CMS search plans and sensitivity to new physics with dijets
Anwar Bhatti, Benjamin Bollen, Marco Cardaci, Frank Chlebana, Selda Esen, Robert M. Harris, Manoj K. Jha, Konstantinos Kousouris, David Mason, Marek Zielinski
aa r X i v : . [ h e p - e x ] O c t CMS search plans and sensitivity to new physicswith dijets
Anwar Bhatti , Benjamin Bollen , Marco Cardaci , FrankChlebana , Selda Esen , Robert M. Harris , Manoj K. Jha ,Konstantinos Kousouris , David Mason and Marek Zielinski Rockefeller University, New York, NY, USA Universiteit Antwerpen, Antwerp, Belgium Fermilab, Batavia, IL, USA Brown University, Providence, RI, USA University of Delhi, India and INFN, Bologna, Italy University of Rochester, Rochester, NY, USAE-mail: [email protected]
Abstract.
CMS will use dijets to search for physics beyond the standard modelduring early LHC running. The inclusive jet cross section as a function of jet transversemomentum, with 10 pb − of integrated luminosity, is sensitive to contact interactionsbeyond the reach of the Tevatron. The dijet mass distribution will be used to search fordijet resonances coming from new particles, for example an excited quark. Additionalsensitivity to the existence of contact interactions or dijet resonances can be obtainedby comparing dijet rates in two distinct pseudorapidity regions.PACS numbers: 12.38.Qk, 12.60.Rc, 13.87.Ce Submitted to:
J. Phys. G: Nucl. Part. Phys.
MS search plans and sensitivity to new physics with dijets √ s = 14 TeV. These dijet events resultfrom parton scattering, produced by the strong interaction of quarks ( q ) and gluons ( g )inside the protons. This paper discusses plans to use dijets in the search for two signalsof new physics: contact interactions and resonances decaying into dijets. Two modelsof quark compositeness have been considered for this generic search. The first model isa contact interaction [1] among left-handed quarks at an energy scale Λ + in the process qq → qq , modeled with the effective Lagrangian L qq = ( ± π/ Λ )( q L γ µ q L )( q L γ µ q L ) with+ chosen for the sign. The second is a model of an excited quark ( q *) [2] in the process qg → q ∗ → qg , detectable as a dijet resonance. All processes presented here have beensimulated using PYTHIA version 6.4 [3].A detailed description of the Compact Muon Solenoid (CMS) experiment can befound elsewhere [4, 5]. The CMS coordinate system has the origin at the center of thedetector, z -axis points along the beam direction toward the west, with the transverseplane perpendicular to the beam. We define φ to be the azimuthal angle, θ to bethe polar angle and the pseudorapidity as η = − ln(tan[ θ/ | η | < < | η | < | η | < . η = ∆ φ = 0 . η and φ width progressively increases at higher values of η . The energy in the HCALand ECAL within each projective tower is summed to find the calorimeter tower energy.Towers with | η | < . . < | η | < . . < | η | < . E , isdefined as the scalar sum of the calorimeter tower energies inside a cone of radius q (∆ η ) + (∆ φ ) = 0 .
5, centered on the jet axis. The jet momentum, ~p , is thecorresponding vector sum: ~p = P E i ˆ u i with ˆ u i being the unit vector pointing fromthe origin to the energy deposition E i inside the cone. The jet transverse momentum, p T , is the component of ~p in the transverse plane. The E and ~p of a reconstructedjet are then corrected for the non-linear response of the calorimeter to a generated jet.Generated jets come from applying the same jet algorithm to the Lorentz vectors ofstable generated particles before detector simulation. On average, the p T of a correctedjet is equal to the p T of the corresponding generated jet. The corrections estimated froma GEANT [6] simulation of the CMS detector increase the average jet p T by roughly MS search plans and sensitivity to new physics with dijets | η | < .
3. The applied correctionsdepend on jet η as well as p T . The jet measurements presented here are within theregion | η | < .
