SI-HEP-2007-16 January 27, 2009 9 0 Complete Michel Parameter Analysis 0 2 of the Inclusive Semileptonic b → c Transition n a J 7 2 Benjamin Dassinger, Robert Feger, Thomas Mannel ] Theoretische Physik 1, Fachbereich Physik, Universit¨at Siegen h p D-57068 Siegen, Germany - p e h [ 2 v 1 Abstract 6 5 3 Weperformacomplete “Michel parameter”analysis ofallpossiblehelicity structures 3. which can appear in the process B Xcℓν¯ℓ. We take into account the full set of → 0 operators parametrizing the effective Hamiltonian and includethe complete one-loop 8 QCD corrections as well as the non-perturbative contributions. The moments of the 0 : leptonicenergyaswellasthecombinedmomentsofthehadronicenergyandhadronic v invariant mass are calculated including the non-standard contributions. i X r a 1 Introduction Theexperimentalandtheoreticaldevelopments inheavyflavourphysics allowustoperform a high precision test of the flavour sector. In particular, the enormous amount of data for semileptonic B decays in combination with very reliable theoretical methods has opened the road for a precision determination of the CKM matrix elements V and V , which are cb ub known at a relative precision of roughly 2% and 10% [1]. Asidefromtesting andextracting itsparametersasprecisely aspossible, asecondmajor goal of heavy flavour physics is to look for possible effects beyond the standard model. It is genenerally believed that flavour changing neutral currents are a good place to search for effects of new physics, since these decays are usually loop-induced and hence sensitive to virtual effects from high-mass states. Thus one expects here possibly an effect which is sizable compared to the standard-model contribution. Semileptonic processes are tree level processes in the standard model and thus the relative effects from new-physics contributions are likely to be small. However, a possible right-handed admixture to thehadronic current is completely absent inthe standard model and hence such an effect would be a clear signal for physics beyond the standard model. In a recent publication [2] we considered a “Michel parameter analysis” [3] of semilep- tonic B decays, where we considered mainly a possible right-handed contribution to the hadronic b c current. In the present paper we complete the analysis of [2] by extending → the analysis to all possible two-quark-two-lepton operators. There is an extensive literature on a possible non-standard model contributions to semileptonic B decays [4, 5, 6, 7, 8]. However, the analysis presented here is different in two respects. First of all, our analysis is completely model-independent; however, we neglect the lepton masses and hence our analysis would need to be extended straightfor- wardly to include e.g. a discussion of a charged Higgs contribution as in [4, 5, 8]. Secondly, we consider different observables (i.e. the moments of spectra) which have become avail- able only recently through the precise data of the B factories; in this way a much better sensitivity to a non-standard contribution is expected. In the next section we perform an effective-field-theory analysis of possible new-physics contribution, which is kept completely model independent. It turns out that only very few operators contribute to the semileptonic b c transition. Compared to the usual → Michel-parameter analysis, well known from muon decay, this effective theory analysis also yields order-of-magnitude estimates of the various contributions. Basedonthese effective interactions werecomputethespectra ofinclusive semi-leptonic b c transitons including the new interactons. We make use of the standard heavy → quark expansion (HQE) and include QCD radiative corrections as well as nonperturbative contributions. In section 3 we shall perform the HQE including the new-physics operators. In subsec- tion 3.1 we compute the QCD radiative corrections for the various helicity combinations of the hadronic current. We adopt the kinetic scheme as it has been used for the calcula- tion of semileptonic moments in [9] and perform the complete one-loop calculation for the new-physics terms. We note that the standard-model calculations have been performed already to order α2 in [10]. s 1 Subsection 3.3 we calculate the nonperturbative contributions of the new physics op- erators to order 1/m2. Finally, in section 4 we quote our results for the various moments b which are frequently used in the analysis of semileptonic decays and conclude. → 2 Effective-Field-Theory Analysis of b cℓν¯ ℓ It is well known that any contribution to the effective Lagrangian of some yet unknown physics at a high scale Λ can be written as contributions of operators with dimensions larger than four. These operators are SU(3) SU(2) U(1) invariant and suppressed × × by an appropriate power of 1/Λ. We note that for such an analysis we have to make an assumption of the yet not established Higgs Sector: We shall stick with our analysis to the single Higgs doublet case; an extension to a type-II two-Higgs doublet as e.g. needed for supersymmetry is straightforward. The lowest dimension relevant for our analysis is six; the list of relevant operators has been given in [11] and we shall use the notations of our previous paper [2]. The quark and lepton fields are grouped into u c t Q = L , L , L for the left handed quarks (1) L d s b L L L (cid:18) (cid:19) (cid:18) (cid:19) (cid:18) (cid:19) u c t Q = R , R , R for the right handed quarks (2) R d s b R R R (cid:18) (cid:19) (cid:18) (cid:19) (cid:18) (cid:19) ν ν ν L = e,L , µ,L , τ,L for the left handed leptons (3) L e µ τ L L L (cid:18) (cid:19) (cid:18) (cid:19) (cid:18) (cid:19) ν ν ν L = e,R , µ,R , τ,R for the right handed leptons (4) R e µ τ R R R (cid:18) (cid:19) (cid:18) (cid:19) (cid:18) (cid:19) where Q and L are doublets under SU(2) and Q and L are doublets under an L L L R R (explicitely broken) SU(2) . Note that we also introduced a right-handed neutrino in R order to complete the right-handed lepton doubletts. The Higgs field and its charge conjugate are written as a 2 2 matrix × 1 φ +iχ √2φ H = 0 0 + (5) √2 (cid:18) √2φ− φ0 −iχ0(cid:19) transforming under SU(2) SU(2) . The potential of the Higgs fields leads to a vacuum L R × expectation value (vev) for the field φ . 0 The dimension-6 operators fall into two classes, the two-quark operators with gauge and Higgs fields and the two-quark-two-lepton operators. In our previous analysis [2] we considered only the first class, and the first step towards a full analysis is to also take into account the second class. 2 The list of two-quark two-lepton operators with SU(2) SU(2) consists of1 L R × (i) = (Q¯ Γ Q )(L Γ L ) (6) OLL,LL L i L L i L (i) = (Q¯ τaΓ Q )(L τaΓ L ) (7) PLL,LL L i L L i L (i) ¯ = (Q Γ Q )(L Γ L ) (8) OLL,RR L i L R i R (i) ¯ = (Q Γ Q )(L Γ L ) (9) ORR,LL R i R L i L (i) = (Q¯ Γ Q )(L Γ L ) (10) ORR,RR R i R R i R (i) = (Q¯ τaΓ Q )(L τaΓ L ) (11) PRR,RR R i R R i R while the operators with explicitely boken SU(2) read R (i) = (Q¯ Γ Q )(L Γ τ3L ) (12) RLL,RR L i L R i R (i) = (Q¯ Γ τ3Q )(L Γ L ) (13) RRR,LL R i R L i L (i) = (Q¯ Γ Q )(L Γ τ3L ) (14) RRR,RR R i R R i R (i) = (Q¯ τaΓ Q )(L τaτ3Γ L ) (15) SRR,RR R i R R i R (i) = (Q¯ τaτ3Γ Q )(L τaτ3Γ L ) (16) TRR,RR R i R R i R Here we have defined Γ Γ = 1 1, γ γµ, γ γ γ γµ, σ σµν (17) i i µ µ 5 5 µν ⊗ ⊗ ⊗ ⊗ ⊗ Note that these operators are not all independent. Furthermore, note that operators with the helicity combinations such as (LR)(LR) cannot appear at the level of dimension six operators, since additional Higgs fields are required for the helicity flip. We shall assume that the right handed neutrino acquires a large majorana mass in which case it can be integrated out at some high scale, which we assume to lie well above Λ. In this case SU(2) is “maximally broken” for the leptons, which means that the R possible operators always have a projection P = (1 τ3)/2 and thus only right handed − − interactions involving the right handed charged leptons remain. For the case at hand we are interested in the charged current interactions containing a b ctransition. Sinceweeliminatedtheright-handedneutrinoandhelicitiesareconserved → for both currents we end up with the conclusion that the charged leptonic current has to be left handed. Thus we have only the operators O = (¯b γ c )(ν¯ γµℓ ) O = (¯b γ c )(ν¯ γµℓ ) (18) 1 L µ L ℓ,L L 2 R µ R ℓ,L L whereℓ = e,µorτ,sinceanyhelicitychangingcombinationhastooriginatefromdimension- 8 operators yielding an additional suppression of a factor v2/Λ2 relative to the dimension-6 contributions. 