Observation of production in collisions at 1.96 TeV
We present an observation for (, = or ) production in collisions at a center-of-mass energy of TeV. Using 1.7 fb of data collected by the D0 experiment at the Fermilab Tevatron Collider, we observe three candidate events with an expected background of events. The significance of this observation is 5.3 standard deviations. The combination of D0 results in this channel, as well as in , yields a significance of 5.7 standard deviations and a combined cross section of pb.
pacs:12.15.Ji, 07.05.Kf, 13.85.Qk, 14.70.Hp
Studies of the pair production of electroweak gauge bosons provide an interesting test of the electroweak theory predictions brown (). In contrast with other diboson processes, boson pair production () does not involve trilinear gauge boson interactions within the standard model (SM). The observation of an unexpectedly high cross section could indicate the presence of anomalous or couplings zz-theory (). The SM prediction for the total production cross section in collisions at the Fermilab Tevatron Collider at TeV is pb Campbell:1999ah (). The requirement of leptonic boson decays reduces the observable cross section, making its measurement rather challenging. The accumulation of integrated luminosities in excess of 3 fb at the Fermilab Tevatron Collider and the development of highly optimized event selection criteria has now made possible the direct observation of production.
Previous investigations of production have been performed both at the Fermilab Tevatron and the CERN (LEP) lep () Colliders. The D0 collaboration reported a search for (, = or ) with 1 fb of data that provided a 95% C.L. limit of pb runiia_zz (). The CDF collaboration reported a signal for production with a significance of 4.4 standard deviations from combined and searches, and measured a production cross section of pb cdf_zz ().
In this Letter, we present a search for boson pairs where the bosons have decayed to either electron or muon pairs, resulting in final states consisting of four electrons (), four muons () or two muons and two electrons (). Data used in this analysis were collected with the D0 detector at the Fermilab Tevatron Collider at TeV between June 2006 and May 2008. The integrated luminosities d0lumi () for the three analyzed channels are about 1.7 fb. This result is later combined with that from an earlier similar analysis runiia_zz () using data collected from October 2002 to February 2006 and corresponding to an integrated luminosity of 1 fb.
The D0 detector run2det () consists of a central tracking system, comprised of a silicon microstrip tracker (SMT), and a central fiber tracker (CFT), providing coverage to pseudorapidity pseudo (), both located within a 2 T superconducting solenoidal magnet. Three liquid argon and uranium calorimeters provide coverage to . Electromagnetic objects are well reconstructed in the regions of the central calorimeter (CC) with coverage to and the end calorimeters (EC) with coverage of . A muon system surrounds the calorimetry, consisting of three layers of scintillators and drift tubes and 1.8 T iron toroids, with a coverage of .
This analysis employs a combination of single and dielectron triggers for the 4e channel. Similarly, single and dimuon triggers are used for the channel. The channel uses a combination of all these triggers, and additional specific electron-muon triggers. The triggering efficiency for events with four leptons having high transverse momentum ( that satisfy all offline selection requirements exceeds 99%.
For the channel, we require four electrons with ordered transverse energies 30, 25, 15, and 15 GeV, respectively. Electrons must be reconstructed either in the CC region or in the EC region, and be isolated from other energy clusters in the calorimeter. Electrons in the CC are required to satisfy identification criteria based on a multivariate discriminant derived from calorimeter shower shape and a matched track reconstructed in the SMT and CFT. Electrons in the EC are not required to have a matched track, but must satisfy more stringent shower shape requirements. At least two electrons must be in the CC region. With no requirement applied on the charge of the electrons at this stage to increase selection efficiency, three possible combinations can be formed for each event. Events are required to have a solution for which one combination has an invariant mass GeV and the other GeV. Finally, events are split into three categories, depending on the number of electrons in the CC region. Subsamples with two electrons, with three electrons and with four or more electrons in the CC are denoted as , , and , respectively. The three exclusive subsamples contain significantly different levels of background contamination and thus the separation of the subsamples provides more sensitivity to the search.
For the channel, each muon must satisfy quality criteria based on scintillator and wire chamber information from the muon system, and have a matched track in the central tracker. We require that the four most energetic muons have ordered transverse momenta 30, 25, 15, and 15 GeV, respectively. At least three muons in the event must be isolated, each passing a requirement of less than GeV of transverse energy deposited in the calorimeter in the annulus centered around the muon track cal_iso (). Finally, the muon is required to be well reconstructed and to originate from the primary event vertex. Of the three possible combinations per event that can be formed without considering muon charge at this stage, a solution is required where one combination has an invariant mass GeV and the other GeV.
