C.Limborg-Deprey LCLS FAC, April April Injector Physics C.Limborg-Deprey, D.Dowell,Z.Li*, J.Schmerge, L.Xiao* RF Gun modifications Linac Sections modifications Risk Mitigation Plans QE studies Pulse Shaping 3D-ellipsoid R&D Laser (see S.Gilevich Presentation) 0.2nC (see P.Emma Presentation) Summary (*)

C.Limborg-Deprey LCLS FAC, April April RF Gun Modifications Z-Coupling 15 MHz  = 2 Feedback: Control signals (Reflected power, metal temperature) Actuator (water T.) Push Pull deformable tuners (No Plunger) November 04 review, Report from J.Wang et al. PIC simulations Bead-drop procedure ? Back-plate dynamically movable Include cell probes

C.Limborg-Deprey LCLS FAC, April April Modified RF gun design Gun fabricated at SLAC RF design complete Mechanical model in progress 120Hz heat calculations under way Dual Feed Suppresses the time dependent dipole mode Matching phase for 2 feeds by holding mechanical tolerances on both arms Z coupling (instead of  -coupling) Pulsed heating reduced + easier machining Racetrack shape compensates for stronger quadrupole mode 15 MHz mode separation adopted Lower cathode voltage for the 0-mode “Suppresses” two degrees of freedom in parameter space Larger radius for coupling cell iris Reduces RF emittance Easy to accomodate elliptical curvature to reduce surface field Shaping of RF pulse for reducing average power 4kW -> 1.8 kW ; cooling channels designed for handling 4kW Reduce reflected power from gun LCLS-TN-05-3.pdf Courtesy L.Xiao

C.Limborg-Deprey LCLS FAC, April April Linac: Dual Coupler at entrance cell Kick is reduced by more than 4 times in output coupler nominal 1nCEnt. L0aExit L0aEnt. L0bExit L0b (  /  ) at 0  single feed in % /0.5 With dual feed reduction head-tail kick reduced by 20 (  /  ) at 0  dual feed in % /0.5 Operating point rms head-tail trans. kick for 10ps bunch Dual feed at entrance cell BUT NOT at exit Quadrupole head-tail not a problem at exit cell Head-tail dipole kick from single feed Generates emittance growth

C.Limborg-Deprey LCLS FAC, April April L0a L0b: New Design With WR284 Waveguide ++ b= w= d= r=1.000 WR284 waveguide a1 a2 a3 a4 t1 t2 t3 t4 b2 b3a b3b/b4 beampipe originalnew Coupler cup2 cup3-a cup3-b R0.5 R1.38 Using standard WR284 waveguide – eliminate all tapers (flanges closer to body, to accommodate linac solenoid ) Coupler cell lengthened to match height of WR284 waveguide Racetrack parameters readjusted Courtesy Z.Li

C.Limborg-Deprey LCLS FAC, April April Waveguide stiffners Courtesy J.Chan PumpOut Linac: New Design With WR284 Waveguide Enough clearance for solenoid Waveguide curvature adjusted to minimize S11 Waveguide cold-tested 2 arms adjusted for identical match

C.Limborg-Deprey LCLS FAC, April April Cathode- QE improvement Courtesy D.Dowell, R.Kirby LCLS QE Spec. 6x10 255nm After H-beam cleaning 1.2x10 -4  Idt for H-ion beam %Carbon on surface initial30 1H C11 2H H H H-ion Cleaning Experiment QE at low voltage (No Shottky Enhancement) Surface unaltered by H-ion beam cleaning contrary to effect of laser cleaning

C.Limborg-Deprey LCLS FAC, April April QE improvement Schottky Enhancement of the QE  Theory(*) Approximations R = 0.34 (reflection) 1phot -> 1 e No e-scattering (before emission) Fermi-Dirac at 0 K no roughness, no surface features Theoretical model can be refined  More tests planned More samples (process) Find Optimal H-ion beam current and integration time (and temperature)  Implement on GTF gun? Courtesy D.Dowell GTF (measured) LCLS Specifications LCLS minimum required (*) Based on J.Schmerge et al., Proc.FEL04,

