Matthias Liepe Cornell University CORNELL Matthias Liepe - - PowerPoint PPT Presentation

▶
matthias liepe cornell university
SMART_READER_LITE
LIVE PREVIEW

Matthias Liepe Cornell University CORNELL Matthias Liepe - - PowerPoint PPT Presentation

Aug 13, 2023 943 likes •1.71k views

- Challenges and Solutions - Matthias Liepe Cornell University CORNELL Matthias Liepe 12/12/2003 - 1 - U N I V E R S I T Y ERL Light Source: Why? The Cornell ERL: An Overview Prototype 5 GeV SR Source SC RF in the Cornell

slide-1
SLIDE 1
  • Challenges and Solutions -

Matthias Liepe Cornell University

U N I V E R S I T Y

CORNELL

  • 1 -

Matthias Liepe 12/12/2003

slide-2
SLIDE 2
  • ERL Light Source: Why?
  • The Cornell ERL: An Overview

– Prototype – 5 GeV SR Source

  • SC RF in the Cornell ERL Injector
  • Challenges and Solutions -
  • SC RF in the Cornell ERL Main Linac
  • Challenges and Solutions -
  • Outlook and Summary

U N I V E R S I T Y

CORNELL

  • 2 -

Matthias Liepe 12/12/2003

slide-3
SLIDE 3

Today's Workhorse Light Sources: Storage Rings Today's Workhorse Light Sources: Storage Rings

  • 1st generation

parasitic SR on high energy physics storage rings

  • 2nd generation

dedicated bending magnet sources, designed for high flux SR

  • 3rd generation

dedicated undulator sources optimized for brilliance, using high current, low emittance

Some rings use superconducting RF

Storage ring light sourses give:

  • Repetition rate
  • Stability
  • High flux, brilliance – average/peak

U N I V E R S I T Y

CORNELL

  • 3 -

Matthias Liepe 12/12/2003

slide-4
SLIDE 4

Synchrotron Radiation at Cornell Synchrotron Radiation at Cornell

  • 1952: 1st accurate measurement of synchrotron radiation power by Dale

Corson with the Cornell 300MeV synchrotron.

  • 1953: 1st measurement of the synchrotron radiation spectrum by Paul

Hartman with the Cornell 300MeV synchrotron.

  • Worlds 1st synchrotron radiation beam line (Cornell 230MeV synch.)
  • 1961: 1st measurement of radiation polarization by Peter Joos with the

Cornell 1.1GeV synchrotron.

  • 1978: X-Ray facility CHESS is being build at CESR
  • 2003: 1st Nobel prize with CESR data goes to R.MacKinnon

Roderick MacKinnon

Dale Corson Cornell’s 8th president

U N I V E R S I T Y

CORNELL

  • 4 -

Matthias Liepe 12/12/2003

slide-5
SLIDE 5

More Users More Users… …

Historical Growth of Synchrotron Publications Worldwide: ISI data 1968 to 2002, Keyword: “synchrotron” (from G. Margaritondo) Protein structures in Protein Data Bank (Mostly from SR)

U N I V E R S I T Y

CORNELL

  • 5 -

Matthias Liepe 12/12/2003

slide-6
SLIDE 6

… … More Demands: What do we need in the future? More Demands: What do we need in the future?

  • 1. High average and high peak
  • Brilliance (photons/s/0.1% bw/mrad2/mm2)
  • Flux (photons/s/0.1% bw)
  • 2. Coherence
  • 3. Flexible pulse structure
  • Programmable pulse trains (interval, bunch size)
  • Adjustable pulse lengths down to the femtosecond regime
  • 3. Small x-ray source size of desired shape, e.g. circular
  • 4. Flexibility of source operation
  • No fill decay
  • Stability & robustness
  • Easily upgraded

FEL ERL

U N I V E R S I T Y

CORNELL

  • 6 -

Matthias Liepe 12/12/2003

slide-7
SLIDE 7

Important Beam Parameters: Important Beam Parameters: A Wish List A Wish List

Low emittance Low emittance

  • Decreasing the

electron beam emittance down to diffraction limit

  • High spectral

High spectral brilliance SR sources brilliance SR sources

  • High coherence

High coherence fraction: fraction:

pc =λ2/(4π)2εxεy B ∝ 1/εxεyτ

Long undulator Low energy spread Low energy spread Short bunches Short bunches High flux High flux

F ∝ Ibeam

High beam current High beam current

U N I V E R S I T Y

CORNELL

  • 7 -

Matthias Liepe 12/12/2003

slide-8
SLIDE 8

How do we get these Beam Parameters? How do we get these Beam Parameters? Limits of Storage Rings Limits of Storage Rings

  • Electron beam emittance, bunch profile and energy spread in a

storage ring is determined by equilibrium between radiation damping and two main diffusion processes:

  • quantum fluctuation of the SR and
  • the intrabeam scattering.

