1. Positron Sources for Linear Colliders* Wei Gai JPOS 2009, Jefferson Lab, March 26, 2009 * Acknowledgement of contributions from the ILC and CLIC e+ collaborations

2. JPOS09, JLab, Newport News, VA, March 26, 2009 Content Overview Undulator Based Positron Source Conventional Positron Source Compton based Positron Source

3. JPOS09, JLab, Newport News, VA, March 26, 2009 Overview

4. JPOS09, JLab, Newport News, VA, March 26, 2009 positron production e- beam (multi MeV – hundreds of GeV) Gamma generation Gamma ray Conversion target Capturing optics Acceleration Planar/Helical Wiggler/Undulator Bremsstrahlung /Channeling radiation Laser Compton scattering Gamma generation schemes Positrons

5. JPOS09, JLab, Newport News, VA, March 26, 2009 Helical undulator Based Scheme: requires very high energy drive beam (~100 GeV) Undulator technology is straightforward. (SC or PM) Supper conducting helix i -i Can produce circularly polarized photon, good for polarized e+ source. Drive beam energy: 150GeV Proposed: A. Mikhailichenko K. Flottmann, et al

6. JPOS09, JLab, Newport News, VA, March 26, 2009 Conventional (Bremsstrahlung) Drive beam energy can be as low as ~ 100 MeV. + E1E1 E2E2 e e h  =E2-E1 6GeV e- e- e+ 4 X0 tungsten target Here, bremsstrahlung refers to radiation from electrons stopping in matter. RF LINAC AMD If the incident electron is polarized, the photon produced will be circularly polarized. And this can give us a possible polarized e+ source using conventional scheme.

7. JPOS09, JLab, Newport News, VA, March 26, 2009 Channeling radiation -- Coherent bremsstrahlung (separateγand e+ production) Schematic illustration of channeling Enhancement can be as high as 40 comparing with incoherent bresstrahlung (R. Chehab et al.) An example of positron source using channeling radiation

8. JPOS09, JLab, Newport News, VA, March 26, 2009 Laser Compton scattering Multi GeV e- Circularly polarized YAG Laser or CO2 Laer Circularly polarized  Mr. Omori-San’s favorite drawing

9. JPOS09, JLab, Newport News, VA, March 26, 2009 Undulator based sources for ILC and CLIC

10. JPOS09, JLab, Newport News, VA, March 26, 2009 ILC (500 GeV CM) Positron Source Layout (undulator based scheme)

11. JPOS09, JLab, Newport News, VA, March 26, 2009 Beam parameters for different machines

12. JPOS09, JLab, Newport News, VA, March 26, 2009 Photon Spectrum and Polarization of ILC baseline undulator 1.Photon energy spectrum and polarization from a ILC “baseline” undulator (K=1, lu=1cm and Edrive =150GeV) up to the 9 th harmonics. 2.Note p hotons close to critical energy (also near axis) for each harmonic have higher polarization. Collimating incoming photons will result polarized e+.

13. JPOS09, JLab, Newport News, VA, March 26, 2009 Target Energy Deposition Profile: Energy deposition profile showing here is calculated per drive e- bunch Energy deposition in target is about 0.5255J per bunch Energy deposition : about 1482J per pulse Power deposition 1482(J)/0.874e-3(s) ~= 1.696MW per pulse Average power deposition: 1482*5=7.4KW Ti target Target has to be rotating at high speed to survive Rotating the 2m diameter target wheel at 1000rpm was estimated for safe operation of the target.

14. JPOS09, JLab, Newport News, VA, March 26, 2009 Energy and polarization distribution e+ source at the target Large energy spread

15. JPOS09, JLab, Newport News, VA, March 26, 2009 Transverse phase space distribution at the target Large divergence, high emittance beam

16. The ILC Collaboration Meeting, IHEP, Beijing, Jan 31 – Feb 2, 2007 Positron collection and acceleration: Adiabatic Matching Device (target immersed in a solenoid B field) L-band Standing Wave Accelerator. AMD field:5T-0.25T in 50cm Accelerating gradient in pre-accelerator: 12 MV/m for first 6 m, 10 MV/m for next 6 m and 8.9 MV/m for the rest.

