How to find a Higgs boson Heather M. Gray, CERN ( へざあ ) S ’ E R E H W ? S G G I H
higgs discovery ( 発⾒覌 ) -1 -1 CMS s = 7 TeV, L = 5.1 fb s = 8 TeV, L = 5.3 fb S/(S+B) Weighted Events / 1.5 GeV 0 Local p Events / 1.5 GeV ATLAS 2011 - 2012 Unweighted Obs. -1 ∫ s = 7 TeV: Ldt = 4.6-4.8 fb 1500 Exp. -1 1500 ∫ s = 8 TeV: Ldt = 5.8-5.9 fb 1 ± σ 1 0 σ -1 1 σ 10 1000 2 σ -2 10 -3 3 10 σ 1000 -4 10 120 130 4 σ -5 10 m (GeV) -6 γ γ 10 5 σ -7 10 -8 10 Data -9 500 10 6 σ S+B Fit -10 10 B Fit Component -11 10 1 ± 110 115 120 125 130 135 140 145 150 σ 2 m [GeV] ± σ H 0 110 120 130 140 150 m (GeV) γ γ Seminar on 4 July 2012 by the ATLAS and CMS collaborations
this talk ( この話 ) • NOT about the Higgs discovery • NOT to discuss the latest Higgs results • There are many and they are interesting • Ask me about them later if you like • But rather, to try to explain how we go about doing a Higgs analysis using a specific example • Example: http://link.springer.com/article/10.1007/ JHEP01(2015)069
choose your channel I ( あなたのチャンネルを選択 ) Gluon fusion Vector Boson Fusion (VBF) H q W, Z Production H Cross-section W, Z q σ [pb] 20 15 10 5 ttH Production 0 ( W/Z) ggF VBF (W/Z)H ttH t Production g q t H W/Z t g t H q
choose your channel I ( あなたのチャンネルを選択 ) Gluon fusion Vector Boson Fusion (VBF) H q W, Z Production H Cross-section W, Z q σ [pb] 20 15 10 5 ttH Production 0 ( W/Z) ggF VBF (W/Z)H ttH t Production g q t H W/Z t g t H q
choose your channel II ( あなたのチャンネルを選択 ) γ Decay Η b t H Probability t H γ t Other b 11% 0% 3% ττ W 6% H bb W WW 58% 22% τ H Z H τ Z
choose your channel II ( あなたのチャンネルを選択 ) γ Decay Η b t H Probability t H γ t Other b 11% 0% 3% ττ W 6% H bb W WW 58% 22% τ H Z H τ Z
build a billion dollar collider ( ⼗卂億ドルの加速器の建設 )
and a couple million dollar detectors ( そして数百万ドルの検出器 ) CMS (Compact Muon Solenoid) ATLAS (A Toroidal ApparatuS)
reconstruction ( 再構築 ) • Reconstruct electrons, muons, photons from energy deposits • Reconstruct jets and tag b- jets with sophisticated algorithms • Use conversation of (transverse) energy to calculate the missing energy (MET) MET jet
jet reconstruction ( ジェット再構築 ) jet reconstruction algorithms group energy deposits together in different ways to form jets
b-jet reconstruction ( ビージェット識別 ) b-quarks have a longer lifetime than other elementary particles identify b-jets by reconstructing displaced vertices from tracks
choose your cuts H W b b l ν • Need events containing two b-jets, 1 lepton and MET • j 1 p T > 45 GeV; j 2 p T > 20 GeV, MV1c > 80% • l p T > 20 GeV; isolated, MET > 20 GeV
choose discriminating variable good discrimination poor discrimination B S B S The better the discriminating variable, the larger the separation between signal and background For the Higgs, a good variable is the mass
background ( 背景 ) • Background events are other events 700 Events / 25 GeV Data 2012 ATLAS VH(bb) ( =1.0) µ ∫ Diboson -1 s = 8 TeV Ldt = 20.3 fb 600 t t that look just like signal 1 lep., 2 jets, 2 Medium+Tight tags Single top V Multijet incl. p >120 GeV T W+hf 500 W+cl W+l Z+hf • Two types of background 400 Uncertainty Pre-fit background × VH(bb) 20 300 • Reducible 200 100 • Experimental: better isolation cut, 1.5 20 40 60 80 100 120 140 160 180 200 220 Data/Pred improved b-tagging algorithm 1 0.5 20 40 60 80 100 120 140 160 180 200 220 m [GeV] • Physics: different final state, e.g. bb top additional lepton, jets W+cl • Irreducible = same final state as signal W+bb • Often different kinematics or WZ need to apply kinematic cuts
background uncertainty ( 背景の不確実性 ) • Large uncertainties -> more difficult to extract the signal • Uncertainties can be both statistical and systematic • Decrease impact by either reducing background or reducing uncertainty: e.g. estimate in a control region
