deflecting cavities for light sources ali nassiri advanced photon source argonne national laboratory...

Deflecting Cavities for Light Sources

Ali Nassiri

Advanced Photon Source

Argonne National Laboratory

ICFA Beam Dynamics Min-Workshop on Deflecting/Crabbing Cavity Applications in Accelerators

April 23 – 25, 2008, SINAP, Shanghai, China

2A. Nassiri Crab Cavities for Light Sources April 23, 2008 - Shanghai

Outline

Scientific Case Scheme Expected Performance and Tolerances Transient Schemes Technology Options Conclusions


Scientific Case


Time Scales: Physical, Chemical, and Biological Changes

10-3 10-8 10-1110-9 10-10 10-12 10-13 10-14 10-15

Nano

Pico FemtoMilli

Radicals

Spectr.

and

Reactions

IVR and Reaction Products

Transition States and Reaction Intermediates

Atomic Resolution

Single Molecule Motion

Femto-chemistry

Radiative Decay Rotational Motion Vibrational Motion Fundamental

Internal Conversion & Intersystem Crossing Vibrational Relaxation

Collisions in Liquids

Physical

Predissociation Reactions

Harpoon Reactions

Norrish Reactions

Dissociation Reactions

Chemical

Proton Transfer

Abstraction, Exchange & Elimination

Diels-Adler

Charge Recomb.

Protein Motion Photosynthesis Biological

106

Period of Moon

PS Source

X-ray Techniques

Storage Ring Sources X-ray FELs

Sec.


A New Era of Ultrafast X-ray Sources

LCLS: 120Hz

SPPS 10Hz

Photo courtesy: D. Reis, UM

APS Concept


Science Enabled by ps Sources

The field of time domain scientific experiments using hard x-rays from synchrotron radiation sources is gaining momentum.

The time range covered by ongoing and future experiments is from sub-picoseconds to thousands of seconds, which is 16 to 17 decades of spread.

The scientific disciplines that will benefit from these studies include:

– Atomic and molecular physics

– Biology and chemical science• Photochemistry in solution

– Condensed matter physics• Ultrafast solid state phase transition

– Engineering and environmental science

– Material and nuclear science


Existing and Future Sources Table-top Plasma Sources

– Short pulse 300 fs - 10 ps– Divergent radiation - low flux– Low rep-rate (10 Hz -1kHz)– Not tunable (target dependent)

Storage Rings– ~100-ps duration pulse– Spontaneous x-ray radiation– High average brightness at high repetition rate

Laser Slicing (ALS, SLS, BESSY)– Short pulse 100-300 fs– Rep-rate kHz– Low flux 105 ph/s @ 0.1% BW– Not effective at high-energy sources

Linacs (LCLS/XFEL)– Short pulse 100 fs– Fully coherent– Extremely high brilliance – Low rep-rate (100 Hz)– Limited tunability


Time-resolved Experiments Today

Pump-probe

– Pump : laser pulses (100 fs – 10 ns), s flash lamps

– Probe: 100-ps x-ray or longer pulse train Data collection

– Slow variable: crystal angular setting

– Fast variable: pump-probe delay time, t• For each crystal orientation collect:

– No laser, t1, t2, t3….Laue frames

Repetition rate depends on:

– Sample (lifetime of intermediates)

– Heat dissipation (laser-induced heating)• 1 – 3 Hz typical

40 – 60 images per data set 2-30 angular increment with undulator sources (few % bandwidth)

X-ray pulse

ns laser pulse


Time-resolved Macromolecular Crystallography

Pulse duration: structural changes to be probed sub-ps – min

– 100 ps available at synchrotron sources

– Longer pulse trains suitable for slow reactions

– Sub-100ps desirable to probe very fast structural changes:• Short-lived intermediates• Fast protein relaxation• Rapid ligand migration

Desired X-ray flux greater than 1010 photons/pulse for single image acquisition Single-pulse acquisition will allow study of fast, irreversible processes X-ray energy: few% bandwidth at 12-15 keV

– Softer X-rays increase radiation damage

– Harder X-rays diffract less strongly and are detected less efficiently


Scheme


Deflecting cavity introduces angle-time correlation into the electron bunch, “crabbing” the beam. Bx kicks head and tail of the bunch in opposite directions in the vertical plane.

