QUASI-STATIC MODELING of PARTICLE –FIELD INTERACTIONS Thomas M. Antonsen Jr, IREAP, University of Maryland College Park MD, 20742 Multiscale Processes in Fusion Plasmas IPAM 2005 Work supported by NSF, ONR, and DOE - HEP Separation of Scales in Wave - Particle Interactions Laser Wake Field Accelerator LWFA Basic Parameters: Plasma Wake = Laser pulse Vacuum Electronic Device (VED) Beam Drive Radiation Source Cavity Electron beam Radiation 8 10-5 cm (100 fsec) = 3 10-3 cm Propagation length = 5 cm Gyromonotron Radiation period (170GHz) Transit time Cavity Time Voltage rise time = 6 10-12 sec = = > 5 10-10 sec 5 10-9 sec 5 10-5 sec Hierarchy of Time Scales • Limited interaction time for some electrons -Transit time Laser - Plasma: - pulse duration Radiation Source: - electron transit time • Radiation period << Transit time Simplified equations of motion Laser - Plasma: - Ponderomotive Force Radiation Source: - Period averaged equations • Transit time << Radiation Evolution Time Quasi Static Approximation Pulse Shape/Field Envelope constant during transit time • Radiation period << Radiation Evolution Time Simplified equations for radiation Envelope Equations LASER-PLASMA INTERACTION APPROACHES / APPROXIMATIONS • Laser Full EM - Laser Envelope • Plasma Particles - Fluid Full Lorenz force - Ponderomotive Dynamic response - Quasi-static Full EM vs. Laser Envelope • Required Approximation for Laser envelope: wlaser tpulse >> 1, rspot >> l wp /wlaser <<1 • Advantages of envelope model: -Larger time steps Full EM stability: Dt < Dx/c Envelope accuracy: Dt < 2p Dx2/lc -Further approximations • Advantages of full EM: Includes Stimulated Raman back-scattering Laser Envelope Approximation • Laser + Wake field: • Vector Potential: E = E laser + E wake A laser = A 0(x ,x ^,t) exp ik 0x • Laser Frame Coordinate: + c.c. x = ct – z • Envelope equation: 2 ˆ 2 ˆ 4p ˆ 2 j ik 0 A 2 2 A ^ Aˆ c t x c t c Necessary for: Raman Forward Self phase modulation vg< c Drop (eliminates Raman back-scatter) AXIAL GROUP VELOCITY v g c(1 True dispersion : Extended Paraxial Requires : : k ^2 c 2 w 2p w 2 )1/2 k ^2 c 2 w 2p v g c / (1 ) 2 2w k^2 c 2,w 2p w 2 Extended Para - Axial approximation - Correct treatment of forward and near forward scattered radiation - Does not treat backscattered radiation Full Lorenz Force vs. Ponderomotive Description • Full Lorenz: d pi v B = q E+ i c dt • Separation of time scales E = E laser + E wake dx i = vi dt g = p2 1+ 2 2 mc • Requires small excursion x(t) E laser < < Elaser x(t) = x(t) + x(t) • Ponderomotive Equations 2 q A laser Fp = – mc mc 2 2g v Bwake dp = q E wake + + Fp c dt g = q A laser p2 1+ 2 2 + mc mc 2 2 2 PLASMA WAKE • Maxwell’s Equations for Wake Fields in Laser Frame 4p q n B p c g E x part icles B E x E 4pq n particles Laser frame coordinate: x = ct -z collapses t and z Time t is a parameter Solved using potentials F, A n0 Quasi - Static vs. Laser Pulse Dynamic Wake PlasmaWake Electron transit time: c Plasma electron c - vz t pulse 1 – vz / c Electron transit time << Pulse modification