1. Plans for a Superconducting Proton Linac at CERN
Part 1
R. Garoby – 10/06/2010

2. OUTLINE Introduction (SC linac)
Low Power SPL in a new LHC injector complex
High power SPL as a proton driver
R & D plans

3. Introduction: Basics of Superconducting RF Linacs*
* Material from M. Vretenar and F. Gerigk

4. Why Linear Accelerators ?
Linacs are mostly used for:
1. Low-Energy accelerators (injectors to synchrotrons or stand-alone)
for protons and ions, linear accelerators are synchronous with the RF fields in the region where velocity increases with energy. As soon as velocity is ~constant, synchrotrons are more efficient (multi-pass instead of single pass).
2. Production of high-intensity proton beams
(in comparison with synchrotrons, linacs can operate at higher repetition rate, they are less affected by resonances and have losses more distributed  more suitable for high intensity beams (but in competition with cyclotrons…).
3. High energy lepton colliders for electrons at high energy,
the main advantage being the absence of synchrotron radiation.

5. Proton and Electron Velocity
b2=(v/c)2 as function of kinetic energy T for protons and electrons.
electrons
Relativistic (Einstein) relation:
Classic (Newton) relation:
“Einstein”
protons
“Newton”
Protons (rest energy MeV): follow “Newton” mechanics up to some tens of MeV (Dv/v < 1% for W < 15 MeV) then slowly become relativistic (“Einstein”). From the GeV range velocity is nearly constant (v~0.95c at 2 GeV)  linacs can cope with the increasing particle velocity, synchrotrons are more efficient for v nearly constant.
Electrons (rest energy 511 keV, 1/1836 of protons): relativistic from the keV range (v~0.1c at 2.5 keV) then increasing velocity up to the MeV range (v~0.95c at 1.1 MeV)  v~c after few meters of acceleration in a linac (typical gradient 10 MeV/m).
Accelerating gradients that we can provide with a linac is nearly constant (2 MeV/m) --- synchrotron radiation ((beta*gamma)**4)/(r**2)

6. Synchronism condition
The distance between accelerating gaps has to be proportional to particle velocity
Example: a linac superconducting 4-cell accelerating structure
Synchronism condition:
t (travel between centers of cells) = T/2
In an ion linac, cell length has to increase (up to a factor 200 !) and the linac will be made of a sequence of different accelerating structures (changing cell length, frequency, operating mode, etc.) matched to the ion velocity.
For electron linacs, b =1, l =l/2  An electron linac will be made of an injector + a series of identical accelerating structures, with cells all the same length
Note: in the example above, the increase in beta inside the structure is neglected!

7. Superconducting Proton Linac
The same cavity shape can be adapted for use at different proton velocities (beta’s), just changing the cell length accordingly to beta.
… but the real estate
accelerating gradient
reduces with b
Need for different
cavity shape/design
at lower b
b=0.52
b=0.7
b=0.8
Cell length has to be matched to the increasing particle velocity
b=1
CERN (old) SPL design, SC linac MeV, 680 m length, 230 cavities

8. Why are sc cavities attractive?
Instead of Q values in the range of ~104 for normal conducting cavities, superconducting cavities easily reach , which drastically reduces the surface losses Pd (basically down to ~0)
➜ large accelerating gradients with minimal peak RF power
(= power required for beam acceleration in the structure)
However, due to the large stored energy, the cavity filling time tl increases (often into the range of the beam pulse length), and hence the average RF power
(only valid for SC cavities)

9. Low Power SPL in a new LHC injector complex
* Detailed during Workshop on
“New Opportunities in the Physics Landscape at CERN”
CERN, May, 2009

10. Motivation for new injectors
1. Reliability ­
The present accelerators are getting old (PS is 50 years old…) and they operate far beyond their initial design parameters
Interest of new accelerators tailored
to the LHC and its future needs
2. Performance ­
Luminosity depends directly upon beam
brightness N/e*
Brightness is limited by space charge at
low energy in the injectors
Þ Need to increase the injection energy in the synchrotrons
New LHC injectors

