ERL Drivers for FELS So Easy Even A

ERL Drivers for FELS - So Easy, Even A Cave Man Could Do IT! D. Douglas JLab

Acknowledgments & Disclaimer • Thanks to you all for the opportunity to participate in this happy occasion, and to recognize the contributions and example of our wonderful friend & colleague • I’ll be relating experience & results from the collective JLab FEL team and our Accelerator • As with many Do. D-funded & Engineering Division co- scientists, I don’t get out much, so rather than trying give a comprehensive overview of ERLworkers – I’m verytograteful to Driven FELs, primarily speak to experiences (and them and to. I’ll JLab for their ongoing supporthere andat JLab… misadventures) opportunities

Historical Context • ERLs 1 st proposed by Tigner (1965), operated at Chalk River (Schriber, Funk, Hodge and Hutcheon, 1975), identified as potential advance for FEL drivers (use of ER at UCSB & LANL, 1980 s, use of same-cell ER at MIT: Flanz and Sargent 1985) • Successful implementation for high-efficiency/high-duty-factor/high -power FELs depends on two further epiphanies: – Use of SRF technology (Todd Smith, 1980 s; Bisognano & Krafft, late 1980 s) • Low peak power, high average power by way of high, CW repetition rate – Longitudinal matching • Bunch compression after acceleration (with correction of higher order effects) – T. Smith, FEL’ 85 • Energy compression during energy recovery (Larry Doolittle ~1991? ; documented by Piot et al (PRST-AB, 2003))

So, Why Use ERL Drivers for FELs? Great Potential for Cost-Effective 4 th GLS! • • Linac quality beam (brightness) Potential for high duty cycle (CW) – High average power from high repetition rate, not high instantaneous power • Much easier • • • Storage ring wall plug efficiency (cost) Operational flexibility (robustness) Entertainment value: numerous beam dynamical effects manifest themselves… – LSC, BBU, CSR, … The perfect combination for an FEL driver: great accelerator performance and lots of distractions to keep physicists occupied… This notion appeared so promising that JLab director Herman Grunder aggressively pursued support for a test system, which led to the IR Demo FEL "With such a beam, we said `My God, there must be something we can do with it other than fundamental physics'" - H. Grunder, Washington Post, 2 March 1997

Indeed, there was… The Jlab IR Demo FEL • USN/ONR funded (1995) construction of SRF ERL testbed: “JLab IR Demo FEL” • Intended to validate a number of concepts – Low peak, high average power paradigm – Use of SRF in “high” CW current application (5 m. A) • BBU management – – High brightness CW injector Beam quality preservation High average power oscillator-based FEL Longitudinal matching scenario • Inject long bunch (alleviate space charge) • Compress length at full energy • Energy compression during energy recovery

Key Concept: Longitudinal Matching in an SRF ERL FEL Driver E “oscillator” E f injector f “amplifier” linac dump E E f f wiggler E E f f

2) 1) E E f linac dump 6) E f 5) 3) injector f 4) E f wiggler E f

System Design for IR Demo • Intention was to leverage investment in CEBAF – Use (pilfer) components from inventory, NP DC R&D gun, … • System needed to accomodate large exhaust energy spread from FEL – so was intended to be a clone of the large-acceptance MIT-Bates recirculator… but the fates (in the person of Slava…), intervened…

CSR (Fear)-Driven Design • Re-worked design to limit bending before wiggler – Risk reduction – Successfully lased CW at various wavelengths with powers up to 2. 1 k. W • Validated design paradigms • Investigated BBU & other effects • Allowed initial work toward THz source • CSR-observant revisions were key to project success – honestly, we really didn’t have a clue how to do the longitudinal matching until after we stumbled over it…

JLab IR Demo Dump core of beam off center, even though BLMs showed edges were centered (high energy tail)

Retrospective on IR Demo • Answered a number of questions, like “can it be done? ” – BBU control, longitudinal matching, baseline on CSR… • “polychromatic” source of radiation Coherent Harmonics – THz – IR (+ coherent harmonics) – Compton X-ray source (Krafft et al. ) • Brought the issue of beam quality preservation to the forefront – CSR nonfatal, but very much an issue • Led immediately to “do it again… with MORE power”…

Jlab IR Upgrade FEL • “That was easy”… so power scale-up (by 10 x) was an obvious next step • Double current, raise FEL extraction efficiency, triple energy to get to 10 k. W • “CSR is your friend” – leverage IR Demo design to provide more flexible longitudinal match, including curvature and torsion correction (not just survivable, but a funding source) – Include THz beamline

Longitudinal Matching Scenario Requirements on phase space: • high peak current (short bunch) at FEL E – bunch length compression at wiggler using quads and sextupoles to adjust compactions • f “small” energy spread at dump – energy compress while energy recovering – “short” RF wavelength/long bunch, large exhaust dp/p (~10%) Þ get slope, curvature, and torsion right (quads, sextupoles, octupoles) E f E f

