S. Prestemon FLS-2010 1

Superconducting undulator options for x-ray FEL applications

Soren Prestemon &

Ross Schlueter

3/1/2010

S. Prestemon FLS-2010 2

Outline

• Basic undulator requirements for FEL’s• Superconducting undulators:

– Superconductor: options and selection criteria– Families by polarization

• Circular• Planar• Variable polarization

– Performance comparison/characteristics• Integration issues

– Spectral scanning rates, field quality correction– Cryogenics

• R&D needs

3/1/2010

S. Prestemon FLS-2010 3

Acknowledgments

Magnetic Systems Group:Ross Schlueter, Steve Marks, Soren Prestemon,

Arnaud Madur, Diego Arbelaez

With much input fromThe Superconducting Magnet Group, Center

for Beam Physics, andThe ALS Accelerator Physics Group

3/1/2010

S. Prestemon FLS-2010 4

Basic undulator requirements for X-ray FELS

• Variable field strength for photon energy tuning– Beam energy and undulator technology must be matched

to provide spectra needed by users– Sweep rate, field stability and reproducibility

• Variable polarization (particularly for soft X-rays)– Variable linear and/or elliptic – Rate of change of polarization

• Field correction capability– Compensate steering errors– Compensate phase-shake

3/1/2010

S. Prestemon FLS-2010 5

Beam energy, spectral range, and undulator performance

3/1/2010

Only for planar undulators

Regime of interest

• For any given technology:– At fixed gap, field increases

with period– Field drops as gap increases

=> Choice of electron energy is closely coupled to undulator technology, allowable vacuum aperture, and spectrum needed

Technology-driven

S. Prestemon FLS-2010 6

Superconductors of interest

• Application needs:– Hi Jc at low field– Low magnetization (small filaments)– Larger temperature margin

3/1/2010

2015

1015

10

20

510 3

5

10 4

10 5

10 6

10 7

tem perature(K )

current density(A /cm )2

N b Sn3

N b-Ti

m agnetic fie ld(T)

critical J-H -Tsu rface

Arno Godeke, personal communication

• ~1 micron YBCO layer carries the current

• Critical temperature ~100K

– 12mm wide tape carries ~300A at 77K

– factor 5-15 higher at 4.5K, depending on applied field

Nb3Sn NbTi

Superconducting materials

Plot from Peter Lee, ASC-NHMFLRegime of interest for SCU’s

S. Prestemon FLS-2010 8

Superconducting undulators

• The first undulators proposed were superconducting – 1975, undulator for FEL

experiment at HEPL, Stanford– 1979, undulator on ACO– 1979, 3.5T wiggler for VEPP

Rev. Sci. Instr., 1979

Ancient history

3/1/2010

S. Prestemon FLS-2010 9

Bifilar helical

• Provides left or right circular polarized light• Continuous (i.e. maximum) transverse acceleration of

electrons• Fabrication

– With or without iron– Coil placement typically dictated by machined path

3/1/2010

S. Caspi

D. Arbelaez, S. Caspi

S. Prestemon FLS-2010 10

Performance• Bifilar helical approaches yield excellent performance:– applicable for “short” periods, λ>~10 (7?) mm, gap>~3-5mm

• wire dimensions, bend radii, and insulation issues– well-known technology (e.g. Stanford FEL Group, 1970’s), but not “mature”– most effective modulator for FEL

• need to consider seed-laser polarization

3/1/2010

Assume Je=1750A/mm2, no Iron

S. Prestemon FLS-2010 11

Planar SCU’s

• “Traditional” approach:– Different methods for coil-to-coil

transitions

• Can use NbTi or Nb3Sn– BNb3Sn/BNbTi~1.4

• HTS concept:– “Winding” defined by lithography– Use coated conductors

• YBCO is best candidate• Use at 4.2K

3/1/2010

Electron beam

• Current at edges largely cancels layer-to-layer; result is “clean” transverse current flow

Soren Prestemon 12July 26, 2006

Performance considerationsMotivation for Nb3Sn SCU’s over NbTi

• Motivation for Nb3Sn– Low stored energy in magnetic system

• “break free” from Jcu protection limitation– Take advantage of high Jc, low Cu fraction in Nb3Sn– “High” Tc (~18K) of Nb3Sn

• provides temperature margin for operation with uncertain/varying thermal loads

S. Prestemon FLS-2010 13

Performance: “Traditional” Planar SCU’s

• Nb3Sn yields 35-40% higher field than NbTi (at 4.2K)– “Raw” performance has been demonstrated at LBNL, with

a 14.5mm period prototype

3/1/2010

Performance curves (calculated)

