The ATLAS Missing ET trigger

Over the last few months, the ATLAS detector collected 900 GeV LHC collision events which allowed for the study the performance of the ATLAS Trigger and Data Acquisition system (TDAQ). With the 7 TeV collision data collected recently, the performance studies of the trigger system are critical for a successful physics program. In particular a large spectrum of physics results will rely on the capacity of the ATLAS TDAQ system to collect events based on the estimate of the missing transverse energy (MET) contained in each event. The MET trigger would be, for example, the primary trigger to be used in new physics searches for processes involving new weakly interacting particles, which could account for the astronomically observed dark matter. In addition to discovery perspectives, the MET trigger can also be used in combination with other triggers to control the rate of signatures involving low energy objects. For example, the MET trigger is necessary in order to measure non-boosted W in the tau channel. Finally, the negligible statistical correlation with lepton triggers make it ideal to estimate the performance of those lepton triggers (and vice versa). The MET trigger requires that the magnitude of the vector sum of all transverse energies exceeds some threshold. At L1, only calorimeter information is used, with ad-hoc hardware computing first the analog sums of all cells aligned along a given $(eta,phi)$ direction (a so-called "trigger tower"), later summing over all trigger tower digital words. Instead of computing MET, a look-up-table is used to accept/reject the event based on the $(x,y)$ components of the measured energy. At L2, the L1 result is refined by applying a correction taking into account the muons reconstructed at L2, whereas at the event filter (EF), the third trigger level, the full calorimeter is accessed to make a more precise MET estimate. Special triggers also exist, which set a minimum threshold to the scalar sum of transverse energies. Such a trigger provides a good estimate of the overall calorimeter activity, and constitutes a good complement to the MET trigger for performance studies as much as for physics purposes. The ATLAS MET trigger has been successfully tested with 2009 and 2010 initial proton-proton LHC collisions data. The comparison with the full offline reconstruction confirms the expected behavior obtained with simulated data and demonstrates the feasibility of a trigger based on global objects, such as the vector and scalar sum of all transverse energies. In the presentation, a quick overview of the key features of the MET trigger will be shown followed by a summary of the results of the MET and SumET measurements obtained at the trigger level with early LHC collisions data. Finally, performance estimates of the different MET trigger algorithms will be presented.