Kicker and RF systems for Damping Rings CLIC Technical Committee Kicker and RF systems for Damping Rings Fanouria ANTONIOU, Mike BARNES, Tony FOWLER, Alexej GRUDIEV, Yannis PAPAPHILIPPOU February3rd, 2009

Design optimisation for CDR (2010) M. Korostelev, PhD thesis, 2006 Outline CLIC damping rings (DR) design goals and challenges Design parameters’ evolution Lattice choice, optics revision and magnet design Wiggler design and power absorption Non-linear dynamics Low emittance tuning e-cloud and other collective effects (IBS) Diagnostics CLIC DR activities Summary CLIC parameter note 2005 CLIC parameter note 2008 Design optimisation for CDR (2010) M. Korostelev, PhD thesis, 2006 03/02/09 CTC, YP

CLIC damping ring layout M. Korostelev, PhD thesis, EPFL 2006

DR bunch structure and timing CLIC DR timing parameters Old (2005) New (2007) Repetition rate [Hz] 150 50 Number of bunches 110 312 Number of trains 4 1 Bunch spacing [ns] 0.533 0.500 Revolution Time [μs] 1.2 >1.2 Machine pulse [ms] 6.67 20 H/V/L damping times [ms] 2.8/2.8/1.4 1.5/1.5/0.76 Reduction of repetition rate from 150 to 50Hz leaves enough time for the emittances to reach their equilibrium Bunch spacing increased almost to same level as for the interleaved train scheme Interleaved train scheme abandoned Extraction kicker rise time relaxed Injection and extraction process simplified 312 bunches with 0.5ns spacing, fill 13% of the rings 4

DR injection/extraction optics Injection and extraction system placed at the same area, upstream of the super-conducting wigglers Kickers placed at maximum beta functions for minimum deflection angle Septa and kickers share the same cell Additional cells to be added in order to increase available space for elements and protection system Phase advance between injection (extraction) septa and kickers of around π/2 M. Korostelev, PhD thesis, EPFL 2006 03/02/09 CTC, YP

Septa parameters DR septa parameters SEP-1 SEP-2 Effective length [m] 0.4 0.5 Bending angle [mrad] 13 42 Field integral [T.m] 0.11 0.34 Blade thickness [mm] 5 Same parameters for inj/ext elements due to optics mirror symmetry Septum parameters are scaled from NLC damping rings Two DC modules with blade thickness of 5 and 13mm Effective length can be increased to 2m if additional cells are added Larger septa blade thicknesses and smaller peak field

Septa and kicker parameters Rise and fall time significantly increased Effective length can be increased to 2m Smaller peak field Kicker stability refers to field uniformity and pulse-to-pulse stability Tolerance of 0.1σ beam centroid jitter DR kicker parameters Old New Rise and fall time [ns] 25 1000 Flat top [ns] 142 ~160 Repetition rate [Hz] 150 50 Effective length [m] 0.4 0.4-2 Aperture [mm] - 20 Kick [mrad] 2.45 3 Field [Gauss] 500 <610 Kicker stability 1.4x10-3 @ inj 1.5x10-4 @ ext Second kicker in transfer line @ phase advance of π for jitter compensation (as in NLC) Third kicker, delay line and RF deflector are removed Kickers’ impedance issues should be adressed during the design (budget of a few MΩ/m in transverse and a few Ω in longitudinal)

DR injection/extraction lattice Most critical the e+ PDR Injected e+ emittance ~ 2 orders of magnitude larger than for e-, i.e. aperture limited if injected directly into DR PDR for e- beam necessary as well A “zero current” linac e- beam (no IBS) would need ~ 17ms to reach equilibrium in DR, (very close to repetition time of 20ms) PDR main challenges Large input momentum spread necessitates large longitudinal acceptance for good injection efficiency Polarised positron stacking time long compared to repetition rate (need fast damping and/or staggered trains) PDR Extracted Parameters CLIC NLC Energy [GeV] 2.424 1.98 Bunch population [109] 4.1-4.4 7.5 Bunch length [mm] 10 5.1 Energy Spread [%] 0.5 0.09 Hor. Norm. emittance [nm] 63000 46000 Ver. Norm. emittance [nm] 1500 4600 Injected Parameters e- e+ Bunch population [109] 4.4 6.4 Bunch length [mm] 1 5 Energy Spread [%] 0.1 2.7 Hor.,Ver Norm. emittance [nm] 100 x 103 9.3 x 106 03/02/09 CTC, YP

