Introduction to Operator Product Expansion

Transcription

1 Introduction to Operator Product Expansion (Effective Hamiltonians, Wilson coefficients and all that... ) Thorsten Feldmann Neckarzimmern, March 2008 Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

2 Outline 1 b cdū decays Born level Quantum-loop corrections Effective Operators Wilson Coefficients 2 b s(d) qq decays Current-current operators Strong penguin operators Electroweak corrections 3 B- B Mixing 4 Summary: Effective Theory for b-quark decays 5 Hadronic matrix elements for B-meson decays B Dπ B ππ and B πk Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

3 ... Some introductory remarks... Physical processes involve different typical energy/length scales Separate short-distance and long-distance effects: New physics : δx 1/Λ NP Electroweak interactions : δx 1/M W Short-distance QCD(QED) corrections : δx 1/M W 1/m b Hadronic effects : δx < 1/m b Sequence of Effective Theories. Perturbative and non-perturbative calculations. Definition / measurement of hadronic input parameters. Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

4 Central Notions to be explained Disclaimer: The dynamics of strong interactions in B-decays is very complex and has many faces. I will not be able to cover everything, but I hope that some theoretical and phenomenological concepts become clearer... Factorization 1 separation of scales in perturbation theory (OPE, ET) 2 simplification of exclusive hadronic matrix elements (QCDF) Operators in the weak effective Hamiltonian (current-current, strong penguins, electroweak penguins) Matching and Running of Wilson coefficients Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

5 Example: b cdū decays Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

6 b cdū decay at Born level Full theory (SM) Fermi model ( ) 2 g αβ + qα q β g 2 (b c) M 2 J W 2 α q 2 MW 2 J (d u) β q M W 2 G F J α (b c) g αβ J (d u) β Energy/Momentum transfer limited by mass of decaying b-quark. b-quark mass much smaller than W -boson mass. q m b m W Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

7 Effective Theory: Analogously to muon decay, transition described in terms of current-current interaction, with left-handed charged currents J α (b c) = V cb [ c γ α (1 γ 5 ) b], J (d u) β = Vud [ d γβ (1 γ 5 ) u ] Effective operators only contain light fields ("light" quarks, electron, neutrinos, gluons, photons). Effect of large scale M W in effective Fermi coupling constant: g 2 8MW 2 G F GeV 2 Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

8 Quantum-loop corrections to b cdū decay Momentum q of the W -boson is an internal loop parameter that is integrated over and can take values between and +. We cannot simply expand in q /M W! Need a method to separate the cases q M W and q M W. OPE / Factorization Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

9 IR and UV regions in the Effective Theory full theory = IR region (M W ) + UV region (m b,c 0) = + I(α s ; m b M w, mc m b )/G F I IR (α s ; m b µ, mc m b ) + I UV (α s ; µ m W ) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

10 IR and UV regions in the Effective Theory full theory = IR region (M W ) + UV region (m b,c 0) + I(α s ; m b M w, mc m b )/G F O loop (α s ; m b µ, mc m b ) + C (α s ; µ m W ) O tree Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

11 IR and UV regions in the Effective Theory full theory = IR region (M W ) + UV region (m b,c 0) + I(α s ; m b M w, mc m b )/G F O loop (α s ; m b µ, mc m b ) + C (α s ; 1-loop matrix element of operator O in Eff. Th. µ m W ) O tree 1-loop coefficient for new operator O in ET independent of M W UV divergent µ independent of m b,c IR divergent µ Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

12 For the curious: Operator Product Expansion Study time-ordered product of two general operators: Z d 4 p e ipx T (φ(x)φ(0)) For small values of x, the product can be expanded into a set of local composite (renormalized) operators O i (0) and c-numbered Wilson coefficients c i (x): T (φ(x)φ(0)) x 0 = X i c i (x) O i (0) Construction is independent of additional field operators appearing in Green functions / scattering amplitudes (i.e. universal for all processes involving the T-product under consideration) In momentum space, this corresponds to the limit of large momenta/energies/masses... Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

