Search for exotic resonances decaying into WZ/ZZ in pp collisions at sqrt(s) = 7 TeV

A search for new exotic particles decaying to the VZ final state is performed, where V is either a W or a Z boson decaying into two overlapping jets and the Z decays into a pair of electrons, muons or neutrinos. The analysis uses a data sample of pp collisions corresponding to an integrated luminosity of 5 inverse femtobarns collected by the CMS experiment at the LHC at sqrt(s) = 7 TeV in 2011. No significant excess is observed in the mass distribution of the VZ candidates compared with the background expectation from standard model processes. Model-dependent upper limits at the 95% confidence level are set on the product of the cross section times the branching fraction of hypothetical particles decaying to the VZ final state as a function of mass. Sequential standard model W' bosons with masses between 700 and 940 GeV are excluded. In the Randall-Sundrum model for graviton resonances with a coupling parameter of 0.05, masses between 750 and 880 GeV are also excluded.


Introduction
In many extensions of the standard model (SM) of particle physics [1][2][3] the spontaneous breaking of the electroweak (EW) symmetry [4][5][6][7][8][9] is associated with new strong dynamics appearing at the TeV scale. For instance, the origin of the new dynamics may be due to new interactions [10][11][12], compact extra dimensions [13,14], or a composite Higgs boson [15,16]. In such scenarios the SM is an effective low-energy theory, valid for energies smaller than a newphysics scale Λ. In these theories, one expects the existence of new resonances coupling to pairs of vector bosons (ZZ, WZ, and WW). A minimal ultraviolet completion of this effective theory for composite models is described in Ref. [17]. Other examples include Randall-Sundrum (RS) gravitons G KK [13,14] coupled to ZZ and WW, or technimesons [18,19] coupled to WZ. Limits from previous searches and from indirect bounds (e.g. in the EW sector and flavor physics) place the masses of these proposed RS resonances at or above the TeV scale [20][21][22][23][24][25][26]. These scenarios could be tested at the Large Hadron Collider (LHC), as long as Λ ∼ O(TeV), as suggested by the EW symmetry breaking scale. This analysis is sensitive to searches for resonances starting at 700 GeV and above. However, there are other theories that predict light resonances (e.g. low-scale technicolor) [18,19].
In this Letter we present a search for heavy resonances decaying to WZ and ZZ final states, with one boson being a Z decaying to leptons, namely Z → + − ( = µ, e) or νν, and the second boson decays to hadrons, V (V=W, Z) → qq. For heavy resonances the decay of each V produces a highly boosted system in which the two fermions are emitted within a small opening angle in the laboratory frame. The hadronization of the V→ qq quarks would then produce two partially overlapping jets reconstructed as a single jet with mass close to the V mass, a topology very different from that of a typical quark or gluon jet. Monte Carlo (MC) simulations suggest that more than ∼70% of the decays would produce a merged-jet topology for resonances heavier than ∼800 GeV. This feature is exploited in a VZ final state, to discriminate a possible signal from the SM background (mainly coming from Z+jets events).
Thus, in this study we consider three final states: one heavy jet and either Z → e + e − , Z → µ + µ − , or Z → E miss T , where E miss T is the characteristic signature of neutrino production. We characterize the signal as a peak in the invariant mass of the VZ system (transverse mass in the case of Z → νν decays). The search is performed with a data sample of pp collisions corresponding to an integrated luminosity of 5.0 fb −1 collected by the Compact Muon Solenoid (CMS) detector at the LHC at √ s = 7 TeV in 2011.
Results are presented in terms of two benchmark scenarios: i) the Sequential Standard Model (SSM) in which a new gauge boson W with the same couplings as the SM W boson decays to a WZ pair; ii) a RS graviton, G KK , decaying to ZZ. In both scenarios we search for resonances heavier than 700 GeV, where the considered boosted topology becomes relevant.  [23,33], G KK → γγ [24,34] and G KK → ZZ → jj [35, 36].