Electrons oscillate along the orbit. Bunch evolution through the lattice results in electrons and photons correlated with

vertical momentum along the bunch length. Second cavity at n phase cancels “kick”; rest of the storage ring unaffected.

A. Zholents, P. Heimann, M. Zolotorev, J. Byrd, NIM A 425, 385, (1999).

Crabbing Scheme


Expected Performance

and Tolerances


Estimating X-ray Pulse Duration

X-ray pulse length can be estimated assuming Gaussian distributions1

Emittance growth matters because it increases the minimum achievable pulse duration.

Electron beamenergy

Deflectingrf voltage &frequency

Unchirped e-beamdivergence (typ.2-3 rad)

Divergence dueto undulator (typ.~5 rad)

For 4 MV, 2800 MHz (h=8) deflecting system, get ~0.6 ps

1M. Borland, Phys. Rev. ST Accel Beams 8, 074001 (2005).

22rad,ye,yrf

Vh

Eid

a

xray,t


Emittance Growth1,2

In the idealized concept, a second set of cavities exactly cancels the effect of the first set

– In reality, it doesn't work exactly and we have emittance growth Sources of growth in an ideal machine:

– Time-of-flight dispersion between cavities due to beam energy spread– Uncorrected chromaticity, if present (normally it is)– Coupling of vertical motion into horizontal plane by sextupoles– Quantum randomization of particle energy over many turns

Additional sources of growth in a real machine– Errors in magnet strengths between the cavities– Roll of magnetic elements about beam axis– Roll of cavities about beam axis– Orbit error in sextupoles– Errors in rf phase and voltage

Emittance growth is not just a worry for brightness.– It also limits how short an x-ray pulse can be achieved

1M. Borland, private communication, 2004.2M. Borland, Phys. Rev. ST Accel Beams 8, 074001 (2005).


Reducing Emittance Growth1,2,3,4

There are several methods of reducing emittance growth:– Don't power cavities past point of diminishing returns

– Manipulate sextupoles between cavities• Turning them off is not the best approach• Minimize emittance directly using particle tracking simulation• Tune sextupoles for zero chromaticity between cavities

– Choose vertical oscillation frequency (“tune”) to facilitate multi-turn cancellation of effects

– Increase separation of horizontal and vertical tunes

1M. Borland, private communication,2004.2M. Borland, Phys. Rev. ST Accel Beams 8, 074001 (2005).3V. Sajaev, private communication, 2005.4M. Borland and V. Sajaev, Proc. PAC 2005, 3886-3888, (2005), www.jacow.org.


Comparison of Emittance Growth for Pulsed, CW

Starting vertical emittance is 13 pm (0.5% coupling) 10-k turn tracking results with parallel elegant1

“1 kHz” shows hybrid bunch emittance only “CW” is for 24-bunch mode, all bunches are affected

1Y. Wang, M. Borland, Proc. PAC07, 3444-3446,www.jacow.org, (2007).


Comparison of Emittance Growth

Starting vertical emittance is 20 pm (0.8% coupling)1

10-k turn tracking results with parallel version of elegant2

Hybrid-mode results are for intense bunch only

1L. Emery, private communication.2Y. Wang, M. Borland, Proc. PAC07, 3444-3446, (2006),.www.jacow.org.


