1 First 13TeV Higgs studies This γγ bumpRescent results on the Higgs & 750 GeV diphoton bump Bill Murray Warwick University / RAL, STFC 12th April 2016 Finishing 8 TeV Higgs First 13TeV Higgs studies This γγ bump
2 Introduction The Higgs boson was discovered back in July 2012Quick review of couplings status A few searches recently released Then results from 2015 running They benefit from the increase in Higgs σ at 13TeV With a factor 2 or 3 – ttH increases by 4.7 So 2015 data has little to say on decays Production cross-sections are priority
3 Background LHC is all about background rejectionEvery event at a lepton collider is physics; every event at a hadron collider is background Sam Ting 1010
4 ATLAS Triggers 2012 Single muon pT>24 GeV llll, lνlν, τlτh, Wbb, Zbb Single electron llll, lνlν, τlτl, τlτh, Wbb, Zbb Muon pair pT > (13,13) GeV llll, lνlν, Zbb Asymmetric Muon pair pT > (18,8) GeV llll, lνlν, τlτl Electron pair pT > (12,12) GeV llll, lνlν, τlτl, Zbb Electron + muon pT > (12,8) GeV τlτl Tau pair pT > (29,20) GeV τhτh Photon pair pT > (35,25) GeV γγ These lepton/photon trigger thresholds are set to retain EW scale objects. They present challenges in Run 2 to maintain them.
5 So what did Run 1 tell about the Higgs?
6 Higgs Mass Estimation ATLAS+CMS Results compatibleCombined mass: ±0.21(stat.)±0.11(syst.) Mass measured to ~0.2% Systematics well below statistics The last parameter of the SM!
7 Total cross-section at 8 TeVmummy
8 Higgs after run 1 Assume we found 1 particleThen we have 5 accesible production and decay modes Combine CMS and ATLAS to measure these rates ttH slightly high? Or Brbb slightly low? correlated
9 Interpreting couplingsWe want to test whether what we have is the Higgs boson This is discussed here: The LHC measures σH*Br(H→ττ) σH is proportional to gHWW2 Br(H→ττ) is proportional to gHWW2 /ΣigHii2 Needs total width LHC cannot directly measure the total width There are unmeasurable decays like H→gluons Some assumption is needed e.g. gHWW
10 Coupling studies Fit with only κV and κfAllows multiple different modes to be compared and contributions assessed All channels are compatible with standard model expectations in fermion and vector coupling The relative sign of the fermion and boson couplings affects e.g. the photon decay loop Opposite signs excluded at 5sigma
11 More general couplingsATLAS-CONF , CMS-PAG-HIG Consistency of the Higgs decays in 8 parameter fit, with: κV constrained < 1 Or BrBSM=0 It is impressive how insensitive fit is to this Upper limit on BSM decay is 95% CL
12 arXiv: Higgs width to W The Higgs decay width to off-shell dibosons is to LO given by: For the case of one on shell and one off shell this become approximately proportional to one power of the width. Thus the Br H→WW is proportional to the W boson width This is currently known to 2% Not totally negligible in analysing Higgs width Djouadi's Anatomy Γ 𝐻 0 → 𝑉 ∗ 𝑉 ∗ = 1 π 𝑀 𝐻 𝑑𝑞 1 2 𝑀 𝑉 Γ 𝑉 𝑞 1 2 − 𝑀 𝑉 𝑀 𝑉 2 Γ 𝑉 𝑀 𝐻 0 − 𝑄 𝑑𝑞 2 2 𝑀 𝑉 Γ 𝑉 𝑞 2 2 − 𝑀 𝑉 𝑀 𝑉 2 Γ 𝑉 2 Γ 0
13 ΓWW/ZZ(q1,q2) for mH=125.09
14 How to interpret? Now calculate ΓWW/ΓZZ=BR(WW)/BR(ZZ)Find BR/BR|lo=7.99 (c/f 8.07 in YR 3) Data ratio is in the LHC CONF on couplings. Extract This can be compared with 2.085±0.042 PDG Ten times worse – but please remember ΓWW is gHWW* ΓW 𝐵𝑅 𝑊𝑊 𝐵𝑅 𝑍𝑍 = 6.8 − Γ 𝑊 = 1.8 − 𝐺𝑒𝑉
15 Focus on invisible Direct and indirect constraints on invisible higgs are independent Combine for best sensitivity Abding visible decays moves BR limit from 25% to 23% Direct search dominates Arguably less model dependent Brs measured in data
16 Higgs invisible v Dark MatterInterpret dark matter in a 'Higgs portal' model Higgs only SM paricle coupled to DM The Spin Independent is very close to this Strong constraints for mχ
17 Higgs Width studies Direct measurement of the peak lineshapeLimited by experimental resolution CMS set 3.4 GeV upper limit from llll mode Extract from peak position Interference with background moves γγ peak c/f llll Or even use high/low pT difference in γγ No results yet Use high-mass tail of BW in llll (& interference) High-mass cross-section stable; take ratio to peak Assumes line-shape is not distorted by new physics 22 MeV limit Use combined invisible, undetected cross-section Assumes relations between couplings 6 MeV upper limit