Collision Data and Monte Carlo Samples 2 The CMS Detector
Here a brief description of the CMS detector is given with an emphasis on the elements most relevant for this analysis. A more detailed description can be found elsewhere [37]. A cylindrical coordinate system about the beam axis is used, in which the polar angle θ is measured with respect to the counterclockwise beam direction and the azimuthal angle, φ, is measured in the x-y plane in radians, where the x axis points towards the center of the LHC ring. The quantity η is the pseudorapidity, defined as η = − ln[tan(θ/2)]. The layout comprises a superconducting solenoid providing a uniform magnetic field of 3.8 T. The bore of the solenoid is instrumented with various particle detection systems. The inner tracking system is composed of a pixel detector with three barrel layers at radii between 4.4 and 10.2 cm and a silicon strip tracker with 10 barrel detection layers extending outwards to a radius of 1.1 m. Each system is completed by two end-caps, extending the acceptance up to |η| < 2.5. A lead-tungstate crystal electromagnetic calorimeter with fine transverse (∆η, ∆φ) granularity and a brass/scintillator hadronic calorimeter surround the tracking volume and cover the region |η| < 3. CMS also has extensive forward calorimetry. The steel return yoke outside the solenoid is instrumented with gas-ionization detectors which are used to identify muons in the range |η| < 2.4. The barrel region is covered by drift tube chambers and the end cap region by cathode strip chambers, each complemented by resistive plate chambers.

Collision Data and Monte Carlo Samples
The preselection of the datasets for the analysis is different for the "dilepton" (VZ→ qq + − , = e, µ) and the "E miss T " (VZ→ qq νν) channels. For the dilepton channels, we consider events that were recorded with double-electron or single-muon triggers. The trigger thresholds changed with time, as a consequence of the increasing peak luminosity and the changes in running conditions. The tightest thresholds used in the trigger (i.e. 40 GeV for the single-muon trigger and 17 GeV for the dielectron trigger) are looser than the corresponding offline analysis requirements. Typical trigger efficiencies exceed 83% (95%) for the electron (muon) triggers. For the E miss T channel, we use triggers requiring at least one calorimetric jet and missing transverse energy. These triggers have efficiencies of more than 99% for events with a leading jet of transverse momentum p T > 160 GeV and E miss T > 300 GeV after offline reconstruction and corrections, which allows resonances heavier than 1000 GeV to be probed with an efficiency above 20%. We use MC samples to study the signal and background. We consider the SM background processes that could contribute with two leptons and a (massive) jet in the final state. The summary of the signal samples is given in Table 1, and the background samples, in Table 2. The PYTHIA 6.424 [38] leading-order (LO) generator with tune Z2 [39] is used to generate the signal events and simulate the parton showering, with a full simulation of the detector based on GEANT4 9.4 package. Mass-dependent K factors are applied. For the G KK analysis, nextto-leading order (NLO) corrections are calculated using the "two cutoff phase space slicing" method [40,41] in the diphoton final state. For the W analysis, the next-to-next-to-leading order (NNLO) corrections are calculated with FEWZ [42] in the leptonic final state. These K factors are used for lack of better (N)NLO calculations for the final states considered. The background samples are generated using the MADGRAPH 5.1.1.0 matrix-element generator [43,44], while PYTHIA is used for the parton showering and hadronization, with the same version and tuning as for signal samples. The parton distribution function (PDF) used is CTEQ6L1 [45]. Jets are matched to partons using the MLM scheme [46].

