In order to build the Compact LInear Collider (CLIC), accelerating structures reaching extremely high accelerating gradients are needed. Such structures have been built and tested using normal-conducting copper, powered by X-band RF power and reaching gradients of 100 MV/m and above. One phenomenon that must be avoided in order to reliably reach such gradients, is vacuum arcs or “breakdowns”. This can be accomplished by carefully designing the structure geometry such that high surface fields and large local power flows are avoided. The research presented in this thesis presents a method for optimizing the geometry of accelerating structures so that these breakdowns are made less likely, allowing the structure to operate reliably at high gradients. This was done primarilly based on a phenomenological scaling model, which predicted the maximum gradient as a function of the break down rate, pulse length, and field distribution in the structure. The model is written in such a way that it allows direct comparison of different criteria, i.e. the peak electric field, the peak local power flow or “modified Poynting vector” Sc, and the global power flow per iris circumference. Using this method, a set of highly optimized accelerating cells were created, as was a C++ library capable of estimating the performance of an RF structure based on these accelerating cells. This library was used in the rebaselinging of the CLIC machine. In addition to this, the thesis also presents a particle in cell simulation of the initial stages of a vacuum arc. This model follows the development of the arc from a small field emitter into a multi-ampere discharge, tracking the evolution of the plasma and the circuit, as well as the interaction between the plasma and the surface. From this model, our understanding of the importance of the ion bombardment under the plasma sheath was improved, especially the ion energy distribution and its effect on sputtering vs. the effective sputtering yields required for the arc to sustain itself and grow.

Avoiding vacuum arcs in high gradient normal conducting RF structures

Analytic RF design of a linear accelerator with a SLED-I type RF pulse compressor

Different types of electron bunch compression schemes are compared for the AXXS FEL design study. The main linac for the proposed machine is based on CLIC x-band structures. This choice leaves several options for the bunch compression schemes which impact the injection system RF band. Both harmonic linearization and phase modulation linearization are considered and their relative strengths and weaknesses compared. Simulations were performed to compare the performance of an s-band injector with a higher harmonic RF linearization to an entirely x-band injector and linearization scheme. One motivation for the study is to optimize the length of the AXXS machine, allowing the linac to fit onto the proposed site and also act as the injector to the existing storage ring at the Australian Synchrotron AXXS AXXS (Australian X-band X-ray Source) [1] is strategically planned to incorporate an upgrade of the current Australian Synchrotron storage ring and provide a new high brightness sub-Angstrom wavelength XFEL. The design includes a 6 GeV linac based around the CLIC x-band linac [2, 3]. PHASE MODULATION LINEARIZATION In attempting to linearize the longitudinal phase space most FELs around the world employ harmonic linearization. When looking to design an all x-band linac, as is the case of AXXS, the high frequency accelerating structures preceding the first bunch compressor make harmonic linearization more difficult to achieve. Phase modulation linearization is the term introduced in this paper to describe a new method of linearization. Through independently varying the phases of two cavities positioned before the first bunch compressor, 2nd order effects encountered through bunch compression can be minimized. These 2nd order effects refer to both the RF curvature impressed onto the beam during accelerating, plus the second order longitudinal dispersion, T566, of the magnetic chicane. Implementation of Phase Modulation Linearization is sketched in Fig. 1. To establish an energy chirp, the RF phase of the first section after the injector, φ0, is chosen to be 70◦. The two additional cavities required for Phase Modulation Linearization are shown as having the parameters V1, φ1, V2 and φ2. 70.0° -63.94° 48.61° 12 GHz 12 GHz 12 GHz W_1 W_2 W_3 R56 , T566 BC1 V0 , φ0 V1 , φ1 V2 , φ2 Figure 1: Layout of an x-band FEL accelerator design, up to end of first bunch compressor, which employs Phase Modulation Linearization. Included are markers showing location of longitudinal phase space plots in Fig. 3. Analytical Approach The beam energy after passing through the accelerating section leading up to the first bunch compressor is, E f = Ei + V0 cos (φ0 + kx z0) + V1 cos (φ1 + kx z0) + ... +V2 cos (φ2 + kx z0), (1) where V0, V1, V2, represent the RF voltages of the three sections shown in Fig. 1, and φ0, φ1, and φ2, are the RF phases of these sections, where the RF phase is defined to be 90◦ at the crest of the RF acceleration, and over the interval 0 < φ < 90◦ defines the negative slope of the RF curve (i.e. the head of the bunch is accelerated less than the tail). z0 is the position of an electron with respect to the bunch center, and kx is the wave number of the RF frequency. The longitudinal position of any electron as it passes through the dispersive region of the chicane is, z f (δ) = zi + R56δ + T566δ +U5666δ + ... (2) Taking the Taylor series expansion of Eq. 1, calculating the relative energy deviation and then substituting the result into to the dispersion relation (Eq. 2), the final longitudinal position is, z f = z0 + R56(az0 + bz2 0) + T566(az0 + bz 2 0) 2 + higher order terms (3) = z0(1 + aR56) + (bR56 + aT566)z 0 +2abT566z 0 + b T566z 0 + higher order terms (4) Proceedings of IPAC2015, Richmond, VA, USA TUPWA005 2: Photon Sources and Electron Accelerators A06 Free Electron Lasers ISBN 978-3-95450-168-7 1399 Co py rig ht © 20 15 CC -B Y3. 0 an d by th er es pe ct iv ea ut ho rs

Comparison of bunch compression schemes for the AXXS FEL

Current horns are ubiquitous in Free Electron Laser (FEL) bunch compression. In this paper, we analyse the formation of these current spikes and identify the cause as caustic formation in the electron trajectories. We also present a possible solution to avoid or mitigate the current horns from developing through inclusion of sextupole and octupole magnets in the compressor design.

Electron Trajectory Caustic Formation Resulting in Current Horns present in Bunch Compression

Electron trajectory caustic formation resulting in current horns present in bunch compression

A linear accelerating structure with an average loaded gradient of 100 MV/m at X-Band frequencies has been demonstrated in the CLIC study. Recently, it has been proposed to use this structure to drive an FEL linac. In contrast to CLIC the linac would be powered by klystrons not by an RF source created by a drive beam. The main advantage of this proposal is achieving the required energies in a very short distance, thus the facility would be rather compact. In this study, we present the structure choice and conceptual design parameters of a facility which could generate laser photon pulses below Angstrom. Shorter wavelengths can also be reached with slightly increasing the energy.

A. Charitonidis

Papers

FEL proposal based on CLIC X-band structure