1. The Thomson Source C. Vaccarezza on behalf of the SPARC_LAB team

2. Outline  SL_Thomson source description  1st commissioning phase result  2 nd commissioning phase results  What’s next EAAC 2015 13-19 September La biodola, Isola d'ELba

4. The FLAME building 4

5. Repetition Rate10 Hz Energy (after compression) up to 6 J (typ. exp. 5.6J) Wavelength 800 nm Pulse duration down to 20 fs (typ.23 fs) Peak power up to 300 TW ASE contrast < 10 10 Pre-pulse contrast < 10 -8 Front end @ FLAME laser: specifications 5

6. SL-Thomson Source Parameters Laser  Wavelength 800 nm  Compressed pulse energy1-5 J  Pulse duration/bandwidth 3 -12 (80) ps/nm  Rep.Rate 10 Hz  Beam quality M2 <1.5  Energy stability 10%  Pointing stability < 2 µm  Synchronization with SPARC photoinjector clock Electron Beam  Bunch charge0.2-1.0 nC  Energy 28-150 MeV  Length10-20 ps  ε n x,y 1-5 µ rad  Energy spread0.1-0.2 %  Spot size at IP10-20 µm  Rep.Rate 10 Hz EAAC 2015 13-19 September La biodola, Isola d'ELba 6

7. Applications 20-500 keV  Mammography  Identification of fissile materials  Cristallography  Microdensitometry 3D for cultural heritage ELI- γ BS test bench for basic machine systems and design EAAC 2015 13-19 September La biodola, Isola d'ELba Operating modes:  High-flux-moderate-monochromaticity mode(HFM2 ) (medical imaging when high-flux sources are needed  Moderate-flux- monochromatic-mode (MFM ) (detection/dose performance optimization)  Short-and-monochromatic-mode (SM) ( pump- and- probe experiments e.g. when tens of fs long monochromatic pulses are needed For the first experiments the source has been optimized for applications in medical imaging, mammography in particular (1 st financed experiment at SL_Thomson: “MAMBO-BEATS”). Highest X-ray flux at about 20 keV, with a relative radiation energy spread < 20% and keeping as low as possible the high harmonics contribution.

8. From theory *-P. Tomassini et al Appl. Phys. B 80, 419-36, 2005 -P. Tomassini et al. IEEE Trans. on Plasma Sci. 36,n.4, 2008 8

9. Total photon distribution Photon energy distribution Spectral-angular distribution  max = 1/  Photon Distribution

10. HFM2 WP EAAC 2015 13-19 September La biodola, Isola d'ELba for Spectral–angular (integrated in the azimuthal angle φ ) distribution of the collected radiation for the optimized parameters w = 15 μ m and duration T = 6 ps, where ϑ M = 8 mrad Spectra of the collected radiation for pulses of duration 3, 6, 9, and 12 ps.

11. 30 MeV WP: 10% 1.5 10 9  e-beam:  Charge Q =0.25-1×10 -9 C  Energy E =30 MeV  σ xy ≈ 15-20 μ m  photons:  energy= 22 keV   = 7 10 -4,  =0.8-1.5mm-mrad  Interaction angle = 0°

12. 1 st commissiong phase (Feb. 2014) 50 MeV Working point 12  Charge (daily avrg):Qav= 200 ± 20 pC  Energy (daily avrg):Eav= 50.6 ± 0.2 MeV  Launch phase: Φ = + 30°  Laser pulse length = 8.3 ps  1 st TW S1 phase: Φ = - 20°  2 nd TW S2 phase: Φ = + 65°  3 rd TW S3 phase: Φ = - 90°  Bunch length rms = 3.1 ps

13. Thomson x-rays signal in red, in black the electron background signal (without FLAME laser), integrated over 120 s (1200 pulses) We calculate that the number of photons per each pulse, coming from poor overlap conditions, and interacting with the detector sensitive area, is in average 6.7x10 3. Spectral density S (MeV -1 ) versus photon energy. (50 MeV electron beam, with 200 pC charge, 5 mm mrad of emittance, 150 mm of rms beam transverse dimension, colliding with the laser with 500 mJ and 30 mm of waist, gives a number of photons of 2×10 5 in a bandwidth of about 19%. The photon energy edge, given by E p ~4E L g 2, is about 63 keV. Experimental results

