SLAC-PUB-7788 hep-phèyymmnnn March 1998 The importance of tau leptons for supersymmetry searches at the Tevatron James D. Wells y Stanford Linear Acce

Size: px

Start display at page:

Download "SLAC-PUB-7788 hep-phèyymmnnn March 1998 The importance of tau leptons for supersymmetry searches at the Tevatron James D. Wells y Stanford Linear Acce"

Mitchell Hood
5 years ago
Views:

1 SLAC-PUB-7788 hep-phèyymmnnn March 998 The importance of tau leptons for supersymmetry searches at the Tevatron James D. Wells y Stanford Linear Accelerator Center Stanford University, Stanford, California 9439 Abstract Supersymmetry is perhaps most eæectively probed at the Tevatron through production and decay of weak gauginos. Most of the analyses of weak gaugino observables require electrons or muons in the ænal state. However, it is possible that the gauginos will decay primarily to ç leptons, thus complicating the search for supersymmetry. The motivating reasons for high ç multiplicity ænal states are discussed in three approaches to supersymmetry model building: minimal supergravity, minimal gauge mediated supersymmetry, and more minimal supersymmetry. The concept of ëe=ç=ç candidate" is introduced, and an observable with three e=ç=ç candidates is deæned in analog to the trilepton observable. The maximum mass reach for supersymmetry is then estimated when gaugino decays to ç leptons have full branching fraction. èsubmitted to Mod. Phys. Lett.è y Work supported by the Department of Energy under contract DE-AC3-76SF55 and DE-FG-5-87ER439.

2 Introduction The eæects of minimal supersymmetry on hadron collider observables are not expected to be spectacular. It would be easier if supersymmetry strongly implied additional resonances in pçp scattering, or narrow invariant mass peaks of two leptons, etc. Perhaps we will ænd high multiplicity lepton, photon or jet signals from R-parity violation or prompt decays to gravitinos in low energy supersymmetry breaking theories. This is not the most likely scenario. Simplicity, proton stability arguments, and dark matter considerations constitute a mild preference for R-parity conservation. Furthermore, it appears most natural in gauge mediated models to have the superpartners feel only a small fraction of the full supersymmetry breaking in nature ëë. Thus, prompt decays of the NLSP ènext lightest supersymmetric partnerè on the time scales of detector sizes are not favored ë2ë. The classic signatures of supersymmetry are not quite as spectacular as those created by promptly decaying NLSPs or R-parity violation. Perhaps the most probing signal ë3, 4ë of supersymmetry at the Tevatron is pçp! ç æ ç 2! 3l+ 6E T èè where l = e or ç. Cutting on the p T values of these leptons and requiring that m l + l, 6= m Z, leaves a small background from W æ Z production followed by W! lç and Z! çç, where the ç's decay leptonically to eçç or ççç. The maximum mass reach of degenerate gauginolike ç æ and ç 2 states is about 2 GeV with 2 fb, of integrated luminosity ë5ë. One of the main purposes here is to demonstrate that it is quite natural to expect that the 3l signal is not present at any reasonable rate at the Tevatron. Instead, one expects over a large region of supersymmetry parameter space to be dominated by multiple ç events from ç + ç 2 production. The reasons for this are explained in the subsequent section for three diæerent approaches to supersymmetry model building èstandard supergravity scenarios, gauge mediated models, and ëmore minimal" supersymmetryè. Because ç's are more diæcult to tag and identify than leptons, and because quark and gluon jets sometimes look like ç's, the searches in this mode are diæcult. Some estimates are made for the search capabilities.

