EPJdirect C3, 117 (2001) DOI 10.

1007/s1010501c0003

EPJdirect

electronic only c Springer-Verlag 2001

Scientic Note:

Search for Z WW at LHC with the ATLAS detector

D. Benchekroun, C. Driouichi, A. Hoummada
Universit Hassan II, e Facult des sciences A Chock, Casablanca, Morocco e n Received: 22 Jan 2001 / Accepted: 2 Mar 2001 / Published online: 20 Mar 2001 Abstract. The potential for measuring the Z WW coupling with the ATLAS detector at the LHC is investigated under conditions of low and high luminosity. It is estimated that detection of the decay channel Z WW ll jj is possible with 300 f b1 , for a Z boson with sequential model couplings and with a mass up to 2.2 TeV.

PACS: not given

Introduction

One of the most important goals of existing and future colliders is to establish the model which fully describes the strong and electroweak interactions. For several years, the experimental data have been in remarkable agreement with the predictions of the Standard Model (SM) [1]. However many fundamental questions have not yet been satisfactorily answered. For instance, there are many arbitrary parameters. Parity violation in the electroweak interactions is introduced in an ad hoc manner by using isospin singlets to describe the right-handed fermions. Recent experimental data give clear evidence that the neutrinos, which are considered as massless particles in the SM, should have very small masses [2]. Several scenarios have been proposed over the last 25 years in which the SM is just an eective low energy version of a somewhat more complex gauge structure which exists at higher energies. If the energy scale of this new physics is not far from the multi-TeV range, the experiments at the Large Hadron Collider (LHC) should nd direct evidence of its existence. The existence of new electroweak gauge bosons, called Z and W , is predicted by many extensions of the SM. Their discovery would be a clean signature for new physics beyond the SM. The largest set of theories with new gauge bosons are those which are based on a GUT group, such as SO(10) and E6 , like the E6 eective rank-5 model (ER5M) which predicts a Z whose couplings depend on a single parameter /2 /2 [3]. An attractive extension of the SM is the Left-Right Symmetric Model (LRM) [4] which extends the SM group SU (2)L U (1) to SU (2)L SU (2)R U (1) and restores parity symmetry at high energy. This model requires also the existence of right-handed neutrinos
http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

and has a free parameter 0.55 K 2 = (gR /gL )2 1 to 2 [5], which is just the ratio of the gauge couplings. The case K = 1 corresponds to a manifest left-right symmetric gauge interaction (Left-Right Symetric Model: LRSM). The search for new gauge bosons and right-handed Majorana neutrinos, in the framework of the LRSM, at LHC with the ATLAS detector has been reported in Ref. [6]. In the Sequential Standard Model (SSM), which will be used as a reference model in this work, the new Z and W are just heavy versions of the SM gauge bosons. The SSM is not a true model (it is not a gauge invariant model) but is commonly used as a guide by experimentalists. Up to now, no evidence has been experimentally found for the production of W and Z . The search for new gauge bosons at existing colliders gives only a limit on the masses. At the Tevatron, direct searches of Z in the channels Z e+ e and Z + led to mZ 690 GeV /c2 [7]. The searches at Tevatron of a W boson decaying leptonically report a limit of mW 720 GeV /c2 if the right-handed neutrinos are light and mW 650 GeV /c2 if they are heavy [8]. The indirect limit on the W mass is given by the analysis of the 0 0 contribution of the right-handed currents to the KL - KS mass dierence which 2 leads to a limit of 1.6 T eV /c [9]. Indirect searches of new gauge bosons are also performed at lepton colliders where the indirect discovery limits are related to the deviations from SM predictions for various cross sections and asymmetries due to the interference between the Z propagator and the and Z 0 propagators [10]. Results from LEP2 ( s = 200 GeV) yield a limit of 1 T eV /c2 for a SSM Z boson [11]. The discovery of the Z boson would constrain models of electroweak symmetry breaking. The decay rate Z W + W is sensitive to the Z - Z mixing angle [12]. The mixing angle is small in models where the Z WW coupling is strongly suppressed, and large in models with a strongly interacting Higgs sector [13] where the new gauge bosons are considered as resonances in the WW or WZ channels and the production amplitude would be of the second electroweak order as for the WW or WZ production. In the present paper, we examine the feasibility of detecting the Z boson via the decay Z W W , which is not the principal discovery channel, but can help to understand the origin of the new gauge bosons. In Sect. 2, we describe the characteristics of the production and decays of the Z boson within a reference model (SSM) at LHC. In Sect. 3 we describe features of the performance of the ATLAS detector which are important for the signal observability in this channel. Sect. 4 is dedicated to the details of the performed analysis and gives nal estimates on the discovery potential of Z boson in this channel. We give a summary and conclusions in the last section.