Arc and wiggler cell TME arc cell chosen for compactness and efficient emittance minimisation over Multiple Bend Structures used in light sources Large phase advance necessary to achieve optimum equilibrium emittance Very low dispersion Strong sextupoles needed to correct chromaticity Impact in dynamic aperture Very limited space Extremely high quadrupole and sextupole strengths FODO wiggler cell with phase advances close to 90o giving Average β’s of ~ 4m and reasonable chromaticity Quad strength adjusted to cancel wiggler induced tune-shift Limited space for absorbers 03/02/09 CTC, YP

New arc cells optics Alternative cell based on SUPERB lattice S. Sinyatkin, et al., BINP P. Raimondi (INFN-LNF) Alternative cell based on SUPERB lattice Using 2 dipoles per cell with a focusing quadrupole in the middle Good optics properties To be evaluated for performance when IBS is included New arc cell design Increasing space between magnets, reducing magnet strengths to realistic levels Reducing chromaticity, increasing DA Even if equilibrium emittance is increased (0 current), IBS dominated emittance stays constant! Dipoles have quadrupole gradient (as in ATF!). 03/02/09 CTC, YP

CLIC DR RF system 1) Main issues: Frequency: 2 GHz Highest peak and average power Very strong beam loading transient effects (beam power of ~5 MW during 156 ns, no beam power during the other 1060 ns) Small stored energy at 2 GHz High energy loss per turn at relatively low voltage results in big sin φs = 0.95 (see also LEP) Wake-fields Pulsed heating related problem (fatigue, …) 2) Recommendations: Reduce energy loss per turn and/or increase RF voltage Consider 1GHz frequency (RF system becomes conventional, RF power reduced, but delay loop for recombination is necessary and emittance budget is tight) A. Grudiev (CERN) 03/02/09 CTC, YP

RF for CLIC DR A very first look 16.10.2008 Alexej Grudiev

Outline CLIC DR energy acceptance Scaling NLC DR RF system to CLIC * Traveling versus Standing wave system RF source issue Conclusions * Parameters of NLC DR RF are taken from “Collective effects in the NLC damping ring design” T. Raubenheimer ,et. al., PAC95

Energy acceptance ±0.8 ±1.7 ±2.6 CLIC DR parameters* Circumference: C [m] 365.2 Energy : E [GeV] 2.42 Momentum compaction: αp 0.8x10-4 Energy loss per turn: U0 [MeV] 3.9 Maximum RF voltage: Vrf [MV] 4.115 RF frequency: frf [GHz] 2.0 Energy acceptance versus rf voltage Vrf [MV] ΔE/E [%] 4.115 ±0.8 4.5 ±1.7 5 ±2.6 * From Yannis CLIC pars WG, 2/10/07

Scaling of NLC DR RF cavity NLC DR RF cavity parameters Frequency: f[GHz] 0.714 Shunt impedance: R [MΩ] 3 Unloaded Q-factor: Q0 25500 Aperture radius: r [mm] 31 Max. Gap voltage: Vg [kV] 500 Scaled RF cavity parameters Frequency: f[GHz] 2 Shunt impedance: R [MΩ] 1.8 Unloaded Q-factor: Q0 15400 Aperture radius: r [mm] 11 Max. Gap voltage: Vg [kV] 180 Five 1 MW CW klystrons feeding 5 SW 5-cells accelerating structures would do it. ηrf-to-beam < 30% Total length ≥ 2m Calculated RF cavity parameters Number of cavities: N Vrf/Vg ~ 23 Total wall losses: POhm [MW] Vrf2/2NR ~ 0.2 Peak beam current: Ib [A] Qb*f ~ 1.3 Peak beam power: Pb [MW] U0*Ib ~ 5 Loaded Q-factor: Qext Q0*POhm /Pb~ 620 Filling time: Tf [ns] Qext/f ~ 310