⇒There is no (affordable) way to decrease the horizontal emittance in storage ring εx< 10 -10 m·rad and energy spread σΕ/E < 10 -3.

  • Beam lifetime limits bunch length to about 10 ps. Too long for many

dynamic processes.

  • Technology well developed. Theoretical limits are being approached.
  • Time structure cannot be tailored to user needs.
  • Fills are necessary, intensity is not constant.

⇒ Equilibrium dynamics determines almost all the parameters on our wish list!

U N I V E R S I T Y

CORNELL

  • 8 -

Matthias Liepe 12/12/2003

slide-9
SLIDE 9

How do we get these Beam Parameters? How do we get these Beam Parameters? The Alternative: Linacs The Alternative: Linacs

  • Injectors can be built with very brilliant e- beams and linacs can accelerate

with very low emittance growth (if we do it right).

  • Emittance & pulse length determined by injector.
  • Single pass non equilibrium device.
  • Easy upgrade path: Better e- source gives higher brilliance.
  • Due to adiabatic damping an emittance ε ~ 10 -11 m·rad and energy

spread σE/E ~ 10 -4 is possible for energies E > 5 GeV.

  • Potential for ultra high brilliance.
  • Complete flexibility of bunch timing.
  • No fill decay, constant intensity.
  • Electron bunches dumped after single pass.

U N I V E R S I T Y

CORNELL

  • 9 -

Matthias Liepe 12/12/2003

slide-10
SLIDE 10

The Alternative: Linacs The Alternative: Linacs Small Beam Size, Coherence, Short Bunches Small Beam Size, Coherence, Short Bunches

U N I V E R S I T Y

CORNELL

  • 10 -

Matthias Liepe 12/12/2003

5 GeV storage ring ERL 5GeV@100mA 100fs 2ps 16ps

Factor 100 more coherent flux for ERL for same x-rays, or provide coherence for harder x-rays 3rd SR coherent coherent ERL

slide-11
SLIDE 11

Linac Light Source: X Linac Light Source: X-

  • Rays Studies in New Regimes

Rays Studies in New Regimes

U N I V E R S I T Y

CORNELL

  • 11 -

plots from Q. Shen frep = MHz … GHz

Matthias Liepe 12/12/2003

Short pulses, high brilliance: Short pulses, high brilliance: High coherent fraction: High coherent fraction:

slide-12
SLIDE 12

Linac Light Source: X Linac Light Source: X-

  • Rays Studies in New Regimes

Rays Studies in New Regimes

Smaller beams lead to better spatial resolution (currently sub mm) 3-D Studies of Structure ERL: 100 to 1000 times smaller area Smaller emittance leads to high brilliance. ERL: 10 to 1000 more brilliance. Shorter bunches allows much higher time resolution.

U N I V E R S I T Y

CORNELL

3D Tomograph of Cells

  • G. Schneider, LBNL

Insect Breathing

Field museum of Chicago & APS, Argonne National Lab.

ERL: 100 times shorter bunches

  • 12 -

Matthias Liepe 12/12/2003

slide-13
SLIDE 13

Linac Light Sources: How to get high currents? Linac Light Sources: How to get high currents?

  • High photon flux ⇒ need high current
  • But: With a simple linac you’d go broke!!
  • Example: 5 GeV * 100 mA = 500 MW

⇒ The energy of the spent beam has to be recaptured for the new beam.

U N I V E R S I T Y

CORNELL

  • 13 -

Matthias Liepe 12/12/2003

slide-14
SLIDE 14

Previous Energy Recovery Machines

U N I V E R S I T Y

CORNELL

Leonardo da Vinci (1452-1519)

  • 14 -

Matthias Liepe 12/12/2003

slide-15
SLIDE 15

Linac Light Sources: How to get high currents? Linac Light Sources: How to get high currents? The Energy The Energy-

  • Recovery

Recovery-

  • Linac

Linac

Solution: Use energy recovery. First proposed by M. Tigner in 19 Solution: Use energy recovery. First proposed by M. Tigner in 1965. 65.

  • Re-use energy of beam after SR generation.
  • Recirculate beam and pass it through the linac a second

time, but 180 deg. out of phase to decelerate beam.

  • ⇒

⇒ “ “Energy Storage Ring Energy Storage Ring” ” but not but not “ “Beam Storage Ring Beam Storage Ring” ”. .

U N I V E R S I T Y

CORNELL

  • 15 -

Matthias Liepe 12/12/2003

slide-16
SLIDE 16

ERLs: What is the trick? ERLs: What is the trick?