17. JPOS09, JLab, Newport News, VA, March 26, 2009 Comparison of positron yield from different undulators High K DevicesLow K Devices BCDUK IUK IIUK IIICornell ICornell IICornell III Period (mm)10.011.511.010.510.012.0 7 K1.000.920.790.640.420.72 0.3 Field on Axis (T)1.070.860.770.650.450.64 0.46 Beam aperture (mm)Not Defined 5.85 8.00 First Harmonic Energy (MeV) 10.710.112.014.418.211.7 28 Yield(Low Pol, 10m drift)~2.4~1.37~1.12~0.86~0.39~0.75~0.54 Yield(Low Pol, 500m drift, 25%) ~2.13~1.28~1.08~0.83~0.39~0.7~0.54 Yield (Pol. 60%)~1.1~0.7~0.66~0.53~0.32~0.49~0.44 Target: 1.42cm thick Titanium

18. JPOS09, JLab, Newport News, VA, March 26, 2009 Proposed ILC target geometry and simulation of the target rotating in magnetic fields. 1m 1.4cm Solenoid positioned at 0.95m The model is checked against known experiments.

19. JPOS09, JLab, Newport News, VA, March 26, 2009 Power vs RPMs for the ILC Target

20. JPOS09, JLab, Newport News, VA, March 26, 2009 Cockroft institute prototype experiment simulation Technical drawing provided by I.Bailey Simulation, Induced field, z-component, 2000RPM z0z0 D – 1m, rim width – 30mm, rim thickness – 14mm, distance between magnet poles is 5cm, field – 1.5Tesla

21. JPOS09, JLab, Newport News, VA, March 26, 2009

22. Another proposed solution: A pulsed flux concentrator Pulsing the exterior coil enhances the magnetic field in the center. –Needs ~ 1ms pulse width flattop –Similar device built 40 years ago. Cryogenic nitrogen cooling of the concentrator plates. –ANL and LLNL did initial rough electromagnetic simulations. Not impossible but an engineering challenge. –No real engineering done so far.

23. JPOS09, JLab, Newport News, VA, March 26, 2009 Advanced Solution: Lithium lens Lithium Lens –Will lithium cavitate under pulsed heating? window erosion –Will lithium flow adequately cool the windows? –Increased heating and radiation load in the capture section –Needs R&D to demonstrate the technology. P.G. Hurh & Z. Tang A. Mikhailichenko A. Mikhailichenko et al.

24. JPOS09, JLab, Newport News, VA, March 26, 2009 What if every capturing magnet technology fails, a safe solution: ¼ wave solenoid Low field, 1 Tesla on axis, tapers down to ¼ T. Capture efficiency is only 25% less than flux concentrator Low field at the target reduces eddy currents This is probably easier to engineer than flux concentrator SC, NC or pulsed NC? ANL ¼ wave solenoid simulations W. Liu

25. JPOS09, JLab, Newport News, VA, March 26, 2009 Summary of Capture Efficiency for Different AMD AMDCapture efficiency Immersed target (6T-0.5T in 20 cm) ~30% Non-immersed target (0-6T in 2cm, 6T-0.5T 20cm) ~21% Quarter wave transformer (1T, 2cm) ~15% 0.5T Back ground solenoid only~10% Lithium lens~29%

26. JPOS09, JLab, Newport News, VA, March 26, 2009 Undulator based e+ for CLIC (3 TeV) J. Sheppard L. Rinolfi, W. Gai

27. JPOS09, JLab, Newport News, VA, March 26, 2009 Undulator K = 0.75 λu = 1.5 cm L = 100 m Pre-Injector Linac G = 12 MV/m E = 200 MeV f RF = 1.5 GHz B = 0.5 T Injector Linac G = 17 MV/m E = 2.424 GeV f RF = 1.5 GHz f rep = 50 Hz A possible CLIC scheme for polarized e + 250 GeV Cleaning chicane NC Linac 2.2 GeV To the IP e - beam Ti alloy 450 m e+e+ e+e+

