RF Deflector Developments and Applications at SLAC

RF Deflector Developments and

Applications at SLAC

Juwen Wang

SLAC Accelerator Seminar

April 12 2011

Many thanks to the following colleagues for

their joint R&D efforts, discussions and

contributions:

G. Bowden, Y. Ding, V. Dolgashev, P. Emma,

J, Frisch, Z. Li, M. Litos, G. Loew, J.

Lewandowski, R. Reed, M. Ross, D. Schultz,

S. Tantawi, D. Xiang

Acknowledgment

1. Introduction to RF Deflectors

2. Basic Theory and Design of RF Deflectors• Basic Theory

• Design Examples

3. Deflector Applications (3 Types and 7 Examples) • Separator for High Energy Physics Experiments

• Time-resolved electron bunch diagnostics for the LCLS

• Measurement of Bunch Time Jitter at LCLS

• Increase slice energy spread σE as well as measure of slice parameters for Upgrade ECHO-7

• Bunch Longitudinal Profile Diagnostics at FLASH, DESY

• Ultra short electron and x-ray beams temporal diagnostics for LCLS

• Drive/Witness Bunch Longitudinal Profile Diagnostics for PWFA at FACET

4. Manufacture and Characterization• Fabrication Technology (Using X-Band Deflector As an Example)

• Tuning and Measurements

5. Future Work• Design Improvements

• More Applications

Outline

1. Introduction

• The RF deflectors are special types of microwave structures in which the

charged particles interact with transversely deflecting modes for a variety of

purposes.

• In 1960’s, SLAC built several RF deflectors called LOLA named by the

designers: Gregory Loew, Rudy Larsen and Otto Altenmueller.

• For fifty years since then, the RF deflectors have been extensively

studied and widely used in the accelerator field for the high energy physics

research and beam diagnostics of FEL and many other projects.

TM Longitudinally Accelerating Mode HEM Transversely Deflecting Mode

RF Deflector versus Accelerator

Snapshot of RF Electrical Field

RF Deflector ApplicationsThree Types and 7 Examples

- Separator for High Energy Physics Experiments

- Time-resolved electron bunch diagnostics for the LCLS

- Measurement of bunch time jitter at LCLS

- Bunch longitudinal profile diagnostics at DESY

- Ultra short e- and x-ray beams temporal diagnostics for

LCLS

- Drive/witness bunch longitudinal profile diagnostics for

PWFA at FACET

- Increase slice energy spread σE as well

as measure of slice parameters for

Upgrade ECHO-7

What the RF Deflectors Look Like?

A LOLA-IV

Ready for

DESY

A Short 13-Cell S-Band LOLA

Structure Under Measurement for

LCLS Injector

Final Assembly

of a 1m X-Band

Deflector for LCLS

Two Short

X-Band

Deflectors for

ECHO-7

2. Basic Theory and

Design of RF Deflectors•Basic Theory

•Design Examples

Principle of TW RF Deflector

dzBvEv

ep

l

o

dzx

Eep

l

z

0

Panofsky-Wenzel Theorem

zP

r

Ec

r

z

/

2

ULr

QVc

UL

QVr z

2

0

3

222

As a measure of the deflecting efficiency, the transverse shunt

impedance r┴ is defined as: where z and r are longitudinal and

transverse axes respectively, Ez is the

electrical field amplitude for the dipole

mode with angular frequency ω, and P

is the RF power as function of z.

Using the simulation codes for electromagnetic field

in RF structures, the transverse shunt impedance

can be calculated from:

“HEM-11 modes revisited”J.W. Wang and G.A. Loew (SLAC). SLAC-PUB-5321, Sep. 1990.

• Top: distribution of

deflection (V/m/4W)

• Bottom: Integrated

deflection (V/4W)

Beam at Maximum Kick Phase In

Deflecting Plane of a Short TW RF Deflector

Zhenghai Li

There are high electric and

magnetic fields in the coupler

regions. Due to their standing

wave characteristic, the total

contribution to the kick is

equivalent to 1/2+1/2 cell.

Design Example – IMaximum Kick 25 MV for Early LOLA S-Band RF Deflector

SLAC-17

• 3.6 m in length

• 2856 MHz

• As a RF separator, the maximum

deflecting voltage ~ 25 MV at 20 MW

input power

2.44 m

.