13 Effective Operators for b cdū short-distance QCD corrections preserve chirality; quark-gluon vertices induce second colour structure. H eff = G F 2 V cb V ud i=1,2 C i (µ) O i + h.c. (b cdū) Current-current operators: (b cdū, analogously for b qq q decays) O 1 = (d i Lγ α u j L )(cj L γα b i L) O 2 = (d i Lγ α u i L)(c j L γα b j L ) The so-called Wilson coefficients C i (µ) contain all information about short-distance physics above the scale µ (SM) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

14 Wilson Coefficients in Perturbation Theory 1-loop result: C i (µ) = { 0 1 } + α s(µ) (ln µ2 4π MW ) { } + O(α 2 s) Wilson coefficients depend on the renormalization scale µ Matching For µ M W the logarithmic term is small, and C i (M W ) can be calculated in fixed-order perturbation theory, since αs(m W ) π 1. Here M W is called the matching scale. Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

15 Anomalous Dimensions In order to compare with experiment / hadronic models, the matrix elements of ET operators are needed at low-energy scale µ m b Only the combination X C i (µ) O i (µ) is µ-independent (in perturbation theory). Need Wilson coefficients at low scale! i Scale dependence can be calculated in perturbation theory: Loop diagrams in ET are UV divergent anomalous dimensions (matrix): ( ) ln µ C αs (µ) i(µ) γ ji (µ) C j (µ) = 4π γ(1) ji +... C j (µ) γ = γ(α s) has a perturbative expansion. Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

16 RG Evolution ( running ) In our case: ( ) { γ (1) 2 6 Eigenvectors: C± = 1 = 2 (C 2 ± C 1 ) 6 2 Eigenvalues: γ (1) ± = +4, 8 Formal solution of differential equation: (separation of variables) ln C±(µ) C ±(M) = Z ln µ ln M d ln µ γ ±(µ ) = Z αs(µ) α s(m) dα s 2β(α s) γ±(αs) Perturbative expansion of anomalous dimension and β-function: γ = αs 4π γ(1) +..., 2β dαs d ln µ = 2β 0 4π α2 s +... «(1) γ αs(µ) ± /2β 0 C ±(µ) C ±(M W ) α s(m W ) (LeadingLogApprox) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

17 Numerical values for C 1,2 in the SM [Buchalla/Buras/Lautenbacher 96] operator: O 1 = (d i Lγ µ u j L )(cj L γµ b i L ) O 2 = (d i Lγ µ Tu i L )(cj L γµ b j L ) C i (m b ): (LL) (LL) (NLL) (NLL) (modulo parametric uncertainties from M W, m b, α s(m Z ) and QED corr.) (potential) New Physics modifications: new left-handed interactions (incl. new phases) C 1,2 (M W ) C 1,2 (M W ) + δ NP (M W, M NP ) new chiral structures extend operator basis (LR,RR currents) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

18 Next Example: b s(d) qq decays Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

19 b s(d) q q decays Current-current operators Now, there are two possible flavour structures: V ub V us(d) (ū Lγ µ b L )( s(d) L γ µ u L ) λ u O (u) 2, V cb V cs(d) ( c Lγ µ b L )( s(d) L γ µ c L ) λ c O (c) 2, Again, α s corrections induce independent colour structures O (u,c) 1. Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

20 b s(d) q q decays strong penguin operators New feature: Penguin diagrams induce additional operator structures Strong penguin operators: O 3 6 Chromomagnetic operator: O g 8 Wilson coefficients numerically suppressed by α s / loop factor. Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

