Search for t t Resonance in the Lepton+Jets Final State in p p Collisions at s = 1.96 TeV

Transcription

1 DØNote Search for t t Resonance in the Lepton+Jets Final State in p p Collisions at s =.9 TeV Thorsten Schliephake, Daniel Wicke Version.. CVS Revision :. (Dated: February, ) A search for a narrow-width tt resonance has been performed using data collected by the DØ detector. This analysis considers tt events in the lepton + jets channel using neural net tagger to identify b-jets and the tt invariant mass distribution to search for evidence of resonant production. The analyzed dataset corresponds to an integrated luminosity of. fb each in both the muon and the electron channel. The observed data are consistent with the background predictions for Standard Model tt production. Therefore, we set upper limits on σ X B(X tt) for different resonance masses using a Bayesian approach.

2 Contents I. Introduction A. Important changes since p7 II. Data Samples and Monte Carlo Simulation A. Data B. Monte Carlo. Resonance Signal. Background III. Object Identification A. b-tag IV. Event Selection A. Common Selection B. Electron Channel C. Muon Channel V. Reconstruction of the t t invariant mass VI. Background estimation A. Estimation of QCD background B. W +jets Background C. Top Background 9 D. Other backgrounds 9 VII. Systematic Uncertainties VIII. Limits 9 A. Limit Calculation 9. Signal Acceptance Correction 9 B. Expected Limits 9 C. Observed Results IX. Summary A. Simulation of Top Pair Resonace B. Control Plots 7 References 77

3 I. INTRODUCTION The top quark has by far the largest mass of all the known fermions. Heavy, yet unknown resonances may play a role in the production of top pairs and add a resonant part to the Standard Model mechanism. Such resonant production are suggested for massive Z-like bosons in extended gauge theories [], Kaluza Klein states of the gluon or Z [, ], axigluons [], topcolor [], and other beyond the Standard Model theories. Independent of the exact model a resonant production mechnism should be visible in the t t invariant mass distribution. In this analysis a model-independent search for a narrow-width heavy resonance X decaying into tt is performed. In the framework of the SM, the top quark decays into a W boson and b quark nearly % of the time. The tt event signature is fully determined by the W boson decay modes. In this analysis, we consider only the lepton+jets (l+jets, where l = e or µ) final state, which results from the leptonic decay of one of the W bosons and the hadronic decay of the other. The event signature is one isolated electron or muon with high transverse momentum (p T ), large transverse energy imbalance ( E T ) due to the undetected neutrino, and at least three jets, two of which result from the hadronization of b quarks. The analyzed dataset corresponds to an integrated luminosity of pb each in both the e+jets channel and the µ+jets channel, collected between August of and Juli 7. The signal-to-background ratio is improved by identifying b-jets using a neural network based b-tagging algorithm. After b-tagging, the dominant physics background for a resonance signal is non-resonant SM tt production. Smaller contributions arise from the direct production of W bosons in association with three or more jets (W +jets), as well as instrumental background originating from multijet processes with jets faking isolated leptons. The search for resonant production is performed by examining the reconstructed tt invariant mass distribution. This analysis has previously been performed in Run I by DØ [] and CDF [7]. For Run II, preliminary conference results exist for DØ [, 9] and CDF [, ] finding no evidence for a tt resonance. The methods applied closely resemble those of the RunIIa analysis currently in review [9]. In these studies a topcolor model was used as a reference to quote mass limits. In this model [] a large top quark mass can be generated through the formation of a dynamical tt condensate, X, which is formed by a new strong gauge force coupling preferentially to the third generation of fermions. In one particular model, topcolor-assisted technicolor [], X couples weakly and symmetrically to the first and second generations and strongly only to the third generations of quarks and has no couplings to leptons. This model results in a predicted cross section for tt production larger than the standard model (SM) prediction. Limits obtained on σ X B(X tt) were used to quote a limit on the mass of such a topcolor Z. The most recent results yield M Z > 9 GeV for the DØ analysis [9] and M Z > 7 GeV for the CDF analysis [, ], both for a resonance with Γ X =.M X. A. Important changes since p7 Since the p7 Analysis the following items have been replaced or added: Data is taken from RunIIb (p) Main backgrounds W +jets are now p alpgen samples Z signal sample is p Multijet/QCD background is taken from p Data with p ε QCD and ε Sig For b-tagging p TRF are used p Taggability-RF s are used In consotency with the new BID Taggaiblity-RF derivation a cpf cut for all jets has been applied requireing each jet to have cpf.7 or cpf=- [] Unless otherwise noted all available systematics have been updated to p

4 II. DATA SAMPLES AND MONTE CARLO SIMULATION A. Data The data sample consists of. pb each in the e+jets channel and in the µ+jets channel. A detailed explanation of the data quality requirements and used trigger lists can be found in Ref. []. B. Monte Carlo. Resonance Signal The resonance signal Monte Carlo is generated with alpgen [7] with the parton distribution functions, PDF, parametrised in CTEQL [] for twelve different resonance masses M X,,,,,,,,7, 7,,, and GeV using m t = 7 GeV. The width of the resonance is set to Γ X =.M X. This qualifies X as a narrow resonance since its width is smaller than the assumed mass resolution of the DØ detector. The width of.% was chosen in order to compare the results with Run I analyses where this width was used. The resonance generated with Pythia is forced to decay to tt only and the tt decays are inclusive. A list of the main parameters used in Pythia is provided in Appendix A. Only these events are processed through the full RunIIa detector simulation and event reconstruction. These semileptonic tt decays include the τ +jets channel and the τ can either decay leptonically or hadronically. Shape comparisons between tt and the different resonance samples can be found in Appendix B.. Background Background samples were produced for various background processes, passed through a full DØ detector simulation and overlaid with zero-bias data. The tt sample was generated with alpgen [7] with the PDF set cteql []. The factorisation scale was set to m t + p T (jets). A top mass of 7 GeV was used. W +jets samples were generated using m t = 7 GeV and alpgen [7] for the hard interaction and Pythia.9 for parton shower, hadronisation and hadron decays. The cteql PDF set was applied. To avoid double counting between the hard matrix element and the parton shower [9] the so called MLM jet-matching algorithm was applied []. Diboson background W W, W Z and ZZ samples were generated with Pythia. using cteql PDF set. Single top production for both s- and t-channel was generated using the COMPHEP single top generator []. Z+jets events were genereated using alpgen [7] for the hard interaction and Pythia. for parton shower, hadronisation and hadron decays. The main backgrounds tt and W +jets are taken from full RunIIb detector simulation. The minor backgrounds W W, W Z, ZZ, single top, and Z+jets are taken from the full RunIIa detector simulation. III. OBJECT IDENTIFICATION In order to isolate tt from background events, fundamental objects like primary vertex, electrons, muons, jets and missing transverse energy E T, need to be defined. These objects, their reconstruction and the respective selection criteria are described in detail in Ref. []. A. b-tag The usage of neural net information to identify b-jets improves the signal to background ratio. The description of the Neural Net Tagger (NN) and its performance is given in Ref. []. To determine the number of background events after b-tagging the efficiencies to have a certain number of tags have to be known. It is therefore necessary to know the efficiencies to tag a light jet, a c- or a b-jet. For this purpose Tag Rate Functions (TRFs) parametrized in E T and η of the jets are used in Monte Carlo []. In this analysis two separate NN working point may be used to select singly and doubly tagged events. First, doubly tagged events are chosen with a less stringent NN working point. From those events which are not selected as doubly