14. The signal is (almost) completely attenuated by 0.5 mm of Pb. This is in agreement with the attenuation characteristics of the expected X-rays

15. 2 nd comm. phase (3 weeks shift on Jun 2015): 30 MeV Working point 15  Charge = 100-200 ± 20 pC  Energy = 29-31 ± 0.2 MeV  Launch phase: Φ = + 30°  Laser pulse length = 4.0 ps  1 st TW S1 phase: Φ = +32°  2 nd TW S2 phase: Φ = - 72°  3 rd TW S3 phase: Φ = - 134°  Bunch length rms = 2.2 ps

16. For a better description of the real data, a quasi-real distribution is generate by using the virtual cathode image (transverse distributin) and the Cross- Correlator (longitudinal distribution). Virtual cathode CrossCorre lator Importing image data by software Points image 2D- histogram 1D- histogram Simulation of the 30 MeV WP from the cathode to the Linac exit

17. Simulations and comparison with real data show a good agreement A small discrepancy exists between the mesured Δ E/E (~1 ‰) and the simulated one (5 ‰). In the simulation the Δ E/E minimum is reached before the exit; this resut is the best trade-off between data and simulations @ 12 meters : linac exit and dogleg entrance. Red dots show envelope measurement values vs calculated ones I TW II TW III TW

18. Envelope and Twiss parameters from Trace3D EAAC 2015 13-19 September La biodola, Isola d'ELba

19. 2 nd comm. Phase: e- beam at IP ( ε nx ≈ 4 μ rad )  Minimum spotsize X Sigma (mm)=7.85E-2±(4.7E-3) Y Sigma (mm)=6.37E-2±(2.7E-3)  Collision spotsize (bkgr limit) X Sigma (mm)=1.11E-1±(9.4E-3) Y Sigma (mm)=1.10E-1±(4.0E-3) EAAC 2015 13-19 September La biodola, Isola d'ELba

20. S2E calculation of e - beam Transverse Distribution (from e - data)  x =112 μ m  y =110 μ m Phase space @ IP

21. Intensity, frequency, frequency variance of the radiation distribution at 1 m Radiation Intensity Photon Energy distribution keV/h ν las Photon Energy Distribution variance

22. For the commissioning beam: EAAC 2015 13-19 September La biodola, Isola d'ELba with: U L ≈ 2 J, Q≈ 200 pC, δ Φ =0.2, h ν =1.55eV σ x,y ≈110 μ m, w o ≈150 μ m

23. Radiation detectors EAAC 2015 13-19 September La biodola, Isola d'ELba Filters and Si PIN diode (monitor) Si PIN diode (flux meas.) and imager CsI scintillator and PMT I.P. @ 201 cm @ 305 cm @ 450 cm

24. Radiation measurements  Measurement of the charge produced in a Si PIN diode (Hamamatsu, 28mm x 28mm x 0.3mm), by an electrometer (Keithley)  No polarisation, low dark current -> high sensitivity  Wide linear dynamic range  Low efficiency -> low background radiation signal  Charge Vs flux, for a known incident spectrum, previously calibrated using monochromatic source (Elettra synchrotron)  Charge measured (depends on conditions, around 10 pC per pulse) compatible with 10 4 photons (@ 20 keV) per pulse on the detector sensitive volume EAAC 2015 13-19 September La biodola, Isola d'ELba

25. Radiation measurements  Hamamatsu imager Flat Panel C9728DK-10  CMOS + 165 micron CsI  Pixel pitch 50 micron, 1032 x 1032 EAAC 2015 13-19 September La biodola, Isola d'ELba Exposure time 1 s, average over 100 images

26. Radiation measurements  Two k-edge filters inserted simultaneously, each covering partially the irradiation field  Nb (50 micron) -> k-edge 19 keV  Zr (75 micron) -> k-edge 18 keV EAAC 2015 13-19 September La biodola, Isola d'ELba Nb filter Without filter Zr filter At this signal level the detector is working at the limit of its dynamic range, noise is quite high and response is not perfectly linear, this implies that quantitative measurements are very difficult.

