Multiobjective genetic algorithm optimizations of a single-shot ultrafast electron diffraction beam line comprised of a $100\text{ }\text{ }\mathrm{MV}/\mathrm{m}$ 1.6-cell normal conducting rf (NCRF) gun, as well as a nine-cell $2\ensuremath{\pi}/3$ bunching cavity placed between two solenoids, have been performed. These include optimization of the normalized transverse emittance as a function of bunch charge, as well as optimization of the transverse coherence length as a function of the rms bunch length of the beam at the sample location for a fixed charge of $1{0}^{6}$ electrons. Analysis of the resulting solutions is discussed in terms of the relevant scaling laws, and a detailed description of one of the resulting solutions from the coherence length optimizations is given. For a charge of $1{0}^{6}$ electrons and final beam sizes of ${\ensuremath{\sigma}}_{x}\ensuremath{\ge}25\text{ }\text{ }\ensuremath{\mu}\mathrm{m}$ and ${\ensuremath{\sigma}}_{t}\ensuremath{\approx}5\text{ }\text{ }\mathrm{fs}$, we found a relative coherence length of ${L}_{c,x}/{\ensuremath{\sigma}}_{x}\ensuremath{\approx}0.07$ using direct optimization of the coherence length. Additionally, based on optimizations of the emittance as a function of final bunch length, we estimate the relative coherence length for bunch lengths of 30 and 100 fs to be roughly 0.1 and $0.2\text{ }\text{ }\mathrm{nm}/\ensuremath{\mu}\mathrm{m}$, respectively. Finally, using the scaling of the optimal emittance with bunch charge, for a charge of $1{0}^{5}$ electrons, we estimate relative coherence lengths of 0.3, 0.5, and $0.92\text{ }\text{ }\mathrm{nm}/\ensuremath{\mu}\mathrm{m}$ for final bunch lengths of 5, 30 and 100 fs, respectively.
Read full abstract