28. JPOS09, JLab, Newport News, VA, March 26, 2009 A possible CLIC complex layout with undulator based e+ source Following the tunnel back to e+ injector e- e+ Booster linac e+ main linac e- main linac e+ capturing optics and preaccelerator undulator Bending assemblies, 20 of them, each one bends the electron beam by 1/20 of the angle between axis of undulator and the axis of the rest of electron main linac target >2m e+ capturing etc undulator

29. Numerical Simulation on the effect of undulator parameter and accelerating gradient Drive e- beam energy: 250GeV Undulator parameters: K = 0.5 - 0.75, λ= 1.3 - 1.5cm, L= 100 m Drift to target: 450m Accelerator L-band Linac, AMD: 7T - 0.5T in 20cm; Target material: 0.4 rl Titanium, Positron capture is calculated by numerical cut using damping ring acceptance window: +/-7.5 degrees of RF(1.3GHz),  x+  y<0.09  m.rad,1% energy spread with beam energy ~2.4GeV

30. Yield and polarization for the CLIC undulator based source Yield is calculated as Ne+ captured/Ne- in drive beam Bottom line: It works

31. Conventional e+ for LCs

32. e- AMD ~ 120 MeV PPA Target Superconducting linacs With quadropole focusing 5 GeV e+ To damping ring The original ILC conventional source schematic layout Target MaterialW23Re Length4.5 RL Electron Beam Energy0.25 - 6 GeV Transverse size, σx = σy2 mm Longitudinal size, σt 1.5 ps Polarized electron →polarized positron (?) After sweeping through the parameter space, this original scheme seems to be not viable for ILC due to the excessive energy deposition in target.

33. Courtesy of M.Kuriki

35. Liquid metal target (BINP design)

36. Liquid metal target development Lead flow

37. Cog-wheel pump test bench (BINP)

38. Temperature distribution using ILC beam time structure: 600MeV drive beam, 1mm spot size, AMD immersed target (130 and 260 bunches) Too hot to handle!!!!!!!!! Ways to improve: higher energy, larger spot size and increasing flow rate 260 bunches 130 bunches x z x z

39. Need 30m/s pumping speed to keep the liquid from boiling.

41. Time structure of 300Hz conventional source Courtesy of T.Omori Output timing structure from DR per ILC specs Advantage: Only deal with 132 pulse each time Low speed target

42. Temperature in target after 2 triplets Target is moving at 10m/s

43. JPOS09, JLab, Newport News, VA, March 26, 2009

45. Compton Based Scheme

46. JPOS09, JLab, Newport News, VA, March 26, 2009

47. Photon spectrums a CO2 laser compton scattering with 3 different drive beam energy Photon number is high but the interaction time is short. Total number of photon produced is small. Stacking is needed.

48. JPOS09, JLab, Newport News, VA, March 26, 2009

50. F. Zimmerman et al.

51. JPOS09, JLab, Newport News, VA, March 26, 2009

53. Summary Three schemes discussed here, each scheme has pluses and minuses, Due to the designed pulse structure, the ILC source is the most difficult one. Intensive R&D are on going, there will be solutions. Seems no fundamental issues with the CLIC scheme. Looking forward to build a linear collider in my life time.

54. Heat transfer simulation up to 2650 bunches, 700MeV, 4mm rms spot size, 30m/s pumping speed, Lithium lens Temperature distribution Interpolated temperature on line (z=1.2cm,x=0) at different time Hot spot temperature changing with time With 4mm rms spot size, when using lithium lens, the yield is about 0.27. Using the yield of about 0.31 when using AMD with 1mm rms spot size and energy deposition of about 2.27J per bunch, the deposited energy 4mm spot with lithium lens can be estimated as 2.626J per bunch. With the scaled energy deposition profile and deposited energy, heat transfer simulation for 700MeV, 4mm rms spot, 30m/s pumping speed and optimized lithium lens shows the temperature after 2650 bunches is still bellow 1600K.