High Mass Standard Model Higgs searches at the Tevatron

We present the results of searches for the Standard Model Higgs boson decaying predominantly to WW pairs, at a center-of-mass energy of sqrt(s)=1.96 TeV, using up to 8.2 fb^{-1} of data collected with the CDF and D0 detectors at the Fermilab Tevatron collider. The analysis techniques and the various channels considered are discussed. These searches result in exclusions across the Higgs mass range of 156.5<mH<173.7 GeV for CDF and 161<mH<170 GeV for D0.


Introduction
The search for the mechanism of electroweak symmetry breaking and a Standard Model (SM) Higgs boson has been a major goal of particle physics for many years. Within the Higgs sector of the SM, the mass of the Higgs boson (m H ) is a free parameter. Constraints on m H come from direct searches at the LEP experiments [1] which conclude that m H >114.4 GeV at 95% Confidence Level (CL) and more recently from combined searches of the ATLAS and CMS experiments using up to 2.3 fb −1 data yielding an exclusion of 141< m H <476 GeV at 95% CL [2]. Additionally, indirect constraints using precision electroweak measurements require m H <185 GeV at 95% CL [3].
Higgs searches at the Tevatron collider are a subject of intense study. By combining CDF and D0 results, the SM Higgs mass range between 158 and 175 GeV was excluded at 95% CL using up to 8.1 fb −1 of data [4].
These proceedings present the status of the searches by CDF and D0 for a SM Higgs boson decaying predominantly to a pair of W bosons and using up to 8.2 fb −1 of data collected until the summer of 2011. All limits will be given at 95% CL

Higgs boson production and decays
At the Tevatron, the dominant production mode is via the gluon fusion process, gg → H, with a cross section ranging between 840-190 fb for a Higgs mass between 130-200 GeV. The associated production, qq → VH (V = W, Z), and vector boson fusion, qq → Hqq, processes with cross sections ranging between 180-32 fb and 57-22 fb respectively for the aforementioned Higgs mass range, are also considered in order to maximise the sensitivity of these searches.
For a Higgs mass m H >135 GeV, the main decay mode is to a pair of W bosons while for m H <135 GeV Higgs decays mainly to a pair of b-quarks. This distinction is what defines the high mass and low mass Higgs searches at the Tevatron. a e-mail: kap01@fnal.gov

Search channels
The high mass searches at CDF and D0 require at least one electron or a muon in the final state in order to suppress the QCD background. Given this requirement all possible decay modes are considered to maximise the signal acceptance. The di-lepton+missing transvere energy (E miss T ) channel requires two electrons or muons (plus neutrinos) of opposite charge in the final state. This represents a small WW decay branching ratio, ≈ 6% (including τ → e, µ decays), but a clean signature offering the highest sensitivity of all the high mass channels. Decays of one of the W bosons to a τ lepton and the other to an electron or muon, with a subsequent hadronic decay of the τ (τ h ) are also considered offering an additional branching ratio of ≈ 4%. The lepton+jets channel requires one W boson to decay hadronically and the other leptonically. This represents a signficant WW decay branching ratio, ≈ 30%, however suffers from a large W+jets background. Dedicated qq → VH searches are also performed by looking for the SM suppressed signature of at least two leptons (electrons, muons) with one same charge pair originating from leptonic decays of the three vector bosons in the final state.

Search strategy
The high mass analyses are each split into categories based on the reconstructed lepton flavour, quality or jet multiplicity. This is done in order to take advantage of the differences in the detector response between the various lepton types, and between the kinematics of the signal production mechanisms and the background processes.
As the signal final states contain neutrinos, selections based on large E miss T are used which account for energy mis-measurements in the detectors. D0 employs Multivariate (MVA) techniques at this stage to reduce low E miss T backgrounds while maximising signal acceptance, by taking advantage of the correlations between the input variables. Selecting events with a high MVA response score, improves the sensitivity by up to 30% compared to conventional square selections. Figure 1 shows the E miss  Even after all selections, the signal/background (S/B) ratio is still in the range of 0.1-1% for m H =165 GeV. This requires the use of MVA discriminants, either boosted decision trees or artificial neural networks, to provide further discriminating power by using their response to derive limits on the Higgs boson yield for the cases where no signallike excess is observed.

Di-lepton+E
The signature of this channel is two high p T , oppositely charged, isolated leptons and E miss T . A small angular separation is also expected due to the spin-0 nature of the Higgs boson, which in turn gives rise to the spin-correlation between the final-state leptons. In contrast, leptons from the irreducible Z → WW background are predominantly backto-back. Figure 2 shows the ∆R = ∆φ 2 + ∆η 2 distribution for a di-lepton+E miss T data sample with zero reconstructed jets. Full use of this topological distinction is made by using variables such as ∆R and ∆φ as inputs to the MVA discriminants. CDF also makes use of likelihood ratios constructed from matrix-element probabilities as input variables to the MVA discriminants. Such a likelihood ratio is shown on the right plot of Figure 2.
This channel is further split according to jet multiplicity into 0, 1 or more than 1 reconstructed jets in the final state. This enables the training of the MVA discriminant to focus on the different signal and background compositions in each of the jet multiplicity categories, such as WW for the 0 jet, Z + jets for the 1 jet and tt for the 2 jet where b-tagging information is used to suppress this background. The statistical analysis of the MVA final discriminant does not exhibit any excess with respect to the background expectations and limits are set as shown in Figure 3. It should be noted that the CDF limit was obtained including the same sign di-lepton and tri-lepton searches described in sections 4.2 and 4.4.

