Search for the Production of a Long-Lived Neutral Particle Decaying within the ATLAS Hadronic Calorimeter in Association with a Z Boson from pp Collisions at √s = 13 TeV

This Letter presents a search for the production of a long-lived neutral particle ( Z d ) decaying within the ATLAS hadronic calorimeter, in association with a standard model (SM) Z boson produced via an intermediate scalar boson, where Z → l þ l − ( l ¼ e , μ ). The data used were collected by the ATLAS detector during 2015 and 2016 pp collisions with a center-of-mass energy of ﬃﬃﬃ s p ¼ 13 TeV at the Large Hadron Collider and correspond to an integrated luminosity of 36 . 1 (cid:2) 0 . 8 fb − 1 . No significant excess of events is observed above the expected background. Limits on the production cross section of the scalar boson times its decay branching fraction into the long-lived neutral particle are derived as a function of the mass of the intermediate scalar boson, the mass of the long-lived neutral particle, and its c τ from a few centimeters to one hundred meters. In the case that the intermediate scalar boson is the SM Higgs boson, its decay branching fraction to a long-lived neutral particle with a c τ approximately between 0.1 and 7 m is excluded with a 95% confidence level up to 10% for m Z d between 5 and 15 GeV.

Search for the production of a long-lived neutral particle decaying within the ATLAS hadronic calorimeter in association with a Z boson from p p collisions at √ s = 13 TeV The ATLAS Collaboration This Letter presents a search for the production of a long-lived neutral particle decaying within the ATLAS hadronic calorimeter, in association with a Standard Model Z boson produced via an intermediate scalar boson, where Z → ℓ + ℓ − (ℓ = e, µ).The data used were collected by the ATLAS detector during 2015 and 2016 pp collisions with a center-of-mass energy of √ s = 13 TeV at the Large Hadron Collider and corresponds to an integrated luminosity of 36.1 fb −1 .No significant excess of events is observed above the expected background.Limits on the production cross section of the scalar boson times its decay branching fraction into the long-lived neutral particle are derived as a function of the mass of the intermediate scalar boson, the mass of the long-lived neutral particle, and its cτ from a few centimeters to one hundred meters.In the case that the intermediate scalar boson is the SM Higgs boson, its decay branching fraction to a long-lived neutral particle with a cτ approximately between 0.1 m and 7 m is excluded with a 95% confidence level up to 10% for m Z d between 5 and 15 GeV.Many extensions to the Standard Model (SM) such as supersymmetry [1,2], inelastic dark matter [3] and hidden valley scenarios [4,5] predict the existence of long-lived neutral particles that can decay hadronically.Searches for the pair production of such particles have been carried out by the ATLAS [6-9], CMS [10,11], and LHCb [12,13] experiments at the Large Hadron Collider (LHC), and the CDF [14] and D0 [15] experiments at the Tevatron.This Letter reports a search for hadronic decays of long-lived neutral particles, denoted by Z d hereafter, singly produced in association with a SM Z boson through an intermediate scalar Φ or Higgs boson, pp → Φ/H → Z Z d , where Z → ℓ + ℓ − (ℓ = e, µ).Production of a new particle in association with a Z boson is a popular scenario in hidden-or dark-sector models with an additional U(1) d dark gauge symmetry [16,17].One such model has been tested by the ATLAS experiment in a search for a new particle that is mediated by the Higgs boson and decays promptly to a lepton pair [18,19].This analysis expands the search to a more general case to include a possible new scalar (Φ) that couples to Z and Z d , instead of only the Higgs boson, and considers the scenario in which the Z d decays hadronically with a cτ between a few centimeters and tens of meters, where c is the speed of light and τ is the Z d proper lifetime.
The analysis uses data from √ s = 13 TeV proton-proton (pp) collisions at the LHC that were recorded by the ATLAS detector in 2015 and 2016 with single-electron and single-muon triggers [20], corresponding to an integrated luminosity of 36.1 fb −1 .The ATLAS detector [21] is a multipurpose particle detector with a cylindrical geometry 1.It consists of an inner detector (ID) [22] surrounded by a solenoid that produces a 2 T magnetic field, electromagnetic and hadronic calorimeters, and a muon spectrometer in a magnetic field produced by a system of toroid magnets.The ID measures the trajectories of charged particles over the full azimuthal angle and in a pseudorapidity range of |η| < 2.5 using silicon pixel, silicon microstrip, and straw-tube transition-radiation tracker detectors.Liquid-argon electromagnetic calorimeters (LArCal) extend from 1.5 m to 2.0 m in radius in the barrel and from 3.6 m to 4.25 m in |z| in the endcaps.A scintillator-tile calorimeter (TileCal) provides hadronic calorimetery and covers the region 2.25 m < r < 4.25 m.The experimental signature searched for is the Z d decaying within the TileCal, thus producing a jet that has little or no energy deposited in the LArCal, and no charged tracks that point to the primary vertex.
