Physics in the Caltech BaBar group

Transcription

1 Physics in the Caltech BaBar group Gregory Dubois-Felsmann Department of Energy grant review 21 July 2004

2 Doing physics with BaBar BaBar has amassed a remarkable dataset: xxx*10 6 B-meson pairs (yyy fb -1 of Υ(4S) data) and developed an effective infrastructure for collecting and processing this data, and using it for physics. Caltech has played an important role in making this possible in general, but we also are involved with a number of specific data analyses, as well as with parts of the BaBar organization that help in producing and releasing results to the physics community. Physics in the Caltech BaBar group 2

3 Overview Organizational contributions Analysis topics Radiative penguin B decays B K π π γ B K ( * ) l + l Leptonic B decays B γ l + ν B invisible (νν or?) (Samuel, Porter) (Samuel, Ryd) (Chen, gpdf) (Albert) Pentaquark search (Narsky) Hadronic B decays B h + h + h (charmless) Dalitz plot analysis B D ( * ) D ( * ) (Dvoretskii, Hitlin, gpdf) (Albert, Piatenko) and A related project CKM fitting (gpdf, Hitlin, Porter) Physics in the Caltech BaBar group 3

4 Organizational contributions Convening analysis working groups Dubois-Felsmann is co-convener of the leptonic b & c decays AWG and convener of the CKM fitting informal working group Providing standards and tools Porter is chair of the Statistics Working Group The SWiG provides recommendations and instruction to BaBar physicists on the appropriate use of statistical techniques in analyses Dubois-Felsmann is providing design leadership for an effort to improve data set bookkeeping tools for analysis users And, as for all other collaborating institutions Serving on analysis review committees Many group members have participated in or chaired these, BaBar s primary internal review procedure for results to be released to the public. Participating in final publication reviews Physics in the Caltech BaBar group 4

5 Radiative penguin B decays These decays involve loop diagrams, and are second-order weak or weak/electromagnetic decays: They can therefore be quite sensitive to the presence of new physics at high mass scales, appearing in the loops. Physics in the Caltech BaBar group 5

6 B Kππ γ (Samuel, Porter) The inclusive rate of the flavor-changing neutral current reaction b s γ has been well-measured by BaBar and other experiments, and the results are in agreement with theoretical predictions. Predictions of exclusive rates, however, are less precise and vary due to model differences. BaBar has measured B(B K* γ, K* K π) and B(B K 2 *(1430) γ, K 2 *(1430) K π), but many other final states have not been studied. Belle has measured B(B + K + π + π γ) = (2.5 ± / 0.2) * 10-5 Predictions for exclusive branching fractions B K X γ for various K X resonances are in the range (1 10) * Physics in the Caltech BaBar group 6

7 Contributing processes The reaction B K π π γ can occur via many intermediate resonances. The lowest contributing kaon resonances beyond the K*(892) are Resonance J^P Mass(GeV) Width(GeV) Decays to K pi pi K1(1270) K rho ( 42%) K*0(1430) pi ( 28%) K* pi ( 16%) K1(1400) K* pi ( 94%) K rho ( 3%) K*(1410) K* pi (>40%) K2*(1430) K* pi ( 25%) All of these processes, plus non-resonant B K π π γ, contribute coherently to the B K π π γ rate. The small mass differences among these resonances and their large widths make them difficult to disentangle. The sizes and relative phases of contributing amplitudes have not been measured. Physics in the Caltech BaBar group 7

8 Polarization measurement The polarization of the photon in b s γ may be measured by observing the variation of the K π π Dalitz distribution with the angle of the K π π decay plane relative to the photon direction. (Gronau, Grossman, Pirjol, and Ryd, Phys. Rev. Lett. 88 : , 2002) In the Standard Model, the photon is approximately fully left-handed for b quark decays, right-handed for anti-b quark decays. Some SUSY models and others beyond the Standard Model predict a significantly different polarization. Physics in the Caltech BaBar group 8

