Inclusive SUSY searches with ATLAS and CMS detectors

Inclusive SUSY searches with ATLAS and CMS detectors at LHC Renaud Brunelière Physikalisches Institut – Universität Freiburg On behalf of ATLAS and CMS Collaborations 08/04/08 R. Bruneliere - Universitat Freiburg

SUSY searches at LHC SUSY is one of the major BSM topics at LHC • it is an interesting extension of SM : • could be dark matter candidate • solves Higgs mass hierarchy problem • it can provide clear experimental signatures • Te. V-scale new physics => direct search • look for excess in tails : high missing Et + high Pt jets, … In this talk, restrict to direct searches : • of models conserving R-parity : • LSP is stable => escape detection => large missing Et • SUSY particles produced by pairs => cascade decays • GMSB models covered by Mark Terwort • focus on early SUSY searches • integrated luminosity of 1 fb-1 (1 month at nominal LHC lumi) 08/04/08 R. Bruneliere - Universitat Freiburg 2

SUSY production at LHC Production at LHC dominated by squarks and gluinos ü Cross-sections mainly depend on SUSY masses (But masses are model dependent) SUSY ? Cross section [pb] ü prospino Huge background to be suppressed SUSY Particle mass [Ge. V/c 2] 08/04/08 R. Bruneliere - Universitat Freiburg 3

SUSY signatures for early searches 1. squark and gluino production cascade decay high Pt jets 2. LSP stable large missing Et 3. Possibly some leptons Generic signature = Large MET + multi-jets + multi-leptons • • • Build mutually exclusive channels depending on number/type of objects Rather general hypothesis => signatures are quite model independent (m. SUGRA, AMSB, NUHM, …) Try to rely on data-driven determination of backgrounds : • Poor understanding of detector (missing Et tails, JES, …) with early data • Rather large theoretical uncertainties on backgrounds in this part of phase space 08/04/08 R. Bruneliere - Universitat Freiburg 4

Jets + MET + 0 -lepton signature 1 3 Meff = MET + PT(jet) 08/04/08 2 ATLAS selection cuts : 1. 4 jets, Pt(Jet 1) > 100 Ge. V, Pt(Jet 4) > 50 Ge. V, MET > 100 Ge. V 2. MET > 0. 2×Meff 3. Transverse sphericity > 0. 2, (MET, jet 1, 2, 3) > 0. 2, no isolated e or (Pt > 20 Ge. V), J 70_X 70 combined trigger Similar CMS cuts : MET > 200 Ge. V, 3 jets(ET > 30 Ge. V, | |<3), | (Jet 1)| < 1. 7 R. Bruneliere - Universitat Freiburg 5

QCD background rejection/estimation QCD jets mimic SUSY events in 0 -lepton channel through : • ‘fake’ MET due to jet mis-measurement => reducible with cuts => cut • ‘real’ MET due to decays into neutrinos (heavy flavor, B hadrons, …) CMS QCD CMS SUSY To estimate remaining QCD background after cleaning, the following method can be used : 1. Measure smearing function (non gaussian tails) in events with large MET associated to a single jet 2. Select seed events with low MET-significance and smear each jet 3. Normalize estimate to data 08/04/08 R. Bruneliere - Universitat Freiburg 6

Z/W + jets background determination • Z + jets is an irreducible background in 0 -lepton channel • But can be estimated with Z l+l- + jets • This method bring good estimate (15% ALTAS, 5% CMS) • Main limiting factor is control sample statistics (Br(Z l+l-)/Br(Z ) ~ 0. 17) • A good tail description requires MC OR extrapolation methods • W + jets background is due to W hadrons (42%) or W e/ with lepton out of acceptance (41%) or W e/ with non-selected lepton (17%) • it can be estimated from Z l+l- + jets or W l + jets control samples 08/04/08 R. Bruneliere - Universitat Freiburg 7

Jets + MET + 0 -lepton discovery reach CMS Reach vs integrated luminosity ATLAS Reach vs n. Jet criteria Relative contributions of the different backgrounds after selection : 08/04/08 Top Z+jets W+jets QCD ATLAS 43% 22% 31% 3% CMS 23% 20% 13% 44% R. Bruneliere - Universitat Freiburg 8

Jets + MET + 1 -lepton MET • Requiring at least 1 isolated lepton provide clean signature as it reduce QCD background • Main background then is top pairs Control Signal sample region A 08/04/08 B MT • Top background can be estimated with many methods • like using an additional discriminating variable MT used to select control sample region. • Normalization obtained with B/A R. Bruneliere - Universitat Freiburg 9

Search comparisons Discovery reach vs channel Discovery reach vs SUSY model 0 -lepton channel • 0 -lepton channel is most efficient channel but 1 -lepton channel is less sensitive to QCD background • in the lower part of the plane, there is a good redundancy between search channels => useful discovery cross-check • discovery reach is rather independent of model & considered scanning grid 08/04/08 R. Bruneliere - Universitat Freiburg 10

Conclusion Recent studies using full simulation (CMS PTDR, ATLAS CSC) show • ATLAS and CMS should discover R-parity conserving SUSY with gluino and squark masses < O(1 Te. V) after having accumulated and understood 1 fb-1. • Many backgrounds can be estimated with data-driven techniques but these analysis are delicate • Next “test” with real data ! Many thanks to the DIS 08 organizers 08/04/08 R. Bruneliere - Universitat Freiburg 11