betaFeedback

WorksPhysics and Astronomy

NeffN_{\mathrm{eff}}, Radiative Dirac Neutrinos, and Flavor Physics from Leptoquarks

Abstract

Leptoquarks provide a common framework for addressing flavor anomalies, rare meson decays, and neutrino-mass generation. We study a model with two mixed scalar leptoquarks and three right-handed neutrinos that generates radiative Dirac neutrino masses and contributes to B+→K+ννˉB^+\to K^+ν\barν, charged-lepton flavor violation, and the effective number of relativistic species. We show that the model can accommodate the Belle~II measurement of B+→K+ννˉB^+\to K^+ν\barν while satisfying the μ→eγμ\to eγ bound and the cosmological constraint on ΔNeffΔN_{\mathrm{eff}}. When both leptoquark masses exceed approximately 1212–14 TeV14\,\mathrm{TeV}, the new contribution to the BB-meson decay becomes too small, and the prediction approaches the SM value, which lies below the Belle~II measurement. Interestingly, the lower bounds on the leptoquark mass scale derived from NeffN_{\mathrm{eff}} can be stronger than those from direct collider searches.

† [email protected]

‡ [email protected]

§ [email protected]

[email protected]

Introduction

Understanding why neutrinos are so much lighter than all other known fermions remains one of the fundamental open questions in particle physics. Although several mechanisms have been proposed to generate tiny neutrino masses, seesaw [1, 2] and radiative mechanisms [3, 4] are among the most widely explored frameworks.

Another fundamental question in neutrino physics concerns their nature: whether they are Majorana or Dirac particles. The Majorana scenario is particularly appealing from a phenomenological perspective, as it allows for lepton-number-violating processes, with neutrinoless double-beta decay [5] providing its most prominent experimental signature. As a consequence, scenarios in which neutrinos are Majorana particles have been studied more extensively than the Dirac case [6].

However, the absence of conclusive evidence for neutrinoless double-beta decay, despite decades of experimental searches [7, 8, 9, 10], has motivated renewed interest in scenarios in which neutrinos are Dirac particles. In this context, seesaw-inspired and radiative mechanisms for generating naturally small Dirac neutrino masses, originally investigated several decades ago [11, 12, 13, 14], have attracted considerable attention in recent years [15, 16, 17, 18].

Whether neutrinos are Dirac or Majorana particles, explaining their tiny masses requires physics beyond the Standard Model (SM). Among well-motivated new-physics scenarios that explain neutrino masses, leptoquark models provide a particularly appealing framework [19, 20] in which tiny neutrino masses can arise exclusively through radiative effects [21, 22, 23, 24, 25, 26, 27].

Leptoquark couplings to quarks and leptons can induce semileptonic meson decays [28, 29, 30], including rare BB-meson decays with neutrinos in the final state, mediated by the flavor-changing transitions b→(d,s)νˉνb \to(d,s)\bar{\nu}\nu [31, 32, 33, 34, 35, 36]. Belle II reported the first evidence for B+→K+ννˉB^+ \to K^+\nu\bar{\nu}, with a combined inclusive- and hadronic-tag result at the 10−510^{-5} level, nearly five times the recent SM prediction at the 10−610^{-6} level and corresponding to a 2.7σ2.7\sigma tension [37, 38]. This result has motivated several new-physics explanations that can enhance the branching fraction while remaining consistent with other flavor data [39, 40, 41, 42, 43, 44].

Scalar leptoquarks are particularly appealing in this context because their interactions can both modify the underlying b→sννˉb \to s\nu\bar{\nu} transition and participate in the radiative generation of neutrino masses [24, 25]. Motivated by this connection, we introduce two electroweak-singlet scalar leptoquarks and three right-handed neutrinos, together with a symmetry structure that forbids tree-level Dirac masses and Majorana mass operators. Neutrinos acquire Dirac masses at one loop through down-type quarks and scalar-leptoquark mixing, while the same Yukawa interactions induce flavor-changing neutral currents. We show that this framework can accommodate the B+→K+ννˉB^+ \to K^+\nu\bar{\nu} branching fraction reported by Belle II while reproducing the measured neutrino mass-squared differences and mixing parameters.

The same leptoquark interactions responsible for neutrino mass generation and rare BBmeson decays also induce charged-lepton flavor violation, particularly μ→eγ\mu\to e\gamma. Moreover, the additional right-handed neutrinos contribute to ΔNeff\Delta N_{\mathrm{eff}} if sufficiently populated in the early Universe. These observables impose complementary constraints on the leptoquark masses and Yukawa couplings. Interestingly, for some Yukawa-coupling configurations, the lower bounds on the leptoquark mass scale from NeffN_{\mathrm{eff}} can be stronger than those from current collider searches.1 We therefore perform a combined phenomenological analysis to identify Yukawa textures and parameter-space regions compatible with neutrino oscillation data and the Belle II measurement while satisfying constraints from charged-lepton flavor violation and NeffN_{\mathrm{eff}}.

This paper is organized as follows. In § 0.2, we introduce the model and derive the radiative Dirac neutrino mass matrix. Section III discusses the perturbativity requirements on the Yukawa couplings. Rare meson decays, charged-lepton flavor violation, and the contribution to ΔNeff\Delta N_{\mathrm{eff}} are studied in § 0.4, § 0.5, and VI, respectively. Section VII presents the combined numerical analysis and illustrative benchmarks, and § 0.8 summarizes our conclusions. Appendix § 0.7.1 provides details of the meson-decay calculation, while Appendix § 0.7.2 discusses rare top decays.

Framework and neutrino masses

We extend the SM by introducing three right-handed neutrinos, νiR\nu_{iR}, and two independent scalar leptoquarks, SS and S~\widetilde{S}, while keeping the SM charged-fermion content unchanged. We denote the standard left-handed lepton and quark doublets by LiLL_{iL} and Q~iL\widetilde{Q}_{iL}, respectively, with i=1,2,3i = 1, 2, 3 labeling the fermion generations. Table I summarizes the fermion and scalar fields, their gauge quantum numbers, and their charges under a discrete Z2×Z3\mathbb{Z}_2 \times\mathbb{Z}_3 symmetry imposed to forbid unwanted interactions, particularly tree-level neutrino mass terms and diquark couplings.

The SM Higgs doublet is parametrized as H=(G+,(v+h+iG0)/2)TH = (G^+,(v+h+iG^0)/\sqrt{2})^T, where vv denotes the electroweak vacuum expectation value, hh is the physical Higgs boson, and G±G^\pm and G0G^0 are the would-be Goldstone bosons associated with electroweak symmetry breaking.

