Swiss Light Source The next 20 years Andreas Streun Paul Scherrer Institut (PSI) Villigen, Switzerland Future Research Infrastructures: Challenges and Opportunities Varenna, Italy, July 8-11, 2015

Outline Portrait of the SLS; history and achievements The new generation of light sources The challenge to upgrade the SLS A new type of lattice cell for lower emittance: longitudinal gradient bends and anti-bends SLS-2 design: performance, challenges, highlights

Paul Scherrer Institut (PSI) 1960 Eidgenössisches Institut für Reaktorforschung (EIR) 1968 Schweizer Institut für Nuklearphysik (SIN) 1988 EIR + SIN = PSI  research with photons, neutrons, muons PSI Accelerators: 590 MeV proton cyloctron: 1.3 MW beam power  spallation neutron source SINQ & muon source SmS 5.8 GeV / 1 Å free electron laser SwissFEL: operation 2017 2.4 GeV synchrotron light source SLS

The SLS Electron beam cross section in comparison to human hair transfer lines 90 keV pulsed (3 Hz) thermionic electron gun 100 MeV pulsed linac Synchrotron (“booster”) 100 MeV  2.4 [2.7] GeV within 146 ms (~160’000 turns) Current vs. time 2.4 GeV storage ring ex = 5.0..6.8 nm, ey = 1..10 pm 400±1 mA beam current top-up operation 1 mA 4 days shielding walls

SLS: beam lines overview

SLS: history Dec. 2000 First Stored Beam July 2001 First experiments 1990 First ideas for a Swiss Light Source 1993 Conceptual Design Report June 1997 Approval by Swiss Government June 1999 Finalization of Building Dec. 2000 First Stored Beam June 2001 Design current 400 mA reached Top up operation started July 2001 First experiments Jan. 2005 Laser beam slicing “FEMTO” May 2006 3 Tesla super bends 2010 ~completion: 18 beamlines

SLS achievements Rich scientific output Reliability > 500 publications in refereed journals/year four spin-off companies (e.g. DECTRIS) Reliability 5000 hrs user beam time per year 97.3% availability (2005-2014 average) Top-up operation since 2001 constant beam current 400-402 mA over many days Photon beam stability < 1 mm rms (at frontends) ‏fast orbit feedback system ( < 100 Hz )‏ undulator feed forward tables, beam based alignment, dynamic girder realignment , photon BPM integration etc... Ultra-low vertical emittance: 0.9 ± 0.4 pm model based and model independent optics correction high resolution beam size monitor developments 150 fs FWHM hard X-ray source FEMTO laser-modulator-radiator insertion and beam line

The storage ring generational change Riccardo Bartolini (Oxford University) 4th low emittance rings workshop, Frascati , Sep. 17-19, 2014 Horizontal emittance normalized to beam energy Storage rings in operation (•) and planned (•). The old (—) and the new (—) generation.

New storage rings and upgrade plans Name Energy [GeV] Circumf. [m] Emittance* [pm] Status PETRA-III 6.0 3.0 2304 4400  1000 85 (round beam) operational MAX-IV 3.0 528 328  200 2015 SIRIUS 518 280 2016 ESRF upgrade 6.0 844 147 2020 DIAMOND upgrade 562 275 started APS upgrade 1104 65 study SPRING 8 upgrade 1436 68 PEP-X 4.5 2200 29  10 ALS upgrade 2.0 200 100 ELETTRA upgrade 260 250 SLS now 2.4 288 5020** SLS-2 2.4 (?) 100-200 ? 2024 ? *Emittance without  with damping wigglers **without FEMTO insertion

The Multi-Bend Achromat (MBA) Miniaturization small vacuum chambers [NEG coated] high magnet gradients more cells in given circumference

SLS upgrade constraints and challenges get factor 20...50 lower emittance (100...250 pm) keep circumference & footprint: hall & tunnel. re-use injector: booster & linac. keep beam lines: avoid shift of source points. “dark period” for upgrade 6...9 months Main challenge: small circumference (288 m) Multi bend achromat: e  (number of bends)─3 Damping wigglers (DW): e  radiated power Low emittance from MBA and/or DW requires space ! Scaling MAX IV to SLS size and energy gives e  1 nm  New lattice concept  e  100...200 pm  ring ring + DW

Theoretical minium emittance (TME) cell dilemma Conditions for minimum emittance (h = 1/r = eB/p curvature) periodic/symmetric cell: b ’ = h’ = 0 at ends  over-focusing of bx  phase advance m min =284.5° 2nd focus, useless overstrained optics, huge chromaticity... long cell better have two relaxed cells of f/2 MBA concept... bx by h f, L, h