HTS conceptHybridPMEPU

Gap=2, 3mm

• Issues considered:– Width of current path - assumed ~1mm laser cuts separating “turns”– Finite-length of straight sections – 83% retained for g=2mm, 12mm wide tape– Gap-period region of strength – most promising in g<3mm, λ<10mm regime– Peak field on conductor & orientation - <~2.5T

• The HTS short period technology compared to PM and hybrid devices:

– Scaling shows regions of strength of different technologies– Assumed Br=1.35 for PM and hybrid devices– Data shown for HTS assumes J=2x105A/mm2, independent of

field• for B>~1.5T, scaling needs to be modified to include J(B) relation

HTS low CuHTS baseline

Hybrid PM

Pure PM

Helical

HTS: 2-2.2mm gapHelical: 3-3.2mm gap, 2kA/mm2

IVID PM: 2-2.2mm gap

S. Prestemon FLS-2010 15

Variable polarization

• Critical for many experiments, particularly in soft X-rays– Photoemission, magnetism (e.g dichroism)

• Variety of parameters define polarization capability– Type and range of polarization control (variable linear,

variable elliptic; spectrum range vs polarization)– Speed at which polarization can be varied

3/1/2010

S. Prestemon FLS-2010 16Soren Prestemon, LBNL ALS SAC meeting, June 24, 2009

Existing PM-EPU vs Conceptual SC-EPU

No iron in SC-EPU-strengths:-Period doubling-No moving parts

Variable polarization capabilities

3/1/2010

S. Prestemon FLS-2010 17Soren Prestemon, LBNL ALS SAC meeting, June 24, 2009

Variable polarization

• Consider a 4-quadrant array of such coil-series.

– If IC=-IA, Coils A and C generate additive –fields.

– Set IC=-IA, ID=-IB; Independent control of IA and IB provides full linear polarization control.

IB IA

IC ID

Beam

For IA=IB=IC=ID:

ψ

Independent control of IA and IB provides variable linear polarization control

- If IA=IB, vertical field, horizontal polarization- If IA=-IB, horizontal field, vertical polarization

BA

3/1/2010

S. Prestemon FLS-2010 18Soren Prestemon, LBNL ALS SAC meeting, June 24, 2009

Superconducting EPU• Add a second 4-quadrant array of such coil-series,

offset in z by λ/4 (coil series α and β)• With the following constraints the eight currents are

reduced to four independent degrees of freedom:

• The α and β fields are 90° phase shifted, providing full elliptic polarization control via C

D

3/1/2010

S. Prestemon FLS-2010 19Soren Prestemon, LBNL ALS SAC meeting, June 24, 2009

Broad spectral range of SC-EPU

• Separating the coils in the α (and β) circuit into two groupings allows for period-halving:

(variable linear, no elliptic)

• Going further… separating the coils in the α1 (and α2, β1, β2) circuit into two groupings allows for period doubling:

Full polarization control

Period-halved linear polarization control

Period-doubled full polarization control

(Full polarization control)

NOTE: Two power supplies (A, B) needed for linear polarization control; four needed for full (linear+elliptic) polarization control; switching network could provide access to the above regimes

3/1/2010

S. Prestemon FLS-2010 20Soren Prestemon, LBNL ALS SAC meeting, June 24, 2009

Nb3Sn superconductor, 24% superconductor in coil-pack cross-section, 90% of Jc, vacuum gap=5 mm

(magnetic gap=7.3 mm for PM-EPU, 6.6 mm for SC-EPU), Br=1.35 T for PM material; block height and width fixed.

Elliptically polarizing undulators

3/1/2010

S. Prestemon FLS-2010 21

Integration issues

• Field correction– Want no beam steering, no beam displacement– Must minimize phase-shake

• Wakefields– What are limitations in terms of bunch stability?– Image current heating: impact on SCU’s

• Modular undulator sections– Allows focusing elements between sections– Requires phase shifters

3/1/2010

S. Prestemon FLS-2010 22

Field correction

• PM systems use “virtual” or magnetic shims• SCU correction methods (proposed):

– Trim “coils”: located on each/any poles• Amplitude of correction (~1%) has been demonstrated at LBNL• Individual control is possible, but becomes complex• Experience with PM devices suggests few “coils” can provide requisite correction =>

locations of corrections determined during undulator testing off-line• Mechanism to direct current using superconducting switches has been tested

– Passive “shims” (ANKA): use closed SC loop to enforce half-period field integral• Should significantly reduce RMS of errors• Some residuals will still exist due to fabrication issues• Possibility of hysteretic behavior from pinned flux – needs to be measured under

various field cycling conditions

3/1/2010

Detailed tolerance analysis is needed to determine amount/type of correction that may be required. Preliminary data (e.g. APS measurements) suggest fabrication errors are smaller than typically observed on PM devices

S. Prestemon FLS-2010 23

Superconducting switches

• Allow active control of current (+/-/0) to each shim coil from one common power supply– Switch produces negligible heat at 4.K while controlling high currents– Can be used to control period-doubling in SC-EPU concept

3/1/2010

Superconducting switches and shim. The current path can be set by combining the switches.