FieldSU(3)CSU(3)_CSU(2)LSU(2)_LU(1)YU(1)_YZ2Z_2Z3Z_3BBLL
Q^iL\widehat{Q}_{iL}3\mathbf{3}2\mathbf{2}16\frac{1}{6}111113\frac{1}{3}00
u^iR\widehat{u}_{iR}3\mathbf{3}1\mathbf{1}23\frac{2}{3}111113\frac{1}{3}00
d^iR\widehat{d}_{iR}3\mathbf{3}1\mathbf{1}−13-\frac{1}{3}111113\frac{1}{3}00
LiLL_{iL}1\mathbf{1}2\mathbf{2}−12-\frac{1}{2}−1-1110011
eiRe_{iR}1\mathbf{1}1\mathbf{1}−1-1−1-1110011
νiR\nu_{iR}1\mathbf{1}1\mathbf{1}0011ω32\omega_3^20011
HH1\mathbf{1}2\mathbf{2}12\frac{1}{2}11110000
SS3‾\overline{\mathbf{3}}1\mathbf{1}13\frac{1}{3}−1-111−13-\frac{1}{3}−1-1
S~\widetilde{S}3\mathbf{3}1\mathbf{1}13\frac{1}{3}11ω3\omega_3−13-\frac{1}{3}−1-1

Table 1. Field content and gauge, discrete, baryon, and lepton charges. We use the hypercharge convention Q=T3+YQ = T_3 + Y and define ω3=e2πi/3\omega_3 = e^{2\pi i/3}. The generation index is i=1,2,3i = 1, 2, 3. Hats denote quark weak eigenstates; SS and S~\widetilde{S} are independent scalar leptoquarks.

The Yukawa interactions consistent with the particle content and all symmetries of the model are given by

LY=yiilL‾iLeiRH+yijdQ^iLd^jRH+yijuQ^iLH~u^jR+YiαLQ^iLCaϵabLαLbS+YiαRd^iRCναRS~+Y~iαRu^iRCeαRS+H.c.,(1)\begin{aligned} \mathcal{L}^{Y}={}&y^l_{ii}\overline{L}_{iL}e_{iR}H+y^d_{ij}\widehat{Q}_{iL}\widehat{d}_{jR}H+y^u_{ij}\widehat{Q}_{iL}\widetilde{H}\widehat{u}_{jR}\\ &+Y^L_{i\alpha}\widehat{Q}^{C a}_{iL}\epsilon^{ab}L^b_{\alpha L}S+Y^R_{i\alpha}\widehat{d}^{C}_{iR}\nu_{\alpha R}\widetilde{S}+\widetilde{Y}^R_{i\alpha}\widehat{u}^{C}_{iR}e_{\alpha R}S+\mathrm{H.c.}, \tag*{(1)} \end{aligned}

where H~=iσ2H∗\widetilde{H}=i\sigma_2H^*, ϵ12=1\epsilon^{12}=1, hats denote quark weak eigenstates, and α\alpha labels lepton flavors.

This set of interactions conserves baryon and lepton numbers with B(S)=B(S~)=−1/3B(S)=B(\widetilde{S})=-1/3 and L(S)=L(S~)=−1L(S)=L(\widetilde{S})=-1. This conservation persists after the soft scalar mixing introduced below, protecting proton stability and the Dirac nature of neutrinos within the renormalizable theory.

We choose the charged-lepton mass basis, so yly^l is diagonal, and a quark weak basis with trivial right-handed quark rotations. The weak and mass eigenstates of the left-handed upand down-type quarks are related by

u^L=VLuuL,d^L=VLddL.(2)\hat{u}_L = V_L^u u_L,\qquad\hat{d}_L = V_L^d d_L. \tag*{(2)}

The matrices VLuV_L^u and VLdV_L^d are unitary, and their mismatch defines the Cabibbo–Kobayashi–Maskawa (CKM) matrix through V≡VCKM=(VLu)†VLdV \equiv V_{\mathrm{CKM}} = (V_L^u)^\dagger V_L^d. In the standard CKM parametrization, we use s12=0.22517s_{12} = 0.22517, s13=0.003763s_{13} = 0.003763, s23=0.04189s_{23} = 0.04189, and δ=1.154\delta= 1.154, as reported by the 2026 PDG [6]. We work in the convenient basis VLu=IV_L^u = I, so that VLd=VV_L^d = V, and independently assume that YRY^R is diagonal. Since q^C=qˉC(VLq)T\hat{q}^C = \bar{q}^C(V_L^q)^{\mathsf{T}}, the leptoquark interactions in the fermion mass basis, with the left-handed neutrinos still labeled by weak flavor, are

LLQY⊃YiαLuˉiLCeαLS−ζiαdˉiLCναLS+YiiRdˉiRCνiRS~+Y~iαRuˉiRCeαRS+H.c.(3)\mathcal{L}_{\mathrm{LQ}}^Y \supset Y^L_{i\alpha}\bar{u}_{iL}^C e_{\alpha L}S-\zeta_{i\alpha}\bar{d}_{iL}^C\nu_{\alpha L}S+Y^R_{ii}\bar{d}_{iR}^C\nu_{iR}\widetilde{S}+\widetilde{Y}^R_{i\alpha}\bar{u}_{iR}^C e_{\alpha R}S+\mathrm{H.c.} \tag*{(3)}

The rotation of the down-type quarks defines ζ≡VTYL\zeta\equiv V^{\mathsf{T}}Y^L, with the inverse relation YL=V∗ζY^L = V^*\zeta. Here u1,2,3=u,c,tu_{1,2,3}=u,c,t and d1,2,3=d,s,bd_{1,2,3}=d,s,b. Although allowed by the symmetries, the coupling Y~R\widetilde{Y}^R does not enter the one-loop mechanism responsible for the Dirac neutrino masses.

