<p>A search for a diffuse flux of cosmic neutrinos with energies in excess of 10¹⁴ eV was performed using two years of AMANDA-II data, collected in 2003 and 2004. A 20% evenly distributed sub-sample of experimental data was used to verify the detector description and the analysis cuts. A very good agreement between this 20% sample and the background simulations was observed. The analysis was optimised for discovery, to a relatively low price in limit setting power. The background estimate for the livetime of the examined 80% sample is 0.035 ± 68% events with an additional 41% systematical uncertainty.</p><p>The total neutrino flux needed for a 5σ discovery to be made with 50% probability was estimated to 3.4 ∙ 10^-7 <i>E</i>^-2 GeV s^-1 sr^-1 cm^-2 equally distributed over the three flavours, taking statistical and systematic uncertainties in the background expectation and the signal efficiency into account. No experimental events survived the final discriminator cut. Hence, no ultra-high energy neutrino candidates were found in the examined sample. A 90% upper limit is placed on the total ultra-high energy neutrino flux at 2.8 ∙ 10^-7 <i>E</i>^-2 GeV s^-1 sr^-1cm^-2, taking both systematical and statistical uncertainties into account. The energy range in which 90% of the simulated <i>E</i>^-2signal is contained is 2.94 ∙ 10¹⁴ eV to 1.54 ∙ 10¹⁸eV (central interval), assuming an equal distribution over the neutrino flavours at the Earth. The final acceptance is distributed as 48% electron neutrinos, 27% muon neutrinos, and 25% tau neutrinos.</p><p>A set of models for the production of neutrinos in active galactic nuclei that predict spectra deviating from <i>E</i>^-2 was excluded.</p>

A search for a diffuse flux of cosmic neutrinos with energies in excess of 1014 eV was performed using two years of AMANDA-II data, collected in 2003 and 2004. A 20% evenly distributed sub-sample of experimental data was used to verify the detector description and the analysis cuts. A very good agreement between this 20% sample and the background simulations was observed. The analysis was optimised for discovery, to a relatively low price in limit setting power. The background estimate for the livetime of the examined 80% sample is 0.035 ± 68% events with an additional 41% systematical uncertainty. The total neutrino flux needed for a 5σ discovery to be made with 50% probability was estimated to 3.4 ∙ 10-7 E-2 GeV s-1 sr-1 cm-2 equally distributed over the three flavours, taking statistical and systematic uncertainties in the background expectation and the signal efficiency into account. No experimental events survived the final discriminator cut. Hence, no ultra-high energy neutrino candidates were found in the examined sample. A 90% upper limit is placed on the total ultra-high energy neutrino flux at 2.8 ∙ 10-7 E-2 GeV s-1 sr-1 cm-2, taking both systematical and statistical uncertainties into account. The energy range in which 90% of the simulated E-2 signal is contained is 2.94 ∙ 1014 eV to 1.54 ∙ 1018 eV (central interval), assuming an equal distribution over the neutrino flavours at the Earth. The final acceptance is distributed as 48% electron neutrinos, 27% muon neutrinos, and 25% tau neutrinos. A set of models for the production of neutrinos in active galactic nuclei that predict spectra deviating from E-2 was excluded.