30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Photocathode lasers generating long trains of flat-top pulses Ingo Will, Guido Klemz Max Born Institute Berlin 800  s 1  s 100 ps (10mm glass plate)

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Optical sampling system for high-resolution measurement of the longitudinal pulse shape Ingo Will, Guido Klemz Max Born Institute Berlin 100 ps (10mm glass plate)

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses The desired pulse trains and pulse energy (superconducting linac, Cs 2 Te photocathode) Spacing of the pulses: 1  s Spacing of the pulses: 1  s In future: 0.2  s = 5 MHz (XFEL) and 0.11  s = 9 MHz (option for FLASH)In future: 0.2  s = 5 MHz (XFEL) and 0.11  s = 9 MHz (option for FLASH) Duration of the pulse train: at least 800  s, variable Duration of the pulse train: at least 800  s, variable n Very reliable synchronization n Rectangular envelope of the pulse trains n Energy: > 100  J in the IR (I.e. = 1047 nm)> 100  J in the IR (I.e. = 1047 nm) –corresponds to >100 W power during the pulse train 15  J in the UV (I.e. = 262 nm)15  J in the UV (I.e. = 262 nm) 800  s 1  s Desired parameters (according to the requirements specified by DESY):

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Desired pulse shape n Desired parameters of the micropulses Wavelength: UV (262 nm)Wavelength: UV (262 nm) Edges: < 2 ps (UV)Edges: < 2 ps (UV) Noise in the flat-top region: < 10…20 %Noise in the flat-top region: < 10…20 % Pulse duration  ~ 20 psPulse duration  ~ 20 ps n Completely remote- controlled laser system n Very reliable synchronization 20 ps < 2 ps

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses First completely diode-pumped pulse-train laser operational at PITZ since April 2005 Conclusion: The duration of the train in the amplifiers must be much larger (>1.5 ms) than the length of the output train

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Pre-compensation of changes of the pulse shape during amplification and conversion to the UV IR =  m green =  m UV =  m

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses The MBI setup of a generator of long flat-top pulse trains for the superconducting linac photo- diode #1 photo- diode #3 wavelength conversion IR -> UV output pulses photo- diode #2 pulse shaper booster amplifier main pulse picker auxiliary pulse picker preamplifier modelocked oscillator 100 ps

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses MBI setup of a generator of long flat-top pulse trains for the superconducting linac photo- diode #1 photo- diode #3 wavelength conversion IR -> UV output pulses photo- diode #2 pulse shaper booster amplifier main pulse picker auxiliary pulse picker preamplifier modelocked oscillator n Effect of all components of the laser on the micropulses should be constant for a duration of 800  s n Components of the laser should work with 1 MHz rep. rate during the train n High average power during the train of the final amplifiers: P train > 100 W

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Selected pulse shaping techniques: suitability for single pulses and pulse trains TypeMethodeschemeeffect Single pulses Pulse trains (bursts) Linear shaping techniques Grating pulse shaper: Spectral shaper Spectral decomposition okok. Grating pulse shaper: Direct space-to-time Diffraction on a grating okok Acoustooptic filter (i.e. DAZZLER) Diffraction on a sound wave package ok Not suitable Birefringent filter filter with a transmission sin(  )/  okok

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Selected pulse shaping techniques/effects: suitability for single pulses and pulse trains TypeMethodschemeeffect Single pulses Pulse trains (bursts) Nonlinear shaping techniques Fiber shaper Self-phase modulation in a fiber okok. Nonlinear amplifying loop mirror (NALM) Nonlinear phase shift in a fiber okok Fast optical power limiter Nonlinear phase shift in a bulk medium ok Limited, depends on NL medium Nonlinear interaction in crystals (SHG, FHG, OPA) Nonlinear interaction in  3 crystals okok output pulses UV

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses output pulses recorded with a streak camera: n Flat-top laser pulses generate electron bunches with a flat-top shape in z-direction -> improved brightness of the electron beam-> improved brightness of the electron beam Simple DST shaper forming flat-top laser pulses

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Simple DST shaper forming flat-top laser pulses output pulses recorded with a streak camera: n Flat-top laser pulses generate electron bunches with a flat-top shape in z-direction -> improved brightness of the electron beam-> improved brightness of the electron beam