Production and decays of a heavy Z boson

In this section, we give the main properties of the dierent decay modes of a heavy Z boson. The usefulness of each decay mode will be also discussed. In the following, we consider that the Z boson is allowed to decay into a pair of jets, leptons and SM bosons. The branching ratios of these decay channels will be estimated in a SM-like situation (SSM).
http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

In the reference model described in [14], the couplings of the Z boson to quarks, leptons and W bosons are a direct transcription of the corresponding standard coupling. In particular, the decay widths into a WW and fermion pair are given in this model by: (Z W W ) (Z fi fi ) = NC where: NC is a colour factor (3 for quarks and 1 for leptons). = e2 /4 and cosW = mW /mZ . Qf is the fermion electric charge. An important aspect of equation (1) is the increase of WW width as m5 which Z leads to a rapidly increasing branching ratio of the WW channel with mZ . However, in a large part of the extended gauge models, while the couplings to fermions are not much dierent from the SM ones, the Z WW coupling is substantially suppressed with respect to the SM. The mixing between the new Z boson and the SM vector boson is typically of the order of m2 /m2 . Thus, Z Z a signicant modication of the reference model is the introduction of a mixing factor 2 in Eq. (1), given by = c m2 Z m2 Z (3) cotgW mZ 48 mZ mW
4

(1) (2)

mZ [1 + (1 4|Qf |sin2 W )2 ] 12 sin2 W cos2 W

leading hence to a linear increase of the partial width (Z W W ) with mZ . The branching ratios of the dierent decay channels do not change signicantly with the Z mass (4) cos3 W sin1 W mZ 48 The Z total width can be approximated by: Z (GeV) = 2.2 GeV + 0.032 * (MZ MZ ) In the case of SSM, the factor c is chosen to be unity [14]. The SSM is a very simple schematization, which cannot cover the wide variety of the SM extensions, but is often used for purposes of comparison. In Fig. 1, the branching ratios, calculated for the various Z decay channels and the Z WW cross section are plotted as a function of mZ . The observability of the hadronic decay mode, which is the dominating one, suers from enormous QCD background, which is the main source of background at the LHC experiments. The possibility of Z boson detection in the di-jet decay mode has been discussed in [15] where it was found that a limit of 3 T eV /c2 can be achieved after a one year run at high luminosity. The leptonic modes have much lower branching ratio than the hadronic ones, but are the cleanest channels to detect the Z signal since the background level is (Z W W ) = c2
http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

Fig. 1. Left: Branching ratios for the dierent Z decay channels. Right: cross section of pp Z WW process as a function of mZ negligible compared to the signal [16]. The intermediate goal after the discovery of the Z boson would be to understand its origin and its properties, including its couplings to ordinary fermions. The forward-backward asymmetry and the ratio of cross sections in dierent rapidity bins in the Z l+ l decay (l = e, ), are well known to be sensitive probe of new physics [17]. In the following, we will concentrate on the estimation of the discovery capability of a SSM Z boson in the WW channel using the decay chain Z WW ll jj (l= e, ). The other possible decay chains Z W + W 4 jets and l+ l l l , would be much more dicult to observe, due respectively to the large QCD background and the missing information related to the two escaping neutrinos. The preliminary results of this study have been published in [18].