1 . I n j e c t i

  • n

: f r e s h b e a m

  • 3. SR generation with

low emittance beam

  • 6. dump beam:

low dump energy, less radioactivity

  • 2. acceleration

4 . R e c i r c u l a t e b e a m 5 . d e c e l e r a t i

  • n

t

  • r

e

  • u

s e e n e r g y

U N I V E R S I T Y

CORNELL

  • 16 -

Matthias Liepe 12/12/2003

slide-17
SLIDE 17

ERLs Worldwide (FEL ERLs Worldwide (FEL-

  • ERLs and SR

ERLs and SR-

  • ERLs)

ERLs)

U N I V E R S I T Y

CORNELL

1970 1980 2000

SCA, Stanford, 1986

  • M. Tigner, 1965

S-DALINAC, 1990 IR FEL Jlab, 1999 JAERI, 2002 Cornell ERL LUX (LBL) PERL (NSLS) 4 GLS (Daresbury) KEK …

1960 1990

  • 17 -

Matthias Liepe 12/12/2003

slide-18
SLIDE 18

ERL Linacs: Why Superconducting Cavities? ERL Linacs: Why Superconducting Cavities?

SRF linacs can deliver beams of superior quality:

  • Smaller emittance (lower impedance) ⇒ higher brilliance
  • Better RF control and stability ⇒ lower energy spread
  • CW operation at high gradient ⇒ flexibility in pulse train,

lower impedance, cost saving In addition, SRF gives

  • Higher power conversion efficiency
  • ERL option (very low wall losses) ⇒ high beam current, high

flux

U N I V E R S I T Y

CORNELL

  • 18 -

Matthias Liepe 12/12/2003

slide-19
SLIDE 19

ERL @ Cornell ERL @ Cornell

U N I V E R S I T Y

CORNELL

  • 19 -

Matthias Liepe 12/12/2003

slide-20
SLIDE 20

Cornell ERL Light Source Cornell ERL Light Source

5 GeV, 100 mA 5 GeV, 100 mA

5 GeV

Neither an electron source, nor an injector system, nor an ERL has ever been built for the required large beam powers and small transverse and longitudinal emittances. A prototype at Cornell should verify the functionality.

U N I V E R S I T Y

CORNELL

  • 20 -

Matthias Liepe 12/12/2003

slide-21
SLIDE 21

Cornell ERL: Phase 1, the Prototype Cornell ERL: Phase 1, the Prototype

30m

100 mA Gun Buncher Dump Bates bends 100 MeV Main linac 5 MeV Injector

100 MeV, 100 mA 100 MeV, 100 mA

s.c. main linac s.c. injector linac

U N I V E R S I T Y

CORNELL

  • 21 -

Matthias Liepe 12/12/2003

slide-22
SLIDE 22

Cornell ERL Prototype: SRF Parameter Cornell ERL Prototype: SRF Parameter

7 5 1 number of cavities 0.06° 0.1° 0.1°

required phase stability (rms)

3·10-4 1·10-3 8·10-3 required relative amplitude stability (rms) 11 130 7 required klystron power per cavity [kW] ≈ 16 (20 MV/m) 1 (3) 0.12

  • acc. voltage per cavity [MV]

2.6·107 for 25 Hz peak detuning 4.6·104 ( 4.1·105) 9,900 Qext > 1010 > 5·109 20,000 Q0 392 109 105 R/Q [Ω] (circuit def.) 7 2 1 cells per cavity 1300 1300 1300 frequency [MHz] ERL s..c. main linac cavities ERL s.c. injector cavities ERL buncher cavity

U N I V E R S I T Y

CORNELL

  • 22 -

Matthias Liepe 12/12/2003

slide-23
SLIDE 23

Cornell ERL @ CESR Cornell ERL @ CESR

Initial idea

Laboratory of Elementary Particle Physics, Cornell University June, 2002 ERL 02-6

Energy Recovery Linac in the Wilson Tunnel

Richard Talman ABSTRACT This is a brief discussion of two out of the many modifications that will be needed to retrofit the Wilson laboratory as an energy recovery linac (ERL). The two issues are: fitting the facility within the existing site boundaries; and designing the approximately isochronous yet adjustable arcs needed to transport ultrashort bunches.

U N I V E R S I T Y

CORNELL

  • 23 -

Matthias Liepe 12/12/2003

slide-24
SLIDE 24

Phase 2: Cornell ERL @ CESR Phase 2: Cornell ERL @ CESR

Injector RF Injector RF Main Linac RF Main Linac RF

U N I V E R S I T Y

CORNELL

  • 24 -

Matthias Liepe 12/12/2003

slide-25
SLIDE 25

Cornell ERL @ CESR: Work in Progress Cornell ERL @ CESR: Work in Progress

U N I V E R S I T Y

CORNELL

  • 25 -

Matthias Liepe 12/12/2003

slide-26
SLIDE 26

520 cm 190 cm 440 cm 80 cm 90 cm 30 cm 125 cm 210 cm

275 cm

Figure 3. Main LINAC, Damping Ring & Klystron Station

Matthias Liepe 12/12/2003

U N I V E R S I T Y

CORNELL

Main Linac Tunnel with Two Linacs Main Linac Tunnel with Two Linacs

  • 26 -
slide-27
SLIDE 27

Injector SRF Injector SRF

U N I V E R S I T Y

CORNELL

  • 27 -

Matthias Liepe 12/12/2003

slide-28
SLIDE 28

Cornell ERL Injector RF Challenges Cornell ERL Injector RF Challenges

  • High RF power transfer to beam for acceleration of high

current beam ⇒ input coupler challenge 1 MeV per cavity * 0.1 A = 100 kW!