The scalar potential consistent with the symmetries of the model is

V0=μ2H†H+μS2S†S+μS~2S~†S~+λ1(H†H)2+λ2(S†S)2+λ3(S~†S~)2+λ4(H†H)(S†S)+λ5(H†H)(S~†S~)+λ6(S†S)(S~†S~)+λ7∣S†S~∣2.(4)\begin{aligned} V_0={}&\mu^2 H^\dagger H+\mu_S^2 S^\dagger S+\mu_{\widetilde{S}}^2\widetilde{S}^\dagger\widetilde{S}+\lambda_1(H^\dagger H)^2+\lambda_2(S^\dagger S)^2+\lambda_3(\widetilde{S}^\dagger\widetilde{S})^2\\ &+\lambda_4(H^\dagger H)(S^\dagger S)+\lambda_5(H^\dagger H)(\widetilde{S}^\dagger\widetilde{S})+\lambda_6(S^\dagger S)(\widetilde{S}^\dagger\widetilde{S})+\lambda_7\lvert S^\dagger\widetilde{S}\rvert^2. \tag*{(4)} \end{aligned}

To preserve color, we adopt the standard electroweak vacuum, ⟨H⟩=(0,v/2)T\langle H\rangle=(0,v/\sqrt{2})^{\mathsf{T}} and ⟨S⟩=⟨S~⟩=0\langle S\rangle=\langle\widetilde{S}\rangle=0. The Higgs mass is mh2=2λ1v2m_h^2=2\lambda_1v^2, and the leptoquark diagonal mass parameters are mS2=μS2+λ4v2/2m_S^2=\mu_S^2+\lambda_4v^2/2 and mS~2=μS~2+λ5v2/2m_{\widetilde{S}}^2=\mu_{\widetilde{S}}^2+\lambda_5v^2/2. To generate neutrino masses, we introduce a soft symmetry-breaking term that mixes the two leptoquarks and induces a non-diagonal mass matrix,

Vsoft=μSS~2S†S~+H.c.,MS2=(mS2μSS~2μSS~2mS~2).(5)V_{\mathrm{soft}}=\mu_{S\widetilde{S}}^2S^\dagger\widetilde{S}+\mathrm{H.c.},\qquad\mathcal{M}_S^2=\begin{pmatrix}m_S^2&\mu_{S\widetilde{S}}^2\\\mu_{S\widetilde{S}}^2&m_{\widetilde{S}}^2\end{pmatrix}. \tag*{(5)}

The mass matrix MS2\mathcal{M}_S^2 is diagonalized by an orthogonal rotation with scalar mixing angle θ\theta,

\begin{pmatrix}S_1\\S_2\end{pmatrix}=\begin{pmatrix}c_\theta&s_\theta\\-s_\theta&c_\theta\end{pmatrix}\begin{pmatrix}S\\\widetilde{S}\end{pmatrix},\qquad\sin(2\theta)=\frac{2\mu_{S\widetilde{S}}^2}{m_{S_1}^2-m_{S_2}^2}. \tag*{(6)}\end{pmatrix}

where cθ=cos⁡θc_\theta=\cos\theta and sθ=sin⁡θs_\theta=\sin\theta. All interactions with physical leptoquarks follow from Eq. (3) by replacing S=cθS1−sθS2S=c_\theta S_1-s_\theta S_2 and S~=sθS1+cθS2\widetilde{S}=s_\theta S_1+c_\theta S_2.

The Yukawa interactions in Eq. (3), together with mixing between the scalar leptoquarks SS and S~\widetilde{S}, generate Dirac neutrino masses at one loop, as illustrated in Figure 1. The required chirality flip occurs along the internal down-type quark line, rendering the resulting neutrino masses proportional to the corresponding quark masses. For diagonal YRY^R, the result is [47]

Figure 1. One-loop Dirac-neutrino mass diagrams for diagonal YRY^R. The external indices are the left-handed weak flavor α\alpha and the right-handed generation qq, which also labels the internal down quark. The cross denotes its mass insertion; the two scalar contributions enter with opposite mixing factors.

(mνD)αq=3sin⁡(2θ)32π2ζqα∗mdqYqqR[F(mS12,mdq2)−F(mS22,mdq2)],(7)\left(m_{\nu}^{D}\right)_{\alpha q} = \frac{3\sin(2\theta)}{32\pi^2}\zeta^{*}_{q\alpha}m_{d_q}Y^{R}_{qq}\left[F\left(m_{S_1}^{2},m_{d_q}^{2}\right)-F\left(m_{S_2}^{2},m_{d_q}^{2}\right)\right], \tag*{(7)}

with no sum over qq and (md1,md2,md3)=(md,ms,mb)\left(m_{d_1},m_{d_2},m_{d_3}\right)=\left(m_d,m_s,m_b\right). The factor of three accounts for the three quark colors. The loop function is

F(M2,m2)=M2M2−m2ln⁡M2m2.(8)F\left(M^2,m^2\right)=\frac{M^2}{M^2-m^2}\ln\frac{M^2}{m^2}. \tag*{(8)}

In the heavy-leptoquark limit, mS1,mS2≫mdqm_{S_1},m_{S_2}\gg m_{d_q}, the difference of loop functions becomes

F(mS12,mdq2)−F(mS22,mdq2)≃ln⁡mS12mS22.(9)F\left(m_{S_1}^{2},m_{d_q}^{2}\right)-F\left(m_{S_2}^{2},m_{d_q}^{2}\right)\simeq\ln\frac{m_{S_1}^{2}}{m_{S_2}^{2}}. \tag*{(9)}

Thus, at finite Yukawa couplings, neutrino masses vanish for zero mixing or exact scalar degeneracy.

In matrix form, we have

mνD=ζ†A,withA=3sin⁡(2θ)32π2mdYR[F(mS12,md2)−F(mS22,md2)].(10)m_{\nu}^{D}=\zeta^{\dagger}\mathcal{A},\quad\text{with}\quad\mathcal{A}=\frac{3\sin(2\theta)}{32\pi^2}m_dY^R\left[F\left(m_{S_1}^{2},m_d^{2}\right)-F\left(m_{S_2}^{2},m_d^{2}\right)\right]. \tag*{(10)}

Here mdm_d is the diagonal down-quark mass matrix, the loop function FF is evaluated separately for each diagonal mass-squared entry mdq2m_{d_q}^{2}, and Aq≡(A)qq\mathcal{A}_q\equiv(\mathcal{A})_{qq}. The Dirac mass matrix admits the biunitary decomposition [48]

mνD=UPMNSDνVR†,Dν=diag⁡(m1,m2,m3).(11)m_{\nu}^{D}=U_{\mathrm{PMNS}}D_{\nu}V_{R}^{\dagger},\quad D_{\nu}=\operatorname{diag}(m_1,m_2,m_3). \tag*{(11)}

When A\mathcal{A} is nonsingular, all left-handed leptoquark couplings follow from

