RFQ

28
Multi-harmonic buncher, 100 keV uranium
RFQ
1.94 m
I0 mA I0.25 mA
29
Multi-harmonic buncher, 12 keV protons
I0 mA I1.0 mA
Multi-harmonic buncher for12 keV, 1 mA protons
does not work !
30
Bunching of 1 mA, 12 keV protons by MHB
31
To improve bunching of high-current protons in
the baseline design of the driver linac

• Install a new buncher operating at 57.5 MHz 50
  cm upstream of the RFQ
• Increase the voltage on the RFQ vanes by 20
  with respect to design value for protons. It
  produces larger separatrix, synch phase -40 deg.

32
Single-harmonic buncher (57.5 MHz) RFQ
RFQ voltage is higher than design voltage by 20,
provideslarger longitudinal acceptance for
high-current beam
200 keV, 0.55 mAExit of the RFQ
12 keV, 1.0 mAEntrance of the RFQ
33
Single-harmonic buncher (57.5 MHz), 12 keV
protons, baseline
Fraction of particles outside of a given
longitudinal emittance as a function of the
emittance.
34
2 MW proton beam, a new injector
RFQ with adiabatic bunching, capture efficiency
is 95-98.
35
Periodic focusing structure, second cryostat
2 mA, proton beam rms and total envelopes
36
High-current proton beam matching, W1 MeV
Betatron and synchrotron oscillations frequencies
as a function of proton beam current in the
second cryostat
37
Beam loading, 400 kW Uranium
38
Beam loading, 2 MW protons
39
Summary on proton beam dynamics

• Baseline design
• Modify bunching system fro protons and other
  light ions
• Rebunchers on post-stripper beamline should be
  matched to proton velocity
• Further BD optimizations and high-statistics
  calculations of possible beam losses are required
  for protons and light ions
• If necessary the driver linac can be upgraded for
  production of 2 MW proton beams
• No limitations from the beam dynamics point of
  view. In the SC linac 2 mA cw beam is emittance
  dominated
• Minor modification of the injector
• Rf power per resonator is higher by 2-15 kW with
  respect to the existing baseline design
• Increased shielding with respect to the existing
  baseline design.

40
Conclusion

• RIA facility is ready for the CD-1 and CDR

Many thanks to our colleagues from ANL and
numerous collaborators AES DESY JLAB
GSI LANL INFN-Legnaro LBL TRIUMF MSU