Low emittance light sources - PowerPoint Presentation

1. Low emittance light sourcesAndreas StreunSwiss Light Source (SLS)Paul Scherrer Institut (PSI) 5232 Villigen, Switzerland“Forum Beschleunigerphysik” WorkshopPerspectives for Accelerators and TechnologyDarmstadt, Feb. 16-17, 2017

2. OutlineI. Storage ring based light sourcesRadiation equilibriumPaths to low emittanceThe multibend achromat (MBA)MBA technologyThe new generation of light sourcesThe dynamic aperture challengeMBA variants and examplesTrends in storage ring designII. Linac based light sources FELSource emittanceEmittance preservation

3. I. Storage ring based light sourcesHorizontal emittance in electron storage ring determined by radiation equilibriumindependent of initial conditionshow to maximize this -- and -- minimize this ?

4. Minimum equilibrium emittanceMaximum radiation dampingincrease radiated power  pay with RF-powerHigh field bending magnets quantum excitation higher too Damping Wigglers (DW): S |deflection angles| > 360Minimum quantum excitationkeep off-momentum orbit close to nominal orbitminimize dispersion at locations of radiation (bends) Horizontal Focusing into bends to suppress dispersion.Multi-Bend Achromat lattice (MBA)many short (= small deflection angle) bends to limit dispersion growth.Longitudinal Gradient Bend (LGB)highest radiation at region of lowest dispersion and v.v.d =Dp/p

5. Emittance scalingHorizontal emittance(for a flat iso-magnetic lattice without damping wigglers)ex  1/6 pm (E [GeV])2 (f [°] )3 F many (n) small dipoles: f = 360°/n focus to magnet center: F  3...6  Fmin end bend center bend longitudinal gradientdispersionBeam energyDeflection angleper dipoleBeam opticsVertical emittance (of a flat lattice)equilibrium emittance small by nature ey < 1 pmdetermined by lattice imperfections ey  1...10 pm 3rd generation (  2015, NSLS-II ) ex  2’000...10’000 pm new generation (  2015, MAX IV) ex  100...400 pm

6. Paths to low emittanceEquilibrium beam parameters of a flat latticeexo natural horizontal emittancesd rms relative momentum spread, d = Dp/pDE energy loss per turnI2 I3 I4 I5 synchrotron radiation integrals (pure lattice properties) constantsto do:I5 (I3 )I2  ?I4  -I2

7. The beam optics pathH = dispersion invarianth, h’ dispersion and slope, b hor. beta function, a = -b’/2 H d 2 = betatron amplitude of particle after momentum change dh = orbit curvature, h = 1/r = B/(p/e) keep dispersion (h, h’) small in dipoles (h  0) !  horizontal focus in each dipole  MBA: small bend angle f = hL , so h can’t grow (h’’ = h)  LGB: longitudinal field variation B (s): compensate H (s) growth by h (s) hdh’dx’x

8. The power pathh = orbit curvature, h = 1/r = B/(p/e) increase radiation loss DE ~ I2  expensive for high energy rings !  high dipole field  large momentum spread sd2 ~ h  Damping Wigglers: length Lw, period lw, peak field Bw = (p/e) hw wiggler: long, short period, weak field, horizontal focusPETRA 3 : Power 1.1 4.9 MW  exo 4.4  1.0 nm

9. The damping partitioning pathh = 1/r = B/(p/e) curvature, b2 = B’/(p/e) focusing strengthDamping partitioning numbers Jx = 1 - I4/I2 Jd = 2 + I4/I2  Jx + Jd = 3separate functions ( b2 = 0 in bending magnets ): |I4| << I2  Jx  1, Jd  2 get ½ emittance on expense of 2 larger momentum spread:  vertical focusing (b2 < 0, |b2| >> h2) in bends (h > 0, h > 0) e.g. MAX IV Jx = 1.85  horizontal focusing (b2 > 0) in anti-bends (h < 0, h > 0)

10. The TME cell what is the lowest possible emittance of a lattice cell ?homogenous (constant h), short (f =hL << 1) bending magnetset ao = ho’ = 0 at bend center (symmetry); find minimum H ( bo, ho ) : Theoretical Minimum Emittance (TME) for periodic symmetric cell: a = h’ = 0 at endsmatching problemTME phase advancem min =284.5° 2nd focus, uselesslong celloverstrained opticsbx by hf, L, hindependent of bending magnet field !