Monte Carlo (MC) simulated events are used to optimize the event selection and to help validate the analysis.Signal samples were generated using the P 8.210 [23] generator with the NNPDF23LO parton distribution functions (PDF) [24] and the A14 set of tuned parameters (A14 tune) [25], with an assumption that the Z d decays only to the highest-mass heavy quark pair (b b or c c) that is kinematically allowed.Nine samples were produced with three different Z d masses for each of three Φ masses (m Z d = {5, 10, 15}, {10, 50, 100} and {20, 100, 200} for m Φ = 125, 250, 500 GeV, respectively), where m Φ = 125 GeV corresponds to the SM Higgs boson.The cτ of the Z d is a free parameter in this model.For each mass hypotheses of Z d and Φ, its cτ is chosen to maximize the probability for Z d to decay inside the TileCal, which is found to be around 20% for all samples, as shown in Figure 1(a).The events were reweighted to produce samples with different cτ(Z d ) [8] between 0.01 m and 100 m.The dominant SM background arises from events with a Z boson produced in association with jets (Z+jets), where a jet mimics the experimental signature of Z d decay inside the TileCal due to the presence of long-lived SM particles (K L , Λ, etc), out-of-time pileup (additional pp collisions occurring in bunchcrossings just before and after the collision of interest), noise, detector inefficiencies, and beam-induced background.Additional SM background processes include the production of top quarks and W+jets.The SM background MC samples are generated with the configurations described in Ref. [26] for W/Z+jets 1 ATLAS uses a right-handed coordinate system with its origin at the nominal interaction point (IP) in the center of the detector and the z-axis along the beam pipe.The x-axis points to the center of the LHC ring, and the y-axis points upward.Cylindrical coordinates (r, φ) are used in the transverse plane, φ being the azimuthal angle around the z-axis.The pseudorapidity is defined in terms of the polar angle θ as η = − ln tan(θ/2).The distance between two objects in η-φ space is ∆R = (∆η) 2 + (∆φ) 2 .
Transverse momentum is defined by p T = psinθ.production, and Ref. [27] for t t and single top production.The effect of multiple pp interactions in the same and neighboring bunch crossings (pileup) is included by overlaying minimum-bias events simulated with P 8.186 on each generated event in all samples.The generated samples were processed through a G 4-based detector simulation [28,29] and the standard ATLAS reconstruction software.
The selected events have a pair of oppositely charged and isolated electrons [30] or muons [31]  ) is required to be between 66 GeV and 116 GeV.Selected jets must have transverse energy E T > 40 GeV and |η| < 2.0 to ensure the jets are completely within the ID.They are reconstructed using the anti-k t algorithm [32,33] with a radius parameter R = 0.4 and calibrated to particle level [34].
Standard ATLAS jet-quality criteria [35] are applied, except the one for the hadronic energy fraction since it removes signal jets.A jet is considered as a Z d candidate, referred to as a calorimeter-ratio jet (CRjet) hereafter, if it satisfies log 10 (E Tile /E LAr ) > 1.2 with no ghost-associated [36] tracks of p T > 1 GeV originating from the primary vertex, where E Tile and E LAr are the jet energy deposited in the TileCal and LArCal, respectively [6], as shown in Figure 1(b).Jets with E T < 60 GeV in the transition region between the barrel and end-cap cryostats (1.0 < |η| < 1.3) are not considered as CR-jet candidates due to noise in the gap scintillator of the TileCal [37].In addition, the timing of the CR-jet is required to be between −3 ns and 15 ns in order to suppress jets arising from out-of-time pileup and beam-induced backgrounds [6].
The timing of a jet is obtained from its constituent calorimeter cells by calculating an average time over cells weighted by cell-energy squared where the cell time is measured according to the bunch crossing clock, relative to the expected time-of-flight from the bunch crossing to the cell [38].After this selection, the number of selected events containing a CR-jet with an E T above a chosen threshold is compared with the predicted total number of background events.The minimum E T requirement of the selected CR-jets is further optimized to achieve the highest experimental sensitivity for each mass hypothesis.It is set to be 40 GeV for m Φ = 125 GeV samples, 60 GeV for m Φ = 250 GeV samples and 80 GeV for m Φ = 500 GeV samples.
The signal efficiency times acceptance (ǫ × A) is defined as the ratio of the number of selected signal events in MC simulations to the number of generated signal events.It is a function of m Φ , m Z d and the cτ(Z d ).
