Rf Deflecting Cavity Design for Generating Ultra-short pulses at APS

1
Rf Deflecting Cavity Design for Generating
Ultra-short pulses at APS

• Geoff Waldschmidt
• Radio Frequency Group
• Accelerator Systems Division
• Advanced Photon Source
• LHC IR Workshop, October 3, 2005

2
Feasibility study group

• Beam dynamics
• M. Borland
• Y.-C. Chae
• L. Emery
• W. Guo
• K.-J. Kim
• S. Milton
• V. Sajaev
• B. Yang
• A. Zholents, LBNL

RF K. Harkay D. Horan R. Kustom A. Nassiri G.
Pile G. Waldschmidt M. White
Undulator radiation x-ray optics L.
Assoufid R. Dejus D. Mills S. Shastri
All affiliated with APS except where noted
3
Outline

• Science case for storage ring ps pulses
• Summary of beam and optics simulations and beam
  studies
• Deflecting cavity designs
• Room temperature vs. superconducting
• Multi-cell and single-cell superconducting cavity
  designs
• RD plan
• Summary

4
Science Drivers for ps X-rays

• APS Strategic Planning Workshop (Aug 2004) Time
  Domain Science Using X-Ray Techniques
• by far, the most exciting element of the
  workshop was exploring the possibility of shorter
  timescales at the APS, i.e., the generation of 1
  ps x-ray pulses whilst retaining high-flux. This
  important time domain from 1 ps to 100 ps will
  provide a unique bridge for hard x-ray science
  between capabilities at current storage rings and
  future x-ray FELs.

• Atomic and molecular dynamics, coherent/collective
  processes
• Atomic and molecular physics
• Condensed matter physics
• Biophysics/macromolecular crystallography
• Chemistry

APS Users Meeting Workshop on Generation and
Use of Short X-ray Pulses at APS (May 2005)
Slide Courtesy K. Harkay
5
Workshop on Time Domain Science Using X-ray
Techniques
D. Reiss, U. Michigan
10-16
10-14
10-12
10-10
10-8
10-6
10-4
10-2
100
102
104
106
Ultrasonic
EPR
NMR
Storage Ring Sources
X-ray FELs
6
Storage Ring ps Sources Offer Unique Capability
Note Femtoslicing not practical at APS (calc per
A. Zholents)
Energy modulation DE7se requires 10 mJ laser
pulse at lL400 nm and t50 fs giving 100 fs
x-ray pulse with 105 photons per pulse BUT
looks difficult at a high rep. rate
7
Generation of ps Pulses
Fig courtesy A. Nassiri
For APS h8, 6 MV deflect. voltage, sy,e 2.2
µrad, and sy,rad 5 µrad the calcd compressed
x-ray pulse length is 0.36 ps rms.
A. Nassiri
8
Parameters / Constraints What hV is Required?
Can get the same compression as long as hV is
constant
V6, h4
V4, h6
Higher V and lower h more linear chirp and less
need for slits
V6, h8
Higher h and lower V smaller maximum deflection
and less lifetime impact
Cavity design and rf source issues h7, Vlt6 MV?
Higher h and maximum V shortest pulse,
acceptable lifetime
Beam dynamics simulation study h
4 (1.4 GHz) V 6
MV (lifetime)
M. Borland, APS ps Workshop, May 2005
9
X-ray Compression Optics Simulation(R. Dejus)
Pulse histogram using optics compression and a
slit of 8.0 mm (half-width). The rms width is
1.19 ps (FWHM 2.8 ps), and the transmission is
62.
2D scatter plot, undulator A irradiance at 30 m,
10 keV. Red 10 level, green 37 (1/e).
20 are thrown away. The tilt angle ß is 55.
10
Ultrashort X-ray Pulse Generation Test 1

• Transient, short pulses are generated using an
  alternate scheme based on synchrobetatron
  coupling (suitable for study, but not operation).
  At left, an image of the beam 0.3 ms after a
  vertical kick, using APS sector 35 optical streak
  camera. At right, a short pulse is observed using
  a 40 µm vertical slit (fit gives 5.8 ps rms,
  nominal bunch length 30 ps).

