Search for the Associated Production of a Higgs Boson and a Top Quark Pair in Multilepton Final States with the ATLAS Detector

ATLAS-CONF-2016-058

3 August 2016

These preliminary results are superseded by the following paper:

HIGG-2017-02
ATLAS recommends to use the results from the paper.

Content

Preview

Main document (CDS record), Physics Briefing - internal

pdf from CDS

Figures Tables

Abstract

The Yukawa coupling of the Higgs boson to the top quark is a key parameter of the Standard Model, and can be constrained using the associated production process $pp \to t\bar t H + X$. A search for this process using final states with multiple leptons, primarily targeting the decays $H\to WW^*$ and $H\to \tau\tau$, has been performed using 13.2 fb$^{-1}$ of data recorded by the ATLAS detector in 2015 and 2016 at a center of mass energy $\sqrt{s} =$ 13 TeV. The best-fit value of the ratio of observed and Standard Model cross sections is 2.5 $\pm$ 0.7 (stat) $^{+1.1}_{-0.9}$ (syst), and an upper limit on this ratio of 4.9 (2.3 expected) is found at 95% confidence level.

Figures

Figure 01a:
Invariant mass of leptons 0 and 1 for the a) tight and b) loose ttZ validation regions. The leptons are labeled in the same way as for the 3ℓ signal region. Events away from the Z peak are those satisfying the Z selection with leptons 0 and 2. Non-prompt lepton backgrounds are estimated using data as described in Section 6.2.

png (86kB) eps (2MB) pdf (39kB)

Figure 01b:
Invariant mass of leptons 0 and 1 for the a) tight and b) loose ttZ validation regions. The leptons are labeled in the same way as for the 3ℓ signal region. Events away from the Z peak are those satisfying the Z selection with leptons 0 and 2. Non-prompt lepton backgrounds are estimated using data as described in Section 6.2.

png (87kB) eps (2MB) pdf (40kB)

Figure 02:
Jet multiplicity in the WZ + 1 b-tag validation region. Non-prompt lepton backgrounds are estimated using data as described in Section 6.2.

png (93kB) eps (23kB) pdf (7kB)

Figure 03a:
Lepton flavor composition (a) and number of jets (b) for events in the ttW validation region. Non-prompt lepton and charge misreconstruction backgrounds (indicated as ``QMisReco") are estimated using data as described in Section 6.2.

png (83kB) eps (18kB) pdf (6kB)

Figure 03b:
Lepton flavor composition (a) and number of jets (b) for events in the ttW validation region. Non-prompt lepton and charge misreconstruction backgrounds (indicated as ``QMisReco") are estimated using data as described in Section 6.2.

png (79kB) eps (18kB) pdf (6kB)

Figure 04:
Pre-fit background and signal predictions and observed data yields for each signal region. The ttH prediction corresponds to the SM expectation (μ_ttH = 1). Charge misreconstruction backgrounds are indicated as ``QMisReco.''

png (94kB) eps (20MB) pdf (86kB)

Figure 05:
Best fit values of the ttH signal strength μ_ttH by final state category and combined. The SM prediction is μ_ttH=1. For the 4ℓ category, as zero events are observed, a 68% CL_s upper limit is shown instead.

png (42kB) eps (2MB) pdf (84kB)

Figure 06:
Upper limits on the ttH signal strength μ_ttH at 95% CL by final state category and combined. The SM prediction is μ_ttH=1. The median upper limit that would be set in the presence of a SM ttH signal (μ=1) is also shown.

png (32kB) eps (2MB) pdf (83kB)

Figure 07a:
Characteristics of events in the 2ℓ0τ_had signal region: (a) lepton flavor composition; (b) 10× the number of b-tagged jets plus the total number of jets. The signal is set to the SM expectation (μ_ttH = 1) and the background expectation is pre-fit (using initial values of the background systematic uncertainty nuisance parameters). The hatched region shows the total uncertainty on the background plus SM signal prediction in each bin. Charge misreconstruction backgrounds are indicated as ``QMisReco.''

png (78kB) eps (2MB) pdf (35kB)

