1. Charm Mixing and D Dalitz analysis at BESIII SUN Shengsen Institute of High Energy Physics, Beijing (for BESIII Collaboration) 37 th International Conference on High Energy Physics July 4 th, 2014 Valencia

2. BESIII Detector and Data NIM A614, 345(2010) Total about 2.9/fb  (3770) data are taken at BESIII, in 2010 and 2011 S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 2 See shuang-shi’s talk

3. S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 3 PRL 110, 101802 (2013)

4. 4

5. Single Tags of CP modes S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 5 unique variable in threshold-production

6. S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 6

7. 7 With external inputs of the parameters in HFAG2013 and PDG, we obtain CLEO measurements with 0.8 fb −1 at ψ(3770 ). [CLEO, PRD 86 (2012) 112001] without external inputs: with external inputs: Phys. Lett. B 734(2014) 227

8. S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 8 One D decays into CP eigenstates and the other D decays semileptonically, the decay rate is: Neglecting terms to order y 2 or higher, we can derive In the limit of no CPV,

9. S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 9 CP- tag flavor tag y CP is measured using CP-tagged semi-leptonic D decays based on 2.9 fb -1 ψ(3770) data Semi-leptonic signal peaks at zero! The observable can be :

10. Single Tags of CP modes S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 10 preliminary preliminary preliminary preliminarypreliminary preliminary

11. S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 11 preliminarypreliminary preliminary preliminary preliminary preliminary signal: MC shape convoluted with an asymmetric Gaussian background: a 1 st -order polynomial function

12. S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 12 preliminary preliminary preliminary preliminary preliminary preliminary

13. S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 13 Signal yields of the full data set preliminary result: result is statistically limited systematic uncertainty is relatively small The average branching ratio over different CP modes by minimizing:

14. Comparison S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 14 BESIII (pre.) -1.6 ±1.3 ±0.6 % World average directly from HFAG2013 (BESIII (pre.) not included) compatible with world average results

15.  In D meson decay, there are many three bodies final states with large branching fraction and including K  and  two body resonances.  K  is a special and interesting system  K  S wave  numerous K excited states: K*(892), K 0 *(1430), K*(1680), etc.  K  S wave and low-mass K  scalar resonance  (800) have been observed significantly in earlier experiments (E791, CLEO) through Dalitz plot analysis.  The D +  K S  +  0 decay as one of gold channels, is needed to obtain relative branching fraction more precisely and to find more intermediate resonances. S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 15

16. Signal and Sideband  ~167k events are selected in signal region.  Shape of combinatorial background on Dalitz plot is estimated by combination of two sidebands (left & right).  A peaking background is very small (~0.6% of signal) is estimated using MC:   + (Ks)  + (D) Mass Recoiling against K S    0 candidates mis-reconstructed signal S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 16 Signal region: 84.9% SIG+15.1% BG Sideband

17. Isobar Model Result  Model without  non-resonant component have been tested. And the modes with small fractions are dropped.  Float parameters of K*0bar(1430) and  (800) S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 17 Modeling errors: shape: angle distribution, form factor and resonance shape; add: additional resonances.

18. Final Results S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 18 PDG 2014: 0.9  0.7 4.8  1.0 1.3  0.6 Phys. Rev. D 89, 052001 (2014)

19. Resonance Parameter  The mass and width of K*0(1430) are consistent with E791 and CLEO-c from D +  K -  +  +.  Another fit to model without  (800) gives m(K*0(1430))=1444  4 MeV,  (K*0(1430))=283  11 MeV, consistent with the value of PDG2012.  The pole of  (800) is consistent with the model C of CLEO-c. PRL 89, 121801(2002) PRD 78, 052001(2008) S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 19

20. Cross-check with Model-Independent PWA  For some interested resonances, a binned amplitude is used. Other resonances are kept same as isobar model.  First by E791  The K*0bar(1430) is destructive interfered with  (800) and non-resonant, which can explain the fraction of K  S wave smaller than the combine of  (800) and non- resonant.  The phase shift can be described by NR+  (800) well. PRD 73,032004(2006) S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 20

21. Summary S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 21

22. Thank you for your attention! S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 22

23. 23  Pure DD final state, no additional particles (Ed=Ebeam)  Double Tag (DT): tag one D in a selected tag mode, study the other D as signal mode.  D->CP ST Yield:  D->CP and D->kpi DT Yield:  Branching fractions:  BESIII: world’s largest sample of  (3770).

24. Momentum-dependent Correction for Efficiency  The differences of efficiency between MC and data are momentum dependent, for Ks/  0 reconstruction and  tracking/PID.  At different position on Dalitz plot, the distributions of momentum are different.  These two cause that efficiency correcting factor should be different at different position (x,y). Therefore, a momentum- dependent correction is perform. A MC study for efficiency correction: black for real, red for uncorrected, blue (matched with black) for corrected. S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 24

25. Shape Approximation for Argus BG  In order to approximate background shape, two sidebands are used to parameterize background  In the right sideband, there are obvious signal components, because of ISR  Parameterized by background + signal  Signal is initialized by left sideband  Iterate to approach the real amplitude of signal more and more Left: (a)(b)(c); right: (d)(e)(f); combined: (g)(h)(i) BESIII Preliminary Signal component S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 25

26. Maximum Likelihood Fit  The log-likelihood function is defined as  p.d.f. is  For efficiency: 3 rd polynomial function  threshold factor or  For Argus BG: resonances  +, K *0, K *+  For signal with background, the efficiency and the backgrounds are fixed as parameterized shapes. PHSP generator DALITZ generator No obvious difference is found. for efficiency by PHSP for efficiency by DALITZ for Argus BG for signal with BG Histogram p.d.f. from MC S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 26

27. Isobar Model and Fit Fraction  Decay matrix element  c R is complex parameter to fit  W R is dynamical function, generally, a Breit-Wigner function.  For special resonance, such as  (800)  For any intermediate resonance, its fraction is calculated by  For combined fraction, Angle distribution Form factor K  S wave is a sum of  (800), K*0bar(1430) and non-resonant. S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 27

28. Fit to Data using Isobar Model  Model with K*bar and  cannot describe our data well, more intermediate resonances are considered.  Float parameters of K*0bar(1430) and  (800) No evidences for DCS channels S.S. Sun, Charm Mixing and D Dalitz Analysis at BESIII@Charm2014 28