Nonlinearity Control Validated By Measurement: Harmonic RF Unnecessary (and Expensive!) Figure 1: Inner sextupoles to 12726 g-cm and trim quads to -215 g Figure 2: trim quads at -185 g with same sextupoles Figure 3: trim quads at -245 g Figure 4: quads at -215, but sextupoles 3000 g below design, at 10726 g-cm Figure 5: where we left it: trim quads -215 g sextupoles at 12726 g-cm launch f arrival f

Injector to Wiggler Transport

If you do it right linac produces stable ultrashort pulses Can regularly achieve 300 fs FWHM electron pulses ~150 fsec rms

Injector to Reinjection Transport

BBU – a bump in the road • Schedule constraints led to use of “The Admiral” – a high gradient prototype SRF module with light HOM damping – Predictions => BBU threshold at 2. 5 m. A – How to fix? • By this time, Slava had arrived at Jlab, and had thoroughly inculcated us all with the outlook that phase space is phase space, not a bunch of disconnected orthogonal transverse and longitudinal subspaces – so it was natural to adopt a fully coupled solution • Rand & Smith, 5 quad rotator interchanging transverse phase spaces; BBU completely stabilized

CSR/THz – Bridge Out • Successfully generating a short bunch at the wiggler lead to a short bunch in the return arc, with significant CSR generation in each location – 10 s of W of THz dumped onto FEL outcoupler… resulting in distortion & power limitations • Initial 10 k. W run at 25% duty cycle: 1 second on, 3 seconds off (cool mirrors) • “The JLab Occasionally 10 k. W FEL (2004)” • Installed “de (actually, over)-bunching” chicane after wiggler; “THz traps”, cryo-cooled OC, thereby alleviating effect • 14. 3 k. W in November 2006

Retrospective on the IR Upgrade • Learned how to manage BBU • Encountered CSR as an unanticipated limit: – Not beam quality dilution – POWER DEPOSITION! • Had 1 st look at halo, other collective effects – Wakes, LSC, RF heating…

Next Step: JLab UV FEL • IR Demo validated – SRF ERL driver – Low peak/high average power paradigm • IR Upgrade validated – Power scaling – BBU control – Role of CSR as performance limit • Issue is not just “beam quality preservation”, its also “power in the wrong place” • Short wavelengths more challenging – Test of beam brightness & beam quality preservation, mirror design, power-flow management, …

System Concept UV FEL “bypass” • ~150 Me. V • 60 p. C x 37 MHz – (5 m. A) Tighter beam quality required at shorter wavelength – Test of beam brightness – Check beam quality preservation

Status • 1 st beam through bypass – Demonstrated bunch compression, beam quality eps x beta x alpha x eps y beta y alpha y 3. 883392503 7. 081035351 9. 527849399 2. 386019152 4. 412957251 8. 681434968 • 1 st CW run 7/29/10: ~1 m. A (~100 k. W) • Installing wiggler chamber • 1 st lasing imminent (we hope…)

State of ERL Performance ERLs provide very high power/high brightness beams • FEL drivers – – E: 10 s of Me. V – few Ge. V Q: 100 s p. C – 1 n. C I: m. A – 10 s m. A enormalized ~ l/4 p • 1 -10 mm-mrad – Pbeam ~ MW • Light sources – – – E: 5 – 10 Ge. V Q: ~10 s p. C – 100 p. C I: 100(s) m. A enormalized < ~1 mm-mrad Pbeam ~ GW • • ***high power=> halo major issue! Can’t lose 10 -5 of beam! implications: tiny spot size, COTR effects, 6 -d systems…

• Higher powers The Future – Higher charge/bunch, shorter bunches => extraction efficiency for (and power from) CSR rivals (exceeds) that of FEL • High rep rates at shorter wavelengths – JLAMP – Hard X-FEL • Multiple FELs driven by single ERL – RF separation as in CEBAF (with recombination)

The Late, Great JLAMP • IR -> IR Upgrade -> UV…. Where next? • JLAMP – yet another upgrade – 2 pass x 300 Me. V linac; seeded amplifier reaching ~10 nm – XFELO test

ERL-Driven X-FELS with apologies to Paul Emma and other people that actually have X-FELs! • Higher energies => longer linacs => higher cost • Recirculation/energy recovery are palliative measures: make systems more affordable • Will require extensive study and creative design to ensure beam quality preserved, optimum cost/benefit achieved – More FELs/unit linac is better… – Multiplicity by way of RF separation (a la CEBAF)?

GERBAL: “Generic Energy-Recovered Bisected Asymmetric Linacs” – Machine configuration: Multiple wigglers (9. 6 Ge. V beam) 1. 2 Ge. V Linac 1. 2 Ge. V ER 10 Me. V 4. 8 Ge. V accel. Injector 6. 0 Ge. V ER 1. 2 Ge. V accel. 4. 8 Ge. V ER 6. 0 Ge. V accel. 1 MW Dump 3. 6 Ge. V Linac – Transverse optics r e c i r c r e c i r c 29

Perspective • “conventional” ERLs – in infancy. FELs (or –“terrible perhapstwos”…) not as advanced, – stone knives but still andvery animal sophisticated – like cathedrals or bridges skins • Rings – very advanced systems – equivalent to nanotechnology or rocket science But at least ERLs are so easy “even a caveman could do it!”

