Development of Non-Scaling FFAG Takeichiro Yokoi John Adams - PowerPoint PPT Presentation

Development of Non-Scaling FFAG Takeichiro Yokoi John Adams Institute for Accelerator Science Oxford University RCNP 研究会「ミュオン科学と加速器研究」 20/10/2008

Introduction ... 
 • FFAG( F ixed F ield A lternating G radient) Accelerator has an ability of rapid particle acceleration with large beam acceptance. ⇒ wide varieties of applications EMMA Particle physics ν -factory, muon source, proton driver Medical PAMELA FFAG Particle therapy, BNCT, X-ray FFAG source (PAMELA) Energy FFAG ADSR, Nucl. Transmutation

CONFORM : Co nstruction of a N on-scaling FF AG for O ncology, R esearch and M edicine PAMELA (PM: K.Peach) EMMA ( PM: R.Edgecock ) Rutherford Appleton Lab Rutherford Appleton Lab Daresbury Lab. Daresbury Lab. Cockcroft Ins. Cockcroft Ins. Manchester univ. Manchester univ. Oxford univ. John Adams Ins. John Adams Ins. Imperial college London BNL (US) Brunel univ. FNAL (US) Gray Cancer Ins. CERN Birmingham univ. LPNS (FR) TRIUMF (CA) FNAL (US) LPNS (FR) TRIUMF (CA)

What is NS-FFAG ? ⇒ Fixed field ring accelerator with “small dispersion linear lattice” ① Orbit shift during acceleration is small ⇒ Small Magnet aperture, energy variable extraction ② Path length variation during acceleration is small ⇒ fixed frequency rf can be employed for relativistic particle acceleration |df/f|~0.1% ~20mm 10MeV TOF/turn(ns) B 0 
 /cell ∆ r/r<1% ① Simple and flexible lattice configuration ⇒ tunability of operating point Δ x 
 20MeV ② Large acceptance B 0 = Δ x × B ③ 1 Large tune drift ( focusing power ∝ B/p ) Kinetic Energy(MeV) ⇒ Fast acceleration is required /cell

EMMA: Electron Model for Many Applications 
 Muon Acceleration  
 Electron NS-FFAG as a proof of principle is to be built as 3-year project. (host lab: Daresbury lab.)  It is also a scaled-down model of muon accelerator for neutrino factory.  Research items are . . . (1) Research of beam dynamics of NS-FFAG (2) Demonstration of NS-FFAG as a practical accelerator (3) Demonstration of fast acceleration with fixed frequency RF Number
of
Cell
 42

(doublet

Q)

 Circumference
 16.57m
 Injection
energy
 10~20MeV(variable)
 Extraction
energy
 10~20MeV(variable)
 RF
 1.3GHz
 Acceptance
 3mm(normalized)


EMMA :Beam acceleration 
 10MeV
 /cell
 20MeV
 /cell
 |df/f|~0.1% 20MeV 








TOF/turn(ns)
 *
 In EMMA, Acceleration completes within 10turns(~500ns) 10MeV EMMA is a unique system to observe transient process of resonance precisely. ⇒ Unique Kinetic
Energy(MeV)
 playground for nonlinear dynamics !!

PAMELA: Particle Accelerator for MEdicaL Applications 
 PAMELA : design study of particle therapy facility for proton and carbon using NS-FFAG ( prototype of slow accelerating NS-FFAG ⇒ Many applications!!! Ex. ADSR ) Difficulty is resonance crossing in slow acceleration

PAMELA:Clinical requirements Dose uniformity should be < ~2% ⇒ To achieve the uniformity, precise intensity 
 SOBP
is
formed
by
 modulation is a must superposing
Bragg
peak 
 IMPT ( I ntensity M odulated P article T herapy) 


 Synchrotron

 Integrated
current 
 &
cyclotron 
 Gate width Beam of FFAG is quantized. controls dose 
 ⇒ At the moment, instead of modulating the time 
 intensity of injected beam, shooting a voxel with multiple bunches is to be employed. “Analog IM” FFAG 
 If 1kHz operation is achieved, more than 100 Integrated
current 
 voxel/sec can be scanned even for the widest SOBP case. Step size ⇒ 1 kHz repetition is a present goal (For controls dose 
 proton machine : 200kV/turn) time 
 “Digital IM” PAMELA meets muon science !!

PAMELA : Beam Dynamics 
 Integer resonance dx: 100µm(RMS) rf: 5kv/cell dx: 10µm(RMS) Half integer resonance dx: 1µm(RMS) Beam blow-up rate can be Field imperfection severely affects estimated quantitatively beam blow up in the resonance crossing

Requirements for lattice eV(MeV/turn) Integer resonance Theoretical value ( ν =6,1 π mm mrad.norm) kV/turn 210 210 260 260 320 320 σ ( µ m ) 70 90 90 70 90 70 σ pos (m) Integer resonance blowup constant  For slow acceleration case, (~200keV/turn) integer resonance crossing should be avoided.  Single half integer resonance crossing would be tolerable  Structure resonance also should be circumvented. Linear NS-FFAG (200kV/turn, average B 0;n, , w/o ∆ B1, σ x =100 µ m)