21 Eff. Hamiltonian for b s(d)q q decays (QCD only) H eff = G F 2 O 3 = ( s i Lγ µ b i L) q t i=1,2 G F λ t 2 ( C i (µ) λ u O (u) i 6 C i (µ) O i i=3 ) + λ c O (c) i G F λ t C g 8 (µ) Og 8 2 ( q j L γµ q j L ), O 4 = ( s i Lγ µ b j L ) q t ( q j L γµ q i L), O 5 = ( s i Lγ µ b i L) q t ( q j R γµ q j R ), O 6 = ( s i Lγ µ b j L ) q t ( q j R γµ q i R), O g 8 = g s 8π 2 m b ( s L σ µν T A b R ) G A µν. virtual u- and c-contributions (with m u = m c 0) : use λ u + λ c = λ t gluon couples to left- and right-handed currents. chromomagnetic operator requires one chirality flip! (m s is set to zero) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

22 Matching and running for strong penguin operators Matching coefficients depend on top mass, x t = m 2 t /M2 W C 3 (M W ) C 5 (M W ) = α s 24π Ẽ0(x t ) +... C 4 (M W ) C 6 (M W ) = + α s 8π Ẽ0(x t ) +... Beware! Different conventions/schemes (BBL,CMM,BBNS) Mixing of O 1 6 under RG evolution described by 6 6 matrix: 0 1 γ = 0 0 B A 0 0 Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

23 Chromomagnetic operator Matching: C g 8 (M W ) = 1 2 E 0(x t ) +... Usually, one considers scheme-independent linear combination: g, eff C8 = C g 8 (M W ) + (z i : scheme-dependent coefficients) g, eff RG mixing between C 1 6 and C8 6 z i C i (M W ) i=1 Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

24 Electroweak Corrections Penguin and box diagrams with additional γ/z exchange: Electroweak Penguin Operators O 7 10 O 7 = 2 3 ( si L γµbi L ) X q t e q ( q j L γµ q j L ), O 8 = 2 3 ( si L γµbj L ) X q t e q ( q j L γµ q i L ), O 9 = 2 3 ( si L γµbi L ) X q t e q ( q j R γµ q j R ), O 10 = 2 3 ( si L γµbj L ) X q t e q ( q j R γµ q i R ). depends on electromagnetic charge of final state quarks! Electromagnetic operators O γ 7 main contribution to b s(d)γ decays. Semileptonic operators O 9V, O 10A main contribution to b sl + l decays. electroweak corrections to matching coefficients Isospin effects in non-leptonic decays (e.g. B πk ) and radiative decays (b s(d)γ, b sl + l ) are particularly sensitive to New Physics! Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

25 Electroweak Corrections Penguin and box diagrams with additional γ/z exchange: Electroweak Penguin Operators O 7 10 depends on electromagnetic charge of final state quarks! Electromagnetic operators O γ 7 O γ 7 = main contribution to b s(d)γ decays. Semileptonic operators O 9V, O 10A main contribution to b sl + l decays. e 8π 2 m b ( s L σ µν b R ) F µν electroweak corrections to matching coefficients Isospin effects in non-leptonic decays (e.g. B πk ) and radiative decays (b s(d)γ, b sl + l ) are particularly sensitive to New Physics! Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

26 Electroweak Corrections Penguin and box diagrams with additional γ/z exchange: Electroweak Penguin Operators O 7 10 depends on electromagnetic charge of final state quarks! Electromagnetic operators O γ 7 main contribution to b s(d)γ decays. Semileptonic operators O 9V, O 10A O 9V = ( s L γ µ b L ) ( l γ µ l), O 10A = ( s L γ µ b L ) ( l γ µ γ 5 l) main contribution to b sl + l decays. electroweak corrections to matching coefficients Isospin effects in non-leptonic decays (e.g. B πk ) and radiative decays (b s(d)γ, b sl + l ) are particularly sensitive to New Physics! Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