In order to characterize the extremely short bunch of the LCLS project, we need to extend

the time-resolved electron bunch diagnostics to the scale of 10-20 fs. The peak deflecting

voltage necessary to produce a temporal bunch resolution Δt is:

d

N Emc

tcneV

2

2

where E is the electron energy and the transverse momentum of the electron at time Δ t (with

respect to the zero-crossing phase of the RF) is py = eV┴/c, n is the kick amplitude in the unit of

nominal rms beam size, λ is the RF wavelength, εN is the normalized rms vertical emittance, c is the

speed of light, and βd is the vertical beta function at the deflector. This is for an RF deflector, which is

π/2 in betatron phase advance from a downstream screen.

Design Example – IIMaximum Kick of 33 MV for LCLS Bunch Length Measurement

X-Band RF Deflector Specifications

Parameter symbol value unit

Electron energy E 13.6 GeV

Desired temporal resolution t 10 fs

Offset of t-particle on screen, in units of rms beam size n 2

RF wavelength of deflector (X-band) 26 mm

Vertical normalized rms emittance N 1 mm

Vertical beta function at the center of the RF deflector d 50 m

Peak vertically accelerating voltage seen by beam V 33 MV

Parameter symbol value unit

Maximum repetition rate f 120 Hz

Minimum iris radius (if located after undulator) r 5 mm

Maximum cavity length (approx.) L 2 m

Minimum RF pulse length tRF 100 ns

RF frequency fRF 11.424 GHz

RF phase stability at f > 1 Hz (rms) jrms 0.05 deg-X

RF relative amplitude stability (rms) V/V0 1 %

Approximate specifications for an X-band RF deflecting cavity

Parameters for a 10-fs temporal resolution using an X-band RF deflecting cavity

Paul Emma

Technical Note

Oct. 2006Deflector Location: After Undulator

Design Examples for a Deflector

Structure type TW DLWG

Mode 2π/3 Backward wave

Aperture 2a 10.00 mm

Cavity diameter 2b 29.74 m

Cell length d 8.7475 mm

Disk thickness 1.45 mm

Quality factor Q 6400

Kick factor k 2.986x1016 V/C/m/m

Transverse shunt

impedance r┴

43.17 MΩ/m

Group velocity Vg/c - 3.165 %

Total length L 1.5 m

Filling time Tf 158 ns

Attenuation factor τ 0.885

Input peak RF power 30 MW

Maximum electric field 129 MV/m

Maximum magnetic field 0.45 MA/m

Deflecting voltage 38.9 MV

Frequency 11.424 GHz

Beam pipe diameter 10 mm

One cell length 8.747 mm

Phase advance per cell 2π/3

Kick per meter

[MeV/Sqrt [MW]]

31 MeV/m/Sqrt

(20 MW)

102 cell structure kick 21.3 MeV/Sqrt(20 MV)

Group velocity/ speed of

light

3.2 %

Filling time 92 ns

Structure length (with

beam pipes)

~94 cm

Structure design for a two-section

system by V. Dolgashev (AAC08,

SLAC-WP-084 )

Design for a one-section system by J. Wang &

S. Tantawi (LINAC2008, SLAC-PUB-13444)

System Layout

One-section systemTwo-section system

Regular X-Band Deflector Cups

Cup Shapes for Stabilization

of Desired Polarized Dipole Modes

Two holes

(LOLA Structures)Two caved-in walls

on cell ID surfacesDeforming using

two more tuning holes

Coupler Design Simulation

Maximum surface electric fields ~100 MV/m.Maximum surface magnetic fields ~400 kA/m,

Pulse heating 22 deg. C for 100 ns pulse.

