1. Re-circulating Linac Option
FLS 2010, ICFA Beam Dynamics Workshop, SLAC, 1-5th March’10
Re-circulating Linac Option
Deepa Angal-Kalinin, Peter Williams & D. Dunning
ASTeC, STFC, Daresbury Laboratory & The Cockcroft Institute

2. Contents Introduction to NLS Re-circulation lattice design
Dogleg
Linac matching
Matching to Arcs and extraction
Arc design
Return path
Switchyard/Spreader
Start to End simulations
Comparison with single pass

3. New Light Source project : FEL source requirements
Repetition rate
1 kHz with an upgrade path to 10 kHz - 1 MHz
Photon energy range and tunability
3 FELs: FEL eV; FEL eV; FEL eV
The required tuning range of FEL3, with the choice of undulator type (APPLE-2) and minimum gap (8 mm)
Pulse length & pulse energy
GW power level in 20 fs (upgrade path to sub-fs pulses)
Transverse and longitudinal coherence
Fully variable polarisation for FEL1 & 2, at least horizontal and circular (R/L) polarisation over the full range for FEL3
=> CW Superconducting RF
=> Machine energy of 2.25 GeV
=> Laser HHG seeded, cascade harmonic FEL
These are the baseline specification as found in the published ODR for the single pass machine.

4. Design optimisation of the facility
Beam energy 2.25 GeV, bunch charge 200 pC
Need constant slice parameters on a length of 100 fs – or longer to accommodate present seed pulse and jitter (See Riccardo Bartolini talk Monday)
No residual (or very small) energy chirp
Baseline option of CW SCRF linac
(Outline Design Report)
Alternative option of re-circulating linac for cost savings

5. Re-circulating linac design
Re-circulation
Reduces capital cost & running cost
Can extract at different energies (1.2 & 2.2 GeV)
Provides natural upgrade path to higher energies
Opportunities for feedback
Additional issues compared to straight through option
Combining and separating different energy beams
CSR and ISR in the arcs, mergers/combiners & extraction
Bunch compression and linearisation scheme restricted
Jitter tolerances due to extra transport

6. Layout of NLS single pass option
This part is common for re-circulation option
High brightness electron gun, operating (initially) at 1 kHz
photoinjector
3rd harmonic cavity
BC1
BC2
BC3
laser heater
accelerating modules
collimation
diagnostics
spreader
FELs
IR/THz undulators
gas filters
experimental stations
2.25 GeV CW superconducting linac
3 free-electron lasers covering the photon energy range 50 eV–1 keV,
Total 18 SC modules

7. Re-circulation linac layout
Laser
Heater
Collimation + beam switchyard
(same as single-pass)
Linac
(2 modules)
BC1
Injection dogleg
Linac
(8 modules)
Extraction/spreader
Gun 3w
BC3
Merger
BC2
~30m
180˚ arc
180˚ arc
FODO + possible path length corrector 0m
50m
100m
150m
200m
Inject at ~200 MeV, two passes through 1 GeV
Total 10 SC modules
Compared to 18 in single pass

8. Layout comparison
Single pass
Recirculating

9. Injector, linac module, LH, 3HC, BC1
3rd harmonic Cavity
Bunch compressor
BC1
2 CW SCRF modules based on the TESLA module
Diagnostic
Gun
L-band NCRF
Possible laser heater
200 pC,18 ps FW, 135 MeV
Normalised emittance 0.3 mm. mrad
Choice of beam energy at injection > 200 MeV for minimising longitudinal space charge & optimum accelerating modules.
Laser heater is a place holder. Detailed microbunching studies not yet carried out.

