$B_s^0-\bar{B}_s^0$ mixing within minimal flavor-violating two-Higgs-doublet models

In the"Higgs basis"for a generic 2HDM, only one scalar doublet gets a nonzero vacuum expectation value and, under the criterion of minimal flavor violation, the other one is fixed to be either color-singlet or color-octet, which are named as the type-III and type-C models, respectively. In this paper, the charged-Higgs effects of these two models on $B_s^0-\bar{B}_s^0$ mixing are studied. Firstly, we perform a complete one-loop computation of the electro-weak corrections to the amplitudes of $B_s^0-\bar{B}_s^0$ mixing. Together with the up-to-date experimental measurements, a detailed phenomenological analysis is then performed in the cases of both real and complex Yukawa couplings of charged scalars to quarks. The spaces of model parameters allowed by the current experimental data on $B_s^0-\bar{B}_s^0$ mixing are obtained and the differences between type-III and type-C models are investigated, which is helpful to distinguish between these two models.

B 0 s −B 0 s mixing are studied. Firstly, we perform a complete one-loop computation of the electro-weak corrections to the amplitudes of B 0 s −B 0 s mixing. Together with the up-todate experimental measurements, a detailed phenomenological analysis is then performed in the cases of both real and complex Yukawa couplings of charged scalars to quarks. The spaces of model parameters allowed by the current experimental data on B 0 s −B 0 s mixing are obtained and the differences between type-III and type-C models are investigated, which is helpful to distinguish between these two models.

Introduction
Thanks to the successful running of the Large Hadron Collider (LHC), particle physics has entered a new era, which is featured by the discovery of a new boson with a mass close to 125 GeV [1,2]. Its measured properties are so far in good agreement with those of the Standard Model (SM) Higgs [3][4][5], suggesting that the electro-weak symmetry breaking (EWSB) is probably realized via the Higgs mechanism implemented through a single scalar doublet. It should be noted, however, that the EWSB is not necessarily induced by just a single scalar.
Interestingly, many new physics (NP) scenarios are equipped with an extended scalar sector.
The search for additional scalars is one of the important programs of the LHC experiments.
One of the extensions of SM scalar sector is the so-called two-Higgs-doublet model (2HDM) [6], in which a second scalar doublet is added to the SM field content. To avoid the experimental constraints on flavor-changing neutral-current (FCNC) transitions, which are forbidden at tree level in the SM due to the GIM mechanism [7], two different hypotheses, natural flavor conservation (NFC) [8] and minimal flavor violation (MFV) [9], have been proposed 1 . In the NFC hypothesis, depending on the Z 2 charge assignments on the scalar doublets and fermions, there exist four types of 2HDM (type-I, II, X and Y) [11]. In the MFV hypothesis, to control the flavor-violating interactions, all the scalar Yukawa couplings are assumed to be composed of the SM ones Y U and Y D . In the "Higgs basis" [12], in which only one doublet gets a nonzero vacuum expectation value (VEV) and behaves as the SM one, the allowed SU (3) C ⊗ SU (2) L ⊗ U (1) Y representation of the second scalar doublet that couples to quarks via Yukawa interactions is fixed to be either (1, 2) 1/2 or (8, 2) 1/2 [13], which implies that the second scalar doublet can be either color-singlet or color-octet. For convenience, they are referred to as type-III and type-C models [14], respectively. Examples of the former include the aligned 2HDM (A2HDM) [15] and the four types of 2HDM reviewed in Ref. [11]. The scalar spectrum of the latter contains, besides a CP-even and color-singlet Higgs boson (the usual SM one), three color-octet particles, one CP-even, one CP-odd and one electrically charged [13].
Although the scalar-mediated flavor-violating interactions are protected by the MFV hypothesis, the type-III and type-C models still present very interesting phenomena in some 1 The NFC and MFV hypotheses are not the only alternatives to avoid constraints from FCNCs; models with controlled FCNCs have also been addressed in the literature and shown to be compatible with the data [10]. low-energy processes, especially due to the presence of a charged Higgs boson [13,14,16]. In this paper, we shall study the B 0 s −B 0 s mixing within these two models and pursue possible differences between their effects. Since the charged Higgs contributes to the process at the same order as does the W boson in the SM, the NP effects might be significant.
It is known that the B 0 s −B 0 s mixing, which is governed by a Schrödinger equation and induced by the b → s transition, plays an important role in accurate tests of the SM and indirect searches for NP. In terms of the off-diagonal elements of the mass and decay matrices, M s 12 and Γ s 12 , the mass and width differences between the two mass eigenstates |B H and |B L are defined, respectively, by where φ s ≡ arg(−M s 12 /Γ s 12 ) is the CP-violating phase. Such two observables have been measured precisely, with the averaged values given, respectively, by [17] ∆M exp.
