1 Summary of the Dijet Topology Group Parallel Session Robert M. Harris Fermilab JTERM III January 16, 2009

Robert Harris, Fermilab2 Agenda l Morning Session (Dijets & Background) 10:30CRAFT Analysis Kwangzoo Chung (University of Iowa) 10:55Inclusive Jet PT David Mason (Fermilab) 11:20Dijet Mass Konstantinos Kousouris (Fermilab) 11:45Dijet Ratio Dan Miner (University of Rochester) 12:10Dijet Angular Distribution Agata Smoron (UIC) l Afternoon Session (mainly Multijets) 13:30Dijet Azimuthal DecorrelationCosmin Dragoiu (UIC) 13:55Multijet QCD Studies Sudaveep Bose (Tata Institute) 14:20Multijet SUSY Studies Burak Bilki (University of Iowa) 14:45Multijet Resonance Search Amitabh Lath (Rutgers) 15:10 Jet Shapes Pelin Kurt (University of Cukurova)

Robert Harris, Fermilab3 Introduction l We had two lively and interesting sessions. è Mainly the speakers attended, but there were a few others here and on EVO. à About 15 participants total è A nice mixture of new and established analyses. è Some new participants to the dijet topology group. l There was a lot of material presented in 10 talks. è I can only show 1-2 slides for each talk in this summary. è There are a lot more results in the full talks on indico at

Robert Harris, Fermilab4 CRAFT Analysis (Kwangzoo Chung & Kerem Cankocak) l L1 jet triggers in CRAFT è Stable rate of 200 Hz. l Cosmic Ray Studies Underway è Craft data comparisons with Cosmic Ray MC are beginning. è Concludes that correct pedestal calculation is crucial ! Cosmic MC CRAFT

Robert Harris, Fermilab5 Inclusive Jet Cross Section (Dave Mason & Pratima Jindal) Event cleanup using MET/  E T è QCD jets will balance in p T. è Backgrounds from cosmic rays, detector noise, and beam halo won’t è These will be a significant background for the highest p T jets. è We plan to employ a cut on MET/SET to reduce these backgrounds.  Previous studies indicated the selection MET/  E T < 0.3 was more than 99% efficient for QCD and remove all the high p T jets in global run data (noise and cosmic triggers). à Recent studies indicate lower efficiency in QCD so we may need to loosen this cut.

Robert Harris, Fermilab6 Inclusive Jet Cross Section (Dave Mason & Pratima Jindal) Started studying MET /  E T in global run data again. è Run demonstrated a problem causing large MET.  Discovered that raw MET /  E T was unusually large for every event! à Only expected occasional large MET. è Caused by a single hot channel in the HF: 25 GeV P T jet every event !  Masking the hot channel in the analysis offline improves MET/  E T. è Cleanup will be required before we can even use MET to reject events ! Hot Channel in HF MET /  E T

Robert Harris, Fermilab7 Dijet Mass (Kostas Kousouris) l Mass Spectrum Measurement è Pythia Summer08 fullsim at 10 TeV  Find two leading jets with |  |<1.3  Correct jet energy vs  and p T è Combine jet triggers for spectrum è Mass reach for 10 pb -1 is 3 TeV è Resolution for SISCone R=0.7 è Resolution unsmearing is small Spectrum and Triggers Resolution Unsmearing Resolution

Robert Harris, Fermilab8 l Theory & Experiment Uncertainties è PYTHIA compared with LO & NLO à K. Hatakeyama using NLOJET++ à ~20-40% variations due to hadronization and higher orders è Experimental error dominated by Jet Energy Scale à Unsmearing uncertainty small. Dijet Mass (Kostas Kousouris) Experimental Uncertainties Corrected Spectrum and Theory

Robert Harris, Fermilab9 Dijet Ratio = N(|  |<0.7) / N(0.7<|  |<1.3) è Small systematic uncertainties and sensitive to angular distributions è Same sample, algorithm and mass binning as dijet mass analysis. è Dijet ratio from corrected calojets agrees with GenJets and is ~ 0.5 è Systematic uncertainty due to relative jet energy scale in barrel is small Dijet Ratio (Daniel Miner) Dijet Ratio from QCD Systematic Error on Dijet Ratio

Robert Harris, Fermilab10 Dijet Ratio (Daniel Miner) l Dijet Resonances with Dijet Ratio è 2 TeV q* resonance produced in summer08 fullsim sample è Pure signal has dijet ratio of 2, QCD has dijet ratio of 0.5 è Combined signal and QCD has dijet ratio shown below è Statistical error bars are for 100 pb -1 ; points below fluctuate with MC statistics. è Signal shown is likely near the edge of our statistical sensitivity (work ongoing). q* QCD 100 pb -1 Dijet Ratio from QCD and 2 TeV Excited Quark

Robert Harris, Fermilab11 Dijet Angular Distribution (Agata Smoron, Len Apanasevich, Nikos Varelas) Angular distribution dN/d  è From CSA08 samples è Uses data in barrel & endcap. è Reconstructed and generated distribution are in good agreements Center of Momentum Frame ** Jet 1 + cos  * 1 – cos  *  =

Robert Harris, Fermilab12 Dijet Angular Distribution (Agata Smoron, Len Apanasevich, Nikos Varelas) l QCD test è Statistical errors with 10 pb -1 è Early estimates of systematic errors are also shown è PYTHIA compares well with LO QCD l Contact Interaction Search è Contact interaction is more isotropic than QCD  Produces more events at low   Can clearly discover a   = 3 TeV contact interaction with only 10 pb -1 at √s = 10 TeV  Best D0 limit is  + > 2.7 TeV from the dijet ratio in run 1. à Preliminary CDF and D0 run 2 results using the angular distribution are not better...

