DHEP Annual Meeting Search for CP violation in the radiative mode

Transcription

1 DHEP Annual Meeting Search for CP violation in the radiative mode D 0 φγ Varghese Babu, G. B. Mohanty, T. Aziz. Tata Institute of Fundamental Research April 7, / 18

2 Motivation CP violation in the charm sector had been investigated for a number of modes in the Belle Experiment. A few recent results are A CP (D 0 π 0 π 0 ) = ( 0.03 ± 0.64 ± 0.10)% A CP (D 0 K 0 s π 0 ) = ( 0.21 ± 0.16 ± 0.07)% (N.K.Nisar et al. arxiv: [hep-ex] 2014) A CP (D + K 0 s K + ) = ( 0.25 ± 0.28 ± 0.14)% (B.R.Ko et al. arxiv: [hep-ex] 2013) A CP (D + K 0 s π + ) = ( ± ± 0.067)% (B.R.Ko et al. arxiv: [hep-ex] 2012) A CP (D + φπ + ) = (+0.51 ± 0.28 ± 0.05)% (M.Staric et al. arxiv: [hep-ex] 2011) The paper by Gino Isidori and Jernej F. Kamenik (2012)(PhysRevLett ) using models beyond SM, predicts sizeable CP Asymmetry in modes of the kind D 0 V γ where V is a vector particle, upto several percent. Thus this can be an important search for physics beyond the standard model. We intend to search for CP violation in the mode D 0 φγ 2 / 18

3 Why study D 0 φπ 0? The modes D 0 φπ 0 and D 0 φη are important backgrounds for our signal mode D 0 φγ as both the π 0 and η decay into two photons and if one of the photons is not reconstructed, it would resemble our signal. The branching fractions of the modes are (PDG, PR D86, (2012)) B(D 0 φπ 0, φ K + K ) = (6.4± 0.4) x 10 4 B(D 0 φη) = (1.4± 0.5) x 10 4 B(D 0 φγ) = (2.7± 0.35) x 10 5 We study the mode D 0 φπ 0 because of its relatively higher branching fraction (more than 20 times the signal D 0 φγ ) and because it will familiarize us with the fitting and extraction procedure for A CP 3 / 18

4 D* + D* - D 0 π + D 0 π - π 0 φ π 0 γ γ K + K - γ γ K + K - Figure : Signal Process Diagrams for D 0 φπ 0 and D 0 φπ 0 φ Cuts Applied (D φπ 0 ) dr K ±,π s < 1.0 cm dz K ±,π s < 3.0 cm R K ± (Kaon Likelyhood) > 0.1 R πs (Kaon Likelyhood) < GeV < M φ < 1.03 GeV E γ > 50MeV (Barrel) E γ > 100MeV (Endcap) 119 MeV < M π 0 < 151 MeV P D > 2.5GeV (optimized) P π 0 > 380MeV (optimized) 1.83 GeV < M D 0 < 1.89 GeV (optimized) 140 MeV < M < 160 MeV 4 / 18

5 Figure : 2-D cut optimization Since the PD P π 0 variables are correlated, a 2-D optimization, varying each cut independently, is performed. The optimizations are performed by minimizing the figure of merit, ɛ = Nsig +N bkg N sig Figure : M distribution. Figure : cos θ hel distribution. 5 / 18

6 How we measure CP asymmetry, A CP A rec is defined in terms of signal yield (N rec ) + D 0 π + A rec = ND s rec N D D 0 πs rec N D + D 0 π + s rec +N D D 0 π s rec A rec is the sum of three terms. A rec = A cp + A FB + A πs ɛ It has been demonstrated from previous CP asymmetry studies that the pion detection efficiency asymmetry A πs ɛ = +0.11% Since A FB is an odd function of cos(θd ),we divide the data in bins of cos(θ D ) and extract A cp and A FB from it by adding and subtracting bins at ±cos(θ D ) A CP = Acor A FB = Acor rec (cos(θd )+Acor rec ( cos(θd ) 2 and rec (cos(θ D ) Acor rec ( cos(θ D ) 2 6 / 18

