Electron cloud studies for CERN particle accelerators and simulation code development

In a particle accelerator free electrons in the beam chambers can be generated by different mechanisms like the ionization of the residual gas or the photoemission from the chamber’s wall due to the synchrotron radiation emitted by the beam. The electromagnetic field of the beam can accelerate these electrons and project them onto the chamber’s wall. According to their impact energy and to the Secondary Electron Yield (SEY) of the surface, secondary electrons can be generated. Especially when the accelerator is operated with closely spaced bunches of positively charged particles, this mechanism can drive an avalanche multiplication process of the electrons with the formation of a so called Electron Cloud (EC) in the chamber. The presence of a large electron density in the beam pipe as well as of a strong electron flux on the chamber’s wall can limit the achievable performance of the accelerator through different effects like transverse instabilities, transverse emittance growth, particle losses, vacuum degradation and heating of the chamber’s surface. EC effects have been recognized among the major performance limitations for the Large Hadron Collider (LHC), presently the world’s largest and most powerful particle accelerator and collider, built by the European Organization for Nuclear Research (CERN) in a 27 km underground tunnel across the Franco-Swiss border near Geneva, Switzerland. EC effects were observed at the LHC during the first three years of beam operation (Run 1, 2010 – 2012), becoming more and more severe while moving to tighter bunch spacing. EC effects with 50 ns bunch spacing could be successfully mitigated through beam induced scrubbing (reduction of the SEY by means of electron bombardment) and this bunch spacing could be used for most of the integrated luminosity production with 7 – 8 TeV Center of Mass (CoM) energy in 2011–12. After the 2013–14 machine shutdown (LS1) the LHC will be able to run at 13–14 TeV CoM energy and it will be necessary to move to the design bunch spacing of 25 ns in order to reach the design luminosity within the pileup limits required by the LHC experiments. Up to now, the 25 ns beam has been used only for test purposes and EC effects proved to be significantly more severe compared to the 50 ns case. The present thesis work addresses EC effects in the LHC and its injector accelerators chain in terms of both numerical simulations and machine experiments. Particular emphasis is put on beams with 25 ns bunch spacing. In particular, the analysis of EC observations in the LHC and its injectors have raised new challenges for the EC build-up simulations. For a correct understanding of machine observations it is often necessary to deal with beams with thousands of bunches and with non-idealities like non-uniform bunch populations and bunch lengths along the beam. Beside the usual simulation scenarios of field free regions and dipole magnets, also more complex situations needed to be addressed, like the EC buildup in quadrupoles or combined function magnets and with two counter-rotating beams in the same chamber. Moreover, the demand for extensive parameter scans gave quite stringent requirements in terms of speed and reliability. CERN’s long experience in the EC build-up simulation, mostly carried out with the ECLOUD code, developed and maintained at CERN since 1997, proved instrumental to respond to the newly arisen needs. However, due to its non-modular structure and to the programming language (FORTRAN 77), the existing ECLOUD code did not appear to be suitable to be extended to fulfill the aforementioned requirements. It was therefore decided to follow a different strategy and write a fully reorganized code, in a more modern and flexible language, considering that the initial effort would be compensated by a significantly increased efficiency in development and debugging. The new code has been called PyECLOUD, since it is almost entirely written in Python and inherits the physical models of the ECLOUD code. During the development we addressed all the known issues of the ECLOUD code were addressed and new features were introduced, necessary to deal with the complex scenarios described above. Modifications on numerical model and implementation were introduced practically everywhere, and key modules of the code, i.e. the MacroParticle (MP) Size Management, the electron space charge evaluation, the MP tracker and the electron/wall interaction have been completely redesigned. The new code has been applied for a full characterization of the EC formation in the main LHC components (including those with the two counter-rotating beams in the same chamber) with respect to different to surface properties (SEY) and beam configurations. In parallel with this modeling and simulation work, an intense experimental activity was carried out, which involved the LHC and the last two synchrotrons of its injector chain, i.e. the Proton Synchrotron (PS) and the Super Proton Synchrotron (SPS), and had three main goals: 1) The qualification of the EC formation in the three accelerators and of its impact on the quality of the proton beam; 2) The collection of experimental data for the validation and the improvement of our simulation models; 3) The definition and experimental validation of possible EC mitigation strategies. As already stated before, at the LHC, EC effects represent the main limitation to the use of the nominal bunch spacing of 25 ns. Experiments with this type of beam took place for the first time in 2011 and more extensively towards the end of the 2012 run. The main goals were to investigate these limitations and to study the process of beam scrubbing as a possible mitigation for future operation. These tests included a 3.5 day scrubbing run at 450 GeV, few test ramps with 25 ns beams in order to study EC effect at high energy (4 TeV), and a pilot physics run with low emittance 25 ns beams. During this period it was possible to collect measurements on several EC observables, e.g. transverse positions for the first seconds after injection, heat load on the cryogenic sections, bunch by bunch intensity, transverse emittance and stable phase. The careful analysis of these data significantly improved our understanding on the EC buildup in the LHC and on its impact on machine performance and beam quality, both at injection and collision energy. Concerning the LHC injectors, both at the PS and at the SPS, several “direct” e-cloud measurements could be collected under different beam conditions (bunch intensity, length, number and spacing) using dedicated devices installed in the rings. Moreover we could observe and qualify the impact of the EC on the vacuum pressure and on the quality of the beams in terms of transverse instabilities, beam losses, emittance growth. For the SPS, the possibility of preparing a dedicated beam for the EC mitigation through beam induced scrubbing has been studied. PyECLOUD simulations have been performed to compare different options and the most promising, the so-called “doublet” beam has been experimentally validated in the accelerator. The thesis is organized in three parts. Part I introduces the main concepts and mechanisms involved in the EC formation and describes in detail our simulation model and its implementation in the PyECLOUD code. Part II addresses the EC effects in the LHC, covering both simulation and experimental studies, with a dedicated chapter focusing on EC effects in the common regions where the two beams share the same chamber. Finally Part III describes simulation and experimental studies on EC effects in the LHC injectors