3 2 Reasons for 3ç events There are several reasons why ç æ may decay preferentially into ç æ çç over l æ çç, and why ç 2 may decay toç + ç, ç over l + l, ç. Firstly, ~ç is lighter than ~ l R. If the ç Yukawa coupling is large then renormalization group eæects will drive ~ç R below ~ l R. More importantly the ~ç mass matrix can have a large mixing from ç ç vç, where v = 74 GeV. This mixing angle will drive the lightest eigenvalue down, thus potentially making m ~ç true for large tan æ which enhances the ç Yukawa coupling, ç ç ç m ~l R. This is especially ' m ç tan æ : è2è v The large tan æ correlation with high ç multiplicity has been recently emphasized in ref. ë6ë. Large mixing also results in models with large A ç triscalar soft mass Yukawa. If the slepton masses are lighter than ç æ and ç 2 masses, then on-shell decays into ~ç will be more likely than ~ l R. Not only is the phase space larger for ~ç, but also the substantially mixing introduces a large ~ç L component to~ç thus coupling much more eæectively to ç æ and ç 2 which can be shown to be mostly SUè2è L gauginos after radiative electroweak symmetry breaking conditions are imposed in minimal supergravity ë7ë and minimal gauge mediated models ë8, 9, ë. èthe result is more robust than these minimal models.è Furthermore, the large ç Yukawa coupling enables interactions with Higgsinos, which is negligible in the case of ~e or ~ç. Another reason why the ç ænal states might dominate can be extracted from the ideas of ëmore minimal" supersymmetry ëë. According to this approach, all the ærst two generations squarks and sleptons must be very heavy ègreater than a few TeVè to suppress large unwanted CP violation and æavor changing neutral current eæects in the electric dipole moment of the neutron, K, ç K mixing, ç! eæ, etc. The ~t i, ~ b L, Higgsinos and SUè2è L æ Uèè Y must be light in order to have natural electroweak symmetry breaking. The remaining æelds è~g, b R and ~ç i è can be either light orheavy without causing problems. It is natural to assume that they are light since ~g is probably tied up with the other gauginos in some uniæcation relation, and it is reasonable to put all scalars of the same generation at the same scale. Since ~t i and ~ b L must be light, then by this argument so should the other third generation scalars ~ br and ~ç i. In this case, decays of ç æ and ç 2 may only be allowed to decay into ç's via ~ç since no other slepton is close in mass. Although this approach may lead to new problems ë2ë, it 2

4 Figure : Branching ratio of ç æ ç 2 into three ç-leptons for the minimal supergravity model described in the text. is an attractive contributing solution to FCNC and CP violation suppression. For minimal supergravity èmsugraè and minimal gauge mediated supersymmetry breaking èmgmsbè models, we can quantify the prevalence of 3ç events in the data with a small number of parameters. For msugra, I have chosen m = 5 GeV, m =2 = 225 GeV, A =, and signèçè = + to be æxed for illustration purposes ë3ë. Both m ç æ and m ç 2 are close to 7 GeV for all tan æ and are mostly the superpartners of W æ and W 3 gauge bosons respectively. In Fig. the value of tan æ is varied from 2 to 3 and the branching fraction of 3ç events expected from ç æ ç 2 production and subsequent decays is plotted ë4ë. Furthermore, the ratio of m ~ç =m ç æ is shown. As expected, the 3ç rate becomes dramatic for larger tan æ and near almost è probability for tan æ é 2, when the ~ç mass dips suæciently below ç +. In minimal gauge mediated supersymmetry èmgmsbè I have chosen æ = M= 72 TeV, one 5+ç5 messenger multiplet, and signèçè = + ë5ë. Again, m ç æ and m ç 2 are close to 7 GeV for all tan æ and are mostly the superpartners of W æ and W 3 gauge bosons respectively. We see in Fig. 2 that the Brè3çè branching rate turns on much more smoothly. Here, the spectrum allows very light ~ç and ~ l R. The dominance of the 3ç ænal state is mostly due to the increasing ~ç L component of ~ç as tan æ increases. When tan æ é 25 the 3ç branching 3

