(Offline JetProb) 10 -Log 0 1 2 3 4 5 6 7 8 9 10 Entries -1 10 1 10 2 10 3 10 4 10 5 10 ATLAS Preliminary =7TeV s QCD MC Data Data with Loose cut at trigger level Data with Medium cut at trigger level Data with Tight cut at trigger level [GeV/c] T Track p 0 5 10 15 20 25 30 Entries 10 2 10 3 10 4 10 5 10 ATLAS Preliminary =7TeV s b-jet c-jet light jet Data 2011 Track multiplicity 0 5 10 15 20 25 30 Entries 1 10 2 10 3 10 4 10 5 10 6 10 ATLAS Preliminary =7TeV s b-jet c-jet light jet Data 2011 Signed transverse IP significance -10 -5 0 5 10 Entries 1 10 2 10 3 10 4 10 ATLAS Preliminary =7TeV s b-jet c-jet light jet Data 2011 The HLT uses the reconstructed tracks information to discriminate b- jets from light jets. b-tagging relies on the relatively long lifetime of B hadrons formed in the hadronisation of b-quarks which allow the B hadron to travel 0.45 mm in its rest frame before it decays. The algorithms exploit this by identifying tracks from the B hadron decay which are displaced from the primary interaction vertex and thus indicate that a long-lived particle was present. Two algorithms based on reconstructing the distance of closest approach, or Impact Parameter (IP), of tracks in both the longitudinal ( z 0 ) and transverse ( d 0 ) direction with respect to the beam line are used. The d 0 and its error depend on the beam spot estimate, so the HLT b-tagging relies on a good estimate of the beam spot width and position. I. Likelihood tagger A likelihood ratio W is used. W = tracks i=1 signal (θ i ) background(θ i ) S (z 0 )= z 0 − z pv σ (z 0 ) S (d 0 )= d 0 σ (d 0 ) II. JetProb tagger A resolution function R(x) is fitted to S (d 0 ) in order to compute the probability that each track comes from the primary vertex: P track = −|S (d 0 )| −∞ R(x)dx A corresponding jet probability is computed from the individual track probabilities giving an estimate of the probability of the jet coming from the primary vertex. For b-jets, this probability is expected to be close to zero. Note the dependence of this tagger on the beam spot estimate at the plot above. • From 2011, the ATLAS b-jet trigger is active and selecting events enriched in heavy flavor. • For the first months of data taking the JetProb b-tagging algorithm is used with other more advanced taggers under commissioning. • One of the main developments currently pursued are data-driven measurements of the online and offline b-tagging efficiencies. LVL2 Jet probability to originate from PV 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Normalised events -2 10 -1 10 ATLAS = 7 TeV s Preliminary Run 160472 / 160530 before HLT beam spot update after HLT beam spot update Due to the high event rate and low cross section of the processes of interest, ATLAS developed a three-tiered trigger system to select events to be recorded. • Level 1 is hardware- based and uses the calorimeters and the muon spectrometer with coarse granularity. It selects regions of interest, which are further analysed at Level 2 and Event Filter. • Level 2 and Event Filter are implemented in software and have a longer time to select events and have access to more detailed information. • The Event Filter has ~ 4 s to take the decision, while Level 2 has ~ 40 ms, allowing a more detailed analysis at the Event Filter. The Level 2 and Event Filter are referred to as the High Level Trigger (HLT). The Large Hadron Collider (LHC) collides protons at 7 TeV center-of- momentum energy. The ATLAS experiment is a general purpose experiment at the LHC, that measures physics observables of the final state particles arising from the LHC collisions. The ATLAS b-jet Trigger Danilo Enoque Ferreira de Lima (University of Glasgow) On behalf of the ATLAS Collaboration ATLAS Trigger b-jet Trigger Selection The ATLAS experiment Data/MC Comparison Outlook The plots show the signed transverse impact parameter, the selected track multiplicity and transverse momenta at the Event Filter, for simulated events and data collected in 2011. The selected tracks are used to calculate the signed impact parameters, which are, in turn, used in the JetProb variable. Trigger Menu The b-jet triggers enhances the ATLAS physics potential selecting events with final states containing several b-jets. Various analyses are currently based on b-jet triggers: • t ¯ t analyses with fully hadronic top quark decays • SUSY with 0 leptons • Searches for Higgs with Higgs decaying into b ¯ b In order to study its performance and measure efficiencies, a set of unbiased triggers exist. Effect of HLT on Offline b-tagging The offline JetProb variable is used during analysis on acquired data to tag b-jets, applying a cut to select jets with -log(JetProb) greater than a value. The effect of the b-jet trigger selection at the Event Filter is shown on data. Three working points of the HLT b-tagging are shown in the plot: “loose”, “medium” and “tight”. 1 2 3 4 5 6 7