1 S. Stone Syracuse Univ. July 2003 Experimental Results in Heavy Flavor Physics Far too much interesting Material to include in 45 min. Apologies in advance. Many references not properly cited here. Apologies in advance. 2 Physics Goals Discover, or help interpret, New Physics found elsewhere using b & c decays - There is New Physics out there: Standard Model is violated by the Baryon Asymmetry of Universe & by Dark Matter Measure Standard Model parameters, the fundamental constants revealed to us by studying Weak interactions Understand QCD; necessary to interpret CKM measurements. 3 b & c quark decays Complete picture requires many studies including rare decays & CP Violation Hitoshi will cover CP I will give an overview and cover rare decays, and show how they can uncover new physics. Interpreting fundamental quark decays requires theories or models than relate quarks to hadrons in which they live and die. I will say something about this. Theoretical issues will be dealt with in more depth by Thomas Mannel this afternoon. Hitoshi will also cover future experiments 4 Some basics: b Lifetimes Note ratio B + / B o = 1.0730.014, a 5.2 difference Also B o According to proponents of the Heavy Quark Expansion model, there should be at most a 10% difference between b-bayron & B o lifetimes b 5 Some basics: Charm lifetimes New precise lifetimes from FOCUS, a challenge to QCD (from Pedrini) 6 The Basics: Quark Mixing & the CKM Matrix A,, and are in the Standard Model fundamental constants of nature like G, or EM multiplies i and is responsible for CP violation We know =0.22 (V us ), A~0.8; constraints on & d s b u c t mass massmass 7 The 6 CKM Triangles From Unitarity ds - indicates rows or columns used There are 4 independent phases: ( can be substituted for or as ) Best measured in B s decays 8 B Meson Decay Diagrams not rare a) is dominant More diagrams for baryons Semileptonic decays are the simplest. Lets start with them 9 Semileptonic B Decays average B sl On Y(4S), use high momentum lepton to tag flavor of 1 st B BaBar V cb or V ub B sl =10.890.23 LEP (most recent EW fit) = 0.22% 10.76% at Y(4S)) 10 B o & B + semileptonic Decays Accurate, separate B sl for B o & B + from Belle, tagged by fully reconstructing one B (preliminary) B (B + X )/ B (B 0 X )=1.14 0.04 0.01 B sl (%) sl (ns -1 ) BoBo 0.32 2.8 B+B 0.26 2.5 (B o + B + ) 0.20 1.9 Y(4S) 1.8 sl (B + )/ sl (B o )= 1.0630.038 11 | V cb | & |V ub | Theory versus Models Theories describe phenomena & make predictions based on general principles. They can have one or two unknown parameters (e.g. coupling constants) and if not exact, must prescribe a convergent series approximation. Examples are Lattice QCD (unquenched) and HQET Models contain assumptions. It is not only that the models may be wrong that causes us a problem, just as serious is that the errors on the predictions are difficult to estimate 12 HQET & V cb Heavy Quark Effective THEORY (HQET) (Isgur & Wise) QCD is flavor independent, so in the limit of infinitely heavy quarks q a q b occurs with unit form-factor [F(1)=1] when the quarks are moving with the same invariant 4-velocity, =1. Example: for B D* All form-factors are related to one universal shape that can be measured Corrections to F(1) due to finite quark masses are calculable along with QCD corrections. These corrections are parameterized in a series: n C n (1/m qi ) n, n=1, 2 To find V cb measure value of decay rate at =1, here D* is at rest in B rest frame 13 F(1)|V cb | using B D* Fit to function shape given by Caprini et al. Yields value of F(1)|V cb | & shape, parameterized by 2. F(1)|V cb |= (38.8 1.1) (HFAG) 2 =1.490.15 (HFAG) F(1)V cb Belle 14 Theoretical calculations of F(1) & |V cb | F(1)= QED QCD (1+ 1/m 2 +) Lukes theorem: no 1/m corrections (would be in D ) QED =1.007, QCD =0.9600.007 at two loops 1/m 2 involves 1/m b 2, 1/m c 2, 1/m c m b First Lattice Gauge calculations (quenched-no light quark loops) ultimate solution PDG (Artuso & Barberio) F(1)=0.910.05 Average: |V cb |=(42.61.2 exp 2.3 thy )x10 -3 15 The Heavy Quark Expansion Predict inclusive distributions