Search for new physics in photon and jet final states
aa r X i v : . [ h e p - e x ] M a y SEARCHES FOR NEW PHYSICS IN PHOTON AND JET FINAL STATES
M. Jaffr´eon behalf of the CDF and D0 collaborations
LAL, Universit´e Paris-Sud, CNRS/IN2P3,91898 Orsay Cedex, France
Recent results from searches of physics beyond the standard model in p ¯ p collisions are reported,in particular, reactions involving high transverse momentum photons or jets in their final state.Data analyzed by the CDF and D0 experiments at the Run II of the Tevatron correspond tointegrated luminosities between 1 and 2 fb − depending of the analyses. At an energy frontier collider, the usual way to search for indices of physics beyond the standardmodel (SM) is to look for the collisions with the highest momentum-transfer particles. Typically,one chooses a particular model, and the event selection is optimized to enhance its contributionagainst the SM expectation. The absence of any deviation in data provides a limit on theproduction cross-section times the branching ratio for the channel under study, which is thentranslated into exclusion limits in the parameter space of this model. However, by nature, anew phenomena is unknown, and it exists a lot of models at disposal. This is the reason whichmotivates the ”signature-based” search strategy which casts a wider look for deviations to theSM.Both strategies will be reported here for final states with photons and jets.
Many models with extra spatial dimensions have been proposed to solve the hierarchy problem.In the RS model 1, the SM brane and the Planck brane are separated by an extra dimensionwith a warped geometry. Only the graviton is allowed to propagate in this extra dimension. Itappears as Kaluza-Klein (KK) towers in the SM brane. This model has only 2 parameters: M ,the mass of the lowest KK excited mode, and k/M P l , a dimensionless coupling constant whosevalue should lie between 0.01 and 0.1.KK towers couple to any boson or fermion pairs. CDF 2 looks separately at γγ and ee finalstates, whereas D0 3 looks at both final states at once as they look similar in the electromagneticcalorimeter. Because of the spin 2 of the graviton, the ratio of the branching ratios to γγ and ee final states is 2. Both experiments have analyzed about the same amount of data ( ∼ f b − ),and found no excess of events over the SM predictions ( Drell-Yan and QCD where jets aremisidentified as photons) excluding graviton masses below 900 GeV /c for k/M P l = 0 .
1. Fig. 1shows the excluded contour in the 2D parameter space as measured by D0. (GeV) Graviton mass M200 300 400 500 600 700 800 9000.010.020.030.040.050.060.070.080.090.1 excluded at 95% CLexpected limitD0 PRL 95, 091801 (2005)excluded by precision ewk (GeV) Graviton mass M200 300 400 500 600 700 800 900 P l M k / -1 DØ 1 fb
Figure 1: 95% C. L. upper limiton k/M Pl versus graviton mass com-pared with the RS expected limit. (GeV) T Missing E E v en t s pe r G e V -1 data =75 TeV L SM + signal =90 TeV L SM + signal gg gg
W/Z+electron mis-IDjet mis-ID (GeV) T Missing E E v en t s pe r G e V -1 -1 ˘ D (TeV) L
70 75 80 85 90 95 100 105 110 ( f b ) s [GeV] c m100 120 140 [GeV] c m180 200 220 240 260 -1 ˘ D NLO cross-sectionobserved limitexpected limit s – expected limit s – expected limit Figure 2: The E / T distribution in γγ data with the various background con-tributions (left). Predicted cross section for the Snowmass Slope model ver-sus Λ. The observed and expected 95% C.L. limits are also shown (right). SUSY 4 is a broken symmetry. Experimental signatures are determined through the manner andscale of the SUSY breaking. In the GMSB scenario, the lightest supersymmetric particle (LSP) isthe gravitino, a very light and weakly interacting particle. The next to lightest supersymmetricparticle (NLSP) is assumed in this analysis to be the neutralino which decays into the LSPand a photon. Assuming R-parity conservation 5, SUSY particles are pair produced and theexperimental signature will be 2 photons and missing energy from the 2 gravitinos. To geta quantitative result, the ”Snowmass Slope SPS 8” model 6 is considered. All the GMSBparameters a are fixed as a function of the effective energy scale Λ of SUSY breaking.In this event topology, the SM background is the Zγγ production where the Z boson decaysinto neutrinos. There is also important instrumental background from events with real E / T ( W boson production) and fake E / T ( QCD where jets are misidentified as photons). Fig. 2 (left)shows the E / T distribution. The observed distribution agrees well with the SM prediction; theentire spectrum is then used to set limits on the GMSB production cross section. Fig. 2 (right)shows the 95% C.L. cross section limit as a function of the effective scale Λ obtained by D0 7.The observed limit on the signal cross section is below the prediction of the Snow-mass Slopemodel for Λ < . T eV , or for gaugino masses m ˜ χ < GeV /c and m ˜ χ ± < GeV /c . The hierarchy problem can also be solved by postulating the existence of n new large extradimensions as proposed first by Arkani-Hamed, Dimopoulos and Dvali8 (ADD); the extra volumeserves to dilute gravity so that it appears weak in our 3D world as the graviton is the only particleallowed to propagate in the extra space. If the extra dimensions are compactified in a torus ofradius R, according to the Gauss law, one can relate the fundamental Planck mass scale M D ,R, the Planck mass and the number of extra dimensions by the relation M P lanck = 8 πM n +2 D R n ,allowing M D to be compatible with the electroweak scale.In this model, the graviton can be produced directly in the reaction q ¯ q → Gγ ; G will remainundetected leaving a signature with a single photon and E / T .The only SM background is the Zγ production where the Z boson decays into a neutrino pair.In addition to the usual instrumental background coming from misidentification of electrons or a The messenger mass M m = 2Λ, the number of messengers N = 1, tan ( β ) = 15, µ > ets into photons, the event topology is rather sensitive to a contribution from beam halos andcosmics where muons produced photons by bremsstrahlung.To fight the latter background, both experiments had to develop specific tools in additionto the usual ones based on the EM shower profile. Special hit finders in the tracker startingfrom the EM cluster increase the track veto efficiency. In addition, D0 uses a EM pointing toolthanks to its preshower detector, and CDF the timing system built within its EM calorimeter.The results of CDF which has analysed about 2 f b − of data, twice as much as D0 9, aredisplayed in Fig. 3. The left plot shows a good agreement for the photon transverse betweendata and the sum of the various backgrounds. This allows to set limits on the fundamentalscale M D (right plot) as a function of the number of extra dimensions. For n > (GeV) T Photon E
40 60 80 100 120 140 160 180 200 E v en t s / B i n (GeV) T Photon E
40 60 80 100 120 140 160 180 200 E v en t s / B i n -1 CDF Run II Preliminary, 2.0 fb nnfi , Z g Z g fi t / m e/ fi W g , lost gg Non-Collision t / m , lost e/ g W/ZADD n=4 m=0.8 TeV
Number of Extra Dimensions Lo w e r L i m i t ( T e V ) D M Number of Extra Dimensions Lo w e r L i m i t ( T e V ) D M CDF Run II Preliminary ) -1 (2.0 fb T E + g CDF ) -1 (1.1 fb T ECDF Jet + LEP Combined
Figure 3: The left figure shows the E / T distribution in the CDF monophoton search. The signal expected fromthe ADD model (n=4, m=0.8TeV) is added on top of the SM backgrounds. The right figure shows the exclusionlimits for the ADD model obtained in this analysis, in comparison with the CDF jet and E / T result and LEPcombined result. Squarks and gluinos can be copiously produced at the Tevatron if they are sufficiently light. Theanalysis is performed within the mSUGRA model 11. The final state is composed of jets with alarge E / T due to the two escaping neutralinos, assumed to be the LSP. According to the relativemass of squarks and gluinos different event topologies are to be expected. If squarks are lighterthan gluinos, a ”dijet” topology is favored. On the contrary if squarks are heavier than gluinos,the final state contains at least 4 jets. Finally, the jet multiplicity is at least 3 if squarks andgluinos have similar masses. After a common event preselection, the three topologies have beenstudied and optimised separately. The left plot on Fig. 4 shows the D0 E / T distribution obtainedin the ”dijet” search, the right one is obtained by CDF in the ”3-jet” search. D0 has analyzed2 . f b − of data without finding any excess over the SM predictions. It allows to extend theexclusion domain in the squark gluino plane (Fig. 5). Using the most conservative hypothesis,D0 12(CDF 13) excludes a gluino lighter than 308(290) GeV /c . (GeV) T E E ve n t s / G e V -1 (GeV) T E E ve n t s / G e V -1 -1 DØ, L=2.1 fbDataSM BackgroundFitted QCDSUSY(a) [GeV] T missing-E
50 100 150 200 250 300 350 eve n t s / G e V -1
3 MET>120 HT>330 ‡ jet N CDF Run II Preliminary ) -1 Data (L = 2.0 fbQCD + non QCD Bkg.non QCD Bkg.Total Syst. Uncertainty = 249 GeV/c g ~ Bkg.+Sig. M = 270 GeV/c s ~ M [GeV] T missing-E
50 100 150 200 250 300 350 eve n t s / G e V -1 Figure 4: Distributions of E / T after applying all analysis criteria except the one on E / T for the “dijet” (D0 left)and “3-jets” (CDF right) squark-gluino analyses; data (points with error bars) and the cumulated contributionsfrom SM background, QCD background and signal MC are shown. Due to the large Yukawa coupling, there could be a large mixing in the 3rd generation of squarks.The lighest of the 2 stops could be the lightest squark and even the NLSP. Furthermore if itsmass is less than the sum of the masses of the b quark, the W boson, and the neutralino, thedominant decay mode is ˜ t → c ˜ χ , a flavor changing loop decay, assumed to be 100% in theanalysis. The final state will then be 2 acoplanar charm jets and E / T . The analysis proceedswith 2 jets detected in the central part of the detector with a loose heavy quark tag for one ofthem. No excess of events has been observed 14 in about 1 f b − of data, which provides a lowerlimit for the stop mass at 149 GeV /c for a neutralino mass of 63 GeV /c (Fig. 6). Gluino Mass (GeV) S qu a r k M ass ( G e V ) -1 DØ, L=2.1 fb <0 m =0, =3, A b tan UA UA LEP CD F I BD Ø I A DØ IB
