CCMS CR-2011/163November 6, 2018
SUSY and high- p T flavor tagging at CMS W. Kiesenhofer on behalf of the CMS Collaboration
Institute of High Energy Physics, Vienna, 1050, Austria
Abstract
We present a result of a search for supersymmetry in final states withmissing transverse energy and b-tagged jets. This search is performedwith data collected by the CMS experiment at the LHC in pp-collisionsat a center-of-mass energy of 7 TeV. Data-driven techniques used to mea-sure the Standard Model background are demonstrated. The result is in-terpreted in terms of the constrained Minimal Supersymmetric StandardModel and compared to a similar search without any b-tag requirement.PRESENTED AT
The Ninth International Conference onFlavor Physics and CP Violation(FPCP 2011)Maale Hachamisha, Israel, May 23–27, 2011 a r X i v : . [ h e p - e x ] S e p Introduction
Searches for new physics beyond the Standard Model (SM) are commonly moti-vated by strong astrophysical evidence for dark matter, and by theoretical prob-lems associated with explaining the observed particle masses while maintaining themass hierarchies in the presence of quantum corrections [1, 2]. Supersymmetry(SUSY) [3, 4, 5, 6, 7, 8, 9] is an extensively studied candidate for new physics with thepotential to solve this problems. The SUSY particle spectrum contains new particlesarising from a correspondence between SM fermions and SUSY partner bosons, aswell as SM bosons and SUSY partner fermions. At the Large Hadron Collider (LHC)supersymmetric particles, if they exist, are predicted to be predominantly producedvia the fusion of two gluons into a: • pair of gluinos (super-partners of the gluons), • pair of squarks (super-partners of the quarks), • gluino and a squark.Gluinos and squarks decay via other supersymmetric particles into quarks and otherSM particles until a lightest supersymmetric particle (LSP) is created. Under theassumption of R-parity conservation the LSP is stable and will generate missingenergy transverse to the beam line ( (cid:54) E T ) when it escapes the detector. The detailsof SUSY decay chains are strongly model dependent and since we aim to designinclusive analysis, which are sensitive over a wide range of SUSY models, we donot consider them in our analysis. However, for a large class of supersymmetricparameter sets squarks can be relatively light. If this is the case for sbottoms orstops, which can decay into b quarks, there may be an abundance of events with large (cid:54) E T and one or more b-quark jets. This proceeding summarizes a search for eventswith two or more hadronic jets in the final state, significant transverse momentumimbalance, and at least one b-tagged jet [10]. It uses the full dataset collected bythe CMS experiment [11] in 2010. This analysis extends a similar one without ab-tag requirement [12]. The momentum imbalance is characterized by the variable α T , definded in the next section.The main backgrounds for this analysis arise due to standard model multi-jetproduction (QCD background), electroweak W and Z boson production (EWK) andtop-quark pair production ( tt ). Their estimation is discussed in section 3.1 The α T variable The α T variable charactarizes the momentum imbalance of jets in the transverseplane. For a two jet system it is defined as α T = E j T M j ,j T (1)where j is the jet with the lower transverse energy ( E T ) and M j ,j T is the invariantmass in the transverse plane of the two jets. Assuming massless jets, this can berewritten for multi-jet events as α T = 12 H T − ∆ H T (cid:113) H T − (cid:54) H T (2)where (cid:54) H T = | Σ i (cid:126)p j i T | , H T = Σ i p j i T and p T denotes a jet’s transverse momentum. Alljets considered in the analysis are grouped into two pseudo-jets such that ∆ H T = | p pseudojet T − p pseudojet T | becomes minimal.The α T variable is very effective in rejecting QCD background, which would other-wise dominate the signal selection. A well measured QCD event will result in α T . α T > .
55 is used for thefinal event selection. A SUSY event with real missing transverse momentum due tothe production of LSPs can exceed this value of α T since (cid:54) H T can be sufficiently largecompared to H T . Events that were considered in this analysis had to pass at least one trigger (based on H T at trigger level) during data taking. These triggers were measured to be over 99%efficient for the final signal selection. Jets were reconstructed using the anti- k T [13]algorithm and required to have E T >
50 GeV and | η | <
3. Since this analysis aimsto be exclusively hadronic, events with an isolated lepton or photon were vetoed. Inaddition the final selection requires two jets with E T >
100 GeV, | η | < . E T jet, H T >
350 GeV, at least one jet tagged as originating from a b quark,and α T > .
55. To discriminate b jets from other flavors the TCHP (Track CountingHigh Purity) algorithm was used [10]. For this discriminator three different workingpoints are defined, namely a loose, a medium and a tight. Only the tight selection wasused in the main analysis, whereas the others were used to collect control samples.2
Estimation of Backgrounds
As stated previously the backgrounds for this analysis can be categorized into threegroups: QCD, EWK and tt .The vast majority of events which belong to the QCD background do not featurelarge momentum imbalance and are therefore rejected by the α T > .