Robert Harris, Fermilab13 Dijet Azimuthal Decorrelation (Cosmin Dragoiu, Len Apanasevich, Nikos Varelas)  of the leading two jets è Sensitive to the presence of initial and final state radiation è Provides test of NLO QCD and a good measurement to tune the amount of radiation in the MC l Analysis è CSA08 sample, SISCone R=0.5 jets. è Six p T bins of the leading jet è No cut on the 2 nd jet pt è Requires |y| < 1.1 for both jets l Systematics è Insensitive to CMS jet position resolution. è Some sensitivity to jet energy resolution for jet p T < 250 GeV  Causes switching between 2 nd and 3 rd jet, giving smaller . 

Robert Harris, Fermilab14 Multijet QCD Studies (Sudaveep Bose) l Full analysis of 3 and 4 jet system underway è Scaled energies of jets in the CM frame: x i = 2E i /Mass è All angles. è Requires lead jet p T > 110 GeV and every other jet to have p T >50 GeV è Currently studying systematic effects. 3 Jet Production4 Jet Production x 3 : Leading Jet in 3 Jet Events

Robert Harris, Fermilab15 Multijet QCD Studies (Sudaveep Bose) l Algorithm Dependence è Scaled energies after jet corrections are sensitive to jet algorithm choice è Some unexpected results on correspondence between K T and Siscone à Here K T with D parameter 0.6 looks like Siscone with R=0.5 à For Dijet system K T with D parameter 0.4 looks like Siscone with R=0.5 x 3 : Leading Jet in 3 Jet Events

Robert Harris, Fermilab16 Multijet SUSY Studies (Burak Bilki) l Comparison of the Multijet variables for SUSY & Background è Hoping to find additional variables to help isolate signal è All the standard multijet variables look similar for SUSY & Background x 5 : Least Energetic Jet in 3 Jet Events Cos  3 : Angle of leading jet in 3 Jet Events QCD SUSY

Robert Harris, Fermilab17 Multijet SUSY Studies (Burak Bilki)  and  T variable for multijet system  The variable  and  T for dijets has been shown to discriminate SUSY from QCD without using MET. è This can be extended to the multi-jet system (3 or greater). à By combining the smaller jets to get a dijet system. à Many methods of combining the smaller jets were studied. Dijet Trijet  T = E T2 / M T  T using hemispheres QCD SUSY

Robert Harris, Fermilab18 Multijet Resonances (Amitabh Lath) l Pair production of particles Q è Each Q decays to 3 jets. è pp  Q Q  3j+3j = 6j è Modeled with PYTHIA gluino pair production followed by r-parity violating decays (No MET) è They investigate a few masses for Q è Model QCD background with Alpgen l Selection for M Q = 290 GeV  Six jets with p T >60 GeV & |  |<3   p T (6J) > 600 GeV è Form all pairs of 3 jets è Require each Q decay be boosted   p T (3J) > M 3J (combinatorics) è Convincing s/√b = 15 for 1 fb -1 à Marginal s/√b = 5 for 100 pb -1 g g ~ ~ q q q q ~ ~ q q q q jet g g ~ ~ q q q q ~ ~ q q q q M 3J QCD 6J QCD + Q (290) 1 fb -1

Robert Harris, Fermilab19 Multijet Resonances (Amitabh Lath) l Higher mass resonances also give signals over QCD è However currently need to tailor the cuts for each resonance in order to beat down the combinatorics. l Also studying systematic uncertainties on QCD & jet resolution 1 fb -1 M Q =290 Cuts: 600_60_200 M Q =420Cuts: 700_90_200 1 fb -1 M Q =660Cuts: 1100_90_300 1 fb -1 Jet Resolution Systematic Study

Robert Harris, Fermilab20 Jet Shapes (Pelin Kurt) l Presentation of mature and approved CMS analysis.   (r) = average fraction of jet energy in a cone of radius r. è Jet shapes are in good agreement for GenJets and Corrected Jets à After generic jet corrections and special jet shape corrections. è The jet shape gets narrower with increasing jet pt à Partly because parton showers are more collimated with increasing p T. à Partly because the number of gluon initiated jets decreases. è Much more in the talk, including new results from NLOJET++. 79% gluon 38% gluon

Robert Harris, Fermilab21 Conclusions l CMS has been studying the classic QCD tests and searches for new physics with dijets. è Inclusive Jet p T è Dijet Mass Measurements. è Dijet Angular Measurements. l CMS is just beginning to study multijet physics. è Basic QCD tests, searches for resonances and supersymmetry. l But there is still a lot to do and all topics need help!! è We welcome new ideas and new people. è Join us at Dijet Topology Group meetings alternate Tuesdays at 1:30. à We are integrated into the CMS JetMET, QCD and Exotica groups. à We have a proven record of getting results approved at CMS. è Pick an analysis and get involved !