Total widths, i.e. lifetimes b c or u semileptonic widths Relies on m b >> QCD Uses Operator product expansion Express decay widths in series in (1/m q ) n & s n Some authors: Chay, Bigi, Uratsalev 16 Problems with HQE Terms in 1/m b 3 are multiplied by unknown functions; hard to evaluate error due to these higher order terms Duality is assumed: integrated over enough phase space the exclusive charm bound states & the inclusive hadronic result will match at quark-level. But no way to evaluate the error Appears to miss b lifetime by 105% & b-baryon by 18 3%; however semileptonic decay may be easier Need experimental tests to evaluate errors Not used for charm as m c is not expected to be large enough, but could be done to see how much it diverged Perhaps use V cb as a test? Then could be used for V ub 17 Parameters of the HQE The heavy quark expansion of the B-meson (B X l ) decay rate is described to the order ( QCD /m b ) 2 by three parameters: 1 = ( M B ) B( v )| h v (iD) 2 h v | B( v ) is the kinetic energy of the residual motion of the b-quark. 2 = ( M B ) B( v )| h v (g/2) G h v | B( v ) the Chromo-magnetic coupling of the b-quark spin to the gluon field. Is determined from (M B* - M B ) mass splitting as 0.12 GeV 2 Decay rate also depends on quark-masses via, M B =m b + 1 2 m b M B* =m b + 1 2 m b (Thus defining ) 18 How to Measure 1 & Can determine 1 and and thus V cb by measuring moments in semileptonic decays Hadronic mass moments (ex: M X 2 - M D 2 , M D is spin-averaged D, D * mass) where B X Semileptonic moments Can also use b s decays, here we use the 1st moment of the photon energy 19 Moments (CLEO) Hadronic Mass & Lepton Energy moments found in semileptonic decays detecting the neutrino using missing energy b s moment determination shown later Fitting this & other data Bauer, Ligeti, Luke Manohar find V cb =(40.80.9)x10 -3 & m b =4.740.10 GeV (hep-ph/ ) Within ~5-10% of |V cb |=(42.61.2 exp 2.3 thy )x from D* - =0.35 0.07 GeV 1 = -0.240.07 GeV 2 exp errors only 20 BaBar Moments Result Using only BaBar hadronic moments & B sl : V cb =(42.11.00.7)x10 -3 again within 7% of D* m b 1S =4.640.090.09 GeV (M x 2 as function of lepton momentum, is now consistent with theory) M x 2 moments Doesnt include 1/m b 3 errors 1 contours See talk by Urs Langenegger 2002 21 Comparison of Hadron & Lepton Moments (BaBar) Lepton & Hadron moments differ somewhat. Does this indicate a Duality violation? Difference of 0.2 GeV in m b leads to 20% difference in V ub 3.5% difference 9% 22 Delphi 2003 Moments Results V cb =(42.40.60.9)x10 -3, using m b 1S =4.560.15, (see Hoang similar to using b s 0.07 GeV 0.07 GeV 2 (only experimental errors) Moments situation is still evolving, expect more results soon 23 | V ub | Use semileptonic decays. Even here, modeling errors will dominate - there isnt a precise theory & our path through this discussion will be perilous One method is to use exclusive decays: B B Currently unquenched lattice error ~10% + quenching error ~20% Also QCD sum rules (Ball) u B 24 V ub from exclusives: B B Use detector hermeticity to reconstruct CLEO finds rough q 2 distributions, good enough to limit model dependence of form-factor in determining branching ratios 25 V ub Results from CLEO theory error averages over quenched lattice results for q 2 >16 GeV 2, light cone sum rules for q 2 2.2 Cut M x +|D ( * ) K> are possible Factorization in B DD sJ seems to be in contradiction with cs 70 Conclusions There have been lots of surprises in Heavy Quark Physics, including: Long b Lifetime B o B o mixing Narrow D s ** states Now finding the effects of New Physics in b & c decays would not be a surprise - we expect to do it! Backups Follow 72 New versus old CLEO & BaBar Moments 73 New BaBar &Delphi Moments BaBar data now agree with theoretical shape (blue curve) using HQE input, although CLEO using b s moment and extrapolating from 1.5 GeV/c is not in great agreement BaBar central value 2002 Delphi 2002 74 Comparison of Bayesian & Scan Method Constraint on sin2 included in both Scan and Bayesian fit of M. Ciuchini et al., JHEP 0107, 13 (2001) Scan Bayes