5 tagged, a possibly more stringent NN working point is applied to separate singly tagged events from no-tag events. The no-tag, single and double tag efficiencies are determined by evaluating the per jet tag efficiency Pf W (E T, η) for a given flavor (f) and working point (W = L, T ) and multiplying the efficiencies for all jets in the event: N jets Pevent(n tag = ) = Pf T j (E T j, η j ) N jets Pevent(>= tag L) = i,j j= N jets Pf L i Pf L j k i,j N jets i=,i j N jets ( Pj L ) + i,j,l ( P L f i (E T i, η i )). () P L f i P L f j P L f l N jets k i,j,l ( P L f j ) +... () P tag event(n = ) = P tag event(n = ) P tag event(n >= ) () For Eq. () terms for upto b-tags are implemented in the computation. Only good jets with p T > GeV and η <. are considered. IV. EVENT SELECTION The event selection aims to filter semi-leptonic top pair events. It requires a high p T tight electron or isolated muon in the e+jets and µ+jets channel, respectively, at least three jets and high missing E T. A. Common Selection The preselection for this analysis is based on the t t cross section measurement []. We require Good data quality At least three jets with p T > GeV, η <., and cpf>.7 or cpf= cpf is defined in terms of: where cpf = cpf(jet, pv) for the hardest pv []. trk ptrack T (jet j, vtx i ) cpf(jet j, vtx i ) = i trk ptrack T (jet j, vtx i ) cpf represents the energy (charged particle fraction) from the hard PV by the jet j. For small numbers the jet has a high probability to orignate from a MB PV. For numbers close to it has a high probability to be from the hard interaction. cpf =- means that there are no tracks assosiated to this jet. (for control plots events with one or two jets matching the above cuts are kept) A leading jet with p T > GeV Missing transvers energy E T > GeV and E T > GeV for e+jets and µ+jets, respectively. Good vertex with z PV cm with at least tracks attached () B. Electron Channel In addition to the above common selection the electron channel requires One tight electron wth p T > GeV in the CC. No second tight electron with p T > GeV in the CC or EC. No isolated muon with p T > GeV with η <.. Triangle cut of.7π. E T < dφ(e, E T )

6 C. Muon Channel In addition to the above common selection the muon channel requires One tight muon wth p T > GeV with η <.. Invariant mass of muon and any second muon m µµ < 7 GeV or m µµ > GeV. No second tight muon with p T > GeV with muon quality MediumNSeg. No tight electron with p T > GeV in the CC. Muon coming from the primary vertex: z(µ, PV) < cm Triangle cut of.. E T < dφ(µ, E T ) Compared to the RunIIa analysis a cut on the cpf of the jets was introduced to give a better description of the Lepton and Jet p T s. For all sanity plots shown in this analysis the Kolmogorov-Smirnov (KS) probability of the tt and background modeling compared to the data is computed. A discussions of these cuts can be found in Ref. [9, ]; Control plots for the preselection before and after b-tagging using the Medium working point of the NN tagger for Standard Model prediction can be found in Appendix B. V. RECONSTRUCTION OF THE t t INVARIANT MASS Various methods have been evaluated to compute the tt invariant mass, M tt from the reconstructed physics objects. In all cases the z-direction of the neutrino has to be inferred using constraints valid for general tt-events. The most direct method is to use the measured E T and lepton momentum. Requiring that the neutrino and lepton stem from the same W boson leads to a quadratic equation that can be solved for the z-component of the neutrino momentum. Of the two solutions the one with the lower absolute value is chosen. If no solution exists it is set to zero. To compute the invariant tt mass the neutrino and lepton momentum need to be added with the jet momenta. This procedure is applicable for events with any number of jets. We separate the data into those with exactly jets and those with or more jets. For events with more than jets we have investigated restricting the computation to the leading jets in p T (method a) as well as using all jets passing the preselection (method b). The performance of the different reconstruction methods can already be judged by comparing the shapes of the reconstructed M t t distribution in Standard Model and signal simulation that are shown in Fig..

7 7 -jet: Fraction/GeV SM tt X->tt M X =GeV Fraction/GeV..... SM tt X->tt M X =GeV Fraction/GeV..... SM tt X->tt M X =GeV M [GeV] tt M [GeV] tt M [GeV] tt jet: Fraction/GeV.. SM tt X->tt M X =GeV Fraction/GeV.. SM tt X->tt M X =GeV Fraction/GeV.. SM tt X->tt M X =GeV (a) M [GeV] tt M [GeV] tt M [GeV] tt (b) Fraction/GeV SM tt X->tt M X =GeV Fraction/GeV.... SM tt X->tt M X =GeV Fraction/GeV.... SM tt X->tt M X =GeV M [GeV] tt M [GeV] tt M [GeV] tt M X = GeV M X = GeV M X = TeV FIG. : Comparison of the shape of the reconstructed invariant t t mass between SM and resonant production. The shapes correspond to the combined lepton+jets channel with at least one b-tag. Each row corresponds to one of the reconstruction methods. The first row shows the direct reconstruction with jets. Below the jets results are shown: From top to bottom the direct reconstruction with leading jets (a) and with all jets (b). Within each line resonance masses of GeV, GeV and TeV(from left to right) are shown.