Description of the ATLAS detector

The ATLAS detector consists of four major subdetectors which are described as follows : the inner detector [19] for particle momentum measurement, the leadliquid argon electromagnetic calorimeter with an accordeon geometry [20], the hadronic calorimeter made with iron and scintillating tiles in the regions where the radiation level is low enough [21] and made with copper and liquid argon elsewhere, and the muon spectrometer [22]. In the present analysis, importance is given to the performance of the detector for identication and reconstruction of electrons, muons and hadronic jets. Also, miss is required to reconstruct the Z boson in the a good reconstruction of ET
http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

WW channel. The latter needs a careful calibration of all calorimeters and an ecient pile-up and electronic noise suppression [23]. miss The results of the detailed detector study, in terms of particles and ET reconstruction, are documented in Ref. [24]. Below, we summarize the most important expected performances of the detector, which will be crucial for the physics measurements. The expected energy resolution for electrons is parametrized as: b a (E) = c E E E (5)

The coecient a of the sampling term is 10% or smaller, except at the end of the barrel where the amount of material in front of the electromagnetic calorimeter is large. The noise term b depends on the pile-up and the electronic noise, the constant term c is smaller than 0.7% over the full electromagnetic calorimeter coverage. For muons, a good momentum resolution can be obtained, in the momentum range 5 < PT < 100 GeV/c, by combining the measurements made in the muon system with the ones from the inner detector. A resolution of (PT )/PT = 2 - 3 % can be achieved in this range of PT with a reconstruction eciency better than 85%. For high PT muons, a degradation of the resolution is expected ((PT )/PT = 11 % for PT = 1 TeV/c). For the hadronic calorimeter, a good energy resolution across the full pseudorapidity range can be obtained. Without pile-up and electronic noise, the energy resolution for jets is parametrized as: 50% 3% for || < 3 E (E) = (6) 100% E 7% for || > 3 E The pile-up and electronic noise contribution to the resolution of the hadronic calorimeter depends on the size of the cone used for the jet energy reconstruction. For a cone size of R = 0.4, the combined contribution of the pile-up and electronic noise to the jet energy resolution is ET = 7.5 GeV. Good measurement of the missing tranverse energy is needed to reconstruct the invariant mass of new particles involving neutrinos among their decay products. If the coherent noise is smaller than 3 MeV per channel in the electromagnetic calorimeter, which is one of the requirements for its electronics, the miss would be: resolution on the ET
miss (ET ) =

0.5 1.1

ET ET

at low luminosity at high luminosity

(7)

In this work, the ATLAS response to simulated events was studied with ATLFAST [25], a fast simulation program which carefully takes into account the expected performance of the ATLAS detector described above. In the simulations, the electrons, muons, jets and the missing transverse energy are reconstructed using the following procedure:
http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