For φ ccs s , the SM prediction ∼ −0.036 [18] agrees with the experimental data −0.015±0.035 [17] within 1σ error bar. For a s sl , on the other hand, the measurement (−0.75 ± 0.41)% [17] is significantly different from the SM estimation ∼ O(10 −5 ) [18], even through they are in agreement with each other at 1.5σ level due to the large experimental error bars. Under the constraints of the above four observables, the allowed NP spaces could possibly be strictly reduced. So, in this paper, we shall evaluate the effects of charged Higgs in type-III and type-C 2HDMs on B 0 s −B 0 s mixing, and pursue possible differences between these two models. 2 The phase φ ccs s appears in tree-dominated b → ccs B s decays, such as B s → J/ψφ, and is generally different from φ s unless the terms proportional to V cb V * cs V ub V * us and (V ub V * us ) 2 in Γ s 12 are neglected [19].  Firstly, for convenience and consistence, we shall give a brief review of the 2HDMs under the MFV hypothesis. In the "Higgs basis", the Yukawa interactions of the two Higgs fields Φ 1 and Φ 2 with quarks are given by [13,14] − where q 0 L , u 0 R and d 0 R are the quark fields in the interaction basis; T where the ellipses denote trivial terms involving higher power of After applying the SM unitary transformations to rotate the fermionic fields from the interaction to the mass-eigenstate basis, one can finally obtain the Yukawa interactions of charged Higgs bosons with quarks in the mass-eigenstate basis [14,16]  where g is the SU (2) L coupling constant, i , j the fermionic generation indices, and m u,d the quark masses; V denotes the Cabibbo-Kobayashi-Maskawa (CKM) matrix [20,21], and P R,L = 1±γ 5 2 are the right-and left-handed chirality projectors. The couplings A i u,d are generally familydependent and read where v = Φ 0 1 = 174 GeV. Since only the couplings of charged Higgs bosons to the top quark are involved for B 0 s −B 0 s mixing, we shall drop the family index in A i u,d from now on. Following the notation used in Ref. [14], we shall denote the model with the second scalar doublet being color-singlet and the one with the second scalar doublet color-octet as the type-III and the type-C model, respectively, both of which satisfy the principle of MFV. Their explicit contributions to the B 0 s −B 0 s mixing will be presented in the next subsection. replaced by the charged-Higgs H ± one(s). After calculating these one-loop box diagrams and applying the standard procedure of matching [22,23], one can obtain the 2HDM corrections to the B 0 s −B 0 s mixing. Together with the SM contribution, the resulting effective weak Hamiltonian responsible for B 0 s −B 0 s mixing can be written as

B
where G F is the Fermi coupling constant, and C i (µ) the scale-dependent Wilson coefficients of the four-quark operators O i , which are defined, respectively, as with α, β being the color indices and σ µν = 1 2 [γ µ , γ ν ]. In addition, the hadronic matrix elements of the four-quark operators can be parameterized as [24] where f Bs is the B s -meson decay constant, and B i (µ b ) are the non-perturbative bag parameters calculated on the lattice at a characteristic scale The Wilson coefficients C i (µ) in Eq. (8) consist of both the SM and 2HDM contributions and the results at the initial scale µ W ∼ O(m W , m t , m H ± ) can be written as Within the SM, the explicit expressions of the Wilson coefficients are computed from the box diagram shown in Fig. 1(a), accompanied by perturbative QCD corrections up to the desired order, details of which could be found, for example, in Refs. [22,23,26]. Including the next-toleading order (NLO) QCD corrections, the SM contribution is given by [22,23,26] , and the leading order (LO) coefficient S 0 (x t ) is the known Inami-Lim function [27]. Generally, the SM also contributes to C SRR and C T RR and, in the absence of QCD corrections, we get with . It is obvious that both C SRR SM (µ W ) and C T RR SM (µ W ) are suppressed by the factor x b and are, therefore, usually neglected in the literature [22].
The charged-Higgs contributions to the Wilson coefficients are computed from the last three box diagrams shown in Fig. 1, and depend where x h = m 2 H ± /m 2 W , and the explicit expressions for f i (x t , x h ) are collected in the appendix. For the color-octet charged-Higgs contributions (type-C model), on the other hand, we get It is noted that the Wilson coefficient Refs. [23,29,31] 3 . For C SRR and C T RR , on the other hand, in order to get a gauge-independent result, the external momenta of the heavy quarks inside the mesons should be taken into account, and the heavy-quark masses should be kept up to the second orde; our results for these two coefficients differ from the ones presented in Refs. [23,29,31].
