QCD dijet analyses at DØ Pavel Demine for the DØ collaboration DAPNIA-SPP, CEA Saclay, France Contents: DØ detector Jet reconstruction at DØ Dijet mass spectrum Dijet azimuthal angle decorrelation XII International Workshop on Deep Inelastic Scattering
DØ Calorimeter uniform and hermetic full coverage up to η < 4.2 compensating (e/π ) fine segmentation: η ϕ = 0. 0. at EM shower max: 0.05 0.05 Changes from Run I Run II: shorter time between bunch crossings (396 ns) faster trigger and readout electronics more material in front of calorimeter (magnet, new tracker) new preshower detector DIS 2004 Pavel Demine DAPNIA-SPP
Jets reconstruction at DØ Clusters: Use Run II cone algorithm (G.C. Blazey et al., hep-ex/000502) Combine particles in a R = η 2 + φ 2 cone Use the 4-vector of every tower with E T > 0.5 GeV as a seed Rerun using the midpoints between pairs of jets as seeds Overlapping jets merged if the overlap area contains more than 50 % of lower p T jet, otherwise particles assigned to nearest jet E-scheme recombination: summing 4-vectors of particles massive jets y J = 2 ln EJ + p J z E J p J z φ J = tan pj y p J x DIS 2004 Pavel Demine DAPNIA-SPP 2
Jet energy scale at DØ Correction of the jet energy measured on the detector level to the jet energy on the particle level Offset (O) energy not associated with the hard interaction (U noise, previous events, additional p p interaction) Response (R jet ) calorimeter response to the jet EM part calibrated on Z ee mass peak measured from E T balance in γ + jet events measured for energies up to 250 GeV Showering (S) losses due to showering the energy in the calorimeter out of the jet cone E jet ptcl = Ejet det O R jet S DIS 2004 Pavel Demine DAPNIA-SPP 3
Dijet mass measurement probe of QCD proton structure at large x resonances existence quark compositeness sample definition: L 43 pb (±6.5 %) R = 0.7 cone jets y jet < 0.5 [nb/gev] dσ/dm JJ - -4-5 -6-7 -8-9 - - NLO QCD (JETRAD) - CTEQ6 at at cone R=0.7, R sep =.3 η < 0.5 jet max = 0.5 E T = µ R µ F s =.8TeV s =.96TeV 200 400 600 800 00 200 400 [GeV] M JJ Jet selection ε 98 %: based on jet energy fraction in EM and Had. parts of the calorimeter Vertex selection ε 80 %: DIS 2004 Pavel Demine DAPNIA-SPP 4
Highest dijet mass event jet jet 2 p T = 66 GeV/c p T = 557 GeV/c η = 0.9 η = 0.25 φ = 0.65 φ = 3.78 M J J 2 = 206 GeV/c 2 DIS 2004 Pavel Demine DAPNIA-SPP 5
Dijet cross section: result <dσ/dm JJ >, pb/gev 2 - - DØ Data, L = 43 pb NLO (JETRAD) CTEQ6M max R sep =.3, µ R = µ F = 0.5 p T -4-5 < 0.5 cone R = 0.7, y jet 200 400 600 800 00 200 400 2 M JJ, GeV/c NLO pqcd is in agreement with the measurement over 7 orders of magnitude DIS 2004 Pavel Demine DAPNIA-SPP 6
Dijet cross section: compare to NLO data / theory 3.5 3 2.5 2 NLO (JETRAD) CTEQ6M max R sep =.3, µ R = µ F = 0.5 p T systematic uncertainty pdf uncertainty.5 0.5 0 < 0.5 cone R = 0.7, y jet 200 400 600 800 00 200 400 2 M JJ, GeV/c NLO pqcd calculations agree with the data within uncertainty systematic uncertainty is dominated by jet energy scale uncertainty: 50 60 GeV: +30 % 23 % 800 400 GeV: +20 % 68 % DIS 2004 Pavel Demine DAPNIA-SPP 7
