Experimental Sensitivity for Majorana Neutrinos Produced via a Z Boson at Hadron Colliders
aa r X i v : . [ h e p - ph ] J a n Experimental Sensitivity for Majorana NeutrinosProduced via a Z Boson at Hadron Colliders
A. Rajaraman and D. Whiteson University of California, Irvine, Irvine, California 92697
We present experimental sensitivity for the production of a fourth-generation Majorana neutrino( N ) via an s -channel Z boson Z → NN → W ℓW ℓ at hadron colliders. This channel has not beenstudied for the Tevatron dataset with R L = 5 fb − , where we find that a single experiment couldsignificantly extend current 95% C.L. mass lower limits, to m N >
175 GeV /c , or report 3 σ evidencefor the N if m N <
150 GeV /c . With 5 fb − , a single LHC experiment at √ s = 10 TeV couldexpect to set a 95% C.L. mass lower limits of m N >
300 GeV /c , or report 3 σ evidence for the N if m N <
225 GeV /c . PACS numbers: 12.60.-i, 13.85.Rm, 14.80.-j
INTRODUCTION
The study of a fourth generation of quarks and lep-tons has undergone a renaissance. While earlier studies[1] claimed that the fourth generation was ruled out byexperimental data, a recent analysis [2] has shown thatthe constraints can be satisfied by appropriate choices ofmass splittings between the new fourth generation par-ticles. Furthermore, the existence of the fourth gener-ation can help to address certain discrepancies betweenthe Standard Model and experimental results in the b-quark sector[3–8], leading to a revival of interest in thispossibility (see [9] for a review).In view of the imminent arrival of LHC data, it is ofparticular interest to search for signals of these fourth-generation fermions at hadron colliders. Such analyseshave been performed for the t ′ and b ′ quarks. CDF hassearched for the t ′ quark [10] using the CKM suppresseddecay t ′ → qW and for the b ′ quark [11] using the pro-cess b ′ b ′ → t ¯ tW − W + which can yield like-sign dileptons.These searches have set limits of 311 GeV for the t ′ [10],and 338 GeV for the b ′ [11] (but see [12]). The LHC willdiscover or exclude fourth generation quarks up to abouta TeV [13–15].However, few detailed analyses have been performed inthe lepton sector. By analogy with the first three gener-ations, one might expect the leptons of the fourth gener-ation to be lighter than the fourth generation quarks andin particular, the fourth generation neutrino might beexpected to be the lightest new particle. It is of interestto see if this neutrino can be found at colliders.Past searches for fourth generation neutrinos havemainly been performed at lepton colliders. In particular,LEP II has looked for neutrinos produced in the process e + e − → Z → N N , where the neutrinos subsequentlydecay via the process N → l ± W ∓ . No excess of suchevents was found, which placed a limit of about 100 GeVfor Dirac neutrinos decaying to electrons, muons or taus.For Majorana neutrinos the corresponding limits wereabout 90 GeV if the neutrino decayed either to electrons or to muons, and about 80 GeV if it decayed to taus.Theoretical analyses of fourth generation neutrinos athadron colliders [16–19] have focused on the process q ¯ q ′ → W ± → N l ± where the fourth generation neu-trino is produced in association with a light charged lep-ton. This process has the significant advantage that onlyone heavy particle is produced, which increases the massreach considerably. Furthermore, the neutrino will de-cay through N → l ± W ∓ which will produce the low-background like-sign dilepton signature in half the events.However, the production cross-section for this processdepends on the mixing between the fourth generationwith the first