V.A. Dolgashev, “Waveguide Coupler for X-band Defectors,” AAC 2008, SALC-WP-084

Fields Normalized 20 MW of transmitted power, or

21.3 MeV kick for 89 cm structure

Input/Output Coupler Assemblies

Completed Coupler AssemblyMechanical Design Model

Assembly of the Deflector

Name of Structure D27 D11

Structure Type 2π/3 Backward wave

Aperture 2a 10.00 mm

Cavity Diameter 2b 29.77 mm

Cell Length d 8.7475 mm

Disk Thickness 2.0 mm

Quality Factor Q 6320

Kick Factor k 2.849x1016 V/C/m/m

Transverse Shunt

Impedance r┴

41.9 MΩ/m

Group Velocity Vg/c - 3.165 %

Total Flange-Flange Length L 43.6 cm 29.6 cm

Filling Time Tf 26.8 ns 12 ns

Attenuation Factor τ 0.147 0.063

Input Peak RF Power 15 MW

for 6 MV Kick

77 kW

for 0.2 MV Kick

Maximum Electric Field 84 MV/m 6 MV/m

Maximum Magnetic Field 0.29 MA/m 0.02 MA/m

Design Example – IIIMaximum Kick 6 and 0.2 MV for ECHO-7 Experiment

3. Deflector Applications (3 Types

and 7 Examples) • Separator for High Energy Physics Experiments

• Time-resolved electron bunch diagnostics for the LCLS

• Measurement of Bunch Time Jitter at LCLS

• Increase slice energy spread σE as well as measure of slice parameters

for Upgrade ECHO-7

• Bunch Longitudinal Profile Diagnostics at FLASH, DESY

• Ultra short electron and x-ray beams temporal diagnostics for LCLS

• Drive/Witness Bunch Longitudinal Profile Diagnostics for PWFA at

FACET

RF Deflector as Separator

In 1960’s at Ends Station B and C, the LOLA RF deflectors

were used for secondary beams separation, where the

secondary particles of different species are naturally phase

shifted by their time of flight due to their different masses.

Two transverse RF deflectors at LCLS

Paul Emma

Bunch Measurement at LCLS

Paul Emma

Bunch Time Jitter Measurement at LCLS

Paul Emma

LOLA-IV at FLASH, DESY

Christopher Gerth

FLASH Seminar

2009

Beam Longitudinal Profile Measurement

at FLASH, DESY

Measurement of y

as a function of the

RF-phase

Estimation of σy

Christopher Gerth

FLASH Seminar

2009

TCAV1 to increase σE TCAV2 to measure slice parameters

Layout and upgrade of the ECHO-7 experimentWHY increasing beam slice energy spread?

• Echo-7 experiment demonstrated the ability to control the phase space correlations (D.

Xiang, et al, PRL, 105, 114801 (2010))

• The advantage of EEHG over HGHG has not been fully demonstrated because the

beam has very small slice energy spread

• Scaling to x-ray FEL requires confidence in harmonic generation with the relative beam

energy spread on the 10-4 level but NLCTA has σE /E~10-5

HOW: Using a rf deflecting cavity to increase beam slice energy spread to ~10 keV

Upgrade ECHO-7 Experiment at NLCTA Increase Slice Energy Spread and Measurement of Longitudinal Profile

Dao Xiang

Longitudinal phase space

before deflecting cavityDeflecting cavity

Longitudinal phase space

after deflecting cavity

Energy spread will be increased to 10~20 keV with TCAV1

Compared to the standard method “laser heater” which requires an additional

undulator and a laser, this scheme has the advantage of less cost, more uniform heating

effect and easier implementation

V ~ 100 kV

TCAV2 with voltage up to 6 MV is for measurement of beam slice parameters

Upgrade ECHO-7 Experiment at NLCTA Increase Slice Energy Spread and Measurement of Longitudinal Profile

Dao Xiang

4. Fabrication and

Characterization• Fabrication Technology (Using X-Band

Deflector As an Example)

• Tuning and Measurements

Annealing of Cups for Final Machining

Surface Flatness Measurement for All

Deflector Cups before Diffusion Bonding

Satisfactory Surface Flatness

Measured by ZYGO Machine

Microwave QC

of Regular Deflector Cups

All 227 cups were microwave QCed with satisfactory results by using the

following setup. The stack includes one half cup, eight regular cups and

one half cylinder to make 9 periodic structure with 10 HEM modes, one of

them is the working 2π/3 mode.

A Typical Microwave Measurement

Results for 9-Period CellStack

Vertical Kick Dipole Modes

Desired Horizontal Kick Dipole Modes

Working 2pi/3 Mode

Unwanted polarized Mode 80 MHz lower

S11 Resonant Modes and Dispersion Curve

for Horizontal Deflecting Modes

Backward Wave 2π/3 horizontal deflecting Mode

Two 57-Cell Stacks are Diffusion

Bonded for Whole Body Brazing

56 Regular CupsOne Matching Cup

Ready for Diffusion Bonding of the 1st

Stack (Left)and 2nd Stack (Right)

Diffusion Bonding

Conditions:

• Temperature 1020°C

• Holding Time 1 hour

• Weight 83 LB (pressure 40 PSI)

All Components Ready for Final Assembly

Microwave Measurement for Whole

Structure before Final Brazing

Output Coupler

Input Coupler

1st Stack

2nd Stack

Connection Cavity

In order to gain confidence to

pursue the final brazing,

microwave measurement for

whole stacked structure was

performed.