9 Polarization measurement II The simulated Dalitz distribution for B + K 1 (1270) + γ, K 1 (1270) + K π π via K* π or K ρ, is shown below for events with a positive value of the decay plane angle cosine. Asymmetry is caused by the interference between the processes K 1 (1270) + K* + π 0, K* + K S0 π + K 1 (1270) + K* 0 π +, K* 0 K S0 π 0 M Kπ,1 We have studied the experimental feasibility of this measurement, and determined that it is out of reach with the current dataset. However, it should be possible with future BaBar datasets, or at a experiment. M Kπ,2 It is nevertheless important to understand the individual resonance contributions with current data. Physics in the Caltech BaBar group 9

10 Analysis Plan for analysis with present / near-term BaBar data: F. Porter and A. Samuel are developing an analysis to measure B(B K π π γ) in four of the six charge modes. We will measure the combined distribution of the Dalitz variables, decay plane angle, and K π π mass. From this distribution, we hope to extract or constrain the relative sizes and phases of the various B K X γ amplitudes. Physics in the Caltech BaBar group 10

11 Status We have developed a Monte Carlo generator for B K π π γ, including angular correlations and a fully coherent treatment of interfering modes. We are developing an analysis procedure based on the BaBar semi-inclusive b s gamma analysis. We are working with toy fits to the kinematic distributions in order to understand how to disentangle the contributing amplitudes. This will be extended to the full simulation when it becomes available. We hope to obtain a preliminary result by this winter. Physics in the Caltech BaBar group 11

12 B K ( * ) l + l (Samuel, Ryd, Porter) Motivation The flavor-changing neutral current decays B K l + l and B K* l + l occur in the Standard Model via one-loop radiative penguin and box diagrams: Study of these decays will help in understanding the basic Standard Model physics of such processes, and it also presents an opportunity for the observation of new-physics effects in the loops. Physics in the Caltech BaBar group 12

13 Theory The Wilson coefficients C 7 eff, C 9 eff, and C 10 governing these decays in the Operator Product Expansion are well-understood, but predictions for rates vary due to modeling variations in long-distance QCD effects. Precise measurements of the B K ( * ) l + l rates can distinguish among these models. BF x 10^6 Prediction Technique Kll K*ee K*mumu Ali et al. '01 LCSR Ali et al. '99 LCSR Chen and Geng '02 PQCD Zhong et al. '02 LCSR Melikhov et al. '98 DQM Faessler et al. '02 RCQM 0.55 Choi et al. '02 RCQM 0.50 Some recent theoretical predictions In addition, the rates and kinematic distributions are sensitive to a wide range of beyond-standard-model physics, such as SUSY models, that predict diagrams with new, heavier particles in the loops. Physics in the Caltech BaBar group 13

14 B K ( * ) l + l measurement The Caltech group (F. Porter, A. Ryd, A. Samuel) participated in BaBar s B K ( * ) l + l study from its inception through the Run 3 (2003) incarnation of the analysis. Using a 113 fb -1 data sample, we established the existence of Kll decays, found evidence for the K*ll decay mode, and measured the branching fractions to be B(B K l + l ) = ( / 0.13 ±0.04) * 10-6 B(B K* l + l ) = ( / 0.29 ±0.10) * 10-6 B fi K* l + l (Published in B. Aubert et al, Phys. Rev. Lett. 91, (2002) and subsequent conference proceedings.) Physics in the Caltech BaBar group 14

15 Leptonic B decays Leptonic decays of B (and D) mesons provide a theoretically clean laboratory for checking certain basic QCD predictions, measuring constants of interest as input to other computations (e.g., f B ), and assisting ultimately in constraining the overall Standard Model description of weak decays of heavy-quark mesons. In certain cases of rare or forbidden processes they could also provide windows on new physics γ Physics in the Caltech BaBar group 15 b W + B +,B 0 u,d ν,l, ν Dubois-Felsmann is co-convener of the BaBar analysis working group on this topic W l +, ν

