Search for prompt neutrinos with AMANDA-II

The investigation performed in this work aims to identify and disentangle the signal of prompt neutrinos from the inclusive atmospheric spectrum. We have analysed data recorded in the years 2000-2003 by the AMANDA-II detector at the geographical South Pole. After a tight event selection, our sample is composed of about 4 . 103 atmospheric neutrinos. Prompt neutrinos are decay products of heavy quark hadrons, which are produced in the collision of a cosmic ray particle with a nucleon in the atmosphere. The technique used to recognise prompt neutrinos is based on a simulated information of their energy spectrum, which appears harder than that of the conventional component from light quarks. Models accounting for different hadron production and decay schemes have been included in a Monte Carlo simulation and convoluted with the detector response, in order to reproduce the different spectra. The background of conventional events has been described with the Bartol 2006 tables. The energy spectrum of our data has been reconstructed through a numerical unfolding algorithm. The reconstruction is based on a Monte Carlo simulation and uses as an input three parameters of the neutrino track which are correlated with the energy of the event. Numerical regularisation is introduced to achieve a result free of unphysical oscillations, typical unfortunate feature of unfolding. The reconstructed data spectrum has been compared with different predictions using the model rejection factor technique. The prompt neutrino models differ in the choice of the hadron interaction model, the set of parton distribution functions and the numerical parameterisation of the fragmentation functions describing the transition from quark to hadrons. Here we considered mainly three classes of models, known in the literature as the Recombination Quark Parton Model, the Quark Gluon String Model and the Perturbative QCD model. Upper limits have been set on the expected flux predictions, based on our observations. The quark gluon string model seems to be disfavoured at 90% confidence level. Theoretical uncertainties strongly affect the predictions, as the deep inelastic QCD scattering process contains quantities that cannot measured at high Q2 and small x. Systematics affecting the measurements are partly ascribed to the ice structure and partly to the detector efficiency. For the former, we use a reference description of the optical properties of the South Pole glacier. As for the latter, the acceptance of the photomultipliers has been estimated in this work with a geometrical method based on the probability of detecting photon at a given distance. Chances to improve upon the current limits are assigned to the future large neutrino telescopes, which will allow to increase the sensitivity to both prompt and extraterrestrial neutrinos. (orig.)