. Early SUSY analyses with ATLAS Giacomo Polesello INFN, Sezione di Pavia
Early analyses at the LHC The LHC will start producing high-energy collisions in the next months Large uncertainties on the data-taking parameters: • Energy (ies) • Integrated luminosity • Luminosity profile However, with a baseline expectation of √ s = 7 TeV and 200 pb − 1 of integrated luminosity, we can expect to cover areas of new physics not explored by the Tevatron Low mass SUSY is an example where we may be able to say something new Combined performance and physics groups in ATLAS have developed a program of work aimed at taking advantage of this possibility Explain today the path through which we plan to address SUSY searches based on the data we will collect in 2010
Before starting searches for new physics With the first few pb − 1 of data, the Collaboration will perform the basic work for understanding of detector performance. Once the reconstruction of the basic building blocks for physics analysis: Jets, electrons, muons, / E T , .... under reasonable control the first physics analyses will start Start with simple analyses of basic SM processes which can be based on a limited level of detector understanding, and in parallel continue the commissioning work As detector understanding improves and statistics cumulates more sophisticated analyses will become possible Aim at detailed measurements of Standard Model cross-sections and first searches when integrated lumi is of order 100 pb − 1
SUSY production at the LHC Production dominated by strongly √ s (TeV) σ SUSY (pb) σ SUSY (pb) σ tt (pb) interacting sparticles: ˜ q , ˜ g q and ˜ ˜ g production cross-section SU3 SU4 ∼ only function of their masses, 7 1.9 36 148 ∼ independent of details of SUSY 10 6.5 103 374 model 14 18.9 264 827 Show LO Cross-sections for two ATLAS m ˜ g (GeV) 717 413 172.5 benchmark points (fHERWIG) and NLO m ˜ q (GeV) 620 410 (MC@NLO) for top SU3: m 0 = 100 GeV, m 1 / 2 = 300 GeV, A 0 = − 300 GeV, tan β = 6 , µ > 0 . SU4: m 0 = 200 GeV, m 1 / 2 = 160 GeV, A 0 = − 400 GeV, tan β = 10 , µ > 0 . Squarks and gluinos are typically the heaviest sparticles ⇒ If R p conserved, complex cascades to undetected LSP, with large multiplicities of jets and leptons produced in the decay.
� � � A SUSY event in ATLAS Multi-jet event in Bulk Region 6 jets 2 high-pt muons Large missing E T
SUSY discovery: basic strategy Basic assumption: discovery from squark/gluinos cascading to undetectable LSP Details of cascade decays are a function of model parameters. Focus on robust signatures covering large classes of models and large rejection of SM backgrounds • / E T : from LSP escaping detection ~ χ 1 — q • High E T jets: guaranteed if squarks/gluinos l + ~ l + ~ 0 χ 2 if unification of gaugino masses assumed. ~ q ~ g l - q ν • Multiple leptons ( Z ): from decays of ~ ~ 0 χ 1 t 1 ~ g Charginos/neutralinos in cascade t W + — l + t W — • Multiple τ -jets or b -jets ( h ): Often abun- b q — b dant production of third generation sparticles q χ 0 Define basic selection criteria on these variables for RPC SUSY with ˜ 1 LSP Optimisation of criteria on parameter space: define set of topologies, and for each define sets of cuts aimed respectively at high and low SUSY masses
Basic analysis cuts For √ s = 10 TeV and on 200 pb − 1 define on low-mass point basic analysis cuts: Perform analyses requiring 2, 3 or 4 jets ans 0, 1 or 2 leptons in the event • P T cuts on jets and leptons depending on topology • / E T > 80 GeV • Cut on ∆ φ ( jet i , / E T ) • Cut on ratio between E T / and i =1 p jet ,i i =1 p lep ,i � 4 M effective ≡ + + / E T � T T • Transverse sphericity S T > 0 . 2 SUSY signal: SU4 point: m ˜ q ∼ m ˜ g ∼ 410 GeV (ATL-PHYS-PUB-2009-084) Observe good S/B background in most of the studied channels In paramters space further optimise statistical significance through additional cut on M effective
Reach in parameter space (200 pb − 1 , 10 TeV) Grid in MSUGRA space, and set of ‘no prejudice’ MSSM points(Tom Rizzo et al.) Reach strongly dependent on assumed value of systematic uncertainty on background evaluation Assume for this study 50% uncertainty on all backgrounds Techniques for assessing backgrounds and evaluating uncertainties are the key to SUSY analysis ⇒ Discuss examples today
0 lepton + jets analysis Large statistical significance, but many backgrounds to keep under control • QCD • top • W+Jets • Z+Jets QCD background particularly insidious as: • Multijet QCD cross-section not well known • / E T from difficult-to-model instrumental effects Look in detail at / E T measurement ATLAS and data-driven estimate of QCD backgrounds
Etmiss and SUSY Etmiss is experimentally difficult variable, as it requires summing over all the detector Any inhomogeneity in the detector performance/calibration reflects onto it Need first of all understand measurement of the gaussian ’core’ of the Etmiss distribution from fluctuations in detector response Next all the possible sources of high / E T events need to be understood and accounted for: • Detector malfunctioning (dead cells, noisy cells...) • Beam Halo • Cosmic rays • Events where particles end up in insensitive parts of the detectors • ........