The maximal values vary between approximately 1% for lowest m Φ samples to 5-7% for samples with larger Φ mass.The main loss is due to the low probability that Z d decays inside the TileCal, as shown in Figure 1(a).The samples for m Φ = 125 GeV suffer further efficiency loss due to the jet E T requirement.
A data-driven approach is used to estimate the background.A control data sample of SM W+jets events with the same event selection criteria of W → ℓν (ℓ = e, µ) in Ref.
[39], is used to derive the probability for a jet to pass the selection of the CR-jet, assuming that the Z d cannot be produced in association with a W boson.The probability is calculated as f CR = N CR-jet /N jet in bins of the jet E T and η, where N CR-jet is the number of jets that satisfy the CR-jet selection criteria and N jet is the total number of jets from the W+jets sample in each bin, as summarized in Table 1.For a selected event in data containing a Z → ℓℓ candidate and N jets, the corresponding probability for it to be identified as a signal event is therefore , where f CR (E i T , η i ) is the probability of the i-th jet in the event to satisfy the CR-jet selection criteria.The sum of the probabilities P for all the selected events is therefore the expected number of background events.
Studies [6] have shown that jets originating from quarks and gluons may have different probabilities of satisfying the selection criteria for CR-jets.MC simulations predict that jets from W+jets and Z+jets production are mostly initiated by quarks with a similar fraction (∼ 73%).However, W+jets data samples are contaminated with a significant fraction of SM multijet events with a misidentified lepton, which is estimated to be approximately 2% in the muon final state and 17% in the electron final state using background-enriched control samples [39].SM multijets originate primarily from gluons and thus introduce a difference between the Z+jets and W+jets samples.The distributions of the track multiplicity of a jet in the W/Z+jets samples, which are sensitive to the quark/gluon jet fraction [40], show a significant difference in Figure 1(c).As a result, the f CR values measured in the muon final state are used for the central value of the background estimate, while the f CR values measured in the electron final state are used as a cross-check to assign a systematic uncertainty due to different quark/gluon jet fractions in the W+jets and Z+jets samples.The measured probabilities, f CR , are found to be dependent on the jet multiplicity in the event.Studies show that this is caused by the presence of jets from pileup interactions which deposit additional energy in the LArCal, suppressing the signature of CR-jets.The jet multiplicity and pileup distributions of events in the W+jets sample are the same as those from the Z+jets sample and therefore the parameterization of the measured f CR as a function of jet multiplicity or pileup is not necessary.
Several studies were performed to validate the background estimation procedure.A Z+jets sideband region is formed from events satisfying all signal selection criteria except the invariant-mass requirement for the Z candidate.The mass is required to be 30 GeV < m ℓℓ < 55 GeV.The events in the higher mass sideband m ℓℓ > 116 GeV are not used as they are still dominated by Z+jets production, as indicated by background MC simulations.Based on the measured CR-jet probability in W+jets, the expected numbers of background events with E T of CR-jets greater than 40, 60, and 80 GeV are estimated to be 2.2 ± 0.2, 0.7 ± 0.1, and 0.3 ± 0.1, where the uncertainties are statistical only.They are consistent with the corresponding observations in data, which have 1, 1, and 0 events, respectively.
The background estimation method relies on an assumption that jets in the W+jets sample have the same characteristics as jets in the Z+jets sample.This assumption is tested using validation jets that are defined to satisfy the selection criteria of the CR-jets except the zero-track requirement.Validation jets must have more than two associated tracks to avoid signal contamination, as MC-simulated signal events show that less than 1% of jets from Z d decays inside the TileCal have more than two tracks.The probability for a jet to be identified as a validation jet is measured in the W+jets sample as a function of jet E T and η and subsequently used to predict the number of events containing a Z → ℓℓ candidate and at least one validation jet.As a result, a global scale factor of 1.24, which is defined as the observed number of events with validation jets divided by the predicted value, is applied to the measured probabilities f CR .A 50% relative correction of the scale factor (± 0.12) is assigned as a systematic uncertainty due to potential bias of the background estimate procedure.