3, where the sensitivity to new physics is expected to be the highest, andwhere the reconstructed jet response variations as a function of η are both moderateand smooth. Further details on jet reconstruction and jet energy corrections can befound elsewhere [5, 7].The dijet system is composed of the two jets with the highest p T in an event (leadingjets), and the dijet mass is given by m = q ( E + E ) − ( ~p + ~p ) . The estimated dijetmass resolution varies from 9% at a dijet mass of 0.7 TeV to 4.5% at 5 TeV.CMS will record events that pass a first level trigger followed by a high leveltrigger. For an instantaneous luminosity of 10 cm − s − , we consider three eventsamples collected by requiring at least one jet in the high level trigger with correctedtransverse energy above 60, 120 and 250 GeV, prescaled by factors of 2000, 40 and 1,respectively. For an integrated luminosity of 100 pb − , the three event samples willeffectively correspond to 0.05, 2.5, and 100 pb − . The first event sample will be used tomeasure the trigger efficiency of the second sample. The second and third event sampleswill be used to study dijets of mass above 330 and 670 GeV, respectively, for which thetrigger efficiencies are expected to be higher than 99% [8].Backgrounds from cosmic rays, beam halo, and detector noise are expected tooccasionally produce events with large or unbalanced energy depositions. They willbe removed by requiring E T / P E T < . P E T <
14 TeV, where E T ( P E T ) isthe magnitude of the vector (scalar) sum of the transverse energies measured by allcalorimeter towers in the event. This cut is estimated to be more than 99% efficient forboth QCD jet events and the signals of new physics considered. In the high p T regionrelevant for this search, jet reconstruction is fully efficient.CMS plans to search for contact interactions using the jet p T distribution. Figure 1shows simulations of the inclusive jet differential cross section as a function of p T , forjets with | η | <
1. Considering first the QCD processes, the reconstructed and correctedquantities are compared with the QCD prediction for generated jets. After corrections,the reconstructed and generated distributions agree. The ratio of the corrected jet crosssection to the generated jet cross section varies between 1.2 at p T = 100 GeV and 1.05at p T = 500 GeV, remaining roughly constant for higher p T . The deviation of this ratiofrom 1 is attributed to the smearing effect of the jet p T resolution on the steeply fallingspectrum. The measured spectrum in data could be further corrected for resolutionsmearing, and this ratio from simulation is an estimate of the size of that correction.The measurement uncertainties are predominantly systematic. The inset in Fig. 1 showsthe effect on the jet rate of a 10% uncertainty in the jet energy correction. Fig. 2 alsoshows the effect of this uncertainty on a lowest order QCD calculation. This level ofjet energy uncertainty could be expected in early running, for an integrated luminosityaround 10 pb − . This experimental uncertainty is roughly an order of magnitude largerthan the uncertainties from parton distributions, as estimated using CTEQ6.1 fits [9]and shown in Fig. 2. Figures 1 and 2 show that the effect of new physics from a contact MS search plans and sensitivity to new physics with dijets + = 3 TeV is convincingly above what could be expected formeasurement uncertainties with only 10 pb − . For comparison, a Tevatron search hasexcluded contact interactions with scales Λ + below 2.7 TeV [10]. The results of thelowest order calculations in Fig. 2 are the same as the simulation results in the inset toFig. 1.CMS plans to search for narrow dijet resonances using the dijet mass distribution.Figure 3 shows the differential cross section versus dijet mass, where both leading jetshave | η | <
1, and the mass bins have a width roughly equal to the dijet mass resolution.Considering first the QCD processes, the cross section for corrected jets agrees with theQCD prediction from generated jets. To determine the background shape either theMonte Carlo prediction or a parameterized fit to the data can be used. The inset toFig. 3 shows a simulation of narrow dijet resonances with a q * production cross section.For q * masses of 0 .
7, 2 . . | η | < .
