Tagging and suppression of pileup jets with the ATLAS detector

The suppression of pileup jets has been a crucial component of many physics analyses using 2012 LHC proton-proton collisions. In ATLAS, tracking information has been used to calculate a variable called the jet-vertex-fraction, which is the fraction of the total momentum of tracks in the jet which is associated with the primary vertex. Imposing a lower limit on this variable rejects the majority of pileup jets, but leads to hard-scatter jet efficiencies that depend on the number of reconstructed primary vertices in the event (NVtx). In this note, new track-based variables to suppress pileup jets are developed in such a way that the resulting hard-scatter jet efficiency is stable as a function of NVtx. A multivariate combination of two such variables called the jet-vertex-tagger (JVT) is constructed. The modeling of JVT is tested in Z(→ μμ)+jets as well as in semileptonic ttbar events. The efficiencies of different JVT criteria are measured in data and compared to simulation. In addition, it is shown that jet-vertex association can be applied to large-R jets, providing a track-based grooming technique that is as powerful as calorimeter-based trimming but based on complementary tracking information. Finally, the performance of track-based grooming is compared with the recently proposed jet cleansing algorithm.