The transverse energies of all stable particles (except neutrinos and muons) are summed in calorimeter cells of a given granularity (0.1 0.1 for || < 3 and 0.2 0.2 for || > 3). Clusters are dened by summing up all the cells in a cone R = 2 + 2 = 0.4 around the cluster initiator (i.e. cell with the highest ET ). If the cluster matches a generated electron passing the selection criteria (pT > 5 GeV/c and || < 2.5) with R < 0.1, it is tagged as an electron and the isolation cuts are applied: the electron must be separated by R > 0.4 from the other clusters and the transverse energy deposition in a cone R = 0.2 around the electron should not exceed 10 GeV. The energy of the isolated electron is smeared by using the expected energy resolution of the electromagnetic calorimeter: 12% noise (E) = 0.7% E ET E the noise coecient noise is parametrized as in [26]. Muons were required to pass the selections PT > 6 GeV and || < 2.5. They were also required to be isolated i.e. the distance from calorimeter clusters R < 0.4 and the transverse energy deposition in a cone R = 0.2 around the muon was required not to exceed 10 GeV. The eciency of the isolation criteria is about 97.8% for muons passing selection criteria. In our analysis, muon tracking is obtained by combining informations from the Inner Detector [27] and the Muon Spectrometer [28]. The momentum resolution is parametrized as a function of PT and . In the nal results, an eciency of 90% is conservatively assumed for the reconstruction of electrons and muons. The clusters which are not identied as electrons or photons are classied as jets. The jet energy is smeared according to a Gaussian distribution given by the expected energy resolution for jets (Eq. 6). At high luminosiy (1034 cm2 s1 ), a pile-up noise term is added in quadrature (pileup = 7.5 GeV for a cone of radius R = 0.4). The missing transverse energy is obtained by summing the transverse momenta of all isolated electrons, photons and muons, of all jets, of clusters which have not been accepted as jets, of non isolated muons which were not added to jets and of all cells which have not been used for clusters reconstruction (the energy of these cells is smeared with the same energy resolution as for jets). (8)

4
4.1

Event selection and background reduction

Signal reconstruction

The PYTHIA 5.7 and JETSET 7.4 event generators [29] have been used for the signal and the SM background generation. The signal Z WW ll jj
http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

Fig. 2. R separation between the two partons coming from the hadronic W decay where one Z decays hadronically and the other leptonically, is characterized by a nal state with one isolated lepton, a large missing transverse energy due to the escaping neutrino and at least two jets. As in many physical processes at the LHC (i.e. search for a heavy Higgs, measurement of the top quark mass ...), the reconstruction of the two jets coming from the W decay will play an important role in the signal discrimination from the background. A detailed description of the dierent methods, which can be used for the reconstruction of the W jetjet decay, is reviewed in [30]. In the case of jets with a small opening angle, two methods can be used. When the two jets are slightly overlapping, the W mass is calculated as the invariant mass of the cluster obtained by summing the energies of cells within a big cone R = 0.5. The energy of each jet is then calculated by reconstructing both jets in smaller cones of R = 0.2, within the big cone of the cluster, and applying a special method to share the energy in the overlap region between the two jets [30]. In the case of a high PT W, which gives two severely overlapping jets, the energy is collected within a single cone. Amongst the cones found, the one with the invariant mass nearest the Ws is chosen. The two jets are not reconstructed separately in this method. Given the experimental limits cited in the introduction of this paper, our analysis will concern the Z masses in the TeV range (from 750 GeV /c2 to 4 T eV /c2 ). The W bosons from the decay of such a heavy Z will have a large boost; typically, the distributions of the two W transverse momenta have a maximum at mZ /2. This results in a pair of jets very close in space (see Fig. 2), and dicult to separate in the detector. An algorithm based on the last method was implemented to reconstruct the hadronic decay of the high PT W. The rst method can be applied for mZ 1 TeV/c2 , but becomes less ecient for higher
http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

Fig. 3. Big cone mass (a) and reconstructed PT of W jj compared to the reconstructed PT of leptonic W (dashed line) (b) at high luminosity masses where the jets overlap is more pronounced. Adopting this method, the signal will be characterized by a nal state with one isolated lepton, a large missing transverse energy and a hadronic cluster reconstructed in a big cone and having an invariant mass compatible with the W nominal mass. The same selection criteria are applied to the backgrounds. In the choice of the cone radius, one must be aware that a small cone can introduce the loss of a big fraction of the energy whereas the pile-up contribution at high luminosity will be amplied if a too big cone is used. A cone radius of R = 0.5 - 0.6 seems to be the optimal choice. The pile-up contribution at high luminosity is taken into account by adding in quadrature a pile-up noise term in the smearing procedure of each cell energy, this term depends on the cone size of the cell and its transverse momentum. The results of the hadronic W reconstruction, obtained for MZ = 1 T eV /c2 and R = 0.5, are shown in Fig. 3. To reconstruct the leptonically decaying W, the PT of the neutrino is taken to be the reconstructed missing transverse momentum. In this way, the transverse mass of the lepton-neutrino system can be reconstructed as well as the transverse

http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

mass of the Z boson, dened as:

mZ = T
l miss + P W jj )2 (P l + P miss + P W jj )2 (P l + P miss + P W jj )2 (PT + PT x y x x y y T