11. Description Present Future Linac2 Linac4 PSB LP-SPL New LHC injectors
Proton flux / Beam power
LP-SPL:
Low Power-Superconducting Proton Linac (4 GeV)
PS2:
High Energy PS (~ 5 to 50 GeV – 0.3 Hz)
sLHC:
“Super-luminosity” LHC (up to 1035 cm-2s-1)
50 MeV
Linac2
Linac4
160 MeV
1.4 GeV
PSB
LP-SPL
4 GeV
New LHC injectors
26 GeV PS
PS2
Main requirements of PS2 on its injector:
50 GeV
Output energy
Requirement
Parameter
Value
2.2 x ultimate brightness with nominal emittances
Injection energy
4 GeV
Nb. of protons / cycle for LHC (180 bunches)
6.7 ´ 1013
Single pulse filling of SPS for fixed target physics
Nb. of protons / cycle for SPS fixed target
1.1 ´ 1014
450 GeV
SPS
LHC /
sLHC
7 TeV

12. Comparison LP-SPL / RCS
Summary table comparing:
RCS (10 Hz) filling PS2 in 14 pulses (1.3 s) + multiple gymnastics in PS2
LP-SPL (2Hz) filling PS2 in 0.6 ms + no gymnastics in PS2
Filling time PS2
Time structure for LHC
Relative proton rate
Fixed target physics
Heavy Ions
Upgrade potential
Relative Cost1
LP-SPL
0.6ms
inherent
2.5
ideal OK
high
1.28
RCS
1.3s
different 1
low
Advantage
SPL
New LHC injectors
1 The relative cost considers only the items that differ between both options ß
The LP-SPL is the best solution for the LHC and it offers a large upgrade potential for the future needs of physics.
Ref.: Comparison of Options for the Injectors of PS2, CERN-AB (PAF),

13. Site layout
SPS
PS2
ISOLDE
New LHC injectors PS
SPL
Linac4

14. LHC beams from PS2 (i) Nominal bunch train at PS2 extraction
h=180 (40 MHz) with bunch shortening to fit SPS 200 MHz.
168 buckets filled leaving a kicker gap of ~ 300 ns (50 GeV!)
Achieved by direct painting into PS2 40 MHz buckets using SPL chopping.
No sophisticated RF gymnastics required.
Beam parameters
Extraction energy: 50 GeV
Maximum bunch intensity: 4E11 / protons per LHC bunch (25 ns)
Bunch length rms: 1 ns (identical to PS)
Transverse emittances norm. rms: 3 mm (identical to PS)
Any other bunch train pattern down to 25 ns spacing
Straightforward with SPL 40 MHz chopping and 40 MHz system
Again without sophisticated RF gymnastics
Same brightness per bunch
New LHC injectors
Chamonix 2010

15. LHC beam from PS2 (ii) Example 25 ns beam from LP-SPL – PS2:
PS2 will provide “twice ultimate” LHC bunches with 25 ns spacing
Bunch train for SPS twice as long as from PS
Only 2 injections (instead of 4) from PS to fill SPS for LHC
PS2 cycle length 2.4 s instead of 3.6 s for PS
Reduces SPS LHC cycle length by 8.4 of 21.6 s (3x3.6 – 1x2.4)
Reduced LHC filling time
New LHC injectors
1 2 Booster
SPS injection plateau 3x3.6 s = 10.8 s
up to 4 consecutive injections PS
LP-SPL
SPS plateau ~2.4 s
2 injections
PS2
Chamonix 2010 15

16. Linac4: main characteristics
H- Þ charge exchange injection and painting in PSB
Ion species H−
Output Energy 160 MeV
Bunch Frequency MHz
Max. Rep. Rate 2 Hz
Max. Beam Pulse Length 1.2 ms
Max. Beam Duty Cycle %
Chopper Beam-on Factor 65 %
Chopping scheme:
222 transmitted /133 empty buckets
Source current 80 mA
RFQ output current 70 mA
Linac pulse current 40 mA
N. particles per pulse 1.0 × 1014
Transverse emittance 0.4 p mm mrad
Max. rep. rate for accelerating structures: 50 Hz
160/50 MeV Þ factor 2 in bg2)  same tune shift with twice the intensity.
Re-use of LEP RF components: klystrons, waveguides, circulators.
New LHC injectors
Chopping at low energy to reduce beam loss in PSB.
Structures and klystrons dimensioned for 50 Hz
Power supplies and electronics dimensioned for 2 Hz, 1.2 ms pulse.