8 VI. BACKGROUND ESTIMATION The two main backgrounds common to both tt pair production and the resonance production are electroweak W boson production accompanied by four or more jets (W +jets) and QCD multi-jet events where a jet fakes an electron or a muon. Also, tt ll, single top production, diboson (W W,W Z and ZZ) as well as Z/γ production are considered as background sources in this analysis. A. Estimation of QCD background The overall normalization of the QCD background is estimated from data using the matrix method. Two samples of events, a loose and a tight set, are needed where the latter is a subset of the first. The loose set (N l ) corresponds to the preselected sample without the tight cut. The tight sample (N t ) requires in addition a tight muon isolation in the µ+jets channel and a likelihood cut in the e+jets channel. Using the number of events with a real lepton (W +jets, top, diboson or Z+jets), N W +top, and the number of events with a fake lepton, N QCD, the number of loose and tight events can be written as: Solving this linear system for N QCD and N W +top yields: N l = N W +top + N QCD N t = ε sig N W +top + ε QCD N QCD () N W +top = N t ε QCD N l ε sig ε QCD and N QCD = ε sign l N t ε sig ε QCD () The values for ε QCD and ε sig are shown in Table and the procedure to derive them can be found in []. The number of loose and tight events after requiring at least one b-tag can be found in Table. The predicted number of multijet events obtained from this method for the various channels is listed in line of Tables. The shape of the QCD background obtained from a bin by bin matrix method suffers from low statistics. Therefore only the total number of QCD events within each jet bin is taken from the matrix method and the shape is derived from the QCD events that pass the loose but not the tight preselection. B. W +jets Background The contribution of W +Jets events before b-tagging is calculated by subtracting all Monte Carlo backgrounds, the multijet contribution and the t t signal from the observed number of events. W +Jet events consist of W bb+jet, W cc+jet and W lp+jet events, with lp denoting light partons. The relative contribution of these three subclasses cannot purely be taken from the alpgen cross section as it does not take into account next-to-leading order corrections. Additionally to the relative fraction determined from Monte Carlo the heavy flavor contributions W bb+jet and before b-tagging with b-tag with b-tags Channel N l N t N l N t N l N t -jets e+jets 79 7 µ+jets -jets e+jets 7 7 µ+jets 7 TABLE : Observed data events for the loose and tight preselection before and after b-tagging as input for the matrix method. Numbers shown for the b-tagged samples correspond to the medium working point of the NN tagger. e+jets µ+jets ɛ QCD.9 ±.(stat) ±, 9(syst).7 ±.(stat) ±, (syst) ɛ sig (= jet). ±.(stat+syst).7 ±.(stat+syst) ɛ sig ( jets). ±.(stat+syst). ±.(stat+syst) TABLE : Efficiencies for the tight selection in both channels.

9 W cc+jet are multiplied by a heavy flavor scale factor of.7 ±. []. The number of W +jets events after b-tagging is calculated as the product of the number of untagged W +jets events and the probability P tag event to have a tag in the event computed from the TRFs. 9 C. Top Background The expected number of events after b-tagging are calculated assuming a Standard Model tt cross section of 7.9 ±.7 pb for the tt l+jets and tt ll backgrounds. This represents the currently best NNLO estimate on the tt cross section (c.f. Ref. []). MC-to-data correction factors [] were applied to the tt background as well as to the resonance signal. D. Other backgrounds To estimate the physics background the number of events after preselection is estimated from Monte Carlo calculations, taking into account the Monte Carlo selection efficiencies and tagging efficiencies. Diboson background: W W, W Z and ZZ samples generated with Pythia are used. The PDF set was CTEQL. The NLO cross section of.,. and. pb are taken respectively. According to [] the uncertainty on the cross section is chosen to be.% for all diboson samples. single top: The single top contribution from the s- and t-channel are used. The samples were generated with the Comphep single top Monte Carlo generator[7]. The cross section is. and.9 pb with.% uncertainty. The top quark mass was set to 7GeV. Z+Jets background: All different available Z+Jets [] samples simulated with alpgen [7] are taken into account. The Z p T reweighting described in [] is applied to increase the data/mc agreement of the Z distribution. To determine the contribution the Monte Carlo cross section is multiplied with a scale factor of. for light Z+Jets Monte Carlo and. consisting of. multiplied with a heavy flavor scale factor of. for heavy flavor Z+Jet samples[9]. As uncertainty on the heavy flavor scale factor % are assigned. On the cross section uncertainties of % for light Z+Jets and % for heavy flavor Z+Jets are used[]. All expected number of events before and after tagging (using the NN Medium and = or tags) can be found in Tables. The expected number of events for the different resonance samples are calculated with an assumed cross section of pb.

10 Yield table Contribution == Jets == Jets == Jets >= Jets Data Multijet 7. ± ± ± 9..9 ±.77 WW.9 ±.9.9 ±.. ±. 7. ±. WZ.7 ±..7 ±..7 ±..9 ±. Wbb. ±.9 7. ±.. ±.. ±. Wcc 9. ± ± ±.9. ±. Wlp 7. ±.7.7 ±.7.7 ±.79. ±. ZZ. ±.. ±.9. ±.7. ±. ZbbEE. ±.7.7 ±.9. ±.. ±. ZbbTauTau.9 ±..79 ±.. ±.. ±. ZccEE. ±.9. ±.9.7 ±.. ±. ZccTauTau.7 ±.. ±.7. ±..7 ±. ZlpEE 9. ±. 79. ±.. ±.. ±.9 ZlauTau 79. ±.. ±.. ±..7 ±.7 tb. ±.. ±.. ±.7. ±. tbq. ±..7 ±..9 ±.. ±. ttbar. ±..7 ±.. ±. 7. ±. sum mc 97. ± ±. 79. ±.. ±.7 Zprime. ±.. ±. 9.9 ±. 7. ±. Zprime. ±.9.7 ±.. ±. 9. ±. Zprime. ±..9 ±.. ±.. ±. Zprime.9 ±.7.9 ±.. ±.. ±. Zprime.9 ±.. ±..7 ±.. ±. Zprime. ±.9. ±..7 ±.7. ±.7 Zprime. ±..7 ±.. ±.. ±. Zprime7. ±..9 ±..99 ±.. ±.9 Zprime7.7 ±. 7. ±.. ±.9. ±. Zprime.7 ±.9 7. ± ±.. ±. Zprime. ±..9 ± ±. 7. ±. Zprime. ±.. ±..7 ±..7 ±. TABLE : Predicted and observed number of events in the e+jets channel before tagging. statistical only. Errors are

