Combination of Higgs boson searches at ATLAS Second MCTP Spring Symposium on Higgs Boson Physics, University of Michigan 16 th -20 th April 2012 Gemma Wooden, University of Michigan On behalf of the ATLAS collaboration
Overview Search for the Higgs boson in multiple channels Different channels contribute in different m H regions Combine searches to maximise sensitivity VBF ggf VH Search for Higgs production via different mechanisms: Gluon-gluon fusion (ggf) Vector boson fusion (VBF) Associated production (VH) 19/04/2012 2
Search channels All channels use the full dataset collected by the ATLAS detector in 2011 H->gg, H->ZZ->llll as well as H->tt, VH(H->bb) important at low mass 3
Limit setting procedure Method: Likelihood Test signal hypotheses m = signal strength parameter No. of signal events Estimate backgrounds using MC and/or data control regions No. of background events Nuisance parameters q Account for systematic uncertainties (q) using auxiliary measurements (e.g. performance groups) Set limits using modified frequentist method (CL s ) 95% C.L. on m found by adjusting m until CL s = 0.05: p m 1 - p b To quantify the significance of an excess, the local p-value can be calculated using same method but with m set to 0 19/04/2012 4
Combination of channels Must ensure that systematics are handled correctly Some will be correlated between channels q gg q 4l q lnln 19/04/2012 Gemma Wooden, University of Michigan, MCTP Higgs Symposium 5
H->gg Use 9 independent categories of varying sensitivity Categories depend on h of photons, whether the photons are (un)converted and the momentum component of the diphoton system transverse to the thrust axis Fit m gg to search for presence of a signal Mass resolution of around 1.7% for m H = 120 GeV 19/04/2012 6
H->ZZ->llll m 4l used as the discriminating variable Clean signature with low background rate Irreducible ZZ* background estimated using MC Z+jets and ttbar backgrounds estimated using data control regions Mass resolution of 1.5%(2%) in 4 muon(electron) channel at m H = 120 GeV 19/04/2012 7
H->ZZ->llnn Use m T as discriminating variable Apply different selections in low (m H < 280 GeV) & high (m H >= 280 GeV) regions Dominant backgrounds (ZZ, WW, WZ, ttbar) normalised using MC Require m ll m Z < 15 GeV to avoid overlap with the H->WW->lnln analysis 8
H->ZZ->llqq Use m llqq as discriminating variable Separate, optimised analyses above and below m H = 300 GeV Normalise dominant Z+jets background using data Divide analysis into tagged and untagged categories: events with two and less than two b-tagged jets respectively Analysis is extendable to lower masses 9
H->ZZ Combination of H->ZZ->llll, H->ZZ->llnn, H->ZZ->llqq 10
H->WW->lnln Use a fit to the m T distribution Split analysis into 0, 1 and at least 2 jet categories Use entirely data-driven method to estimate W+jets background Normalise top, WW and Z+jets backgrounds with data Separate analysis into low, intermediate and high mass regions 19/04/2012 11
H->WW->lnqq Fit the m WW distribution to search for Higgs signal M(en) is constrained to the W mass Mass resolution is quite good in this channel (~9% at m H = 400 GeV) Divide search into events with 0 or 1 jets in addition to the two arising from the W decay Backgrounds parametrised using forms motivated by MC studies 12
H->tt Use three different search channels: H->tt->ll4n, which uses m eff in 0j channel and collinear approx. in 1 and 2j channels H->tt->lt had 3n, which uses Missing Mass Calculator to calculate di-tau mass H->tt->t had t had 2n + jet, which uses collinear approx. to calculate the mass 13
VH(H->bb) Search for the Higgs boson in 3 channels: ZH->llbb WH->lnbb ZH->nnbb Use leptonic vector boson decays to trigger on and to reduce backgrounds m bb is used as the discriminating variable 19/04/2012 14
Systematic uncertainties All uncertainties are considered as either 100% correlated or 100% uncorrelated Typical theoretical uncertainties Detector uncertainties 15
Combination results limits Expect to exclude the Standard Model Higgs boson at 95% C.L. between: 120 < m H < 555 GeV Observed exclusion at 95% C.L.: 110 < m H < 117.5, 118.5 < m H < 122.5, 129 < m H < 539 GeV Observed exclusion at 99% C.L.: 130 < m H < 486 GeV Observe an excess in the low mass region 16
Combination results limits (ii) 17
Combination results local p 0 Investigate local p 0 the probability for the background to fluctuate and give an excess of events as large or larger than that observed Observed local p 0 is most consistent with the SM Higgs expectation at m H = 126 GeV Observed local significance 2.5σ (expected 2.9σ) 18
A closer look at the local p 0 19
Combination results signal strength Calculate best fit signal strength parameter, m = s/s SM m is most consistent with the SM Higgs production cross section at a mass of 126 GeV At m H = 126 GeV, the best-fit signal strength is 0.9+0.4-0.3 Probability of a fluctuation of this magnitude occurring across full range of m H (110 < m H < 600 GeV) is 30% or 10% in the mass range from 110 < m H < 146 GeV 20
Conclusions Searches for the Higgs boson have been undertaken in a wide range of channels using the full 2011 dataset of up to 4.9 fb -1 Stringent limits have been placed on the allowed mass of the Standard Model Higgs boson Allowed regions: 117.5 < m H < 118.5 GeV, or 122.5 < m H < 129 GeV Excess most compatible with a Higgs boson with a mass of 126 GeV Observe a local significance at m H = 126 GeV of 2.5s Expected significance in the presence of a SM Higgs boson is 2.9s However, more data is needed to arrive at a definite conclusion Already looking forward to analysing the new 8 TeV 2012 data, which is coming in fast 21
Backup
References Limit setting and combination procedure: ATL-PHYS-PUB-2011-011 Higgs combination using 1.04-4.9 fb -1 : arxiv:1202.1408 Higgs combination using full 2011 dataset: ATLAS-CONF-2012-019 H -> gg: arxiv:1202.1414 H -> ZZ -> llll: arxiv:1202.1415 H -> ZZ -> llnn: ATLAS-CONF-2012-016 H -> ZZ -> llqq: ATLAS-CONF-2012-017 H -> WW ->lnln: ATLAS-CONF-2012-012 H -> WW -> lnqq: ATLAS-CONF-2012-018 H -> tt: ATLAS-CONF-2012-014 VH(H -> bb): ATLAS-CONF-2012-015 23