PAMELA : Lattice Integer resonance crossing must be circumvented. ⇒ Tune-stabilization by introducing higher order multipole field is required One option : Non-Linear NS-FFAG (simplified scaling FFAG) : B=B 0 (R/R 0 ) k ⇒ B=B 0 [1+k ∆ R/R 0 +k(k-1)/2 ( ∆ R /R 0 ) 2 ···· ] * Eliminating higher order multipoles ~2m (1) Long straight section (~2m) (2) Small tune drift ( <1) (3) Short beam excursion(<20cm) by S. Machida(RAL) (4) Limited multipoles (Up to decapole)

PAMELA : Magnet  Applicable to superconducting magnet  Well-controlled field quality  Present lattice parameters are within engineering limit ⇒ Feasible option for magnet !! 40cm Sectapole Quadrupole Dipole ~17cm Octapole Decapole by H.Witte (JAI)

PAMELA : Magnet (cnt’d) Quadrupole Dipole Sextapole Octapole

Acceleration Rate (1) Half integer resonance eV/turn(MeV)   :200kV/turn ε 1 / ε 0 1 :50kV/turn ε 1 / ε 0 ε 1 / ε 0 ∆ B 1 /B 1 1 ∆ B 1 /B 1 ∆ B 1 /B 1 eV/turn ∆ B 1 /B 1 (MeV) (2) 3rd integer resonance  Nominal blow-up margin : 5 (1 π mm mrad → 5 π mm mrad) ε 1 / ε 0 -1  With modest field gradient error (2 × 10 -3 ), acceleration rate of 50kV/turn can suppress blow up rate less than factor of 5.  For the considered range, 3rd integer resonance will not cause serious beam blow-up eV/turn ∆ B 2 /B 2 (MeV) ⇒ Required accelerating rate : >50kV/turn

PAMELA: Beam Acceleration Repetition rate: 1kHz ⇔ min. acceleration rate : 50kV/turn (=250Hz) ⇒ How to bridge two requirements ?? Option 2 Energy Option 1 Energy time time 1ms 1ms Low Q cavity (ex MA) can mix wide range of frequencies ( Σ V ) 2 ∫ P = dt R ( Σ V ) 2 ≡ ( Σ V i sin [ f i ( t )]) 2 Option 1: P ∝ N rep 2 i ( V i sin [ f i ( t )]) 2 + Σ i ≠ j ( V i sin [ f i ( t )] ⋅ V j sin [ f j ( t )]) = Σ Option 2: P ∝ N rep 1 ∫ dt  →  0 T Multi-bunch acceleration is preferable from the viewpoint of efficiency and upgradeability

Multi-bunch acceleration Multi-bunch acceleration has already been demonstrated ∆ f ≥ 4 f sy 2-bunch acceleration using POP-FFAG : Mori et al. (PAC 01 proceedings p.588) In the lattice considered, typical synchrotron tune <0.01 ⇒ more than 20 bunches can be accelerated simultaneously (6D Tracking study is required) “Hardware-wise, how many frequencies can be superposed ??”

Test of multi-bunch acceleration Extraction (5.5MHz) 50kV Injection � (2.3MHz) 50kV PRISM RF  PRISM rf can provide 200kV/cavity  It covers similar frequency region  B rf -wise, MA can superpose more than 20 bunches ⇒ Now, experiment using PRISM cavity is under planning ( in this October )

Applications for ADS Accelerator Driven System  ADS will be used for ADSR, nucl. transmutation.  ADS will employ high power low energy proton accelerator as proton driver (<1GeV, >1mA)  FFAG, cyclotron, LINAC are the candidates  Key issues are cost and reliability (how to realize redundancy ?)  From the view point of redundancy, FFAG is a competitive candidate. ⇒ Proton driver for ADS is one of main applications for PAMELA type FFAG. EADF parameters

Multi-turn extraction in NS-FFAG Why?  Circulating bunch = extracted bunch ⇒ Low bunch intensity for spot scanning  For energy variable extraction, extraction system is required to be moves mechanically due to the radial orbit shift especially for HI ring (problems: response time, reliability)  Number of bunch accelerated simultaneously is limited by kicker aperture. ( For the kicker aperture of 2cm, minimum orbit separation is ~4cm. ) ⇒ charge exchange injection is preferable from this point of view  ( Life time of kicker ? : ex 10 6 msec = 1000 sec = 17min )  For the application of ADSR, pulsed beam structure might not be preferable from the viewpoint of ADS core damage B kicker ∆ x ~ aperture

Multi-Turn Extraction in NS-FFAG (cnt’d) ν H ν v ∆ ν v <0.5 ~2% of F/D ratio can change the vertical tune more than 0.5 ⇒ In a lattice with vertical tune drift, by changing the F/D ratio, resonance energy can be varied ⇒ Half integer resonance can be used for the extraction : “ Energy variable multi-turn extraction in fixed field accelerator” ” With present design strategy, is it possible to develop a lattice with vertical tune drift of less than 0.5? ” ⇒ If it is realized, it will solve almost all the problems in PAMELA

a layout Fast extraction (horizontal) Slow extraction (vertical) Charge exchange injection(horizontal) p HI Proton ring Fast extraction (horizontal) 1turn injection HI ring (horizontal)