  • Strong damping of HOMs essential for beam stability

and to reduce monopole power.

  • Emittance preservation ⇔ space charge, small

transverse kick fields

U N I V E R S I T Y

CORNELL

  • 28 -

Matthias Liepe 12/12/2003

slide-29
SLIDE 29

Low Emittance Preservation in a Linac Low Emittance Preservation in a Linac

Wakefields:

  • longitudinal wakes generate energy

spread

  • transverse wakes generate time

dependent kick fields ⇒ transverse emittance growth

Coupler kicks:

  • Rotation asymmetry of input

coupler and HOM coupler gener- ates time dependent transverse kick fields on the cavity axis bunch

∆tb

Emittance growth Space Charge

U N I V E R S I T Y

CORNELL

  • 29 -

Matthias Liepe 12/12/03

slide-30
SLIDE 30

Cornell ERL Injector Cornell ERL Injector

merger into main linac

2-cell s.c. cavities

DC gun buncher

  • DC gun produces low emittance beam (Qb= 77 pC).
  • Fast acceleration in first cavity to reduce space charge

effects.

  • Short 2-cell cavities to reduce RF power per input

coupler and to achieve strong HOM damping.

U N I V E R S I T Y

CORNELL

  • 30 -

Matthias Liepe 12/12/2003

slide-31
SLIDE 31

Cornell ERL Injector: 2 Cornell ERL Injector: 2-

  • Cell Cavity

Cell Cavity

Symmetric twin input coupler Reduced iris to maximize R/Q of accelerating mode

Large 106 mm diameter tube to propagate all TM monopole HOMs and all dipole modes

facc = 1.3 GHz First copper model

U N I V E R S I T Y

CORNELL

  • 31 -

Matthias Liepe 12/12/2003

slide-32
SLIDE 32

Cornell ERL Injector Cryomodule Concept

He gas return pipe

77 K shield 5 K shield

2-phase 2 K He pipe

vacuum vessel

cavity interconnection with bellow and HOM ring absorber 2-cell cavity beam

U N I V E R S I T Y

CORNELL

  • 32 -

Matthias Liepe 12/12/2003

slide-33
SLIDE 33

Cornell ERL Injector: 2-Cell Cavity with LHe Vessel

2-cell cavity inside LHe vessel input coupler flange 2-phase He pipe

U N I V E R S I T Y

CORNELL

  • 33 -

Matthias Liepe 12/12/2003

slide-34
SLIDE 34

Cornell ERL Injector: Cavity with Support and Alignment Structure

sliding support with alignment screws sliding support with alignment screws frequency tuner invar rod He-gas return pipe

U N I V E R S I T Y

CORNELL

  • 34 -

Matthias Liepe 12/12/2003

slide-35
SLIDE 35

Sliding Cavity Support (TTF Module Style)

  • Sliding supports define transverse cavity positions.
  • Longitudinal position is defined by invar rod.
  • Cavity stays in place during cool-down to 2 K.

U N I V E R S I T Y

CORNELL

  • 35 -

Matthias Liepe 12/12/2003

slide-36
SLIDE 36

Cornell ERL Injector Cryomodule: Cavity with HOM Absorber

U N I V E R S I T Y

CORNELL

Matthias Liepe 12/12/2003

  • 36 -
slide-37
SLIDE 37

Cornell ERL Injector: 50 kW CW Input Coupler

cold window in standing wave minimum 62 mm diameter outer conductor ⇒ MP free air-cooled bellow Design studies under progress, high power cw test planed.

U N I V E R S I T Y

CORNELL

  • 37 -

Matthias Liepe 12/12/2003

slide-38
SLIDE 38

Symmetric Twin Input Coupler Symmetric Twin Input Coupler

Symmetric twin input coupler

  • 50 % less power

50 % less power per coupler. per coupler.

  • Reduced emittance

Reduced emittance growth in the growth in the injector. injector.

U N I V E R S I T Y

CORNELL

  • 38 -

Matthias Liepe 12/12/2003

slide-39
SLIDE 39

Cornell ERL Injector Cryomodule: HOM Damping

  • Accelerating mode does not propagate in tubes.
  • All TM monopole HOMs and all dipole modes

propagate in tubes and are damped by ferrite ring absorber between cavities.