16 B + γ l + ν (Chen, Dubois-Felsmann, Hitlin) The purely leptonic decay B + l + ν l provides a theoretically clean means of testing a QCD calculation of a simple process: b u g l + W G 2 2 ( + + F Vub m ν) = τ l B B l f B BmBml π m B ν ~ Calculable in Lattice QCD But the rate is low and expected to be hard to observe, due to helicity suppression for light leptons (we expect B( B µν µ ) ) or a multiple-neutrino final state for the larger B( B τν τ ) ). None of these modes have yet been seen. New limits will be available from BaBar shortly. Only the tau lepton analyses appear to be nearing the SM value. The radiation of a photon relieves the helicity suppression, at the cost of additional theoretical uncertainty (at least for the time being). Physics in the Caltech BaBar group 16

17 Analysis possibilities Theoretical predictions: B(B + γ l + ν) ~ (1 4) 10-6 Analysis techniques: Neutrino reconstruction is required Recoil against fully reconstructed hadronic B decay very clean ν reco`n Recoil against semileptonic B decay more events, poorer constraint Inclusive analysis: sum up all missing E, p in event high stats, high bkgd Too few tags yet for recoil analyses wait for at least ~400 fb -1 Previous measurement: CLEO (PRD ) using 2.5 fb -1 : Simple inclusive neutrino reconstruction B( B eν e γ ) < (90% C.L.) (apparent unidentified backgrounds) B( B µν µ γ ) < (90% C.L.) Prospects: In 82 fb -1 sample, w/ CLEO 2% eff., expect ~6 ev/mode. Backgrounds? Physics in the Caltech BaBar group 17

18 Analysis scheme Blind analysis, starting with electron mode Validate simulation with control samples before unblinding Refine CLEO s inclusive approach Improved continuum rejection by adding thrust cut, Rejection of two-photon background using electron charge-angle correlation Iterative cut optimization procedure Obtain final result with maximum-likelihood fit to missing E missing p distribution (peaks at zero for a single neutrino) Physics in the Caltech BaBar group 18

19 Current sensitivity Signal efficiency for electron mode: 4.5% in a cut-and-count approach Signal scaled to BR of Expected scaled signal/background for lumi of 81.9 fb 1 assuming BF of : 11.6 / 82.4 FOM S/ B = % CL UL of CLEO (1997): For 500 fb 1, expect a 3σ measurement Physics in the Caltech BaBar group 19

20 Analysis Plans Finish control-sample studies of discriminating variables B D π, D K π B D* l ν Finalize maximum-likelihood fit Evaluate systematics Unblind on on-resonance data Also look at radiative muonic mode Planning for results at the winter conferences! Physics in the Caltech BaBar group 20

21 B invisible (+ γ) (Albert) B 0 fi invisible (+ g) J. Albert Caltech Physics in the Caltech BaBar group 21

22 Motivation from SUSY and elsewhere B 0 invisible (+ gamma) is sensitive to new physics Several new physics models predict a small but significantly nonzero branching fraction for B 0 invisible Dedes, Dreiner, & Richardson hep-ph/ (SUSY model, attempt to explain NuTeV anomaly, B (B 0 invisible) = O(10-6 )) Agashe, Deshpande, & Wu hep-ph/ Agashe & Wu hep-ph/ Davoudiasl, Langacker, & Perelstien hep-ph/ (Large extra dimensions, B(B 0 invisible) = O(10-7 )) Diagrams (in SM): Standard model contamination is small and well-predicted. Helicity suppression kills SM decay B 0 νν. Even without such suppression (B 0 ννγ), these are still second-order weak decays. Lu & Zhang B (B 0 ννγ) = 2.5 x 10-9 in SM hep-ph/ Physics in the Caltech BaBar group 22

