Linac Design Update P. Emma LCLS DOE Review May 11, 2005 LCLS

Linac Design is Mature and Stable Single bunch, 1-nC charge, 1.2-mm slice emittance, 120-Hz repetition rate… 6 MeV z  0.83 mm   0.05 % 135 MeV z  0.83 mm   0.10 % 250 MeV z  0.19 mm   1.6 % 4.30 GeV z  0.022 mm   0.71 % 13.6 GeV z  0.022 mm   0.01 % Linac-X L =0.6 m rf= -160 rf gun new Linac-1 L 9 m rf  -25° Linac-2 L 330 m rf  -41° Linac-3 L 550 m rf  -10° Linac-0 L =6 m 21-1b 21-1d X 21-3b 24-6d 25-1a 30-8c undulator L =130 m ...existing linac BC-1 L 6 m R56 -39 mm BC-2 L 22 m R56 -25 mm DL-1 L 12 m R56 0 LTU L =275 m R56  0 SLAC linac tunnel research yard (RF phase: frf = 0 is at accelerating crest)

New Since Last Review Optional low-charge operating point (0.2 nC) Laser-heater low FEL gain “lock-in” detection Bunch length & energy feedback simulations Dark current, beam loss, and collimation study complete Detailed linac tuning simulations done Engineering constraints worked into design

Low Charge Operating Point (1) Motivation: Less charge  less wake Same compression factor  ~same jitter Lower gun current  lower emittance Chosen Scaling: Charge: 1 nC  0.2 nC Gun Current: 100  30 A (10 ps  6.5 ps) Sliced gun emittance: 1 mm  0.8 mm Final current: 3 kA  2 kA (same Lsat)

Low Charge Operating Point (2) gex,y (mm) 1.0 nC rms errors: 300-mm struct. 200-mm quad 200-mm BPM steer 10 seeds gex,y (mm) 0.2 nC Linac Alignment Eased

Low Charge Operating Point (3) 0.2 nC 1 nC Bane/Stupakov AC-conductivity model

Low Charge Operating Point (4) 1012 photons B. Fawley, S. Reiche

Low Charge Operating Point - Summary BC2 CSR De  1/5 Linac quad/BPM align. tol.’s  2 L2 transverse wake De  1/16 Peak current jitter  ½ X-ray pulse 80 fs (was 200 fs) Weak undulator RW-wake FEL power: 20 GW & ~1012 photons 1-nC operation still fully supported

Low FEL Gain ‘Lock-in’ Detection* (1) Simulate LCLS (Linac + M. Xie) with: Linac jitter (DQ, t0, f, V, etc.) Large emittance (3 mm)  low gain Spontaneous radiation background Laser-heater modulated at 7 Hz Use FFT to ‘lock-in’ on very weak FEL signal in spont. background * Idea from K. Robinson and comments by J. Rossbach (FAC, April 2004)

Low FEL Gain ‘Lock-in’ Detection (2) sE PFEL at 13.6 GeV

Low FEL Gain ‘Lock-in’ Detection (3) total signal Gain of ~25 detectable spontaneous signal FEL signal FFT of total signal PAC’05 paper: (P.E., Z. Huang, J. Wu)

Bunch Length and Energy Feedback Feedback OFF Feedback ON Juhao Wu (SLAC)

Dark Current, Beam Loss, Collimation (1) Model cathode dark current (Fowler-Nordheim and Parmela), scaling charge from GTF measurements Add dark current in critical RF structures along linac, based on K. Bane work in NLC (a non-issue) Track dark current through linac and through und. Include aperture restrictions and collimators Assess collimation scheme in terms of undulator protection and average power loss on collimators Evaluate wakefield effect of each collimator PE, J. Wu

Dark Current, Beam Loss, Collimation (2) 1 new BC2 E-coll. 36-mm (d = 10%) 2 new E-coll. 2.5 mm (d = 2%) under ground BC1 BC2 undulator 4 existing x-coll.’s 4 existing y-coll.’s 1.6 & 1.8 mm 3 new x-coll.’s 3 new y-coll.’s 2.2 mm… 1 new BC1 E-coll. 45-mm (d = 20%) 2.4 pC/pulse 3.3 W (120 Hz, 11.3 GeV) 1.0 pC/pulse 1.6 W (120 Hz, 13.6 GeV) 0.2 pC/pulse 0.3 W (120 Hz, 13.6 GeV) DE/E of 1 dropped klystron = -1.7%

Dark Current, Beam Loss, Collimation - Summary Undulator is protected from gun and structure dark current Maximum collimated beam power in ‘above-ground’ section is 0.3 W (well below safe level) Results still look safe even for 10-times more dark current (but already used worst-case GTF) Collimator wakefields should not be an issue (~0.5-mm alignment tolerances) Shower calculations were done (20 W/coll. was assumed, now ~70-times smaller)

Linac Tuning Simulations (1) Start with low-charge (0.2 nC, no CSR) Use elegant to automate tuning; use only ‘real’ diagnostics Add large errors to linac systems (e.g., magnets, RF, beam) Assume rough corrections already made (see LCLS Commissioning Workshop, Sep. 2004) Track through linac many times, each step simulating one particular correction (e.g., b-matching or RF phasing) Use correction devices already built into design (e.g., BC2 correction quads, trajectory controls) Evaluate final beam quality, correction convergence, dynamic range, problem areas, etc.

Linac Tuning Simulations (2) Element rms errors (Gaussian, 3-s cutoff) value unit Quads x and y misalignments 300 mm z misalignments 5 relative gradient errors 0.5 % roll angle errors 2 mrad Bends relative field errors 0.3 anomalous field gradients (BC’s) 3 tol. BPMs RF strucs. phase errors (static) deg relative voltage errors (static) 1 e- beam random charge error 10 initial beta mismatch in x and y 2.0 z

Linac Tuning Simulations (3) 0.2 nC before tuning gex,y  1.4 mm zx = 3.3 !

Linac Tuning Simulations (4) after tuning gey= 0.91 mm gex = 0.88 mm zx,y  1

Linac Tuning Simulations (5) DESIGN DESIGN DESIGN TUNEUP TUNEUP TUNEUP 2 kA

Comments System in mature, well studied, and reasonably ready for construction Many more slides are available to answer questions