W. Guo, K. Harkay, B. Yang, M. Borland, V.
Sajaev, Proc. 2005 PAC
Carry out a demonstration short-pulse x-ray
experiment on APS beamline Structural molecular
reorganization following photoinduced
isomerization/ dissociation can be studied on a
finer timescale. Transient pulse requires
acquisition of entire x-ray absorption spectrum
in a single shot. (L. Young and colleagues)
11
Room Temperature (RT) vs. Superconducting (SC)
RF(K. Harkay, A. Nassiri)

• No installed crab cavity in existing
  synchrotron facilities
• Need to study and understand transients during
  pulsing of RT structure and its effect on SR
  beam. How does it affect non-crab beamline
  users?
• RT pulsed system allows user to turn off ps
  pulse via timing (1 ?s pulse, 0.1 1 kHz rep
  rate) (M. Borland, P. Anfinrud)
• KEKB plans to install two SC single cell cavity
  in March 2006
• SC system runs CW, rep rate up to bunch spacing
  some experiments desire high rep rate (no pump
  probes) (D. Reiss)
• Either option requires dedicated RD effort

12
9 Cells SW Deflecting Structure
V. Dolgashev, SLAC, APS seminar, June 2005

• Pulsed heating lt 100 deg. C
• BMAX lt 200 kA/m for 5 µs pulse (surface)

• Limited available power 5 MW
• EMAX lt 100 MV/m (surface)

13
SC RF Cavity Study for APS (G. Waldschmidt, G.
Pile, D. Horan, R. Kustom, A. Nassiri, K. Harkay)
Frequency 2.81 GHz
Deflecting Voltage 6 MV
Qo 1 x 109
BMAX lt 100 mT
RF loss at 2 K lt 100 W
HOM Rt _at_ 100 mA lt 2.5 MW/m
HOM Rsfp _at_ 100 mA lt 0.8 MW - GHz
Available length 2.5 m

• Single-cell vs. multiple-cell SC cavity
  configurations
• Orbit displacement causes beam loading in
  crabbing mode adopt KEKB criterion of ?y 1 mm
  (for orbit distortions 0.1 mm)

Superconducting Deflecting Cavity Design
Parameters
14
Squashed RF Crab Cavity Designs (based on KEK
design)
1-cell 4-cell
Frequency (GHz) 2.81 2.81
No. of cavities 10 4
Cavity radius (cm) 7.8 7.8
Iris radius (cm) --- 1.8
Beam pipe rad (cm) 1.9 2.1 / 1.8
Defl voltage (MV) 6 6
Defl gradient (MV/m) 11.3 7.0
Qo 1 x 109 1 x 109
RT/Q (W/m) 50.6 10 220 4
Active length (cm) 5.3 10 21.3 4
BMAX (mT) 100 90
Pbeam (300mA) _at_ 1mm (kW) 100 100
RF loss at 2K (W) 7.1 10 10.2 4
15
Single-Cell Deflecting Cavity Coaxial Beam Pipe
Damper

• Multiple cells produce multiplicity of
  parasitic modes. Single-cell cavity chosen due
  to stringent HOM requirements.
• Coaxial beam pipe damper (CBD) is located on
  one side of the cavity and extracts LOMs and
  HOMs from cavity
• LOM / HOMs can couple to the coaxial beam pipe
  as a TEM mode. HOMs can also couple as higher
  order coaxial modes
• HOMs above beam pipe cutoff, propagate along
  CBD and / or through other beam pipe

16
Single-Cell Deflecting Cavity Rejection Filter

• Deflecting mode creates surface currents along
  the coaxial beam pipe damper, but does not
  propagate power.
• When a resistive element is added, there is
  substantial coupling of power into the damping
  material.
• A radial deflecting mode filter rejects at -10
  dB.
• Performance improvement pursued as well as
  physical size reduction.