The QCD renormalization group (RG) evolution of these Wilson coefficients from the matching scale µ W down to the lower scale µ b has been calculated in Ref. [24]. One can then obtain the corresponding Wilson coefficients at the scale µ b through [24] where the explicit expressions of the evolution factors η could be found in Ref. [24].
Equipped with the above information, the off-diagonal mass matrix element M s 12 is given as where A V LL , A SRR and A T RR denote the contributions induced by the three four-quark operators defined by Eq. (9), respectively. Within the SM, the off-diagonal decay matrix element Γ s 12 can be written as [28,33] Γ s 12 (SM ) = − λ 2 t Γ cc 12 + 2 λ t λ u (Γ cc 12 − Γ uc 12 ) + λ 2 u (Γ cc 12 − 2Γ uc 12 + Γ uu 12 ) , with the CKM factors λ i = V ib V * is for i = u, c, t. The explicit expressions for Γ cc,uu,uc 12 could be found in Refs. [28,33]. It should be noted that Γ s 12 is dominated by the CKM-favored tree-level b → ccs transition within the SM, and the NP effects are generally negligible [28]. Hence Γ s 12 = Γ s 12 (SM ) holds as a good approximation, which will be assumed throughout this paper.

Numerical results and discussions
We now proceed to present our numerical results and discussions. Values of the relevant input parameters used throughout this paper are summarized in Table 1. Our SM predictions for the   observables of B 0 s −B 0 s mixing are given in the third row of Table 2, in which the experimental data averaged by the HFAG [17] are also listed in the second row for comparison. As mentioned already in the introduction section, there is no significant deviation between the SM predictions and the experimental data for the observables at the current level of precision, even through a bit disagreement appears for a s sl . Therefore, these observables are expected to put strong constraints on the parameter spaces of 2HDMs with MFV.
From the analytic expressions of the charged-Higgs contributions to the B 0 s −B 0 s mixing calculated in section 2.2, one can find that the model parameters relevant to our study include the two Yukawa coupling parameters A u and A d , as well as the charged-Higgs mass m H ± . For the case of complex couplings, one could equivalently choose |A u | and A d A * u = |A d A * u |e −iθ as the independent parameters, with θ being the relative phase between the two Yukawa coupling parameters. For the parameter |A u |, as detailed in Ref. [14], an upper bound can be obtained from the Z → bb decay. The parameter A d is, however, much less constrained phenomenologically [14,16]. Concerning the charged-Higgs mass, we shall use the LEP lower bound m ± H > 78.6 GeV (95% CL) [35], which is obtained under the assumption that H ± decays dom-inantly into fermions and does not refer to any specific Yukawa structure. Direct searches for charged-Higgs bosons are also performed by the Tevatron [36], ATLAS [37] and CMS [38] collaborations. However, most of the limits on m ± H depend strongly on the assumed Yukawa structure. In this paper, we shall generate randomly numerical points for the model parameters in the ranges whereas no severe constraints for |A d A * u | (or |A d |) and θ.  Table 1, we get in the case of type-III model, and in the case of type-C model, respectively. The number 3.73 in Eqs. (26) and (29) is the SM contribution, while the SM contributions to A SRR and A T RR are suppressed by the factor x b , making them numerically smaller by about three orders than A V LL SM and hence negligible. From the above numerical results, we make the following observations: (i) In both the type-III and type-C models, the charged-Higgs contributions to A V LL (Eqs. (26) and (29)) depend only on the Yukawa coupling parameter A u via |A u |, and hence are always constructive to the SM one. For a value |A u | ∼ O(1), the type-III contribution could be comparable with the SM one, while the type-C model provides a relatively smaller correction.
(ii) Comparing Eqs. (27)-(28) with (26) (for the type-III model) and Eqs. (30)-(31) with (29) (for the type-C model), one can see that the NP contributions to A SRR and A T RR are much smaller than to A V LL , especially when |A d | ∼ |A u |. This is because the Wilson coefficients C SRR (µ W ) and C T RR (µ W ) are always suppressed by the factor x b with respect to C V LL (µ W ), both within the SM and in the 2HDMs with MFV.
(iii) In the case with large complex values of A d A * u , however, the charged-Higgs contributions to A SRR and A T RR could provide a large imaginary part to the off-diagonal mass matrix element M s 12 , which may result in a significant correction to the CP-violating observables, such as φ s and φ ccs s .
(iv) Different from the type-C model, the type-III contribution to A T RR is induced only by the RG evolution effect, and is numerically much smaller. There are, however, cancelations between the charged-Higgs contributions to A SRR and A T RR in the type-C model.
It is therefore expected that the current experimental data on B 0 s −B 0 s mixing could put some constraints on the model parameters and be used to distinguish between these two models.