01 and0 . − a q * dijet resonance with a mass of 2 TeV would produce aconvincing signal above the statistical uncertainties from the QCD background. Forcomparison, a Tevatron search has excluded q * dijet resonances with mass, M, below0 .
87 TeV [11]. The heaviest dijet resonances that CMS can discover (at five standarddeviations) with 100 pb − of integrated luminosity, using this search technique andincluding the expected systematic uncertainties [12, 13], are: 2.5 TeV for q *, 2.2 TeVfor axigluons [14] or colorons [15], 2.0 TeV for E diquarks [16], and 1.5 TeV for coloroctet technirhos [17]. Studies of the jet η cut have concluded that the optimal sensitivityto new physics is achieved with | η | < . q ¯ q [18].CMS plans to search for both contact interactions and dijet resonances using thedijet ratio, r = N ( | η | < . /N (0 . < | η | < . N is the number of events withboth jets in the specified | η | region. The dijet ratio is sensitive to the dijet angulardistribution. For the QCD processes, the dijet ratio is the same for corrected jetsand generated jets, and is constant at r = 0 . − of integratedluminosity, including trigger prescaling. The signal from a contact interaction with scaleΛ + = 5 TeV rises well above the QCD statistical errors at high dijet mass. Systematicuncertainties in the dijet ratio are expected to be small, since they predominantly cancelin the ratio as previously reported [12, 19]. Using the dijet ratio, CMS can discover acontact interaction at scale Λ + = 4, 7 and 10 TeV with integrated luminosities of 10,100, and 1000 pb − , respectively [18]. The signal from a 2 TeV spin 1/2 q * producesa convincing peak in the dijet ratio, because it has a significant rate and a relativelyisotropic angular distribution compared to the QCD t -channel processes. Fixing thecross section of the 2 TeV dijet resonance for | η | < . . q * model),the dijet ratio in the presence of QCD background increases by approximately 6% when MS search plans and sensitivity to new physics with dijets q ¯ q and gg (such as a Randall-Sundrumgraviton [20]), and the dijet ratio decreases by approximately 4% when considering aspin 1 resonance decaying to q ¯ q (such as a Z ′ , axigluon, or coloron) [18]. Hence, thesensitivity to a 2 TeV dijet resonance depends only weakly on the spin of the resonance.To measure the spin, we need both the dijet ratio and an independent measurement ofthe cross section of the resonance, for example, from the dijet mass differential crosssection. Nevertheless, with sufficient luminosity, this simple measure of the dijet angulardistribution, or a more complete evaluation of the angular distribution, can be used tosee these small variations and infer the spin of an observed dijet resonance.In conclusion, CMS plans to use measurements of rate as a function of jet p T anddijet mass, as well as a ratio of dijet rates in different η regions, to search for new physicsin the data sample collected during early LHC running. With integrated luminositysamples in the range 10–100 pb − , CMS will be sensitive to contact interactions anddijet resonances beyond those currently excluded by the Tevatron. Acknowledgments
We gratefully acknowledge the assistance of our colleagues at CMS in preparing thispaper. For their constructive comments and guidance, we would like to thank ouranalysis review committee: Richard Cavanaugh, Dan Green and Luc Pape. Weappreciate the many useful comments on prose and style provided by Carlos Lourencoand the CMS publications committee. Finally, we would like to thank Gigi Rolandi forhis encouragement and for leading the collaboration wide review of this paper, whichserved as a test of the publication process at CMS.