Conventional cells = relaxed TME cells Deviations from TME conditions Ellipse equations for emittance Cell phase advance  Real cells: m < 180°  F ~ 3...6 MBA: F > 10

how to do better ? disentangle dispersion h and beta function bx release constraint: focusing is done with quads only. use “anti-bend” (AB) out of phase with main bend suppress dispersion (ho  0) in main bend center. allow modest bxo for low cell phase advance. optimize bending field for minimum emittance release constraint: bend field is homogeneous. use “longitudinal gradient bend” (LGB) highest field at bend center (ho = (e/p) Bo) reduce field h(s) as dispersion h(s) grows sub-TME cell (F < 1) at moderate phase advance

step 1: the anti-bend (AB) General problem of dispersion matching: dispersion is a horizontal trajectory dispersion production in dipoles  “defocusing”: h’’ > 0 Quadrupoles in conventional cell: over-focusing of beta function bx insufficient focusing of dispersion h  disentangle h and bx use negative dipole: anti-bend kick Dh’ =  , angle  < 0 out of phase with main dipole negligible effect on bx , by dispersion: anti-bend off / on bx by relaxed TME cell, 5°, 2.4 GeV, Jx  2 Emittance: 500 pm / 200 pm

AB emittance effects AB emittance contribution h is large and constant at AB  low field, long magnet Cell emittance (2AB +main bend) main bend angle to be increased by 2| |  in total, still lower emittance AB as combined function magnet Increase of damping partition Jx vertical focusing in normal bend horizontal focusing in anti-bend. horizontal focusing required anyway at AB  AB = off-centered quadrupole  half quadrupole  Disp. h bx by

AB impact on chromaticity Anti-bend  negative momentum compaction a  Head-tail stability for negative chromaticity! small large < 0 negative side note: AB history 1980’s/90’s: proposed for isochronous rings and to increase damping - but  PAC 1989

step 2: the longitudinal gradient bend (LGB) Dispersion’s betatron amplitude Orbit curvature Longitudinal field variation h(s) to compensate H (s) variation Beam dynamics in bending magnet Curvature is source of dispersion: Horizontal optics ~ like drift space: Assumptions: no transverse gradient (k = 0); rectangular geometry Variational problem: find extremal of h(s) for numerical optimization h(s) = B(s)/(p/e)

LGB numerical optimization Half bend in N slices: curvature hi , length Dsi Knobs for minimizer: {hi}, b0, h0 Objective: I5 Constraints: length: SDsi = L/2 angle: ShiDsi = F/2 [ field: hi < hmax ] [ optics: b0 , h0 ] Results: hyperbolic field variation (for symmetric bend, dispersion suppressor bend is different) Trend: h0   , b0  0 , h0  0 Results for half symmetric bend ( L = 0.8 m, F = 8°, 2.4 GeV ) optimized hyperbola fit homogeneous I5 contributions

LGB optimization with optics constraints Numerical optimization of field profile for fixed b0, h0 Emittance (F) vs. b0, h0 normalized to data for TME of hom. bend F = 1 F = 2 F = 3 F  0.3 small (~0) dispersion at centre required, but tolerant to large beta function

The LGB/AB cell } at R = 13 mm Conventional cell vs. longitudinal-gradient bend/anti-bend cell both: angle 6.7°, E = 2.4 GeV, L = 2.36 m, Dmx = 160°, Dmy = 90°, Jx  1 conventional: e = 990 pm (F = 3.4) LGB/AB: e = 200 pm (F = 0.69) Disp. h Disp. h bx by bx by dipole field quad field total |field| longitudinal gradient bend } at R = 13 mm anti-bend

SLS-2 lattice layout TBA  7BA lattice: ½ + 5 + ½ cells of LGB/AB type SLS-2 arc SLS arc TBA  7BA lattice: ½ + 5 + ½ cells of LGB/AB type periodicy 3: 12 arcs and 3 different straight types: 6  4 m  6  2.9 m 3  7 m  3  5.1 m split long straights: 3  11.5 m  6  5.1 m beam pipe: 64 mm x 32 mm   20 mm  magnet aperture  26 mm

SLS-2 lattice db02l (one superperiod = 1/3 of ring) optics and magnetic field  (field at poletips for R = 13 mm) Superbends in arcs 2/6/10 3  2.9 T  3  5.0 T