ζ=(A−1)†VRDνUPMNS†→YL=V∗(A−1)†VRDνUPMNS†.(12)\zeta=\left(\mathcal{A}^{-1}\right)^{\dagger}V_RD_{\nu}U_{\mathrm{PMNS}}^{\dagger}\quad\to\quad Y^L=V^{*}\left(\mathcal{A}^{-1}\right)^{\dagger}V_RD_{\nu}U_{\mathrm{PMNS}}^{\dagger}. \tag*{(12)}

For the oscillation inputs, we use the NuFIT 6.0 best fit including atmospheric information [49]:

sin⁡2θ12=0.308,δCP=212∘,sin⁡2θ23=0.470,Δm212=7.49×10−5 eV2,sin⁡2θ13=0.02215,Δm312=2.513×10−3 eV2.(13)\begin{aligned} \sin^2\theta_{12} &= 0.308, & \delta_{\mathrm{CP}} &= 212^\circ,\\ \sin^2\theta_{23} &= 0.470, & \Delta m^2_{21} &= 7.49 \times10^{-5}\,\mathrm{eV}^2,\\ \sin^2\theta_{13} &= 0.02215, & \Delta m^2_{31} &= 2.513 \times10^{-3}\,\mathrm{eV}^2. \tag*{(13)} \end{aligned}

The quoted values correspond to normal ordering [49], which we adopt in the numerical analysis. We fix m1=0m_1 = 0, so m2=Δm212m_2 = \sqrt{\Delta m^2_{21}} and m3=Δm312m_3 = \sqrt{\Delta m^2_{31}}, and retain the PMNS phase. We additionally choose VR=IV_R = I, aligning the right-handed neutrino mass basis with the basis in which YRY^R is diagonal. The first column of mνDm_\nu^D must then vanish:

(mνD)α1=ζ1α∗A1=0for every α.(14)\left(m_\nu^D\right)_{\alpha1} = \zeta^*_{1\alpha}\mathcal{A}_1 = 0 \qquad\text{for every } \alpha. \tag*{(14)}

For nondegenerate scalars and nonzero sin⁡(2θ)\sin(2\theta), this requires either Y11R=0Y^R_{11} = 0 or ζ1α=0\zeta_{1\alpha} = 0 for all α\alpha, but does not require both. If Y11R≠0Y^R_{11} \ne0, the first row of ζ\zeta must vanish; if Y11R=0Y^R_{11} = 0, that row is unconstrained by the mass relation.

For the combined numerical analysis in Sec. VII, we adopt the simple texture

YR=diag⁡(0,y2,y3),y2,y3>0,ζ1α=0,Y~R=0.(15)Y^R = \operatorname{diag}(0,y_2,y_3), \qquad y_2,y_3 > 0, \qquad\zeta_{1\alpha} = 0, \qquad\widetilde{Y}^R = 0. \tag*{(15)}

Thus we set both factors in Eq. (14) to zero as an additional simplifying choice. Since A1=0\mathcal{A}_1 = 0, the inverse in Eq. (12) is used only on q=2,3q = 2,3, giving

ζqα=mνqUαq∗Aq∗ (q=2,3),YL=V∗ζ.(16)\zeta_{q\alpha} = \frac{m_{\nu_q}U^*_{\alpha q}}{\mathcal{A}^*_q}\ (q = 2,3), \qquad Y^L = V^*\zeta. \tag*{(16)}

Here, for q=2,3q = 2,3, mνqm_{\nu_q} denotes the qqth neutrino mass, whereas mdqm_{d_q} denotes the corresponding down-quark mass. The CKM rotation generally leaves the first row of YLY^L nonzero even though ζ1α=0\zeta_{1\alpha} = 0.

Thus, the phenomenology is determined by the scalar masses, the scalar mixing angle, and the Yukawa couplings in Eq. (3). Using a diagonal YRY^R as input, the reconstruction in Eqs. (12) and (16) expresses ζ\zeta and YLY^L in terms of YRY^R and the neutrino oscillation and quark mixing inputs. In the following section, we study the perturbativity of the Yukawa couplings within this texture.

Perturbativity of Yukawa couplings

We apply the texture in Eq. (15) and the reconstruction in Eq. (16). We apply the conventional elementwise perturbativity cutoff 4π\sqrt{4\pi} to the Yukawa matrices YRY^R, ζ\zeta, and YLY^L, following Ref. [50]. For the chosen texture, this condition reads

Λ(y2,y3)≡max⁡{y2,y3,max⁡q=2,3α∣ζqα∣,max⁡i,α∣YiαL∣}<4π.(17)\Lambda(y_2,y_3) \equiv\max\left\{y_2,y_3,\max_{\substack{q=2,3\\\alpha}}\lvert\zeta_{q\alpha}\rvert,\max_{i,\alpha}\lvert Y^L_{i\alpha}\rvert\right\}<\sqrt{4\pi}. \tag*{(17)}

We use Eq. (17) as a selection criterion for the numerical scan. Partial-wave unitarity bounds on Yukawa couplings depend on the scattering channels and field representations [51]. Through the reconstruction in Eq. (16), each entry of YLY^L generally depends on both y2y_2 and y3y_3. The full criterion therefore requires a joint test of the two couplings.

In this section, we evaluate Aq\mathcal{A}_q in Eq. (10) using the logarithmic approximation of Eq. (9), valid for mS1,mS2≫mbm_{S_1},m_{S_2}\gg m_b. The scalar dependence is then contained in sin⁡(2θ)\sin(2\theta) and the mass ratio. It is useful to define the relative scalar-mass splitting ΔS≡2∣mS1−mS2∣/(mS1+mS2)\Delta_S\equiv2\lvert m_{S_1}-m_{S_2}\rvert/(m_{S_1}+m_{S_2}), with 0≤ΔS<20\leq\Delta_S<2. The logarithmic factor can then be written as ∣ln⁡(mS12/mS22)∣=4arctanh⁡(ΔS/2)\lvert\ln(m_{S_1}^2/m_{S_2}^2)\rvert=4\operatorname{arctanh}(\Delta_S/2).