27 Electroweak Corrections Penguin and box diagrams with additional γ/z exchange: Electroweak Penguin Operators O 7 10 depends on electromagnetic charge of final state quarks! Electromagnetic operators O γ 7 main contribution to b s(d)γ decays. Semileptonic operators O 9V, O 10A main contribution to b sl + l decays. electroweak corrections to matching coefficients Isospin effects in non-leptonic decays (e.g. B πk ) and radiative decays (b s(d)γ, b sl + l ) are particularly sensitive to New Physics! Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

28 Electroweak Corrections Penguin and box diagrams with additional γ/z exchange: Electroweak Penguin Operators O 7 10 depends on electromagnetic charge of final state quarks! Electromagnetic operators O γ 7 main contribution to b s(d)γ decays. Semileptonic operators O 9V, O 10A main contribution to b sl + l decays. electroweak corrections to matching coefficients Isospin effects in non-leptonic decays (e.g. B πk ) and radiative decays (b s(d)γ, b sl + l ) are particularly sensitive to New Physics! Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

29 B- B Mixing Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

30 B- B Mixing B = 2 operators require two W ± exchanges box diagrams: H B=2 eff = G2 F 16π 2 M2 W (V tbv td ) 2 C(x t, µ) ( b L γ µ d L )( b L γ µ d L ) + h.c. only one colour structure due to symmetry of operator analogously for B s- B s mixing Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

31 Summary: Effective Theory for b-quark decays Full theory all modes propagate Parameters: M W,Z, M H, m t, m q, g, g, α s... µ > M W C i (M W ) = C i tree + δ (1) i α s(m W ) 4π +... matching: µ M W Eff. theory low-energy modes propagate. High-energy modes are integrated out. Parameters: m b, m c, G F, α s, C i (µ)... µ < M W ln µ C i(µ) = γ ji (µ) C j (µ) anomalous dimensions Expectation values of operators O i at µ = m b. All dependence on M W absorbed into C i (m b ) resummation of logs Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

32 From b cdū to B 0 D + π In experiment, we cannot see the quark transition directly. Rather, we observe exclusive hadronic transitions, described by hadronic matrix elements, like e.g. D + π Heff b cdū B d 0 = V cbvud G F C i (µ) r i (µ) 2 i r i (µ) = D + π O i B 0 d µ The hadronic matrix elements r i contain QCD (and also QED) dynamics below the scale µ m b. Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

33 "Naive" Factorization of hadronic matrix elements r i = D + J (b c) i B d 0 }{{} π J (d u) i 0 }{{} Quantum fluctuations above µ m b already in Wilson coefficients Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

34 "Naive" Factorization of hadronic matrix elements r i = D + J (b c) i B d 0 }{{} π J (d u) i 0 }{{} form factor decay constant Part of (low-energy) gluon effects encoded in simple/universal had. quantities Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

35 "Naive" Factorization of hadronic matrix elements r i (µ) = D + J (b c) i B d 0 }{{} π J (d u) i 0 }{{} + corrections(µ) form factor decay constant Gluon cross-talk between π and B D QCD corrections Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

36 QCD factorization [Beneke/Buchalla/Neubert/Sachrajda 2000] light quarks in π have large energy (in B rest frame) gluons from the B D transition see small colour-dipole corrections to naive factorization dominated by gluon exchange at short distances δx 1/m b New feature: Light-cone distribution amplitudes φ π (u) Short-distance corrections to naive factorization given as convolution r i (µ) j F (B D) j 1 0 du ( 1 + α ) sc F 4π t ij(u, µ) +... f π φ π (u, µ) φ π (u) : distribution of momentum fraction u of a quark in the pion. t ij (u, µ) : perturbative coefficient function (depends on u) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

37 QCD factorization [Beneke/Buchalla/Neubert/Sachrajda 2000] light quarks in π have large energy (in B rest frame) gluons from the B D transition see small colour-dipole corrections to naive factorization dominated by gluon exchange at short distances δx 1/m b New feature: Light-cone distribution amplitudes φ π (u) Short-distance corrections to naive factorization given as convolution r i (µ) j F (B D) j 1 0 du ( 1 + α ) sc F 4π t ij(u, µ) +... f π φ π (u, µ) φ π (u) : distribution of momentum fraction u of a quark in the pion. t ij (u, µ) : perturbative coefficient function (depends on u) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

