Theories of exclusive B meson decays Hsiang-nan Li Academia Sinica Presented at Mini-workshop Nov. 19, 2004 Outlines • • • • • • • Naïve factorization and beyond QCDF vs. PQCD Parton kT? Scales and penguin enhancement Strong phase and CP asymmetry SCET Remarks Naïve factorization and beyond Naïve factorization (BSW) f fD F BD F B Color-allowed A( B D ) a1 f F a1 , a2 Color-suppressed BD a2 f D F : universal Wilson coefficients B Success due to “color transparency” Lorentz contraction Small color dipole D B Decoupling in space-time From the BD system To be quantitative, nonfactorizable correction? Large correction in color-suppressed modes due to heavy D, large color dipole Generalized naïve factorization Exp shows that the Wilson coefficients are not really universal Due to nonfactorizable correction? a1 a1 1 , a2 a2 2 Fine tune the mode-dependent parameters Equivalently, effective number of colors in a1( 2) C2(1) C1( 2) N C to data N C N Ceff 2 ~ 6 Not very helpful in understanding decay dynamics Strong phase and CP asymmetry When entering the era of B factories, CP asymmetries in charmless decays can be measured u b W b d g W q Tree q Penguin Interference of T and P A( B 0 ) T Pei ei2 Data ACP sin sin 2 Theory Extraction In naïve factorization, strong phase comes from the BSS mechanism Only source? Important source? Nonfactorizable correction, strong phase,… Need a systemic, sensible, and predictive theory Expansion in S , 1 mb Factorization limit… Predict not yet observed modes Explain observed data QCDF vs. PQCD • OCD-improved factorization=naïve factorization + QCD correction TII TI b b F B (a) Factorizable emission Leading F B (b) Vertex correction (c ) Non-spectator Sub-leading (d ) Exchange & Annihilation QCDF amplitude: AB TI F B TII B Two questions: The emission diagram is certainly leading…. But why must it be written in the BSW form ? Has naïve factorization been so successful that what we need to do is only small correction ? Both answers are “No” There is another option for factorizing the leading term, and naïve factorization prediction could be modified. However, the subleading calculation shows an end-point singularity 1 0 ( x) dx 2 , x ( x) x(1 x) in twist-3 nonspectator and in annihilation Need to introduce arbitrary cutoffs mB mB i H xC ln 1 H e , ln 1 Aei A Same singularity appears in the form factor This is the reason the form factor is not factorizable (calculable), and treated as a soft object (BSW form) Curiosity: 0 Why are the form factor O( S ) and the annihilation O( S ) , though none is calculable ? Want to calculate subleading correction?..... An end-point singularity means breakdown of simple collinear factorization Use more conservative kT factorization Include parton kT to smear the singularity 1 0 ( x) dx x( x kT2 mB2 ) The same singularity in the form factor is also smeared Then the form factor also becomes factorizable b b F B (a) F B (a ) (b) Perturbative QCD approach (b) Parton kT? Beneke’s 6 comments (ICHEP, Osaka, 2000) 1.Parton kT must be small, no help 2.kT breaks gauge invariance 3.kT factorization needs a proof 4.Twist-3 contribution is not complete 5.DA models should come from sum rules 6…..Could not remember all of them 1.Parton kT must be small, no help? Sudakov factors S Describe the parton Distribution in kT kT accumulates after infinitely many gluon exchanges Similar to the DGLAP evolution up to kT~Q 2.kT breaks gauge invariance? • kT factorization still starts with on-shell external particles • Decay amplitudes are gauge invariant • Parton kT is gained by exchanging gluons • Try to construct a gauge-invariant kTdependent wave function • Then hard kernels H are gauge-invariant • Convolution of H with WF models (prediction) is gauge-invariant 3.kT factorization needs a proof • Have proved it for semileptonic decays • Leading-power proof is easy: dynamics of different scales decouples • Proof for nonleptonic decays follows • Learned how to construct a gaugeinvariant kT-dependent WF from proof • ……. Scales and penguin enhancement Fast partons In QCDF this gluon is off-shell by O(mB2 ) F B In PQCD this gluon is off-shell by b O( mB ) Slow parton Fast parton For penguin-dominated modes, PQCD QCDF ~ 1.5 2 2 Strong phase and CP asymmetry Annihilation is similar to BSS mechanism Loop line can go on-shell Strong phase kT Sudakov gluons kT: loop momentum with the weight (Sudakov) factor Pinch-induced strong phase=FSI? Inclusive decay B X u l , , p, n u ~ b Cut quark diagram ~ Sum over final-state hadrons Off-shell hadrons On-shell Our concerns in 2000 • Is kT factorization an appropriate theory? • Is a pinched singularity the correct way to produce the strong phase? • Is the annihilation the only important source of strong phases? • Do we have the guts to present the prediction, large CP asymmetries with definite signs? Soft-collinear Effective Theory • An effective theory at large energy E • Effective degrees of freedom: collinear fields, soft fields,… • Expansion of Lagrangian in 1/E in terms of effective operators • Wilson coefficients: hard kernels • Convenient for factorization proof. Effective operators define nonlocal matrix elements (wave functions) mB QCD At lower energy, detailed structure of form factor can be seen Effective (soft) operator for energy < mB nonpert SCET • SCET is more careful in scale separation. • A form factor is split into two pieces: soft and hard contributions. • No annihilation contribution. • Need Acc (nonperturbative charming penguin) to introduce large strong phases. • All the above parameters are from fitting. T can be chosen to be real, and C is assumed to be real. 0.016-0.064 BBNS 04 In fact, charming penguin is factorizable (no IR divergence) and small Li, Mishima 04 BBNS 04 Acc is large My personal comments • A bit disappointed by that SCET was led to this direction. • I can get the same “prediction” using T, C, P, assuming C to be real---4 parameters with 4 inputs. • The pi0pi0 amplitude is fixed by the isospin relation. • A stringent test will be Kpi modes. Need more parameters. pi+pi-: T+P pi+pi0: T+C pi0pi0: C-P Amplitude topologies Remarks • Compard to HQET, exclusive theories are still not yet well established: Matrix elements (wave function) not known Subleading corrections not clear Mechanism not explored completely ……… • It is definitely a much richer and challenging field. Experimental data Exact solution PQCD 0.23 e 0.07 0.05 (1435) o i 0.2 0.14 pi0pi0 branching ratio gets smaller. P/T approaches theory. New data: ~0.38 B->K pi amplitudes and data K pi data imply large Pew ? • The updated data imply a large C, instead of a large Pew. K+piLarge strong phase between P and T is confirmed T exp(i phi3) P T exp(-i phi3) (T+C) exp(i phi3) Pew K+pi0 (T+C) exp(-i phi3) Buras’s picture K+pi0 T exp(i phi3) P T exp(-i phi3) Pew This is a possible solution, but ruled out by the pi pi data • Charming penguin: need many Acc for each polarization and for each mode. • Rescattering: hard to accommodate rho K*, phi K* simultaneously. • b->sg: negligible due to G parity. • Annihilation: not sufficient for phi K*, but able to explain rho K*. • rho+ K*0: P rho0 K*+: P+T • Interference between P and T enhances the longitudinal polarization
© Copyright 2026 Paperzz