S. Prestemon FLS-2010 24

Passive shimming

• Passive scheme – does not have/need external control– Will compensate errors independent of error source– Assumes “perfect conductor” model for superconductor

• Pinned (i.e. trapped) flux may yield some hysteresis – needs measurements

3/1/2010

D. Wollman et al., Physical Review Special Topics-AB, 2008

S. Prestemon FLS-2010 25

Measurements

• Any field correction depends on ability to measure fields with sufficient accuracy– “traditional” Hall probe schemes not applicable– Need system compatible with cryogenic temperatures:

• System must work with integrated vacuum chamber• Hall probe “on a stick” or “pull”:

– most common and basic approach;– suffers from uncertainty in knowledge of Hall probe location– Could use interferometry to determine location– Could use Hall probe array to provide redundancy to compensate spatial uncertainty

• Pulsed wire: – need to demonstrate sufficient accuracy– benefits from vacuum for reduced signal noise

• In-situ:– Use electron beam=>photon spectrum as field-quality diagnostic– Fourier-transform – loss of spatial information – recoverable?

3/1/2010

Soren Prestemon 26July 26, 2006

Cryogenic design options

• Can use liquid cryogens or cryocoolers– Liquid cryogen approach requires liquifier + distribution system or user refills– Cryocoolers require low heat load and (traditionally) incur temperature gradients through conduction

path and impose vibrations from GM cryocooler• Limits operating current due to current-lead heat load (despite HTS leads; typical limit is <1kA)• Solution: heat pipe approach (C. Taylor; M. Green)

• Need to know the heat loads under all operating regimes

Aggressive spacings:

Dw~0.75mm

Dgv~1mm

Dgv

20-60K

Dw

Yoke

Vacuum chamber

4.2-12K

•Vacuum chamber and magnet can be thermally linked; magnet and chamber operate at 4.2-8K

•Vacuum chamber and magnet can be thermally isolated; chamber operates at intermediate temperature (30-60K); magnet is held at 4.2K

M. Green, Supercond. Sci. Tech.16, 2003M. Green et al, Adv. in Cryogenic Eng., Vol. 49

Expected for FEL applications

Soren Prestemon 27July 26, 2006

Beam heating impact on performance: Example of ALS

0 2 4 6 8 10 12 14 160

1

2

3

4

5

Assumes Asc

/Atot

=0.25, with no Jc margin. Based on existing Nb

3Sn material Jc data.

Performance evaluated for 4.2K, 5K, 6K, 7K, 8K

15mm period

20mm period

25mm period

30mm period

Pea

k ax

ial f

ield

[T]

Magnetic gap [mm]

Dgv

20-60K

Dw

Yoke

Vacuum chamber

4.2-12K

Intermediate intercept model

Cold bore model

0T(Q) T +aQ

02.51static imQ Q Q Qh

Ref: Boris Podobedov, Workshop on Superconducting Undulators and Wigglers, ESRF, June, 2003

2 2 / 3 1/ 3( )05 / 3( )im e

lI sQ Zh lb

α λ

• In synchrotron rings, image current heating impacts design• In FEL’s, low duty-factor typically implies low image currents

→ Other heating sources will dominate

Cold, extreme anomalous skin effect regime:ALS: ~ 2 W/mLCLS: ~ 3.e-4 W/m

Principal SCU challenges/Readiness

• Principal challenges – Fabrication of various SCU design types– vacuum, wakefields, heating -> acceptable gap?– Shimming/tuning– Cold magnetic measurements

• Readiness– Prototypes: three SCU LBNL prototypes; ANL prototypes– Concepts: for SC-EPU, stacked HTS undulator & micro-

undulators, Helical SCU’s

Undulator R&D plan

• SCU – NbTi and subsequently Nb3Sn-based planar and bifilar helical– demonstrate reliable winding, reaction, & potting process for Nb3Sn– develop trajectory correction method– magnetic measurements

• Stacked HTS undulator :– demonstrate effective J (i.e. B)– evaluate image-current issues– determine field quality / trajectory drivers– current path accuracy, J(x,y) distribution– accuracy of stacking– develop field correction methods [consider outer layer devoted to field correction (ANKA passive shim)]

Undulator R&D plan, cont.(initial cut- undulator R&D list)

• Stacked HTS Micro-undulator– demonstrate ability to fabricate layers– demonstrate effective J (i.e. B)– evaluate image-current issues

• SC-EPU– develop integrated switch network– Demonstrate performance

• All SCU concepts:– Detailed tolerance analysis– Need reliable measurements