Performance of / E T experimental measurement E T measurement based on assumption that all the energy is measured in the calorimeters or seen as / muons in the spectrometers Multi-step procedure correcting for experimental effects, starting from vector sum of E T deposition in calorimeter cells Measurement resolution estimated on MC by plotting the difference between true and estimated / E T separately on each of the components Resolution can be fitted as 0 . 57 · √ � E T Linearity of response 40 Resolution (GeV) 0.3 QCD Jets 35 SUSY 0.2 30 t t 0.1 A 25 → τ τ 0 20 15 -0.1 miss E calibration at EM scale T 10 miss E global calibration -0.2 T miss E global calibration+cryostat 5 T ATLAS miss ATLAS E refined calibration -0.3 T 0 0 200 400 600 800 100012001400160018002000 0 50 100 150 200 250 300 E (GeV) Σ miss T True E (GeV) T
Etmiss commissioning with random events Basic check: look at random triggers, and plot / E T distribution Use two different algorithms for cell noise subtraction: simple cut at 2 σ , 3-D energy clusters (topoclusters) Much narrower distribution for topoclusters Arbitrary Units Arbitrary Units ATLAS COSMIC 2008 PRELIMINARY ATLAS COSMIC 2008 PRELIMINARY -1 10 Cells, |E|>2 σ Cells, |E|>2 σ Run 91639 -2 10 Topo clusters 4/2/0 Gaussian noise model -2 10 -3 10 -3 10 -4 10 -4 10 -5 10 -5 10 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 miss 2 miss 2 miss 2 miss 2 (E ) + (E ) (GeV) (E ) + (E ) (GeV) X Y X Y Observe excellent agreement between measured / E T and simple gaussian model of noise in calo cells (tail now understood as detector mafunctioning) Good stability observed over 1.5 months period
Fake Etmiss: cosmic rays High energy cosmic ray muons undergoing hard bremsstrahlung can produce localised high-energy dposit in calorimeter, and thence fake / E T Observe good agreement with detailed simulation Discrepancy in tails due to MC statistics and from cosmic ray air showers not modelled in MC
TeV event from single cosmic ray muon
TeV event from cosmic ray air shower
Cleaning cuts for cosmic rays Jet EM fraction ( F EM ) : Typically 0 or 1 for muons undergoing bremsstrahlung in Tilecal of LARG Number of clusters ( N clus ) : lower for cosmics Resulting rejection after requiring: • 0 . 2 < F EM < 0 . 97 • N clus ≥ 7
Cleaning of detector malfunctions in / E T sample T E T from mismeasured multi-jet events: / Populated by detector and machine problems Example of / E T cleaning in D0 • Reject runs with detector malfunctioning • Reject events with noise in the detector • Remove bad cells Fraction of Events Fraction of Events Dead Regions Region 1 (2EM+1HAD) -1 -1 10 10 Region 2 (1EM+1HAD) ATLAS example: assume a few HV channels dead Region 3 (Good) -2 -2 10 10 ATLAS in calorimeters -3 -3 10 10 Tools being prepared to monitor and correct event- -4 -4 by-event, very active area of work 10 10 0 0 100 100 200 200 300 300 400 400 500 500 600 600 700 700 800 800 900 900 1000 1000 Missing Et (GeV) Missing Et (GeV)
Instrumental background: definition of fiducial region for jets Use a sample of 2-jet events ( p T > 280 GeV), apply basic cuts to reject events containing neutrinos E T / √ � E T , ∝ / • For each event calculate S = / E T significance • For each jet in the event, take η ( jet ) , and fill one entry in the plot • For each bin in η calculate the average value of S <S> ATLAS 1.6 1.4 1.2 1 0.8 0.6 -5 -4 -3 -2 -1 0 1 2 3 4 5 η Observe rise in significance for events with jets at interface between calorimeters Reject high / E T events with a jet falling in yellow regions
Recommend
More recommend