17. Linac4: layout New LHC injectors
Linac4 is a normal-conducting H− linac at 160 MeV energy, made of 4 types of 352 MHz accelerating structures, matched to the increasing beam energy. A beam chopper at low energy allows modulating the linac beam pulse to minimise losses in the ring. A beam dump at linac end allows setting-up of the beam, will be displaced when connecting to the SPL.
The Linac4 project includes important modifications to the PSB injection region (higher injection energy, H- stripping).
160 MeV
New LHC injectors
100 MeV
50 MeV
Linac4 tunnel and surface equipment building

18. Linac4: Block diagram H- RFQ CHOPPER DTL CCDTL PIMS New LHC injectors
Linac4: 80 m, 18 klystrons
45keV
3MeV
3MeV
50MeV
94MeV
160MeV H-
RFQ
CHOPPER
DTL
CCDTL
PIMS
RF volume
source
(DESY)
45 kV
Extrac.
Radio
Frequency
Quadrupole
3 m
1 Klystron
550 kW
Chopper
& Bunchers
3.6 m
11 EMquad
3 cavities
Drift Tube
Linac
18.7 m
3 tanks
3 klystrons
4.7 MW
111 PMQs
Cell-Coupled Drift Tube
Linac
25 m
21 tanks
7 klystrons
7 MW
21 EMQuads
Pi-Mode Structure
22 m
12 tanks
8 klystrons
~12 MW
12 EMQuads
New LHC injectors
Ion current:
40 mA (avg.),
65 mA (peak)
RF accelerating structures: 4 types (RFQ, DTL, CCDTL, PIMS)
Frequency: MHz
Duty cycle: 0.1% phase 1 (Linac4), 3-4% phase 2 (SPL), (design: 10%)

19. Linac4: Beam dynamics New LHC injectors
The Linac4 design (machine architecture, beam optics) allows for high beam power operation  it incorporates modern linac technologies developed for high-power projects (SNS, JPARC, ESS,…) – and has contributed to the development of some of these technologies ! – providing an operational margin for PSB and LP-SPL.
Beam optics design to minimize beam loss.
Chopping at low energy to reduce longitudinal capture losses in the synchrotron.
Charge exchange injection.
Remaining losses concentrated on defined spots (collimation)
Measures used for keeping beam loss < 1W/m (for hands-on maintenance) at high beam power:
Smooth phase advance transitions.
Operating point far from resonances.
Longitudinal to transverse phase advance ratio (no emittance exchange).
Smooth variation of transverse and longitudinal phase advance.
Large apertures (> 7 rms beam size)
New LHC injectors

20. Linac4: Ion source (DESY design)
Improved version of the DESY RF volume source (antenna outside vacuum higher reliability).
45 kV extraction
2 MHz excitation at increased power
(100 kW for 80mA)
First tests in 2009.
Plasma chamber (Al2O3)
Antenna
New LHC injectors
Ignition + gas
Electron dump

21. Linac4: RFQ 352 MHz RFQ, 3 MeV energy New LHC injectors
Decision (summer 07) to build a taylor-designed CERN RFQ optimised for Linac4, using the experience gained in IPHI and TRASCO:
45 kV injection, 3 m length
beam dynamics and RF design jointly by CEA & CERN
based on the IPHI (CEA) & TRASCO (INFN) general mechanical design
detailed design, construction and brazing by CERN (EN/MME); availability during the first half of 2011
New LHC injectors
Main RFQ characteristics