11. Deviations from TME conditionsEllipse equationsfor emittanceCell phase advance real cells: m < 180°  F ~ 3...6Relaxed TME cellsELETTRA: F = 4.1( this is for a periodic cell, the dispersion suppressor cell is different, with Fmin = 3 ) results independent of particular cell design!

12. The Multi-Bend Achromat conceptrelax optics: allow large Flow phase advance mexploit e  f 3 dependencemany small angle bending magnets, many lattice cellsMAX IV: m ~ 78°, b ~ 6, d ~ 10, F ~ 15, f = 3°  e = 328 pm Miniaturizationsmall bend angle f  small dispersion hminimum aperture for momentum acceptance(with regard to Touschek lifetime)MAX IV: 20  7BA arcsmove here! D. Einfeld & M. Plesko,NIM A 335 (1993), 402-416

13. The multi-bend achromat optimization cycleM. Eriksson et al., Some small-emittance light-source lattices with multi-bend achromats, NIM A 587 (2008) 221. Miniaturization of components until reaching physical or technical limits

14. MBA technology: vacuumBeam pipe cross section:6532 mm2   22 mmMiniaturization  3 smaller beam pipeUltra high vacuum requirements: 10-12 barCoulomb scattering beam lifetimeBremsstrahlung background to experimentsNEG* coating of small vacuum chambers.beam pipe = getter pumplow radiation desorptionMAX IV, SIRIUS et al.large resistive wall impedance  beam instabilitiesAlternative: ante-chamber and discrete pumps ante chamber required for radiation anywaycopper [plated] beam pipenew generation of compact getter pumpsESRF-EBS, SLS-2 et al. S. Dos Santos, LER-WS, Frascati, 2014 SLS-2 vacuum chamber  100 l/s NEXTorr® pump MAX IV vacuum chamber*Non Evaporable Getter 1 mm Ti-Zr-V layerwww.saesgetters.com

15. MBA technology: magnetssmall gap/bore magnets (bore radius R)high gradients: 2n-pole momentincreasing use of permanent magnetslittle flexibility required (one energy, one mode)dense lattice: no space for coilssave cabeling and electric energynew magnet types: longitudinal gradient dipoles, octupoles etc.ESRF-EBS quadrupole90 T/m, R = 12.5 mmESRF-EBS permanentmagnet dipoleSIRIUS permanent magnet 3.2 T superbend,  F. H. de Sa, LER-WS, SOLEIL, Oct. 2016 D. Einfeld, LELD-WS, Lund 2016

16.  S. C. Leemann at al., PRST AB 12, 120701MBA technology: alignmentTolerances:  50...100 m  10..50 mCombination of magnets in blocks MAX IV magnet block Precise reference surface on girders ESRF-EBS girder Dynamic alignments systemsgirder mover motors & encodershydrostatic leveling systemsdigital encoder based sensor systemsgirder alignement with stored beam SLS girder Extensive use of beam based alignmentand optics correction methods (BBA, LOCO etc.)required for commissioning!