1In order to have a streamlined notation we suppress all flavour indices in the following 3 However, as has been discussed in our previous paper, helicity violating combinations suchas(LR)(LL)operatorscanappearfromthetwo-quarkoperatorswithgaugeandHiggs fields. These operators induce anomalous gauge-boson couplings which are suppressed by a factor v2/Λ2. They originate from two-quark operators, which are (at the scale of the weak bosons) (1) ¯ O = Q L/ Q (19) LL L L O(2) = Q¯ L/ Q (20) LL L 3 L with Lµ = H(iDµH)† +(iDµH)H† (21) Lµ = Hτ (iDµH)† +(iDµH)τ H† (22) 3 3 3 The terms proportional to τ have once again been included to break the custodial sym- 3 metry explicitely. In the same spirit we define RR-operators O(1) = Q¯ R/ Q (23) RR R R O(2) = Q¯ τ ,R/ Q (24) RR R { 3 } R O(3) = iQ¯ [τ ,R/] Q (25) RR R 3 R O(4) = Q¯ τ R/τ Q (26) RR R 3 3 R with Rµ = H†(iDµH)+(iDµH)†H (27) Using an odd number of Higgs fields we can construct invariant LR operators. For our analysis the relevant operators are O(1) = Q¯ (σ Bµν)H Q +h.c. (28) LR L µν R O(2) = Q¯ (σ Wµν)HQ +h.c. (29) LR L µν R O(3) = Q¯ (iD H)iDµQ +h.c. (30) LR L µ R After spontaneous symmetry breaking the LL and RR operators contain anomalous quark-boson couplings of the order of magnitude v2/Λ2. For the LR operators the field strenghts of the weak bosons appear, inducing an additional factor of a quark momentum p, and hence the order of magnitude is pv/Λ2 m v/Λ2. q ∼ At the scale of the bottom quark theses anomalous coupling terms have the same power counting as the two-quark two-lepton operators: Integrating out the weak bosons, their propagatortogether with thegaugecouplings reduce toa pointlike interaction proportional to g2/M2 = 1/v2. Combining this with the order-of-magnitude of the anomalous coupling W of LL and RR v2/Λ2 we find that at the scale of the bottom mass these contributions 4 scale in the same way as the two-quark-two-lepton operators directly induced at the high scale Λ. For the case of LR the additional momentum p of the quark is of the order of its mass, and hence is as well of the order v2/Λ2, possibly further suppressed by a small quark Yukawa coupling. This conclusion may be altered in a two-higgs doublet model in the case of large tanβ, i.e. of a large ratio of Higgs vacuum expectation values. After integrating out the heavy degrees of freedom tanβ will play the role of a coupling constant which then may be enhanced by a large value. In e.g. [8] such a scenario has been considered, where a sizable value of tanβ overcomes the supression of the factor (m m )/m2 in the amplitude; in ℓ b H+ this case also scalar contributions to the leptonic current have to be taken into account. To this end, the parametrization introduced in [2] remains valid also in the general case, if only dimension-6 operators are included. Thus the effective Hamiltonian reads 4G V = F cbJ Jµ, (31) Heff √2 q,µ l where Jµ = e¯γµP ν is the usual leptonic current and J is the generalized hadronic l − e h,µ b c current which is given by → J = c c¯γ P b+c c¯γ P b+g c¯i←D→P b+g c¯i←D→P b (32) h,µ L µ − R µ + L µ − R µ + +d i∂ν(c¯iσ P b)+d i∂ν(c¯iσ P b), L µν − R µν + where P denotes the projector on positive/negative chirality and D is the QCD covariant ± µ derivative. Notethatthetermproportionaltoc contains thestandard-modelcontribution L as well as a possible new-physics contribution and c may now also contain a contribu- R tion from a two-quark-two-lepton operator induced at the high scale Λ. The gauge part ig Aaλ /2of the QCD covariant derivative D gives rise to a new quark-quark-gluon-boson 3 µ a µ vertex. 3 Operator Product Expansion The operator product expansion (OPE) for inclusive decays has become textbook material [12]. For the case of inclusive semileptonic decays the OPE is formulated for the T-product of the two hadronic currents T = d4xe−ix(mbv−q) B(p)¯b (x)Γ c(x)c¯(0)Γ†b (0) B(p) (33) µν h | v µ ν v | i Z where Γ is the combination of Dirac matrices and derivatives given in (32), v = p/M B is the four velocity of the decaying B meson and q is the momentum transferred to the leptons. The quantity T is expanded in inverse powers of the scale of the order m , where µν b m is the the heavy quark mass. Technically this procedure is an OPE for the product of b the two currents. 