SUSY Searches at CMS in the Fully Hadronic Channel Project B2 - Supersymmetry at the Large Hadron Collider Christian Autermann, Sergei Bobrovskyi, Ulla Gebbert, Kolja Kaschube, Friederike Nowak, Benedikt Mura, Christian Sander, Peter Schleper, Matthias Schröder, Torben Schum, Jan Thomsen Institut für Experimentalphysik, Universität Hamburg SFB 676 - Visit from India - 9th July 2008 C. Sander (Hamburg) Susy at the LHC 9th Jul 08 1 / 27
The Project B2 Supersymmetry at the Large Hadron Collider The Goals 1 Discovery of supersymmetry 2 Determination of model parameters (in particular the sparticle masses) The Challenges Understanding of the detector (reconstruction efficiencies, energy resolutions... ) Separation of SUSY signal against SM background Suppression of combinatorial and SUSY background to access SUSY mass parameters by kinematic properties C. Sander (Hamburg) Susy at the LHC 9th Jul 08 2 / 27
Outline Introduction SM background rejection and discovery of SUSY Constraining the msugra parameter space by: Event rates and weak gauge boson production rates Reconstruction of mass edges in invariant mass spectra Summary C. Sander (Hamburg) Susy at the LHC 9th Jul 08 3 / 27
Supersymmetry Why is it so nice? Problems of the Standard Model No gauge coupling unification Hierarchy problem Fine tuning problem No DM candidat Simultanous Solution with Supersymmetry (SUSY) SUSY particles change running of couplings Hierarchy/fine tuning: SUSY-contributions have opposite sign cancellation logarithmic scale dependence DM: lightest neutralino is (often) perfect candidat (massive, stable, only weak interaction) C. Sander (Hamburg) Susy at the LHC 9th Jul 08 4 / 27
Supersymmetry So far no discovery, only hints supersymmetry is broken msugra 5 new Parameters m 0 : unified mass breaking term of the fermion partners m 1/2 : unified mass breaking term of the gauge boson partners tan β: ratio of the VEVs of the two Higgs doublets unified trilinear coupling A 0, sign(µ) Contraints on the Parameter Space Higgs mass m h > 114.4 GeV Br(b X sγ) = (3.43 ± 0.36) 10 4 a µ = (27 ± 10) 10 10 Ω DM h 2 = 0.113 ± 0.008 SUSY mass limits from LEP, Tevatron, etc. Theoretical constraints: EWSB, LSP neutral Direct detection limits for DM from nuclear recoils and ν fluxes from sun or earth... C. Sander (Hamburg) Susy at the LHC 9th Jul 08 5 / 27
Supersymmetry Status of the preferred regions of the msugra parameter space Natural prior for tan β flat in other (more fundamental) parameters (µ and Bµ) Flat prior for tan β Large regions of multi dimensional parameter space allowed B. Allanach and D. Hooper, hep-ph/0806.1923 C. Sander (Hamburg) Susy at the LHC 9th Jul 08 6 / 27
Supersymmetry Agreed msugra Benchmark Points Point LM1 LM2 LM3 LM4 LM5 LM6 LM7 LM8 LM9 LM10 HM1 HM2 HM3 HM4 m0 [GeV] 60 185 330 210 230 85 3000 500 1450 3000 180 350 700 1350 m1/2 [GeV] 250 350 240 285 360 400 230 300 175 500 850 800 800 600 tan β 10 35 20 10 10 10 10 10 50 10 10 35 10 10 A0 [GeV] 0 0 0 0 0 0 0-300 0 0 0 0 0 0 σtot NLO(LO) [pb] 55(43) 9.4(7.3) 45(34) 25(19) 7.7(6.0) 4.9(3.8) 6.8(3.8) 12(8.8) 40(23) 0.076(0.041) 0.045(0.043) 0.065(0.061 0.047(0.043) 0.10(0.077) sign(µ) = + G. L. Bayatian et al., Physics TDR vol. II, J. Phys. G 34 (2007) 995 C. Sander (Hamburg) Susy at the LHC 9th Jul 08 7 / 27
Susy Signature In the Fully Hadronic Decay Mode Example diagram: Signature Cascade decay of primary produced SUSY particles R-parity conserving models LSP at end of decay chain is stable E T Many jets Jet pairs compatible with a W ± or a Z C. Sander (Hamburg) Susy at the LHC 9th Jul 08 8 / 27