23 Invisible B decays? Wouldn t something as unusual as an invisible decay have been noticed already? There are no significant limits on invisible decays of heavy flavor at all. Invisible decays can easily elude notice branching fractions could even be up to order 1-5% without being in conflict with any present B physics results. If a B were to just disappear, how would one reconstruct it? The Y(4S) is a unique environment for searching for these types of decays. If one reconstructs a B, one knows that another must have been produced. One reconstructs one B and looks for nothing else in the event. Physics in the Caltech BaBar group 23

24 Semileptonic tagged B reconstruction The Run 1+2 dataset (release 10) is used fb -1, 88.9 million BB events D (*) lν where D* D 0 ð or D K ðð is reconstructed D 0 is reconstructed to 3 modes: K ð, K3ð, and K ðð 0 Mass cuts used in reconstruction: ± 40 MeV for all modes except for K ðð 0 ( ± 70), 130 < Ä m(d*- D 0 ) < 170 MeV. Also reconstruct D (*0) lν for B ± invisible check Common BToDlnuX production with Knunu and TauNu. The following loose cuts are applied: -2.5 < cosè B,Dl < 1.1 Prob(D vtx) > B 0 D (*)- l + n (K + p - ) p - (K + p - p 0 ) p - (K + p - p + p - ) p - K + p - p - Efficiency for tag reconstruction: 0.2% Then, rest of event is observed for consistency with nothing (+ machine background) Prob(Dl vtx) > Kaon must not pass loose electron selection If mode=k ð, K must be tight kaon, and ð must not be NotAPion If the lepton in the Dlυ is an electron, then the pions in the Dlυ should not pass tight electron cuts Number of remaining tracks 2, # remaining photons < 10 Amount of remaining energy < 1GeV [for B 0 invisible +γ, use (amount of remaining energy) (energy of highest- energy γ)] R2all < 0.6 Physics in the Caltech BaBar group 24

25 Signal side reconstruction Make sure there are no additional reconstructed charged tracks, sharply restrict the number of additional reconstructed K L, photons, π 0, etc. Look at amount of remaining (neutral) energy in the event: Strongly peaked near zero for signal. Backgrounds dominated by larger amounts of remaining energy. Two analysis techniques: 1) make a set of optimized cuts to maximize signal sensitivity, then subtract remaining background; 2) perform a likelihood fit to combined signal and background distributions for remaining energy. Both analysis techniques are performed. Physics in the Caltech BaBar group 25

26 Unblinded B0 invisible signal plots Likelihood fit: N signal = 20.6 ± 9.0 (stat.) D mass Data Data Data K ð K ððð K ðð 0 MC MC MC D mass Data K ðð MC Physics in the Caltech BaBar group ETotLeft 26 ETotLeft

27 Conclusions B 0 invisible (+ gamma) is sensitive to new physics, and had not previously been measured or constrained. Experimentally-motivated SUSY and extra-dimensions models can give a significant rate for this process. We obtain the following limits: B (B 0 fi invisible) B (B 0 fi invisible + g) < 23.8 x 10-5 < 5.0 x 10-5 A report of this result has been accepted for publication in PRL (preprint available at hep-ex/04xxxxx) Physics in the Caltech BaBar group 27

28 Pentaquark search (slides not yet received from Ilya) Physics in the Caltech BaBar group 28

29 Hadronic B decays The Caltech group is involved in studies of two families of hadronic B meson decays: D ( * ) D ( * ) (b ccd) and the charmless h + h + h, h=k,π. These decays are interesting both as probes of the hadronic dynamics of B decays and, in time-dependent analyses, of their CP(-violating) structure. D ( * ) D ( * ) decays probe the unitarity triangle angle β (albeit with additional hadronic uncertainties compared to the charmonium golden modes ) and, it s been more recently realized, the angle γ (with D S ( * ) D ( * ) information). h + h + h decays, if their resonant substructure is understood, probe the angle α, with the possibility of resolving hadronic uncertainties by combining a variety of measurements related by isospin and SU(3). Physics in the Caltech BaBar group 29

