John Kelley, UW-Madison, May 2008
After applying the Nch-dependent cuts described in the previous section, we revisit the observable distributions, Nch and cos(Zenith). We note that these plots are binned more finely than in the likelihood analysis, and we have not normalized MC to data.
Figure 9.1: Data vs. atmospheric MC, cos(Zenith) for (20 <= Nch <= 120) (2000-2006 Zeuthen+2D purity cuts). MC is generated with Photonics AHA and 85% OM sensitivity. |
Figure 9.2: Data vs. atmospheric MC, Number of channels hit (2000-2006 Zeuthen+2D purity cuts). MC is generated with Photonics AHA and 85% OM sensitivity. |
The new limits on VLI are presented in figure 9.3. The upper limit at maximal mixing (2.8e-27) is slightly worse, as expected, than the limit calculated after the first unblinding, and it is no longer better than the SuperK+K2K limit. However, it is still a 40% improvement on the previous AMANDA-II limit.
Figure 9.3: Allowed region of VLI parameter space (90%, 95%, and 99% contours), preliminary limits from 2000-2006 data. The purple solid curve is the previous AMANDA-II limit (2000-03, J. Ahrens), and the yellow dotted curve the IceCube sensitivity (90% CL, 10 years of IC80).
Figure 9.4: Allowed region of QD parameter space (90%, 95%, and 99% contours), preliminary limits from 2000-2006 data.
We have also calculated the upper limits for similar VLI and QD models using different energy dependences. The expectation for a signal does not appreciably change -- only the steepness of the new physics transition. We present VLI and QD limits in the table below for effects proportional to E, E^2, and E^3. VLI upper limits are for maximal mixing; QD upper limits are for all decoherence parameters equal. Numbers in bold are the models shown above in the 2D contour plots.
| n | VLI limit | QD limit | Units |
| 1 | 2.8e-27 | 1.2e-27 | -- |
| 2 | 2.7e-31 | 1.3e-31 | 1/GeV |
| 3 | 1.9e-35 | 6.3e-36 | 1/GeV^2 |
The two-parameter unfolding (or "forward folding") results for the conventional atmospheric flux are presented below. The contours suggest an atmospheric flux that has a slightly higher normalization and spectral slope than the Bartol model (although we do note the latter is extrapolated via NeutrinoFlux past 700 GeV). We are currently checking the results based on the Honda 2006 model to make sure they are comparable.
Figure 9.5: Allowed region of conventional parameter space (90%, 95%, and 99% contour levels), that is, normalization and slope change relative to Bartol.
The 90% contour can be represented as an allowed band of possible fluxes -- the envelope of curves generated by the different allowed normalizations and spectral slopes. The energy region for the band is the intersection of the 5-95% regions for each particular possibility.
Figure 9.6: Angle-averaged nu_mu + nu_mu_bar atmospheric flux (90% CL, dotted line is best-fit curve). SuperK results (as determined by González-García et al.) are also shown for comparison. Fluxes are multiplied by E^3 to enhance the features.
Finally, we show a comparison of this result with the 2000-03 Münich et al. unfolding result. The energy ranges are complementary, and the agreement in the overlap region is rather good. The data points describing the 90% allowed forward-folding flux can be found here.
Figure 9.7: Allowed nu_mu + nu_mu_bar atmospheric flux from this analysis with 2000-03 unfolding result, SuperK results, and Fréjus results. Fluxes are multipled by E^3.