Compact Muon Solenoid Multi purpose detector at the LHC (pp collider, CM energy s = 14 TeV, 2πr 27 km) Inner magnetic field B 3.8 T Calorimeter mostly inside the magnet 7... 10 interaction length High granularity ECal; largest Si-Tracker ever built... C. Sander (Hamburg) Susy at the LHC 9th Jul 08 9 / 27
Compact Muon Solenoid Event Display Physic objects (jets or tracks) are composed from detector signals Response of HCAL towers is important for jet physics and E T HCAL calibration Pile-Up from other interactions ( 20 events at high lumi) makes it more difficult C. Sander (Hamburg) Susy at the LHC 9th Jul 08 10 / 27
SM Backgrounds QCD, t t, W +jets and Z +jets Preselection cuts N jets > 3 (signal signature) E T > 200 GeV (signal signature) φ( E T, j 1,2,3 ) > 0.3 (QCD rejection) E T + E t,j2 + E t,j3 + E t,j4 > 500 GeV (S/B optimization) E t,j1 > 180 GeV and E t,j2 > 110 GeV (S/B optimization) Indirect lepton veto, cleanup... More than 7 orders of magnitude larger background (in optimistic SUSY scenarios) C. Sander (Hamburg) Susy at the LHC 9th Jul 08 11 / 27
SUSY Discovery S/B 300 can be achieved at LM1 with a signal efficiency of 10% For such optimistic scenarios with light SUSY masses almost bg free SUSY sample; discovery might be more difficult for heavier masses Jan Thomsen C. Sander (Hamburg) Susy at the LHC 9th Jul 08 12 / 27
QCD Background from Data E T tails from MC: incomplete understanding of the detector and insufficient statistics Technique Signal region (C) defined by cuts on various variables Do all cuts except on two (uncorr.) vars with sufficient separation power Challenges Variables are not uncorrelated Norm. factor N B /N A has to be extrapolated Signal contamination in A, B and D Use N C = N D N B /N A to estimate bg events in C Estimate should be conservative including all uncertainties C. Sander (Hamburg) Susy at the LHC 9th Jul 08 13 / 27
QCD Background from Data MHT vs. φ( E T, j 1,2,3 )) Signal contamination leads to overestimation (no problem for discovery) Data driven background estimation for other channels (Z νν, t t... ) also under development C. Sander (Hamburg) Susy at the LHC 9th Jul 08 14 / 27
Determination of Mass Parameters In the Fully Hadronic Decay Mode For one decay chain: combine following objects to invariant squark mass LSP χ 0 Two W -jets Quark jet of the decaying squark LSP momenta not measured no peak in the invariant trijet mass but... Distribution of m jjj with a lower and upper mass edge Big problem: combinatorial background (e.g. 7 jets 21 dijet and 35 trijet combinations) C. Sander (Hamburg) Susy at the LHC 9th Jul 08 15 / 27
Find Hadronic W ± s and Z s Defining Candidates Jetalgorithm: Iterative Cone 0.5 Jet cuts: p t > 20 GeV, η < 2.5 Boson candidate: dijet object with 70 GeV<M inv <110GeV Large combinatorial Background detector level: Reconstruction Efficiency Low W reco efficiency at low p T due to low jet reco eff. Low W reco efficiency at high p T due to W -jet merging C. Sander (Hamburg) Susy at the LHC 9th Jul 08 16 / 27
Find Hadronic W ± s and Z s Suppression of Combinatorial Background detector level: Discriminating Variables θ is the angle (in the W rest frame) of a W -jet to the flight direction of the W p T of W candidate Angle φ between E T and W C. Sander (Hamburg) Susy at the LHC 9th Jul 08 17 / 27
Boson Candidate Rate Tops in SUSY generator level: One hadronic decaying top has one true dijet W combination Wrong dijet W combinations have invariant mass "near" the W mass more than one W candidate per top due to detector smearing Separation of candidates in two classes Top-W -candidates which can be combined with one further jet to top quark SUSY-W -candidates which can not be combined to top C. Sander (Hamburg) Susy at the LHC 9th Jul 08 18 / 27