55 requirement.Sources of artificial momentum imbalance in QCD events, which can lift α T over thesignal requirement are jet under-measurement caused by inactive ECAL regions ormultiple jets falling below the jet E T threshold of 50 GeV. Such events were identifiedand removed by control variables implemented for this purpose [15].The EWK background processes (production of Z/W+jets) with real missing en-ergy from the production of neutrinos are greatly suppressed by the requirement ofat least one b jet. Thus tt becomes the dominant background in this analysis.Three independent estimation procedures have been employed to estimate the SMbackground in the final event selection. The main method is a data-driven procedureestimating all backgrounds simultaneously. In this method the fraction of all eventswith α T > .
55, denoted F ( α T > . H T control regionand applied in the singal region (the jet E T threshold is scaled according to H T ).As studied extensively in Ref. [12] F ( α T > .
55) is independent of H T in sampleswere (cid:54) H T comes predominantly from real sources. This can be seen in Fig. 1, whichshows F ( α T > .
55) vs H T in simulation of all SM processes combined. In a QCDdominated sample F ( α T > .
55) is expected to be a decreasing function of H T dueto the H T dependence of the factors contributing to artificial (cid:54) H T , such as jet energyresolution and jet E T threshold effects. Loosening the α T requirement to α T > . F ( α T > .
51) as afunction of H T . This is shown in Fig. 2. In data F ( α T > .
55) is consistent withhaving no H T dependence, indicating that EWK and tt backgrounds dominate. Alsothe anti-b-tagged data sample is consistent with having no H T dependence. Sincethe tight b-tag requirement only further suppresses the QCD background, the tighttagged sample is expected to have a negligible QCD contribution. To predict thenumber of background events in the final signal region F ( α T > .
55) is measured ina lower H T control region (250-350 GeV) and multiplied by the number of events inthe signal region before the α T > .
55 requirement. The results of this procedure aresummarized in Table 1.The two other background prediction methods estimate the Z → νν and the tt contribution to the background and serve as a cross check to the first method.To estimate Z → νν , which is expected to be the dominant EWK background,a sample of Z → µ + µ − events with ≥ BR ( Z → νν ) BR ( Z → µ + µ − ) ≈
6. The result of this procedure shows a slight over-prediction of the number of Z → νν events due to looser selection cuts, but is consistent with the prediction of the formermethod and with the prediction from simulation.Since tt is expected to be the dominant background it is important to cross checkits contribution with a second estimation method. Simulation studies indicate that72% of the tt background consists of events with hadronic tau decays. To estimatethe hadronic tau decay yield, events with one or two muons are selected and used toemulate the hadronic decays of taus. For each muon, the presence of a jet from thehadronic decay of a tau lepton is emulated. The jet E T is determined as fraction of themuon p T , where the fraction is drawn from a distribution extracted from simulation.Again a less stringent version of the final event selection is used in which the E T threshold of the leading two jets is 80 GeV, the H T requirement is 280 GeV, and amedium b-tag is required. The measured value of F ( α T > .
55) in this sample ismultiplied by the total number of emulated events in the signal region with no α T requirement. This value is corrected for the muon selection efficiency and acceptanceand the hadronic tau decay branching ratio to obtain the hadronic tau decay yield.To account for the entire tt background, the predicted hadronic tau decay yield isscaled by a simulation derived factor. A 50% systematic uncertainty is assigned forthis factor. This procedure again yields a slight over prediction in tt simulation.This two cross checks confirm that the total background is small and consistentwith the inclusive prediction.Figure 1: F ( α T > .
55) vs H T in SM simulation. In all cases, F ( α T > .
55) isconsistent with having no H T dependence.4igure 2: F ( α T > .
51) vs H T in data. F ( α T > .
51) is a decreasing function of H T for lower H T , indicating that QCD dominates these regions.N-jets MC Background Prediction Data LM0 ≥ . ± .
26 0 . +0 . − . (stat) ± .
13 1 14 . ± . pb − . The prediction comes from the α T vs H T extrapolation described in Section 3. The LM0 uncertainty is statisticalonly. The observation of 1 data event in the signal region is consistent with the backgroundpredictions summarized in the last section. Together with an estimated systematicuncertainty of 24% on the signal selection efficiency [14] this allows the determinationof a 95% confidence level upper limit on the predicted number of observed signalevents N obs = 4 .