11 Yield table Data Multijet. ± ±.9.7 ±.9. ±.7 WW.9 ±. 79. ±.9. ±.7.7 ±. WZ. ±.. ±.. ±..7 ±. Wbb. ±.9.7 ±..9 ±.. ±.9 Wcc 7.99 ±..9 ± 7.7. ±..7 ±. Wlp.9 ± 7..9 ±. 99. ±.9. ±. ZZ.99 ±.9. ±.. ±.7. ±. ZbbMuMu 7. ±.. ±.. ±.7. ±. ZbbTauTau. ±.7. ±..7 ±..7 ±. ZccMuMu 9.9 ±.7. ±.. ±.7. ±. ZccTauTau.9 ±..7 ±.. ±..7 ±. ZlpMuMu 9. ±. 7. ±. 7.9 ±.9.7 ±. ZlauTau 9. ±.99. ±. 9.9 ±.. ±. tb. ±.7.9 ±..9 ±.7.7 ±. tbq 9. ±.. ±.. ±.. ±. ttbar. ±..9 ±.9.9 ±.7. ±. sum mc. ±.. ± 7.9. ±.. ±. Zprime. ±.7. ±.. ±..9 ±. Zprime. ±.. ±. 7. ±.. ±.7 Zprime.79 ±.7.7 ± ±. 7.7 ±. Zprime.7 ±..9 ±..7 ±.. ±. Zprime.79 ±.7. ±.7. ±..9 ±. Zprime.7 ±..7 ±.. ±. 9. ±. Zprime. ±.. ±. 7.9 ±. 9. ±. Zprime.77 ±.7.97 ±..9 ±.. ±. Zprime7. ±.. ± ±.. ±. Zprime. ±.. ±. 7. ±. 7. ±. Zprime. ±.. ±.7.7 ±..9 ±. Zprime. ±.. ±.. ±.9. ±. TABLE : Predicted and observed number of events in the µ+jets channel before tagging. Errors are statistical only.

12 Yield table Contribution == Jets == Jets == Jets >= Jets Data 7 Multijet. ±.. ±..7 ±.9 9. ±.9 WW. ±.7. ±..7 ±.7. ±. WZ. ±.. ±..9 ±..7 ±. Wbb. ±.. ±.. ±..9 ±. Wcc. ±. 9.9 ±.7. ±.77.9 ±.9 Wlp 7.9 ±.7 7. ±..9 ±.. ±. ZZ. ±.. ±.. ±..7 ±. ZbbEE.9 ±.. ±.. ±.. ±. ZbbTauTau.7 ±.. ±.7. ±.. ±. ZccEE. ±..77 ±.. ±.9. ±. ZccTauTau. ±.7.7 ±.9. ±.. ±. ZlpEE. ±.. ±.7. ±.. ±. ZlauTau.7 ±.. ±.. ±.. ±. tb. ±.. ±..97 ±..9 ±. tbq. ±.7.79 ±.. ±..9 ±. ttbar. ±.. ±.9 9. ±.. ±. sum mc 9. ±.7. ±.99. ±. 7.9 ±. Zprime.7 ±.. ±.. ±..9 ±. Zprime. ±.. ±.9. ±.. ±. Zprime. ±..7 ±.. ±.. ±. Zprime. ±.. ±.. ±.. ±. Zprime. ±..7 ±.. ±.. ±. Zprime. ±..9 ±.9. ±.. ±. Zprime. ±.. ±..9 ±.. ±. Zprime7. ±.. ±.. ±.. ±. Zprime7. ±.. ±..7 ±..7 ±. Zprime. ±..7 ±.9. ±..97 ±. Zprime. ±..7 ±.. ±.. ±. Zprime.9 ±.. ±.9.9 ±.. ±. TABLE : Predicted and observed number of events in the e+jets channel after tag with NN medium. Errors are statistical only.

13 Yield table Contribution == Jets == Jets >= Jets Data 7. Multijet. ±..9 ±.9. ±.9 WW. ±.. ±..9 ±. WZ. ±.. ±.. ±. Wbb 7. ±..7 ±.. ±. Wcc. ±.. ±.. ±. Wlp.9 ±.. ±.. ±. ZZ. ±.. ±.. ±. ZbbEE. ±.. ±.. ±. ZbbTauTau. ±.. ±.. ±. ZccEE. ±.. ±.. ±. ZccTauTau. ±.. ±.. ±. ZlpEE. ±.. ±.. ±. ZlauTau. ±.. ±.. ±. tb. ±.. ±..7 ±. tbq. ±.. ±.. ±. ttbar. ±.. ±. 9. ±. sum mc.9 ±. 9. ±.. ±.9 Zprime.77 ±..7 ±..7 ±.7 Zprime.79 ±.. ±.. ±. Zprime. ±.. ±.. ±. Zprime.7 ±.. ±..9 ±. Zprime.7 ±..9 ±..7 ±. Zprime. ±.. ±.. ±.7 Zprime.7 ±.. ±.. ±. Zprime7. ±.. ±..9 ±.7 Zprime7.7 ±.. ±..7 ±.7 Zprime.9 ±.. ±..7 ±. Zprime. ±..7 ±.. ±. Zprime. ±.. ±..9 ±. TABLE : Predicted and observed number of events in the e+jets channel after tags with NN medium. Errors are statistical only.

14 Yield table Contribution == Jets == Jets == Jets >= Jets Data 7 7 Multijet. ±.. ±.7. ±.. ±.9 WW.9 ±..9 ±.. ±..7 ±. WZ. ±.9.9 ±.. ±.. ±. Wbb 9. ± ±.99. ±.. ±. Wcc. ±..9 ±. 9. ±.9.9 ±.7 Wlp.7 ±..7 ±.9.7 ±..7 ±. ZZ. ±.. ±.. ±.. ±. ZbbMuMu. ±..7 ±.9. ±.. ±. ZbbTauTau.7 ±..9 ±.. ±.. ±. ZccMuMu. ±.9. ±.9. ±..7 ±. ZccTauTau. ±.. ±.. ±.. ±. ZlpMuMu. ±..97 ±.. ±.. ±. ZlauTau.9 ±.. ±.. ±.. ±. tb.7 ±.. ±.. ±.. ±. tbq. ±. 9. ±..97 ±..7 ±. ttbar. ±.9. ±.7. ±.. ±.9 sum mc. ±.. ±..99 ±.. ±.7 Zprime. ±.. ±.. ±..7 ±. Zprime. ±.. ±.7.7 ±..99 ±. Zprime. ±..9 ±.. ±..9 ±. Zprime. ±..7 ±..9 ±.. ±. Zprime.9 ±.. ±.. ±.. ±. Zprime.7 ±.. ±.7.79 ±.. ±. Zprime. ±.. ±..9 ±.. ±. Zprime7.9 ±..9 ±..7 ±.. ±. Zprime7. ±.. ±.. ±..7 ±. Zprime.7 ±.. ±..9 ±..9 ±. Zprime. ±..7 ±..97 ±.. ±. Zprime. ±.. ±.7.7 ±.9. ±.9 TABLE 7: Predicted and observed number of events in the µ+jets channel after tag with NN medium. Errors are statistical only.