time Trapped electron Advantages: te = d = + c –v +v z ^ ^ dt t x fewer particles, less noise (particles marched in x= ct-z) Disadvantages: particles are not trapped CODE STRUCTURE Laser Fp 2 w 2 p 2 ˆ ˆ ˆ ik a a 0 ^ 2 a c t x c Particles and Wake dp^ 1 v B qE F p dx c v z c ^ dr^ v ^ dx c v z w 2p B 4p q n p c g Note: t is a parameter part icles E x Plasma Particle Motion and Wake Become 2D Particles marched in x Motion in r^ - x plane dx c vz dt r ri +1 dp^ 1 v B q E F p dx c v z c ^ ri dr^ v ^ dx c v z xj xj + 1 x = ct –z Density n(r^,x ) c n(r ,0) c vz ^ 0 PARTICLES CONTINUED : • Hamiltonian: H H (x ct z,r^, p) mc 2g q • Weak dependence on “t” in the laser frame Pz pz • Introduce potentials • Algebraic equation: B A H cPz const. q A c z E A x 2 pz pz (p ^, Az, a ) 2 g g (p ^, Az , a ) WAKE FIELDS • Maxwell’s Equations for Potentials 2^ A 4p j A c x E 2 A 4p x • Iteration required for EM wake 2 ^ A^ 4p j c ^ dp^ A E ^ ^ ^ dx x GAUGE Lorentz QUICKPIC ^ A^ Az x 4p A j c Transverse Coulomb WAKE ^ A^ 0 2 ^ 2^ 4p Pro: Simple structure Compatible with 2D PIC Con: A carries “electrostatic” field 2^ A Az 4p j c x Pro: A^ = 0 in electrostatic limit Con: non-standard field equations Numerical Simulation of Plasma Wave 2D WAKE Mora and Antonsen, Phys Plasmas 4, 217 (1997) Viewed in laser frame Particle trajectories Density maxima WAKE - Particle Mode Intensity Density Trajectories 5 4 3 r 2 1 0 Cavitation and Wave Breaking 0 10 20 x 30 40 QUICKPIC 3D UCLA/UMD/USC Collaboration UCLA UCLA: Chengkun Huang, V. K. Decyk, C. Ren, M. Zhou, W. Lu, W. B. Mori UMD: J. Cooley, T. M. Antonsen Jr. USC: T. Katsouleas 3D simulation - Laser pulse evolution - Plasma particles - Beam particles Numerics Parallel Object Oriented Beam particles equations: 3D dPb ^ qb ^ dt dPb || qb dt x dxb^ P b^ dt gmc A|| dx b P 1 b|| dt gmc Beam charge and current b 1 q Volume i b i jb 1 q V Volume i bi bi Axial Electric Field UCLA x ct z Laser Pulse x 1.8 nC electron bunch 25 MeV injection energy Reduced amplitude due to effects of beam loading UCLA x ct z Electron Distribution and Axial Field Laser Pulse Electron Bunch Distribution ~1.7x108 electrons Axial Electric Field VED Modeling Interaction Circuit Types Interaction requires: Wiggler FEL Beam Structure Synchronism S. H. Gold and G. S. Nusinovich, Rev. Sci. Instrum. 68 (11), 3945 (1997) Synchronism in a Linear Beam Device E(x,t) = Re {E exp [ik zz – wt]} w w = kz vz Dispersion curve w(kz) Doppler curve kz v z BWO TWT -p/d 0 p/d 2p/d kz Time Domain Simulation Standard PIC time dz vz dt Trajectories Positions interpolated to a grid in z Signal and particles injected t+dt Carrier and its harmonics must be resolved Fields advanced in time domain t System state specified 0 z L Frequency & Time Domain Hybrid Simulation RF phase sampled time Trajectories Vs (t)exp[iwt] t+dt dp q v B E dz vz c dt 1 dz vz Signal Period T=2p/w Ensemble of particles samples RF phase 0 z z+dz L Separate Beam Region from Structure Region Cavities coupled through slots Cavity fields Penetrate to Beam tunnel trough