18 Forbidden decays?
19 So what did we get?
20 MSSM constraints MSSM Higgs sector is defined by mA and tanβ at tree level But radiative corrections are important Previous MSSM benchmark scenarios (e.g. mHmax) fix all other parameters Do not match mh=125 One new approach is to absorb radiative corrections into a single parameter used to fit mh at each point Only approximately true Deduce mA>400 GeV Within assumptions 𝑚 ℎ 0 , 𝐻 0 2 = 𝑚 𝐴 𝑚 𝑍 2 ∓ 𝑚 𝐴 𝑚 𝑍 −4 𝑚 𝑍 2 𝑚 𝐴 cos 2 2β 𝑚 ℎ 0 < 𝑚 𝑍 2 cos 2 2β ATLAS-CONF
21 Higgs lepton flavour violationH→μτ from CMS 0/1/2 jets x τe/τh The most powerful is 0 jets x τe Shown right 4/6 channels have small excesses 2.5σ excess Best fit Br is 0.84±0.38%
22 ATLAS LFV H→μτh only Divided into two caregories of mT<,>40 GeVThey are combined in the plot right: Br is 0.77±0.62% Remember CMS found the most powerful is 0 jets x τe
23 H→τμ LFV Both ATLAS and CMS have excesses 2.5 sigma in CMS,H→ μτ lim. ATLAS CMS Expected Observed μτe n.a. 1.32/1.66/3.77% 2.04/2.38/3.84% μτh 1.24% 1.85% 2.34/2.07/2.31% 2.61/2.22/3.68% Combined n.a.1 0.75% 1.51% Both ATLAS and CMS have excesses 2.5 sigma in CMS, small in ATLAS Small, but very interesting
24 Out today: ATLAS H→μτe Plots were put out for Moriond Paper out todayArXiv: H→μτe two categories No jets is more powerful Shown right No sign of any signal I read approximate limit numbers off a plot... arXiv:
25 H→τμ LFV Both ATLAS and CMS have excesses 2.5 sigma in CMS,H→ μτ lim. ATLAS CMS Expected Observed μτe 1.73% 1.79% 1.32/1.66/3.77% 2.04/2.38/3.84% μτh 1.24% 1.85% 2.34/2.07/2.31% 2.61/2.22/3.68% Combined 1.01% 1.43% 0.75% 1.51% Both ATLAS and CMS have excesses 2.5 sigma in CMS, 1σ in ATLAS Perhaps less interesting than it was before?
26 Run 2? The increased beam energy increases cross-sections by afactor 2+ Factor 5 for ttH 100fb-1 in 3 years will give ~10 times the Higgs production Measurements becoming increasingly precise Testing the SM in a new sector With a very different structure from the W, Z bosons
27 Higgs to γγ at 13 TeV Different approaches:σfid=52±34(stat) (sys)±3(lumi) c/f theo. μ= in CMS
28 Higgs to llll at 13 TeV Fiducial cross section – different mll, mllll, pT, ΔR (c/f 2.74±0.28 theo) ATLAS (c/f 2.39±0.25 theo) CMS
29 ttH at 13 TeV ttH benefits has high mass systemCross-section gain by x4.7 at 13 TeV It was also on the edge of sensitivity in run 1 CMS search for multileptons from H→WW,ZZ or ττ combined Same sign ll (top) and trilepton (bot) Best fit is close to SM expected
30 ttH at 13 TeV: bb mode H to bb is highest rate - but large backgrounds. No sign of any excess here either
31 Higgs to invisible: VBFVBF was essential for the H→ττ discovery The high-mass forward jet pair gives an improved s/b Tagging the jet pair allows a search for the invisible Higgs decay Much higher cross-section than ZH But not as clean a tag
32 VBF H to invisible Jet pair mass > 1.0TeV (CMS, ATLAS main signal)Delta eta cut on tag jets Observed (expected) 0.28 (0.31) in ATLAS Observed (expected) 0.49 (0.65) in CMS
33 CMS VBF H to invisible @ 13 TeV!CMS-PAS-HIG CMS VBF H to 13 TeV! 8 regions of pT and jet pair mass used for VBF Limit set at 69% (62% expected) Brings CMS combined to 32% Compare ratios of accepted σ at 13 and 8 TeV Generally below PDFs – preliminary? 8 TeV 13 TeV ratio Z→νν 158 62 3.2 W→ lν 255 51 1.7 Top/qcd/VV 26 6 1.9 Total 439 119 2.2 ggH 23 5.4 VBF H 273 53.5
34 γγ
35 Dark Photons: ZdZd If target if pair productionb of Zd start from 4l search, but relax m12~ mZ Search mass spectrum for ZDZD modes 4 events with both pairs below 62.5 GeV Constraint of equal pair masses has just 2 events survive Br(H→ZdZd→llll)<3x10-4 15
36 h→aa→γγγγ A light nMSSM a might be produced in h→aaarXiv: v1 A light nMSSM a might be produced in h→aa With a→γγ a possible signature Select 3 photons pT>17 GeV for lowest Gives efficient signal reconstruction 4th photon likely soft Total 3γ rate sets limits Improve using m23 and vary ma Br (H→aa) * Br(a→γγ)2 below 10-3 Is it worth trying 4 photons?