5 ν¯e ν¯e ν¯e W− W− W− e e e c c c g b (cid:1) b (cid:2) g b (cid:3)g Figure 1: Real Corrections ν¯e ν¯e ν¯e W− W− W− e e e c c c g g g (cid:1) (cid:2) (cid:3) b b b Figure 2: Virtual Corrections The standard-model calculation has been performed at tree level up to order 1/m4 and b it turns out that the non-perturbative corrections are small. The radiative corrections have been computed to order α , β α2 and recently also to order α2 for the leading (i.e. the s 0 s s parton model) term [10] and to order α for the term of order 1/m2 involving µ2. s b π In the following we shall consider the perturbative and non-perturbative contributions to the OPE, performed with the modified current (32). We shall compute the complete one-loop contributions as well as the leading non-perturbative corrections proportional to µ2 and µ2. π G 3.1 QCD Corrections and Renormalization Group Analysis The calculation ofthe QCD radiative corrections hasbeen performed in[13]andthe results in the kinetic scheme have been given in [9] for the semileptonic moments in the standard model. In order to perform an analysis of possible non-standard contributions we have to calculate the QCD radiative corrections for the current (32) to order α . Thus we have to s evaluate the Feynman diagrams shown in fig. 1 and 2 for the real and virtual corrections respectively. Note that the scalar current (i.e. the terms proportional to g ) induces L/R new vertices shown in the Feynman diagrams at right. The real and virtual corrections are individually IR-divergent. We regulated the IR-divergence by introducing a gluon mass which drops out upon summation of the real and virtual correction being IR-convergent. In the calculation of the virtual corrections the wave function renormalizations of the b 6 and c quark field also have to be included. The total amplitude consists of the sum of the standard-model contribution and the one from the new-physics operators. Since the new-physics piece is of order 1/Λ2, we shall include only the interference term of the standard model with this contribution. The square of the new-physics term is already of order 1/Λ4 and has to be neglected, since we compute only up to this order. Thus we compute 1 dΓ = clν c SM b clν b ∗ + clν b clν c SM b ∗ dφ (34) 2m h | LHeff | ih |Heff| i h |Heff| ih | LHeff | i PS b (cid:0) (cid:1) where dφ is the corresponding phase space element and PS 4G V SM = F cb c¯γ P b e¯γµP ν Heff √2 µ − − e (cid:0) (cid:1)(cid:0) (cid:1) is the standard-model effective Hamiltonian, which has the same helicity structure as the new-physics contribution proportional to c . L The relevant Feynman rules for the new-physics operators at tree level can be read off from (32); note that the terms involving g and g yield a boson-gluon-quark-antiquark L R vertex in order to maintain QCD gauge invariance. It is well known that the left- and right-handed currents do not have anomalous dimen- sions and hence the parts of (32) with c and c are not renormalized. However, the scalar L R and tensor contributions have anomalous dimensions and hence we need to normalize these operators at some scale and run them down to the scale of the bottom quark. To this end, we have to calculate the anomalous-dimension matrix of these currents to set up the renormalization group equation. It can be obtained from the requirement that the physical matrix elements must not depend on the renormalization scale µ: d 0 = cℓν b (35) e eff dlnµh |H | i Inserting the OPE for the Hamiltonian we get: 4G V cℓν b = F cb cℓν [c (c¯γ P b)(e¯γµP v )+c (c¯γ P b)(e¯γµP v )] b e eff e L µ − − e R µ + − e h |H | i √2 ·h | | i 4G V F cb ~ ~ + C cℓν b , (36) e √2 ·h |O| i with g (c¯i←D→P b)(e¯γµP v ) L µ − − e g (c¯i←D→P b)(e¯γµP v )  R  µ + − e  d (i∂ν(c¯iσ P b))(e¯γµP v ) L µν − − e     d  (i∂ν(c¯iσ P b))(e¯γµP v ) C~ =  R ~ =  µν + − e . (37) cmLb O  (mbc¯γµP−b)(e¯γµP−ve)      cmb  (m c¯γ P b)(e¯γµP v )   R   b µ + − e  cmc  (m c¯γ P b)(e¯γµP v )   L   c µ − − e      