30 B h + h + h (h=k,π) (Dvoretskii, Hitlin, Dubois-Felsmann) Physics motivation Charmless hadronic decays of B meson can be used to: Constrain parameters of the CKM matrix. SU(3) symmetry in the light-quark sector can be used to reduce theoretical uncertainties. Study hadronization dynamics Three-body decays can proceed via multiple interfering channels. An amplitude analysis allows the determination of contributing quantum mechanical amplitudes. SU(3) relations can be used to place model-independent constraints on penguin contributions. Amplitude analysis can be used to determine UT angles á, ã. Physics in the Caltech BaBar group 30

31 B h + h + h References Theoretical: M. Beneke, G. Buchalla, M. Neubert and C.T. Sachrajda, Phys.Rev.Lett 83 (1999) 1914 Y. Grossman, Z. Ligeti, Y. Nir, H. Quinn, Phys. Rev. D68 (2003) Experimental BABAR: quasi two-body or inclusive Phys. Rev. Lett. 91 (2003) , hepex/ BELLE-CONF-0338 (amplitude analysis of B+ K+K+K-, B+ K+ð+ ð-) Physics in the Caltech BaBar group 31

32 B h + h + h Datasets, final states Using BaBar Run 1-3 datasets: fb - 1 on-resonance, 12.1 fb MeV below Õ(4S) resonance Using three body charmless preselection, n-tuples shared with other analyses Final states include pi+pi+pi-, K+pi+pi-, K+pi+K-, and K+K+K-. For completeness and cross-checks, we include the highly suppressed pi+pi+k-, K+K+pi- Simultaneous analysis of multiple final states because of K/pi misidentification Branching fractions are of order 10-5, no signal is expected in the last two modes, they are used to control K/pi misidentification Kaon loose PID selector used to classify candidate into final states Candidate multiplicity per event is 1.02, random candidate is chosen if multiple candidates are present Physics in the Caltech BaBar group 32

33 B h + h + h Classification probabilities, loose K Mode out/in pi+pi+pi- K+pi+K- K+pi+pi- K+K+K- pi+pi+pi K+pi+pi K+pi+K K+K+K pi+pi+k K+K+pi Total eff Physics in the Caltech BaBar group 33

34 B h + h + h Continuum background suppression Three layer, four input perceptron neural network is used to suppress the dominant qqbar continuum background. The variables used are: Cosine between candidate thrust axis and z axis: Cosine between candidate momentum in c.m. and z axis L 0, L 2 Legendre moments of rest-of-event momentum flow w.r.t. candidate thrust axis Some correlation is observed between neural network output and Dalitz plot coordinates. Mode dependent cut on neural network output is applied to achieve qqbar background suppression Physics in the Caltech BaBar group 34

35 B h + h + h Neural network output Simulation Data Red - signal MC, green - B-bkg, blue - sideband Physics in the Caltech BaBar group 35 Black data, blue fit

36 B h + h + h m ES fits Distributions of m ES allow visual and quantitative tests of purity Physics in the Caltech BaBar group 36

37 B h + h + h Event counts Mode Events selected Sideband background pi+pi+pi ± 46 K+pi+pi- [pi+pi+k-?] ± 22 K+pi+K- [K+ pi+ pi-?] ± 38 K+K+K- [K+pi+K-?] pi+pi+k- [K+ K+ pi-?] ± ± 11 Focus on these for now K+K+K ± 16 Begin with the lower-background final states, especially K+K+K- Ultimately full six-mode fit with PID crossfeeds will be valuable Physics in the Caltech BaBar group 37

38 B h + h + h Ready for amplitude fits Dalitz plots after all event selections and D-decay vetoes: Currently working on numerical stability of fits (phases!) Physics in the Caltech BaBar group 38

