Final results on tt H H Akihiro

Final results on tt. H H + Akihiro, ITO, Reisaburo TANAKA Okayama University Shoji ASAI, Junichi TANAKA ICEPP, University of Tokyo Junichi KANZAKI KEK Many thanks to Lidija ZIVKOVIC (Weizmann) 2003/9/22 ATLAS Higgs Working Group Meeting, September 22, 2003 Final Results on tt. H (H->tau+tau)

Contents 1. 2. 3. 4. 2003/9/22 tt. H (H + -) analysis tag algorithm ATLFTAU Summary Final Results on tt. H (H->tau+tau) 2

1. tt. H (H + -) analysis l Physics Motivation: direct determination of top-Yukawa coupling in the mass range MH=[100, 130]Ge. V/c 2. Complementary to H bb. l Signal l Small cross section ( Br 40 fb @MH=120 Ge. V/c 2) l tt. H Wb. Wb … 9 different modes, followings are analyzed. l l b+jjb +l +h channel … same charge di-lepton l l Br=14% Israel l l b+jjb + h +h channel … single lepton Br=13% 5 modes l l b+jjb +l +l channel … tri-lepton Br= 4% Japan Br= 2% 2 modes l l b+l +h channel … tri-lepton l l b+h +h channel … di-lepton Br= 2% l Backgrounds l HUGE tt ! (tt. H) Br(H ) / (tt) 10 -4, tt. W l tt. Z, tt. WW, tt. H(H WW) … irreducible backgrounds. Key issue: -tag, fake-jet rejection 2003/9/22 Final Results on tt. H (H->tau+tau) 3

top-Yukawa coupling in tt. H(H bb, + -, W+W-) Athens Physics Workshop (R. T. ) L = 30 fb-1 gtt. H/gtt. H Preliminary H bb poor S/N H W+W 20 30% H + (Weizmann) MH [Ge. V/c 2] 2003/9/22 Final Results on tt. H (H->tau+tau) 4

2 issues ü tag algorithm ATLFTAU v. s. Weizmann - ATLFTAU: parametrized 50% fixed efficiency - Weizmann: inspired by TDR algorithm. ü Is tt. H(H-> ) l b+jjb + h +h possible? 2003/9/22 Final Results on tt. H (H->tau+tau) 5

2. tag algorithm l Issue: tag algorithm optimization for tt. H ATLFTAU … 50% fixed efficiency for low Pt taus is not good. Leads to unreasonable low rejection factor. l Emulated Weizmann tau identification method l Rj-track < 0. 07 (distance between jet and nearest track) l Ntrack = 1 or 3 (# of tracks with PT>5 Ge. V/c in DR = 0. 07) l PT(track) > 5 Ge. V/c (PT of track nearest to jet) l I < 0. 03 I = { ET(0. 4) – ET(0. 2) } / ET(0. 2) l PT( -jet) > 20 Ge. V/c l Ncell < 6 (# of cells in jet, COMMON/ATLFEVENT/ K(nj, 3) ) 2003/9/22 Observed good agreements on these variables. Though differences in PT(track) and I distributions. Final Results on tt. H (H->tau+tau) 6

E. Gross, June 25, 2003@Higgs WG tag efficiency and jet rejection m. H = 120 Ge. V 50% algorithm real 2003/9/22 fake Final Results on tt. H (H->tau+tau) p. T- -jet 7

Weizmann cut on l b + jjb + h l High Level Trigger l Lepton PT > 20 Ge. V/c, |η| < 2. 5, Nlepton ≥ 1 l Jet PTj 1 > 35 Ge. V/c, PTj 2 > 25 Ge. V/c |η| < 4. 5, Njet ≥ 2 l b-jet ( efficiency = 70%) PT > 20 Ge. V/c, |η| < 2. 5, Nb-jet ≥ 2 l -jet PT > 20 Ge. V/c, |η| < 2. 0, N -jet ≥ 2 2003/9/22 l Mass windows 60 < Mqq < 90 Ge. V 150 < Mbjj < 195 Ge. V 50 < M -jet, -jet < 120 Ge. V l Missing ET > 25 Ge. V Φ( , missing ) < 1 opposite charge of -jet l B-tag efficiency 70% Final Results on tt. H (H->tau+tau) 8

Labeled -jet Japan ATLFAST, tt. H MH=120 Ge. V, Low lumi. Other jet Rj-track< 0. 07 Ntrack= 1 or 3 (Ge. V) Israel 2003/9/22 PT> 5 Ge. V/c Final Results on tt. H (H->tau+tau) 9

ATLFAST, tt. H MH=120 Ge. V, Low lumi. Japan I< 0. 03 Cone R<0. 3 PT> 20 Ge. V/c N 5 (Ge. V) Very powerful cut ! Israel 2003/9/22 Final Results on tt. H (H->tau+tau) 10

Results with Weizmann tau ID Japan Israel efficiency 44% rejection 80 84 Efficiency 38% 33% rejection 72 100 tt. H(H ) tt Ncell Remaining difference YPAR(93) - fraction of tau-jet in a cone Rtau-jet Japan: default 0. 9 Israel : 0. 0 (optimistic? ) 2003/9/22 Ncell tt. H l b+jjb + h +h MH=120 Ge. V, Low lumi. 100 fb-1 Japan Israel tt. H 2. 9 4. 7 tt 8. 9 5. 9 Final Results on tt. H (H->tau+tau) 11