4.2

Background reduction

The main sources of background are described by: The W + jets background, with the W boson decaying leptonically (W l). This backgound is potentially the largest one and becomes dangerous when the two jets have an invariant mass close to the W mass. In [31, 32], the PYTHIA and VECBOS Monte Carlo program [33] predictions for the W + jets production were compared, and an underestimation of the cross section was xed for PYTHIA. A ratio (VECBOS)/(PYTHIA) of about 1.2 was found. Furthermore, the ratio of the cross section measured at the Tevatron to the VECBOS prediction is around 1.3 [34]. Thus, the cross section of W + jets production given by PYTHIA has been increased by W a K-factor of 1.5, leading to a total cross section of 270 pb for PT > 100 GeV/c. tt background: The total cross section of this background is 600 pb. The b-quarks produced in the decay chain pp tt W bW can be used to b background using a jet veto in the central region. further reduce the tt The WW and WZ continuum: in this case, there are two real W and Z, with a total cross section about 96.4 pb. This background has a much lower rate than the W + jets and tt backgrounds but is irreducible in the central region. Shown in Fig. 4 are few characteristic distributions of the three backgrounds compared to those for the signal of mZ = 1 T eV /c2 . One can notice the clear peak in the big cone jj mass around the W mass in the case of the tt and WW-WZ backgrounds due to the presence of real W bosons in these background types. 4.2.1 Central cuts To ensure an ecient background rejection, one has to select events with high PT central Ws and leptons, which are expected to come from the Z decay, and with a large missing transverse energy. At this stage of the analysis, not knowing the Z mass, some of the cuts are dened as a function of the reconstructed invariant mass mrec in each event. Imposing the constraint Z that the invariant lepton-neutrino mass be equal to MW gives two solutions for the longitudinal momentum of the neutrino. The solution with the smaller value was found to agree generally better with the true value generated, and was therefore used to estimate the invariant mass mrec in each event at this stage of Z the analysis. A set of cuts, called central or high PT cuts, has been optimised in terms of statistical signicance of the signal above the background:
l miss > mrec /10 (GeV/c) and |l | < 2.0; Lepton cuts: PT , PT Z W W High PT W cuts: PT l , PT jj > mrec /3 (GeV/c); Z

http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

10

Fig. 4. Characteristics of 1 TeV/c2 Z signal (hashed histogram) compared to the three dominant background types for an integrated luminosity of 100 f b1 : (a) lepton transverse momentum, (b) reconstructed PT of leptonic W, (c) invariant mass of the two jets reconstructed in a cone R = 0.5 and (d) pseudorapidity of the hadronic W

http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

11

Table 1. S/B and S/ B before and after central cuts for MZ = 1 T eV /c2 Cuts before cuts hadronic W cuts Leptonic W cuts S/B 30 f b1 1.9103 6.9102 0.18 S/ B 30 f b1 2.6 10. 13.3 S/B 300 f b1 2.103 5.6102 0.144 S/ B 300 f b1 8.7 31.3 40.

