Selection of the Sample for Data-Driven $Z \to \nu\bar{\nu}$ Background Estimation in SUSY-Searches

The topic of this study was to improve the selection of the sample for data-driven Z → ν ν background estimation, which is a major contribution in supersymmetric searches in ̄ a no-lepton search mode. The data is based on Z → + − samples using data created with ATLAS simulation software. This method works if two leptons are reconstructed, but using cuts that are typical for SUSY searches reconstruction efficiency for electrons and muons is rather low. For this reason it was tried to enhance the data sample. Therefore events were considered, where only one electron was reconstructed. In this case the invariant mass for the electron and each jet was computed to select the jet with the best match for the Z boson mass as not reconstructed electron. This way the sample can be extended but significantly looses purity because of also reconstructed background events. To improve this method other variables have to be considered which were not available for this study. Applying a similar method to muons using tracks instead of jets as probe one obtains again poor purity. Furthermore only a few additional events are added to the sample leading to neglectable increase in efficiency. To understand this the distribution of the pseudorapidity |η|, the transverse momentum pT and the muon isolation ∆R were studied for true and reconstructed muons. This led to the conclusion that most muons which are not reconstructed have too large pseudorapidity. Since the tracker does not cover this region either, using it for a tag and probe approach does not increase the efficiency of the selection of the sample.