5 Figure 2: Branching ratio of ç æ ç 2 into three ç-leptons in the gauge mediated model described in the text. fraction climbs above 9è. Even if the decays of ç æ! W æ ç and ç 2! Zç are on-shell, they do not compete with on-shell ~ç decay modes in large tan æ. The 3ç ænal state is most expected when tan æ is rather large. Many grand uniæed theories based on b, ç, t uniæcation ë6ë prefer large tan æ ' m t =m b. In gauge mediated models, unwanted large CP violation eæects are suppressed by requiring B = at the messenger scale ë, 9, 8, 7ë. This also implies large tan æ, hence high ç-multiplicity. Therefore, it is worthwhile to examine how eæectively one can search for 3ç ænal states. 3 Searching in the 3ç mode Analysis for ç lepton ænal states is possible for both CDF and D in the subsequent runs. The ç identiæcation rate might be as high as 3è in the hadron decay modes when p T èçè é 5 GeV ë8ë. This is encouraging for analysis of ænal state ç lepton signatures of new physics, such as the 3ç signature being considered here. The main background for 3ç events is WZ! 3ç and Zg where Z! çç and the g fakes a ç with the right jet-charge and low hadron multiplicity. The fake rate for this is expected to be approximately è or less with p T é 5 GeV. 4

6 One unifying approach to all trilepton signals, including 3l and 3ç, is to deæne an ëe=ç=ç candidate." An e=ç=ç candidate is deæned to be an isolated electron or muon or a fully tagged ç-jet in a hadronic decay mode of the ç. Therefore, the identiæcation of a primary e or ç as an e=ç=ç candidate is near è. The identiæcation of a primary ç lepton is the sum of its branching ratios into eçç è8èè and ççç è8èè plus the ç identiæcation rate è3èè that utilizes the hadronic decay modes of the ç. The combined e=ç=ç candidate eæciency is then approximately 66è for a ç lepton. The estimate here is made assuming p T èçè é 5 GeV, and çèçè é. For three ç's in an event satisfying these p T rate is as high as è:66è 3 ' 3è. and ç requirements, the identiæcation The visible decay products of a ç can be signiæcantly softer in p T than the original ç itself. For this reason, it is best to insist that all three p T èçè in the signal events are greater than 5 GeV when analyzing search capabilities ë9ë. In the end, a detailed detector simulation with well-deæned ç identiæcation requirements for each decay mode of the ç will be required for complete understanding of the capabilities. making sure that one ç satisæes p T Here, I will approximate this process by é 2 GeV, and the other two ç's satisfy p T èçè é 5. The trigger could occur by the substantial missing E T in the events, andèor large enough p T of an isolated leptonèsè, andèor a dedicated ç trigger. The trigger issues at both collaborations have not been settled but the hope here is that requiring p T èçè = f2; 5; 5g GeV and çèçè é f; ; g will yield a high overall trigger eæciency. To reduce the WZ background all events with two opposite sign same-æavor leptons should satisfy jm l + l,, m Z j é GeV. A small reduction in Zg can be obtained by cutting against back-to-back ç's in the azimuthal plan; however, the high p T requirements on the ëç" imposed above make this background reduction rather insigniæcant. Table 3 contains the estimated background to three e=ç=ç candidate events after cuts. A veto on additional jet activity abovee T é5 GeV has been applied to help reduce additional background. For example, the gg! b ç b background rate is expected to be small with an eæective veto in place. To estimate the maximum supersymmetry mass reach for the 3ç mode, the following choices should be made: The 3ç branching ratio is è which is valid for high tan æ. The light charginos and neutralinos are mostly gaugino so that m ç æ = m ç 2 = 2m ç, which is generally true in supersymmetric models with radiative electroweak symmetry breaking, and is certainly valid for msugra and mgmsb. And lastly, for maximum kinematic eæciency we assume that m ~ç is halfway between m ç and m ç 2. With these choices, the total signal 5