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Amplification of flat-top pulses from an Yb:YAG oscillator Record of flat-top pulses with a synchroscan streak camera (Optronis, ~3...4 ps resolution) at 515 nm wavelength n Parameters of the pulses shown: length of the train: 1.5 ms (1500 pulses)length of the train: 1.5 ms (1500 pulses) Energy in the train: 27 mJEnergy in the train: 27 mJ Energy per micropulse: 18  J (at 1030 nm)Energy per micropulse: 18  J (at 1030 nm) Streak camera measurement taken with SHG (at 515 nm)Streak camera measurement taken with SHG (at 515 nm) n Energy is ~ 4…5 times smaller than in the present Nd:YLF phothocathode laser n Increasing this energy is a major challenge to the laser designer 100 ps (10mm glass plate)

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses The MBI setup of a generator of long flat-top pulse trains for the superconducting linac photo- diode #1 photo- diode #3 wavelength conversion IR -> UV output pulses photo- diode #2 pulse shaper booster amplifier main pulse picker auxiliary pulse picker preamplifier modelocked oscillator 100 ps

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Some amplification techniques: suitability for single pulses and pulse trains TypeMaterial Single pulses Amplifiers for pulse trains (bursts) Optical-parametric amplifiers (OPA) BBO, LBO, KTP Ok Ok (implemented in the FLASH pump/probe laser) Laser amplifiers: Nd:YLFOkOk Edges limited to > ps (see present PITZ photocathode laser) Ti:SaOk (regenerative amplifier required)Unknown: - very strong thermal lens - Sophisticated pump laser needed Yb-doped media (Yb:KGW, Yb:YAG, Yb:CaF) Ok (regenerative amplifier required) Under development, - moderate thermal lens - regen required, but too low saturation during a single micropulse

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Part 2: OPCPA stage generating femtosecond pulses output pulse train which contains 700 micropulses OPCPA: OPCPA: Optical Parametric Chirped-Pulse amplification Generates femtosecond pulses   150 fs FWHM Generates femtosecond pulses   150 fs FWHM pulse energy available at present : E micro = 100  J (before compressor) E micro = 50  J (behind compressor) pulse energy available at present : E micro = 100  J (before compressor) E micro = 50  J (behind compressor) n Available wavelength: = 790…830 nm = 790…830 nm on request: = 395…415 nmon request: = 395…415 nm

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses OPCPA stage Pump laser Scheme of the Pump-Probe laser

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Regenerative amplifiers can be made to work at 1 MHz repetition rate n Specialty in burst mode: Each micropulse can extract only a small fraction (~ 0.2%) of the stored energy Low stability (2% fluctuation during the train)Low stability (2% fluctuation during the train) Reduced efficiency (~50%) in comparison to single pulsesReduced efficiency (~50%) in comparison to single pulses Failure in the trigger will damage the amplifierFailure in the trigger will damage the amplifier –sophisticated software solution, (present DOOCS not save) –NL limiter –Fast repair technology Pulse traveling in the resonator Output pulses 20  s 1ms (1000 pulses) 50 ns

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Two-stage regenerative amplifier concept n Thermal lens in the power regen leads to a drop of the intensity to 50% during 2000 pulses n The two- or three-stage regen concept may enable us to apply advanced amplifier techniques (i.e. thin-disk amplifiers) First regen Second regen E micro = 15  J E micro = 3  J 2ms (2000 pulses) Yb:KGW oscillator Yb:YAG regen Yb:YAG power regen DST shaper Drop due to thermal lensing First regen Second regen Compensation of the drop by the drive current of the pump diodes, but the „pumping“ of the beam diamter remains! 2ms (2000 pulses) E micro = 15  J E micro = 3  J

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Amplification of flat-top pulses from an Yb:YAG oscillator Record of flat-top pulses with a synchroscan streak camera (Optronis, ~3...4 ps resolution) at 515 nm wavelength n Parameters of the pulses shown: length of the train: 1.5 ms (1500 pulses)length of the train: 1.5 ms (1500 pulses) Energy in the train: 27 mJEnergy in the train: 27 mJ Energy per micropulse: 18  J (at 1030 nm)Energy per micropulse: 18  J (at 1030 nm) Streak camera measurement taken with SHG (at 515 nm)Streak camera measurement taken with SHG (at 515 nm) n Energy is ~ 4…5 times smaller than in the present Nd:YLF phothocathode laser n Increasing this energy is a major challenge to the laser designer 100 ps (10mm glass plate)

