Operational experience with Crab Cavities at KEKB

Y. FunakoshiKEK

KEKB B-FactorySuperconducting

cavities (HER)

e-

e+

ARES copper cavities (LER)

8 GeV e- 3.5 GeV e+ Linac e+ target

ARES copper cavities (HER)

Belle detector

KEKB B-Factory

TRISTAN tunnel

♦World-highest Peak Luminosity• 2.11 x 1034cm-2s-1

• Twice as high as design value

♦World-highest Integrated Luminosity• Total: 1041fb-1 as of June 30th 2010

Crab cavities1 for each ring

♦Crab crossing (f = 11 mrad) ♦Skew-sextupole magnets

The KEKB operation was terminated at the end of June 2010 for theupgrade toward SuperKEKB. Operation of SuperKEKB will start in Jan. 2015.

(mA)

Motivation of crab cavity at KEKBHead-on (crab)

First proposed by R. B. Palmer in 1988 for linear colliders.

Crab Crossing can boost the beam-beam parameter higher than 0.15 ! (K. Ohmi)

22mrad crossing angle

Strong-strong beam-beam simulation

nx =.508

Head-on } y ~0.15

After this simulation appeared, the development of crab cavities wasrevitalized.

Luminosity would be doubled with crab cavities!!!

Finally two crab cavities were installed in KEKB,

one for each ring in January 2007

HER (e-, 8 GeV) LER (e+, 3.5 GeV)

…..after 13 years’ R&D from 1994

Input couplerMagnetic Shield 　（ Jacket Type ）

80 K LN2 Radiation Shield

Coaxial Coupler

Stub Support

BellowsLiq. HeMonitor Port

RF Absorber

Frequency Tuningby Adjusting Distance

Crab Mode Reject Filter

RF Absorber

I.D. 240 I.D.100

Liq. He

Structure of KEKB Crab Cavity

Top View

Crab mode: TM110: By on beam axisLower mode: TM010: dumped through coaxial couplerSquashed cell shape to split TM110 modes

Skew-sextupoles

Beam lifetime problem

Evidence of crabbing motion (1): Streak camera

9

Evidence of crabbing (2): Beam-beam deflection

Bunch Current: 0.73/0.42 mA Bunch Current: 0.64/0.47 mA

HER current was lost from 15 to 13.5 mA during scan.

Crab ON ( Vc = 1.31/0.92 MV)Crab OFF

Horizontal effective size at IP reduces to 72% by the crab.

Σx_x’=00=167 +/- 3 m (ON)

x L/H=18/24 nm

Σx_x’=11=230 +/- 3 m (OFF) Ratio of

Slope:

T. Ieiri

( ): design of KEKB

Horizontal Tune close to half-integer

Efforts to improve performance with crab cavities

1. Beam lifetime problem2. Introduction of skew-sextupole

magnets

Beam lifetime problem• Observation

– When the HER bunch current increased, the LER beam lifetime decreased. The HER bunch current was limited by this phenomena.

– It turned out that the lifetime decrease was caused by dynamic beam-beam effects related to the physical aperture at the crab cavity.

• Cures– Change linear optics near crab cavities.– Enlarge IP horizontal beta functions which enabled us to

decrease horizontal beta functions at crab cavities.• bx*: 0.8 or 0.9m to 1.2 or 1.5m

Dynamic-b and dynamic emittanceby beam-beam (calculation)

LER

The focusing force of the beam-beaminteraction not only squeezes thebeam at the interaction point, butincreases the emittance drastically.

Optics deformation with dynamic beam-beam effect (LER)

Black: w/o beam-beamRed: with beam-beam

Beta’s with dynamic beam-

beam effect(LER)

before optics change bx*= 0.9m

After optics change bx*= 0.9m

crab

crab

with/without beam-beam effects

Tuning with skew-sextupole magnets

• Ohmi et al. showed that the linear chromaticity of x-y coupling parameters at IP could degrade the luminosity, if the residual values, which depend on machine errors, are large.

• To control the chromaticity, skew sextupole magnets were installed during winter shutdown 2009.

• The skew sextuples are very effective to increase the luminosity at KEKB.

• The gain of the luminosity by these magnets is ~15%.