The big cone mass must be compatible with the nominal W mass: MW jj = MW 15 GeV /c2 ; | W jj | < 2.0: hadronic W must be central. Table 1 gives the signal to background ratios and the signicances before and after the central cuts; the values of S/B and S/ B calculated before the application of cuts, are obtained by requiring one isolated lepton and at least one hadronic cluster in the event. 4.2.2 Central jet veto A central jet veto is applied to reject the tt background which is expected to have large central jet activity due to the presence of two b quarks from top pair decay. The eciency of such a veto has been conrmed by the studies concerning the heavy Higgs search in the channel H WW [35, 36]. After central events selection, the events containing central jets thr (|j | < 2.0), with PT > PT , are rejected. After identication of the cluster coming from the hadronic W decay, the central jets are reconstructed using a standard cone algorithm with R=0.4. thr The eciency of such a cut depends on the choice of the PT value. At low luminosity the jet veto threshold can be lowered without loss of signal eciency while keeping a good rejection of the background. At high luminosity minimumbias events tend to generate additional low PT jets and the jet veto threshold thr has to be raised. A PT of 20 GeV at low luminosity and 30 GeV at high luminosity has been chosen. In the presence of high luminosity pile-up, the surviving eciency for the signal is 60%. For the W+jets background, the eciency is signicantly lower than for the signal (37%) because of the larger jet activity in these events. As predicted before, the tt background is strongly reduced by the jet veto with an eciency of 15%. For the WW-WZ background, the eciency is not much dierent from the signal one (55% at high luminosity) because the central cuts applied to this background select real W or Z with high PT similar to those from Z decay making this background irreducible in the central region. The eect of the jet veto on the statistical signicance is illustrated in Table 2, where S and B denote respectively the eciencies for signal and the background. The signicance is improved by the veto if the ratio S / B is greater than 1. The eect of the veto on the signal to background ratio is clear but no improvement of the signicance is observed for mZ > 1.5 T eV /c2 . The relative contribution of each background type to the total one in the central region with high mass can be used to explain such a behavior. From Table 2, one
http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

12

Table 2. S / B and the contribution of WW-WZ continuum to the total background at high luminosity mZ (GeV /c2 ) S/ B % of tt % of WW-WZ
S/ B

1000 2.2 1.15 45.5% 4.4%

1500 1.8 1.02 18.4% 12.73%

2000 1.5 0.94 6.2% 25.76%

2500 1.37 0.9 2.04% 40.8%

can notice a clear decrease of the tt contribution and an increase of the WW-WZ continuun contribution, leading to an increase of B . 4.3 Discovery potential

After applying all cuts described above, we can estimate the expected statistical signicance as a function of MZ after three years running at high and low luminosity. The limit obtained for Z WW coupling, corresponding to a significance of 5 has been determined. In Fig. 5, we show the ratio of ( B)limit , the cross section required for a signicance of 5, to ( B)SSM , the cross section obtained using the reference model, for MZ ranging from 750 GeV to 3 TeV. We assume that, for a discovery in the SSM scenario, the ( B)limit/( B)SSM 1 and at least 10 signal events are expected. For an integrated luminosity of 30 f b1 , the channel Z W W ljj allows discovery of the SSM Z boson up to masses about 1.6 TeV/c2 . This limit is about 2.2 TeV/c2 for an integrated luminosity of 300 f b1 . For higher masses, it seems dicult to detect the SSM Z boson in this channel due essentially to the rapid drop of the associated cross section. Shown in Fig. 6 is the transverse mass spectra for a 1.5 TeV Z boson together with the contribution from the residual background. Alternatively, in the general case where the coupling of Z to WW pair is suppressed by a factor c(mZ /mZ )2 compared to the standard ZWW coupling, if no signal is observed, one can obtain an upper bound clim , at the 3 level, on the ratio of the Z WW coupling to the SSM values (Fig. 7). The discovery limits found in this section can be applied to extract new limits for other generic resonances, in the WW channel. Such resonances appear in strong symmetry breaking models where the Higgs is very heavy [35], in technicolor models [37] or in more general Chiral Lagrangian models. Given the production cross section, the width of the resonance and the branching ratios, the limits found in the present study can be re-interpreted in terms of limits for the new models. If the resonance is produced by vector boson fusion, forward jet tagging provides an additional selection criterion.