Think about ET cuts in cells ? ETcell > 0. 5 Ge. V (noise & pile-up) Signal -jet signal N 4 ET>0. 5 Ge. V tt (Ge. V) tt. H tt efficiency 44 44% 38 38% rejection 80 87 72 78 2003/9/22 tt. H l b+jjb + h +h MH=120 Ge. V, Low lumi. 100 fb-1 Japan tt. H 3. 0 tt 4. 4 Final Results on tt. H (H->tau+tau) 12

Without ETcell cut With ETcell cut Number of events for 100 fb-1 (percentage wrt previous cut) Japan tt. H tt precuts 38. 8 1555 precuts 40. 3 1256 Trigger 37. 5 (97) 1405 (90) Trigger 39. 2 (97) 1131 (90) Q 1*Q 2 < 0 28. 3 (75) 658 (47) Q 1*Q 2 < 0 30. 4 (78) 516 (46) F 18. 6 (66) 399 (61) F 20. 3 (67) 248 (48) MW wind. 11. 2 (60) 155 (39) MW wind. 12. 2 (60) 86. 1 (35) Mtop wind 8. 8 (79) 77. 2 (50) Mtop wind 10. 1 (83) 39. 0 (45) |h. W| 7. 8 (89) 56. 8 (74) |h. W| 8. 8 (87) 29. 2 (75) Missing ET 6. 0 (77) 39. 0 (69) Missing ET 6. 6 (75) 16. 0 (55) Ptlep 5. 5 (92) 34. 6 (89) Ptlep 5. 9 (89) 15. 1 (94) PTj 1 5. 1 (93) 24. 0 (69) PTj 1 5. 3 (90) 8. 9 (59) PTj 1 3. 5 (69) 10. 7 (44) PTj 1 3. 4 (64) 4. 4 (50) M -jet, -jet wind 2. 9 (80) 8. 9 (83) M -jet, -jet wind 3. 0 (88) 4. 4 (100) 2003/9/22 Final Results on tt. H (H->tau+tau) 13

3. ATLFTAU Atlftau Results Japan tt. H 2. 7 2. 5 tt • • Israel 126. 7 22. 1 6 times! Reasons Fake -jet charge Japan Method 2003/9/22 Israel Random number [0, 1] < 0. 5 Q=-1 > 0. 5 Q=+1 Sum up charged track in R = 0. 07 from Jet Final Results on tt. H (H->tau+tau) 14

l Generator Japan Israel Generator PYTHIA 6. 2 PDF CTEQ 5 L tt event number 500 k 300 k 492 pb 490 pb cross section l ATLFAST parameter Japan Israel YPAR(90) 10. 000 YPAR(91) 2. 500 YPAR(92) 0. 300 YPAR(93) 0. 900 0. 000 YPAR(93) : fraction in a cone R_tauj for tau-jet 2003/9/22 Final Results on tt. H (H->tau+tau) 15

Q 1*Q 2 tt. H tt Israel Drops the tt. H signal. But also artificial cut on tt. Unphysical effect. Explains factor 3 difference Japan 2003/9/22 Final Results on tt. H (H->tau+tau) 16

ATLFAST Number of events for 100 fb-1 (percentage wrt previous cut) Now identical condition YPAR(93)=0. 0, Φcut good agreement. Japan Israel tt. H tt precuts 61. 6 19532 75 20330 Trigger 59. 0 (96) 18115 (93) 68. 5 (91) 17671 (87) Q 1*Q 2 < 0 20. 1 (34) 1427 (7. 9) 24 (35) 1426 (8. 1) F 12. 5 (62) 685 (48) 15 (63) 704 (49) MW wind. 8. 1 (65) 308 (45) 8. 4 (56) 301 (43) Mtop wind 6. 7 (83) 190 (62) 6. 16 (73) 191. 5 (64) |h. W| 6. 0 (90) 167 (88) 5. 35 (87) 150 (78) Missing ET 4. 8 (80) 133 (80) 4. 5 (84) 117 (78) Ptlep 4. 5 (94) 108 (81) 4. 3 (96) 96. 5 (82) PTj 1 3. 9 (87) 85. 6 (79) 4. 08 (95) 84. 6 (88) PTj 1 2. 8 (72) 48. 7 (57) 3. 08 (75) 44. 5 (53) M -jet, -jet wind 2003/9/22 tt 2. 1 (75) 26. 6 (55) 2. 5 (81) 22. 3 (50) Final Results on tt. H (H->tau+tau) 17

4. Summary l Compared two analyses on tt. H(H ) l Tau hadronic mode: l b + jjb + h l -tag algorithm l ATLFTAU (50% eff. Fixed) - Unreasonable at low Pt. Problematic for tt. H. l Weizmann algorithm - Ncell is a powerful parameter for tt rejection but pile-up and noise should be taken into account. ü Full simulation with noise&pile-up under study. l ATLFAST (parametrized ATLFTAU) l Good agreement between Japan and Israel results now. l Discrepancies due to -jet charge assignment , YPAR(93), Φ. l These are not critical for our final results … 2003/9/22 Final Results on tt. H (H->tau+tau) 18

Our Results for Athens L=300 fb-1 MH=115 Ge. V/c 2 tt. H (H + -) t( l b ) + ( h ) + ( l ) t(l b ) + t(jjb) + ( h ) + ( l ) For 2 300 fb-1 S/B = 0. 8 S/ B = 4. 5 (Poisson) Sig. Bkg. tt tt. Z tt. W tt. H 2 -lepton 11. 5 16. 7 4. 1 4. 5 1. 4 3 lepton 3. 7 2. 4 0. 3 1. 1 0. 3 0. 7 t(l b) + t(jjb) + (h ) Similar sensitivity as di-lepton channel, but suffers from irreducible tt. Z, tt. WW …background. 2003/9/22 Final Results on tt. H (H->tau+tau) 19