Temporal shaping of narrow-band picosecond pulses via noncolinear sum-frequency mixing of dispersion-controlled pulses

A long-sought-after goal for photocathode electron sources has been to improve performance by temporally shaping the incident excitation laser pulse. The narrow bandwidth, short wavelength, and picosecond pulse duration make it challenging to employ conventional spectral pulse shaping techniques. We present a novel and efficient intensity-envelope shaping technique achieved during nonlinear upconversion through opposite-chirp sum-frequency mixing. We also present a numerical case study of the linac coherent light source-II photoinjector where transverse electron emittance is improved by at least 25%.

Figure 1

DCNS method in the spectral domain demonstrating two pulses of equal on opposite chirp mixing in a nonlinear medium to generate chirped second harmonic (SHG) copies of each pulse along with the narrow band, flat phase, sum frequency (SFG) pulse.

Figure 2

Pulse envelopes resulting from different combinations of SOD and TOD. Each column has a the same magnitude of the TOD value at the top and each row has the same magnitude of SOD. The blue plot results from combinations with negative SOD and positive TOD while the red plot is positive SOD and negative TOD.

Figure 3

Genetic algorithm optimized 515 nm laser profile for a bunch length of 1.22 mm resulting in an emittance value of $0.30\mu \mathrm{m}$. Notably, the genetic algorithm settled on a profile which has slower rise and fall times than an ideal flat top as emittance is optimized rather than profile shape. The two plots shown are (a) the laser profile in time before (grey) and after (blue, dashed) a 0.5 nm spectral filter and (b) the spectrum of the pulse before (grey) and after filtering (blue, dashed) with the super-gaussian spectral filter in black.

Figure 4

Pareto front comparison of DCNS and Gaussian performance for the LCLS-II injector. DCNS pulses in combination with a 1.0 nm spectral filter, achieves the lowest emittance values at most bunch lengths. The lowest achieved emittance value is $0.30\mu \mathrm{m}$ at a bunch length of 1.22 mm, using a 0.5 nm filter.

Figure 5

Histogram showing simulation density of all GA solutions, not only Pareto optimal, for Gaussian and DCNS cases. Lighter colored areas indicate a higher number of simulations with valid solutions and thus regions where different methods might be more effective.