1. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 1 Measurements of Hadronic, Semileptonic and Leptonic Decays of D Mesons at E cm =3.77 GeV in CLEO Steven Blusk Syracuse University Outline  Introduction  Hadronic Branching Fractions  Semileptonic Decays  D +   +   Conclusion CLEO-c

2. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 2 Inner Drift Chamber CLEO-c CESR-c/CLEO-c The CLEO program has migrated from running on the  resonances to the region around the   Charm factory to study D, D s mesons -- Broad program of charm physics (3 fb -1 goal)  Additional running at J/  search for exotics in radiative decay) in year 3 Accelerator changes - installation of SC wigglers to improve damping  higher L  (3770) ’’ E beam CLEO-c detector largely same as CLEO-III,  Silicon replaced with drift inner chamber  B field reduced from 1.5T  1.0T  Tracking (93% of 4   16 axial, 31 stereo layers   p /p ~ 0.6 %  CsI (93% of 4   6144 crystals (barrel only):   E /E ~ 5% at 100 MeV ~2.2% at 1 GeV  Particle ID  RICH (80% of 4  ) + dE/dx   K >90% for  fake  Log

3. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 3 Some highlights of the CLEO-c Charm Program  Precision measurements of D branching fractions  Precise measurements of f D and f Ds, the D decay constants.  When combined with LQCD will enable ~5% determinations of V td and V ts  Pave the road for a more accurate extraction of V ub  Measurements of D   l and D   l form factors will provide “tested” lattice QCD predictions on heavy-to-light FFs.  Extraction of |V cd |, |V cs |  Unitarity Triangle  Once V ub & V td are measured to O(5%)  Allows for a stringent test of CKM angles (ie., sin2  ) vs sides  Precision measurements of D branching fractions  Precise measurements of f D and f Ds, the D decay constants.  When combined with LQCD will enable ~5% determinations of V td and V ts  Pave the road for a more accurate extraction of V ub  Measurements of D   l and D   l form factors will provide “tested” lattice QCD predictions on heavy-to-light FFs.  Extraction of |V cd |, |V cs |  Unitarity Triangle  Once V ub & V td are measured to O(5%)  Allows for a stringent test of CKM angles (ie., sin2  ) vs sides

4. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 4 CLEO & D Tagging D tag D sig K   Tag one D meson in a selected tag mode.  Dictates whether final state is D + D - or D 0 D 0  Study decays of other D, (signal D) E D  E beam improves mass resolution by ~10X e+e+ e + e -   (3770)  DD e-e-  Pure DD final state, no additional particles (E D = E beam ).  Low particle multiplicity ~ 5-6 charged particles/event  Good coverage to reconstruct in semileptonic decays  Pure J PC = 1 - - initial state  Hadronic BF: Use double-tagged and single-tagged yields  Semileptonic decays: D tag + (D sig  Xe e ), reconstruct e using P miss  Leptonic Decays: D tag + (D sig    ) Analysis Preview Analyses shown today based on 57 pb -1

5. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 5 Absolute D Hadronic Branching Fractions D +  K -     D +  K -       D +  K s   D +  K s     D +  K s       D +  K - K    D +  K -     D +  K -       D +  K s   D +  K s     D +  K s       D +  K - K    D 0  K -   D 0  K -     D 0  K -       D 0  K -   D 0  K -     D 0  K -       D -  K +     D -  K +       D -  K s   D -  K s     D -  K s       D -  K - K    D -  K +     D -  K +       D -  K s   D -  K s     D -  K s       D -  K - K    D 0  K +   D 0  K +     D 0  K +       D 0  K +   D 0  K +     D 0  K +       e+e+ e-e- D D Single Tags: Reconstructed one D meson Double Tags: Reconstruct both D mesons Since  ij   i  j, correlated systematics cancel in N DD To first order, B i is independent of tag modes’ efficiencies, , L. N DD and B i ’s extracted using a  2 minimization technique  Validated on toy MC and 50X simulated data.

6. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 6 Fits to Data Mode N D (x10 3 )  D  (%) KK 5.11±0.075.15±0.0765.7±0.3 K   9.51±0.119.47±0.1133.2±0.1 K  7.44±0.097.43±0.0944.6±0.2 K  7.56±0.09 51.7±0.2 K   2.45±0.072.39±0.0727.2±0.2 KsKs 1.10±0.041.13±0.0445.6±0.4 K s   2.59±0.072.50±0.0723.4±0.2 K s  1.63±0.061.58±0.0631.4±0.2 KK  0.64±0.030.61±0.0342.6±0.5 D 0 Modes D + Modes KK K   K  K  K   KsKs K s   K s  KK  Signal shape includes:   (3770) line shape, ISR, beam energy spread & momentum resolution * Efficiency includes FSR losses

7. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 7 Systematic Uncertainties  For pion:  Look at mass recoiling against J/  in  ’  J/     events  Peak at M  2 for J/      Count the number of times the track is found versus not found. Tracking,  0 and K s all use similar “missing mass” technique. MC DATA  track found  track not found Uncertainty  0.7% / (  K  SourceValue (%) Tracking / K s /  0 0.7 / 3.0 /2.0 Particle ID  (0.3) / K (1.3)  E selection 1.5  (3770) 0.6 Final State Radiation0.5 ST / 1.0 DT Resonant Substructure0.4 – 1.5 Double DCSD Interf.0.8 Fit functions0.5 Data Processing0.3

8. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 8 Preliminary Results ParameterFitted Value (%) N(D + D - )(1.558±0.038±012)x10 5 B (D +  K -  +  + ) (9.52±0.25±0.27) % B (D +  K -  +  +  0 ) (6.04±0.18±0.22) % B (D +  K s  + ) (1.55±0.05±0.06) % B (D +  K s  +  0 ) (7.17±0.21±0.38) % B (D +  K s  +  +   ) (3.20±0.11±0.16) % B (D +  K + K -  + ) (0.97±0.04±0.04) % ParameterFitted Value (%) N(D 0 D 0 )(2.006±0.038±016)x10 5 B (D 0  K -  + ) (3.91±0.08±0.09) % B (D 0  K -  +  0 ) (14.94 ±0.30±0.47) % B (D 0  K -  +  +    ) (8.29±0.17±0.32) % D 0 Modes D + Modes As many of the systematics are evaluated using data, they will shrink as  L Normalized to PDG to be submitted to PRL

9. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 9 Semileptonic Decays c e+e+ e W+W+ |V cs |, |V cd |  Test LQCD on shape of f + (q 2 )  Use tested Lattice for norm.  From B (D  Xe ) extract |V cd |  D   FF related to B   FF by HQS  Precise D   FF’s can lead to reduced  theory in |V ub | at B factories  Similar for D  V l  except 3 FF’s enter  Can also form ratios, where theory should be more precise LQCD, PRL 94, 011601 (2005)

10. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 10 Basic Technique K-+K-+ K-+0K-+0 K -  +  0  0 K -  +  +  - KS+-KS+- K S  +  -  0 KS0KS0 -+0-+0 K-K+K-K+ e+e+ e-e- D tag  (3770) D 0 Tag Modes X = K, K*,  ( )  Require no extra tracks in event  Average X and X modes  Reconstruct from p miss Fit to U distribution Fit to M BC distributions Efficiency from MC M BC

11. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 11 Pseudoscalar Modes: D  Pe e U = E miss – |P miss | (GeV) Events / ( 10 MeV ) (~110 events) (~1400 events) Events / ( 10 MeV ) (~60 events) (~500 events) U = E miss – |P miss | (GeV) c  s Cabibbo Favoredc  d Cabibbo Suppressed

12. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 12 Vector Modes: D  Ve e U = E miss – |P miss | (GeV) (~30 events) (~400 events) U = E miss – |P miss | (GeV) c  s Cabibbo Favored c  d Cabibbo Suppressed (~90 events) (~30 events) First Observation (~8 events) First Observations (5  ) 57 pb -1 Data

13. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 13 Preliminary Results   V cd /V cd ~ 1.7% from D   e   V cs /V cs ~ 1.6% from D   e Assuming  FF ’s to ~3% from LQCD, 3 fb -1 RatioCLEO-cPDG - - - - CLEO-c goal (3 fb -1 ): to be submitted to PRL

14. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 14 Inclusive Semileptonic Electron Spectra CLEO-c Preliminary D 0 & D +

15. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 15 Leptonic Decay Goal: Extract f D, and eventually f Ds (with precision)  Test LQCD, if it passes then trust it in predicting f B, f Bs  Critical to measuring |V td |/|V ts |, one of the sides of the UT  (B   ) ~ 10 -4 – 10 -5 difficult

16. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 16 The Technique Form a “missing mass” (= 0 for ) MM 2 (K s  ) Data MC Test on K s , by ignoring the K s  Require 1 additional track from IP (cos  < 0.8)  Momentum too low for  system, require E cal <300 MeV  No additional showers with E>250 MeV “Extra” shower energy -studied with double-tagged events Energy deposition of muons Data MC Tag D Signal D e+e+ e-e- ++

17. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 17 CLEO-c Yellow Book: 1 fb -1 Mostly K L  ± background DATA Backgrounds ModeBF(%)# Events +0+0 0.13±0.020.31±0.04 ++ 2.77±0.180.06±0.05        2.64*B(D +   + ) 0.30 ±0.07  0   0.25±0.15negligible D 0 /D 0 0.16±0.16 Continuum -0.17±0.17 Total 1.00±0.25 8 events Results BES CLEO-c Lattice 2004 Isospin Mass Splittings Potential Model Rel. Quark Model QCD Sum Rules QCD Spectral Sum Rules MILC UKQCD

18. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 18 Summary  CLEO-c is off to a great start.  With only 57 pb -1 on  (3770) (3 fb -1 proposed), measurements are already comparable or better than world average.  Many more analyses are in the pipeline which I haven’t had time to discuss.  Many more exclusive BR’s being investigated  Several variants of inclusive and exclusive SL analyses  Techniques for estimating systematics established using data.  With more data, they will be reduced.  Look forward to many precision results in charm physics coming from CLEO.

19. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 19 Backup Slides

20. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 20 Particle ID  Use modes where the particle content is unambiguous.  For  : D +  K -    +, D 0  K s  D 0  K    For  : D +  K -    +, D 0  K     Then apply tagging requirements  If both  & K hypotheses analyzed,  3  dE/dx consistency, and   = (  (dE/dx)  2 -  (dE/dx)  2 + LogLik(  )- LogLik(  ) 2)  Drop RICH if:  RICHDONE is false, or, p(  ) 0.8 Particle Data  ID (%)MC  ID (%)  DATA -  MC (%)  97.57±0.1397.90±0.04-0.33±0.13 K89.70±0.5790.84±0.10-1.14±0.58 Correction applied: (0.3 ± 0.3)% for  and (1.3 ± 1.3)% for K

21. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 21 D Hadronic Systematics   requirement: compare yields with & without  E cut (1.5%)  FSR: Validated using J/y  , conservatively 0.5% for ST, 1% for DT   3770 : Lowe from 30.6 MeV  23 MeV and take shift in data as systematic (0.6%)  Resonant substructure: affects efficiencies, depending on mode: 0.4 – 1.5%  Trigger efficiency: trigger simulation  0.1%(K s  ) & 0.2% (K -     )  Multiple candidates:  Multiple candidates can result in choosing the wrong combination resulting in a loss in efficiency.  MC does not model the number of multiple candidates/event well.  Affects modes with  0 ’s:  1.30% for K s      0.44% for K -      , 0.32 for K -      Double DCSD: Unknown relative phase between DCSD & CAD amplitudes (0.8%)  Fit functions: 0.5%  Data processing 0.3%  Quantum (CP) correlations: Negligible

22. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 22 Yield Extraction in D Hadronic  Signals are fit using:   (3770) line shape, ISR, beam energy spread, momentum resolution (  3770 set to 30.6 MeV, as determined from data; reduced to WA for systematics)  Fit double tags first, using D  X / D  X, then fit single tags with sig pars fixed  Disentangle momentum resolution from beam energy spread Beam Energy, inc. ISR Signal region Signal Resolution (Signal MC) (5 parameters per D)  3 Gaussian widths (  1,  2,  3 ) 1.5  1 <  2 < 4  1 1.5  2 <  3 < 4  2  Two fractions: f 2, f 3, (1-f 2 -f 3 )  Fixes resolution for double & single tag fits in MC & data Background  1 correct D +1 incorrect D (f sig * ARGUS)  Mispartitioning of daughters ARGUS( )*GAUSS(  M)  Both D’s are background ARGUS(M D1 )*ARGUS(M D2 ) All candidates Candidate with best (M D +M D )/2 K  0 Double Tag Fit to Signal MC CBX 05-06, A. Ryd.

23. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 23 Branching Fraction Fitting CBX 04-36 (W. Sun)  N = Fitted yields of single & double tags ~ N DD *B i  E = Efficiency matrix  diagonal elements are efficiencies  off-diagonal are cross-feed probabilities  F = background probability matrix  n = Raw yields of single & double tags  b = estimated backgrounds from “other” D modes V is the variance matrix, and contains both statistical & systematic uncertainties Since  ij   i  j, correlated systematics cancel in N DD To first order, B i is independent of tag modes’ efficiencies. Corrected yields are given by: Test using Toy MC - 3 neutral + 2 charged modes - no biases - proper error estimation

24. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 24 Backgrounds Cross-feed “External” – Not simulated in MC Backgrounds that are dependent on fit parameters, ie., N DD, are updated after each iteration.. Single Tags Double Tags: assume only 1 fake contributes, since P(2 fakes) very small

25. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 25 Backgrounds - II M BC distributions for generic MC after signal modes and backgrounds considered are removed M BC distributions for non-DD MC

26. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 26 Fit to Generic MC (50X Data!) Worst difference is 2.1 , for K  But this is for 50X data  Scale by  50 for data  0.3  stat. Deemed acceptable by committee, and noted in PRL. Worst difference is 2.1 , for K  But this is for 50X data  Scale by  50 for data  0.3  stat. Deemed acceptable by committee, and noted in PRL.

27. Steven Blusk, Syracuse UniversityRecontres de Moriond, March 2005 27 Form Factor Shapes No efficiency corrections, resolution ~ 0.025 GeV 2 D 0 →K - e + νD0→π-e+νD0→π-e+νD 0 →K * - e + ν Future goal: slopes ~ 4%, form factors over all q 2