17
LOM Damping

• Damping load is placed outside of cryomodule.
• Ridge waveguide and coaxial transmission lines
  transport LOM / HOM to loads
• Efficiency of deQing was simulated by creating
  the TM010 mode with an axial antenna.
• Stability condition for LOM achieved when Q lt
  12,900 for 100 mA beam current.
• Unloaded Q of LOM was 4.34e9.
• Coaxial beam pipe damper with four coaxial
  transmission lines, damped the LOM to a loaded Q
  of 1130.

18
Instability Thresholds from Parasitic Mode
Excitation(per Y-C. Chae)

• APS parameters assumed I 100 mA, E 7 GeV,
  a2.8e-4, ws/2p2 kHz, ns0.0073, bx 20 m

1 A. Mosnier, Proc 1999 PAC. 2 L.
Palumbo, V.G. Vaccaro, M. Zobov, LNF-94/041 (P)
(1994 also CERN 95-06, 331 (1995).
19
Damping Parasitic Modes f lt fc
20
Space Requirements

• Total available space at the APS for deflecting
  cavity assembly is 2.5 m.
• Assuming ten single-cell cavities, and 0.4 m on
  each end for an ion pump/valves/bellow assembly,
  the total space required by the following
  physical arrangement is 2.6 m.
• Additional dampers may also be required.
• Coupling between cavities must be addressed
• Impedance bump in coaxial beam pipe damper or
  shorting plane may be used
• Beam impedance considerations may require
  different cavity configuration

21
Summary of RF Design Considerations

• SC CW option is more attractive
• May offer a greater degree of compatibility with
  normal SR operation
• Compatible with future development of higher rep
  rate pump probe lasers
• Opportunity for APS to gain SC rf expertise
• Goal ?1 ps in crab insertion
• No impact of crab cavities on performance outside
  insertion
• Availability of 100-kW class rf amplifiers limits
  study to h 8 (2.8 GHz)
• Available insertion length for cavities nominally
  2.5 m (could be extended)
• Effects of errors emittance growth or orbit
  kicks (M. Borland)
• Intercavity phase error lt 0.04 for ltygt/sy lt 10
  
• Intercavity voltage difference lt 0.5
• Significant parasitic mode damping requirement
  (Y-C Chae)

22
SC RF Issues to be Resolved

• Physically realizable coaxial beampipe damper and
  optimized filter.
• Extraction of power from parasitic modes
• Removal of heat load outside of cryo-module.
• Quantify power handling requirements
• Additional dampers and/or enlarged beam pipe on
  input coupler side
• Examination of inter-cavity coupling
• Compare coaxial and waveguide input couplers for
  ease of implementation and HOM damping.
• Cavity optimization
• Tuner design
• Beam impedance calculation
• Evaluation of microphonics
• 2.5 m space limitation !!

23
RD Draft Plan

• Feasibility study completed
• SC rf technology chosen

• Model impedance effects (parasitic modes,
  head-tail)
• Finalize RF system design, refine simulations
• System review of crab cavities at KEKB during
  assembly and testing
• Refine x-ray compression optics design,
  end-to-end ray tracing
• Conduct proof of principle tests (beam dynamics,
  x-ray optics)
• Chirp beam using synchrobetatron coupling
  (transient) (W. Guo)
• Install 1 MV RT S-band structure, quarter
  betatron tune (M. Borland, W. Guo)
• Install warm model of SC rf cavity (passive),
  parasitic mode damping (K. Harkay)

24
Summary

• We believe x-ray pulse lengths 1 ps achievable
  at APS
• SC RF chosen as baseline after study of
  technology options
• RF priorities include parasitic mode damping and
  cavity assembly optimization with physical length
  constraints
• Beam impedance calculation may have appreciable
  effect on final design
• Optics design performed in parallel with RF.
  Optics compression increases throughput 2-15x
  better than slits alone.
• Proof of principle RD is underway beam/photon
  dynamics
• Operational system possibly 3 yrs from project
  start