For the channel, the two most energetic electrons and muons in an event must have GeV. All muons and electrons must satisfy the single lepton selection criteria defined for the and final states. In addition, electrons and muons are required to be spatially separated by to remove background with muons radiating photons giving events with two muon and two trackless electron candidates. At least one muon must satisfy the same calorimeter isolation requirement imposed in the final state. A solution is required where one pair of same flavor leptons has an invariant mass GeV, and the other GeV. Finally, events are split into three categories depending on the number of electrons in the CC region. Subsamples with no electron, with one electron and with two or more electrons in the CC are denoted as , and , respectively. As in the channel, these subsamples have significantly different levels of background contamination.
A Monte Carlo (MC) simulation is used to determine the expected number of signal events in each subchannel. The small contribution from events with at least one boson decaying into tau pairs is also included in the signal. Simulated events are generated using pythia pythia () and passed through a detailed geant-based simulation geant () of the detector response. Differences between MC simulation and data in the reconstruction and identification efficiencies for electrons and muons are corrected using efficiencies derived from large data samples of inclusive (, or ) events. The systematic uncertainty in the signal is dominated by the uncertainty in the theoretical cross section (6.25%), the uncertainty on the lepton identification and reconstruction efficiencies (% for the and subchannels and % for the subchannels) and a 6.1% uncertainty on the luminosity measurement d0lumi (). Additional smaller sources of systematic uncertainty arise from energy and momentum resolutions and MC modeling of the signal process.
Backgrounds to the signal originate from top quark pair () production and from events with and/or bosons that decay to leptons and additional jets or photons. The jets can then be misidentified as leptons or contain true electrons or muons from in-flight decays of pions, kaons, or heavy-flavored hadrons.
The background from production is estimated from MC, assuming the cross section of pb ttbar_xsec () for a top quark mass of 170 GeV. The systematic uncertainty includes a 10% uncertainty on , as well as contributions from the variation in cross section and acceptance originating from uncertainties on the mass of the top quark.
To estimate the misidentified lepton background, we first measure the probability for a jet to produce an electron or muon that satisfy the identification criteria from data using a “tag and probe” method runiia_zz (). The probability for a jet to mimic an electron, parameterized in terms of jet and , is equal to for the case of CC electrons with a matched track and in the case of EC electrons for which no track matching is applied. The probability for a 15 GeV (100 GeV) jet to produce a muon of GeV is without requiring muon isolation, and it is when the muon is required to be isolated. A systematic uncertainty of 30% on the jet-to-lepton misidentification probabilities is estimated by varying the selection criteria of the control samples used.
The probabilities for jets to be misidentified as electrons are then applied to jets in +jets and +jets data to determine the background to the and channels, respectively. This method takes into account contributions from +jets, ++jets, +jets, +jets, +jets, and events with jets. However, this method double counts the contribution from +jets. A correction is measured, amounting to %, to correct for the double counting. The probabilities for jets to contain a muon are applied to jets in +jets data to determine a background estimate for the channel. Systematic uncertainties on this background arise from the 30% uncertainty in measured misidentification rates, and from the limited statistics of the data remaining in the samples after selection.
Table 1 summarizes the expected signal and background contributions to each subchannel, as well as the number of candidate events in data. The total signal and background expectations are events and events, respectively. We observe a total of three candidate events in the data, two in the subchannel and one in the subchannel. Table 2 summarizes some of their kinematic characteristics. The quoted dilepton masses in the table correspond to the one out of the three possible combinations having opposite charge within the pairs and having a dilepton mass closest to that of the boson. Figure 1 shows the distribution of the four lepton invariant mass for data and for the expected signal and background.
We extract the significance of the observed event distributions using a negative log-likelihood ratio (LLR) test-statistic collie (). As input, we use the expected yields (number of events) from signal and background, separated into the seven subchannels compared to the observed yields. The modified frequentist method returns the probability (-value) of the background-only fluctuating to give the observed yields or higher. In 5 background pseudo-experiments, we find 213 trials with an LLR value smaller or equal to that observed in data. This gives a -value of 4.310 which corresponds to a 5.3 standard deviation () observed significance ( expected). The probability for the signal plus background hypothesis to give less signal-like observations than the observed one is 0.87. A correction factor of 0.93, derived using pythia, is used to convert the measured cross section for into that for production. Minimizing the LLR function we obtain a cross section of pb for this analysis.