C.Limborg-Deprey LCLS FAC, April April Minimum Emittance  perfect machine ~ 0.9  m.rad (for nominal 1nC tuning) Only ~ 0.1  m.rad margin for emittance growth Contributions to emittance Large cathode emittance for copper measured 0.6  m.rad per mm of r laser (theoretical is 0.3  m.rad ) Minimum set by space charge limit Minimum r laser or electrons cannot leave cathode (for metal cathodes) r min. = 0.82 mm at 54 MV/m for a 1nC  cathode > 0.5  m.rad RF emittance small ~0.15  m.rad  space charge can be supressed by appropriate “emittance compensation” uniform distribution inside an ellipsoid produces linear space charge force Linear “emittance compensation” corrects for this term Should we investigate on 3D-ellipsoid pulse shaping ?

C.Limborg-Deprey LCLS FAC, April April Ellipsoidal Emission pulse “Beer Can” is not the optimal distribution Electrons uniformly distributed in 3D ellipsoid volume r max = 1.2 mm Pulse length Radial profile = half-circle fwhm = 10 ps Pulse length Line Density = parabola

C.Limborg-Deprey LCLS FAC, April April Standard “Beer can” against “3D ellipsoid” r max = 1.2mm r= 1.2mm  cath. = 0.69mm.mrad per mm  cath. = 0.6 mm.mrad per mm

C.Limborg-Deprey LCLS FAC, April April “Beer Can” vs “3D ellipsoid” Best Tunings for ~ 100A at end of injector  = 1.02 mm.mrad;  80% = 0.95 mm.mrad  = 0.57 mm.mrad ;  80% = 0.58 mm.mrad  using standard  “cathode” = 0.6 mm.mrad per mm radius !! Simulations with similar numerical meshing parameters and 200k particles

C.Limborg-Deprey LCLS FAC, April April Sensitivity and Safety Margin Solenoid  RF r laser Pulse length “Beer Can”  1%< 5  ~0.1 mm<1ps “3D-ellipsoid” >  3%>10  > 0.3 mm>4ps Solenoid  2% Tuning + Stability of injector are eased; very large margin below 1mm.mrad Margin for emittance below 1 mm.mrad for the  80% 0.67 mm.mrad for “3D-ellispoid” (  projected =  80% ) 0.9/1.0 mm.mrad for “beer can” (  80% /  projected )

C.Limborg-Deprey LCLS FAC, April April D-Ellispoid Feasibility ? Two solutions proposed Pulse Stacker With 12 Gaussians of alternating polarities Too lossy, uses too much space, unweildy Awkward but easy control on individual components Technically feasible with many $$$$$$$ for controls, to achieve alignment, timing measurement to adjust amplitude coefficient Spectral Control technique Masking technology for IR exists Probably better for space and money than previous solution Before or after amplifier ? Before = recover lost energy but shape might not be preserved through chain After = difficult power handling (high losses in gratings and masking) Direct UV might be more appropriate; masking technology needs to be developed (transmissive or reflective scheme)- need to solve high damage threshold issue z y x X mask y mask Chirped input, temporally t x y t To cathode (z,y) plane (z,x) plane Fluence < 150 mJ/cm 2, E = 50mJ BW < 15 nm, Chirp = THz/ps  = 2200 groves per mm,  = 6.7  D pencil beam (1m) =11.7 cm D y = 2 waist y = 2  = 0.9 cm Fluence < 150 mJ/cm 2, E = 50mJ BW < 15 nm, Chirp = THz/ps  = 2200 groves per mm,  = 6.7  D pencil beam (1m) =11.7 cm D y = 2 waist y = 2  = 0.9 cm Courtesy of P.Bolton