15 Yield table Contribution == Jets == Jets >= Jets Data 9 Multijet.9 ±.. ±.. ±.9 WW. ±.. ±.. ±. WZ. ±.. ±.. ±. Wbb 7.9 ±..77 ±.. ±. Wcc. ±.9.9 ±.. ±. Wlp. ±.. ±.. ±. ZZ. ±.. ±.. ±. ZbbMuMu.7 ±.. ±.. ±. ZbbTauTau. ±.. ±.. ±. ZccMuMu. ±..9 ±.. ±. ZccTauTau. ±.. ±.. ±. ZlpMuMu. ±.. ±.. ±. ZlauTau. ±.. ±.. ±. tb.99 ±.. ±.. ±. tbq. ±.. ±.. ±. ttbar. ±.9. ±. 7. ±. sum mc. ±.9. ±.. ±. Zprime. ±.. ±.. ±. Zprime. ±..9 ±.. ±. Zprime. ±..9 ±..7 ±. Zprime. ±.. ±.. ±. Zprime. ±.. ±.. ±. Zprime.7 ±.. ±..99 ±. Zprime. ±..7 ±..97 ±. Zprime7. ±.. ±..79 ±. Zprime7. ±.. ±.. ±. Zprime. ±..7 ±.. ±. Zprime. ±..9 ±.. ±. Zprime.7 ±..7 ±.. ±. TABLE : Predicted and observed number of events in the µ+jets channel after tags with NN medium. Errors are statistical only.

16 VII. SYSTEMATIC UNCERTAINTIES The analysis presented here relies on the shapes and the overall normalization. The systematics which only change the normalisation include the uncertainies on the MC-to-data correction factors, the cross sections and luminosity. Shape changing systematics are the jet energy scale and those related to b-tagging, among others. The following systematic effects of this kind have been studied on all samples including the signal: Uncertainty associated to the jet energy scale (JES): The uncertainty on the Jet Energy Scale (JES) is determined by changing the JES correction up and down by one standard deviation. The impact on the preselection efficiency is taken into account. Systematics on jet energy resolution: The uncertainty on the Jet Energy Resolution (JER) is determined by changing the JER correction up and down by one standard deviation. The impact on the preselection efficiency is taken into account. Uncertainty associated to the jet reconstruction and identification efficiency The uncertainty on the Jet Reconstruction and Jet Identification Efficiency (JetID) is determined by changing the JetID down by one standard deviation. The impact on the preselection efficiency is taken into account. The upward correction is obtained by using an opposite shift of the same size in each bin of the final M t t distributions. Uncertainty associated with the tt cross section: The uncertainty of 7.9 ±.7 pb on the tt cross section is taken into account. Uncertainty associated with the b-trfs: The uncertainty on the b-trf is calculated by raising and lowering the efficiencies associated with the b-trf by one standard deviation. Uncertainty associated with the c-trfs: The uncertainty on the c-trf is calculated by raising and lowering the efficiencies associated with the c-trf by one standard deviation. Uncertainty associated with the light-trfs: The uncertainty on the light-trf is calculated by raising and lowering the efficiencies associated with the light-trf by one standard deviation. Uncertainty associated with the Taggability Rate Function: The uncertainty on the Taggability Rate Function is calculated by raising and lowering the efficiencies associated with the Taggability by one standard deviation. Uncertainty associated with the Heavy flavor scale factor on W +Jets: The uncertainty on the W +Jets heavy flavor scale factor is.%, according to []. Uncertainty associated with the b-fragmentation reweighting: The systematics on the reweighting of the b-fragmentation function from default in Pythia to the tuned value was assumed to be the difference from the AOD tune to the SLD tune []. Therefore only a one sided systematic is assigned. Uncertainty associated with ɛ qcd : The uncertainty of ɛ qcd on the multijet yield calculated with the Matrix Method is taken into account. Top mass uncertainty: The uncertainty due to the top mass uncertainty is estimated by modifying the m t for the simulation of the dominating t t l+jets background. Samples with m t = GeV and m t = 7 GeV were used to estimate a σ variation. This is then scaled to obtain a one standard deviation uncertainty. Luminosity uncertainty: As uncertainty of the measured luminosity ±.%was taken into account []. Electron ID scale factor systematics. A systematic uncertainty of. % is used. This uncertainty includes the dependence of the electron ID scale factor on the variables ignored in the parameterization. We estimate the contributions to the electron ID scale factor systematics to come fr om the jet multiplicity dependence (.%) of the H-matrix, track match

17 and likelihood scale factor and from dependences not taken into account (on p T and φ of the em object) (.%). We estimate the uncertainty due to the jet multiplicity dependence by calculating the difference between the efficiency with taking into account the dependence on R between electron and a jet and without that dependence, which results in.%. Muon ID and track scale factor systematics. According to the studies summarized in the muon ID certification note [] uncertainties on muon ID and tracking are both.7 %. Muon isolation scale factor systematics. An uncertainty on the efficiency of the muon isolation criteria is extracted from [] to be %. Z vertex distribution difference between data and MC. An uncertainty of. % on the difference between the z vertex distribution in simulated events and the data is used, as given in []. PV scale factor. The preselection efficiency is scaled with 9.9 % to correct for the difference of the primary vertex selection efficiency between data and Monte Carlo []. A relative uncertainty on this scale factor is estimated to be. %. In Tables 9 and, the relative change of the total number of predicted events in the Standard Model background is listed for all systematics. The numbers listed are for e+jets and µ+jets channels combined and reflect the influence of the systematic uncertainties on the estimated number of background events. The impact on the normalization and in case of JES and others also the differences in preselection efficiencies are taken into account. A comparison of the nominal M tt distributions to +σ and σ jet energy scale systematic variations are shown in Figure for the dominant tt background. 7 invariant mass of the ttbar system without hitfit (all jets).. ttbar central ttbar JES_plus invariant mass of the ttbar system without hitfit (all jets).. ttbar central ttbar JES_minus..... M tt [GeV]..... M tt [GeV] invariant mass of the ttbar system without hitfit (lead jets).. ttbar central ttbar JES_plus invariant mass of the ttbar system without hitfit (lead jets).. ttbar central ttbar JES_minus..... M tt [GeV]..... M tt [GeV] FIG. : Shape comparison of the M tt distributions of the nominal, +σ (right), and σ jet energy scale systematic (left). The upper row corresponds to jet events the lower row corresponds to or more jet events.