37 h→aa→μμττ If ma>2mτ the τ decay opensAnalsis uses good μμ mass to identify peak μ pT>18 (1st) & 5-9 (2nd) Identify τ in e/μ/had modes pT>5-15 GeV 19 events observed, 20 expected Older results looked for 4-tau mode – no sign of signal
38 Combining a→μμ and a→ττCombination needs relative rate Here assume given by mass Upsilon region is covered by 4τ J/ψ and not covered μμbb mode is also searched for
39 Width via ZZ mass distributionUse observed lineshape Need to understand interference with gg→ZZ Assume K-factors same to 10% Take measured on-peak σH Using ZZ in llll and llνν modes ME in llll suppresses qq→ZZ ΓH<22MeV (both ATLAS and CMS) Both experiments 'lucky' This assumes line shape is SM-like ArXiv
40 Conclusions: no new physicsBSM couplings analyses H→BSM Br<34% ATLAS+CMS, κV<1 assumed Loops with virtual particles (gg→H,H→γγ) good to 10% H→Invisible Br<25% direct (23% in combination) Non-SM couplings of the H125 searched for: Br(H→Z(d)Zd→llll)<(3x)10-4 for 15
41 The surprise: γγ spectrumCuts a la H(125) pT>40%(30%) mγγ Peak at 750 GeV EOYE Dec 2015 Significance is 3.9σ For ΓX=45 GeV The Global siginificance is 2σ Across 200
42 ATLAS alternative: spin 2Spin-0 analysis is complemented by spin-2 pT thresholds lowered as graviton is more forward pT>55 GeV Nearly twice as many events Significance 3.6σ WJM estimate: You expect ~2.8 if noise This strengthens case
43 8 TeV comparison Surely we would have seen it at 8 TeV?Re-analyses 2012 using 2015 methodology Excess for spin-0 at 750 GeV, Γ/m=0.06 of 1.9σ Only 1.2σ tension for gg production Spin 2 analysis has 2.7σ tension assuming gg
44 OK, so what about CMS? Combining EBEB and EBEE gives 2.4σ for a narrow resonance in the spin-0 interpretation
45 Add the field-off data Adding spin-0 takes 2.4σ to 2.85σ for a narrow resonance in the spin-0 interpretation
46 What of CMS 8 TeV? For a narrow resonance 8 and 13 TeV combine to give 3.4σ For a broad one 8 TeV dilutes - marginally
47 Broad or narrow? ATLAS results favour a 45 GeV widthCMS have highest significance for narrow I have done some private toy studies, mocking up the ATLAS search. Use Breit-Wigner signal on an exponential background Generate toy MC, and fit to see what is found
48 Broad or narrow For example, use mX=750 GeV ΓX= 45 GeV,23 signal events expected If you fit using true m and Γ you see pink distribution, mean 22.8 events If you fit m, Γ to data mean is OK, but signal very spread
49 Broad or narrow Signal varies because width varies a lotRight: fraction of fitted Γ 30-60, or <5, as a function of true width With Nsig tuned to give ~3.5 sigma on average Narrow signal rarely appears wide. But a wide signal can appear narrow
50 Conclusions H125 is looking more and more like SM HiggsH→γγ seen: SM rate H→ZZ seen: SM rate H→WW seen: SM rate H→ττ seen: SM rate gg→ H seen: SM rate VBF VV→H seen: SM rate H→bb and ttH are 2-3σ measurements Show some tension No clear discrepancies in non-SM modes...but keep looking 750 GeV resonance seems rather unexpected Significance of order 2σ But two experients are possibly consistent: more data!
51 Virtual Higgs decays ATLAS-CONF , CMS-PAG-HIG Search for BSM Higgs particle by assuming all SM but allowing arbitrary strength on Higgs loops Despite early γγ final hits the SM nail Not a trace of new particles here 4th chiral fermion generations rarely considered now
52 CMS ttH→bb at 13TeV