Monte Carlo Modeling of Crystal Channeling at High Energies

Charged particles entering a crystal close to some preferred direction can be trapped in the electromagnetic potential well existing between consecutive planes or strings of atoms. This channeling effect can be used to extract beam particles if the crystal is bent beforehand. Crystal channeling is becoming a reliable and efficient technique for collimating beams and removing halo particles. At CERN, the installation of silicon crystals in the LHC is under scrutiny by the UA9 collaboration with the goal of investigating if they are a viable option for the collimation system upgrade. This thesis describes a new Monte Carlo model of planar channeling which has been developed from scratch in order to be implemented in the FLUKA code simulating particle transport and interactions. Crystal channels are described through the concept of continuous potential taking into account thermal motion of the lattice atoms and using Moliere screening function. The energy of the particle transverse motion determines whether or not it is trapped between the crystal planes while single Coulomb scattering on lattice atoms can lead to dechanneling. The volume capture and reflection applying to quasi-channeled particles are also modeled. Analogously to dechanneling, single scattering is used to determine the occurrence of volume capture. The parameters of the crystals, such as torsion or miscut, are described as well. For channeled particles, the suppression of electromagnetic and nuclear collisions is implemented. It is stronger for particles oscillating close to the center of the channel and is crucial for a correct evaluation of the rate of dechanneling without having recourse to the use of a macroscopic dechanneling length. The UA9-H8 experiment conducted at CERN aims at investigating new crystal physics as well as characterizing crystals that can be of interest in view of the implementation of crystal collimation at CERN, including in the LHC. This experiment uses silicon strip detectors situated on both sides of the crystal. Putting together upstream and downstream tracks in coincidence and matching at an identical fitted location on the crystal, it yields information about the deflections given to the beam population. Several runs from the UA9-H8 experiment are analyzed and compared to the model results. Channeling and dechanneling rates, as well as angular distributions at crystal exit are shown to be in a very encouraging agreement both for strip and quasi-mosaic crystals.