three generations due to the W → lN ver-tex. In many models, this mixing angle can be small; forexample, if the mixing is generated by GUT or Planck-scale suppressed operators, the angle may be as small as10 − [12]. For mixing parameters smaller than about10 − , the neutrino production rates in this channel aretoo small to be observable at colliders [16]. In modelswith small mixing angles, the dominant production mech-anism becomes pair production through an s -channel Z , for which the production rate is model-independent.These signals have been studied at various benchmarkpoints for the LHC [20] and for future linear colliders [21].However, the analysis of the s -channel Z process hasnot been performed for the Tevatron. In this Letter, wepresent a sensitivity study for the Tevatron and arguethat the LEP limits on Majorana neutrinos can be sig-nificantly improved with an analysis of the data alreadytaken. It would be of great interest to perform a fullanalysis of this data. We also perform a similar study forthe LHC, which can probe the parameter space to muchhigher energies. PRODUCTION AND DECAY
We consider an extension to the standard model bya fourth generation of fermions and a right-handed neu-trino. The mass term for the neutrinos can be writtenas L m = −
12 ( Q cR N cR ) (cid:18) m D m d M (cid:19) (cid:18) Q R N R (cid:19) + h.c. (1)where ψ c = − iγ ψ ∗ . This theory has two mass eigen-states of masses m = − ( M/
2) + p m D + M / , m =( M/
2) + p m D + M /
4. In addition, the mass of thefourth generation lepton is constrained to be close in massto the neutrinos by precision electroweak constraints [2].We consider processes where only the lightest neu-trino is produced, providing the most model-independentbound on this theory. In future work, we will study theeffect of the second neutrino and fourth generation lep-ton. For this analysis, we treat the lepton and secondneutrino as infinitely massive, corresponding to a limitwhere
M, m D go to infinity with m D M fixed.The lighter neutrino mass eigenstate is the Majoranafermion N = N cL + N L . The coupling for pair productionvia the Z is through the coupling L Z = Z µ J µ where J µ = e θ W cos θ W ( ¯ N γ µ γ N )The heavy neutrino will decay through N → W ± l ∓ (the neutral current process is suppressed.) Note that N can decay to either sign of lepton, giving like-sign leptonsin half of the events, see Fig 1. We assume that the heavyneutrino decays promptly; this will be the case unless themixing between the fourth generation and the first threeis extremely small [12].We consider the possible decay modes N → W ( e, µ, τ ).In a hadron collider, backgrounds to τ leptons are muchlarger and efficiencies are much lower than for e and µ ,giving the τ decay mode little power. We consider twocases, (a) µµ , in which the non- τ decays appear solely asmuons: BR( N → W µ ) = 1 - BR( N → W τ ); and (b) ℓℓ ,with ℓ = e, µ in which the non- τ decays appear as bothelectrons and muons: BR( N → W µ ) + BR( N → W e )= 1 - BR( N → W τ ). The µ ± µ ± mode has significantlysmaller background rate than ℓ ± ℓ ± . FIG. 1: Pair production of the heavy Majorana neutrino N via a Z boson, and subsequent decay W ± l ∓ . If the N decays to ℓW , then the decay of N N → ℓW ℓW can be categorized by the decays of the W bosons.If both W s decay hadronically, we expect approximately 4 jets. If one decays leptonically, we expect approxi-mately 2 jets and a third lepton. If both decay leptoni-cally, we expect approximately zero jets but four leptons.All but the fully leptonic mode, the smallest fraction,contribute to the ℓ ± ℓ ± jj signature and allow for directreconstruction of the N . EXPERIMENTAL SENSITIVITY
