Temporal overlapping for HHG- seeded EUV-FEL operation by using EOS-based timing-drift controlling system H. Tomizawa 1,4 *, S. Matsubara 1, T. Togashi 1, E. J. Takahashi 2, K. Midorikawa 2, M. Aoyama 3, K. Yamakawa 3, T. Sato 4, 5, A. Iwasaki 5, S. Owada 5, K. Yamanouchi 5, K. Togawa 4, 1, T. Hara 4, 1, K. Ogawa 4, T. Ohshima 4, 1, Y. Okayasu 1, 4, Y. Otake 4, 1, H. Tanaka 4, 1, T. Tanaka 4, 1, T. Watanabe 1, 4, M. Yabashi 1,4, and T. Ishikawa 4 1. Japan Synchrotron Radiation Research Institute 2. RIKEN Advanced Science Institute 3. Japan Atomic Energy Agency 4. RIKEN SPring-8 Center 5. The University of Tokyo

Overview & History EOS Setup to improve the seeded FEL performance Recent Results Summary “HHG seeded EUV-FEL operation at SCSS Test Accelerator” Outline

Development history of X-ray light source Seeded X-ray FELs “SACLA” XFEL (JAPAN) “SPring-8” 3 rd generation SR-Ring (JAPAN)

Spontaneous radiation e-beam 電子銃 加速管 e-gunAcceleratorUndulator X-ray laser ☺ Intense light pulse (  J ~ mJ ) * ☺ Wide wavelength range ( THz ~ x-ray ) ☺ Short pulse ( 10 fs ~ 1 ps ) ☺ Spatial coherence ☹ Shot-to-shot fluctuation * FLASH (Germany), SCSS (Japan), LCLS (USA), SACLA (JAPAN), … Self-amplification of spontaneous emission (SASE)

SCSS EUV-FEL Accelerator SCSS accelerator ( 250MeV, 50-60nm ) – EUV-FEL machine is operated for user experiment. – Development of HHG-seeded FEL 5 Electron Gun Bunch Compressor C-band Accelerator Undulator

Fluctuating spectra – SASE operation at SCSS SASE-FEL starts up from noise Energy spectra and temporal profiles are fluctuating shot-to-shot! (originated from the SASE process) External seeding scheme is one of the solution to suppress shot-noise for reliable full-coherent light source (user operation).

Our history of Seeding Developments DateEventConditionReference June 2006 The first SASE amplification with our new machine concept 250 MeV, 49nm Dec Seeding at 160 nm 150 MeV, HHG 5 th G. Lambert et al., Nat. Physics 4, 296 (2008) Sept SASE saturation250 MeV, 50~60nm T. Shintake et al., Nat. Photonics. 2, 559 (2008) Oct Seeding at 61 nm Hit rate: ~0.3% Pulse energy : ~2  J 250 MeV, 300 fs HHG 13 th T. Togashi, et al., Opt. Exp. 19, 317 (2011) March 2011 The first test of Arrival time monitor (relative timing btw. e-bunch and HHG with EO sampling) H. Tomizawa, BIW2012, Newport News, VA (2012) July 2012 Seeding at 61 nm with EO sampling Hit rate: ~30% Pulse energy : ~20  J 250 MeV, 600 fs HHG 13 th H. Tomizawa, et al., LINAC2012, Tel-Aviv (2012)

Task force in our collaboration for HHG-seeding Supports for this projects: RIKEN/JASRI XFEL project SCSS test accelerator operation team (Engineers) Financial supports : RIKEN extreme photonics MEXT X-ray free electron laser utilization research (The University of Tokyo) Japan Atomic Energy Agency, Quantum Beam Science Directorate A. Iwasaki K. Ogawa, Y. Okayasu, H. Tomizawa, T. Togashi, T. Sato, S. Owada T. Watanabe, E. J. Takahashi, S. Matsubara

6D phase space overlap for seeded FEL 9 Size (x, y) FWHM Time FWHM Wavelength ( Energy) Electron bunch ~ 500  m~ 600 fs61.2 nm HHG seed pulse ~ 1 mm~ 50 fs61.4 nm 6D Phase Space X,  x Y,  y t, E

Seeding results at 61 nm in 2010 (1) 61nm-2nJ Undulator HHG Spectrum ← Temporal overlap ↓ Spatial overlap

Spectra of sequent 50 shots T. Togashi et. al., Optics Express, Vol. 19 Issue 1, pp (2011) Trend graph of peak intensity Hit rate 〜0.3% (Intensity >4σ of SASE) SASE FEL: 0.7  J Seeded FEL: 1.3  J but, hit rate ~ 0.3% Duration of seeded operation < 10 min SASE FEL: 0.7  J Seeded FEL: 1.3  J but, hit rate ~ 0.3% Duration of seeded operation < 10 min Seeding results at 61 nm in 2010 (2)

Improvement of Hit Rate (~ 2012 ) (1) 61nm-2nJ Arrival time monitor by means of EO-sampling implemented

Principle of EOS ( Electro-Optic Sampling ) Pockel’s effect (ZnTe) Z(110) (001) p p By detecting the retardation (Modulation of Polarization), we can know the temporal information of electron bunch (Electric field). Spectral Decoding

61nm-2nJ Late Early Wavelength (time) Improvement of Hit Rate (~ 2012 ) (2)

Relative timing-drift monitored by EOS “Relative” timing drift between the electron bunch and the laser pulse The spectra of EOS signal pulses decoded as the timing shifts from the best seeding condition (On-time) Late +1 ps Early -1 ps The arrival-time drift of electron bunch with respect to the central wavelength of EO-probe pulse: ~50 ps for ½ day The arrival-time drift is calculated automatically with the computer program in terms of the peak position of the EOS signals. On-time

Seeded FEL Performances (2012) (1) Seeded SASE SASE+4  SASE+6  Seeded SASE SASE+4  SASE+6  +4  hit rate: 24% Improved by 100! Seed HHG: 2 nJ x 10 4 gain Improved by 10! +4  hit rate: 24% Improved by 100! Seed HHG: 2 nJ x 10 4 gain Improved by 10!

Seeded FEL Performances (2012) (2) The correlation data plot between the normalized intensity and central wavelength for shot data The correlation data plot between the normalized intensity and central wavelength for shot data The standard deviation of the central wavelength of SASE (blue) and the seeded pulse (red) were 0.15 nm and nm, respectively.

Seeded FEL Performances (2012) (4)

Improvements of FEL Performances (2010  2012) w/o feedback w/ feedback Previous result (2010) This result (2012) By using the EOS-based timing-drift system, the HH seeded FEL succeed to continuously operate about a half day which is the machine time of SCSS accelerator with 20-30% hit rate. ･ Seeded FEL output was 1.3  J ･ The seeding operation was only obtained less than 10 minutes.

Summary Recent R&D to improve the hit rate at the SCSS FEL test accelerator has proven the capability of stable HHG-seeded FEL. – EOS-based timing-drift controlling system was constructed for the continual seeded FEL operation.  The seeding laser pulse has large timing drift with respect to electron bunch of ~50 ps for ½ day.  The relative timing-drift controlling system was needed to stable operation. – Improved HHG seeded FEL operation was succeed with 20  J and ~30% hit-rate.