10. Injection dogleg sextupoles
Need to merge with the high energy beam; Ratio of energies = 6
Achromatic and isochronous dogleg design
Optimised number, locations and strengths of
energy chirp ~0.9%)using ‘Simplex algorithm’ to minimise the emittance growth due to chromatic effects.
sextupoles
Last dipole of high energy chicane

11. Re-circulation linac
FODO between modules optimised at two different energies to minimise CSR blow up in first extraction dipole whilst simultaneously keeping 1.2 GeV twiss sensible.
__ bx
__ by
__ bx
__ by

12. Matching to arcs and extraction
1.2 GeV
Extraction Dipole after Linac
2.2 GeV
First dipole after the Linac separates 1.2 and 2.2 GeV beams.
Re-designed to bring down the betas in both lattices
R56 = 5 mm for matching to arcs and mm for extraction.
Optimised locations and strengths of sextupoles in matching to the arc to minimise emittance.
1.2GeV
2.2GeV

13. Arc design
Considered arcs from 4GLS, BESSY, LUX & DLS low-alpha – BESSY best
Four triple bend achromats, low dispersion
Wider footprint, non-isochronous (R56~8 mm, same sign as BCs)
ISR emittance growth from both arcs 3.6%
Two sextupole families to correct the chromaticities

14. Return pass
Return transport connecting arcs is simple FODO-cells with phase advance of 45˚ (can be increased to reduce number of magnets). Possible to include few screens to study the beam properties at 1.2 GeV.
Plan to have isochronous path length adjuster. 4GLS design, based on moving girders - would give independent phase control on second pass of linac (this control has been assumed in the simulations)
Path length corrector

15. Beam switchyard/spreader
Beam switchyard design based on fast kicker and scheme like LBNL has been chosen as it provides easy path
to include more FELs
to switch the required bunches (or all the bunches) to any one particular FEL branch using a DC dipole at kicker location
First branch is dedicated diagnostics branch for tomography
Arc 1 : TBA
Kicker
Septum
Arc 2 : TBA
Diagnostics line
FEL1 
FEL2 
FEL3  20 40 60 2 4 6 8
Take-off section of FELs
Details of design of one spreader line

16. Machine optics (start to exit of the spreader)
Arc
Arc
Linac -2nd pass
Spreader
Linac -1st pass
Coll
BC1
Dogleg
Return FODO

17. Simulation results : projected emittance
Extraction, BC3
Collimation
Return FODO
BC1
Dogleg
Linac -1st pass
Match to arc, arc
arc,BC2, merger
Linac -2nd pass
Spreader
Final normalised projected emittances in x/y ~ 0.6/0.43 mm.mrad

18. Emittance Start to End of spreader
Recirculation
x/y ~ 0.6/0.43 mm.mrad
Single pass
x/y ~ 0.6/0.3 mm.mrad

19. Parameters exit spreader (Nov 09)
Tracked 100k particles. The data is binned in 1fs slices.

20. Luus-Jaakola optimisation (Feb 10)

21. Parameters exit spreader (Feb 10)
Tracked 100k particles. The data is binned in 1fs slices.

22. Parameters exit spreader (Nov 09)
Tracked 100k particles. The data is binned in 1fs slices.

23. Bunch parameters at the exit of spreader (Single Pass)
Longitudinal current distribution
Current (A)
rms relative energy spread
Normalised emittance
Normalised emittance(m)
rms relative energy spread(%)
R. Bartolini

24. Longitudinal current distribution
Single pass
Recirculation

25. Emittance
Single pass
Recirculation

26. Relative energy spread
Recirculation
Single pass

27. Comparison with single pass time-dependent results
Comparison between radiation profiles of the re-circulating and single pass cases at a comparable power level

28. Summary
Many interesting beam dynamics challenges have been understood and solved for the NLS re-circulation option.
The design is very close to achieving simultaneous bunch properties required for the FEL operation.
Continuing with the automatic optimisation process to tailor bunch to “flat top” at 1kA and obtain less energy chirp on the bunch. First results show that the optimisation process gives a flat region (but energy chirp remains rather high)
The design will be documented in the planned Conceptual Design Report of NLS (to be published in May’10) to be re-visited in the future if required.

29. R. Bartolini Diamond Light Source Ltd.
Thanks to
J. Jones (ASTeC/STFC)
B. Fell (EID/STFC)
R. Bartolini Diamond Light Source Ltd.
The NLS Physics and Parameters Working Group
&
D. Douglas, G. Krafft (JLab) for useful discussions