  • Example: Lowers frequency dipole mode:

symmetric ferrite ring ⇒ no kick

U N I V E R S I T Y

CORNELL

  • 39 -

Matthias Liepe 12/12/2003

slide-40
SLIDE 40

ERL Injector: HOM Monopoles (CLANS Model)

4.34 m

ferrite #1 ferrite #4 ferrite #5 ferrite #3 ferrite #2 ferrite #6

beam

U N I V E R S I T Y

CORNELL

  • 40 -

Matthias Liepe 12/12/2003

slide-41
SLIDE 41

ERL Injector: Damping of HOM Monopoles (CLANS Simulation)

1000 1500 2000 2500 3000 3500 4000 4500 10

  • 2

10 10

2

10

4

10

6

10

8

10

10

10

12

f [MHz] Qferrite

accelerating mode strongly damped HOMs

U N I V E R S I T Y

CORNELL

  • 41 -

Matthias Liepe 12/12/2003

slide-42
SLIDE 42

Main Linac SRF Main Linac SRF

U N I V E R S I T Y

CORNELL

  • 42 -

Matthias Liepe 12/12/2003

slide-43
SLIDE 43

Cornell ERL Main Linac RF Challenges Cornell ERL Main Linac RF Challenges

  • Strong damping of HOMs essential for beam stability

and to reduce monopole power.

  • HOM power extraction at temperature with good cryo-

efficiency.

  • Emittance preservation ⇔ small transverse kick fields.
  • RF field control and efficient cavity operation.
  • Cavity module operation at high cw field gradient with

large cryo-losses.

U N I V E R S I T Y

CORNELL

  • 43 -

Matthias Liepe 12/12/2003

slide-44
SLIDE 44

HOM Damping Challenge 1: HOM Damping Challenge 1: High Beam Currents and HOM Monopole Power High Beam Currents and HOM Monopole Power

In average the total HOM losses per cavity are given by the single bunch losses (7cell ERL cavity, 77 pC bunch charge, 2.6 GHz bunch repetition rate, σb= 600 µm):

W 140 A 0.2 77pC V/pC 1 . 9

|| ||

= ⋅ ⋅ = =

beam bunchI

Q k P

But: If a monopole mode is excited on resonance, the loss for this mode can be much higher:

2

2

beam

QI Q R P         =

Need strong HOM damping! ⇒ Example: To stay below 200 W (I=2*100 mA) 200 W (I=2*100 mA):

  • achieve (R/Q)Q < 2500

(R/Q)Q < 2500 Ω Ω,

  • or avoid resonant excitation of the mode.

U N I V E R S I T Y

CORNELL

  • 44 -

Matthias Liepe 12/12/03

slide-45
SLIDE 45

HOM Damping Challenge 2: HOM Damping Challenge 2: HOM Monopole Power at High Frequencies HOM Monopole Power at High Frequencies

600 µm bunch length in the linac

100 200 300 50 100 150 200 250 f [GHz] power [W] integrated power up to frequency f

HOM power HOM power up to about 100 GHz. up to about 100 GHz.

U N I V E R S I T Y

CORNELL

  • 45 -

Matthias Liepe 12/12/03

slide-46
SLIDE 46

HOM Damping Challenge 3: HOM Damping Challenge 3: Avoid Beam Breakup (BBU) Avoid Beam Breakup (BBU)

  • In an ERL the feedback system formed between cavities

and the beam is closed. ⇒ Instability can result at sufficient high currents.

Q Q R IBBU ) / ( ω ∝

  • simple model for instability beam current:

Strong HOM damping! Need IBBU > 100 mA!

U N I V E R S I T Y

CORNELL

  • 46 -

Matthias Liepe 12/12/03

slide-47
SLIDE 47

BBU Limit for the Cornell 5 GeV, 100 mA ERL BBU Limit for the Cornell 5 GeV, 100 mA ERL

threshold current [mA]

] /GHz /cm [ f / ) / (

2

Ω Q Q R

104 103 106 105 104 103 102 Simulation by I. Bazarov

⇒ (R/Q)Q/f < 2·105 Ω/cm2/GHz required for BBU instability current > 100 mA (with safety-factor).

U N I V E R S I T Y

CORNELL

  • 47 -

Matthias Liepe 12/12/03

slide-48
SLIDE 48

HOM Damping in the Cornell ERL Main Linac: RF Structure Candidates

2x7 2x7-

  • cell superstructure:

cell superstructure:

good for 10 mA

7 7-

  • cell cavities:

cell cavities:

good for 100 mA

2x5 2x5-

  • cell superstructure:

cell superstructure:

good for 100 mA ?

9 9-

  • cell cavities:

cell cavities:

good for 10 mA

U N I V E R S I T Y

CORNELL

  • 48 -

Matthias Liepe 12/12/2003

slide-49
SLIDE 49

HOM Damping in the Cornell ERL Main Linac: 7-Cell Cavity

2 K reduced iris to maximize R/Q of accelerating mode

facc = 1.3 GHz

large 106 mm diameter tube to propagate all TM monopole HOMs and most dipole modes small 78 mm beam tube 7-cell s.c. cavity, TESLA shaped center cells

U N I V E R S I T Y

CORNELL

  • 49 -

Matthias Liepe 12/12/2003

slide-50
SLIDE 50

HOM Damping in the Cornell ERL Main Linac: Damping Concept

2 HOM loop couplers 4 HOM loop couplers 2 K 80 K 80 K ferrite ring absorber ferrite ring absorber

  • Enlarged beam tube on one side to propagate all TM monopole modes

and most dipole modes.