27. Radiation measurements EAAC 2015 13-19 September La biodola, Isola d'ELba Without filter Zr filter Nb filter  Using the radiography of k-edge filters it would be possible to see the discontinuity of the absorption corresponding to lower signals in the regions where the energy is higher than the k-edge. In our case Zr and Nb (19 and 18 keV) do not show a clear discontinuity of absorption. Considering that the emission centre doesn’t appear in the radiograph, the energy distribution of the radiation on the detector could be lower than the k-edge’s used.  The measurement of the absorption by different filters allows to estimate the incident energy distribution, comparing the measured values with the theoretical ones. As shown in the plot below, i n the region highlighted in green, the energy is compatible with 13.5 keV

28. Spectral simulation with a real electron beam transported up to the IP E peak =22.5 KeV  E/E (FWHM)=21% Spectrum of the photons at the IP Acceptance angle: 9 mrad (parabolic mirror hole acceptance) Jitters analysis

29. Beam-alignment procedure EAAC 2015 13-19 September La biodola, Isola d'ELba 0  The reference line for the Flame laser beam alignment is the one passing through the magnetic centers of the focusing quadrupole triplet and the IP solenoid as coming from the magnetic measurements.  This reference line is adopted to set the He-Ne laser at the entrance of the last dogleg dipole to sight the light path up to the Thomson radiation window exit  In this way the alignment laser image on the two screens provide the guide line for the Flame laser beam.  Unfortunately due to the achievable setting precision of the He-Ne laser and its spot enlargement at the radiation exit window the final offset of this line can be of the order of ± 4-5 mm respect to the electron beam reference line.  This means that a superposition of the two beams on the IP screen by means of the steerer magnets can lead to a non zero angle of the electron beam trajectory at the entrance of the focusing solenoid. 012345 z (m) Dip Q Q Q Sol BPM CHV PMMA window

30. IP Focusing Solenoid EAAC 2015 13-19 September La biodola, Isola d'ELba

31. Off axis trajectory effects w the solenoid field map A 4-5 mm offset at the exit of the final window (about 3.5m) can mean a non negligible offset and angle at the entrance of the focusing solenoid, much more severe as long as lower is the electron beam energy. Qualitative example: Δ x,y= 1 mm Δ yp=1mrad E= 30 MeV -150 MeV Perspex window Φ =35 mm Figures: centroid trajectory up to the window EAAC 2015 13-19 September La biodola, Isola d'ELba 30 MeV 150 MeV

32. Conclusions  The 2 nd Commissioning phase of the SL_Thomson source took place in the June 2015 dedicated shift.  The 30 MeV e - beam energy WP has been addressed as foreseen for the first imaging experiment.  With the available hw (phase shifters on 3 TW’s) the applied acceleration/deceleration scheme worked well enough to produce a low energy spread e - beam at 30 MeV, even though resulting in a strong sensitivity for the e - beam to the machine imperfections/stability.  Nevertheless there is still room for improvements towards the nominal performance EAAC 2015 13-19 September La biodola, Isola d'ELba

33. Next: Collision Optimization  4-button BPM installation downstream the IP and on the dumping pipe to control the e beam trajectory inside the solenoid  Referenced Pinhole insertion upstream and downstream for FLAME laser beam alignment  Radiation exit window replacement DN100 vs DN40 (done)  Tapered pipe replacement w straight one (DN100) (done) to reduce background and allow to squeeze the e-beam down to nominal size  1 cm shift (vs lattice model) in one 2 nd dogleg quad EAAC 2015 13-19 September La biodola, Isola d'ELba

36. The synchronization system The synchronous arrival of electrons and photons at the IP is obtained by locking precisely the oscillators of the photo-cathode laser and interaction laser systems, and the phase of the RF accelerating fields to a common Reference Master Oscillator (RMO). The RMO is a low phase noise (60 fs RMS integrated in the 10 Hz ÷ 10 MHz range) μ –wave oscillator tuned at the Linac main frequency 2856 MHz. The laser oscillators are locked through a PLL architecture to the 36 th sub-harmonics of the RMO, while the output RF phase of the linac klystrons is downconverted to baseband by mixing with the RMO signal, and deviations are corrected both within the 4 μ s RF pulse duration (jitter feedback) and pulse-to-pulse (drift feedback). EAAC 2015 13-19 September La biodola, Isola d'ELba

37. The synchronization system 37

38. The synchronization system 38

39. Characterization of the source Insurgence of non linear effects Quantum electron grouping X two colors radiation Experiments on Thomson@SPARC-LAB X EeEe

40. Transverse coherence Coherence length= R/d 0.7 10^-10*10/100 10^-6 = 7 10^-6 m