Eric R. Colby* SLAC National Accelerator Laboratory

Transcription

1 Eric R. Colby* SLAC National Accelerator Laboratory Work supported by DOE contracts DE AC03 76SF00515 and DE FG03 97ER41043 III.

2 Overview of the Technology Likely Performance Characteristics Present State of R&D Illustration of the Concept 500 ev Strawman 50 kev Strawman Conclusions

3 Compact: High Gradient Acceleration is Possible Dielectrics have excellent short pulse breakdown resistance >1 GV/m Net accelerating gradients of GV/m are expected to be possible, based on material damage threshold data Microjoule class rack mounted lasers are needed, and are readily available Efficient: High Efficiency Acceleration is Possible Efficient DPSS and fiber lasers (>30%) are commercially available and are getting better (DARPA/SHEDS) Coupling structures that tightly yguide the laser have been designed and fabricated using conventional techniques Power recirculation is possible to further enhance efficiency Rapidly Evolving Technology: Large scale industry R&D can be leveraged Fiber lasers, optics, and PCF fiber telecomm, cutting/welding/engraving Silicon structures chip industry Structure Fabrication is by inexpensive mass scale industrial manufacturing methods Attosecond, point like sources: Short wavelength acceleration naturally leads to attosecond bunches and point like radiation sources

4 Fully Integrated Accelerator System on a Chip Quantum Well InGaAsP lasers and hybrid silicon lasers both candidates for on board power sources for accelerators Fully integrated waveguiding system between laser and structure Photonic band gap gpaccelerator structures wake is dominated by Cerenkov contribution and 1 2 propagating modes On board phase and amplitude control Optically based diagnostics with on board receivers The next R&D step is to demonstrate high gradient, efficient acceleration in a microstructure

5 Structure Candidates for High Gradient Accelerators Projected maximum gradients based on measured material damage threshold data Photonic Crystal Fiber Silica, λ=1890 nm, E z =400 MV/m cylindrical lens vacuum channel cylindrical lens laser beam top view Photonic Crystal Woodpile Silicon, λ=2200nm,, E z=400 MV/m electron beam z y x Transmission Grating Structure Silica, λ=800nm, E z =830 MV/m λ/2 λ

6 Y (μm) Synchronous (β=1) Accelerating Field * Accelerating mode in planar photonic bandgap structure has been located and optimized * Developed method of optical focusing for particle guiding over ~1m; examined longer range range beam dynamics * Simulated several coupling techniques * Numerical Tolerance Studies: Nonresonant nature of structure relaxes tolerances of critical dimensions (CDs) to ~λ/100 or larger S. Y. Lin et. al., Nature 394, 251 (1998) (99) X (μm) Vacuum defect beam path is into the page silicon This woodpile structure is made by stacking gratings etched in silicon wafers, then etching away the substrate.

7 Fabrication of Woodpile Structures in Silicon Silicon woodpile structure produced at the Stanford Nanofabrication Facility (SNF) Detailed Tolerance Studies of CDs Process Version Rod width base Rod width top Taper Angle Layer Thickness Alignment Offset Period average std version 3 mean version 3 std version 2 mean version 2 std Layer Structure (6/08) 4 Layer Structure (10/08) Best achieved: Width Variation: <40 nm RMS (~λ/125) Layer Thickness: <65 nm RMS (~λ/75) Layer Alignment: <65 nm RMS (~λ/75) Measurement Technique Granularity: 7nm

8 cylindrical lens vacuum channel cylindrical lens laser beam top view electron beam y z x λ/2 λ r 1 E ~ 2 E laser F r = 0 T. Plettner et al, Phys. Rev. ST Accel. Beams 4, (2006) Simple Variant: Fast Deflector Silica, λ=800nm, E z =830 MV/m T. Plettner, submitted to Phys. Rev. ST Accel. Beams

9 The next level of integration: A Single-Pulse 32 MeV-Gain Woodpile Accelerator Chip (1 chip 1 ILC cavity) input beam L eff =2mm beam Distribution, delay, and mode shaping lines ~80 mm Silicon Chip beam Cutaway sketch of coupler region Input waveguide Simulation work in collaboration with Tech X (SBIR, Phase I submitted this year). Image courtesy of B. Cowan, Tech X. Fiber coupled input λ=2 μm 20 μj/pulse 1 ps laser pulse 5μm 4 layer Structure Fabrication i (completed at SNF)