SLS-2 lattice parameters Name SLS*) db02l fa01f status operating baseline fallback Emittance at 2.4 GeV [pm] 5022 137 262 Lattice type TBA 7BA 5BA Total absolute bending angle 360° 585° 488° Working point Qx/y 20.42 / 8.74 38.38 / 11.28 28.29 / 10.17 Natural chromaticities Cx/y -67.0 / -19.8 -67.5 / -36.0 -64.1 / -39.9 Optics strain1) 7.9 5.6 8.9 Momentum compaction factor [10-4 ] 6.56 -1.39 -1.86 Dynamic acceptance [mm.mrad] 2) 46 10 17 Radiated Power [kW] 3) 205 228 271 rms energy spread [10-3 ] 0.86 1.05 1.15 damping times x/y/E [ms] 9.0 / 9.0 / 4.5 4.5 / 8.0 / 6.4 5.0 / 6.8 / 4.1 product of horiz. and vert. normalized chromaticities C/Q max. horizontal betatron amplitude at stability limit for ideal lattice assuming 400 mA stored current, bare lattice without IDs *) SLS lattice d2r55, before FEMTO installation (<2005)

Non-linear optimization 13 sextupole & 10 octupole families step 1: perturbation theory: insufficient 1st & 2nd order sextupole terms 1st order octupole terms up to 3rd order chromaticities step 2: multi-objective genetic optimizer objectives: dynamic aperture at Dp/p = 0, 3% contraints: tune fooprint within ½ integer box   Lattice acceptance results (ideal lattice) horizontal acceptance 10 mm·mrad sufficient for off-axis multipole injection from existing booster synchrotron Touschek lifetime 3.2 hrs 1 mA/bunch, 10 pm vertical emittance, 1.43 MV overvoltage further increase to 7-9 hrs by harmonic RF system.

More challenges... work just started Collective effects large resistive wall impedance ( aperture-3) low momentum compaction factor |a|, and a < 0 close thresholds for turbulent bunch lengthening  head-tail stability for chromaticity  0  intrabeam scattering  15-30% emittance increase  Alignment tolerances common magnet yoke = girder initial mechanical alignment will be insufficient extensive use of beam based alignment methods longer commissioning than 3rd gen. light source

Advanced options Round beam scheme A new on-axis injection scheme Wish from users (round samples...) Maximum brightness & coherence Mitigation of intrabeam scattering blow-up “Möbius accelerator”: beam rotation on each turn to exchange transverse planes SLS SLS-2 SLS-2 RB A new on-axis injection scheme cope with reduced aperture (physical or dynamic) interplay of radiation damping and synchrotron oscillation forms attractive channel in longitudinal phase space for off-energy off-phase on-axis injection. Figure taken from R. Hettel, JSR 21 (2014) p.843

Longitudinal gradient superbend Photon energy range YBCO1) HTS2) tape in canted-cos-theta configuration on hyperbolic mandrel  hyperbolic field profile open for radiation fan > 5T peak field 1) Yttrium-Barium-Copper-Oxide 2) High Temperature Superconductor Courtesy Ciro Calzolaio, PSI

Time schedule Jan. 2014 Letter of Intent submitted to SERI (SERI = State secretariat for Education, Research and Innovation) schedule and budget 2017-20 studies & prototypes 2 MCHF 2021-24 new storage ring 63 MCHF beamline upgrades 20 MCHF Oct. 2014 positive evaluation by SERI: SLS-2 is on the “roadmap”. Concept decisions fall 2015. Conceptual design report end 2016.

Summary The Swiss Light Source is successfully in operation since 15 years... ...but progress in storage ring design enforces an upgrade. Upgrade of the Swiss Light Source SLS has to cope with a rather compact lattice footprint... ... but the new LGB/AB cell provides five times lower emittance than a conventional lattice cell: Anti bends (AB) disentangle horizontal beta and dispersion functions. Longitudinal gradient bends (LGB) provide minimum emittance by adjusting the field to the dispersion. The baseline design for SLS-2 is a 12  7BA lattice providing 30-35 times lower emittance. The design is challenged by non-linear optics optimization, beam instabilities and correction of lattice imperfections. A conceptual design report is scheduled for end 2016.

Acknowledgements Beam Dynamics: Michael Ehrlichman, Ángela Saá Hernández, Masamitsu Aiba, Michael Böge Instabilities and impedances: Haisheng Xu, Eirini Koukovini-Platia (CERN), Lukas Stingelin, Micha Dehler, Paolo Craievich Magnets: Ciro Calzolaio, Stephane Sanphilippo, Vjeran Vrankovic, Alexander Anghel Vacuum system: Andreas Müller, Lothar Schulz General concept and project organisation: Albin Wrulich, Lenny Rivkin, Terry Garvey, Uwe Barth