Boson Candidate Rate Probing the msugra Parameter Space L = 3 fb 1 LM3 LM4 LM8 LM10 HM3 HM4 N SUSY,tot 136400 71500 36300 230 140 310 N presel 11600 9020 5740 15 48 80 Ntot Cand 49600 33100 29000 88 140 514 ɛ Boson,tot 0.09 0.12 0.13 0.12 0.16 0.13 N Cand 790 670 360 1.0 2.5 4.3 SUSY cut ɛ SUSY Boson 0.24 0.45 0.42 0.3 0.4 0.21 N Cand top cut 2640 1560 1520 3.9 5.0 24 ɛ top Boson 0.14 0.11 0.28 0.33 0.26 0.47 R 0.30 0.43 0.57 0.26 0.5 0.18 Purity of Boson from SUSY decays increased by a factor > 3 Absolute number of W and Z candidates in different classes and ratios R = NSUSY Cand cut /NCand top cut are varying for different msugra parameters Possible to distinguish between different models? Friederike Nowak C. Sander (Hamburg) Susy at the LHC 9th Jul 08 19 / 27
Probing the msugra parameter space χ 2 (N E, N B, R) χ 2 (N E ) Scan hypothesis and compare with pseudo data of (m 0 = 800 GeV, m 1/2 = 600 GeV) Boson candidate rate contains information in addition to absolute event rate larger parts of the parameter space can be excluded Idea: Use other observables (e.g. invariant 3-jet (squark) mass) C. Sander (Hamburg) Susy at the LHC 9th Jul 08 20 / 27
Reconstruction of Mass Edges Signal definition: hadronic decay of squarks of 1./2./(3.) Generation SUSY events with 3 or more jets W candidate: 2 jets with dijet mass 70 and 120 GeV Trijet combinations: dijet objects + one of the two p t hardest jets (squark jet candidat) Up to 20 combinations per event Problem: Signal combination almost not visible At beginning: S/B 1/100 C. Sander (Hamburg) Susy at the LHC 9th Jul 08 21 / 27
Suppression of Combinatorial Background With Likelihood Ratio Invariant dijet mass Overall 17 variables with different probability densities, e.g. Invariant mass of boson candidate Invariant mass of remaining jets of the event R between boson candidate and remaining jets... For each parameter set: determine up to 5 best variables which seperate signal and combinatorial background R between boson candidate and remaining jets C. Sander (Hamburg) Susy at the LHC 9th Jul 08 22 / 27
Suppression of Combinatorial Background All combinations Best combination per event Cut on likelihood ratio > 0.9 Choose only best combination per event Increase S/B from 1/100 to 1/10 Background might be signal like C. Sander (Hamburg) Susy at the LHC 9th Jul 08 23 / 27
Determination of SUSY Parameters So far trijet mass is not used to improve bg supression Idea: Use this variable for SUSY parameter determination Method: 1 Definition of pseudo data 2 Scan over different hypothesis 3 Choose best variables and likelihood ratio cut to optimize S/ B for each hypothesis 4 Compare trijet mass distribution for hypothesis and pseudo data (with binned maximum likelihood) A 0 = 0 GeV, tan β = 50 and sign(µ)=+ C. Sander (Hamburg) Susy at the LHC 9th Jul 08 24 / 27
Determination of Parameters Binned Likelihood L = N bins i=1 e H i H D i i D i! with hypothesis normalized on data: H i = D i Good hypothesis Bad hypothesis C. Sander (Hamburg) Susy at the LHC 9th Jul 08 25 / 27
Determination of Parameters First Preliminary Results Pseudo data: m 0 = 600 GeV and m 1/2 = 400 GeV True squark mass (of pseudo data) m q 1030 GeV Hypothesis agreeing best with true squark mass has smallest negative log(likelihood) Ulla Gebbert C. Sander (Hamburg) Susy at the LHC 9th Jul 08 26 / 27
Summary SUSY Searches in the Fully Hadronic Channel Discovery of SUSY might be easy (if nature is kind) Data driven background estimations under development Understanding of detector and suppression of combinatorial background is crucial Access to SUSY parameters might be much more difficult (if nature is unkind) Probing the msugra parameter space by new observales like weak gauge boson production rate or invariant mass distributions C. Sander (Hamburg) Susy at the LHC 9th Jul 08 27 / 27