39 B D ( * ) D ( * ) Physics in the Caltech BaBar group 39

40 B D ( * ) D ( * ) (Albert, Piatenko) ( ) B 0 t = 0 D ( *)- D ( *) + ( ) B 0 D (*)- D ( *) + Time-dependent CP-violating asymmetry from tree amplitude is proportional to sin(2â) Penguin amplitude can add different phase. Phase correction due to penguins expected to be small in SM (< 0.1 correction to measured sin(2â) ) however models which can enhance loops can produce large corrections. 0 B B mixing 0 CP decay A f CP A f CP f CP -sin(2â)sin(ä mät) Grossmann & Worah, Phys. Lett. B395, 241 (1997) Physics in the Caltech BaBar group 40 D*+D*- is not a CP eigenstate Can measure CP-odd fraction by separating out the angular momentum states Result: R = D*D* is mostly CP-even! 0 A A + A + A R = 0.07± 0.06± 0.03

41 Gamma from D (s) ( * ) D ( * ) decays! Datta and London present a method for extracting gamma from measurements of D (s) (*) D (*), using a combination of branching fraction and CP asymmetry information. hep-ph/ (Phys.Lett.B (2004)) The CP asymmetry from the tree amplitude measures sin2â, so where does γ come in? g comes from the u- and t-penguin terms: Physics in the Caltech BaBar group 41

42 Gamma from D (s) ( * ) D ( * ) decays! For a given B D (*) D (*) decay, there are 3 observables: a branching fraction, a direct CP asymmetry, and a time-dependent CP asymmetry *: This is 3 equations in 5 unknowns. More information required The additional information can be obtained by inputting two things: 1) beta, taken from charmonium decays(!), and 2) branching fractions of B D (*) s D (*) decays. * Note that the BF s for the charged and neutral modes are the same by isospin, and the direct CP asymmetries are the same in the charged and neutral modes (within the SM). Physics in the Caltech BaBar group 42

43 D (*)+ D (*)- previous data sample (88 million BB events) 126 ± 12 events in D*D* and 113 ± 13 events in D*D (combinatoric & peaking bkgd. subtracted): D *+ D * - D *+ D s - D*D signal D *± D ± D *+ D *- ÄE Expect ~300 events in each mode this time around! New yields in runs 1-3: Yields in runs 1-4 data will be coming shortly Physics in the Caltech BaBar group 43

44 B D ( * ) D ( * ) 2002 Results C D *- D + S D *- D + C D *+D - S D *+D - = ± 0.37(stat.) ± 0.10(syst.) = ± 0.69(stat.) ± 0.12(syst.) = ± 0.40(stat.) ± 0.12(syst.) = ± 0.75(stat.) ± 0.14(syst.) Im(ë + ) D *+ D *- ë + D *+ D *- A = 0.05 ± 0.29(stat.) ± 0.10(syst.) = 0.75 ± 0.19(stat.) ± 0.02(syst.) = ± 0.11(stat.) ± 0.05(syst.) Physics in the Caltech BaBar group 44

45 B D ( * ) D ( * ) 2004 Prospects We will be measuring time-dependent asymmetries in D(*)D(*) with runs 1-4 (corresponding to better than 230 fb -1 of data in all). Toy and signal MC studies have begun. We have made fit-input-data files in the new BaBar computing model for signal and generic MC. We are presently checking that efficiencies and background in the new system are similar to our previous data (Tim Piatenko). We will proceed with tagging and vertexing studies when this has been validated Goal (end of summer) is a comprehensive D(*)D(*) CP asymmetry publication with gamma measurement! With 2002 data (London & Datta method): rad ? Physics in the Caltech BaBar group 45

46 CKM fitting Physics in the Caltech BaBar group 46

47 Prospects Physics in the Caltech BaBar group 47