http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

13

Fig. 5. Discovery limit of a Z boson in the Z WW ll jj channel for an integrated luminosity of 30 f b1 and 300 f b1

Fig. 6. Transverse mass distribution of the two hadronic jets, the lepton and the escaping neutrino for the decay of 1.5 TeV/c2 Z boson and the corresponding background for an integrated luminosity of 300 f b1

http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

14

Fig. 7. Upper bound on the c parameter, at 3 level, for an integrated luminosity of 300 f b1

Reconstruction of the Z invariant mass

l The mass of the Z can be obtained by constraining P , the longitudinal momentum of the neutrino, to yield an invariant mass of the lepton-neutrino system equal to the mass of the W. In Sect. 4.2, the smaller of the two calculated values l of P was chosen. If the Z boson has been discovered, a second method would be to choose the solution which yields the invariant mass closer to the expected Z mass extracted from the transverse mass distribution. The two methods are compared in Fig. 8, for the signal (Figs. 8a and b) and for the signal together with the residual background (Figs. 8c and d). The second method gives a better resolution on the Z mass but creates an articial peak in the background distribution around the Z mass since a value closer to mZ was chosen. After the discovery of the Z boson, using the transverse mass distribution, one could be able to extract the signal, the articial peak will then disappear and the Z width could be estimated.

Conclusions

The possibility to observe a new Z boson in the Z WW channel was examined. Even if the Z boson is rst observed via the purely leptonic decay mode, the measurement of the Z WW vertex would give independent information on the new physics. The diculty of reconstructing a Z in this channel comes from the need to identify a high PT hadronic W giving two very close jets. A single cone method seems to be the optimal procedure in this case. This decay channel suers also from the enormous background coming essentially from the WW, WZ continuum, the W + jets production and tt pairs. Central cuts selecting
http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

15

Fig. 8. Invariant mass distribution obtained with the two methods for the signal and for the signal with the residual background. The distributions correspond to MZ = 1.5 TeV/c2 and an integrated luminosity of 300 f b1

http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

16

a high PT WW pair are used to reject a large part of the backgound. These cuts can be supplemented with the central jet veto to reject the tt background. The usefulness of this veto is limited to the Z boson with mass lower than 1.5 T eV /c2 . For an integrated luminosity of 300 f b1 , the SSM Z signal can be seen clearly up to masses of 2.2 T eV /c2 . The limits found in this study can be re-interpreted to obtain discovery limits of other generic resonances, arising in other models, like technicolors signal. A detailed study of the characteristics of the resonance could help to distinguish between dierent models. Acknowledgements This work has been performed within the ATLAS Collaboration, and we thank collaboration members for helpful discussions. We have made use of the physics analysis framework and tools which are the result of collaboration-wide eorts. Special thanks go to G. Azuelos and E. Richter-Was for careful reading of this article and for fruitful discussions and helpful suggestions.