7. Figure 3 shows the excluded region in the ( m / , m ) planefor CMSSM (constrained Minimal Supersymmetric Standard Model) [16] parameters A = 0 GeV, tan β = 50, and µ >
0. The excluded region is extended with respect tothat of Ref. [12] without b tagging, also shown, values of m above 350 GeV, where bproduction is frequent. For models with infrequent b production, Ref. [12] sets morestringent limits, whereas this analysis has greater sensitivity to models with frequentb production. 5 (GeV) m
200 250 300 350 400 450 500 ( G e V ) / m (500)GeVq~ (500)GeV g~ (650)GeVq~ (650)GeV g~ (800)GeVq~ (800)GeV g~ = 7 TeVs, -1 = 35 pb int CMS L > 0 µ = 0, = 50, A β tan LMB <0 µ =5, β , tan q~ , g~ CDF <0 µ =3, β , tan q~ , g~ D0 ± χ∼ LEP2 = L SP τ ∼
95% CL Limits:Observed Limit, NLOExpected Limit σ ± Expected Limit Observed Limit, LO, NLO T α CMS (GeV) m
200 250 300 350 400 450 500 ( G e V ) / m (GeV) m
200 250 300 350 400 450 500 ( G e V ) / m Figure 3: Exclusion regions in the ( m / , m ) plane for one set of CMSSM parameters,for this analysis (red), and the non-b tagged version [12] (green). In this proceeding we present a search for events with multiple jets, at least one ofwhich must be b tagged, and signicant transverse momentum imbalance, using the35 pb − of the CMS 2010 dataset. No evidence for new physics is observed. Theresult of the search is characterized as an exclusion region in SUSY parameter spaceand compared to an equivalent analysis without b tagging. References [1] E. Witten, “Dynamical Breaking of Supersymmetry,” Nucl. Phys. B (1981)513.[2] S. Dimopoulos and H. Georgi, “Softly Broken Supersymmetry and SU(5),” Nucl.Phys. B (1981) 150.[3] S. P. Martin, “A Supersymmetry primer,” In “Kane, G.L. (ed.): Perspectives onsupersymmetry” 1-98. [hep-ph/9709356].64] J. Wess, B. Zumino, “Supergauge Transformations in Four-Dimensions,” Nucl.Phys.
B70 (1974) 39-50.[5] H. P. Nilles, “Supersymmetry, Supergravity and Particle Physics,” Phys. Rept. (1984) 1-162.[6] H. E. Haber, G. L. Kane, “The Search for Supersymmetry: Probing PhysicsBeyond the Standard Model,” Phys. Rept. (1985) 75-263.[7] R. Barbieri, S. Ferrara, C. A. Savoy, “Gauge Models with Spontaneously BrokenLocal Supersymmetry,” Phys. Lett.
B119 (1982) 343.[8] S. Dawson, E. Eichten, C. Quigg, “Search for Supersymmetric Particles inHadron - Hadron Collisions,” Phys. Rev.
D31 (1985) 1581.[9] J. R. Ellis, J. S. Hagelin, D. V. Nanopoulos, K. A. Olive, M. Srednicki, “Super-symmetric Relics from the Big Bang,” Nucl. Phys.
B238 (1984) 453-476.[10] CMS Collaboration, “Commissioning of b-jet identification with pp collisions atsqrt(s) = 7 TeV”, CMS PAS BTV-10-001[11] R. Adolphi et al. [CMS Collaboration], “The CMS experiment at the CERNLHC,” JINST (2008) S08004.[12] V. Khachatryan et al. [CMS Collaboration], “Search for Supersymmetry in ppCollisions at 7 TeV in Events with Jets and Missing Transverse Energy,” Phys.Lett. B698 (2011) 196-218. [arXiv:1101.1628 [hep-ex]].[13] M. Cacciari, G. P. Salam, G. Soyez, “The Anti-k(t) jet clustering algorithm,”JHEP (2008) 063. [arXiv:0802.1189 [hep-ph]].[14] S. Chatrchyan et al. [CMS Collaboration], “Search for Supersymmetry in Eventswith b Jets and Missing Transverse Momentum at the LHC,” JHEP (2011)113 [arXiv:1106.3272 [hep-ex]].[15] CMS Collaboration, “Search for Supersymmetry in Final States with b Jets andMissing Energy at the LHC”, CMS PAS SUS-10-011[16] G. L. Kane, C. F. Kolda, L. Roszkowski and J. D. Wells, “Study of constrainedminimal supersymmetry,” Phys. Rev. D49