We select events with the ℓ ± ℓ ± jj signature: • two like-signed reconstructed leptons ( e or µ ), eachwith p T >
20 GeV and | η | < . • at least two reconstructed jets, each with p T > | η | < . N can be reconstructed from the two jets and eitherof the leptons. In the case that three jets are recon-structed, we form the mass of the N from the invariantmass of the two jets which are closest to the W mass,and either of the leptons. In the case that four jets arereconstructed, the mass of the two N s come from the ljj assignments that give the best W masses and the small-est difference between the two reconstructed m N s. SeeFigure 2. ] Neutrino Mass [GeV/c80 100 120 140 160 180 200 E ve n t s FIG. 2: Reconstructed Majorana neutrino ( N ) mass in eventswith 2, 3 or at least 4 jets for m N = 150 GeV/ c . Backgrounds
At the Tevatron, the largest backgrounds to the ℓ ± ℓ ± jj signature come from W γ or W Z production ormisidentified leptons [22] either from semi-leptonic t ¯ t de-cays or direct W + jets production.For the Tevatron, we extrapolate the number of ex-pected backgrounds events from Ref. [22] to a datasetwith 5 fb − , use madgraph [23] to model the kinemat-ics of the events, pythia [24] for showering and a versionof pgs [25] tuned to describe the performance of the CD-FII detector.At the LHC, the diboson contribution includes an ad-ditional process, qq → W ± W ± q ′ q ′ , which directly pro-duces the ℓ ± ℓ ± jj signature. For the LHC, we calculatethe size and kinematics of each contribution using mad-graph , use pythia for showering and a version of pgs tuned to describe the performance of the ATLAS detec-tor.Figure 3 shows the reconstructed mass shape for N pair production and for the backgrounds in the µ ± µ ± jj case. ] Neutrino Mass [GeV/c
80 100 120 140 160 180 2000.20.40.60.811.21.4 NN fi ZDibosonFakes ] Neutrino Mass [GeV/c
100 150 200 250 300 350 400 450 5000510152025 NN fi Zdibosonfakes
FIG. 3: Expected reconstructed neutrino mass for N produc-tion with m N = 150 GeV/ c , and backgrounds to the µ ± µ ± jj signature in 5 fb − of Tevatron data (top) or 10 TeV LHCdata (bottom). Expected Limits and Discovery Potential
We perform a binned likelihood fit in the reconstructed N mass, and use the unified ordering scheme [26] to con- struct frequentist intervals. If the N does not exist andno excess is seen, the median expected upper limits onthe cross-section are given in Table I and Fig. 4. In5 fb − , with BR( N → µW ) = 100%, a single Tevatron(LHC) experiment could expect to set a 95% lower limitof m N >
175 (300) GeV. The limits as a function ofBR( N → τ W ) are given in Fig. 5. If the N does exist,a 3 σ excess would be observed in the regions shown inFig. 5. ] Mass [GeV/c n
100 150 200 C r o ss - s e c t i on [f b ] -1 ] Mass [GeV/c n
100 200 300 400 C r o ss - s e c t i on [f b ] -1 FIG. 4: Theoretical cross-section for N production and de-cay to ℓ ± W ℓ ± W and median expected 95% C.L experimentalcross-section upper limits in 5 fb − of Tevatron data (left) orLHC data (right), assuming BR( N → µW ) = 100%. ] Neutrino Mass [GeV/c
100 150 ) t W fi n B R ( ] Neutrino Mass [GeV/c
200 250 300 ) t W fi n B R ( ] Neutrino Mass [GeV/c
100 120 140 ) t W fi n B R ( ] Neutrino Mass [GeV/c
100 150 200 ) t W fi n B R ( FIG. 5: Median expected 95% C.L experimental exclusion(top) or 3 σ evidence (bottom) in 5 fb − of Tevatron (left)LHC data (right) as a function of BR( N → W τ ). Two decaycases are shown: µ ± µ ± (black) or ℓ ± ℓ ± (red), as defined inthe text. CONCLUSIONS