References
1. S.L Glashow, Nucl. Phys. 22 (1961) 579. S. Weinberg, Phys. Rev. Lett. 19 (1967) 1264. 2. Super-Kamiokande Collaboration, Y. Fuduka et al. Phys. Rev. Lett. 81 (1998) 1562. 3. J.L. Hewlett and T.G. Rizzo, Phys. Rep. 183 (1989) 193. 4. J.C Pati and A. Salam, Phys. Rev. D10 (1974) 275. R.N. Mohapatra and J.C. Patti, Phys. Rev. D11 (1975) 566. 5. D. Chang, R.N. Mohapatra and M. Parida, Phys. Rev. D30 (1984) 1052. 6. A. Ferrari et al., Phys. Rev. D62 (2000) 013001. 7. CDF Collaboration, F. Abe et al., Phys. Rev. Lett. 79 (1997) 2191. 8. D0 Collaboration, S. Abachi et al., Phys. Rev. Lett. 76 (1996) 3271. 9. G. Beal, M. Bander and A. Soni, Phys. Rev. Lett. 48 (1982) 848. 10. M. Cveti and S. Godfrey, published in Electroweak Symmetry Breaking and c Physics Beyond the Standard Model, eds T. Barklow, S. Dawson, H. Haber and J. Seigrist (World Scientic 1996). 11. E. Merle, PhD Thesis, Universit de Haute Savoie (1999). e 12. M Cveti, P. Langacker and B. Kayser, Phys. Rev. Lett. 68, 2871 (1992). c 13. A.C. Longhitano, Phys. Rev. D22 (1980) 166; Nucl. Phys. B188 (1981) 118. 14. G. Altarelli et al. Z. Phys. C45 (1989) 109. 15. A. Henriques and L. Poggioli, ATLAS internal note ATLAS-PHYS-92-010 (1992). 16. P. Camarri et al., Proceedings the LHC workshop, Vol. II, p. 704, October 1991. 17. F. Del Aguila et al., Phys. Rev. D R969 (1993) 48. 18. ATLAS Collaboration, Detector and Physics Performance , Technical Design Report vol. II, ATLAS TDR 15, CERN/LHCC 99-15. 19. ATLAS Collaboration, Inner Detector Technical Design Report, ATLAS TDR 9, CERN/LHCC 97-16. 20. ATLAS Collaboration, Liquid Argon Calorimeter Technical Design Report, ATLAS TDR 1, CERN/LHCC 97-40.

http://link.springer.de/link/service/journals/10105/index.htm

EPJdirect C3, 117 (2001) Springer-Verlag

17

21. ATLAS Collaboration, Tile Calorimeter Technical Design Report, ATLAS TDR 3, CERN/LHCC 96-42. 22. ATLAS Collaboration, Muon Spectrometer Technical Design Report, ATLAS TDR 10, CERN/LHCC 97-22. 23. D. Cavalli, ATLAS Internal Note ATLAS-PHYS-96-080 (1996). 24. ATLAS Collaboration, Tetector and Physics Performance , Technical Design Report vol. I, ATLAS TDR 14, CERN/LHCC 99-14. 25. D. Froidevaux, L. Poggioli and E. Richter-Was, ATLAS Internal Notes ATLASPHYS-96-079 and ATLAS-PHYS-98-131. 26. ATLAS Collaboration, Technical Proposal, CERN/LHCC 92-4. 27. L. Poggioli, D. Froidevaux and C. Guyot, ATLAS Internal Note, ATLAS-PHYSNo-076 (1995). 28. E.J. Buis et al. ATLAS Internal Note ATLAS-INDET-98-215 (1998). 29. T. Sjstrand, Comp. Phys. Comm. 82, 74 (1994). o 30. P. Savard, ATLAS Internal Note ATLAS-CAL-97-092 (1997). 31. M. Cobal, D. Costanzo and S. Lami, ATLAS Internal Note, ATLAS-PHYS-96-84 (1996). 32. V. Ilyin and A. Skachkova, proceeding of the 14th International Workshop On High Energy Physics And Quantum Field Theory (QFTHEP 99), 27 May - 2 Jun 1999, Moscow, Russia, p. 323-326. 33. F.A. Berends et al., Nucl. Phys. B357, 32 (1991). 34. CDF Collaboration, F. Abe et al., Phys. Rev. Lett. 79 (1997) 4760. 35. P. Savard and G. Azuelos, ATLAS Internal Note ATLAS-PHYS-98-128 (1998). 36. S. Smushko, D. Froidevaux and L. Poggioli, ATLAS Internal Note ATLAS-PHYS97-103 (1997). 37. G. Azuelos et al., ATLAS Internal Note, ATLAS-PHYS-99-020 (1999).

http://link.springer.de/link/service/journals/10105/index.htm