TW versus SW acc. structure Several fully beam loaded travelling wave accelerating structures with shorter filling time ~20ns could increase efficiency significantly but only at fixed (nominal) current and voltage. SW structure would require tunable coupler in order to change the loaded Q-factor and maintain efficiency when changing beam current In summary, both systems are possible

RF power source In case of a klystron, It must be pulsed with DR revolution frequency of ~1 MHz repetition rate in order to maintain efficiency OR a better option would be to do pulse compression on each turn, this will also reduce peak power requirements on klystrons at the expenses of pulse compression efficiency (~70%) though And it must have certain bandwidth in order to be able to shorten the filling time to increase efficiency (more of an issue in TWS). Tf~50ns => df~20MHz => df/f~1% IOTs are better choice from the point of view efficiency and bandwidth (-> efficiency). But they have less power per tube and lower gain (two stages will be required). An R&D item at 2 GHz. Solid state rf power amplifier showed 50% efficiency from the plug at 500 MHz (SOLEIL). BUT Efficiency and power at 2 GHz -?

Vacuum Electron Device Limitations for High-Power RF Sources Considerable part of former klystron domain claimed by IOTs Why? 0.1 1 10 100 1000 Frequency (GHz) Heinz Bohlen,Thomas Grant, CPI

Wakefields of the rf system Loss/kick factors of NLC DR rf cavity for bunch length of 3.3 mm and aperture radius of 31 mm from the reference: Total loss factor: kl = 1.7 V/pC Transverse kick factor: kt = 39.4 V/pC/m Scaling to CLIC DR rf cavity for bunch length of 3.3 mm and aperture radius of 11 mm: Total loss factor: ~1/d: kl ~ 4.8 V/pC per cell Transverse kick factor: ~1/d3: kt ~ 873 V/pC/m per cell Number of cavities (cells) is also higher for CLIC: ~23 In summary: it is ~10 times higher for longitudinal wake and ~100 times higher for transverse wake for the whole rf system One good thing is that at 2 GHz HOM damping is more compact and could be done more efficient. Q-factor of HOM could of the order of few tens or so.

CLIC DR RF system issues Frequency: 2 GHz Highest peak power High average power Very strong beam loading transient effects: Peak beam power of ~5 MW during 156 ns No beam power during the other 1060 ns Small stored energy at 2 GHz High energy loss per turn at relatively low voltage results in big sin φs = 0.95 (any examples of operation ?) Wakefields Pulsed heating related problem (fatigue, …)

Recommendations Reduce energy loss per turn This will help anyway Consider frequency reduction down to 1 GHz It makes rf system a conventional high power rf system (other DRs, B-factories, etc.) One can take advantage of a superconducting RF system Reduce beam peak power by 2, so the SR peak power, less pulsed heating, less rf peak power, etc. It makes life easier for positron capture BUT Recombination of bunches is necessary at extraction or in a separate delay loop. Potential impact on the beam emittance must be addressed.

Summary Detailed design of the CLIC damping rings, delivering target emittance with the help of super-conducting wigglers Prototype to be built and tested at ANKA synchrotron Radiation absorption protection Collective effects evaluation including electron cloud and fast ion instability Lattice revision with respect to space and magnet parameters Parameter scan for conservative beam emittances for 500GeV collider Active collaboration with ILC, test facilities, B-factories, synchrotron light sources and other interested institutes Critical items for the performance of the damping rings Super-conducting wigglers E-cloud and fast ion instability Low emittance tuning Intra-beam scattering 03/02/09 CTC, YP