X-ray Slicing Results (2.4-m U33, 10keV)

Two slits at 26.5 m

– Vertical slit is varied from ±100 mm to ±0.010 mm

– Fixed horizontal slit of ±0.25 mm (E. Dufrense)


Results for Constant 1% Transmission

24-bunch mode has a slight edge due to smaller emittance

Effect of emittance increase is clear in comparison of 2 MV and 4 MV results

No compelling reason to go above 4 MV


Details of X-ray Slicing Results for Hybrid Mode1

Slits: H=0.5 mm, V=0.2 mm1M. Borland, private communication.2007.

2nd harmonicradiation

back-chirp back-chirp

Back-chirppulses haveabout 2.5%of the intensityof the centralpulse.


Details of X-ray Slicing Results for 24 Bunch Mode

Slits: H=0.5 mm, V=0.2 mm

2nd harmonicradiation

Back-chirppulses haveabout 0.02%of the intensityof the centralpulse and arenot seen here.


Summary of Tolerances1

Quantity Driving Requirement 24-bunch Hybrid

Common-modevoltage

Keep intensity and bunch length variationunder 1%

±1% ±1%

Differential voltage Keep emittance variation under 10% ofnominal

±0.44% ±0.43%

Common-modephaserelative to buncharrival

Constrain intensity variation to 1% ±10 deg ±10 deg

Differential phase Keep centroid motion under 10% ofbeam size

±0.07 deg ±0.09 deg

Rotational alignment Emittance control ~1 mrad ~1 mrad

Tolerance on timing signal from crab cavity to users: ±0.9 deg

1M. Borland, “Long-Term Tracking, X-ray Predictions, and Tolerances,” SPX Cavity Review, 8/23/07.


Transient Short Pulse

via

Beam Manipulation


Transient Short Pulse Generation via Beam Manipulation

Studied various transient alternate short pulse schemes (i.e., pulsed) that manipulate beam and rely on radiation damping to restore emittance, bunch length. Potentially useful for beam and beamline diagnostics development, possibly experiments (during machine intervention/studies).

Synchrobetatron coupling W. Guo et al., Phys. Rev. ST Accel. Beams 10, 020701 (2007)

– Chirp is produced via a magnet kick: A sin(x + (z)), rather than deflecting cavity: A(z) sin(x + )

– Beam tilt (y-t) in ID, rather than (y’-t) as with deflecting cavity Rf phase modulation

G. Decker et al., Phys. Rev. ST Accel. Beams 9, 120702 (2006)

– Bunch length actually compressed – no tilt

– Bunch shape oscillation at 2x synchrotron frequency Quarter-integer betatron resonance

W. Guo, private communication, M. Borland, private communication, 2005

– Same chirp as deflecting cavity, except build-up over several turns using resonant excitation at frequency: 8frf + 0.25frev

– Drive at much lower power: ~1 MV


Comparison: Transient short pulse generation

Pulse compression

achieved

Repetition rate limit

Pro Con

Synchro-betatron

3x (avg)

6.5x (w/o jitter)

~40 Hz (1 kHz

possible with fast kickers)

Available hardware

Bunch current limited to few mA; sensitive to tune jitter &

wakefields

Rf phase modulation

2x ~40 Hz Available hardware,

should allow ~50

mA

Limited pulse compression

Quarter-integer

resonance

TBD

(simul. 50x)

~20 Hz Same as RT

deflecting cavity

Needs hardware

Slide courtesy K. Harkay; Figs. courtesy B. Yang, G. Decker, M. Borland


Technology Options


APS operating modes, 100 mA

1.59 s

1x1 (16 mA)

8x7 (86 mA)

1.3 s

Deflecting cavity rf voltage


Cavity Design Evolution – A“ warm” system

June 05*

Nov 07* V. Dolgashev, SLAC


APS Short-Pulse X-Ray Normal-Conducting Cavity Design*

Frequency 2.815 GHz

Deflecting Voltage 2 MV

Peak Power 2.8 MW

Working Mode Qo 12000

Rt / Q 117

Iris Radius 22 mm

Phase Advance π

Structure Length w/o beam pipes

11.17 cm

Duty Factor 0.147%

Pulse Rate 1.0 kHz

Kick / (Power) 1/2 1.19 MV/MW1/2

Beam Current 100 mA

Input coupler

Rectangular damping waveguide

Water header

Normal-conducting 3-cell cavity with damping waveguide and dual input couplers

Damping material is attached to each damping waveguide flange

Tuning pins Ridged

damping waveguide

*In collaboration with V. Dolgashev (SLAC)