Note that, at fixed (sin⁡θ,ΔS)(\sin\theta,\Delta_S), increasing an entry yqy_q of YRY^R suppresses the corresponding row of ζ\zeta and its contribution to YLY^L, while decreasing yqy_q has the opposite effect. These competing tendencies constrain the allowed values of y2y_2 and y3y_3. At each point in the (sin⁡θ,ΔS)(\sin\theta,\Delta_S) plane, we ask whether at least one pair satisfies the common perturbativity bound. This existence condition can be written as

Λmin⁡(sin⁡θ,ΔS)=min⁡y2,y3>0Λ(y2,y3),Λmin⁡<4π.(18)\Lambda_{\min}(\sin\theta,\Delta_S)=\min_{y_2,y_3>0}\Lambda(y_2,y_3),\qquad\Lambda_{\min}<\sqrt{4\pi}. \tag*{(18)}

For nondegenerate scalar masses and nonzero sin⁡(2θ)\sin(2\theta), the reconstruction in Eq. (16) reproduces the input neutrino masses and mixing for each positive pair (y2,y3)(y_2,y_3). Equation (18) tests whether any of these pairs yields perturbative couplings. Figure 2 shows Λmin⁡\Lambda_{\min} on the color scale in the perturbative region. Light gray indicates points with no perturbative solution, and the black contour marks the boundary Λmin⁡=4π\Lambda_{\min}=\sqrt{4\pi}.

Figure 2. Perturbativity map for the texture in Eq. (15). The color scale gives Λmin⁡\Lambda_{\min} where a perturbative texture exists; light gray denotes no perturbative solution for the common bound on YRY^{R}, ζ\zeta, and YLY^{L}. The black contour marks Λmin⁡=4π\Lambda_{\min}=\sqrt{4\pi}.

At exact scalar degeneracy or either mixing endpoint, sin⁡θ=0\sin\theta=0 or 11, the radiative mass vanishes for finite Yukawa couplings, so no texture reproduces m2,m3≠0m_2,m_3\ne0. Approaching these limits suppresses the loop factor and requires larger Yukawa couplings to maintain the neutrino masses. The joint criterion above quantifies the corresponding loss of perturbative solutions.

Rare meson decays

At the quark level, B+→K+ννˉB^{+}\to K^{+}\nu\bar{\nu} proceeds through bˉ→sˉννˉ\bar{b}\to\bar{s}\nu\bar{\nu}, the charge conjugate of b→sννˉb\to s\nu\bar{\nu}. Combining its inclusive- and hadronic-tag analyses, Belle II obtained [37]

B(B+→K+ννˉ)=[2.3±0.5  (stat)−0.4+0.5  (syst)]×10−5.(19)\mathcal{B}(B^{+}\to K^{+}\nu\bar{\nu})=\left[2.3\pm0.5\;(\mathrm{stat}){}^{+0.5}_{-0.4}\;(\mathrm{syst})\right]\times10^{-5}. \tag*{(19)}

Combining the statistical and systematic uncertainties in quadrature separately on each side gives the experimental 1σ1\sigma interval B(B+→K+ννˉ)∈[1.6596876,3.0071068]×10−5\mathcal{B}(B^{+}\to K^{+}\nu\bar{\nu})\in[1.6596876,3.0071068]\times10^{-5}. A recent short-distance SM prediction based on the HPQCD form factors and an inclusive determination of ∣Vcb∣\lvert V_{cb}\rvert gives [38]

B(B+→K+ννˉ)SMSD=(4.78±0.30)×10−6.(20)\mathcal{B}(B^{+}\to K^{+}\nu\bar{\nu})_{\mathrm{SM}}^{\mathrm{SD}}=(4.78\pm0.30)\times10^{-6}. \tag*{(20)}

Our numerical implementation is consistent with this result.2 Thus, the Belle II central value is approximately 4.8 times the SM prediction. Figure 3 shows the tree-level leptoquark contribution: the bˉ\bar{b} transitions to a sˉ\bar{s}, while the uu quark is a spectator.

Tree-level contribution to $B^+ \to K^+\nu_i\bar{\nu}_j$ from physical leptoquarks $S_a$

Figure 3. Tree-level contribution to B+→K+νiνˉjB^+ \to K^+\nu_i\bar{\nu}_j from the physical leptoquarks SaS_a (a=1,2a = 1, 2). The choices A,B=L,RA, B = L, R include both equal-chirality vector and mixed-chirality scalar and tensor contributions.

Since the neutrino flavors are not resolved, we sum the differential rate, including the SM and leptoquark contributions, over all νiνˉj\nu_i\bar{\nu}_j final states, including i≠ji \ne j. Integrating this inclusive rate over the allowed range of the neutrino-pair invariant mass squared gives

B(B+→K+ννˉ)=τB+∫0(mB−mK)2dΓdq2 dq2,(21)\mathcal{B}(B^+ \to K^+\nu\bar{\nu}) = \tau_{B^+}\int_0^{(m_B-m_K)^2} \frac{\mathrm{d}\Gamma}{\mathrm{d}q^2}\,\mathrm{d}q^2, \tag*{(21)}

where mBm_B and mKm_K are the charged-meson masses, q2q^2 is the invariant mass squared of the neutrino pair, and τB+=(1.638±0.004)×10−12 s\tau_{B^+} = (1.638 \pm0.004) \times10^{-12}\,\mathrm{s} is the charged-BB lifetime [6].

The down-quark-neutrino couplings of the physical leptoquarks are

YL(1)=cθζ,YL(2)=−sθζ,YR(1)=sθYR,YR(2)=cθYR.(22)\mathcal{Y}_L^{(1)} = c_\theta\zeta,\quad\mathcal{Y}_L^{(2)} = -s_\theta\zeta,\quad\mathcal{Y}_R^{(1)} = s_\theta Y^R,\quad\mathcal{Y}_R^{(2)} = c_\theta Y^R. \tag*{(22)}

Here ζ\zeta and YRY^R are defined in Eq. (3). At tree level, two vertices of the same chirality generate vector operators, while opposite chiralities generate scalar and tensor operators. We evolve the contribution of each scalar from its respective matching scale down to μb=4.8 GeV\mu_b = 4.8\,\mathrm{GeV} and employ the HPQCD lattice-QCD form factors [52, 53]. The differential rate, matching coefficients, QCD evolution, and form-factor inputs are given in Appendix § 0.7.1.

Figure 4 shows the branching fraction for sin⁡θ=10−9\sin\theta= 10^{-9}, 3×10−93 \times10^{-9}, and 0.10.1, with YR=diag⁡(1,1,1)Y^R = \operatorname{diag}(1,1,1). For these choices, points compatible with Belle II occur mainly for mS2m_{S_2} of a few TeV, while mS1m_{S_1} can extend into the tens-of-TeV range. As scalar degeneracy is approached, the loop-function difference in Eq. (7) vanishes, and reproducing the neutrino masses requires increasingly large Yukawa couplings. The gray region identifies points that fail the perturbativity criterion in Eq. (17). This region narrows as the mixing angle increases across the three panels because, at fixed YRY^R and scalar masses, the required left-handed couplings scale as 1/∣sin⁡(2θ)∣1/\lvert\sin(2\theta)\rvert.