Dijet azimuthal angle decorrelation φ between two leading jets provides a clean laboratory for testing QCD simple to define and understand much easier to measure a jet direction than its energy φ distribution is directly sensitive to higher-order QCD radiation without explicitely measuring a third jet Dijet production in lowest order pqcd: three-jet production in lowest order pqcd: p T balance φ,2 = π hard third jet: (k large) φ,2 << π soft third jet: (divergence in LO pqcd) (k 0) φ,2 π DIS 2004 Pavel Demine DAPNIA-SPP 8
Dijet φ decorrelation: measurement (/σ) dσ/d 3 2 - p T jet > 80 GeV ( 25) 30 < p T jet < 80 GeV ( 25) 0 < p T jet < 30 GeV ( 5) 75 < p T jet < 0 GeV ( ) R cone p = 0.7 - s =.96 TeV, L = 50 pb > 40 GeV, y jet < 0.5 dijet azimuthal angle defined as φ = φ jet φ jet2 the φ distribution measured only for φ > π/2 to avoid φ 2 R cone (overlapping jets) change in the shape of the φ distribution proportion of back-to-back jets is higher for higher p T.6.8 2 2.2 2.4 2.6 2.8 3 DIS 2004 Pavel Demine DAPNIA-SPP 9
Dijet φ decorrelation: compare to LO (/σ) dσ/d (/σ) dσ/d - NLOJET++ NLO LO CTEQ6.M jet,2 75 GeV < p < 0 GeV T jet.6.8 2 2.2 2.4 2.6 2.8 3 - NLOJET++ NLO LO CTEQ6.M jet,2 0 GeV < p < 30 GeV T jet.6.8 2 2.2 2.4 2.6 2.8 3 (/σ) dσ/d (/σ) dσ/d - NLOJET++ NLO LO CTEQ6.M jet,2 30 GeV < p < 80 GeV T jet.6.8 2 2.2 2.4 2.6 2.8 3 - NLOJET++ NLO LO CTEQ6.M jet,2 p > 80 GeV T jet.6.8 2 2.2 2.4 2.6 2.8 3 poor overall description by LO divergence towards φ π where third jet is soft: p T 0 phase space limitation: φ > 2π/3 very good description by NLO exceptions: extreme φ regions large φ: third and fourth jets are soft small φ: no hard phase space restriction, but still limited phase space for four-jet configurations DIS 2004 Pavel Demine DAPNIA-SPP
Dijet φ decorrelation: HERWIG, PYTHIA HERWIG & PYTHIA: (/σ) dσ/d - PYTHIA6.28 - default PYTHIA6.28 - Tune A HERWIG6.500 - default CTEQ5L (/σ) dσ/d - PYTHIA6.28 - default PYTHIA6.28 - Tune A HERWIG6.500 - default CTEQ5L different implementations of parton shower default versions: (/σ) dσ/d jet,2 75 GeV < p < 0 GeV T jet.6.8 2 2.2 2.4 2.6 2.8 3 - PYTHIA6.28 - default PYTHIA6.28 - Tune A HERWIG6.500 - default CTEQ5L jet,2 30 GeV < p < 80 GeV T jet.6.8 2 2.2 2.4 2.6 2.8 3 (/σ) dσ/d jet,2 0 GeV < p < 30 GeV T jet.6.8 2 2.2 2.4 2.6 2.8 3 - PYTHIA6.28 - default PYTHIA6.28 - Tune A HERWIG6.500 - default CTEQ5L jet,2 p > 80 GeV T jet.6.8 2 2.2 2.4 2.6 2.8 3 not good but reasonable description also for extreme φ: φ π/2 and φ π PYTHIA tune A : tuned to CDF data on the underlying event significant improvement over default still not as good as NLO in intermediate φ region DIS 2004 Pavel Demine DAPNIA-SPP
Summary Dijet mass measurement: dijet mass distribution was measured for cone R = 0.7 jets in the central region of the calorimeter ( y jet < 0.5) using 43 pb of the NLO pqcd calculations agree with the measurements within uncertainties systematic uncertainty is dominated by jet energy scale uncertainty Dijet azimuthal angle decorrelation: φ distribution was measured for cone R = 0.7 jets in the central region of the calorimeter ( y jet < 0.5) for four bins of leading jet p T HERWIG & PYTHIA, MC generators that model parton shower evolution describe the data better than LO pqcd and can be tuned to improve the agreement NLO pqcd calculations give good description of the data but fails in the largest/smallest φ regions DIS 2004 Pavel Demine DAPNIA-SPP 2