Observations As we’re way too early in the game to draw conclusions… • 35 years of ERL operation experience – Chalk River, MIT, LANL, JLAB, JAERI, Novosibirsk, JLAB, Daresbury, JLAB, … • Successful trend toward shorter & shorter wavelengths and higher & higher powers • Many unresolved issues, but thanks to great leadership – by our guest of honor and those he’s influenced – there’s good reason to expect excellent outcome!

The Late, Great JLAMP • IR -> IR Upgrade -> UV…. Where next? • JLAMP – yet another upgrade – 2 pass x 300 Me. V linac; seeded amplifier reaching ~10 nm – XFELO test

Design Requirements • Generate, accelerate, and deliver properly configured drive beam to FEL – 1 mm-mrad x 50 ke. V-psec x 200 p. C – Ipeak ~ 1 k. A (200 fsec FWHM x 0. 1% dp/p) • Recover (degraded) exhaust beam • Preserve beam quality, manage losses, avoid instabilities, etc • Fit in vault (an upgrade) • Cost < 100 M$

Design Parameters (F. Hannon, IPAC 2010) 2010 2012 Bunch charge (p. C) 135 200 Bunch rep. rate (MHz) 75 4. 68 Average current, max (m. A) 10 1 Norm. transverse emittance at FEL (µm) 10 1 Longitudinal emittance at FEL (ke. V ps) 60 50 Energy spread at FEL (% rms) 0. 4 0. 1 Bunch length at FEL, rms (fs) 150 80 Bunch energy (Me. V) 100 600

Reality Check • As defined by these requirements, JLAMP will – Be a low cost user facility meeting significant scientific need – Test numerous concepts critical to next generation light sources • High brightness/high duty factor sources • Beam quality preservation in SRF environment – LSC, CSR, MBI, … • Multi-pass recirculation/energy recovery • Very high risk, very high return…

Beam Dynamics Issues • • • space charge BBU other wakes/impedances – linac, vacuum chamber, diagnostic impedences • • ISR – emittance, dp/p. . . • Error analysis – Alignment • Microwave. Studio modeling of all components impedance budget, policy, enforcement (impedence policing) Magnets, cavities, diagnostics – Powering • Excitation, ripple, reproducibility – resistive wall – field tolerance – Ions – gas scattering – timing & synchronism – phase & gradient – diagnostic errors • vacuum effects intrabeam scattering – IBS – Touschek • – Formation – gas scattering – beam formation processes • halo CSR – CSR basic ("elegant") – 3 -d modeling – microbunching instabilities Homogeniety, calibration RF drive – transient analysis Operational simulations – – threading, orbit correction emittance measurement lattice function tuning longitudinal matching • • • phase transfer function bunch length compression tuning energy compression tuning

JLAMP Recirculator Beam Dynamics

ERL-Driven X-FELS with apologies to Paul Emma and other people that actually have X-FELs! • Higher energies => longer linacs => higher cost • Recirculation/energy recovery are palliative measures: make systems more affordable • Will require extensive study and creative design to ensure beam quality preserved, optimum cost/benefit achieved – More FELs/unit linac is better… – Multiplicity by way of RF separation (a la CEBAF)?

FEL-Seeded ERL-Driven XFEL Two bunch trains UV seed, XFEL drive RF separation in 1 st pass UV bypass l. RF/2 longer (recovers bunch train) Issues: SYNCHRONISM UV seed pulse energy, up-conversion

Synchronization

GERBAL – Machine configuration: Multiple wigglers (9. 6 Ge. V beam) 1. 2 Ge. V Linac 1. 2 Ge. V ER 10 Me. V 4. 8 Ge. V accel. Injector 6. 0 Ge. V ER 1. 2 Ge. V accel. 4. 8 Ge. V ER 6. 0 Ge. V accel. 1 MW Dump 3. 6 Ge. V Linac – Transverse optics r e c i r c r e c i r c 43

Perspective • “conventional” ERLs – in infancy. FELs (or –“terrible perhapstwos”…) not as advanced, – stone knives but still andvery animal sophisticated – like cathedrals or bridges skins • Rings – very advanced systems – equivalent to nanotechnology or rocket science But at least ERLs are so easy “even a caveman could do it!”

Observations As we’re way too early in the game to draw conclusions… • 35 years of ERL operation experience – Chalk River, MIT, LANL, JLAB, JAERI, Novosibirsk, JLAB, Daresbury, … • Successful trend toward shorter & shorter wavelengths and higher powers • Many unresolved issues, but thanks to great leadership – by our guest of honor and those he’s influenced – there’s good reason to expect excellent outcome!