The s -channel pair production of heavy Majorana neu-trinos via a Z boson ( Z → N N → W ℓW ℓ ) is a powerfuldiscovery mode at hadron colliders. With 5 fb − of data,the Tevatron can significantly extend the limits on suchneutrinos to 175 GeV/ c , and a 3 σ evidence is possible if TABLE I: Theoretical cross section, σ Theory at the Teva-tron or LHC, including branching ratio to like-sign leptons;selection efficiency ǫ for the µ ± µ ± jj channel; number of ex-pected events in 5 fb − of data; and median expected exper-imental cross section 95% CL upper limits, σ Limit , assumingBR( N → W µ ) = 100%. TevatronMass [GeV/ c ] 100 125 150 175 200 225 σ Theory [fb] 26.7 9.8 4.1 1.8 0.9 0.4 ǫ σ Limit [fb] 8.3 2.5 2.0 1.8 1.6 1.0LHC, 10 TeVMass [GeV/ c ] 100 150 200 250 300 350 400 σ Theory [fb] 195 39 12 5.2 2.3 1.2 0.6 ǫ σ Limit [fb] 10.7 4.5 2.7 2.6 2.3 2.0 1.5 the mass is less than 150 GeV /c . A dataset of the samesize at the LHC would have an 95% C.L. exclusion reachof 300 GeV /c and 3 σ evidence potential for m N < /c . ACKNOWLEDGEMENTS
We acknowledge discussions with M. Bondioli, L. Car-penter, K. Slagle, T. Tait, and R. Porter,. A.R. is sup-ported in part by NSF Grant No. PHY-0653656. D.W.is supported in part by the U.S. Dept. of Energy. [1] C. Amsler et al. [Particle Data Group], Phys. Lett. B , 1 (2008).[2] G. D. Kribs, T. Plehn, M. Spannowsky and T. M. P. Tait,Phys. Rev. D , 075016 (2007) [arXiv:0706.3718 [hep-ph]].[3] T. Aaltonen et al. [CDF Collaboration], “First Flavor-Tagged Determination of Bounds on Mixing-InducedCP Violation Phys. Rev. Lett. , 161802 (2008)[arXiv:0712.2397 [hep-ex]].[4] V. M. Abazov et al. [D0 Collaboration], “Measurementof B s mixing parameters from the flavor-tagged decay Phys. Rev. Lett. , 241801 (2008) [arXiv:0802.2255[hep-ex]].[5] W. S. Hou, M. Nagashima and A. Soddu, “Differ-ence in B+ and B0 direct CP asymmetry as ef-fect of a fourth Phys. Rev. Lett. , 141601 (2005)[arXiv:hep-ph/0503072].[6] W. S. Hou, M. Nagashima and A. Soddu, “Enhanced K ( L ) → π ν ¯ ν from direct CP violation in B → Kπ Phys. Rev. D , 115007 (2005) [arXiv:hep-ph/0508237].[7] A. Soni, A. K. Alok, A. Giri, R. Mohanta and S. Nandi,arXiv:0807.1971 [hep-ph].[8] [The Belle Collaboration], “Difference in direct charge-parity violation between charged and neutral Nature , 332 (2008).[9] B. Holdom, W. S. Hou, T. Hurth, M. L. Mangano,S. Sultansoy and G. Unel, PMC Phys. A , 4 (2009)[arXiv:0904.4698 [hep-ph]].[10] D. Cox [CDF Collaboration and for the CDF Collabora-tion], arXiv:0910.3279 [hep-ex].[11] T. Aaltonen et al. [The CDF Collaboration],arXiv:0912.1057 [hep-ex].[12] P. Q. Hung and M. Sher, Phys. Rev. D , 037302 (2008)[arXiv:0711.4353 [hep-ph]].[13] V. E. Ozcan, S. Sultansoy and G. Unel, Eur. Phys. J. C , 621 (2008).[14] O. Cakir, H. Duran Yildiz, R. Mehdiyev and I. TurkCakir, Eur. Phys. J. C , 537 (2008) [arXiv:0801.0236[hep-ph]].[15] B. Holdom, JHEP , 069 (2007) [arXiv:0705.1736[hep-ph]].[16] T. Han and B. Zhang, Phys. Rev. Lett. , 171804 (2006)[arXiv:hep-ph/0604064].[17] A. Atre, T. Han, S. Pascoli and B. Zhang, JHEP ,030 (2009) [arXiv:0901.3589 [hep-ph]].[18] F. del Aguila, S. Bar-Shalom, A. Soni and J. Wudka,Phys. Lett. B , 399 (2009) [arXiv:0806.0876 [hep-ph]].[19] F. del Aguila, J. A. Aguilar-Saavedra and R. Pittau,JHEP , 047 (2007) [arXiv:hep-ph/0703261].[20] T. Cuhadar-Donszelmann, M. K. Unel, V. E. Ozcan,S. Sultansoy and G. Unel, JHEP , 074 (2008)[arXiv:0806.4003 [hep-ph]].[21] A. K. Ciftci, R. Ciftci and S. Sultansoy, Phys. Rev. D , 053006 (2005) [arXiv:hep-ph/0505120].[22] CDF Collaboration, Phys. Rev. Lett. , 221803 (2007)[23] J. Alwall et al. , JHEP
028 (2007)[24] T. Sj¨ostrand et al. , Comput. Phys. Commun. et al. arXiv:hep-ph/0010338v2[26] G. J. Feldman and R. D. Cousins, Phys. Rev. D57