22. Linac4: 3 MeV chopper line
Chopper = fast electrostatic deflector, removing the fraction of the linac beam pulse which would not be captured in the PSB bucket, in order to reduce beam losses.
Compact line design 3.7 m
Dynamic range 20 – 60 mA
Small e growth 4% long.,
8% trans.
Tolerant to alignment errors
RF bunchers
Choppers inside quads
Dump
New LHC injectors
Conical dump: dumping of chopped beam and collimation
Chopper structure: double meander strip line, 400mm length, metallized ceramic plate. 2 ns rise/fall time for bunch selectivity (352 MHz beam structure), ±500V between deflecting plates, provided by a special pulse generator (FID technology).

23. Linac4: Accelerating structures
Distributed focusing by PMQs (Permanent Magnet Quadrupoles) at low energy, where space is tight. Use of conventional EMQs above 50 MeV, for flexibility-reliability.
352.2 MHz RF frequency.
Maximum RF efficiency (shunt impedance ZT2 )  3 different RF structures (two 0-mode, one p-mode).
High accelerating gradients because of space constraints, but still in a safe range (<1.8 Kilpatrick peak surface field).
Safety margins: 25% on klystron max. power, 20% on theoretical Q-value.
New LHC injectors
E0 (MV/m)
Max. field (Kilpatrik)
E0 (MV/m, cost optimum for 10 yrs. op.)
DTL
3.2
1.6
4.3
CCDTL
1.7
3.1
PIMS
4.0
1.8
2.7

24. Linac4: Alvarez DTL (“Drift Tube Linac”)
Typical geometry
3 tanks ( cells). Overall length m.
f=352 MHz  l=85 cm
W=3…10 MeV  b= 0.08…0.145
for E0=3.3 MV/m, 28 cells
Cell length: bl = 6.8cm … 12.3cm
PMQ equipped
Drift Tube
New LHC injectors
RF resonator – TM0,1,0
3D view
Cell#1
6.8cm
Cell#28
12.3cm

25. Linac4: CCDTL (“Cell Coupled DTL”)
21 short (~1m) 2-drift tube tanks with quadrupoles in between, coupled in 3-tanks modules by coupling cells
New LHC injectors
CERN development since 2000.
Useful above ~40 MeV where the length of the focusing period allows to bring quadrupoles out of the drift tubes.
Advantages wrt to DTL: easy use of EMQs (possible adjustment for different currents), relaxed tolerances on machining and alignment, simpler and more economic construction.

26. Linac4: PIMS (“PI Mode Structure”)
12 tanks (~1m) made of 7 cells operating in Pi-mode, coupled via slots. Overall length 22 m.
New LHC injectors
Similar to LEP copper cavities.
Pi-mode required for efficient acceleration above 100 MeV.
Compact size at 352 MHz (no need for 704 MHz).

27. Linac4: planning 2013 + 2014 2015 New LHC injectors Milestones
End CE works: December 2010
Infrastructure:
2011
Installation:
Commissioning: till 2014
Modifications PSB: shut-down 2014/15
Operation:
Spring 2015
New LHC injectors
project duration: ~ 7 years

28. LP-SPL: Main characteristics
Required for flexibility and low loss in PS2
Ion species H−
Output Energy 4 GeV
Bunch Frequency MHz
Max. Rep. Rate 2 Hz
Max. Beam Pulse Length 0.9 ms
Max. Beam Duty Cycle 0.2 %
Nominal chopping factor 65 %
(Flexible chopping scheme)
Source current 40 mA
Linac pulse current 20 mA
Number of ions per pulse 1.1 × 1014
Transverse emittance 0.4 p mm mrad
Max. rep. rate for
accelerating structures
and klystrons: Hz
Required by space charge tune spread at the specified beam brightness
Re-use of LEP RF components in Front-end (Linac4)
Required for flexibility and low loss in PS2 (linac4 chopper with new driver)
New LHC injectors
Structures and klystrons dimensioned for 50 Hz
Power supplies and electronics dimensioned for 2 Hz, 2 ms pulse.