Reconstruction and Event Selection
Events are required to have at least one primary vertex of good quality, where the vertex is reconstructed within ±24 cm of the nominal interaction point along the beam axis, with a transverse distance from the beam spot of less than 2 cm [47]. The events are reconstructed with the particle-flow (PF) technique [48]. The PF algorithm reconstructs a complete list of particle candidates in each event from the measurements in all the components of the CMS detector in an integrated fashion. The algorithm separately identifies muons, electrons, photons, charged and neutral hadrons. Charged hadrons that are consistent with primary vertices other than the leading one (defined as the vertex with the largest sum of track p 2 T ) are removed from the collection of particle candidates used to reconstruct the jets, to mitigate the effects of multiple protonproton interactions within the same bunch crossing (pileup). Electrons are reconstructed as isolated objects in the calorimeters which satisfy requirements on the shower shape and the ratio of the hadronic to the electromagnetic energy deposits. Due to the boosted topology of this analysis, some care is needed when reconstructing the Z → e + e − decay: each reconstructed electron interferes with the isolation definition of the other electron and has to be excluded from the isolation calculation in order to avoid introducing inefficiencies. The isolation criterion for electrons is the combined relative isolation R iso built upon information from the tracker, ECAL and HCAL. In calculating R iso , the track momenta and energy deposits, excluding those associated with the electron itself, are summed in a cone of radius ∆R < 0.3 around the electron direction, where ∆R ≡ (∆φ) 2 + (∆η) 2 , and divided by the electron transverse momentum. Muon tracks are built by combining a track from the inner tracker and a track from the outer muon system. No explicit isolation requirement is imposed on the muon candidates. Lepton (electron and muon) candidates are required to have a transverse (longitudinal) distance to the leading vertex smaller than 2 (5) mm. Jets are clustered from the reconstructed PF particles using the infrared-safe anti-k T [49] algorithm with a distance parameter of 0.7, as implemented in FASTJET [50,51]. The jet momentum is determined as the vector sum of all particle momenta in the jet, and is found in the simulated data to be within 5% to 10% of the true momentum of the generator-level jet over the whole p T spectrum and detector acceptance [52]. An area-based correction is applied to take into account the extra energy clustered in jets due to additional proton-proton interactions within the same bunch crossing, and for the average effect of outof-time pileup interactions [53,54]. Jet energy corrections are derived from the energy balance of dijet and photon+jet events. Additional selection criteria are applied to each event to remove spurious jet-like features originating from isolated noise patterns from the hadron or the electromagnetic calorimeters. The offline missing-transverse-momentum vector ( p T miss ) is calculated as the negative vector sum of the transverse momenta of all PF particles reconstructed in the event, and its magnitude is denoted by E miss T .

Dilepton channels
Candidate events are required to have at least two good quality reconstructed leptons within the detector acceptance (|η| < 2.5 for electrons and |η| < 2.4 for muons, with at least one muon within |η| < 2.1 at the trigger level) with p T > 45 GeV. We also require at least one jet in the event reconstructed with p T > 30 GeV within |η| < 2.4.
Whenever two same-flavor, opposite-sign leptons are found in the event, a Z candidate is formed summing the four-momenta of the leptons. We select the Z candidates by requiring their invariant mass to be in the range 70 < M Z < 110 GeV and with a transverse momentum p Z T > 150 GeV. If there are multiple Z candidates, the one with mass closest to the nominal Z mass is selected. The requirement that the dimuon mass is consistent with a Z → µ + µ − decay strongly suppresses non-prompt muons from jets.
The V candidate is selected by requiring a reconstructed jet with p T > 250 GeV and |η| < 2.4, having an invariant mass M j (computed from the jet energy and momentum calculated as the vector sum of the four-momenta of the constituent PF particles) such that 65 < M j < 120 GeV. We require the jet to be well separated from the two leptons forming the Z candidate: ∆R ≡ (∆η) 2 + (∆φ) 2 > 1.0 for each lepton, where ∆η (∆φ) is the pseudorapidity (azimuthal) distance between the jet and the lepton directions. The selection has been optimized by maximizing the quantity N S / √ N S + N B (where N S and N B are the number of expected signal and background events) for the lowest W mass point (700 GeV) considered in this search.
Once the Z → and (mono-jet) V→ qq candidates have been reconstructed, we combine their four momenta to compute the mass of the parent particle, M VZ . This variable is used to evaluate the hypothesis of the signal presence in the data sets analyzed.