7 ç ëpbë Brèë3ç"è æ kin æ id ç cuts ëfbë W æ Z Zg 82 3:3 æ, total 7.2 fb Table : Main backgrounds to 3ç events at 2 TeV center of mass energy. The Brèë3ç"è indicates the branching fraction into three e=ç=ç candidates, but not allowing Z! ee; çç. The æ kin column is the eæciency of selecting all three e=ç=ç candidates with p T = f2; 5; 5g GeV and çé. The æ id column is the estimated probability of experimentally identifying all three leptons in the ænal state as e=ç=ç candidates after the kinematic cuts have been applied. Note that the WZ background mainly sums over eçç, ççç, and ççç ænal states, and the æ id averages over these ænal states. Also, the fake rate for g! ç is incorporated in Brèë3ç"è not æ id. R dtl mç æ reach fb, 35 GeV Table 2: Estimated maximum mass reach for ç æ in the 3ç channel of çæ ç 2 production and decay. rate after cuts is ç cuts è3ç signalè ' :4çèç 2 çæ è: è3è Combing this equation with a background rate of 7:2 fb, it is required that çèç æ ç 2è é ç 33 fb q R dtl in fb, è4è to get a 5ç signal above background over the applicable range for m + ç. Table 3 indicates what the corresponding mass reach is for these events for diæerent integrated luminosity at 2 TeV center of mass energy. The background and signal estimates require further consideration, especially after a detailed study has been completed on the possibility of triggering on ç leptons. The results above depend on having high eæciency triggers for central ç leptons with p T é ç 2 GeV. Because the leptonic decay products of the ç lepton are usually quite soft, mass reaches of the chargino depend very sensitively on the trigger capabilities of high ç-multiplicity events ë6ë. 6

8 4 Conclusion The use of ç leptons in supersymmetry goes well beyond the 3ç signal discussed here. High ç multiplicityevents are present in some gauge mediated models when the ~ç is the NLSP ë8, 2, 2ë. In this case two prompt decays of ~ç! ç + G ~ may be present in all superpartner events. Without these ç leptons in the ænal state the supersymmetry events may be quite ordinary, and mimicked by large standard model backgrounds. Tau leptons can also be used to gain signiæcance in supersymmetry Higgs physics observables. For example, many studies of Higgs boson detection at the Tevatron require Wh! lçb ç b, where l = e or ç. Identifying ç leptons in the decays of the W andèor h ë22, 23ë can only help in the search for the Higgs boson. Since supersymmetry generally requires a light Higgs boson with mass less than about 3 GeV, the Tevatron is well-poised to discover it. This requires high luminosity ë24ë and taking advantage of all possible production and decay modes of the Higgs. Nevertheless, the search reach of the pure 3ç signal of supersymmetry is a good benchmark for ç studies. Not only is this ænal state ubiquitous in highly motivated large tan æ models of supergravity and gauge mediated supersymmetry breaking frameworks, but it also may be the only way to see supersymmetry in the more minimal supersymmetry approaches where the ~ç is the only light slepton generation in the spectrum. Acknowledgements: I thank D. Chakraborty and S. Martin for helpful comments. References ëë M. Dine, A. Nelson, Y. Shirman, Phys. Rev. D 5, 362 è995è; M. Dine, A. Nelson, Y. Nir, Y. Shirman, Phys. Rev. D 53, 2658 è996è. ë2ë For models allowing prompt decays see, K. Intriligator, S. Thomas, Nucl. Phys. B 473, 2 è996è. ë3ë H. Baer, K. Hagiwara, X. Tata, Phys. Rev. D 35, 598 è987è; P. Nath, R. Arnowitt, Mod. Phys. Lett. A 2 è987è 33; R. Barbieri, F. Caravaglios, M. Frigeni, M. Mangano, Nucl. Phys. B 367, 28 è99è. 7