Coupler

Coupler

DamperFlange

DamperFlange

DamperFlange

Slide courtesy: L. Morrison


Two-Sector Layout

Sector 6, section 6 Sector 7, Girders 1 through 5 Sector 7, section 6

Upstream end ID chamber

Downstream end ID chamber

Gate valve

Gate valve


APS 2.8 GHz Superconducting Single-Cell Deflecting Cavity1

Frequency (GHz) 2.815

Deflecting Voltage 4 MV * 2

Qo (2K) 3.8 * 109

G 235

RT / Q ( m) 37.2

Beam Radius 2.5 cm

No. Cavities 12 * 2

Operation CW

Beam Current (mA) 100

Esp/Vdefl (1/m) 83.5

Bsp/Vdefl (mT/MV) 244.1

HOM dampers

LOM/ HOM damper

Input Coupler / HOM damper

Compact single-cell cavity / damper assembly

Deflecting cavity

Waveguide damper replaces KEK coaxial coupler

1 In collaboration with JLab and LBL


Deflecting Cavity Layout - Schematic

B B T1T1 B

2920 mm

VV

8000 mm

4592.7 mm

T2 ID VC

P P

T2

190 mm190 mm

107.3 mmSpace available for cryo-modules + bellows

Created:1/16/08Rev: 00

4100 mm

12 cavities + cryomodule

Gate valve Bellows

Thermal intercept

400 mmBellows

2

3


Conclusions Short X-ray pulse generation at the synchrotron light sources will open up new

frontiers in time domain science using X-ray techniques to study structural dynamics included but not limited to:

– Condensed Matter, Chemical and Biological, Gas Phase Dynamics Both normal-conducting room-temperature and SRF options are feasible, with

the advantages of SRF being:

– Not limited to SR bunch train fill patterns

– Higher flux and higher repetition rates up to CW Tracking studies have been performed for pulsed and CW system For CW system

– Presented studies cover only single-particle dynamics Emittance growth for 4 MV is acceptable

– Present results start from base of 20 pm, which seems to be minimum presently achievable

– We stay under 50 pm (2% coupling)

– Little benefit from going to higher voltages We can achieve below 2 ps FWHM with ~1% of nominal intensity


Acknowledgements

V. Dolgashev (SLAC)

R. Rimmer (JLab)

H. Wang (JLab)

P. Kneisel (JLab)

L. Turlington (JLab)

Derun Li (LBL)

J. Shi ( Tsinghua University- Beijing), PhD Candidate

B. Adams, A. Arms, N. Arnold, T. Berenc, M. Borland, T. B. Brajuskovic, D.

Bromberek, J.Carwardine, Y-C. Chae, L.X. Chen, A. Cours, J.Collins, G. Decker, P.

Den Hartog, N. Di Monti, D. Dufresne, L. Emery,M. Givens, A. Grelick, K. Harkay,

D. Horan, Y. Jaski, E. Landahl, F. Lenkszus, R. Lill, L. Morrison, A. Nassiri, E.

Norum, D. Reis, V. Sajaev, G. Srajer, T. Smith, X. Sun, D. Tiede, D. Walko, G.

Waldschmidt, J. Wang, B. Yang, L. Young

Collaborators

deflecting cavities for light sources ali nassiri advanced photon source argonne national laboratory...

Documents

ps sources

nassiri crab cavities

synchrotron radiation

shanghai science

plasma sources short

future sources table

shanghai scientific

shanghai time scales