Figure 4. Branching fraction for B+→K+ννˉB^+ \to K^+\nu\bar{\nu} in the (mS1,mS2)(m_{S_1},m_{S_2}) plane for YR=diag⁡(1,1,1)Y^R = \operatorname{diag}(1,1,1). The panels show sin⁡θ=10−9\sin\theta= 10^{-9} (left), 3×10−93 \times10^{-9} (center), and 0.10.1 (right). The black line indicates mS1=mS2m_{S_1} = m_{S_2}. In the gray region, at least one entry of YRY^R, ζ\zeta, or YL=V∗ζY^L = V^*\zeta reaches or exceeds 4π\sqrt{4\pi}. The orange band is the experimental 1σ1\sigma interval from the Belle II measurement in Eq. (19), with statistical and systematic uncertainties combined in quadrature.

Charged-lepton flavor violation

Charged-lepton flavor violation (CLFV) provides a particularly clean probe of the leptoquark Yukawa sector. In the SM extended only to include light neutrino masses and mixing, the predicted branching fraction for μ→eγ\mu\to e\gamma is of order 10−5410^{-54} [54, 55, 56], far below experimental sensitivity. Consequently, any observable signal would constitute unambiguous evidence for physics beyond the SM. In the present model, the same couplings YLY^L reconstructed in Eq. (12) couple charged leptons to up-type quarks and generate the electromagnetic dipole transition at one loop, thereby testing the flavor structure required to reproduce neutrino oscillation data. Figure 5 shows the relevant diagrams.

Figure 5. One-loop diagrams contributing to the electromagnetic dipole transition μ→eγ\mu\to e\gamma. The photon is emitted from the physical leptoquark SaS_a in the left panel and from the charge-conjugate up-type quark uicu_i^c in the right panel. Here a=1,2a=1,2.

For the decay μ→eγ\mu\to e\gamma, the branching fraction is [57]3

B(μ→eγ)≃3αem256πGF2(cos⁡2θmS12+sin⁡2θmS22)2∣∑i=13Yi2LYi1L∗∣2.(23)\mathcal{B}(\mu\to e\gamma) \simeq\frac{3\alpha_{\mathrm{em}}}{256\pi G_F^2}\left(\frac{\cos^2\theta}{m_{S_1}^2}+\frac{\sin^2\theta}{m_{S_2}^2}\right)^2\left|\sum_{i=1}^{3}Y^L_{i2}Y^{L*}_{i1}\right|^2 . \tag*{(23)}

Here αem\alpha_{\mathrm{em}} is the electromagnetic fine-structure constant and GFG_F is the Fermi constant.

Figure 11. Cumulative constraints in the (y2,y3)(y_2,y_3) plane. Columns show the indicated scalar spectra; rows correspond to sin⁡θ=10−3\sin\theta= 10^{-3}, sin⁡θ=1/2\sin\theta= 1/\sqrt{2}, and cos⁡θ=10−3\cos\theta= 10^{-3}. White corresponds to a nonperturbative neutrino reconstruction; light gray has a perturbative neutrino reconstruction; orange also satisfies Belle II; blue additionally satisfies MEG II; and green satisfies all four conditions. The constraints are applied cumulatively in the order shown in the legend.

The flavor-violating combination ∑iYi2LYi1L∗\sum_i Y^L_{i2}Y^{L*}_{i1} is fixed once the neutrino masses and mixings, the right-handed Yukawa matrix YRY^R, and the scalar-sector parameters are chosen. Since YL†YL=ζ†ζY^{L\dagger}Y^L=\zeta^\dagger\zeta, this flavor factor is independent of the CKM rotation in the approximation of Eq. (23). This makes μ→eγ\mu\to e\gamma a direct and stringent consistency condition on the neutrino-mass solution. The MEG II search using the 2021–2022 data found no significant excess and established the most stringent upper limit to date [58],

B(μ+→e+γ)<1.5×10−13(90% C.L.).(24)\mathcal{B}(\mu^+ \to e^+\gamma) < 1.5 \times10^{-13} \qquad(90\% \ \mathrm{C.L.}). \tag*{(24)}

Figure 6 shows the MEG II constraint for the fixed benchmark YR=diag⁡(1,1,1)Y^R=\operatorname{diag}(1,1,1). For sin⁡θ=10−9\sin\theta=10^{-9}, the lower bound on mS1m_{S_1} increases from approximately 17.4 TeV at mS2=1m_{S_2}=1 TeV to 51.9 TeV at mS2=20m_{S_2}=20 TeV. For sin⁡θ=3×10−9\sin\theta=3\times10^{-9}, these bounds decrease to 8.0 and 33.0 TeV, respectively, showing that larger scalar mixing weakens the CLFV constraint.

Figure 6. Constraint from the YLY^L contribution to μ→eγ\mu\to e\gamma in the (mS1,mS2)(m_{S_1},m_{S_2}) plane for YR=diag⁡(1,1,1)Y^R = \operatorname{diag}(1,1,1), using ζ=VCKMTYL\zeta= V_{\mathrm{CKM}}^{T}Y^L, mνD=ζ†Am_\nu^D = \zeta^\dagger A, and the restricted normal-ordering texture ζ1α=0\zeta_{1\alpha}=0. The panels show sin⁡θ=10−9\sin\theta= 10^{-9} (left), sin⁡θ=3×10−9\sin\theta= 3 \times10^{-9} (center), and sin⁡θ=3×10−8\sin\theta= 3 \times10^{-8} (right). Blue regions satisfy the MEG II bound in Eq. (24), while orange regions are excluded. The black contours correspond to B(μ→eγ)=1.5×10−13\mathcal{B}(\mu\to e\gamma) = 1.5 \times10^{-13}.

For sin⁡θ=3×10−8\sin\theta=3\times10^{-8}, the exclusion is concentrated near mS1=mS2m_{S_1}=m_{S_2}, while either mass hierarchy is allowed sufficiently far from degeneracy. This follows from Eqs. (12) and (9), since YLY^L scales inversely with sin⁡(2θ)ln⁡(mS12/mS22)\sin(2\theta)\ln(m_{S_1}^2/m_{S_2}^2). Scalar degeneracy therefore enhances the dipole contribution and the sensitivity to the MEG II limit.