T channel
For the E miss T channel, background from W-boson decays is reduced through rejection of events with isolated electrons or muons with p T > 20 GeV. In order to further reduce leptonic backgrounds, we veto on the presence of isolated tracks. For all tracks with p T > 10 GeV and |η| < 2.4, a hollow cone of 0.02 < ∆R < 0.30 is constructed. The isolation parameter of each track is defined as the scalar sum of all tracks with p T > 1 GeV inside the cone, divided by the p T of the track. Events containing a track with its isolation parameter smaller than 0.1 are discarded. Events are then selected if the jet with the highest transverse momentum has p T > 300 GeV and |η| < 2.4, and E miss T is larger than 300 GeV. In order to reduce the number of QCD multijet background events in the signal region, events with more than two jets with p T > 30 GeV and |η| < 2.4 are discarded. Events with exactly two jets above 30 GeV are retained, if the azimuthal angle ∆φ between the two jets is smaller than 2.8 radians. This condition improves the signal over background ratio by reducing the number of QCD dijet background events.
The signal sample is defined as the set of events that meet two extra requirements: the invariant mass of the leading jet, M j , is larger than 70 GeV, and the jet-E miss T transverse mass, defined as is larger than 900 GeV. Figure 1 on the left (right) shows the two-dimensional M T vs M j distribution for the simulated SM backgrounds (for a simulated signal sample with M G KK = 1250 GeV).
In contrast to the approach used for the dilepton channels, here we perform a single "event counting" experiment by comparing the number of expected background and observed events integrated over the region M j > 70 GeV and M T > 900 GeV.

Background Estimation
We are discussing the background estimation separately for the dilepton and E miss T channels.

Dilepton channels
The analysis of the simulated data shows that the dominant (∼90%) background after all selection requirements is the inclusive Z production ("Z+jets"), with additional contributions from tt+jets and the continuum SM diboson production (WZ and ZZ). The shape and the overall normalization of the expected background M VZ distributions are derived from data, with additional cross-checks carried out with the inclusive simulated background samples. Effects caused by pileup are modeled by adding to the generated events multiple proton-proton interactions with a multiplicity distribution matched to the luminosity profile of the collision data.
The background is modeled using a control region consisting of a sideband in M j (30 < M j < 65 GeV). The remaining selections are applied unmodified to these events, providing a sample that is kinematically equivalent to the nominal selection. The robustness of this method against pileup effects, jet energy scale uncertainties, and variations in the sideband range has been confirmed with dedicated studies (Section 6).
The procedure is as follows: we first produce the M VZ distribution for the sideband selection. We define the ratio α(M VZ ) as the total number of Monte Carlo background entries in the M VZ spectrum with the nominal (65 < M j < 120 GeV) and sideband (30 < M j < 65 GeV) selections: where N NS (M VZ ) is the number of events in the signal region and N SB (M VZ ) is the number of events in the sideband region, contained in a bin of the VZ mass distribution centered at a given value M VZ . We then use the product of the M VZ distribution made with the sideband selection in the data and the ratio α(M VZ ) to derive an estimate of the background M VZ distribution with the nominal selection. Following the example of other resonance searches [55], we fit the α-corrected sideband data M VZ distribution to the following analytic function f A (M VZ ), and the fit result is used to parametrize the expected SM background distribution: where √ s is the collision energy, p i , i = 0, ..., 3 are free parameters of the fit, and M VZ is expressed in GeV. The fit determines both the shape and the overall normalization of the expected background as a function of M VZ . The fitting functions are then used to describe the expected background in any subregion of the M VZ spectrum in the electron and muon channels. There are several advantages in using the ratio α(M VZ ) for the background modeling of the M VZ distributions: the background estimation becomes insensitive to effects such pile-up corrections and integrated luminosity uncertainty which cancel out in the ratio; α(M VZ ) is less sensitive to improper modeling of the matrix element calculation for the background and to theory systematics (e.g. normalization and factorization scale, PDFs, etc.) since the background composition is similar in the two regions.  The comparison of the estimated background with the prediction from the simulation and the data M VZ distributions is shown in Fig. 2. No significant excess of events is observed, with the largest deviation appearing in the ∼900 GeV region in the muon channel. The tail of the M VZ distribution, which is the region of interest for the new resonance search, is well described by the fit. A discrepancy is observed at low M VZ values. Any modeling imperfections, quantified as the difference between the best-fit function and the MC simulation estimation, are taken into account in the limit calculation by assigning a systematic uncertainty.