Chromaticity of x-y coupling at IP

D. Zhou, K. Ohmi, Y. Seimiya,

Luminosity

Horizontal size

Vertical size

Definition of x-y coupling parameters (SAD notation)

Normal (decoupled)coordinate

Usual coordinate

Examples of scan of chromatic x-y coupling at IP

2005/June/06

Measurement on chromaticity of x-y coupling at IP (HER)

blue: without skew-sextuples

red: with skew-sextuples (after luminosity tuning)

Y. Ohnishi

dotted line: model opticswithout machine errors

Effectiveness of skew-sextupole magnets (crab on)

Effectiveness of skew-sextupole magnets (crab off)

Machine parameters (before/after crab)

Date Nov.15 2006before crab

Jun. 17 2009with crab

LER HER LER HERCurrent 1.65 1.33 1.64 1.19 ABunches 1389 1584

Bunch current 1.19 0.96 1.03 0.750 mAspacing 2.10 1.84 mA

emittance εx 18 24 18 24 nmβx* 59 56 120 120 cmβy* 6.5 5.9 5.9 5.9 mm

σx @IP 103 107 147 170 μmσy @IP 1.8 1.8 0.94 0.94 μm

νx 45.505 43.509 45.506 44.511νy 44.534 41.565 43.561 41.585νs -0.0246 -0.0226 -0.0246 -0.0209

beam-beam ξx 0.117 0.070 0.127 0.102beam-beam ξy 0.108 0.058 0.129 0.090

Luminosity 17.6 21.08 1033cm-2s-1

Effect of the crab cavities on the luminosity and the beam-beam

parameter

Specific luminosity (crab on/off)

Luminosity improvement by crab cavities is about 20%.Geometrical loss due to the crossing angle is about 11%.

Beam-beam parameter (crab on/off)

Crab on Crab off

RL 0.828 0.763

Ry(HER) 1.15 0.993

Calculation of beam-beam parameter

• Reduction factor for beam-beam parameter

– 2 sources of reduction• hourglass effect and finite crossing angle

Montague’s factor

Calculation of beam-beam parameter [cont’d]

• Reduction factor for luminosity

– Luminosity

– We use calculated values for x* and calculate

y* and y0 from observed luminosity.

Beam-beam parameters crab on/off

Possible causes of lower luminosity with crab than simulations

Possible causes of lower luminosity with crab than simulations

Short beam lifetime prevents us from approaching a better parameter set?

Synchro-betatron resonance lines near half-integer?

Vertical crab?Beam-beam + impedance?Some fast noise?Too many tuning parameters to find out an

optimum set of parameters?Crosstalk between beam-beam and lattice

nonlinearity?

35

Crab phase error

Span 200 kHzSideband peaks at 32kHz and 64kHz.

Span 10 kHz Span 500 HzSideband peaks at 32, 37, 46, 50, 100 Hz.

• Spectrum of pick up signal is consistent with phase detector data.• Phase fluctuation faster than 1 kHz is less than ± 0.01°, and slow fluctuation from ten to

several hundreds of hertz is about ± 0.1°. • They are much less than the allowed phase error obtained from the beam-beam simulations

for the crabbing beams in KEKB.

According to b-b simulation by Ohmi-san, allowed phase error for N-turn correlation is 0.1×√N (degree).

Spectrum around the crabbing mode measured at a pick up port of the LER crab cavity. Beam current was between 450 and 600 mA.

LER crab phase

HER crab phase

± 1 deg

Phase detector signal. Beam current was 385mA (HER) and 600 mA (LER).

K. Akai

2005/June/06

Too many tuning parameters?

• Many tuning parameters used to be tuned one-by-one by scans with only observing the luminosity and the beam sizes.

• There was a possibility that we could not reach an optimum set of parameters with this method.

• A method of parameter search based on downhill simplex method was developed in 2007, which can search 12 parameters simultaneously. However, with this method the achievable luminosity was not improved.

IP coupling and dispersion tuning knob

39

Side effect of a large tuning knob

iSize Bump 3 mm to check response of size monitor

R2=37.29 units

R2=-21.36 units

10 units

Original settingR2=7.99

Knob zero

The vertical beam size is enlarged due to the side effect of a large tuning know such as 30 units of R2 knob. If a large x-y coupling remains at IP, the luminosity degradation may not be recovered by the tuning knobs.

LER R2

Vertical beam size [m

]

Strong-weak beam-beam simulation with lattice nonlinearity

K. Ohmi

The strong-weak beam-beam simulation was done using SAD with full lattice (head-on).Sextupole mis-alignments were introduced so that the emittance ratio (V/H) of ~ 1% was created.

Strong-strong (head-on)

Strong-strong (22mrad crossing angle)

experiment

With IP x-y coupling and vertical dispersionCorrect IP coupling and dispersion

Strong-weak beam-beam simulation with lattice nonlinearity

K. Ohmi

The crab cavities were implemented in the strong-weak beam-beam simulation (crab crossing).We found a remarkable difference between the head-on and crab crossing.