38 Light-cone distribution amplitude for the pion ΦΠ u 1.4 (for illustration only) u Exclusive analogue of parton distribution function: PDF: probability density (all Fock states) LCDA: probability amplitude (one Fock state, e.g. q q) Phenomenologically relevant u 1 π 3.3 ± 0.3 [from sum rules, lattice, exp.] Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

39 Complication: Annihilation in B d D + π Second topology for hadronic matrix element possible: Tree (class-i) Annihilation (class-iii) annihilation is power-suppressed by Λ/m b difficult to estimate (final-state interactions?) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

40 Still more complicated: B D 0 π Second topology with spectator quark going into light meson: Tree (class-i) Tree (class-ii) class-ii amplitude does not factorize into simpler objects (colour-transparency argument does not apply) again, it is power-suppressed compared to class-i topology Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

41 Non-factorizable: B 0 D 0 π 0 In this channel, class-i topology is absent: Tree (class-ii) Annihilation (class-iii) The whole decay amplitude is power-suppressed! Naive factorization is not even an approximation! Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

42 B ππ and B πk Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

43 B ππ and B πk Naive factorization: Both final-state mesons are light and energetic. Colour-transparency argument applies for class-i and class-ii topologies. B π(k ) form factors fairly well known (QCD sum rules) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

44 QCDF for B ππ and B πk decays (BBNS 1999) Factorization formula has to be extended: Vertex corrections are treated as in B Dπ Include penguin (and electroweak) operators from H eff. Take into account new (long-distance) penguin diagrams! ( Fig.) Additional perturbative interactions involving spectator in B-meson ( Fig.) Sensitive to the distribution of the spectator momentum ω light-cone distribution amplitude φ B (ω) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

45 Additional diagrams for hard corrections in QCDF (example) additional contributions to the hard coefficient functions t ij (u, µ) r i (µ) hard j F (B π) 1 j (mk 2 ) du 0 ( 1 + α ) s 4π t ij(u, µ) +... f K φ K (u, µ) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

46 Additional diagrams for hard corrections in QCDF (example) additional contributions to the hard coefficient functions t ij (u, µ) r i (µ) hard j F (B π) 1 j (mk 2 ) du 0 ( 1 + α ) s 4π t ij(u, µ) +... f K φ K (u, µ) Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

47 Spectator corrections with hard-collinear gluons in QCDF additive correction to naive factorization ( αs ) r i (µ) = du dv dω spect. 4π h i(u, v, ω, µ) +... f K φ K (u, µ) f π φ π (v, µ) f B φ B (ω, µ) Distribution amplitudes for all three mesons involved! Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

48 New ingredient: LCDA for the B-meson ΦB Ω Ω GeV Phenomenologically relevant: ω 1 B (1.9 ± 0.2) GeV 1 (at µ = m b Λ 1.5 GeV) (from QCD sum rules [Braun/Ivanov/Korchemsky]) (from HQET parameters [Lee/Neubert]) Large logarithms ln m b can be resummed using SCET Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

49 Complications for QCDF in B ππ, πk etc. Annihilation topologies are numerically important. BBNS use conservative model estimates. Some power-corrections are numerically enhanced by chiral factor µ π = m2 π f π 2f π m q Many decay topologies interfere with each other. Many hadronic parameters to vary. Hadronic uncertainties sometimes quite large. Th. Feldmann (Uni Siegen) Introduction to OPE March / 40

50 Summary [from C. Berger s homepage] When looking for new physics, do not forget about the complexity of the old physics! Th. Feldmann (Uni Siegen) Introduction to OPE March / 40