Di-lepton+E miss
T channel with a τ h Additional signal acceptance can be obtained by considering hadronically decaying τ leptons from W decays. Muon (electron)+τ h channels are a new addition to the D0 high mass program. In contrast to CDF, D0 splits these channels into ≤1 or >1 jets due to the different signal production mechanisms and background compositions. The ≤1 jet category does not include an electron+τ h final state due to the very large backgrounds and the >1 jet is also part of the low mass Higgs searches.
Both CDF and D0 see no excesses with respect to the background expectation in the statistical analysis of the respective MVA discriminants and therefore set limits. Using 8.2 fb −1 of data CDF obtains an observed (expected) limit for m H =165 GeV of σ 95 /σ SM =17.5(13.0). For the muon +τ h +E miss T +≤1 jet and using 7.3 fb −1 D0 obtains a corresponding observed (expected) limit of σ 95 /σ SM =6.6(7.8).

Lepton+jets channel
The signature of this channel is one isolated high p T electron or muon, high E miss T and two high p T jets [5]. For m H ≥ 160 GeV, the Higgs decays to two on-shell W bosons, thus offering the ability to reconstruct the kinematics of the full event including the longitudinal momentum of the neutrino, up to a two-fold ambiguity. Although this signature suffers from large W+jets backgrounds, the large branching ratio of hadronic W boson decays provide 52 expected signal events surviving all selections in 5.4 fb −1 of D0 data. The statistical analysis of the MVA discriminant does not exhibit any excess with respect to the background expectation and limits are set. Using 5.4 fb −1 of data, D0 obtains an observed (expected) limit for m H =165 GeV of σ 95 /σ SM =5.2 (5.1).

Same charge di-leptons and tri-leptons
In these channels, the defining characteristic is the presence of at least two isolated, high p T electrons or muons, which can form a pair of the same charge, and high E miss T . Charge misidentification can lead to a significant migration of opposite sign charge backgrounds into the same sign region. Therefore high quality tracking criteria are also required to suppress this instrumental background. CDF also requires the presence of at least 1 jet in the same charge di-lepton final state, since the decay of the third boson will most likely result in the production of an additional jet.
In contrast to CDF, D0 does not include a dedicated tri-lepton search. Tri-lepton events can occur naturally in WH → WWW events with all Ws decaying leptonically, or in ZH → ZWW where the Z and one of the Ws decay leptonically and the other hadronically. This distinction allows the separation of tri-lepton events into cases where the same flavour opposite sign leptons, form an invariant mass compatible to the Z mass or not. If a mass compatible to the Z mass is found, the probability that one of the Ws decayed hadronically, allows this channel to be further split by requiring one or more than one reconstructed jets in the event. If more than one jets are found then the event is fully reconstructed and the mass of the Higgs can be determined. CDF ensures orthogonality between the same charge di-lepton and tri-lepton search by vetoing the presence of a third lepton in the di-lepton case. As D0 has no dedicated tri-lepton search, no such veto is required. Figure 4 shows the E miss T distributions in same charge di-lepton and tri-lepton events after all selections.
No excess was observed in the statistical analysis of same charge di-lepton and tri-lepton MVA discriminants, with respect to the background expectation, and thus limits are set. The same charge di-lepton and tri-lepton limits of CDF are combined with the the di-lepton+E miss T searches of Section 4.1. The most sensitive channel is the same charge di-lepton + E miss T +> 1 jets which gives an observed (expected) limit for m H =165 GeV of σ 95 /σ SM =3.69

Systematic uncertainties
The systematic uncertainties considered account for detector resolutions, reconstruction efficiencies, background normalizations and shapes, both theoretical and data-driven. Where appropriate, these uncertainties are correlated between CDF and D0. Careful consideration has also been taken to account for the theoretical uncertainties related to the various Higgs production mechanisms and particularly for gg → H, which is the main signal in the most sensitive di-lepton+E miss T channel.
The Higgs boson signal is normalized to the most recent highest-order calculations available for all production mechanisms considered. The gluon fusion cross section, σ(gg → H), is calculated to NNLO in QCD with soft gluon resummation to NNLL [7] [8] and uses the MSTW 2008 NNLO PDF set as it is the only NNLO set which results from a global fit to all relevant data [9]. Since the di-lepton +E miss T analyses are split in categories depending on the number of reconstructed jets, the QCD scale uncertainties on σ(gg → H) are estimated following the prescreption described in [10]. By propagating the uncorrelated un-  certainties of the NNLL inclusive [7] [8], NLO one or more jets [12] and NLO two or more jets [13] rate to the exclusive gg → H + 0jet, 1jet and > 1jet rate, the error matrix shown in Table 1 is built. The PDF uncertainty on σ(gg → H), obtained using the prescription in [12] [11], is summarised in Table 2.

Results
CDF and D0 set limits by combining all of the SM high mass channels with up to 8.2 fb −1 of integrated luminosity. CDF excludes the range 156.5< m H <173.7 GeV (157< m H <172.2 GeV expected) and D0 excludes the range 161< m H <170 GeV (159< m H <170 GeV expected). In order to further improve the sensitivity of these searches, CDF and D0 combined their individual high mass channels in the Tevatron high mass combination, which is discussed elsewhere in these proceedings [15] [14].