We set Y~R=0\widetilde{Y}^R=0 in the CLFV analysis, defining a minimal benchmark in which neutrino data fix the dipole amplitude once YRY^R and the scalar-sector parameters are specified. The matrix Y~R\widetilde{Y}^R introduces independent flavor parameters that are not required for neutrino mass generation or the tree-level b→sννˉb\to s\nu\bar{\nu} transition. For nonzero Y~R\widetilde{Y}^R, the amplitude also receives pure right-handed contributions proportional to ∑iY~iμRY~ieR∗\sum_i\widetilde{Y}^R_{i\mu}\widetilde{Y}^{R*}_{ie} and mixed-chirality terms involving YLY^L and Y~R\widetilde{Y}^R. The latter involve an internal-quark chirality flip, are weighted by the up-type-quark mass, and can interfere with contributions to the same dipole chirality.

A diagonal Y~R\widetilde{Y}^{R} eliminates the pure right-handed flavor-changing combination, but not the mixed terms because YLY^L is non-diagonal.

Neff

We study ΔNeff≡Neff−NeffSM\Delta N_{\mathrm{eff}} \equiv N_{\mathrm{eff}} - N_{\mathrm{eff}}^{\mathrm{SM}}, the additional contribution to the effective number of relativistic neutrino species. In this model, the YRY^R interactions can populate the right-handed neutrinos in the early Universe. If they thermalize and later decouple, their temperature is diluted relative to that of the active neutrinos, yielding a contribution controlled by their decoupling temperatures [59, 60].

For the cosmological constraint, we use the Planck 2018 TT,TE,EE+lowE+lensing+BAO result in the Λ\LambdaCDM+NeffN_{\mathrm{eff}} extension, Neff=2.99−0.33+0.34N_{\mathrm{eff}} = 2.99^{+0.34}_{-0.33} at 95% confidence level [61]. Together with the most recent SM prediction NeffSM=3.0440±0.0002N_{\mathrm{eff}}^{\mathrm{SM}} = 3.0440 \pm0.0002 [62, 63], this motivates the approximate upper cut ΔNeff≤0.286\Delta N_{\mathrm{eff}} \le0.286 used below. More recent results combining ACT DR6, Planck, and DESI Year-1 data find Neff=2.86±0.13N_{\mathrm{eff}} = 2.86 \pm0.13 at 68% confidence level [64]; we retain Planck 2018+BAO as a conservative reference and to facilitate comparison with previous studies.

To determine whether the right-handed neutrinos thermalize, we consider the tt-channel annihilation νiRνˉjR→dˉidj\nu_{iR}\bar{\nu}_{jR}\to\bar{d}_i d_j, with d1,2,3=d,s,bd_{1,2,3}=d,s,b, shown in Figure 7. The inverse reactions populate the right-handed states.4

Figure 7. tt-channel annihilation νiRνˉjR→dˉidj\nu_{iR}\bar{\nu}_{jR}\to\bar{d}_i d_j through SaS_a (a=1,2a=1,2). Its charge-conjugate channel gives dˉidˉj\bar{d}_i\bar{d}_j; the inverse reactions populate the right-handed neutrinos.

We neglect neutrino masses in the reaction kinematics while retaining the down-quark masses and work at the free-quark level.5 In terms of the Mandelstam variables ss and tt, the cross section for fixed initial helicities, summed over final spins and colors, is

σij(s)=3∣YiiRYjjR∣216πs2[sin⁡4θf(mS12)+cos⁡4θf(mS22)+2sin⁡2θcos⁡2θI12].(25)\sigma_{ij}(s)=\frac{3\lvert Y_{ii}^{R}Y_{jj}^{R}\rvert^{2}}{16\pi s^{2}}\left[\sin^{4}\theta f(m_{S_1}^{2})+\cos^{4}\theta f(m_{S_2}^{2})+2\sin^{2}\theta\cos^{2}\theta I_{12}\right]. \tag*{(25)}

The function associated with each diagonal scalar contribution is

f(mSa2)=(t+−t−)+(2mSa2−mdi2−mdj2)ln⁡(t+−mSa2t−−mSa2)+(mSa2−mdi2)(mSa2−mdj2)(1t−−mSa2−1t+−mSa2).(26)\begin{aligned} f(m_{S_a}^{2})={}&(t_{+}-t_{-})+(2m_{S_a}^{2}-m_{d_i}^{2}-m_{d_j}^{2})\ln\left(\frac{t_{+}-m_{S_a}^{2}}{t_{-}-m_{S_a}^{2}}\right)\\ &+(m_{S_a}^{2}-m_{d_i}^{2})(m_{S_a}^{2}-m_{d_j}^{2})\left(\frac{1}{t_{-}-m_{S_a}^{2}}-\frac{1}{t_{+}-m_{S_a}^{2}}\right). \tag*{(26)} \end{aligned}

The corresponding interference function is

I12=(t+−t−)+Aln⁡(t+−mS12t−−mS12)+Bln⁡(t+−mS22t−−mS22).(27)I_{12}=(t_{+}-t_{-})+A\ln\left(\frac{t_{+}-m_{S_1}^{2}}{t_{-}-m_{S_1}^{2}}\right)+B\ln\left(\frac{t_{+}-m_{S_2}^{2}}{t_{-}-m_{S_2}^{2}}\right). \tag*{(27)}

The kinematic quantities entering these functions are

t±=mdi2+mdj2−s±λ(s,mdi2,mdj2)2,(28)t_{\pm}=\frac{m_{d_i}^{2}+m_{d_j}^{2}-s\pm\sqrt{\lambda(s,m_{d_i}^{2},m_{d_j}^{2})}}{2}, \tag*{(28)}
λ(x,y,z)=x2+y2+z2−2xy−2xz−2yz,\lambda(x,y,z)=x^{2}+y^{2}+z^{2}-2xy-2xz-2yz,

where λ\lambda is the Källén function. Finally, the coefficients in Eq. (27) are

A=(mS12−mdi2)(mS12−mdj2)mS12−mS22,B=(mS22−mdi2)(mS22−mdj2)mS22−mS12.(29)A=\frac{(m_{S_1}^{2}-m_{d_i}^{2})(m_{S_1}^{2}-m_{d_j}^{2})}{m_{S_1}^{2}-m_{S_2}^{2}},\qquad B=\frac{(m_{S_2}^{2}-m_{d_i}^{2})(m_{S_2}^{2}-m_{d_j}^{2})}{m_{S_2}^{2}-m_{S_1}^{2}}. \tag*{(29)}

The cross section vanishes below the threshold s=(mdi+mdj)2s=(m_{d_i}+m_{d_j})^2.6

Let nνReq(T)n_{\nu_R}^{\mathrm{eq}}(T) denote the equilibrium density of one generation, including both νiR\nu_{iR} and νˉiR\bar{\nu}_{iR}. At a common temperature TT and vanishing chemical potentials, this density is the same for all generations:

nνReq(T)=34ζ(3)π2gνRT3.(30)n_{\nu_R}^{\mathrm{eq}}(T)=\frac{3}{4}\frac{\zeta(3)}{\pi^2}g_{\nu_R}T^3. \tag*{(30)}

Here, ζ\zeta is the Riemann zeta function and gνR=2g_{\nu_R}=2 counts one helicity state for the neutrino and one for its antiparticle. Consequently, nνiReq=nνˉiReq=nνReq/2n_{\nu_{iR}}^{\mathrm{eq}}=n_{\bar{\nu}_{iR}}^{\mathrm{eq}}=n_{\nu_R}^{\mathrm{eq}}/2.

