The IceCube Neutrino Observatory IV: Searches for Dark Matter and Exotic Particles

The intriguing question of why less than one fifth of the matter in our universe seems to be of the ordinary sort that emits electromagnetic radiation and the rest being invisible by all other means than gravitational inference, has been an outstanding problem in physics for almost 80 years. The most popular current hypotheses involve Weakly Interacting Massive Particles (WIMPs) that are distributed in and around galaxies as a halo. In several theories these particles will accumulate in massive objects, like the Sun, by scattering (weakly) multiple times. With each scatter momentum is lost and the particle becomes gravitationally trapped in the centre regions of the object. Should the WIMPs then be Majorana particles they will pair-wise annihilate and, in some channels, give rise to a neutrino signal that is detectable on Earth. In this paper we present an explicit search for Solar WIMPs described by the Minimal Supersymmetric Standard Model (MSSM) [1], where the lightest supersymmetric particle χ0

1 , henceforth denoted χ, is a promising dark matter candidate. The two most extreme annihilation channels are studied; χ → b b producing the softest possible neutrino spectrum, and χ → W + W -producing the hardest possible spectrum. Note that for m χ < m W we consider instead the channel χ → τ + τ -. The same event selection is also applied to search for another potential dark matter candidate, the Lightest Kaluza-Klein Particle (LKP), arising from theories of Universal Extra Dimensions (UED) [2]. The 5-dimensional model we considered here is described by the mass m

γ of the LKP, the first photon excitation, and the mass splitting ∆ q (1) = (m q (1)m γ (1) )/m γ (1) , where m q (1) is the mass of the first quark excitation. The Monte Carlo defining the signal here, with ∆ q (1) = 0, was taken from [3]. The world’s largest neutrino observatory, IceCube [4], was halfway to completion in 2008 with 40 strings, each containing 60 digital optical modules, deployed in the lay-out shown in figure 1. AMANDA, the predecessor detector to IceCube, underwent a low-energy trigger upgrade and integration of its Data Acquisition System (DAQ) with that of IceCube. As seen in Figure 1, AMANDA, with a markedly lower energy threshold than IceCube, was horizontally enclosed by the 2008 IceCube array. Data from this year thus provided a unique opportunity to use the two detectors as one. In particular, IceCube could be used as an efficient veto for horizontal atmospheric muons entering AMANDA. This opportunity was not missed, and we present here the result of a search for Solar dark matter using this data, and follow up by combining the result with previous AMANDA and IceCube limits.

An analysis was undertaken on the 2008 IceCube-AMANDA dataset amounting to a total livetime of 149 days. The livetime considers the period when the sun was below the horizon (March to September). A triggered sample of 1.7×10 10 events exists in this dataset, the vast majority of which are atmospheric muons. The chosen event-sample was to be reduced by a factor of ∼10 -7 using variable cuts and a machine learning algorithm in an effort to maximise the separation of signal and background. When optimising the event selection, described further in section 3, data from November 2008 was used and then discarded.

The signal Monte Carlo was generated with DarkSUSY [5], simulating WIMPs and LKPs for a range of different particle masses. As mentioned in section 1, we generated two MC samples for each MSSM mass corresponding to soft and hard neutrino spectra. The atmospheric muon background was simulated with CORSIKA [6] using the Hörandel [7] model for the cosmic ray spectrum and composition. The background of atmospheric ν µ was generated with the Honda et al. [8] conventional flux model and the Enberg et al. [9] prompt flux model. After simulating charged particle propagation with MMC [10] and light propagation in the ice with PHOTON-ICS [11], the detector response was finally simulated using the IceCube simulation software IceSim.

This analysis divided the data into two data streams: One stream used IceCube as the main detector searching for high WIMP masses, using as a signal template the 1000 GeV WIMP model with the hard spectrum. The second stream used AMANDA as the main fiducial volume and IceCube as a sophisticated veto, with the 100 GeV softspectrum WIMP as a signal template. In each stream a series of linear cuts were applied to data and MC, optimised individually using the signal templates and data from November 2008 as the estimate for the background. In the AMANDA-stream no more than 4 hits were allowed in IceCube, effectively favouring low-energy events. In both streams we made log-likelihood reconstructions of the particle tracks, requiring them to be upwardgoing and good fits to the hit information. The progression of data reduction is visualised in figure 2. After applying the linear cuts a Support Vector Machine (SVM) [12] was trained to s

…(Full text truncated)…

This content is AI-processed based on ArXiv data.