  • Ferrite broadband absorbers at 80 K between cavities to damp

propagating modes at temperature with good cryo-efficiency.

  • 6 HOM loop coupler per cavity to reduce power per coupler and to

damp quadrupole modes reliable.

  • Opposite HOM couplers to reduce transverse kicks.

U N I V E R S I T Y

CORNELL

  • 50 -

Matthias Liepe 12/12/2003

slide-51
SLIDE 51

Setup to measure Ferrite Absorber Losses at Setup to measure Ferrite Absorber Losses at ≈ ≈ 100 K 100 K absorber material inside cooled air-line RF power transmitted power reflected power

Absorber

RF losses in the absorber material can be calculated from the reflected and transmitted power as function of the RF frequency.

U N I V E R S I T Y

CORNELL

  • 51 -

Matthias Liepe 12/12/2003

slide-52
SLIDE 52

Measured Ferrite Absorber Losses Measured Ferrite Absorber Losses

U N I V E R S I T Y

CORNELL

  • 52 -

Matthias Liepe 12/12/2003

slide-53
SLIDE 53

Example: Ferrite TT2 Example: Ferrite TT2-

  • 111R Absorber Properties

111R Absorber Properties at 300 K and at 300 K and 80 K 80 K

5 10 15 x 10

9

8 10 12 14 16 Frequency / Hz Re ε / ε0 5 10 15 x 10

9

  • 1.5
  • 1
  • 0.5

0.5 Frequency / Hz Im ε / ε0 5 10 15 x 10

9

5 10 15 Frequency / Hz Re µ / µ0 5 10 15 x 10

9

  • 15
  • 10
  • 5

Frequency / Hz Im µ / µ0 297K 100K

Magnetic Resonance Magnetic Resonance

U N I V E R S I T Y

CORNELL

  • 53 -

Matthias Liepe 12/12/2003

slide-54
SLIDE 54

HOM Damping in the Cornell ERL Main Linac: CLANS Simulations

TT2-111R ferrite absorber at 80 K TT2-111R ferrite absorber at 80 K TT2-111R ferrite absorber at 80 K bellow

1000 2000 3000 4000 5000 6000 7000 8000 10

  • 2

10 10

2

10

4

10

6

10

8

10

10

10

12

10

14

R/Q*Q per cavity (circ. def.) [ Ω] frequency [MHz] fundamental mode monopole limit for first beam harmonics monopole limit for second beam harmonics monopole limit for third beam harmonics

Monopole Modes Monopole Modes

BBU limit BBU limit

1000 1500 2000 2500 3000 3500 10

  • 2

10

  • 1

10 10

1

10

2

10

3

10

4

10

5

R/Q*Q/f per cavity [Ω /cm2GHz] frequency [MHz]

Dipole Modes Dipole Modes

slide-55
SLIDE 55

HOM Damping in the Cornell ERL Main Linac: Bellow-Ferrite-Bellow Section

53 mm ferrites (TT2-111R) flange gap bellow

TT2-111R ferrite absorber at 80 K TT2-111R ferrite absorber at 80 K TT2-111R ferrite absorber at 80 K bellow

longitudinal loss factor k = 1.4 V/pC

U N I V E R S I T Y

CORNELL

  • 55 -

Matthias Liepe 12/12/2003

slide-56
SLIDE 56

HOM Damping in the Cornell ERL Main Linac: Bellow-Ferrite-Bellow Section

TT2-111R ferrite absorber at 80 K TT2-111R ferrite absorber at 80 K TT2-111R ferrite absorber at 80 K bellow

Trapped modes in bellows: HOM power goes into wall, not into cooled ferrite!

U N I V E R S I T Y

CORNELL

  • 56 -

Matthias Liepe 12/12/2003

slide-57
SLIDE 57

HOM Damping in the Cornell ERL Main Linac: Improved Bellow-Ferrite-Bellow Section

TT2-111R ferrite absorber at 80 K TT2-111R ferrite absorber at 80 K TT2-111R ferrite absorber at 80 K bellow

longitudinal loss factor k = 1.2 V/pC

U N I V E R S I T Y

CORNELL

  • 57 -

Matthias Liepe 12/12/2003

slide-58
SLIDE 58

HOM Damping in the Cornell ERL Main Linac: Improved Bellow-Ferrite-Bellow Section

TT2-111R ferrite absorber at 80 K TT2-111R ferrite absorber at 80 K TT2-111R ferrite absorber at 80 K bellow

HOM power goes into cooled ferrite!