C.Limborg-Deprey LCLS FAC, April April What we would like to have ? Optically Controlled Spatial Filtering Spatial frequency mask in Fourier Plane with sub-ps dynamics for switching Easy to generate flat disk of fixed radius in image plane by masking in Fourier plane a(t) controlled by driver pulse Courtesy of P.Bolton Object Driver with temporal shape = half-disk Mask in Fourier Plane ImageTransmissive or Reflective Optics FFT -1 Pulse length a(t)

C.Limborg-Deprey LCLS FAC, April April Summary on 3D pulse shaping Ideal emission pulse = “3d-Ellipsoid” not “Beer Can” Perfect emittance compensation in high charge regime Impressively less sensitive to tuning parameters Much larger tolerances than those defined for “beer can” pulse Much easier to tune  projected as low as 0.6 mm.mrad Ellipsoidal Laser Pulse is a Technical challenge maybe only slightly more challenging than “beer can” generation if direct UV shaping is considered for “beer can”, the “ellipsoid generation” shares many of the same difficulties Solution has (by construction) adaptive correction Of Shottky Of non-uniformity on cathode Solution should be corrected on e-beam measurement (using Genetic Algorithm as suggested at ERL05)

C.Limborg-Deprey LCLS FAC, April April Conclusion Gun RF design completed Mechanical design under way (thermal analysis on-going) Linac RF design completed Mechanical design in progress Risk Mitigation Plans for 1nC Cathode studies: H-ion cleaning for higher QE 3D-pulse shaping Tuning of 0.2nC completed (see P.Emma) On-Going studies Laser steering stabilization Feedback for Laser Energy/ RF gun (P,  ) Commissioning Plans Beam BA strategy

C.Limborg-Deprey LCLS FAC, April April BACK-UP

C.Limborg-Deprey LCLS FAC, April April Early thought : Stacking pulses 6+6 beamlets of different radii Gaussians Wash out discrete steps of rms value Interferences

C.Limborg-Deprey LCLS FAC, April April Fighting interferences in Stacker Alternating polarization + appropriate choice of , interference effect is minimized No interferenceInterferences random phases ~<15 % for all draws

C.Limborg-Deprey LCLS FAC, April April PARMELA simulations using stacker distributions Beer Can Direct beer can Ellipsoid ideal 50 Beamlets no interference Stacker 12 Beamlets and random phase  = 1.02 mm.mrad;  80% = 0.95 mm.mrad (with standard  “cathode” =0.6)  = 0.71 mm.mrad ;  80% = 0.71 mm.mrad (with overestimated  “cathode” =0.7)  = 0.80 mm.mrad;  80% = 0.80 mm.mrad (with overestimated  “cathode” =0.7) IDEAL

C.Limborg-Deprey LCLS FAC, April April Stacker Layout Profile shape r pulse energy control delay line imaging optics collimator- magnifier delay slide grating spectral filters photocathode half waveplate polarizing cube launch mirror Courtesy of P.Bolton “what to try to avoid…” from P.Bolton

C.Limborg-Deprey LCLS FAC, April April Spectral Control Principle Create a time-space correlation Ideally better in UV but masking technology does not exists yet Constraints 1- Fluence < Damage threshold 2- BW not too large (for THG) 3- Space 4- if possible after amplifiers Create a time-space correlation Ideally better in UV but masking technology does not exists yet Constraints 1- Fluence < Damage threshold 2- BW not too large (for THG) 3- Space 4- if possible after amplifiers Chirped input, temporally In z,x plane Vertical mask z y x Fluence F < 150 mJ/cm 2 pulse energy E = 50mJ BW < 15 nm L gratings->mask < 2m Chirp = THz/ps  = 2200 groves per mm   = 6.7  D pencil beam (1m) =11.7 cm D y = 2 waist y = 2  = 0.9 cm Beam dispersed enough to have beam size negligible Fluence F < 150 mJ/cm 2 pulse energy E = 50mJ BW < 15 nm L gratings->mask < 2m Chirp = THz/ps  = 2200 groves per mm   = 6.7  D pencil beam (1m) =11.7 cm D y = 2 waist y = 2  = 0.9 cm Beam dispersed enough to have beam size negligible In z,y plane Courtesy P.Bolton