9 ë4ë D Collaboration èb. Abbott et al.è, Phys. Rev. Lett. 8, 59 è998è; CDF collaboration èf. Abe et al.è, hep-exè9835. ë5ë H. Baer, C.-h. Chen, C. Kao, X. Tata, Phys. Rev. D 52, 565 è995è; S. Mrenna, G.L. Kane, G. Kribs, J. Wells, Phys. Rev. D 53, 68 è996è. ë6ë H. Baer, C.-h. Chen, M. Drees, F. Paige, X. Tata, Phys. Rev. Lett. 79, 986 è997è; hep-phè ë7ë M. Drees, M. Nojiri, Nucl. Phys. B 369, 54 è992è; V. Barger, M. Berger, P. Ohmann, Phys. Rev. D 49, 498 è994è; G.L. Kane, C. Kolda, L. Roszkowski, J.D. Wells, Phys. Rev. D 49, 673 è994è. ë8ë S. Dimopoulos, S. Thomas, J. Wells, Nucl. Phys. B 488, 39 è997è. ë9ë K. Babu, C. Kolda, F. Wilczek, Phys. Rev. Lett. 77, 37 è996è. ëë J. Bagger, K. Matchev, D. Pierce, R.-J. Zhang, Phys. Rev. D 55, 388 è997è. ëë A. Cohen, D. Kaplan, A. Nelson, Phys. Lett. B 388, 588 è996è; S. Dimopoulos, G. Giudice, Phys. Lett. B 357, 573 è995è. ë2ë N. Arkani-Hamed, H. Murayama, Phys. Rev. D 56, 6733 è997è; K. Agashe, M. Graesser, hep-phè ë3ë See, for example, ref. ë7ë for an explanation of the minimal supergravity parameters. ë4ë The signal events were simulated using ISAJET: F. Paige, S. Protopopescu, in Supercollider Physics, ed. D. Soper èworld Scientiæc, 986è; H. Baer, F. Paige, S. Protopopescu, X. Tata, in Proceedings of the Workshop on Physics at Current Accelerators and Supercolliders, ed. J. Hewett, A. White, D. Zeppenfeld èargonne National Laboratory, 993è hep-phè ë5ë See, for example, ref. ë8ë for an explanation of the minimal gauge mediated parameters. ë6ë L. Hall, R. Rattazzi, and U. Sarid, Phys. Rev. D 5, 748 è994è; G. Anderson, S. Dimopoulos, L. Hall, S. Raby, Phys. Rev. D 47, 372 è993è. 8

10 ë7ë R. Rattazzi, U. Sarid, Nucl. Phys. B 5, 297 è997è. ë8ë D. Chakraborty, private communication. ë9ë Backgrounds are simulated using PYTHIA, T. Sjíostrand, Comp. Phys. Comm. 82 è994è 74. ë2ë S. Ambrosanio, G.L. Kane, G. Kribs, S. Martin, S. Mrenna, Phys. Rev. D 54, 5395 è996è; S. Ambrosanio, G. Kribs, S. Martin, hep-phè9727. ë2ë B. Dutta, S. Nandi, hep-phè9795. ë22ë S. Mrenna, G. Kane, hep-phè ë23ë For other ç observables in extended Higgs sector searches, see M. Drees, M. Guchait, P. Roy, Phys. Rev. Lett. 8, 247 è998è. ë24ë A. Stange, W. Marciano, S. Willenbrock, Phys. Rev. D 5, 449 è994è; S. Kim, S. Kuhlmann, W.-M. Yao, ëimprovement of Signal Signiæcance in Wh! lçb ç b Search at TeV33," in ëproceedings of the 996 DPFèDPB Summer Study on New Directions for High Energy Physics" è996è; S. Mrenna, ANL-HEP-PR è997è. 9

Searches for Physics Beyond the Standard Model at the Tevatron

FERMILAB-CONF-10-704-E-PPD Proceedings of the XXX. Physics in Collision Searches for Physics Beyond the Standard Model at the Tevatron Chris Hays 1 for the CDF and D0 Collaborations (1) Oxford University,