17. The old (—) and the new (—) generation.The storage ring generational changeHorizontal emittance normalized to beam energyFigure from P. Raimondi, LELD WS, Barcelona, April 2015SLS-2ALS-U

18. New rings and upgrades (incomplete list)NameEnergy [GeV]Circumf. [m]Emittance* [pm]StatusPETRA-III(ex-HEP) PETRA-IVDE6.03.06.023044400  1000285 (round beam)2(10-30) (round beam)operationalstudyMAX-IV SE3.0528328  200operationalSIRIUS BR3.05182402019ESRF-EBS EU6.08441352020APS-U US6.0110441-90advancedSPRING-8 II JP6.01436140advancedHEPSCN6.0129660studyPEP-X(ex-HEP)US4.5220029  10(study)DIAMOND-II UK3.0562125studySLS-2 CH2.4 290140studyELETTRA 2.0 IT2.0259250studyALS-U US2.0197250 (round beam)study*Emittance (no IBS) without  with damping wigglers

19. MBA challenge: dynamic apertureChromaticity correctionminimize chromatic tune footprint: MBA require 5% momentum acceptancesuppression of head-tail instability: all modes suppressed for x = 0MBA lattice: small dispersion h  strong sextupoles (b3)  reduced dynamic aperture  reduced lattice momentum acceptance  reduced beam life time (Touschek scattering)  reduced transverse acceptance  off-axis injection may become impossibleLocal aperture = beam pipePhysical aperture = “shadow” of local apertures[linear] projection of local apertures to track pointDynamic aperture = area of stable particle motionobtained from tracking including local aperture limitsyx

20. MBA variantsMBA design concepts w.r.t. dynamic apertureLarge circumference (MAX IV, SIRIUS, PETRA-IV, PEP-X...): very many cells, relaxed optics  moderate chromaticity Hybrid-MBA (ESRF-EBS  APS-U, HEPS, DIAMOND II ...): dispersion bumps (local high h) to place sextupoles Longitudinal gradient / anti-bend - MBA (SLS-2): optimum emittance matching in compact lattice Cope with reduced dynamic aperture (ALS-U): swap-out on-axis injection and bunch lengthening in any case: extensive dynamic aperture optimization procedures lattice design: align sextupole betatron phases to cancel adverse effects...semi-analytic: minimization of perturbation terms up to any order...numeric: optimization of dynamic apertures from tracking (MOGA etc.)...

21. Large circumference ring: PEP-XPEP: HEP collider. PEP-II: B-factory 1999-2008PEP-X: 2.2 km ring @ 4.5 GeV, 11 pm emittance ! MAX-IV type7-bend achromat: 6 x 8 arcs long straights from HEP times: places for damping wigglers large bx  200 m for off-axis injection emittance 29 pm, with damping wigglers: 11 pmLattice courtesy Y. Cai, SLAC48  7BAY. Cai et al., PRST AB 15, 054002 (2012)

22. Hybrid-MBA: ESRF-EBSdispersion bumps to place chromatic sextupolescancellation of most sextupole terms by phase advance (-I transformer)no sextupoles in central sectionall beam line source points maintained in upgradeconcept also used for APS-U, Spring-8-II, HEPS , Diamond-II [, ALS-U] J.-L. Revol, LER WS, Frascati, 2014Dmx = 3p Dmy = p ESRF-EBS

23. Longitudinal gradient / anti-bend cell: SLS-2bx by Dispersion hfor q = 0.78°for q = 0°dipole fieldquad fieldtotal |field| }R = 13 mmAB TGB LGB TGB ABsextupolestransverse gradient bends (combined function)LGB longitudinal variation of curvature h(s) to minimizeAB angle q < 0 anti-bend provides optimum matching (ho 0) for LGBAB-angle q -0.78° 0° Emitt. [pm] 135 535Damp. Jx 1.80 1.15E- loss [keV] 7.9 5.8MCF a [104] 2.1 +4.8S|bend angles| > 360 high peak field in LGBvert. foc. in TGB (b2< 0)horiz. foc. in AB (b2 > 0) A. Streun, LELD-2 WS, Lund, Dec. 2016

24. Optimization of field profile By(s) for fixed b0, h0 Emittance (F) vs. b0, h0 normalized to parameters for the “theoretical minimum emittance” of a homogenous bend.Emittance reduction in LGB/AB cellF = 1F = 2F = 2F = 3F = 3F = 1LGB requires small (~0) dispersion at centre, but tolerates large beta function !F  0.3 bbddusual operating region for TME cells135°180°225°phase advancein TME cellHomogeneous Bending MagnetLongitudinal Gradient BendABLGB constant field