References [1] Eichten E, Lane K and Peskin M 1983
Phys. Rev. Lett. Int. J. Mod. Phys. A2 preprint hep-ph/0108264[4] CMS Collaboration 2008 The CMS experiment at the CERN LHC (accepted for publication in JINST) [5] CMS Collaboration 2006 Physics TDR Volume I CERN-LHCC-2006-001 [6] Agostinelli S et al. (GEANT4 Collaboration) 2003
Nucl. Instrum. Meth. A
CMS Physics Analysis Summary
JME-07-003 athttp://cms-physics.web.cern.ch/cms-physics/public/JME-07-003-pas.pdf[8] Esen S and Harris R 2006
CMS Note [9] Pumplin J et al. (CTEQ Collaboration) 2002,
JHEP et al. (D0 Collaboration) 1999 Phys. Rev. Lett. et al. J. Phys. G: Nucl. Part. Phys. CMS Note [14] Bagger J, Schmidt C and King S 1988
Phys. Rev. D Phys. Rev. D Phys. Rep.
193 and references therein
MS search plans and sensitivity to new physics with dijets [17] Lane K and Mrenna S 2003 Phys. Rev. D CMS Physics Analysis Summary
SBM-07-001 athttp://cms-physics.web.cern.ch/cms-physics/public/SBM-07-001-pas.pdf[19] Esen S and Harris R 2006
CMS Note [20] Randall L and Sundrum R 1999
Phys. Rev. Lett. MS search plans and sensitivity to new physics with dijets (GeV) T Jet p0 500 1000 1500 2000 2500 3000 3500 ( pb / G e V ) T / dp s d -7 -6 -5 -4 -3 -2 -1
10 110 (GeV) T Jet p
200 400 600 800 1000 1200 1400 N u m b e r o f Je t s / G e V CMS Simulation
Figure 1.
The inclusive jet p T differential cross section expected from QCD for | η | <
1, for generated jets (points), reconstructed jets (triangles), and corrected jets(open circles). The inset shows the number of generated jets expected in 50 GeV binsfor an integrated luminosity of 10 pb − . The standard QCD curve (solid) is modified bya signal from contact interactions with scale Λ + = 3 TeV (dotted) and 5 TeV (dashed).The shaded band represents the effect of a 10% uncertainty on the jet energy scale. MS search plans and sensitivity to new physics with dijets (GeV) T Jet p200 400 600 800 1000 1200 1400 F r ac t i on a l D i ff e r e n ce f r o m Q CD -1-0.500.511.522.53 LO CalculationsQCD & Stat. Err.Energy Err. (10%)PDF Err. (CTEQ 6.1) = 3 TeV + L = 5 TeV + L Figure 2.
The fractional difference from the QCD jet rate resulting from a 10%uncertainty on the jet energy scale (dashed), uncertainties in parton distributions(dotted), and signals from contact interactions with scale Λ + = 3 TeV (boxes) andΛ + = 5 TeV (triangles). Statistical uncertainties expected for an integrated luminosityof 10 pb − (vertical bars) are shown on the QCD prediction (points). MS search plans and sensitivity to new physics with dijets Dijet Mass (GeV)0 1000 2000 3000 4000 5000 6000 7000 / d m ( pb / G e V ) σ d -6 -5 -4 -3 -2 -1
10 110
10 CMS Simulation
Dijet Mass (GeV) ( S i gn a l - Q CD ) / Q CD -0.100.10.20.30.40.50.60.7 Figure 3.
The dijet mass differential cross section expected from QCD for | η | < q *signals [13] of mass 0 .
7, 2, and 5 TeV. The fractional difference (histogram) betweenthe q * signal and the QCD background is compared to the statistical uncertainties inthe QCD prediction (vertical bars) for an integrated luminosity of 100 pb − . MS search plans and sensitivity to new physics with dijets Dijet Mass (GeV)500 1000 1500 2000 2500 3000 3500 | < . ) η | < . ) / N ( . < | η D ij e t R a t i o = N ( | Figure 4.
The dijet ratio for corrected jets expected from QCD (horizontal line),with statistical uncertainties (vertical bars) for an integrated luminosity of 100 pb − ,is compared to QCD + contact interaction signals with a scale Λ + = 5 TeV (dashed)and 10 TeV (dotted), as well as to QCD + dijet resonance signals (histogram) with qq