29. SC-linac (160 MeV ® 4 GeV) with ejection at intermediate energy
LP-SPL: Block diagram
SC-linac (160 MeV ® 4 GeV) with ejection at intermediate energy
0 m
0.16 GeV
110 m
0.73 GeV
186 m
1.4 GeV
427 m
4 GeV
Medium b cryomodule
High b cryomodules
High b cryomodules
Debunchers
From Linac4
Ejection
To PS2
New LHC injectors
9 x 6
b=0.65 cavities
5 x 8
b=1 cavities
14 x 8
b=1 cavities
TT6 to
ISOLDE
Length: ~430 m

30. Elliptical 5 cell bulk Niobium cavities
LP-SPL: RF technology
Elliptical 5 cell bulk Niobium cavities
(e.g.: b=0.47)
Auxiliary equipment
(e.g.: 1 MW RF coupler)
New LHC injectors
from G. Devanz (HIPPI meeting Nov. 2007)

31. LP-SPL: Cryomodules Medium b cryomodule High b cryomodule
Energy range: 160 MeV – 732 MeV
5 cell cavities
Geometrical b: 0.65
Maximum energy gain: 19.4 MeV/m
54 cavities (9 cryomodules)
Length of medium b section: ~ m
Energy gain (MeV/m)
High b cryomodule
Energy range: 732 MeV – 4 GeV
5 cell cavities
Geometrical b: 1
Maximum energy gain: 25 MeV/m
152 cavities (19 cryomodules)
Length of medium b section: ~286.2 m
Position (m)

32. LP-SPL parameters: RF frequency
704 MHz
1408 MHz
Length (5 GeV)
472 m
+12%
Ncavities
246
+15%
Nβ-families 2 3
ε-growth (x/y/z)
5.6/8.2/6.8
6.3/7.8/12.1
Longitudinal beam loss
none in simulations
lossy runs for realistic RF gradient/phase variations
BBU (HOM)
IBBU,704
1/(8..128)
Trapped modes
normal risk
2..4 higher risk
RF power density limit (RF distribution) ok
problematic
Klystrons
comfortable: MBK
difficult
Overall power consumption (RF+cryo, nom. SPL)
28 MW
up to -30%
Power converter
more bulky
saves tunnel space
Synergy with ESS
yes no
New LHC injectors

33. LP-SPL parameters: temperature
@ 704 MHz
T [K]
Eq. 4.5 K [kW]
Electrical power
[MW]
HP-SPL, 2% beam d.c. (4% cryo d.c.) 2
19.4
4.48
4.5
104
26.0
LP-SPL, 0.24% beam d.c. (0.32% cryo d.c.)
6.1
1.5 11
2.75
New LHC injectors
+ not clear that 25 MV/m can be achieved at 4.5 K!

34. LP-SPL parameters: Summary
Frequency/temperature:
704 MHz and 2 K are confirmed,
Cavity gradient:
25 MV/m “on average” (= with a high yield) is very challenging and may be costly (in terms of reprocessing),
20 MV/m seems more achievable but will have an impact on linac length (or energy).
New LHC injectors ß
High-power RF cavity tests of fully equipped cryo-modules are mandatory for realistic SPL layout estimates!!
Ref.: Assessment of the basic Parameters of the CERN SPL, CERN-AB BI-RF,

35. Max. pulse duration (ms) Max. current during pulse
LP-SPL: beam characteristics for ISOLDE
Beam energy (GeV)
Max. pulse duration (ms)
Max. current during pulse
(mA)
Repetition
period
(s)
Max.
protons
/pulse
(x1013)
beam power (kW) 4
0.9 20
0.6 11
120
1.4
~ 0.6
(3 out of 4 pulses) 31 1
0.35 28
~ 0.3
(7 out of 8 pulses)
6.1 29
~ 0.1
(23 out of 24 pulses) 94
PS2
Basic performance
New LHC injectors
ISOLDE
Need phase modulation
Needs higher power klystron modulators