Evaluation of a compact electron preinjector using a low beta (β) acceptance X-band accelerating structure

At the University of Melbourne X-LAB we are investigating the use of a low β acceptance X-band accelerating structure as part of the design of an all X-band RF electron preinjector optimised for the production of low emittance electron bunches for medical physics applications and compact light source development. In this work we will elaborate on the estimated performance, design issues, and optimisation methodology of the preinjector beamline.


Introduction
With the commissioning of the University of Melbourne X-band Laboratory for Accelerators and Beams (X-LAB) now underway [5] simulations of the performance of a future beamline are required.This beamline will leverage high gradient X-band (11.9942GHz) Radio Frequency (RF) linear accelerators to form a compact preinjector section to maximise space available for the rest of the beamline.The final expected use of this electron preinjector is to produce low transverse emittance bunches for use in a potential Inverse Compton Scattering (ICS) X-ray light source, and also for dosimetry studies.
For optimal beam capture, acceleration, and current an RF photogun is planned for the final beamline design but is not available for immediate procurement.Previously a design using a relatively low-cost 100 keV DC photogun was investigated, but would require an additional acceleration and bunching section.The extra bunching section was required for capture into the X-band accelerating structures currently present at X-LAB which are matched to β ≈ 1 [6].There are some practical issues with this approach, namely mixing and synchronising S-band (2.9986 GHz) and X-band RF accelerating structures, and required extra infrastructure that may affect future upgrade pathways.
However based on the work of the CQT group at the Eindhoven University of Technology an alternative approach has become available; replacing one of the current X-band accelerating structures with a modified structure with low β acceptance.

Description of electron preinjector
The beamline preinjector to be simulated consists of, in order: (i) A 100 keV DC photogun with focusing solenoid.
(ii) A low β acceptance Travelling Wave (TW) X-band RF accelerating structure (to be referred to as the low β structure).(iii) An X-band TW RF accelerating structure similar to the CLIC T24 structure [2].
Simulations from the electron gun to a short distance beyond the second X-band accelerating structure are carried out in ASTRA [3].Preinjector parameters such as focusing solenoid strength, structure placement, and structure phases have been chosen so as to minimise the final transverse emittance and minimise total length of the preinjector.
A total of 40 MW of RF input power is expected to be available to be distributed between the two structures, with a maximum of 20 MW each.This corresponds to an average gradient in both structures of approximately 70 MV m −1 .As part of these simulations a reduced input power to the first structure of has been considered, 11.25 MW, which corresponds to an average gradient of approximately 52.5 MV m −1 .This was considered as it corresponds to reduced power availability during the early commissioning stages of the beamline.

Description of low β structure
The proposed structure is similar to the second accelerating structure of the preinjector, but the first few cells at the entrance are modified to accept 100 keV electrons.This structure has been designed and simulated by the CQT group at the Eindhoven University of Technology (working on the SmartLight project [4]) in collaboration with the CLIC project at CERN and is briefly discussed in [1].It is currently being investigated whether the irises of this accelerating structure could be manufactured locally at the Australian National Fabrication Facility (ANFF), with rest of structure assembly to be carried out using CERN facilities.
A key difference between the beamline preinjector simulations to be documented here and those of the CQT group at the Eindhoven University of Technology is that they have utilised an extra RF buncher before the structure, which is not expected to be a part of the X-LAB design at this stage.This was to minimise extra RF infrastructure required at X-LAB.
However, the lack of an extra bunching section constrains the design of the electron preinjector significantly.Namely, special care must be taken that space charge does not cause the bunch to longitudinally extend to the point where it splits across the RF period.
Two approaches to this are shown in the following sections.The first is a study of the maximum charge that can be transmitted assuming a small laser spot size (and initial particle distribution) of σ x = 0.25 mm.The second is where a initial particle distribution with larger transverse size is considered.This is because by increasing the transverse dimensions of the bunch space charge forces that drive the longitudinal growth of the bunch prior to the first structure are decreased.

Situation 1 -simulations with small initial size
This situation considers the maximum bunch charge that can be transmitted through the preinjector with particle distribution of small initial transverse size.The unstable 5 pC simulations have been included to show how 5 pC is close to the maximum charge that can be transmitted, as it barely makes it through the first structure but starts to blow up after the second.The range of appropriate accelerating structure phases for 5 pC is also small in comparison to the other simulations in this situation.
The different line colours and styles of Figs. 1, 2, 3, and 4 represent different simulation parameters with a legend in Fig. 1.These different curves show how various statistical measures of the bunch vary along the length of the preinjector.A conventional coordinate system has been used where Z represents the longitudinal distance along the preinjector and X one the transverse dimensions.In Figures 1, 2 and 3 configurations with the same bunch charge have similar shapes, with input power linked to smaller variation in comparison.Figures 1 and 3     Kinetic energy development, situation 1. Legend in Fig. 1 the transverse dynamics of the bunch as it is accelerated for the different configurations, with the 5 pC configurations showing large transverse size and emittance growth compared to the other configurations.Figure 2 demonstrates how the longitudinal extent of the bunch increases with increasing bunch charge.Figure 4 shows the mean particle kinetic energy growth of the bunch and indicates the position of the DC photogun and accelerating structures.
For user applications of the beamline a greater bunch charge and beam current would be desired.This provides motivation to reduce the effect of space charge forces on the longitudinal extent of the bunch, leading to the simulations of situation 2.  extreme scenario where the initial spot size is greater than the flat section of the DC photogun cathode, but are kept as an upper bound estimate of the bunch charge that could be transmitted and the expected emittance.As well, in the 25 pC configurations the the RMS transverse size of the bunch is starting to approach the size of the irises in the second accelerating structure.

Situation 2 -simulations with larger spot size
By increasing the initial distribution transverse size the effect of space charge on the longitudinal dimension is decreased and bunches with greater charge than those in situation 1 can be transmitted through the preinjector.However, as initial size increases so does the final transverse emittance.

Conclusion
Different configurations of the electron preinjector have been investigated, and the tradeoffs between beam size, current, and emittance have been observed.The use of the low β structure allows for the assembly of a compact electron preinjector suitable for the X-LAB.A mitigating strategy to allow for higher charge bunches despite the lack of an extra bunching section, at the tradeoff of a slightly higher final emittance has also been shown.From these simulations a configuration using the low β structure and a low energy DC photogun is an alternative to an RF photogun able to produce stable bunches with transverse emittance less than 1π mm mrad. show