18 rel. sys. uncertainty (%) source Standard Model processes (%) Resonance M X = 7 GeV jets and more jets jets and more jets σ + σ σ + σ σ + σ σ + σ Jet energy scale.%.%.7%.%.%.7%.%.7% Jet energy resolution.%.%.%.%.%.%.%.% Jet identifcation.7%.7%.%.%.%.%.7%.7% σ t t(m t =7 GeV).%.%.%.9% b tag TRF.7%.7%.%.%.9%.99%.7%.% c tag TRF.%.%.9%.%.99%.99%.7%.7% light tag TRF.7%.%.%.9%.99%.99%.%.% Taggability.7%.%.%.%.99%.9%.7%.7% Heavy flavor scale factor for W +jets.%.9%.%.7% b fragmentation.%.%.9%.9%.%.%.7%.7% Multijet lepton fake rate.7%.9%.9%.7% Luminosity.%.%.%.%.9%.7%.%.7% Top mass variation.%.%.%.% Selection efficiencies (incl. leptonid).%.9%.9%.%.9%.9%.7%.% TABLE 9: Summary of the relative systematic change on the overall normalization of the Standard Model background and for an resonance mass of M X = 7 GeV with b-tag. rel. sys. uncertainty (%) source Standard Model processes (%) Resonance M X = 7 GeV jets and more jets jets and more jets σ + σ σ + σ σ + σ σ + σ Jet energy scale.7%.7%.%.79%.%.%.%.% Jet energy resolution.7%.%.%.% Jet identifcation.%.%.7%.7%.%.%.9%.9% σ t t(m t =7 GeV).9%.% 7.% 7.7% b tag TRF.%.9%.%.%.7%.%.%.% c tag TRF.%.77%.%.%.%.97%.%.9% light tag TRF.%.9%.%.%.%.%.%.% Taggability.%.9%.%.%.77%.77%.%.% Heavy flavor scale factor for W +jets.%.%.%.9% b fragmentation.%.%.%.%.%.%.%.% Multijet lepton fake rate.%.%.%.% Luminosity.%.%.7%.97%.%.7%.%.% Top mass variation.9%.%.%.% Selection efficiencies (incl. leptonid).%.%.%.%.%.%.%.% TABLE : Summary of the relative systematic change on the overall normalization of the Standard Model background and for an resonance mass of M X = 7 GeV with b-tags.

19 9 VIII. LIMITS A. Limit Calculation The tt invariant mass, as shown in Fig. and, is used to perform a binned likelihood fit of the signal and background expectations compared to data. The backgrounds, including Standard Model tt production, are normalized to the predictions, as described in Sec. VI. In case of tt, σ tt = 7.9 pb is assumed. A Bayesian approach leads to upper limits on σ X B(X tt) for the several different resonance masses. A Poisson distribution is assumed for the number of observed events in each bin, as well as flat prior probabilities for the signal cross sections. The prior for the combined signal acceptance and background yields is a multivariate Gaussian with uncertainties and correlations described by a covariance matrix. More information about the limit calculation can be found in the single top analyses Ref. [] and Ref. [], and the CVS package top statistics.. Signal Acceptance Correction No tt resonance contributions were assumed when estimating the W +jets background. But when calculating the limits, the presence of the resonance has to be taken into account, since in this case the number of estimated W +jets events would be reduced. This is due to the fact that in addition to the backgrounds normalised to their cross-section (SM top, Diboson, Z+jets) also the expected resonance events have to be subtracted from N untagged W +top obtained from the Matrix Method in Eq. (). To correctly treat the W +jets background in the presence of a resonance the following considerations were made: The untagged data sample consists of N untagged untagged W +top+x and NQCD : N untag data = N untag W +top+x + N untag QCD (7) N untag W +top+x is the sum of the W +jets, other SM backgrounds with real leptons (dominated by SM top, but including Diboson and Z+jets) and resonance contributions. So the W +jets contribution is calculated by: This leads to: N untag data N untag W +jets = N untag W +top+x N untag top N untag X () = (N untag W +top+x N untag top N untag X ) + N untag top + N untag X + N untag QCD (9) After applying the b-tagging, the number of untagged events for each sample is multiplied with the b-tag efficiency in that particular sample, which is then the number of events in the tagged sample. N tag data = εw b (N untag W +top+x N untag top N untag X Solving this equation for the number of resonance events N X gives: ) + ε top b N untag top + ε X b N untag X + ε QCD b N untag QCD () (ε X b ε W b )N untag X = N tag data εtop b N untag top ε W b (N untag W +top+x N untag top ) ε QCD b N untag QCD () Thus, instead of correcting the number of W +jets events for the limit calculation, the resonance input can be corrected for the b-tag efficiency in a W +jets sample. This was done for all resonance samples and also for all systematics which potentially change the b-tag efficiency. B. Expected Limits The invariant mass distribution for the and jet bin are shown in Fig. and Fig.. Attributing the differnce between expected and observed to ttbar only and scaling the nominal sigma ttbar by their ratio yields observed cross-sections of 7.pb and.pb in the - and -jet bins, respectively.

20 # tagged events - L =. fb KS =.. data 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet M [GeV] tt FIG. : Distribution of the t t invariant mass reconstructed from -jet events. Combined e+jets and µ+jets data (dots) with at least NN Medium b-tag are compared to the Monte Carlo expectation (histograms). # tagged events - L =. fb KS =.9. data.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet M [GeV] tt FIG. : Distribution of the t t invariant mass reconstructed from -jet events. Combined e+jets and µ+jets data (dots) with at least NN Medium b-tag are compared to the Monte Carlo expectation (histograms) using direct reconstruction of the neutrino momentum and include the four leading jets to compute the invariant mass. Expected limits with statistical errors only as well as including full systematics were computed by using the SM Monte Carlo expectation as input data in the limit calculation. They were computed for the various analysis options. The option to be applied to real data was chosen based on the expected performance. For consistency with other l+jets analyses the Medium working point was chosen for further analysis. Expected limits from statistical errors only were also to reconstruct the invariant top pair mass, M t t. Finally we note that among the methods with the direct reconstuction the curves show a slight preference for including only the leading jets. In computing the limit the distributions for and or more tags can be handled as separate channels or as a combined one of more tag channel. The two methods give almost identical results as was shown in [9]. Following the results of the comparisons made in this section the final result will be computed using the Medium b-tag working point and the direct M t t reconstruction. For events with more than jets only the leading jets p T are used.