10 The E-163 Facility at the NLCTA Cl. 10,000 Clean Room (Commissioned March 2007) E S B Counting Room (b. 225) Ti:Sapphire Laser System RF PhotoInjector Gun Spectrometer E 163 Optical Microbuncher Next Linear Collider Test Next Accelerator Linear Collider Test Accelerator The E163 program has advanced rapidly due to three factors: A decade of experience conducting this type of experiment at LEAP Extensive NLCTA infrastructure required modest extension to make a functioning facility Experienced help from the Test Facilities staff at every step Experimental Hall

11 800 nm 400 nm First- and Second-Harmonic COTR Output as a function of Energy Modulation Depth ( bunching voltage ) λ=800 nm Inferred Electron Pulse Train Structure 400 nm 800 nm Left: First- and Second-Harmonic COTR output as a function of temporal dispersion (R 56 ) Bunching parameters: b 1 =0.52, 052 b 2 = C. M. Sears, et al, Production and Characterization of Attosecond Electron Bunch Trains, Phys. Rev. ST AB, 11, , (2008).

12 8 6 Energy Gain/Loss (kev) After Correction for Slow Drift 1.5 Energy Gain/Loss (kev) roid Shift Centroid Shift (ke kev) Cent minutes/7000 points Phase at 800nm (radians) 0 π 2π Phase of Accelerator (radians) Binned 500/events per point Centroid Shif ft (kev) Phase at 800nm (radians) C. M. Sears, Production, Characterization, and Acceleration of Optical Microbunches, Ph. D. Thesis, Stanford University, June (2008). The first demonstration of staged particle acceleration with ihvisible iibl light! h! Effective averaged gradient: 6 MeV/m (poor, due to the ITR process used for acceleration stage)

13 Electron Source: Example chosen: FEM Metal Tip

14 I~11 e /optical cycle <I>~500 μa B>10 13 A/m 2 sr 2 500nm Scale bar 1μm Scale bar

15 Accelerator: Transversely Powered Grating Structure

16 cylindrical lens vacuum channel cylindrical lens laser beam top view electron beam y z x λ/2 λ r 1 E ~ 2 E laser F r = 0 T. Plettner et al, Phys. Rev. ST Accel. Beams 4, (2006) Peak Field [GV/m] Fluence [J/cm 2 ] Efficient: Produces 150w optical from 800 W wall plug. Pulse Length [ps]

17 Radiator: Laser Driven Deflector Structure, 1 mm period, CO2 powered

18 pulse-front tilted laser beams deflection structure sections undulator period λ u undulator period λ u

19 Resultant Photon Output Note: X ray pulse format for this example is: 6.6 fsec 1 μsec 1 μsec 50 pulses per train 33 as 5x ev 1x kev ev 4λ ev 400λ

20 Small Apertures! Small charge high repetition rate required for average brightness However, Peak currents are reasonable high (~10A) due to extremely short bunches Rayleigh Range for x ray mode tends to be short compared to the required undulator length increase field strength or find another way to shorten the gain length Different Realm: [fc, nm, fsec] vs. [nc, mm, psec] Accuracies tighter, signals weaker, but: Bunchwakefields aeedspossess harmonics well up into the EUV Perturbing the beam results in the emission of light Excellent diagnostics are available in this range Demonstrate t high gradient h acceleration and deflection in microstructures t Development of EO deflector to couple in the laser power synchronously Source development to increase current from 10 e/cycle to 1000 e/cycle Development of suitable focusing element to guide the beam through the accelerator an undulator

21 Optical acceleration naturally leads to a attosecond scale x ray pulses trains spaced at the femtosecond scale Very small emittances can be exploited to provide point like source qualities High gradient and short gain length lead to more compact machines A 500 ev class machine appears possible on a small laboratory scale ~5 m A 50 kev class machine appears possible on a the scale of a single floor of a building