Search for VBF H → ττ in CMS with 1 fb − 1 and commissioning of the particle-flow with data Lorenzo Bianchini 1 LLR, ´ Ecole Polytechnique 02 April 2010, CMS France, Saclay 1 on behalf of the VBF H → ττ subgroup at LLR/Saclay Lorenzo Bianchini Search for VBF H → ττ in CMS 1 / 26
Outline Search for the Standard Model H → ττ in weak boson fusion (VBF) 1 at the LHC. The VBF H → ττ → l + τ jet channel in CMS: strategies with 1 fb − 1 . 2 Commissioning of the particle-flow event reconstruction with the 3 December 2009 data at √ s = 900 GeV. Conclusions and future plans. 4 Lorenzo Bianchini Search for VBF H → ττ in CMS 2 / 26
Why VBF H → ττ → l + τ jet ? At the LHC: σ ( qqH ) ≈ 0 . 1 · σ ( pp → H + X ); 3% < BR ( H → ττ ) < 7% for 115 < m H < 145 GeV; BR ( ττ → l + τ jet ) ≈ 2 · BR ( τ → e ) BR ( τ → hadrons) ∼ 45%; l + τ jet ⇒ trigger efficiency; ǫ ( τ ID) > 45% ⇋ QCD τ fake-rate < 2%; Lorenzo Bianchini Search for VBF H → ττ in CMS 3 / 26
The VBF H → ττ → l + τ jet signature jet rapidity gap e / μ τ τ rapidity gap hadrons jet No color exchange two forward jets from the interacting quarks, with large η separation; one lepton plus tau-jet in the central region; suppressed central hadron activity due to lack of color exchange between the scattering quarks. Lorenzo Bianchini Search for VBF H → ττ in CMS 4 / 26
Search for VBF H → ττ → l + τ jet at a hadron collider Signal: 1 isolated lepton ( e or µ ) + 1 tau-jet + ≥ 2 jets + � E miss . T Mass of the tau-pair ( M ττ ) ⇒ sensitive to the presence of the Higgs boson. Two main sources of SM background: physics background: (QCD+EWK) Z → ττ + jets; reducible background: W + jets, QCD multi-jets, t ¯ t , γ + jets; Use the VBF signature and di-tau constraints as a handle to suppress the background via a cut-based event selection: at least two jets (e.g. p T > 30 GeV/ c ) in opposite emispheres, with large rapidity separation and invariant mass (e.g. | ∆ η | > 4 , m jj > 500 GeV / c 2 ), veto on extra-jets between the tag jets (e.g. E extra jet < 20 GeV), T physical reconstruction of M ττ (e.g. collinear approximation valid, m T ( l , MET ) < 40 GeV/ c 2 ). Lorenzo Bianchini Search for VBF H → ττ in CMS 5 / 26
Search for VBF H → ττ → l + τ jet in CMS Strategies for the VBF H → ττ → l + τ jet channel with 1 fb − 1 at √ s = 14 TeV can be found in CMS PAS HIG-08-008 . Data-driven modeling of the M ττ spectrum: Z → ττ : from real Z → µµ events with τ embedding (and with only marginal systematics); QCD: from data, using ABCD methods on OS/SS samples; W and t ¯ t : from their leptonic channels with τ embedding (for real taus), from MC simulation and measured tau fake-rates (otherwise); τ -ID efficiency: from Z → ττ events with ∼ 5% systematic; τ fake-rate: from Z + jets ( Z → µµ ) with 10% systematic; e , µ trigger&offline efficiencies, e → τ fake-rate: from Z → ee with < 1% systematic using tag-and-probe. Jet veto for Z → ττ : from Z → µµ events with 5% systematic. JES and MET scale: from γ +jet and/or QCD di-jets, with ∼ 5% systematic. Lorenzo Bianchini Search for VBF H → ττ in CMS 6 / 26
Search for VBF H → ττ at CMS in the particle-flow language We have revisited the perspective studies on the VBH channel incorporating a new technique for the global event reconstruction recently developed by the CMS Collaboration: it is known as particle-flow ( PAS PFT-09-001 ). On top of this new technology: optimization of the analysis reach (fine-tuning the selection cuts 1 against the background), improvement on the di-tau reconstruction efficiency ( rescue 2 algorithm for events failing the collinear approximation). You can find all details in AN-2010/073 . Lorenzo Bianchini Search for VBF H → ττ in CMS 7 / 26
Partcle-flow objects Vs calo-objects Particle-flow ⇔ global ( i.e. unique and complete) event description at the level of the single reconstructed and identified particles exploiting in an optimal way the redundancy of the CMS detectors. Better peformances on � = − � all particles E miss p i � T (in both magnitude T i and direction), as well as on jets and taus. ⇒ direct impact on M ττ . σ/ M =14% σ/ M =9.7% Lorenzo Bianchini Search for VBF H → ττ in CMS 8 / 26