21 σ X *B(X->tt) [pb].. SM expected limit 9% CL Observed limit 9% CL Topcolor Z (CTEQL). - L =. fb [GeV] M X FIG. : The expected and observed 9% confidence level upper limits on σ X B(X tt) as a function of resonance mass M X. C. Observed Results After all selection cuts events remain in the e+jets channel and 77 events in the µ+jets channel. Invariant mass distributions are computed for events with exactly one b-tag and for those with more that one b-tag. Additionally, the distributions are separated into jet and or more jet events. Corresponding plots for the individual channels can be found in Figures and in Appendix B. The expected and observed 9% C.L. upper limits on σ X B(X tt) as a function of M X are summarized in Table and displayed in Fig.. This figure also includes the predicted σ X B(X tt) for a leptophobic Z boson with Γ Z =.M Z computed using CTEQL parton distribution function which, combined with the experimental limits, excludes M Z < GeV at 9% C.L. The expected cross-section limits as a function of M X have improved with respect to a previous analysis performed with 9 pb [9] which gave an expected limit to the Z mass of 7 GeV. With combining the -jet and -jet data the expected limit on the Z mass is 79 GeV. In Figures 7 the posterior probability densities are shown as a function of σ X B(X tt) for the expected and observed limits. For illustration purposes, the tt invariant mass distribution is shown with an assumed resonance contribution where the resonance cross section is set to the observed limit including systematics. Using a model of topcolor assisted technicolor, the existence of a leptophobic Z boson with M Z < GeV and width Γ Z =.M Z can be excluded at 9% C.L. In Figure, the observed limits at 9% C.L. were already shown together with the predicted cross section of a topcolor Z boson for a width of Γ Z =.M Z. statistical only including systematics M X [GeV] exp. limit [pb] obs. limit [pb] exp. limit [pb] obs. limit [pb] TABLE : Expected and observed limits for σ X B(X tt) at the 9% confidence level when combining all channels with and without systematic uncertainties taken into account.

22 σ X *B(X->tt) [pb].. SM expected limit 9% CL SM expected limit % CL Observed limit 9% CL Observed X cross-section Topcolor Z (CTEQL). - L =. fb [GeV] M X FIG. : The expected % and 9% confidence level upper limits on σ X B(X tt) (shaded bands) as a function of resonance mass M X compared to observed best resonance cross-section with uncertainty and exclusion limits at 9%CL.

23 ] ] ] ] ] ] - Posterior Prob. Density [pb.. σx 9 <. pb σx 9 exp <. pb M X = GeV limit exp. limit # tagged events - L =. fb KS =.9. data. Zprime.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet # tagged events - L =. fb KS =.. data. Zprime 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet..... σ X *BR [pb] M [GeV] tt M [GeV] tt - Posterior Prob. Density [pb.... σx 9 exp σx 9 <. pb <. pb M X = GeV limit exp. limit # tagged events - L =. fb KS =.9. data 7. Zprime.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet # tagged events - L =. fb KS =.. data. Zprime 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet..... σ X *BR [pb] M [GeV] tt M [GeV] tt - Posterior Prob. Density [pb. σx 9 exp <.7 pb M X = GeV limit exp. limit σx σ X *BR [pb] # tagged events <. pb - L =. fb KS =.9. data. Zprime.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet M [GeV] tt # tagged events - L =. fb KS =.. data 9. Zprime 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet M [GeV] tt - Posterior Prob. Density [pb. σx 9 exp <. pb M X = GeV limit exp. limit σx 9... σ X *BR [pb] # tagged events - L =. fb KS =.9. data. Zprime.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet M [GeV] tt # tagged events - L =. fb KS =.. data. Zprime 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet M [GeV] tt - Posterior Prob. Density [pb.. σx 9 exp <.9 pb σx 9 M X = GeV limit exp. limit <.9 pb # tagged events - L =. fb KS =.. data 9. Zprime.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Wbb. Wcc. Wlp. Multijet # tagged events - L =. fb KS =.. data. Zprime 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb 9. Wbb. Wcc 7. Wlp.7 Multijet... σ X *BR [pb] M [GeV] tt M [GeV] tt - Posterior Prob. Density [pb.. σx 9 exp <.7 pb σx 9 <. pb M X = GeV limit exp. limit # tagged events - L =. fb KS =.9. data 9. Zprime.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet # tagged events - L =. fb KS =.. data 7. Zprime 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet... σ X *BR [pb] M [GeV] tt M [GeV] tt FIG. 7: In the left plots, the posterior probability densities are shown as a function of σ X B(X tt). On the right side, M tt is shown with an assumed resonance contribution in the or more jet distribtion (middle) and the jet distribution (right). Each row corresponds to a different resonance mass assumption, starting from GeV (top) to GeV (bottom).

24 ] ] ] ] ] ] - Posterior Prob. Density [pb σx 9 exp σx 9 <. pb <.9 pb M X = GeV limit exp. limit # tagged events - L =. fb KS =.9. data. Zprime.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet # tagged events - L =. fb KS =.. data.9 Zprime 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet... σ X *BR [pb] M [GeV] tt M [GeV] tt - Posterior Prob. Density [pb σx 9 <. pb σx 9 exp <.7 pb M X =7 GeV limit exp. limit # tagged events - L =. fb KS =.9. data 9.7 Zprime7.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet # tagged events - L =. fb KS =.. data.7 Zprime7 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet... σ X *BR [pb] M [GeV] tt M [GeV] tt - Posterior Prob. Density [pb σx 9 <.7 pb σx 9 exp <. pb M X =7 GeV limit exp. limit # tagged events - L =. fb KS =.9. data.7 Zprime7.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet # tagged events - L =. fb KS =.. data. Zprime7 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet... σ X *BR [pb] M [GeV] tt M [GeV] tt - Posterior Prob. Density [pb σx 9 <.9 pb σx 9 exp <. pb M X = GeV limit exp. limit # tagged events - L =. fb KS =.9. data. Zprime.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet # tagged events - L =. fb KS =.. data. Zprime 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet... σ X *BR [pb] M [GeV] tt M [GeV] tt - Posterior Prob. Density [pb σx 9 <. pb σx 9 exp <. pb M X = GeV limit exp. limit # tagged events - L =. fb KS =.9. data.7 Zprime.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet # tagged events - L =. fb KS =.. data.9 Zprime 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet... σ X *BR [pb] M [GeV] tt M [GeV] tt - Posterior Prob. Density [pb σx 9 <. pb σx 9 exp <. pb M X = GeV limit exp. limit # tagged events - L =. fb KS =.9. data.7 Zprime.7 ttljet. ZZ.7 WZ.7 WW. tbq. tb. Z+jets. Wbb. Wcc. Wlp. Multijet # tagged events - L =. fb KS =.. data. Zprime 9. ttljet. ZZ.7 WZ 7. WW. tbq. tb.9 Z+jets 9. Wbb. Wcc 7. Wlp.7 Multijet... σ X *BR [pb] M [GeV] tt M [GeV] tt FIG. : In the left plots, the posterior probability densities are shown as a function of σ X B(X tt). On the right side, M tt is shown with an assumed resonance contribution in the or more jet distribtion (middle) and the jet distribution (right). Each row corresponds to a different resonance mass assumption, starting from GeV (top) to GeV (bottom).