Figure 07b:
Characteristics of events in the 2ℓ0τ_had signal region: (a) lepton flavor composition; (b) 10× the number of b-tagged jets plus the total number of jets. The signal is set to the SM expectation (μ_ttH = 1) and the background expectation is pre-fit (using initial values of the background systematic uncertainty nuisance parameters). The hatched region shows the total uncertainty on the background plus SM signal prediction in each bin. Charge misreconstruction backgrounds are indicated as ``QMisReco.''

png (85kB) eps (2MB) pdf (37kB)

Figure 08a:
Characteristics of events in the 2ℓ1τ_had signal region: (a) lepton flavor composition; (b) 10× the number of b-tagged jets plus the total number of jets. The signal is set to the SM expectation (μ_ttH = 1) and the background expectation is pre-fit (using initial values of the background systematic uncertainty nuisance parameters). The hatched region shows the total uncertainty on the background plus SM signal prediction in each bin. Charge misreconstruction backgrounds are indicated as ``QMisReco.''

png (72kB) eps (2MB) pdf (35kB)

Figure 08b:
Characteristics of events in the 2ℓ1τ_had signal region: (a) lepton flavor composition; (b) 10× the number of b-tagged jets plus the total number of jets. The signal is set to the SM expectation (μ_ttH = 1) and the background expectation is pre-fit (using initial values of the background systematic uncertainty nuisance parameters). The hatched region shows the total uncertainty on the background plus SM signal prediction in each bin. Charge misreconstruction backgrounds are indicated as ``QMisReco.''

png (72kB) eps (2MB) pdf (36kB)

Figure 09a:
Characteristics of events in the 3ℓ signal region: (a) lepton flavor composition; (b) 10× the number of b-tagged jets plus the total number of jets. The signal is set to the SM expectation (μ_ttH = 1) and the background expectation is pre-fit (using initial values of the background systematic uncertainty nuisance parameters). The hatched region shows the total uncertainty on the background plus SM signal prediction in each bin.

png (79kB) eps (1MB) pdf (19kB)

Figure 09b:
Characteristics of events in the 3ℓ signal region: (a) lepton flavor composition; (b) 10× the number of b-tagged jets plus the total number of jets. The signal is set to the SM expectation (μ_ttH = 1) and the background expectation is pre-fit (using initial values of the background systematic uncertainty nuisance parameters). The hatched region shows the total uncertainty on the background plus SM signal prediction in each bin.

png (80kB) eps (1MB) pdf (20kB)

Figure 10:
Post-fit background and signal predictions and observed data yields for each signal region. Background expectations have been updated to reflect the values of systematic uncertainty nuisance parameters after the fit to data. The ttH prediction corresponds to the best-fit value (μ_ttH = 2.5 ^+1.3_-1.1) and the displayed total uncertainties reflect the uncertainty in ttH as well as the backgrounds. Charge misreconstruction backgrounds are indicated as ``QMisReco.''

png (90kB) eps (20MB) pdf (86kB)

Figure 11:
Effect of the fifteen most important systematic uncertainty nuisance parameters θ on the signal strength μ=μ_ttH and constraints on the nuisance parameters from the fit. The blue and cyan bars show the ± 1σ impact of the nuisance parameter on the signal strength (shown on the top axis). The points and associated error bars show the best-fit values of the nuisance parameters and post-fit uncertainties on the nuisance parameters (shown on the bottom axis). The open bars show the effect of the systematic uncertainties on μ_ttH before the fit, and the solid bars show the effect after the fit. The nuisance parameters are initially normalized to 0 ± 1. The dotted vertical lines show ± 1 σ excursions of the nuisance parameters from their initial values.

png (99kB) eps (61kB) pdf (18kB)

Figure 12:
Event display for a candidate ee event in the 2ℓ0τ_had category. The blue tracks are the two selected electrons. Green and yellow bars indicate energy deposits in the electromagnetic (liquid argon) and hadronic (tile) calorimeters, respectively. In the inset display, the three azure cones are b-tagged jets and the yellow cones are the six non-b-tagged jets.

png (2MB)