gaps Electrodynamic structures e- Beam region simulation boundary Cavity fields , jth cavity: E(x,t) Vs (t)e s (x)exp[iw t] c.c. j smodes B(x,t) s modes Beam tunnel fields E T Vk z,t e k rT ,z exp[iw n t] k ,n I sj (t)bs (x)exp[iw t] c.c. B T I k z,tbk rT ,zexp[iw nt] k ,n Code Verification: Comparison with MAGIC Operating Frequency 3.23 GHz Input Power (Pin) 49.06 kW B0=1kG rwall=1.4 cm rbeam=1.0 cm Q= R/Q= fres Output Power (Pout) MAGIC 214.2 kW TESLA 214.0 kW Dzgap = 1 cm 115 85.6 3.225 GHz 115 85.6 (on axis) 3.225 GHz TESLA : Sub-Cycling to Improve Performance Frequency of trajectory update with respect to field update (each nth step) CPU Time [sec] Pentium IV 2.2 GHz 10 7.4 5 11.7 2 24.6 1 46.0 MAGIC 2D MAGIC 3D ~ 2 hours ~72 hours Output Power [W] MAGIC: Pout=214.2 kW TESLA: Pout=214.0 kW 2.5 10 5 2 10 5 1.5 10 5 every time step 1 10 nd each 2 5 time step th each 5 time step 5 10 th each 10 time step 4 0 0 1 10 -8 2 10 -8 3 10 -8 Time [sec] 4 10 -8 5 10 -8 6 10 -8 Parallelization - Multiple Beam Klystrons (MBK) Input Power Output Power Beams surrounded by individual beam tunnel Beam Tunnels Resonators (Common) Code development for multiple beam case Code is being developed to exploit multiple processors Each beam tunnel assigned to a processor Communication through cavity fields Each processor evolves independently cavity equations Technical Challenge: Simulations of Saturated Regimes Phase Space NRL 4 cavity MBK Saturated regime of operation: Particles may stop Analogous situation in LWFA: plasma electrons accelerated Resonators: 1, 2, 3, 4 Particles with small z Reflected Particles Equations of particles motion (EQM): d/dz representation d pi q dz v z ,i d ri v i dz vz ,i vi B E c If vz 0 right hand side of EQM Numerical solution of EQM lost accuracy currently these particles are removed d/dt representation d pi vi B q E dt c d ri vi dt Switch to d/dt equations for selected particles with small vz,i Particle Characteristics and Current Assignment t Followed in t trajectories j(x ^ ,z j ) Iiv i (x ^ x^i )D(z j zi )e dt i Sum over time steps of duration dt Followed in z j(x ^ ,z j ) Ii i zj zj+dz z vi (x ^ x ^i )e iwt i v zi iwti Sample Trajectories in MBK 3 10 -9 0.6 1 10 -10 8 10 -11 0.1 t-integr 2.5 10 -9 0.08 0.5 t, sec 0.06 -9 0.04 0.02 0.3 1 10 -9 0.2 z 1.5 10 -9 t [sec] 4 10-11 2 10-11 0 19.35 0 19.36 19.37 19.38 19.39 -0.02 19.4 z [cm] z 5 10 -10 0.1 0 0 0 5 10 15 z [cm] 20 Direction reverses z 0.4 z 2 10 t [sec] 6 10-11 Accelerated Plasma Particles Plasma particles with E > 500 keV promoted to status of passive test particles 1.5 1 p^/mc 0.5 0 -0.5 -1 -1.5 -10 0 10 20 30 pz/mc 40 50 60 70 Conclusions • Reduced Models based on separation of time scales yield efficient programs • Simplifications take various forms - Envelope equations - Ponderomotive force - Resonant phase - Quasi-static fields • Breakdown of assumptions can cause models to fail - Reflected particles - Accelerated electrons - Spurious modes (VEDs) • Ad hoc fixes are being considered. Is there are more general approach?