Low intrinsic emittance in modern photoinjector brightness

Reducing the intrinsic emittance of photocathodes is one of the most promising routes to improving the brightness of electron sources. However, when emittance growth occurs during beam transport (for example, due to space charge), it is possible that this emittance growth overwhelms the contribution of the photocathode, and, thus, in this case source emittance improvements are not beneficial. Using multiobjective genetic optimization, we investigate the role intrinsic emittance plays in determining the final emittance of several space-charge-dominated photoinjectors, including those for high-repetition-rate free electron lasers and ultrafast electron diffraction. We introduce a new metric to predict the scale of photocathode emittance improvements that remain beneficial and explain how additional tuning is required to take full advantage of new photocathode technologies. Additionally, we determine the scale of emittance growth due to point-to-point Coulomb interactions with a fast tree-based space-charge solver. Our results show that, in the realistic high-brightness photoinjector applications under study, the reduction of thermal emittance to values as low as $50\mathrm{pm}/\mu \mathrm{m}$ (1 meV mean transverse energy) remains a viable option for the improvement of beam brightness.

Figure 1

An example of the initial longitudinal and transverse spatial distributions of the beam for each system. These examples were selected from the 0 meV Pareto front and are the same individuals plotted in Fig. 5.

Figure 2

The on-axis electric and magnetic field as seen by a reference particle in the center of the electron bunch. In each subfigure, the cavity and magnet parameters are taken from an individual along the 0 meV Pareto front of the respective beam line. Fields are output directly from general particle tracer. In the case of the FEL, fields are computed from the energy change ($\mathrm{d}E/\mathrm{d}s$) and Larmor angle output from astra as a function of the position.

Figure 3

The Pareto fronts of each beam line for the $\sim 150$ and 0 meV MTE photocathodes and their characteristic MTE. The UED examples show between a factor of 10 and 100 improvement in brightness between the two Pareto fronts. The characteristic MTE calculated from a simulation including the effects of Coulomb scattering is included for the dc and NCRF gun UED examples as a yellow cross.

Figure 4

The probability distribution of initial spot sizes among the optimized individuals. The three example beam lines are labeled by color, and individuals from the $\sim 150\mathrm{meV}$ fronts are in dashed lines, while the individuals from the 0 meV fronts are represented by solid lines.

Figure 5

Emittance and beam sizes for an individual along the 0 meV Pareto front of each example. The UED individuals were selected at 16 fC bunch charge. The projected emittance is the typical rms normalized transverse emittance, and the slice emittance is the average of the emittance evaluated over 100 longitudinal slices. The beam width and length are also plotted for reference. The total projected emittance in (a) is clipped at 500 pm for clarity.

Figure 6

Individuals from the 0 meV beam line were resimulated with a photocathode MTE equal to their characteristic MTE. The probability distribution of the ratio of the new final emittance to the original final emittance is plotted.

Figure 7

An illustration of how reoptimization may be required upon insertion of a new photocathode. In black is the emittance due to transport (${ϵ}_{\mathrm{T}}$) as a function of the initial spot size. Around the optimal spot size ${\sigma }_{x,i,0}$, this is approximately quadratic. The sensitivity in this example is roughly $x\approx 0.001$. The solid lines represent the initial emittance (${ϵ}_{\mathrm{C}}$) for three different thermal emittances. The dashed lines are the final emittance (${ϵ}_{\mathrm{F}}$) or the quadrature sums of initial and transport emittance. The optimal spot size with the 150 meV photocathode is significantly smaller than with a 0 or even 1 meV photocathode.

Figure 8

The rms and core emittance of an individual with ${10}^5$ electrons per bunch from the dc gun UED and NCRF gun UED 0 meV MTE Pareto fronts. In the row labeled “beam dynamics”, the yellow lines were computed with the point-to-point space-charge algorithm and the blue lines with smooth space charge. The solid lines are the rms normalized emittance, and the dashed lines are the core emittance. Below are plots of the beam’s transverse phase space at the sample location computed with the smooth and point-to-point methods. Linear $x-p_x$ correlation has been removed, and the ellipse of phase space second moments is plotted in addition to the particle density.