7 Full Υ(4S) MC M Fit: Shift in cos θ hel distribution Figure : Simultaneous fit of the Full Υ(4S) generic MC (711 fb 1 ) For the Helicity component, the signal shape is modeled by an (x a) 2 function. The fitted value obtained for the helicity shift parameter is a shift = ± / 18

8 Full Υ(4S) Data 2D Fit: Shift in cos θ hel distribution Figure : Simultaneous 2-D fit of the Full Υ(4S) Data (711 fb 1 ) The fitted value obtained for the helicity shift parameter is a shift = ± ( 20σeffect) 8 / 18

9 Results : A uncorrected CP at the Υ(4S) resonance (711 fb 1 ) Sample A uncorr CP (1-D Fit) % A uncorr CP (2-D Fit) % MC ± ± 0.83 MC ± ± 0.84 MC ± ± 0.83 MC ± ± 0.83 MC ± ± 0.82 MC ± ± 0.83 Data 0.047± ± 0.81 Since we do not correct for the slow Pion detection efficiency asymmetry, we measure A uncorrected CP = A CP + A πs ɛ 9 / 18

10 D 0 φγ Study : Efficiency of the Extended π 0 Veto D* + D* - D 0 π + D 0 π - γ φ γ K + K - K + K - φ The candidate photon in in our main channel D φγ is clubbed up with all other photons in the event and a probability is assigned for each combination on whether they come from a decayed π 0. We have tried to improve the Vetoing efficiency by including in our list of photons in the given event, photons that have converted into an e + e pair. 10 / 18

11 Summary of cuts Variable Cut-Υ(4S) Eff. (%) Cut-Υ(5S) Eff.(%) Loss Loss dr K +,K,π s < 1.0 cm < 1.0 cm dz K +,K,π s < 3.0 cm < 3.0 cm R K +,K > > R πs < < M φ (GeV) (GeV) ( ±2.9σ) ( ±2.9σ) M D (GeV) (GeV) 0.53 M (MeV) (MeV) 0.19 π 0 probability < < η probability < < E9/E25 > > E γ > 580 MeV > 610 MeV (2-D) (2-D) P D > 2.55 GeV > 3.10 GeV 11 / 18

12 D 0 φγ: Backgrounds, 2-D optimization 12 / 18

13 Comparison of different fitting schemes Table : Comparison of fitting results Fit type A raw measured % M-cos θ hel 2-D Fit -4.5 ± 8.5 M D -cos θ hel 2-D Fit -2.7 ± D Fit 0.2 ± 6.9 For the M-cos θ hel fit, we use an optimized M D signal window ( GeV), and for the M D -cos θ hel fit, we use an optimized M signal window ( MeV) The 3-D fit has the smallest uncertainty in A CP 13 / 18

14 Projections of the 3-D fit to the variables M, M D and cos θ hel 14 / 18

15 Results : A uncorr CP for Υ(4s) + Υ(5s) combined sample MC Stream A uncorr CP (3-D Fit)(%) ± ± ± ± ± ± / 18

16 The method of reweighting: A Raw for D K S π 0 Since we do not expect large statistics for the main mode D φγ, the method of binning in the D production direction to correct for A FB is not viable. Instead, we use a high statistics mode like D K S π 0 to estimate A FB bin by bin, and we use these values to appropriately assign weights to our signal mode, during the simultaneous fitting, in order to correct for A FB. 16 / 18

17 The method of reweighting Table : Comparison of total A CP measurements for the mode D K S π 0 Number of bins Total A CP (%) by the Total A CP (%) by the binning method reweighting method ± ± ± ± ± ± ± ± ± ± 0.15 Table : Comparison of total A CP measurements for the mode D φπ 0 Fit method Total A uncorr CP (%) by the Total A uncorr CP (%) by the binning method reweighting method 1-D Fit ± ± D Fit ± ± / 18

18 Current status Extensive Monte Carlo studies have been performed on the sources of backgrounds and the A CP estimation procedure. We have also estimated the sensitivity of our final A CP measurement using various streams of generic MC. The studies performed are under collaboration wide review, and once the scrutiny is completed, we expect to look at A CP in real data. THANK YOU! 18 / 18