U N I V E R S I T Y

CORNELL

  • 58 -

Matthias Liepe 12/12/2003

slide-59
SLIDE 59

Ferrite Shielded Bellow

LN cooling loop bellows ferrite tile

U N I V E R S I T Y

CORNELL

  • 59 -

Matthias Liepe 12/12/2003

slide-60
SLIDE 60

HOM Loop Coupler Studies: HOM Loop Coupler Studies:

Microwave Studio model: Copper model:

Coupler model: superconducting pick-up antenna superconducting pick-up loop capacitive coupling

  • utput to

room temp. load capacitor of the 1.3 GHz notch filter

U N I V E R S I T Y

CORNELL

  • 60 -

Matthias Liepe 12/12/2003

slide-61
SLIDE 61

Main Linac Accelerating RF Field Main Linac Accelerating RF Field

  • Flexibility and stability with small ∆E/E
  • Efficient high QL operation with energy recovery:
  • efficient energy recovery and RF control of random

beam loading

  • RF control in presents of microphonics and Lorentz-

force detuning ⇒ optimal loaded Q?

  • Minimum RF power required?

U N I V E R S I T Y

CORNELL

  • 61 -

Matthias Liepe 12/12/03

slide-62
SLIDE 62

ERL: RF Control of Random Beam Loading ERL: RF Control of Random Beam Loading

Ibeam Idec Ig Vg

⇒ Path length errors (phase errors) and non-zero recirculation time will result in beam loading with fluctuation! Re

U N I V E R S I T Y

CORNELL

  • 62 -

Matthias Liepe 12/12/03

slide-63
SLIDE 63

Microphonics Microphonics

  • Microphonics: modulation of resonance frequency by

external mechanical disturbances

  • thin wall thickness and small bandwidth of

superconducting cavities ⇒ sensitive to microphonics

U N I V E R S I T Y

CORNELL

time [s] cavity detuning [Hz]

Example: TTF 9-cell cavity in a horizontal test cryostat (cw operation)

Matthias Liepe 12/12/03

Cornell ERL: Cavity bandwidth = 25 Hz ! ⇒ Large field errors without field control!

  • 63 -
slide-64
SLIDE 64

Cornell ERL: Required RF Power without Beam Loading Cornell ERL: Required RF Power without Beam Loading

106 107 108 109

Qext power per cavity [kW]

10 Hz 20 Hz 30 Hz 40 Hz 0 Hz

f f Qopt ∆ 2 1 =

f f Q R V P

acc g

∆ /

2 min ,

=

2 2 / 1

1 4         − + = f f f P Q Q R V

⇓

7-cell main linac cavity at 20 MV/m:

3.9 7.8 11.7

Example: 310 cavities for 5 GeV, 50 % klystron efficiency 25 Hz peak detuning ⇒ 3.9 MW wall plug power 10 Hz peak detuning ⇒ 1.6 MW wall plug power

g L

Peak microphonics detuning important! Peak microphonics detuning important!

U N I V E R S I T Y

CORNELL

  • 64 -

Matthias Liepe 12/12/2003

slide-65
SLIDE 65

Cornell ERL: Frequency Tuner

piezo +/- 1 kHz, 1 Hz resol. stepping motor +/- 200 kHz at 300 K!

⇒ Slow and fast frequency control!

JLab upgrade tuner

DESY blade tuner

slide-66
SLIDE 66

Ultra-Fast Digital RF Field Control System for the Cornell ERL

2 DACs 4 ADCs Virtex II FPGA DSP

RF system synthesizer 1300 MHz vector modulator RF switch

klystron

I Q RF on/off, trip

fast control slow control + DAQ

piezo-tuner fast interlock card link ports LO Master Oscillator 130 MHz ADC ADC ADC DAC ADC I Q Pt1 Pkly Pdrive Pr Pf phi

cavity

DAC FPGA 1300 MHz + 13 MHz ADC ADC ADC DAC DAC ADC FPGA

memory

sample- buffer DSP

memory

sample- buffer DSP

  • very low delay in the

control loop (< 1 µs)

  • Field Programmable Gate

Array (FPGA) design combines the speed of an analog system and the flexibility of a digital system

  • high computation power

allows advanced control algorithms

  • all boards have been

designed in house

  • generic design: digital

boards can be used for a variety of control and data processing applications

slide-67
SLIDE 67

Test with Copper Cavity

Matthias Liepe 12/12/2003

U N I V E R S I T Y

CORNELL

  • 67 -

500 MHz CESR copper cavity: 500 MHz CESR copper cavity: Digital Boards: Digital Boards:

slide-68
SLIDE 68

Cryogenic Challenges Cryogenic Challenges

  • High gradient cw operation: dynamic head load dominates:

Example: 20 MV/m, Q0 = 1010 ⇒ 40 W/m

  • Module design:
  • Heat transfer through LHe ⇒ need large enough pipes
  • Mass transport of helium gas ⇒ need large enough pipes
  • HOM losses ⇒ need cooling of absorbers
  • Cavity:
  • Cavity treatment for high Q0
  • Optimal bath temperature?`