This result is combined with the results from an independent search llnunu (), and the previous analysis runiia_zz () which used a separate data sample with a looser mass requirement GeV. The earlier search contributes no signal events and we have scaled its background estimate to the tighter kinematic range used in the recent analysis. The combination of the three analyses is performed taking into account the correlations of systematic uncertainties between subchannels and among analyses. The resulting -value is 6.210, and the significance for observation of production increases to ( expected). The probability for the signal plus background hypothesis to give less signal-like observations than the observed one is 0.71. We therefore report the observation of a ZZ signal at a hadron collider with a combined cross section of pb, consistent with the standard model expectation.
We thank the staffs at Fermilab and collaborating institutions, and acknowledge support from the DOE and NSF (USA); CEA and CNRS/IN2P3 (France); FASI, Rosatom and RFBR (Russia); CNPq, FAPERJ, FAPESP and FUNDUNESP (Brazil); DAE and DST (India); Colciencias (Colombia); CONACyT (Mexico); KRF and KOSEF (Korea); CONICET and UBACyT (Argentina); FOM (The Netherlands); STFC (United Kingdom); MSMT and GACR (Czech Republic); CRC Program, CFI, NSERC and WestGrid Project (Canada); BMBF and DFG (Germany); SFI (Ireland); The Swedish Research Council (Sweden); CAS and CNSF (China); Alexander von Humboldt Foundation (Germany); and the Istituto Nazionale di Fisica Nucleare (Italy).
- (a) Visitor from Augustana College, Sioux Falls, SD, USA.
- (b) Visitor from The University of Liverpool, Liverpool, UK.
- (c) Visitor from INFN Torino, Torino, Italy.
- (d) Visitor from ECFM, Universidad Autonoma de Sinaloa, Culiacán, Mexico.
- (e) Visitor from II. Physikalisches Institut, Georg-August-University, Göttingen, Germany.
- (f) Visitor from Helsinki Institute of Physics, Helsinki, Finland.
- (g) Visitor from Universität Bern, Bern, Switzerland.
- (h) Visitor from Universität Zürich, Zürich, Switzerland.
- (‡) Deceased.
- (10) R.W. Brown and K.O. Mikaelian, Phys. Rev. D 19, 922 (1979).
- (11) U. Baur and D. Rainwater, Phys. Rev. D 62, 113011 (2000).
- (12) J. M. Campbell and R. K. Ellis, Phys. Rev. D 60, 113006 (1999).
- (13) R. Barate et al. [ALEPH Collaboration], Phys. Lett. B 469, 287 (1999); J. Abdallah et al. [DELPHI Collaboration], Eur. Phys. J. C 30, 447 (2003); M. Acciarri et al. [L3 Collaboration], Phys. Lett. B 465, 363 (1999); G. Abbiendi et al. [OPAL Collaboration], Eur. Phys. J. C 32, 303 (2003).
- (14) V. M. Abazov et al. [D0 Collaboration], Phys. Rev. Lett. 100, 131801 (2008).
- (15) T. Aaltonen et al. [CDF Collaboration], Phys. Rev. Lett. 100, 201801 (2008).
- (16) T. Andeen et al. , FERMILAB-TM-2365 (2007).
- (17) V. M. Abazov et al. [D0 Collaboration], Nucl. Instrum. Meth. Phys. Res. A 565, 463 (2006).
- (18) The D0 coordinate system is cylindrical with the -axis along the proton beamline and the polar and azimuthal angles denoted as and respectively. The pseudorapidity is defined as .
- (19) The variable between two objects and is defined as
- (20) T. Sjöstrand et al., Comput. Phys. Commun. 135, 238 (2001).
- (21) R. Brun and F. Carminati, CERN Program Library Long Writeup W5013, 1993 (unpublished).
- (22) N. Kidonakis and R. Vogt, Phys. Rev. D 68, 114014 (2003).
- (23) W. Fisher, FERMILAB-TM-2386-E (2007).
- (24) V. M. Abazov et al. (D0 Collaboration), Phys. Rev. D 78, 072002 (2008).