cmc  (m c¯γ P b)(e¯γµP v )   R   c µ + − e      7 ~ where the operators are of dimension seven. O In the following we consider the renormalization group mixing of these dimension-four operators. The calculation of the one-loop anomalous dimension is standard. We define the anomalous dimension matrix γ by: dC~ = γT(µ)C~ (38) dlnµ and compute γ from the divergencies of the renormalization constants in the usual way. We obtain 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0   1 0 1 0 0 0 0 0 2α (µ) 0 1 0 1 0 0 0 0 γT(µ) = s   (39) 3π 0 3 0 0 3 0 0 0   3 0 0 0 0 3 0 0   3 0 0 0 0 0 3 0   0 3 0 0 0 0 0 3     The renormalization group equation for the Wilson coefficient is ∂ ∂ +β(α ) C~ = γT α (µ) C~. (40) s s ∂lnµ ∂α (cid:18) s(cid:19) (cid:0) (cid:1) We seek a solution of this equation with the initial conditions cmb (Λ) = 0 = cmc (Λ), (41) L/R L/R since the matching of the left- and right-handed currents is performed by fixing the coef- ficients c and c and all additional contributions are only due to renormalization group L R running. Inserting the one-loop expressions we obtain c (µ) = c (Λ) L/R L/R g (µ) = g (Λ) L/R L/R 4 αs(Λ) 3β0 d (µ) = g (Λ)+d (Λ) g (Λ) L/R L/R L/R L/R α (µ) − (cid:18) s (cid:19) (42) (cid:0) (cid:1)4 cmb (µ) = g (Λ) αs(Λ) β0 1 L/R R/L α (µ) − (cid:18) s (cid:19) ! 4 cmc (µ) = g (Λ) αs(Λ) β0 1 L/R L/R α (µ) − (cid:18) s (cid:19) ! One may reexpand (42) using the one-loop expression for the strong coupling constant and obtain the logarithmic terms of the one-loop calculation. However, the straight-forward 8 one-loop calculation also yields finite terms, which depend on the choice of the renormal- ization scale µ. It is well known that in order to fix this dependence on the renormalization scale, one would need to include the running at two loops, which, however, goes beyond the scope of the present paper. Rather we shall fix this scale to be µ = m , which is the b relevant scale of the decay process, assuming that the full NLO calculation would fix a scale of this order. The advantage of this procedure is that the kinematic effects, which lead to a distortion of the spectra and thus have an impact on the moments, are given by these finite terms. We expect that a full NLO calculation will lead to very similar results. 3.2 Mass Scheme The calculation of the process is usually set up with pole masses of the particles, which is known not to be a well defined mass. The problems manifest themselves by abnormally large radiative corrections when the pole mass scheme is used. It has been discussed extensively in the literature that an appropriately defined short-distance mass is better suited for the OPE calculation of an inclusive semileptonic rate. In the present analysis we will use the kinetic mass scheme, where the mass is defined by a non-relativistic sum rule for the kinetic energy [14]. At one-loop level the kinetic mass is related to the pole mass by 4α 4 µ µ2 mkin(µ ) = mpole 1 s f + f (43) q f q − 3 π 3m 2m2 " (cid:18) b b(cid:19)# where µ is a factorizationscale for removing contributions below from the mass-definition. f The factorization scale is set to 1GeV since this is the typical energy release in the process. This low renormalization scale is in fact the reason why the MS scheme is inappropriate. The ratio ρ = m2/m2 is rather stable under the choice of schemes (provided that the c b same scheme is chosen for both m and m ) and thus the choice of the mass scheme enters b c only through the m5 dependence of the rates. It is well known from the calculation in the b standard model that the (α ) corrections from the relation of the kinetic mass with the s O pole mass mpole 5 α q s 1+2.0899 . (44) mkin(1GeV) ≈ π (cid:18) q (cid:19) compensate to a large extent the radiative corrections to the rates computed in the pole scheme, leaving only small QCD radiative corrections. It turns out that this also is the case in our calculation including the anomalous couplings. 3.3 Non-perturbative Corrections The nonpertubative corrections at tree level, including the modified current (32), have been studied in [15], however, these results have not yet been published, and hence we shall quote these results in the following. 9