A Search for nu_mu to nu_e oscillations in the NOMAD experiment

The NOMAD experiment is a neutrino oscillation experiment. capable of identifying Vμ, -Vμ, Ve, -Ve and Vt for use in oscillation analyses. A search for Vμ --> Ve oscillations is conducted, emphasising the development of two separate beam simulators, to provide the background, (no oscillation), lie signal. Both beam descriptions include fits to the results of the SPY experiment that measured hadron production from a 450 GeV proton beam on beryllium target. An independent analysis of the raw SPY data to produce the particle yield is reported. A series of criteria are described for the selection and classification of neutrino events. These produce the data samples necessary for both tuning the beam simulation and determining the oscillation signal. The development of a GEANT and FLUKA based Monte Carlo beam simulator is presented; providing good agreement to the measured neutrino beam. This simulation method has sizeable variations depending on the beamline geometry, which is not knonm precisely. This causes large systematic errors. An empirical parametrisation is proposed and used for the first time in a NOMAD oscillation search. It uses the measured neutrino spectra at NOMAD, except the lie, to infer the meson production at the target, and then predict the Ve spectrum. This method has good agreement with the data and is also insensitive to alterations of the beamline geometry, resulting in much smaller systematic errors. The reduction of the systematic errors allows the Vμ --> Ve lie oscillations search to be performed with much greater precision. Comparisons of the Ve/Vμ ratio between the empirical parametrisation and data yields no evidence for Vμ --> Ve oscillations, setting a limit on the mixing parameter, sin2(W) < 1.9 x 10-:~(90% CL) at high Δm2. The present sensitivity of the analysis on the mixing parameter is 0.0017