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Amplification of long pulse trains for the cold linac by an Yb:YAG booster n Stable pulse train: control of the ramp of the current of the pump diodes n Stable beam diameter: beam-shaping aperture at the output Can the beam-shaping aperture in the beamline play this role?Can the beam-shaping aperture in the beamline play this role? n Technology for lossless stabilisation of the beam diameter: fast deformable mirror E micro = 15  J E micro = 3  J n Energy per micropulse: ~ 15  J n Amplification (two stages): G = ms (5000 pulses) Without compensation by pump current 5 ms (5000 pulses) with compensation by pump current

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Shortest pulses and bandwidth of this amplifier combination n Output pulses of the KGW oscillator:  = 0.5 ps n Output pulses of the regen combination:  = 1.8 ps n Can pulses of this duration efficiently be transferred to the UV (forth harmonics, = 258 nm) ? E micro = 2x7  J E micro = 3  J Yb:KGW oscillator Yb:YAG regen Yb:YAG power regen E micro = 2x0.1  J 12ps 2ps d = 1.2mm

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Shortest pulses and bandwidth of this amplifier combination E micro = 2x7  J E micro = 3  J Yb:KGW oscillator Yb:YAG regen Yb:YAG power regen E micro = 2x0.3  J 12ps 2ps n Output pulses of the KGW oscillator:  = 0.5 ps n Output pulses of the regen combination:  = 1.8 ps n Can pulses of this duration efficiently be transferred to the UV (forth harmonics, = 258 nm) ?

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Two-channel mixing scheme: reduced energy requirements to the broadband laser amplifier  75% of the total laser energy delivered by the Nd:YLF long- pulse system  only 25% need to be delivered by broadband channel Broadband pulse: from Yb:KGW laser - sharp edges - E micro = 20  J - =1038 nm Narrowband pulse from Nd:YLF laser - slow edges - E micro = 100  J - = 349 nm BBO crystal UV output pulse - sharp edges - E micro > 20  J - = 260 nm to photocathode beam stop beam stop

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses present scheme pre-shaping by an aspherical Lens pair Pre-shaping the beam of the photocathode laser may significantly reduce losses in the beamline

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Summary: Long trains of flat-top laser pulses for the superconducting linac n Pulse shaping techniques: Only minor limitations n Most linear pulse-shaping techniques (gratings, filters etc) work for long trains n Techniques based on travelling acoustic waves (i.e. DAZZLER) cannot be used nNonlinear techniques: limited duration of the pulse train ( micropules) (only for nonlinear techniques using bulk materials) n Amplifiers: 1.Optical-parametric amplifiers (OPA): work without restrictions, but large pump laser required 2.For laser amplifiers: The broadband laser materials require regenerative amplifiers in the first stages. These regens can work at 1 MHz for Yb:YAG, Yb:KGW: n With somewhat reduced stability and with slightly less less efficiency (~ 50%) than for single pulses Reason: low saturation during each micropulse (0.2% energy extraction per pulse) 3.Linear power amplifiers: work well with pulse trains n Some problems arise from the thermal lens, that drifts during the train Solution: Dynamic correction with fast deformable mirrors 4.Ti:Saphire: No solutions for long, intense pulse trains available  We have made the correct choice for the laser material: Diode-pumped Yb:KGW and Yb:YAG instead of Ti:Saphire

30 Nov. 06 I.Will et al., Max Born Institute: Long trains of flat-top laser pulses Amplification of flat-top pulses from an Yb:YAG oscillator Record of flat-top pulses with a synchroscan streak camera (Optronis, ~3...4 ps resolution) at 515 nm wavelength n Parameters of the pulses shown: length of the train: 1.5 ms (1500 pulses)length of the train: 1.5 ms (1500 pulses) Energy in the train: 27 mJEnergy in the train: 27 mJ Energy per micropulse: 18  J (at 1030 nm)Energy per micropulse: 18  J (at 1030 nm) Streak camera measurement taken with SHG (at 515 nm)Streak camera measurement taken with SHG (at 515 nm) n Energy is ~ 4…5 times smaller than in the present Nd:YLF phothocathode laser n Increasing this energy is a major challenge to the laser designer 100 ps (10mm glass plate)