To get the explicitly allowed parameter spaces, we perform the analysis with the following procedure: we scan the parameter spaces within the ranges specified by Eq. (25), with the value of m H ± fixed at 100, 250 and 500 GeV, respectively. At each point in the parameter spaces, we evaluate the theoretical prediction for an observable, together with the corresponding theoretical uncertainty induced by the input parameters listed in Table 1. The theoretical range for an observable at each point is obtained by varying each input parameter within its respective allowed range and then adding the individual uncertainty in quadrature. If the obtained theoretical range has overlap with the 2σ range of the experimental data, the point is regarded as allowed. In addition, we consider two different cases: real and complex couplings with respect to A u and A d . Under the combined constraints from ∆M s , φ ccs s and a s sl , the allowed parameter spaces of the type-III and type-C models are shown in Figs. 2 (for the case of real couplings) and 3 (for the case of complex couplings), respectively.
For the case of real couplings, it can be seen from Fig. 2  (i) In the type-III model, as shown in Fig. 2(a), the module of Yukawa coupling parameter A u is severely constrained by the good agreement between the SM prediction and the experimental data for ∆M s ; for instance |A u | < 1 is obtained with m H ± = 500 GeV.
There are, however, almost no constraints on the coupling A d , because the contribution involving it is negligible with respect to the one involving only A u .
(ii) In the type-C model, because the charged-Higgs contribution to A V LL is relatively small and large cancelation effects exist between the terms involving A d and A u , the allowed values of A d and A u could be large simultaneously, with either the same or the opposite signs, as shown by the four "legs" in Fig. 2(b).
(iii) Besides the "legs" in Fig. 2(b), the difference between the two models is also featured by the different shapes of the allowed parameter spaces. The current data on B 0 s −B 0 s mixing generally puts a stronger constraint on the type-III model; for instance, with the assumption |A d | ∼ |A u | and choosing m H ± = 500 GeV, the upper bound |A u | ∼ 1.5 obtained in type-C model is obviously looser than the one |A u | ∼ 1 in type-III model.
For the case of complex couplings, one more model parameter θ is introduced. From Fig. 3, it is found that:  This is due to the fact that large cancelation effects appear among the charged-Higgs contributions when θ ∼ ±90 • , which can also be seen from Eqs. (27) and (28). Moreover, as shown in Fig. 3(a), an approximately linear relationship is observed between |A * u A d | and |A u | when |A u | 0.5.
(ii) As shown in Figs. 3(b) and 3(d), similar observations could also be made in the type-C model, except for the fact that the constraints on the model parameters are now much looser. In addition, the cancelation effects among the charged-Higgs contributions occur around θ ∼ 0 • and ±180 • , which is different from that observed in the type-III model.
From the above discussions, we conclude that the type-III and type-C models exhibit some significantly different behaviors under the experimental constraints from B 0 s −B 0 s mixing. However, due to the large theoretical and experimental uncertainties, the differences in the small |A u | and |A d | ranges are hardly to be distinguished from each other. The future refined measurement and precise theoretical evaluation for B 0 s −B 0 s mixing might show a much clearer phenomenological picture for the type-III and type-C models.
As a final comment, it should be noted that the same analysis could also be applied to the

Conclusion
In this paper, we have calculated the one-loop electro-weak corrections to the B 0 s −B 0 s mixing within the type-III and type-C 2HDMs with MFV, in which the second scalar doublet is fixed to be color-singlet and color-octet, respectively. It is noted that, in order to get a gaugeindependent result, the external momenta of the heavy quarks inside the mesons should be taken into account, and the heavy-quark masses should be kept up to the second order.
Based on the obtained short-distance Wilson coefficients of the four-quark operators appearing in the effective weak Hamiltonian, and combining the up-to-date experimental data on B 0 s −B 0 s mixing, we then performed a detailed phenomenological analysis of the charged-Higgs effects on this process. Our main conclusions are summarized as follows: (i) While the type-C model gives a nonzero contribution to the Wilson coefficient C T RR C (µ W ), the type-III contribution to the amplitude A T RR is induced only by RG evolution effect.
(ii) In the case of real couplings, the allowed spaces of the Yukawa coupling parameters A u and A d in the two models are obviously different, as shown in Fig. 2. (iii) In the case of complex couplings, due to the cancelation effects among the charged-Higgs contributions, large values of |A u | and |A d | are still allowed around θ ∼ ±90 • in the type-III and around θ ∼ 0 • , ±180 • in the type-C model, which is shown in Fig. 3.
The observed differences could be used to distinguish the two models. It should be noted, however, that their differences in the small |A d | and |A u | ranges are hardly to be distinguished, due to the large theoretical and experimental uncertainties. More refined theoretical and experimental efforts are therefore needed for a much clearer phenomenological picture.