25 IX. SUMMARY A search for a narrow width resonance in the l+jets final states has been performed using data corresponding to an integrated luminosity of about. fb, collected with the DØdetector during Run IIb of the Tevatron collider. By analyzing the reconstructed tt invariant mass distribution assuming a cross section of 7.9 pb and using a Bayesian method, model independent upper limits on σ X B(X tt) have been obtained for different hypothesized masses of a narrow-width heavy resonance decaying into tt. Within a topcolor-assisted technicolor model, the existence of a leptophobic Z boson with M Z < GeV and width Γ Z =.M Z can be excluded at 9% C.L..

26 APPENDIX A: SIMULATION OF TOP PAIR RESONACE Top pair resonance simulation X tt were produced in D MC requests,,,,, 7,, 7, 7, 7,, and. The requests required Pythia from drelease p.9., which contains Pythia v.. For each of the resonance masses individual cardfiles are needed to maintain the desired width of.m X. The requests asked for cardfileversion v-9-. For M X = GeV/c and Γ X =. M X the following parameters were used: MSTP() was set to use the LHPDFCTEQL parton distribution functions. MSTP() = PMAS(, ) = 7 : this sets the top mass to 7 GeV/c MSEL = : this selects the process f i fi γ/z /X MSTP() = : only the Z boson is considered in the matrix elements,no interference terms. CKIN() =, CKIN() =, CKIN() =, CKIN() = : the invariant mass of tt varies between GeV/c and p t <. GeV CKIN(9) =, CKIN() =, CKIN() =, CKIN() = : η ranges for process and for decay products. PMAS(, ) = : this sets the mass of the X boson to GeV/c MDME(9, ) = MDME(9, ) =, MDME(9, ) = MDME(, ) =, MDME(9, ) = : only tt decay mode is allowed for the X boson Vector and axial couplings of the X boson to first generation quark and leptons PARU() =., PARU() =., PARU() =., PARU() =., PARU() =., PARU() =., PARU(7) =., PARU() =. Vector and axial couplings of the X boson to second generation quark and leptons PARJ() =., PARJ() =., PARJ() =., PARJ() =., PARJ() =., PARJ() =., PARJ() =., PARJ(7) =. Vector and axial couplings of the X boson to third generation quark and leptons: PARJ() =., PARJ(9) =., PARJ(9) =., PARJ(9) =., PARJ(9) =., PARJ(9) =., PARJ(9) =., PARJ(9) =. The settings for the vector and axial couplings are needed to set the width of the X boson to the desired value of Γ X =.M X since the direct way of setting the width in Pythia doesn t work because one needs to include the full interference structure. In Pythia, the default setting for the couplings of the X boson to quarks and leptons is the same as that for the Standard Model Z boson.

27 7 As a sanity check the following control plots are shown: APPENDIX B: CONTROL PLOTS Figures 9 show a comparison between tt and resonances of GeV, GeV, and GeV. Figures - show data-mc comparisons after all preselection cuts except for the b-tag. Figures - show data-mc comparisons of jet events after all preselection cuts including b-tag using the Medium working point of the NN tagger. Figures - show data-mc comparisons of jet eventsafter all preselection cuts including or more b-tags using the Medium working point of the NN tagger. Figures 7 - show data-mc comparisons of or more jet events after all preselection cuts including b-tag using the Medium working point of the NN tagger. Figures - show data-mc comparisons of or more jet events after all preselection cuts including or more b-tags using the Medium working point of the NN tagger. Figure shows the tt invariant mass distributions for or more jet events in the e+jets and µ+jets channels. Figure shows the tt invariant mass distributions for or more jet events in the e+jets and µ+jets channels. Figures - show data-mc comparisons after all preselection cuts except for the b-tag for jet events. Figures - show data-mc comparisons after all preselection cuts except for the b-tag for jet events. Figures - show data-mc comparisons of jet events after all preselection cuts including b-tag using the Medium working point of the NN tagger. Figures - show data-mc comparisons of jet events after all preselection cuts including b-tag using the Medium working point of the NN tagger. Figures 7-9 show data-mc comparisons of jet events after all preselection cuts including or more b-tags using the Medium working point of the NN tagger. Figures - show data-mc comparisons of jet events after all preselection cuts besides cpf and b-tag for jets. Figures - show data-mc comparisons of jet events after all preselection cuts besides cpf and b-tag for jets. Tables and show the Cutflow for Data in, channel

28 p Zprime for jet bin p Zprime Aplanarity for jet bin 9 7 p Zprime missing transverse energy for jet bin p Zprime p Zprime Leading jet for jet bin Lepton for jet bin p Zprime for jet bin p Zprime Aplanarity for jet bin 9 7 p Zprime missing transverse energy for jet bin p Zprime p Zprime Leading jet for jet bin Lepton for jet bin p Zprime number of jets FIG. 9: Shape comparison between tt and a resonance of GeV in the channel. The first rows correspond to or more jets the second rows to jet events.

29 9 p Zprime for jet bin p Zprime Aplanarity for jet bin p Zprime missing transverse energy for jet bin p Zprime p Zprime Leading jet for jet bin Lepton for jet bin p Zprime p Zprime p Zprime for jet bin Aplanarity for jet bin missing transverse energy for jet bin p Zprime p Zprime Leading jet for jet bin Lepton for jet bin p Zprime number of jets FIG. : Shape comparison between tt and a resonance of GeV in the channel. The first rows correspond to or more jets the second rows to jet events.