Photoinjector Lasers for UltraBright Electron Sources Graeme Hirst

Photoinjector Lasers for Ultra-Bright Electron Sources Graeme Hirst STFC Central Laser Facility

Background The CTF 3 photoinjector laser (developed with Marta Divall and Ian Ross - picture shows Ian Musgrave and Gabor Kurdi, 2006) The ERLP photoinjector laser (with Marta Divall, Gary Markey and Fay Hannon 2005)

Low-Emittance PI Laser Requirements Laser beams can be characterised in terms of four parameters: WAVELENGTH (photon energy, tunability) POLARISATION (may be dictated by technology choices) TEMPORAL PROFILE (pulse shape, time-structure of pulse train, timing jitter) TRANSVERSE PROFILE (intensity distribution, ‘pointing’ stability, could, perhaps, be dynamic ? )

Low-Emittance PI Laser Requirements Cs: Ga. As WAVELENGTH Photon energy should exceed the photocathode work function by as little as possible. NLO allows energy multiplication (2 w, 3 w. . . ) and full tunability via OPA. Cs 2 Te Mg Cu 2 3 4 Photon energy (e. V) 5 OPA Ti: S Yb Nd 1 But it adds complexity, reduces efficiency and can compromise stability, beam quality and reliability. POLARISATION Doesn’t affect photoelectron production so is, in principle a free parameter. But in practice NLO is polarisation-sensitive and cathode absorption may be too.

Low-Emittance PI Laser Requirements TEMPORAL PROFILE Repetitive picosecond/femtosecond pulses are generated by phase-locking the discrete frequency-domain modes of an optical cavity. Fourier relates the pulse shape to the individual modes’ amplitudes and phases which are limited by the laser medium’s gain profile but which are also independently controllable. n Rapid changes in the pulse need broad spectral bandwidth from the laser. Low emittance electron bunches may need unusually short drive laser pulses. Mode control hardware can be complex and challenging but is effective for pulse shaping, even if NLO is involved. On a picosecond timescale pulse shaping by division, delay and stacking is also effective. Data from M. Danailov, 2007

Low-Emittance PI Laser Requirements TEMPORAL PROFILE Pulse shaping systems are now becoming commercially available. Dazzler AO phase and amplitude modulator from Fastlite Coherent’s ‘Silhouette’ provides feedback control of spectral amplitude and phase.

Low-Emittance PI Laser Requirements TEMPORAL PROFILE But in the end there is no point in temporally shaping the laser pulse on timescales much faster than the response time of the photocathode.

Low-Emittance PI Laser Requirements TRANSVERSE PROFILE Laser oscillators tend to deliver Gaussian profile beams Amplifier saturation tends to ‘square off’ beams in the near field provided: • Gain media are uniform and pumping is stable and well-profiled • Transport optics and media are good-quality and clean • Diffraction is managed • NLO self-focusing is managed Data from M. Danailov, 2007 NLO frequency conversion efficiency is sensitive to beam direction which may be rapidly varying near a high-intensity focus and tends to reverse the squaring Optical squaring (refractive shaping or simple clipping) must balance inefficiency, limited depth of field, chromaticity and sensitivity to input beam variations

Low-Emittance PI Laser Requirements TRANSVERSE PROFILE Laser oscillators tend to deliver Gaussian profile beams. Amplifier saturation tends to ‘square off’ beams in the near field provided: • Gain media are uniform and pumping is stable and well-profiled • Transport optics and media are good-quality and clean Data from C. S. Chou • Diffraction is managed et al, 2009 • NLO self-focusing is managed NLO frequency conversion efficiency is sensitive to beam direction which may be rapidly varying near a high-intensity focus and tends to reverse the squaring Optical squaring (refractive shaping or simple clipping) must balance inefficiency, limited depth of field, chromaticity and sensitivity to input beam variations Transporting sharp-edged beams requires large numerical aperture and benefits from e. g. adaptive optics and image-relaying

Low-Emittance PI Laser Requirements TRANSVERSE PROFILE Laser oscillators tend to deliver Gaussian profile beams. Amplifier saturation tends to ‘square off’ beams in the near field provided: • Gain media are uniform and pumping is stable and well-profiled Data from D. H. Dowell et al, FEL 09, Paper WEOA 03 2009 • Transport optics and media are good-quality and clean • Diffraction is managed • NLO self-focusing is managed NLO frequency conversion efficiency is sensitive to beam direction which may be rapidly varying near a high-intensity focus and tends to reverse the squaring Optical squaring (refractive shaping or simple clipping) must balance inefficiency, limited depth of field, chromaticity and sensitivity to input beam variations Transporting sharp-edged beams requires large numerical aperture and benefits from e. g. adaptive optics and image-relaying

Low-Emittance PI Laser Requirements TRANSVERSE PROFILE Laser oscillators tend to deliver Gaussian profile beams. Amplifier saturation tends to ‘square off’ beams in the near field provided: • Gain media are uniform and pumping is stable and well-profiled Data from D. H. Dowell et al, FEL 09, Paper WEOA 03 2009 • Transport optics and media are good-quality and clean • Diffraction is managed • NLO self-focusing is managed NLO frequency conversion efficiency is sensitive to beam direction which may be rapidly varying near a high-intensity focus and tends to reverse the squaring Optical squaring (refractive shaping or simple clipping) must balance inefficiency, limited depth of field, chromaticity and sensitivity to input beam variations Transporting sharp-edged beams requires large numerical aperture and benefits from e. g. adaptive optics and image-relaying