Figure 13:
Event display for a candidate eμτ_had event in the 2ℓ1τ_had category. The blue track is the selected electron; the red track is the selected muon; and the white cone is the τ_had candidate. The azure cone is the selected b-tagged jet, and the three yellow cones are the non-b-tagged jets. Green and yellow bars indicate energy deposits in the electromagnetic (liquid argon) and hadronic (tile) calorimeters, respectively.

png (2MB)

Figure 14:
Event display for a candidate 3μ event in the 3ℓ category. The red tracks are the selected muons. The two blue cones are the selected b-tagged jets, and the five yellow cones are the non-b-tagged jets. Green and yellow bars indicate energy deposits in the electromagnetic (liquid argon) and hadronic (tile) calorimeters, respectively.

png (720kB)

Figure 15a:
Characteristics of events in the 2ℓ0τ_had signal region, by flavor category. The variable plotted is 10× the number of b-tagged jets plus the total number of jets for (a) ee events, (b) eμ events, (c) μμ events. The signal is set to the SM expectation (μ_ttH = 1) and the background expectation is pre-fit (using initial values of the background systematic uncertainty nuisance parameters). The hatched region shows the total uncertainty on the background plus SM signal prediction in each bin. Charge misreconstruction backgrounds are indicated as ``QMisReco.''

png (80kB) eps (2MB) pdf (36kB)

Figure 15b:
Characteristics of events in the 2ℓ0τ_had signal region, by flavor category. The variable plotted is 10× the number of b-tagged jets plus the total number of jets for (a) ee events, (b) eμ events, (c) μμ events. The signal is set to the SM expectation (μ_ttH = 1) and the background expectation is pre-fit (using initial values of the background systematic uncertainty nuisance parameters). The hatched region shows the total uncertainty on the background plus SM signal prediction in each bin. Charge misreconstruction backgrounds are indicated as ``QMisReco.''

png (83kB) eps (2MB) pdf (37kB)

Figure 15c:
Characteristics of events in the 2ℓ0τ_had signal region, by flavor category. The variable plotted is 10× the number of b-tagged jets plus the total number of jets for (a) ee events, (b) eμ events, (c) μμ events. The signal is set to the SM expectation (μ_ttH = 1) and the background expectation is pre-fit (using initial values of the background systematic uncertainty nuisance parameters). The hatched region shows the total uncertainty on the background plus SM signal prediction in each bin. Charge misreconstruction backgrounds are indicated as ``QMisReco.''

png (82kB) eps (2MB) pdf (37kB)

Figure 16a:
Additional characteristics of events in the 2ℓ1τ_had signal region: (a) number of jets; (b) number of tracks in the τ_had candidate. The signal is set to the SM expectation (μ_ttH = 1) and the background expectation is pre-fit (using initial values of the background systematic uncertainty nuisance parameters). The hatched region shows the total uncertainty on the background plus SM signal prediction in each bin. Charge misreconstruction backgrounds are indicated as ``QMisReco.''

png (71kB) eps (2MB) pdf (35kB)

Figure 16b:
Additional characteristics of events in the 2ℓ1τ_had signal region: (a) number of jets; (b) number of tracks in the τ_had candidate. The signal is set to the SM expectation (μ_ttH = 1) and the background expectation is pre-fit (using initial values of the background systematic uncertainty nuisance parameters). The hatched region shows the total uncertainty on the background plus SM signal prediction in each bin. Charge misreconstruction backgrounds are indicated as ``QMisReco.''

png (46kB) eps (2MB) pdf (35kB)

Figure 17:
Expected contribution to the background in each channel from various sources, using values of the background estimates before the fit. Charge misreconstruction backgrounds are indicated as ``QMisReco."

png (124kB) eps (20MB) pdf (89kB)

Tables

Table 01:
Configurations used for event generation of signal and background processes. If only one parton distribution function (PDF) is shown, the same one is used for both the matrix element (ME) and parton shower generators; if two are shown, the first is used for the matrix element calculation and the second for the parton shower. ``V" refers to production of an electroweak boson (W or Z/γ^*). ``Tune" refers to the underlying-event tune of the parton shower generator. ``MG5_aMC" refers to MadGraph5_aMC@NLO 2.2.1; ``Pythia 6" refers to version 6.427; ``Pythia 8" refers to version 8.2; ``Herwig++" refers to version 2.7. Samples using Pythia 6 and Pythia 8 have heavy flavor hadron decays modeled by EvtGen 1.2.0 [10]. All samples include leading-logarithm photon emission, either modeled by the parton shower generator or by PHOTOS [11].

