+--------------------------------------------------------------------+ | LCLS ICFA03 S2E FEL SIMULATION TEST CASE DESCRIPTIONS | | (last modified: 28 June 2003) | +--------------------------------------------------------------------+ Electron Beam Parameters (as general information and to be used in the "zero-order simulation case described below) : Beam current: 3400 Amps (1.0-nC beam charge case) 1500 Amps (0.2-nC case) Relativistic Gamma: 28082.00 (equivalent energy: 14.35 Gev) RMS Instantaneous (slice) energy spread (delta-gamma/gamm0): 1.00e-4 (Gaussian distribution) Transverse 4-D phase space distribution: Gaussian Instantaneous (slice) projected normalized transverse emittance in x-x' or y=y': 1.20 mm-mrad (1.0-nC case) 0.65 mm-mrad (0.2-nC case) Twiss parameters at entrance to undulator: For 0.2 & 1.0-nC beam charge cases: (from Paul Emma Web page for S2E) betaX = 20.150 m, alphaX = -1.140 betaY = 14.947 m, alphaY = 0.818 Note: these Twiss parameters will give an approximate (but NOT exact!) matched beam solution to the FODO lattice described below Effective laser seed input power specification (to be used in monochromatic amplifier mode calculations as a "replacement for shot noise): nominal wavelength: 0.1500 nm P_in = 3.0 kW (1-nC case) P_in = 1.0 kW (0.2-nC case) For BOTH cases: Gaussian mode with waist at z=0 m Rayleigh range = 50.0 m (i.e. e-field waist size = 48.86 microns) LCLS Undulator Specification: Linearly-polarized, 30-mm period; wiggle-motion in x-plane Nominal K: 3.7110 (RMS a_w = 2.6241) Note: For amplifier mode simulations ("zero- and first-order" cases below), EITHER fix beam gamma at 28082. and then EMPIRICALLY vary K to give maximum gain at 0.1500 nm **OR** alternatively fix K at 3.711 and then vary gamma to maximize monchromatic gain at 0.1500 nm Constant pole tip strength (i.e. K) along undulator Flat-pole face (i.e. no curved pole tip focusing) Undulator poles to be organized in blocks of 3.360 m (112 periods), separated by "drift spaces" with length 0.240 m Focusing Lattice Specification: quadrupole FODO focusing with simple period of 7.200 m quadrupole magnetic length = 0.240 m each to be contained in a "drift space" (i.e. undulator break with K=0) also 0.240 m long (i.e. quadrupole has full occupancy of drift) each quad has integrated focusing strength of 5.40 T (i.e. for a 0.24-m length, this is equivalent to gradient of 22.5 T/m) Lattice begins with a focusing quad in x-plane (i.e. wiggle plane) at z=0.00 m (i.e. first undulator block begin at z=0.240 m) Full undulator length = 118.80 m (including quad/drift spaces); equivalent to 33 half FODO periods (this is close to nominal LCLS undulator length) Beam pipe specifications (for wake calculations): Inner diameter: 5.00 mm material: copper RMS roughness: 100 nm Longitudinal scale length: 50 microns (i.e. aspect ratio 500:1) Beam pipe to be considered continuous including through quad/drift spaces which seaparate the undulator blocks Suggested simulation runs (arranged by difficulty): "Zero-order": Monochromatic (i.e. "time-steady") amplifier mode run using base LCLS e-beam parameters , lattice configuration for both 1.0 and 0.2-nC cases. Input laser seed as specified above. This run serves as a check for the most difficult following cases. Suggested diagnostic output: Exponential gain length, saturation power and length (defined to be the z-position of the peak of microbunching) "First-order": Multi-slice calculations using time-varying e-beam parameters given in the ASCII "elegant2envelope" output file to be obtained from P. Emma's S2E Web page following these "rules": (a) Slippage-effects are artificially suppressed - i.e. each slice is run as an ***isolated*** amplifier-mode run (b) The same K is used for all slices and the central gamma of each slice is set to that which gave peak gain for the "zero-order", amplifier mode run parameters described above; i.e. variations in gain from slice-to-slice will be due to only to variations in current, Twiss parameters, instantaneous energy spread, etc. and NOT due to a centroid energy chirp (c) No shot noise microbunching and initial laser seed to be base case specified above (e.g. 3.0 kW for 1.0-nC, 1.0 kW for 0.2 nC) Suggested output: Exponential gain length, saturation power and length as functions of time; snapshots of P(t) at various z-positions (e.g. 25, 50, 75, 100, 122.4 m) "Second-order": Full time-dependent (i.e. polychromatic) SASE simulations for isolated segments of the LCLS beam. The following two segments are of particular interest: (1) A 12-micron (or longer if possible) z-length (e.g. 40.0 fs) beam "head" segment. For the 1.0-nC case, we will use Emma's Elegant macroparticle output file (http://www-ssrl.slac.stanford.edu/lcls/s2e/LCLS10JUN03_1nC_wCSR.out) and define the head to begin at -1.0400E-13 seconds. The current and beam energy are rapidly varying with time in this segment and its SASE behavior should be "rich in phenomena". (2) A 12-micron (or longer) segment beginning at -1.00e-14 seconds. This is in the "main" body of the beam and should be representative of the behaviour of the majority of the LCLS pulse. (3) Main body case for 0.2-nC case ***WILL GET FROM EMMA*** Suggested simulation method: If possible, please reconstruct a time-varying 5D phase space for z=0 by intelligent interpolation/use of the actual macroparticles in the Emma-produced elegant file and then superimpose the correct amount of shot noise microbunching for a SASE start. Do not presume periodic boundary conditions in time if possible. For the beam head segment, one will need to be sure that the simulation spectral bandpass both appropriately centered AND sufficiently wide to cover the time-varying FEL resonance wavelength. If time permits, include wakefield losses for the head segment case; for the body segment, we presume microtapering will be used to compensate for any wake and spontaneous emission losses. Use the same undulator K-strength as in the previous calculations. Suggested output: P(t) and P(omega) at various z-locations. "Third-order": Monochromatic amplifier runs with several different sets of quadrupole/BPM transverse offset errors. These runs can be done both for the single slize ("zero-order" case and multislice, non-slippage case (first-order" cases). Tables of the quad offset errors are obtainable from files linked to the LCLS S2E Web page (http://www-ssrl.slac.stanford.edu/lcls/s2e/) under the heading "LCLS Undulator Beam-Based Alignment Simulation Quadrupole Offset Files". These files also give an initial beam centroid and tilt which should be used for the e-beam at entrance to the undulator assembly (i.e. z=0 m). Suggested diagnostic output: sensitivity of to quad/BPM errors, a plot of and versus z as a check on beam position.