Journal Club Meeting Minutes (5/21/2019) Paper: Jet energy measurement and its systematic uncertainty in proton-proton collisions at sqrt(s) = 7 TeV with the ATLAS detector (https://arxiv.org/pdf/1406.0076.pdf ) * Plan * Weeks 1-2: What goes into building reconstructed jets, In situ corrections * Week 3: Particular topics of interest. Certain types of jets, track jets, particle flow, heavy quark jets. * See end of minutes for new plan * Why not the modern paper? * Ben recommended this paper because it’s more detailed (good introduction to the essentials). Maybe talk about the newer paper (the improvements, changes) in later weeks. * Detectors * What are the two types of pileup? * In Time -- Pileup from the same bunch crossing as your event. Jets not related to the event you’re interested in. * Out of time -- Pileup from other bunch crossings. * Liquid argon calorimeter (600ns → 12 bunch crossings for 2011). * LHC designed for bunch crossing=25ns. 2011 was 50ns. * EM calorimeter -- Soft scatter hidden by hard scatter objects. * 2.2. End of first paragraph: Fast, bipolar shape. * No pileup noise in 2010. * Noise in calorimeters. * 2011: Pileup=8 * 2012: Pileup=20 * Run 2: Pileup=80. * Neutral particles * Run 1: Jet definition based on jet activity * Run 2: Jet activity from tracks in Inner Detector; lose information from neutral jets. Affects MET calculation. * Count activity around tracks when doing soft-term (underlying event activity) for MET. * Modern paper → Big change in pileup * Section 5.1 * Topoclusters -- Parts of the calorimeter next to each other. * More important for large pileup. * Signal / noise * Assume pileup is symmetric background and baseline you can subtract from. * Is this a good assumption? Can you do better? * Is it equally distributed among φ? Depends if beam is centered around detector. Can be biased if beam is off-center. We have a good idea of the beamspot in x direction, but not y direction. * Only care about beamspot being centered for super precise measurements (e.g. W mass). * Section 5.2 * Anti-kT algorithm. Clustering by minimize value (look at all pseudo-jets in calorimeter and calculate the distance between them (pT to some power (^2--kT algorithm, ^0?, ^-2?) times distance). Algorithm that clusters jets. Will go into in more detail in later weeks. * Can choose different radii for jets. R=0.4/0.6 is standard for Run 1 and 2. Run 2 standard is R=0.4, except it is R=1.0 for large jets. * What are we trying to correct by calibrating the jets? * Get reconstructed jets to match underlying objects. * Measurement vs true reference. * Example: Let’s say you have a 50GeV jet in a detector, and you want to calibrate the jet. * What is the energy of the underlying object? Correctly identify and add the constituent particles of the jet. * Is this the same as MC? Do you want to add any calibration to MC? Reference is truth jet, not simulated particles through detector; initiating particle and stable particles. * Step 1: Match the energy of the object that you’re measuring. * Step 2: If calibration is perfect, would you still expect to see differences between data and MC? Check for mismodeling in MC. * Went over Figures 2 and 3. * Step 1: Real detector calibration. Step 2: MC calibration. * Calorimeters -- Is it a sampling or non-sampling calorimeter? Is it compensating or non-compensating? * Sampling/non-sampling : Radiator vs Absorber. All of our calorimeters is sampling. * EM cal and Hadron cal. Have to calibrate separately because two different scales; EM compensating but Hadron not? LCW wouldn’t be necessary if we had a compensating calorimeter, but compensating calorimeter is harder to make. * Difference between energy scale and energy resolution * EM scale vs LCW scale jets? * EM is pretending all energy is the same. * LCW incorporates hadronic information by including signal shapes, etc. * Corrections are same with modern technique, but the order is different. * Karol showed modern plot. * Global sequential calibration. * Composition of jet different between quark and gluon. * There’s a paper on the ratio in equation 1. (Section 5.2) Correction is by inverting the formula. * Numerical Inversion (Ben’s paper): https://arxiv.org/pdf/1609.05195.pdf * Most of the jet calibration is this. Not as much time is spent on this as the in situ corrections because this process is well known. * In situ corrections -- Momentum balance. * 5.3. Jet quality selection * Jet cleaning -- Is your jet actually real or not? * Beam-gas events, beam-halo events, cosmic ray muons, calorimeter noise * Selections: Looser, loose, medium, tight * Looser vs tight? Things you want -- tight cut. Background -- looser criteria. * Can think of like a ROC (receiver operating characteristic) curve. * High pT -- Good selection efficiency for medium and tight. 15% and 4% for low pT. pT dependence. * Dijet measurement -- Loose because no background * High precision measurement -- Tight for the signal; lowers efficiency but less systematics. Not statistically limited. * If MET is important variable, use tight selections to make sure MET is as accurate as possible. * What do you want to get from MET/sqrt(sum of ET)? * Tag and probe. Good example is ttbar. For lepton efficiency, use Z and J/ψ. * What are you rejecting with MET cut to get good dijet sample? See if there are any irregularities in phi to regulate phi dependence. Dijet want something perfectly balanced. * pT balance. Angular balance and energy balance. * What to do next week: * Cover 6.1 and 6.2 to talk about pileup correction. Then cover the different in situ corrections. Figure out how they do the in situ correction, and then rest will be similar process. * Modern method for 3rd week.