Final Brazing for Full Structure

Deflector Tuning for an Assembled RF Deflector

Deflector Tuning Methods

Perturbation in a resonant system

Selection of perturbation for structure tuning:

• Metallic bead on axis

• Metallic bead off axis

• Dielectric bead on axis – We used

• Dielectric bead off axis

tm

dH

Ev

)2

(2

02

0

D27 Amplitude and Cumulated Phase Shift

Zhenghai Li

Field Simulation to Understand the Bead-

pulling Measurement Results

Electric field Amplitude in the deflecting plane

Electric field Amplitude in the plane perpendicular to deflecting plane

Magnetic field Amplitude in the deflecting plane

Zhenghai Li

Amplitude and Phase for Electric and Magnetic

Fields in Deflecting Plane of a 17-Cell Deflector

Amplitude for Electric (Red) and Magnetic (Green) Fields in Deflecting Plane.

Phase for Electric (Red) and Magnetic Fields (Green) in Deflecting Plane.

Final S Parameters after Tuning

S11 shows a nice match for TW structure S12 shows consistency with theory

5. Future Work

1. Design Improvement

SW option for higher RF efficiency of short deflectors

2. X-Band RF System in LCLS Undulator Hall

3. More Applications• Ultra short electron and x-ray beams temporal

diagnostics for LCLS

• Deflector system for drive/witness bunch longitudinal

profile diagnostics for PWFA at FACET

SW option for higher RF

efficiency of short deflectors

Maximum surface electric fields ~105 MV/m.Maximum surface magnetic fields ~420 kA/m,

Pulse heating 24 deg. C for 100 ns pulse.

Fields normalized to 1.5 MW of input power, deflection 2 MeV

V. Dolgashev

Future Application – IUltra short electron and x-ray beams temporal

diagnostics with X-band deflector

Yuantao Ding

2

022

0

2 cossin

s

RFzsdxx

E

eVk 22

0

2 )( yyy D

e

z

2×1 m

ds

90°

V(t)RF

„streak‟Dipole

X-band TCAV

ener

gyDy

High resolution, ~ few fs;

Applicable for all radiation wavelength;

Wide diagnostic range, few fs to few hundred fs;

Profiles, single shot;

No interruption with LCLS operation;

Both e-beam and x-ray.

head

LCLS Hard x-ray case (250pC, 3kA)

OTRDMP, FEL ONOTRDMP, FEL OFF

Reconstruction e-beam profile Reconstruction x-ray profile

Yuantao Ding

PWFA

[1] M.J. Hogan, et al., New J. Phys. 12 055030 (2010)

[2] R.J. England, et al., Phys. Rev. Lett. 100 214802 (2008)

Plasma Wakefield Acceleration

(PWFA) experiments at FACET

will use unique two-bunch

beam.

Acceleration of witness bunch

depends strongly on

longitudinal profile of the two

bunches as well as the inter-

bunch spacing.

Will also try non-Gaussian

“ramped” driver bunch, which

could provide especially high

transformer ratio.

Good single-shot

measurement of longitudinal

profile is essential for

interpreting experimental

results.

(not real data)

Ramped Driver Concept

witness

ramped driver Ez

Mike Litos

Future Application – IIPlasma Wakefield Acceleration (PWFA) experiments at FACET

Measurement of longitudinal profile for both Drive Beam and Witness Beam

z,d=30µm

z,w=20µm

zd-w=140µm

Simulated TCAV

Response

TCAV Configuration

Frequency 11.424 GHz

Length 2 × 1m

Power 2 × 20 MW

Kick 2 × 20 MeV

Placement of TCAV and

Screen must be chosen

carefully such that TCAV

is at large y and Screen

is at /2 phase advance

downstream (at a waist).

Elegant simulation for a

pair of adjacent X-band

TCAVs shows terrific

capability to resolve the

longitudinal profile.

Placement of TCAV and Screen in FACET

y,T = 93m

y,T = 72cm

TCA

V

SCRE

EN

s (m)

y (m)

y≈/2

Sector 20 Plasma

(beam dir.)

Plasma Wakefield Acceleration (PWFA) experiments at FACET Measurement of longitudinal profile for both Dive Beam and Witness Beam

Mike Litos

Thank You Very Much!