Speaker
Description
Searches for new BSM physics at the ATLAS experiment typically target particles that decay promptly, very close to the collision point. New physics has so far evaded these searches. Multiple BSM models predict new long-lived particles (LLPs) with decay lengths ranging from millimetres to kilometres. In new physics scenarios where LLPs couple preferentially to heavy particles or to leptons, tau lepton final states would offer high experimental sensitivity.
A search for displaced tau decays within the ATLAS Inner Detector is presented, using a partial ATLAS Run 2 dataset. The search targets final states with two taus, where one or both taus decay hadronically. The fully leptonic scenario is experimentally covered by other ATLAS searches. A machine learning-based identification model was developed to identify displaced tau jets, reducing the dominant background contribution from QCD-induced jets mis-reconstructed as tau jets. The analysis design is model-independent, allowing for both the placement of model-independent and model-dependent exclusion limits. Future searches are planned using the ATLAS Run 3 datasets, and Run 3-specific developments of the identification model will be presented.