T channel
By analyzing simulated data we determine that the dominant backgrounds in this channel after all selections are inclusive Z → νν (∼70%) and W → ν (∼30%) production, with the charged lepton remaining undetected in the latter. To estimate the SM background, we use a sideband-based technique similar to that described above, which utilizes events that meet all other requirements but the M j and M T thresholds. In particular, the events which meet all the selection requirements are classified into four regions according to two thresholds in jet mass and two thresholds in M T : • Signal region A: M j > 70 GeV, M T > 900 GeV; • Sideband region B: 20 < M j < 70 GeV, M T > 900 GeV; • Sideband region C: 20 < M j < 70 GeV, 700 < M T < 900 GeV; • Sideband region D: M j > 70 GeV, 700 < M T < 900 GeV; The numbers of events observed in the above regions are denoted as N A , ..., N D .

5 Background Estimation
The estimated total background B est in Region A is given by the expression where ρ is a correction factor to account for the correlation between the jet mass and the jet-E miss T transverse mass. The ρ parameter is estimated from the simulated SM samples by rearranging Eq. (1) in the following way: and setting the values of N A , . . . , N D to the ones from the SM prediction. Using the values reported in Table 3 gives  Events / 10 GeV

Systematic Uncertainties
The systematic uncertainties that are considered in this analysis can be divided into two main categories: the uncertainty in the determination of the SM background and the uncertainty in the expected yields of signal events. All the systematic uncertainties are summarized in Tables 4  and 5 for the dilepton and E miss T channels, respectively. The total systematic uncertainty is the combination of the signal and background systematic effects, assuming they are completely uncorrelated.

Background systematic effects
As we employ a method based on control samples in data for the background determination, several systematic effects are eliminated. In the following, we consider the remaining relevant uncertainties in detail for the dilepton and E miss T channels.

Dilepton channels
The expected number of background events in each mass window is determined by the integral of the function f A (M VZ ) in the corresponding region. The statistical uncertainty is calculated by employing the full covariance error matrix of the fit parameter uncertainties in the integral of the fitting function in the mass window. The pileup and jet energy scale (JES) systematics can potentially affect the background determination through the α(M VZ ) ratio and are considered separately. The former is found to have a negligible effect. For the latter, the uncertainty is evaluated by varying the jet p T to p T ± σ JES (p T , η), where σ JES (p T , η) is the total jet uncertainty, and applying the full fitting procedure. The yield differences, in each mass window, between the expected background with the positive (N +Bgd ) and negative (N −Bgd ) jet energy scale variation with respect to the nominal selection and fit are taken as the ±1 σ estimates for the JES systematic uncertainty. We also consider several variations in the fitting procedure (fitting range, functional form, and sideband definition). These variations are compared to the difference in the number of expected background events in the mass window as estimated from data and with MC simulation. The largest of the two is used as the systematic uncertainty in the background determination.

