Test of color-reconnection models in Z 3 jets G.Rudolph - PowerPoint PPT Presentation

Test of color-reconnection models in Z → 3 jets G.Rudolph Inst.f.Experimentalphysik, Uni Innsbruck Fachtagung Kern- u. Teilchenphysik, Weyer, 26 Sept 2004

Motivation: precise measurement of M W at LEP-2 (to better than 5 · 10 − 4 ) Color reconnection (CR) can modify the hadronic final state in the process e + e − → W + W − → ( q ¯ q )( q ¯ q ) and thus the reconstructed W mass. CR contributes the largest systematic uncertainty (90 MeV at present). Direct information from - particle momentum spectra - interjet particle flow (LEP comb.) - M W ( q ¯ qq ¯ q ) − M W ( lνq ¯ q ) (LEP comb.) - differential behaviour of M W with Pcut and Cone (new development) shows no convincing evidence for CR. Data consistent with no CR. Data not sensitive enough to test all models. Only extreme models (SK1 100%) can be excluded.

How to calculate CR effects ? Perturbative QCD prediction for WW is of order (( α s /N C ) 2 · Γ W /M W ) , from which δM W = few MeV (negligible). Possible non-pert. CR ⇒ phenom. models General case : multiple gluon emission Leading Log parton shower determines a color sequence (in the large N C limit) for drawing the string : R − ¯ RG − ¯ GB − ¯ BG − ¯ GR − ¯ R.... Some models assume that non-perturbative rearrangement is possible among identical colors, with probability 1 /N 2 C , and takes place if the total string length (e.g. λ ) decreases. Leads to effects also within color singlet systems.

CR Monte Carlo models : All but the last run with Pythia model criterion of free effect in reconnection parameter value Z → q ¯ q SK1 space-time overlap 0.6 not k I of flux tubes implemented SK2 crossing of - - NO vortex lines ARIADNE reduce total P reco 1/9 Y AR1, AR2 string length λ GAL reduce area with prob. R 0 0.1 Y (Rathsman) P = R 0 (1 − exp( − b ∆ A )) HERWIG reduce cluster size P reco 1/9 Y in space-time Predicted M W bias < 100 MeV for reasonable parameter values

Test of CR in Z → hadrons : String drawing : possible ? normal CR effect localised in gluon jet because it often fragments in isolation = ⇒ select 3 jet events Sensitive variables : • Interjet particles (L3, PLB 581 (2004) 19) reconnected problem: no-CR Jetset and Ariadne predictions differ • Particle momentum spectra in jets : not clear what is tested. • Rapidity gaps, Jet charge ⇒ clear signal (OPAL, Eur.Phys.J.C35 (2004) 293, and Eur.Phys.J.C11 (1999) 217)

ALEPH data analysis : From 3.4 million multihadronic Z events (LEP-1, 1992-1995) select 3-jet events using Durham cluster algorithm, with resolution y cut = 0 . 02 , ⇒ R 3 = 0 . 228 Energy-ordered analysis : Order the jets as x 1 > x 2 > x 3 , with x j = 2 E j /E cm More cuts : require Φ jk > 40 ◦ , x 3 > 0 . 1 , | cos Θ j | < 0 . 9 → 539000 events gluon purity < E jet > GeV P g (from MC) jet 3 17.7 0.69 gluon enriched jet 1 40.8 0.06 quark enriched Rapidity : refers to respective jet axis Charged particles: pion mass assumed; Neutral particles: pseudorapidity used.

Multiplicity distributions (data, Jetset, GAL) (data, AR0, AR1) in fixed rapidity interval (0 - 1.5) : 80000 80000 ID 3022331 ID 3022331 Entries 538775 Entries 538775 70000 70000 60000 60000 Rate of jet 3 with a gap 50000 50000 40000 40000 (7% in data) 30000 30000 is sensitive to CR, 20000 20000 10000 10000 but not used as observable. 0 0 0 2 4 6 0 2 4 6 N(c+n), y in 0 - 1.5, jet 3 N(c+n), y in 0 - 1.5, jet 3 ID 3022311 ID 3022311 90000 90000 Entries 538775 Entries 538775 80000 80000 70000 70000 60000 60000 50000 50000 40000 40000 30000 30000 20000 20000 10000 10000 0 0 0 2 4 6 0 2 4 6 N(c+n), y in 0 - 1.5, jet 1 N(c+n), y in 0 - 1.5, jet 1