Head-on

Crab crossing

Crab crossing(vertical crab)

Summary• The crab cavities were installed in Feb. 2007 at KEKB and

worked very well until the end of the KEKB operation.• The highest luminosity with the crab cavities is about

23% higher than that before crab (prediction by b-b simulation: ~100% increase).

• The tuning with skew-sextupole magnets were effective to increase the luminosity w/ crab (~15% gain).

• We found that the skew-sextupole magnets are also effective to increase the luminosity when the crab cavities were switched off.

Summary [cont’d]• The luminosity difference between the crab on and off

was about 20% and is larger than the geometrical loss of the luminosity due to the crossing angle (~11%).

• The achieved vertical beam-beam parameter was ~ 0.09. This value is rather high but is lower than the predicted value by the beam-beam simulation (~0.15).

• Possible causes for this discrepancy:– Large machine errors which can not compensated by the usual

tuning knobs?– Lattice nonlinearity + beam-beam?– We are still studying this issue.

Spare slides

45

Phase stability

Span 200 kHzSideband peaks at 32kHz and 64kHz.

Span 10 kHz Span 500 HzSideband peaks at 32, 37, 46, 50, 100 Hz.

• Spectrum of pick up signal is consistent with phase detector data.• Phase fluctuation faster than 1 kHz is less than ± 0.01°, and slow fluctuation from ten to

several hundreds of hertz is about ± 0.1°. • They are much less than the allowed phase error obtained from the beam-beam simulations

for the crabbing beams in KEKB.

According to b-b simulation by Ohmi-san, allowed phase error for N-turn correlation is 0.1×√N (degree).

Spectrum around the crabbing mode measured at a pick up port of the LER crab cavity. Beam current was between 450 and 600 mA.

LER crab phase

HER crab phase

± 1 deg

Phase detector signal. Beam current was 385mA (HER) and 600 mA (LER).

K. Akai

46

LER Nikko New Optics

Before 2/21(maintenance)βx max 199 m(νx,νy)=(45.505, 43.59)

After 2/21(maintenance)βx max 91 m(νx,νy)=(44.505, 43.59)Large β distortion in wiggler section

crab crabβx @ crab ~85 m as is before

βx

βy

H. Koiso

QC2L

QC2R

QW4NP.1

QW4NP.2

QW4OP.2

QW4OP.1

MD06H1

MD06H2

MD06H3

MD06H4

MD03H1

MD03H2

MD03H3

MD03H4

Beam size calculation with dynamic beam-beam effects

@crab (aperture < 5 x)

Synchro-betatron resonance• The horizontal tune is set nearby

the half integer resonance and its synchrotron side bands.

• On the resonances, some harmful effects are observed.

– Single-beam beam size blowup– Tow-beam beam size blowup– Beam lifetime reduction (or beam

loss)• The resonance is stronger in HER

where no local chromaticity correction is installed.

• Strength of the resonance is strongly dependent on a choice of sextupole setting. The luminosity also changes by changing the sextupole setting.

• Even in the off resonance tunes, it affects the luminosity due to the tune footprint by the beam-beam?

2nx + ns = integer

2nx + 2ns = integer

LERHER

Negative- Optics• Motivation

– To weaken the synchro-betatron resonance particularly in HER

– To shorten the bunch length

• Results– We have succeeded to weaken the synchro-betatron resonance line in HER.

We could operate the machine with nx below the resonance line.– We have successfully shorten the bunch length of both beam.

• ~6mm -> ~4.5mm– However, we found unexpectedly large synchrotron oscillation in LER

(microwave instability) and gave up the trial of the negative- optics.

2νx + νs = integer2νx + 2νs = integer

2νx - νs = integer2νx - 2νs = integer

νx: .5112, .5224with given νs ~ -.0224

Comparison of beam lossKEKB (design) KEKB (achieved) SuperKEKB

LER HER LER HER LER HER

Radiative Bhabha 21.3h 9.0h 6.6h 4.5h 28min. 20min.

Beam-gas 45ha) 45ha) 24.5min.b) 46min. b)

Touschek 10h - 10min. 10min.Total 5.9h 7.4h ~133min. ~200min. 6min. 6min.Beam current 2.6A 1.1A 1.6A 1.1A 3.6A 2.6A

Loss Rate 0.12mA/s 0.04mA/s 0.23mA/s 0.11mA/s 10mA/s 7.2mA/s

a) Bremsstrahlungb) Coulomb scattering, sensitive to mask setting

Beam loss accompanied with the beam injection should be added.