T channel
To evaluate the robustness of the evaluation of the expected background two tests are conducted.
The first test studies the dependence of the correction factor ρ on the definition of the sideband regions. We vary the definition of the sideband regions by changing the thresholds in the M j and M T variables in the intervals 20-70 GeV and 650-750 GeV, respectively. We find that the resultant variation in the mean estimated background is typically 5% or less, confirming the robustness of the sideband method.
A second test is used to check the propagation of all the uncertainties involved in the B est calculations. We generate a series of pseudo-experiments with the number of events constrained to be equal to that of the actual experiment. We obtain a value of ρ and calculate the mean estimated background in each case. The distribution of the values of B est thus obtained has a variance of 20 events. This result is in agreement with the estimated uncertainty on B est obtained in Section 5.2, using the yields of N B , N C and N D observed in the data.
The mean expected SM background in Region A, within the uncertainties calculated above, is compatible with the observed event yield in the signal region, N A = 138 events. Table 4: Systematic uncertainties in the dilepton channels for given mass point and optimized mass window for the background (columns 3-6) and signal (columns 7-8) expected yields, following the procedures described in the text. In addition to the estimated signal uncertainties listed in the Table, constant uncertainties are considered on the integrated luminosity (2.2%), the lepton reconstruction and trigger efficiencies as determined by the "tag and probe" method (2%) [56] and the V mass selection as determined from a sample of boosted tt events (9%).

Signal systematics
There are several systematic uncertainties in the expected signal yields that are common across channels. These uncertainties are on the luminosity measurement, the JES effects on jets, the PDF, and the trigger and reconstruction efficiencies. A value of 2.2% was taken for the uncertainty in the measurement of the integrated luminosity [57].
To determine the effect of JES uncertainty, we vary the jet p T to p T ± σ JES (p T , η), where σ JES (p T , η) is the total jet uncertainty, and apply the full analysis selection. The differences in the signal yields N +sig and N −sig with respect to the nominal selection N sig are taken as the ±1 σ estimates for the JES systematic uncertainty. For W and RS signals with the mass in the range [700,2000] GeV in the dilepton channels and in the range [1250,2000] GeV in the E miss T channel, this systematic uncertainty is less than 1%. However, for resonance masses in the range [700, 1200] GeV in the E miss T channel this systematic uncertainty is found to be between 2 and 9%, owing to threshold effects. To estimate the systematic uncertainty associated with the choice of the PDF used for the simulated samples, three scenarios are considered: CTEQ6.6, MSTW2008 and NNPDF2.0 [58]. The systematic uncertainty is set to half of the difference between the maximum and the minimum PDF values predicted for each mass point [59].

Dilepton channels
To account for differences in trigger and reconstruction efficiencies between the Monte Carlo simulation and data, we determine scaling factors by using data control samples of Z → µµ and Z → ee candidate events [60,61]. We derive corrections for the muon (0.974 ± 0.001) and the electron (0.960 ± 0.004) channels and we apply them to the expected signal yields. These numbers assume that the efficiency does not vary with p T (E T ). However, we observe a small decrease (increase) in the efficiency in the asymptotic high-p T (high-E T ) region for muons (electrons) of about 2%. This small difference is used as the systematic uncertainty in the expected number of signal events for each mass point considered in this study. Finally, we assign a

Results
9% systematic uncertainty on the V mass selection efficiency. This is determined by studying an independent sample of boosted tt → Wb Wb events in which one of the W bosons decays leptonically and the other hadronically.

T channel
Propagating the jet energy scale effects to the calculation of E miss T , and accounting for the anticorrelation between jets and E miss T itself, we estimate a systematic effect of around 3% for all values of M G studied, except for the lowest M G = 750 GeV. In this case, because of threshold effects, the systematic effect is found to be around 7%.
Summing in quadrature the uncertainties above, we arrive at a final 5% systematic uncertainty on the signal acceptance and efficiency except for M G = 750 GeV, where a value of 10% is obtained for the final systematic uncertainty on the signal acceptance and efficiency.