U N I V E R S I T Y

CORNELL

  • 68 -

Matthias Liepe 12/12/03

slide-69
SLIDE 69

Maximum Heat Flux in a 2 K LHe II Bath

2 2.05 2.1 2.15 2.2 1 2 3 4 5 qmax for L=x cm [W/cm2] max tolerable T [K] L=5 cm tube length L=1 cm tube length L=1 m tube length L=20 cm tube length

⇒ need large enough LHe pipes

U N I V E R S I T Y

CORNELL

  • 69 -

Matthias Liepe 12/12/03

slide-70
SLIDE 70

U N I V E R S I T Y

CORNELL

Cornell ERL Cryomodule Concept: TTF Module modified for CW Operation

5 K forward 77 K forward 77 K HOM forward and return 2 K 2-phase He pipe

s.c. cavity in LHe vessel He-gas return pipe

Matthias Liepe 12/12/2003

70 K shield 77 K return 8 K return 5 K shield

  • 70 -
slide-71
SLIDE 71

C C-

  • ERL: Cryogenic Loads per Module with 10 7

ERL: Cryogenic Loads per Module with 10 7-

  • cell

cell Cavities (20 MV/m, 100 mA, Q Cavities (20 MV/m, 100 mA, Q0

0=10

=1010

10)

)

588 W/W 2 K efficiency (TESLA TDR) 213.5 kW 213.5 kW Total wall plug power 363.1 W Total 2 K 6 W HOM Couplers 15 W HOM absorber 1.6 W Input couplers 330 W Cavity RF load 2 K dynamic loads 0.03 W HOM Couplers 9.6 W HOM absorber 0.2 W Input Couplers 0.21 W

  • Instrum. cables

0.5 W Supports 2 K static loads 168 W/W 5 to 8 K efficiency (TESLA TDR) 1.7 W Radiation 33.8 kW 33.8 kW Total wall plug power 201.1 W Total 5 K 12 W HOM Couplers 150 W HOM absorber 12 W Input couplers 0.4 W Resistive wall 5 K dynamic loads 1 W HOM Couplers 19.5 W HOM absorber 1.3 W Input Couplers 1.2 W

  • Instrum. cables

2 W Supports 5 K static loads 17 W/W 77 K efficiency (TESLA TDR) 5 W Current leads 5 W Supports 38.4 W Radiation 29.6 kW 29.6 kW Total wall plug power 1742.3 W Total 77 K 60 W HOM Couplers 1335 W HOM absorber 210 W Input couplers 77 K dynamic loads 6.4 W HOM Couplers 0 W HOM absorber 65 W Input Couplers 4.5 W

  • Instrum. cables

13 W Current leads 77 K static loads

Example: 5 GeV, 20 % cryo overhead Q0 = 1·1010 ⇒ 10.3 MW wall plug power Q0 = 2·1010 ⇒ 6.7 MW wall plug power

U N I V E R S I T Y

CORNELL

  • 71 -

Matthias Liepe 12/12/2003

slide-72
SLIDE 72

R RBCS

BCS and LHe Bath Temperature

and LHe Bath Temperature

  • RBCS decreases significantly if T is lowered.
  • Important: The residual resistance must be low to

make use of this! ⇒ Very good shielding of Earth’s magnetic field (< some mOe).

  • Examples:
  • T = 2.0 K ⇒ Q = 2.6·1010
  • T = 1.8 K ⇒ Q = 6.3·1010
  • T = 1. 6K ⇒ Q = 1.9·1011
  • A dream?…

U N I V E R S I T Y

CORNELL

  • 72 -

Matthias Liepe 12/12/03

slide-73
SLIDE 73

High Q High Q0

0 : Let

: Let’ ’s dream s dream… …

⇒ 2 W/m losses at 1.6 K instead of 20 to 40 W/m losses at 2 K?

U N I V E R S I T Y

CORNELL

  • 73 -

Matthias Liepe 12/12/03

slide-74
SLIDE 74

Outlook and Summary: SRF @ Cornell ERL

  • ERLs have the potential to produce photon beams with ultra high

brilliance, coherence and ultra short pulses.

  • Cornell is proposing to built an 100 MeV, 100 mA ERL prototype

and we study a 5 GeV SR ERL light source.

  • SRF is a key technology for these machines.
  • Many challenges need our attention:

HOM damping, emittance preservation, RF field control, cw cavity

  • peration at high fields, high power input couplers, ...
  • We have started to work in these areas and hope to start with the

construction of the injector soon.

  • Next year: first 2-cell cavity, first 50 kW input coupler, first HOM

ferrite ring-absorber, injector module design.

U N I V E R S I T Y

CORNELL

  • 74 -

Matthias Liepe 12/12/2003

slide-75
SLIDE 75

Et facta est lux. (And there was light.)