systematic uncertainties ( 系統誤差 ) Z +jets background background Zl normalisation, 3/2-jet ratio 5% Zcl 3/2-jet ratio 26% normalisation shape Z +hf 3/2-jet ratio 20% Z +hf/ Zbb ratio 12% ∆ φ (jet 1 , jet 2 ), p V T , m bb S W +jets Wl normalisation, 3/2-jet ratio 10% Wcl , W +hf 3/2-jet ratio 10% Wbl / Wbb ratio 35% Wbc / Wbb , Wcc / Wbb ratio 12% ∆ φ (jet 1 , jet 2 ), p V T , m bb S tt 3/2-jet ratio 20% High/low- p V T ratio 7.5% Top-quark p T , m bb , E miss S T Single top Cross section 4% ( s -, t -channel), 7% ( Wt ) Acceptance (generator) 3%–52% m bb , p b 1 S T signal Diboson Cross section and acceptance (scale) 3%–29% scale Cross section and acceptance (PDF) 2%–4% S m bb Multijet 0-, 2-lepton channels normalisation 100% 1-lepton channel normalisation 2%–60% Template variations, reweighting S
improving sensitivity: mass resolution ( 質量分解能 ) • The better the mass resolution, the Events / 4.0 GeV ATLAS Simulation Resolutions ( - )/ 0.1 σ σ σ GSC GSC Pythia VH, H b b MC → 16.4 GeV -- 2 lep., 2 jets, 2 b-tags 14.4 GeV 12% smaller the amount of background V p inclusive 0.08 14.1 GeV 14% T that needs to be considered 0.06 Global Sequential Calib. (GSC) + Muon-in-Jet Correction + Resolution Correction 0.04 • 14% improvement in resolution 0.02 0 0 20 40 60 80 100 120 140 160 180 200 m [GeV] bb
improving sensitivity: MVA ( 多変量解析 ) Events / 25 GeV Data 2012 ATLAS VH(bb) ( =1.0) µ 60 ∫ Diboson -1 s = 8 TeV Ldt = 20.3 fb t t 1 lep., 2 jets, 2 Tight tags Single top Events / 0.2 Events / 25 GeV Events / 20 GeV V 1000 900 Multijet 1800 50 160<p <200 GeV Data 2012 Data 2012 Data 2012 ATLAS ATLAS ATLAS VH(bb) ( =1.0) VH(bb) ( =1.0) VH(bb) ( =1.0) T W+hf µ µ µ ∫ -1 Diboson ∫ -1 Diboson ∫ -1 Diboson 800 s = 8 TeV Ldt = 20.3 fb s = 8 TeV Ldt = 20.3 fb s = 8 TeV Ldt = 20.3 fb Z+hf 1600 t t t t t t 0 lep., 2 jets, 2 tags 0 lep., 2 jets, 2 tags 1 lep., 2 jets, 2 tags Single top Single top Single top Uncertainty 800 V Multijet V Multijet V Multijet p >120 GeV p >120 GeV 700 p >120 GeV 1400 40 Pre-fit background T W+hf T W+hf T W+hf × W+cl W+cl W+cl VH(bb) 10 W+l W+l 600 W+l 1200 Z+hf Z+hf Z+hf 600 Z+cl Z+cl Uncertainty 500 1000 Z+l Z+l Pre-fit background 30 × Uncertainty Uncertainty VH(bb) 50 Pre-fit background Pre-fit background 400 800 VH(bb) × 10 VH(bb) × 50 400 600 300 20 200 400 200 200 100 10 0 50 100 150 200 250 300 350 400 450 500 0 50 100 150 200 250 300 350 400 450 500 2 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Data/Pred Data/Pred Data/Pred 1.5 1.5 1.5 1 1 1 0.5 20 40 60 80 100 120 140 160 180 200 220 0.5 0.5 0 Data/Pred 0 50 100 150 200 250 300 350 400 450 500 0 50 100 150 200 250 300 350 400 450 500 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 1.5 ∆ m [GeV] miss R(b ,b ) E [GeV] bb 2 (a) (b) T (c) 1 1 0.5 Events / 20 GeV 4500 Events / 20 GeV Events / 0.2 Data 2012 Data 2012 Data 2012 20 40 60 80 100 120 140 160 180 200 220 300 ATLAS ATLAS ATLAS VH(bb) ( =1.0) VH(bb) ( =1.0) VH(bb) ( =1.0) µ µ µ ∫ Diboson ∫ Diboson ∫ Diboson 4000 -1 -1 -1 s = 8 TeV Ldt = 20.3 fb s = 8 TeV Ldt = 20.3 fb 250 s = 8 TeV Ldt = 20.3 fb m [GeV] t t t t t t bb 1 lep., 2 jets, 2 tags 2 lep., 2 jets, 2 tags 2 lep., 2 jets, 2 tags Single top Single top Single top 250 3500 V Multijet V Z+hf V Z+hf p >120 GeV p >120 GeV p >120 GeV T W+hf T Z+cl T Z+cl 200 W+cl Z+l Z+l 3000 W+l Uncertainty Uncertainty 200 Z+hf Pre-fit background Pre-fit background Uncertainty VH(bb) × 60 VH(bb) × 60 2500 Pre-fit background 150 × VH(bb) 90 150 2000 100 1500 100 1000 50 50 500 Events / 0.14 Data 2012 0 50 100 150 200 250 300 350 400 450 500 50 100 150 200 250 300 350 400 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 450 Data/Pred Data/Pred Data/Pred 1.5 ATLAS 1.5 1.5 VH(bb) ( =1.0) µ 1 1 1 ∫ Diboson -1 s = 8 TeV Ldt = 20.3 fb 0.5 0.5 400 t t 0.5 1 lep., 2 jets, 2 Tight tags 0 50 100 150 200 250 300 350 400 450 500 50 100 150 200 250 300 350 400 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Single top p V [GeV] p (b ) [GeV] ∆ η V Multijet (V,bb) p >120 GeV (d) T (e) T 1 (f) 350 T W+hf Z+hf Uncertainty 300 Pre-fit background × VH(bb) 20 250 200 Multivariate techniques combine 150 100 information from kinematic distributions 50 2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Data/Pred into a single discriminating variable 1.5 1 0.5 0 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 BDT VH
Recommend
More recommend