The systematic uncertainties of the background estimation include the statistical uncertainty from the W+jets sample (2-8%), potential difference in the quark/gluon jet fractions between the W/Z+jets samples (7-20%), and the scale factor uncertainty (∼ 10%) measured using the validation jets.The uncertainty of the integrated luminosity is 2.1% [41,42].Uncertainties resulting from detector effects such as the trigger efficiencies, the energy scale and resolution of jets [34], lepton identification, reconstruction and isolation efficiencies, lepton momentum scales and resolutions [30,31,43] only affect the calculation of the selection efficiencies of Z d signal events, since the background is estimated from the data.They are typically small (< 1-5%).Pileup adds extra tracks and electromagnetic energy to jets.The systematic uncertainties associated with reweighting the pileup distribution from the generated MC simulations to the data are found to be typically small (< 5%) except for the samples with m Φ = 125 GeV (∼ 13%), in which case the Z d have small energies and additional energy deposition in the LArCal from pileup can significantly affect their selection efficiencies.Since the CR-jets in this analysis have a very small fraction of their energies inside the LArCal, the in situ jet energy intercalibration [6, 34] is repeated using the p T balance method in dijets events, and the observed difference between the data and MC simulation is used to derive an additional systematic uncertainty of the jet energy scale.The corresponding effect on the signal efficiencies is found to be approximately 5-9% for samples with m Φ = 125 GeV, and negligible for samples with higher m Φ values.The effects on the signal efficiency and acceptance due to theoretical uncertainties, such as PDF choice and initial-and final-state radiation modeling, are found to be very small (< 1%).
Table 2 shows the predicted numbers of background events and the observed data events with different minimum E T requirements for the selected CR-jets.The data are well-described by the background estimate.In the absence of any significant data excess, upper limits (UL) on the signal yield of pp → Φ → Z Z d at the 95% confidence level (CL) are derived using the CL s method [44] taking into account both the statistical and systematic uncertainties.The results are listed in Table 2.The results are further reinterpreted as the UL on the production cross section of Φ times the decay branching fraction  In conclusion, a search for a long-lived neutral particle Z d produced in association with a SM Z boson via coupling to an intermediate scalar boson is presented.The analysis is based on 36.1 fb −1 of pp collisions at √ s = 13 TeV collected in 2015 and 2016 with the ATLAS detector at the LHC.No excess over the expected background was observed.Upper limits on the production cross section of the scalar boson times its branching fraction to the long-lived neutral particle at 95% CL are derived as a function of the particle proper lifetimes for different masses of the scalar boson and the Z d .In the case that the intermediate scalar boson is the SM Higgs boson, its decay branching fraction to a long-lived neutral particle with a cτ approximately between 0.1 m and 7 m is excluded with a 95% CL up to 10% for m Z d between 5 and 15 GeV.We thank CERN for the very successful operation of the LHC, as well as the support staff from our institutions without whom ATLAS could not be operated efficiently.The crucial computing support from all WLCG partners is acknowledged gratefully, in particular from CERN, the ATLAS Tier-1 facilities at TRIUMF (Canada), NDGF (Denmark, Norway, Sweden), CC-IN2P3 (France), KIT/GridKA (Germany), INFN-CNAF (Italy), NL-T1 (Netherlands), PIC (Spain), ASGC (Taiwan), RAL (UK) and BNL (USA), the Tier-2 facilities worldwide and large non-WLCG resource providers.Major contributors of computing resources are listed in Ref. [

Figure 1 :
Figure 1: (a) The probability of a Z d boson to decay within the TileCal as a function of the cτ for each choice of m Φ and m Z d .As m Z d increases (for a fixed m Φ ) the Z d becomes less boosted and therefore travels less distance into the detector before decaying.(b) The distributions of log 10 (E Tile /E LAr ) for jets in background and signal MC simulations (see legend of Figure 1(a) for signal labels) and W+jets data (prior to any requirements on the track multiplicity of jets or jet timing).The threshold for this variable is shown as a solid black line.(c) The distributions of the track multiplicity for jets prior to the selection of CR-jets in the W+jets and Z+jets data samples.
as a function of m Φ , m Z d , and cτ of the Z d .In the case of the SM Higgs boson, where m H = 125 GeV, the UL on B(H → Z Z d ) are evaluated using the SM Higgs boson cross section σ SM = 48.5 +4.6 −6.7 pb [45] of the gluon-gluon fusion process; other production modes are ignored.The results, reweighted to other cτ [8], are shown in Figure 2.

Figure 2 :
Figure 2: (a) Observed 95% CL limits on the decay branching fraction of B(H → Z Z d ) for the SM Higgs boson as a function of the cτ(Z d ).(b) and (c) Observed 95% CL limits on the production cross section (σ) of Φ times its decay branching fraction to Z Z d as a function of the cτ(Z d ).

Table 1 :
The numbers of jets satisfying different requirements on minimum jet E T and their corresponding averaged CR-jet selection probabilities in the W → ℓν samples.

Table 2 :
Event yields for the predicted backgrounds and data, and the expected and observed UL on the signal yields at the 95% CL.The reported errors include both the statistical and systematic uncertainties.