Low-Emittance PI Laser Requirements TRANSVERSE PROFILE Laser oscillators tend to deliver Gaussian profile beams. • Gain media are uniform and pumping is stable and well-profiled Data from D. H. Dowell et al, FEL 09, Paper WEOA 03 2009 • Transport optics and media are good-quality and clean Projected emittance (microns rms) Amplifier saturation tends to ‘square off’ beams in the near field provided: • Diffraction is managed • NLO self-focusing is managed NLO frequency conversion efficiency is sensitive to beam direction which may be rapidly varying near a high-intensity focus and tends to reverse the squaring Optical squaring (refractive shaping or simple clipping) must balance inefficiency, limited depth of field, chromaticity and sensitivity to input beam variations Transporting sharp-edged beams requires large numerical aperture and benefits from e. g. adaptive optics and image-relaying

Low-Emittance PI Laser Requirements TRANSVERSE PROFILE Laser oscillators tend to deliver Gaussian profile beams. • Gain media are uniform and pumping is stable and well-profiled Data from D. H. Dowell et al, FEL 09, Paper WEOA 03 2009 • Transport optics and media are good-quality and clean Projected emittance (microns rms) Amplifier saturation tends to ‘square off’ beams in the near field provided: • Diffraction is managed • NLO self-focusing is managed If required laser designers can generate transverse profiles which are better controlled than the ‘standard’ commercial product But in the end there is no point in spatially shaping the laser pulse to make it much more uniform than the QE profile of the photocathode.

Requirements for Practical PI Lasers RELIABILITY AND UPTIME Favours design simplicity, mature technologies, commercial laser systems, over-specification, low photon energy, high thermal efficiency AVERAGE POWER Proportional to average beam current and to photocathode QE, affects cost and reliability, removing heat from the cathode may be an issue 1 m. A with 1% QE requires 6× 1017 ph/s which is 0. 25 W (green) or 0. 5 W (UV) ~10 W (IR) short pulse lasers are commercially available Militates against the use of low-QE cathodes STABILITY AND CONTROL For low emittance the laser must stay within a very small parameter space, requiring high intrinsic stability plus a multi-parameter feedback control system (timing jitter, temporal pulse shaping, adaptive beam shaping and pointing, environmental control e. g. temp, vibration, utilities (power, cooling, gas purge)) applied to the whole optical transport system, not just the laser. Individual FCS’s are commercially available but integrated suites are not.

Laser System Options Nd: crystal (YAG, YLF, YVO 4. . . ) Pros: High power, mature, commercially available, diode or flash pumped, compatible with fibre systems Cons: Slow temporal response (<1 ps), thermal beam quality issues, low hn FLASH amplifier chain Nd: YLF photoinjector lasers are in use e. g. at FLASH, PITZ and CERN CTF 3 and Nd: YVO 4 at ALICE PITZ

Laser System Options Ti: S Pros: Fastest temporal response of conventional lasers (~10 fs), mature, commercially available, some tunability, higher hn Cons: Complex, thermally inefficient, laser pumped, noisy (broad bandwidth, sensitive modelocking, short tupper), needs CPA SPARC Ti: S photoinjector lasers are in use e. g. at SPARC, LCLS and [email protected] FERMI LCLS

Laser System Options Yb: glass, crystal (YAG, SFAP. . . ) or ceramic Pros: High power, diode pumped, can be largely fibre-based, efficient, quite fast temporal response (~100 fs) Cons: Less mature but with some commercial availability, low hn, may need cryo cooling, may need CPA Clark-MXR Impulse Yb: fibre photoinjector laser is in use at Cornell ERL test facility and Yb: YAG is in use with SC Pb cathode at HZB (T Kamps) Commercial Yb: fibre laser 20 Wave, 100 m. J/pulse, 250 fs FWHM Recently deployed at HZB BESSY (but not for photoinjection)

Laser System Options OP(CP)A driven by Nd or Yb or Ti: S Pros: Tunable hn (real-time tuning not generally required) Cons: Inefficient, complex, can be noisy, can be prone to optical damage, temporal control is less mature, low QE may demand very high power STFC CLF ULTRA OPA systems are in wide use for spectroscopy and selective processing and are being sold into industry and medicine RIKEN

Conclusions • Photoinjector lasers have been developed over many years and are now driving guns with state-of-the-art electron bunch brightness • Further improvements to increase the brightness are likely to involve control of at least three of the lasers’ fundamental parameters: Wavelength (fine-tuned close to work function) Temporal profile (shortened and/or shaped) Spatial profile (smoothed and shaped to reduce electrons’ transverse momentum) • There is ‘headroom’ left to do this • Keeping the laser’s performance inside the necessarily small parameter space will require both high intrinsic stability and tight feedback control • As well as delivering the technical advances photoinjector laser scientists need to satisfy demanding operational needs. The inevitable conflict complicates the choice between commercial and non-commercial vendors