Improving the collinear approximation If � p ν ∝ � p τ : p ττ from a µ parallelogram construction. If � E miss lies outside the T ”parallelogram” ⇒ unphysical E ν < 0 solutions. For not back-to-back taus, this may happen when 2 /10GeV/c 0.18 Mass distributions for qqH(135) after VBFPF cut. E T ν τ >> ( << ) E T ν τ + ν l ; we 0.16 CA -1 rescue the event assuming 0.14 Events/fb ICA 0.12 E miss = E T ν τ ( E T ν τ + ν l ). T 0.1 Use Z → ττ to correct the 0.08 residual underestimation . 0.06 0.04 Up to 30% more signal events. 0.02 0 0 50 100 150 200 250 300 2 mass [GeV/c ] Lorenzo Bianchini Search for VBF H → ττ in CMS 9 / 26
Fine-tuning of the cut-flow Need for further requirements to characterize the VBF process; Rectangular-cuts on the kinematics of the tag jets are quasi-optimal againts Z / W +jets. The central jet-veto can be relaxed to increase the signal efficiency at no price for the significance. 8 1 2 (j-j)| fb 1 Z2+3j, PF AK4 η 0.9 Background Efciency 7 eff(qqH)/Sqrt[eff(ttbar)] 1.8 ∆ Z2+3j, PF IC5 jets | 0.8 Z2+3j, Calo IC5 jets for rect. cut on [ Δη, Mjj] 6 0.8 1.6 0.7 5 1.4 0.6 relative to S=1.23 0.6 4 0.5 S=1.25 max signif. 1.2 0.4 3 0.4 1 VBF relaxed 0.3 2 VBF full 0.8 0.2 0.2 VBF loose eff(qqH)/Sqrt[eff(Z+jets)] 1 0.6 0.1 best rectangular for central jet-veto 0 0 500 1000 1500 2000 2500 3000 0 0.4 0.4 0.5 0.6 0.7 0.8 0.9 1 M [GeV] j,j Signal Efciency Lorenzo Bianchini Search for VBF H → ττ in CMS 10 / 26
Exclusion limits and projection for √ s = 7 TeV Exclusion limits at 95% c.l. for the combined e + µ channel using the CLs method. Lorenzo Bianchini Search for VBF H → ττ in CMS 11 / 26
Commissioning of the particle-flow at √ s = 900 GeV With 1 fb − 1 of integrated luminosity, the particle-flow event reconstruction will be fully optimized and commissioned, thus providing powerful analysis tools; The first collisions delivered by the LHC at √ s = 900 GeV, and recorded by CMS in December 2009, already served for a succesful commissioning of the building bricks of the algorithm. In the next months we expect to complete and re-commission the particle-flow at the higher c.o.m energy. In the following, I will show show some results from the recent commissioning paper of the particle-flow PAS PFT-10-001 with the purpose of giving a basic idea of the algorithm and to stress the excellent data/simulation agreement. Lorenzo Bianchini Search for VBF H → ττ in CMS 12 / 26
The work-flow calo/tracker/muons hits clustering/tracking clusters-tracks linking Pflow blocks reconstruction + identification Pflow particles isolation taus jets jets . . . . MET deposits Lorenzo Bianchini Search for VBF H → ττ in CMS 13 / 26
Commissionig of the link algorithm Figure: η − φ view of two charged Figure: η − φ view of two charged tracks all their way to ECAL. tracks all their way to HCAL. p track = 14 . 64 GeV/c → calibrated E ECAL + HCAL = 14 . 33 GeV; T p track = 10 . 94 GeV/c → calibrated E ECAL + HCAL = 9 . 19 GeV. T Lorenzo Bianchini Search for VBF H → ττ in CMS 14 / 26
Energy calibration 500 500 Number of photon pairs Number of photon pairs fit 2 m = 131.6 ± 0.3 MeV/ c π 0 400 400 2 σ = 12.9 ± 0.3 MeV/ c m 0 π 900-GeV Data 300 300 CMS Preliminary 2009 200 200 100 100 0 0 0.05 0.05 0.1 0.1 0.15 0.15 0.2 0.2 0.25 0.25 0.3 0.3 0.35 0.35 0.4 0.4 0.45 0.45 0.5 0.5 2 2 Mass (GeV/c Mass (GeV/c ) ) Figure: Average calibrated calorimeter Figure: π 0 mass plots. All PF photons response as a function of the track with | η | < 1 and E > 400 MeV are momentum. Linear fit: (0 . 920 ± 0 . 037) paired. The mass peak is only ∼ 2% smaller than the world average. For the energy-scales probed so far, the calibration is adequate within 2% (5%) for ECAL (HCAL). Lorenzo Bianchini Search for VBF H → ττ in CMS 15 / 26
Commissioning of the particle-flow jets Jet 1 p T = 22 GeV/c Jet 2 p T = 42 GeV/c HF em NH PH HF hadr MET 1.9 GeV CH Jet 3 p T = 38 GeV/c Figure: PF event at √ s = 2 . 36 TeV: Figure: Fractional components of PF the small recorded E miss (1 . 9 GeV) jets measured at √ s = 900 GeV. T validates the PF reconstruction Lorenzo Bianchini Search for VBF H → ττ in CMS 16 / 26
Recommend
More recommend