DØ II no mSUGRAsolution – c~ LEP2 – l~ LEP2
Figure 5: Excluded region in the squark and gluinomass plane; newly excluded domain by D0 is shownin dark shading. The region where no mSUGRAsolution can be found is shown hatched. (GeV) t ~ m
40 60 80 100 120 140 ( G e V ) c~ m c + m c ~ = m t ~ m o c ~ + m b + m W = M t ~ m o = 56 q LEP o = 0 q LEP -1 CDF Run II 295 pb -1 D0 Run II 360 pbObserved Expected -1 DØ, L = 995 pb
Figure 6: Region in the ˜ t – ˜ χ mass plane excludedat the 95% C.L. by the D0 search. The yellow bandrepresents the theoretical uncertainty on the scalartop quark pair production cross section due to PDFand renormalization and factorization scale choice. Signature-based searches
In its quest for “signature-based” excess, CDF has searched for anomalous production of eventsin the γγ + E / T topology. In this analysis, use is made of the E / T resolution model. The aimof this model is to discriminate events with large mismeasured E / T from events with real E / T .It has been shown to provide a better background rejection power than a simple E / T cut. Thismodel is based on the assumption that individual particle’s energy resolution has Gaussian shapeproportional to particle’s √ E T . Only two sources of fake E / T are considered : soft unclusteredenergy ( from underlying event and multiple interactions ), and jets. The latter is responsiblefor most of the E / T as it is collimated energy in contrast to the former which is spread outall over the calorimeter. According to this model, each event is given a E / T significance value.Most of the QCD background is eliminated by requiring a significance above 5, leaving only theexpected number of SM events with real E / T (Fig. 7), and not much room for an extra signal. Figure 7: Distribution of missing transverse energysignificance for diphoton candidates. Figure 8: The measured dijetmass spectrum andresults of the fit to the parametrization form 1.
Many classes of models beyond the SM predict the existence of new massive particles decayinginto 2 partons which would appear as resonances in the dijet mass spectrum. Such classes includeexcited quarks, techniparticles, new W’ or Z’ bosons, RS graviton,... Jets are reconstructed bythe cone-based midpoint jet algorithm 15 with a cone radius of 0 . | y | < . f b − of data and measured the dijet differential crosssection (Fig. 8). The spectrum is fitted by the smooth parametrization : dσdm = p (1 − x ) p /x p + p log( x ) , x = m/ q ( s ) . (1)This parametrization is found to fit well the dijet spectra from PYTHIA and HERWIG MCevents as well as from NLO pQCD. As no evidence for existence of a new massive particle isobserved, limits on new particle production cross sections can be derived as a function of thedijet mass. These limits are then translated into mass exclusion limits, see Table 1. No hints of physics beyond the SM have been found so far. As the Tevatron is continuing toprovide experiments with more data to analyze, the quest for indices will be pursued by CDF able 1: Mass exclusion ranges for several models.
Model description Observed mass exclusion range(
GeV/c )Excited quark ( f = f ′ = f s = 1) 260-870Color octet technirho(top-color-assisted-technicolor couplings) 260-1110Axigluon and flavor universal coloron(mixing of 2 SU(3)’s cot(theta)=1) 260-1250E6 diquark 290-630W’ (SM couplings) 280-840Z’ (SM couplings) 320-740 and D0. Some analyses presented in this talk have already been published, for the others, furtherdetails can be found at: CDF D0 Acknowledgments
The author would like to thank the CDF and D0 working groups for providing the material forthis talk, and the organizers of the
Rencontres for a very enjoyable conference and the excellentorganization.
References
1. L. Randall and R. Sundrum,
Phys. Rev. Lett. , 3370 (1999); , 4690 (1999).2. T. Aaltonen et al (CDF Collaboration), Phys. Rev. Lett. , 171801 (2007); , 171802(2007).3. V.M. Abazov et al (D0 Collaboration), Phys. Rev. Lett. , 091802 (2008).4. H.E. Haber and G.L. Kane, Phys. Rep. , 75 (1985).5. G.R. Farrar and P. Fayet,
Phys. Lett. B , 442 (1978).6. S.P. Martin, S. Moretti, J.M. Qian, and G.W. Wilson, “Direct Investigation of Supersym-metry: Subgroup summary report,” in Proceedings of the APS/DPF/DPB Summer Studyon the Future of Particle Physics (Snowmass 2001), edited by N. Graf, eConf