png (34kB) pdf (62kB)

Table 02:
Tight and loose light lepton definitions. The lepton identification working points are documented in Refs. [44,45]. Selections for tight leptons are applied on top of the selections for loose leptons. ``99% eff" refers to isolation working points designed to be 99% efficient for isolated leptons at all p_T. (*) An additional ``gradient" isolation working point is defined with efficiency and fake/non-prompt lepton rejection intermediate between the loose and tight isolation selections.

png (18kB) pdf (56kB)

Table 03:
Fraction of the expected ttH signal arising from different Higgs boson decay modes in each analysis category and acceptance times efficiency (A×ε) for ttH signal in each category. The decays contributing to the ``other" column are dominantly H→μμ and H→bb. Rows may not add to 100% due to rounding. The acceptance times efficiency includes Higgs boson and top quark branching fractions, detector acceptance, and reconstruction and selection efficiency, and is computed relative to inclusive ttH production.

png (11kB) pdf (47kB)

Table 04:
Selections for the signal regions (SR) and validation regions (VR). The variable H_T,jets is the scalar sum of transverse momenta for the considered jets. Same-flavor, opposite-charge lepton pairs are referred to as SFOC pairs. Trigger-matched leptons correspond to an object reconstructed by the trigger, and must have p_T > 25 GeV (21 GeV for muons in 2015 data). In all regions at least one selected light lepton is required to be trigger-matched.

png (103kB) pdf (64kB)

Table 05:
Expected and observed event yields in validation regions (VR). The quoted uncertainties in the expectations include all systematic uncertainties. ``Purity" indicates the fraction of events in the VR expected to arise from the targeted process (ttZ for the first two VRs, WZ for the third, and ttW for the fourth).

png (9kB) pdf (32kB)

Table 06:
Summary of the effects of the systematic uncertainties on μ. Due to correlations between the different sources of uncertainties, the total systematic uncertainty can be different from the sum in quadrature of the individual sources. The impact of the systematic uncertainties is evaluated after the fit described in Section refsec:result.

png (25kB) pdf (43kB)

Table 07:
Expected and observed yields in the six signal region categories in 13.2 fb^-1 of data at √s = 13 TeV. Uncertainties in the background expectations due to systematic effects and MC statistics are shown. ``Other" backgrounds include tZ, tWZ, tHqb, tHW, tttt, ttWW, and triboson production. Values are obtained pre-fit, i.e., using the initial values of background systematic uncertainty nuisance parameters.

png (27kB) pdf (48kB)

Table 08:
Best-fit values of the signal strength μ_ttH and 95% CL upper limits obtained using the CL_s method. For the best-fit values, the first uncertainty is statistical and the second systematic. For the expected upper limits, the median is reported, and the uncertainties give the 68% expected range. The signal-injected upper limit is the median expected upper limit that is set by this procedure, testing against the μ_ttH=0 hypothesis, in the presence of a SM ttH signal. For the 4ℓ category, as zero events are observed, a 68% CL_s upper limit is shown in place of the best-fit value.

png (21kB) pdf (51kB)

Table 09:
Post-fit background, signal, and observed yields in the six signal region categories in 13.2 fb^-1 of data at √s = 13 TeV. Uncertainties in the background expectations due to systematic effects and MC statistics are shown. ``Other" backgrounds include tZ, tWZ, tHqb, tHW, tttt, ttWW, and triboson production. Background expectations have been updated to reflect the values of systematic uncertainty nuisance parameters after the fit to data. The prediction and uncertainties for ttH reflect the best-fit production rate of 2.5 ^+1.3_-1.1 times the Standard Model expectation. The uncertainty on the total background estimation is smaller than for the pre-fit values due to anticorrelations between the nuisance parameters obtained during the fit.

png (27kB) pdf (48kB)

© 2016 CERN for the benefit of the ATLAS Collaboration. Reproduction of the article, figures and tables on this page is allowed as specified in the CC-BY-4.0 license.
2024-05-19 01:23:29