Results
We do not observe any significant excess over the expected background. We employ the modified frequentist CL S statistical method [62,63] to search for exotic VZ resonances. For the dilepton channels we use a series of search windows corresponding to different mass hypotheses. Each mass window is optimized to give the best exclusion limit, a procedure which is also appropriate for establishing a new resonance discovery. The mass windows optimization has been carried out separately for the W and RS graviton hypotheses to account for differences in the width and efficiencies. For the E miss T channel we perform a single counting experiment in the M T > 900 GeV and M j > 70 GeV region. We calculate 95% confidence level (CL) exclusion limits on the combined products of the cross section times the branching ratio σ(pp → W ) × B(W → WZ) and σ(pp → G KK ) × B(G KK → ZZ) for the three final states under study (separately and combined) as a function of the mass of the hypothetical resonance. We interpret these exclusion limits in two benchmark signal models: SSM W and RS graviton.
The limit setting is performed by looking for an excess over the expected background in the VZ mass distributions for the three channels separately. Tables 6 and 7 show the search windows for each mass point with the corresponding signal efficiency and the numbers of expected background and observed events in the electron and muon channels, respectively. These numbers are used as input for the calculation of the expected and observed exclusion limits on cross section times branching ratios at 95% CL that are also reported in the same Tables. Table 8 shows the signal efficiency and the observed and expected exclusion limits as a function of the signal mass in the E miss T channel. The combined results are reported in Table 9. The exclusion limits as a function of the VZ resonance mass can be seen in Fig. 4, where a linear interpolation is used between the benchmark mass values. These limits can be interpreted in the theoretical framework of the W and RS graviton models. We exclude SSM W bosons with masses in the range 700-940 (890) GeV in the SSM at NNLO (LO) at 95% CL These results are complementary to the ones obtained in the tri-lepton analysis (with M W > 1143 GeV in the SSM [64]). The exclusion limit calculated in the RS graviton model rules out masses (M G KK ) in the range 750-880 (800) GeV for k/M Pl = 0.05 at NLO (LO). Assuming the resonance width is much smaller than the experimental resolution for the range of k/M Pl considered here, the limit can be translated into the M G KK -k/M Pl plane. We do this by using the quadratic dependence of the cross section on k/M Pl , and by assuming that the signal efficiency remains the same. The result is shown in Fig. 5.
These results are particularly relevant in the context of RS models proposed in recent studies [65], with SM fields propagating in the extra dimension where the graviton coupling to light fermions is strongly suppressed. This opens the possibility to an enhancement of the branching fractions for final states with V pairs, and motivates the investigation of large values of k/M Pl [65]. In this scenario the previously-published searches for RS gravitons decaying to γγ and + − final states [24] do not impose stringent bounds, since the branching fraction for these final states is suppressed. The results derived in this analysis are currently the most stringent in the V pair channel, and provide important constraints that are complementary to the ones from the search for resonances decaying to boosted top pairs [66]. Table 6: Electron channel: Search window for each mass point with the corresponding signal efficiency (" sig ") and the numbers of mean expected background ("N bgd ") and observed ("N obs ") events. The uncertainties include both statistical and systematic effects. These numbers are used as input for the calculation of the expected and observed exclusion limits on σ(pp → W ) × B(W → WZ) and σ(pp → G KK ) × B(G KK → ZZ) at 95% CL which are reported in the last two columns.       [27] ATLAS Collaboration, "Search for a heavy gauge boson decaying to a charged lepton and a neutrino in 1 fb −1 of pp collisions at √ s = 7 TeV using the ATLAS detector", Phys. Lett. B 705 (2011) 28, doi:10.1016/j.physletb.2011.09.093, arXiv:1108.1316.
[36] CMS Collaboration, "Search for a narrow spin-2 resonance decaying to a pair of Z vector bosons in the semileptonic final state", (2012). arXiv:1209.3807. Submitted to Phys. Lett. B.