25. On-axis swap-out: ALS-U9BA lattice for 100 pm at 2 GeV dynamic and physical apertures too small for off-axis injection additional accumulator ring for on-axis swap-out injectionlow momentum acceptance  low Touschek lifetimeharmonic RF system for bunch lengtheningmitigation of intrabeam scatteringincrease of instability threshold current frequent ( 30 sec) swap-out injections D. Robin, M. Venturini, LER-WS, Soleil, Oct. 2016

26. Trends New injection schemesoff-axis multipole kicker, on-axis longitudinal, on-axis swap-out etc.new injection elements: fast kickers, nonlinear kickers, thin septa etc. Permanent magnet designshigh gradients, no cables, zero powerlimited tunability, radiation damage, temperature stability – o.k. Vacuum technologyNEG coating of small beam pipes; ante-chambers & small pumps etc. Multiple harmonic RF systemslong bunches required to mitigate Touschek losses, IBS and instabilities. Precise and interactive alignment systems beam based alignment is part of commissioning. (NBEND)-3  Miniaturization  densely packed lattice  small but strong magnets  small dynamic aperture  small beam pipe  small physical aperture  large impedance

27. II. Linac based light sources  FELQualitative summary with some data from SITF (SwissFEL test facility) T. Schietinger et al., PR AB 19, 100702 (2016) comprehensive report on SITF commissioning and emittance minimization

28. Free electron lasermost basic FEL relationsQ bunch charge; LSAT, PSAT saturation length and power for SASE-FELFEL = diffraction limited source per se : l  1Å  e < 40 pm relevant: slice emittance, not projected emittance.small emittance e & short pulse length ss & low E-spread seno use of damping ringslimitation of particle density in bunch due to intra beam scattering and turbulent bunch lengtheningtoo large energy spread ( 0.1%)emittance reduction only by adiabatic damping  accelerationnormalized emittance stays constant eN = bg esource intrinsic emittance  electron gun design (70%) emittance degradation during transport and acceleration (30%)

29. Electron source emittanceElectron gun type: DC static, RF multi-cell, DC pulsedCathode type: thermionic, photo effect, field emitterintrinsic emittance  laser spot size  (electron kinetic energy)½electron kinetic energy = laser photon energy + Schottky effect  field at cathode (metal cathode) – work function (semiconductor) – (band gap + electron affinity)cathode: CsTe coated Cu vs. pure coppersame intrinsice normalized emittance eN  100 nm for Q = 200 pCbetter quantum efficiency – but less robust.slower response  smoothing of laser pulse temporal substructure  prevents microbunching in bunch compression.20% fluctuation for different cathode specimens.

30. Emittance preservationSources of emittance (& energy spread) degradation:space charge (at low energy)fast acceleration, high gradient (85 MV/m at cathode).linear forces  homogeneous laser pulse (radial & temporal).compensation (“Carlsten scheme”)  solenoid & drift space (3 m)wakefields centering of beam in linac, etc.coherent synchrotron radiation low dipole deflection angles  long buncher chicanefocusing of beam in last dipole of buncher chicaneminimize chromaticity of energy-chirped beam SITF results for sliced [projected] normalized emittancesbunch charge 200 pC 10 pCintrinsic source emittance 100 nm 40 nmafter acceleration to 250 MeV 200 [300] nm 100 [150 ] nmafter compression 200 [400] nmSwissFEL: e  20 pm at 5.8 GeV 

31. SummaryFELs[normalized] emittance determined by sourcestate-of-the-art close to theoretical limits1 Å FEL operational lower emittance ? longer linacs (€)  higher gradients - novel acceleration techniques Storage ringsemittance function of latticeongoing “MBA-revolution” most projects not yet diffraction limited: e  10 pmlower emittance ? larger rings (€)  further miniaturizationinjection schemes and elementspermanent magnetsdynamic alignmentTHE END