Jet charge Q j = � i q i x 10 2 data, Jetset, GAL distributions, 1800 20000 ID 3020032 ID 3122332 Entries 538775 Entries 39037 18000 normalized to same area, 1600 Mean 0.4268E-01 Mean 0.1885E-01 RMS 1.595 RMS 1.055 16000 1400 for all jets 14000 1200 12000 1000 and for those with a rapidity gap 10000 800 8000 (i.e. no particles in 0 ≤ y ≤ y u , 600 6000 400 4000 with y u = 1 . 5 ) 200 2000 0 0 -5 -2.5 0 2.5 5 -5 -2.5 0 2.5 5 Several factors determine Q j : Q(jet 3), all Q(jet 3), y gap x 10 2 1800 20000 - charge compensation in fragment. ID 3020012 ID 3122312 Entries 538775 Entries 42307 18000 1600 Mean 0.5470E-01 Mean 0.1066E-01 RMS 1.463 RMS 0.9716 - cluster algorithm 16000 1400 14000 1200 - jet environment 12000 1000 10000 - detection effects 800 8000 600 6000 400 4000 Fraction f ( Q j = 0) of neutral 200 2000 0 0 gluon jets -5 -2.5 0 2.5 5 -5 -2.5 0 2.5 5 Q(jet 1), all Q(jet 1), y gap is sensitive to CR !

Similar for data vs. Ariadne x 10 2 data, AR0, AR1 1800 20000 ID 3020032 ID 3122332 Entries 538775 Entries 39037 18000 1600 16000 1400 14000 1200 12000 1000 10000 800 8000 600 6000 400 4000 200 2000 0 0 -5 -2.5 0 2.5 5 -5 -2.5 0 2.5 5 Q(jet 3), all Q(jet 3), y gap x 10 2 1800 20000 ID 3020012 ID 3122312 Entries 538775 Entries 42307 18000 1600 16000 1400 14000 1200 12000 1000 10000 800 8000 600 6000 400 4000 200 2000 0 0 -5 -2.5 0 2.5 5 -5 -2.5 0 2.5 5 Q(jet 1), all Q(jet 1), y gap

Result for relative model-data difference model-data difference δ = f ( Q j = 0) MC − f ( Q j = 0) data 0.2 0.2 δ δ JETSET+GAL HERWIG CR ARIADNE AR1 f ( Q j = 0) data 0.15 0.15 0.1 0.1 as a function of y u : 0.05 0.05 Data vs Jetset and Ariadne : 0 0 -0.05 -0.05 • jet 1: well described JETSET HERWIG ARIADNE -0.1 -0.1 0 1 2 0 1 2 • jet 3: CR models clearly disfavoured. y u of gap (c+n), jet 3 y u of gap (c+n), jet 3 0.2 0.2 δ δ • jet 3: interesting effect (??) : JETSET+GAL HERWIG CR ARIADNE AR1 0.15 0.15 More neutral jets with a y-gap seen 0.1 0.1 than expected without CR. 0.05 0.05 Also reported in Delphi 2002-053. 0 0 Data vs Herwig : -0.05 -0.05 JETSET HERWIG ARIADNE • cannot describe jet 1 ⇒ doubtful MC -0.1 -0.1 0 1 2 0 1 2 y u of gap (c+n), jet 1 y u of gap (c+n), jet 1 • no CR sensitivity in jet 3

Systematic checks: Variation of the jet definition : • jet resolution parameter y cut varied from 0.005 to 0.05 • a different jet finder : Durham = ⇒ Jade • re-assigning particles to jets on basis of smallest angle • define y-gap with charged only ⇒ result does not change qualitatively Systematics from detector simulation : • varied track selection cuts and | cos Θ j | cut ⇒ δ remains within 1 σ • event charge Q ev = � i q i (= 0 ideally) used as a test quantity

B-tag analysis : 0.4 δ = (f(Q=0) MC - f(Q=0) data ) / f(Q=0) data select Z → b ¯ bg events by JETSET+GAL requiring lifetime signals in 2 jets ARIADNE AR1 0.3 and no signal in 1 jet ⇒ the gluon jet 0.2 High purity : P g = 97 % < E jet > = 19 . 8 GeV 0.1 Less statistics : 24600 events 0 Effects numerically larger. Same result, JETSET -0.1 but statistically inferior to the energy- ARIADNE ordered analysis -0.2 -0.3 0 0.5 1 1.5 2 2.5 y u of gap (c+n), gluon jet

Discussion The CR models AR1 and GAL fail to describe gluon jet data at the Z (confirming L3 and OPAL results) Does this have consequences for WW ? • Yes, according to the above models, the physics is the same • Not necessarily, according to Sj¨ ostrand Note also: ALEPH does not see Bose-Einstein correlations between pions from different W’s. My conclusion: large CR effects are unlikely