Using Fermi–Dirac distributions for both incoming particles and the Møller flux factor vMv_{\mathrm{M}} in the plasma rest frame [65, 66], we write the thermal average as

⟨σijvM⟩=4[nνReq(T)]2∫d3p1(2π)3d3p2(2π)3fFD(E1,T)fFD(E2,T)σij(s)vM.(31)\langle\sigma_{ij}v_{\mathrm{M}}\rangle=\frac{4}{[n_{\nu_R}^{\mathrm{eq}}(T)]^2}\int\frac{d^3p_1}{(2\pi)^3}\frac{d^3p_2}{(2\pi)^3} f_{\mathrm{FD}}(E_1,T)f_{\mathrm{FD}}(E_2,T)\sigma_{ij}(s)v_{\mathrm{M}}. \tag*{(31)}

Here, fFD(E,T)=(eE/T+1)−1f_{\mathrm{FD}}(E,T)=(e^{E/T}+1)^{-1}, Ea=∣pa∣E_a=|\mathbf{p}_a|, and vM=1−p^1⋅p^2v_{\mathrm{M}}=1-\hat{\mathbf{p}}_1\cdot\hat{\mathbf{p}}_2. The momenta p1p_1 and p2p_2 correspond to the incoming neutrino and antineutrino, respectively. We neglect final-state Pauli blocking, so that the vacuum cross section in Eq. (25) can be used.

The annihilation rate per νiR\nu_{iR} is obtained by summing over the antineutrino species,

Γi(T)=nνReq(T)2∑j⟨σijvM⟩.(32)\Gamma_i(T)=\frac{n_{\nu_R}^{\mathrm{eq}}(T)}{2}\sum_j\langle\sigma_{ij}v_{\mathrm{M}}\rangle. \tag*{(32)}

We compare this rate with the expansion rate during radiation domination, where g∗(T)g_*(T) counts the effective energy degrees of freedom,

H(T)=1.66g∗(T)T2MPl.(33)H(T)=1.66\sqrt{g_*(T)}\frac{T^2}{M_{\mathrm{Pl}}}. \tag*{(33)}

Here MPl=GN−1/2≃1.22×1019 GeVM_{\mathrm{Pl}}=G_N^{-1/2}\simeq1.22\times10^{19}\ \mathrm{GeV} is the unreduced Planck mass, with GNG_N the gravitational constant. For both g∗(T)g_*(T) and the entropy degrees of freedom g∗s(T)g_{*s}(T) below, we use Table A1 of Ref. [67], with its TQCD=214 MeVT_{\mathrm{QCD}}=214\ \mathrm{MeV} prescription. We interpolate linearly in ln⁡T\ln T within each phase, retaining the tabulated entropy jump at the transition; both functions approach 106.75 at high temperature.

For species that reach thermal equilibrium, we use the instantaneous decoupling approximation and compute the subsequent entropy dilution following Ref. [68]. We estimate the decoupling temperature of νiR\nu_{iR} from

Γi(Tdec,i)=H(Tdec,i).(34)\Gamma_i(T_{\mathrm{dec},i}) = H(T_{\mathrm{dec},i}). \tag*{(34)}

We select the low-temperature crossing at which Γi/H\Gamma_i/H falls below unity as the Universe cools. Because the full cross section depends nontrivially on ss, this equation is solved numerically. We assume a negligible initial abundance. For species that never reach thermal equilibrium over the assumed thermal history, we neglect production by freeze-in [60]. The entropy-dilution formula below applies to the species that did thermalize.

The tree-level thermal rates depend only on YRY^R, the scalar masses, and the mixing angle θ\theta. Figure 8 illustrates the resulting evolution of Γi/H\Gamma_i/H for two representative scalar spectra; the relevant decoupling temperatures are the low-temperature crossings of unity.

Figure 8. Annihilation rates Γi/H\Gamma_i/H as functions of inverse temperature for YR=diag⁡(1,1,1)Y^R = \operatorname{diag}(1,1,1) and sin⁡θ=0.5\sin\theta= 0.5. Solid curves correspond to (mS1,mS2)=(1,1.5) TeV(m_{S_1},m_{S_2}) = (1,1.5)\,\mathrm{TeV} and dashed curves to (10,15) TeV(10,15)\,\mathrm{TeV}. Markers identify the decoupling crossings Γi/H=1\Gamma_i/H = 1.

Entropy conservation after instantaneous decoupling fixes the temperature of each right-handed species relative to that of the active neutrinos:

TνiRTνL=[g∗s(TdecνL)g∗s(Tdec,i)]1/3.(35)\frac{T_{\nu_{iR}}}{T_{\nu_L}} = \left[\frac{g_{*s}(T_{\mathrm{dec}}^{\nu_L})}{g_{*s}(T_{\mathrm{dec},i})}\right]^{1/3}. \tag*{(35)}

Here TdecνLT_{\mathrm{dec}}^{\nu_L} denotes active-neutrino decoupling. Taking g∗s(TdecνL)=10.75g_{*s}(T_{\mathrm{dec}}^{\nu_L}) = 10.75, before e±e^\pm annihila tion, the νiR,νˉiR\nu_{iR},\bar{\nu}_{iR} pair contributes

Figure 9. Individual and total ΔNeff\Delta N_{\rm eff} contributions as functions of mS1m_{S_1} for mS2=15 TeVm_{S_2}=15\,\mathrm{TeV}, sin⁡θ=0.5\sin\theta=0.5, and YR=diag⁡(1,1,1)Y^R=\operatorname{diag}(1,1,1). The horizontal dotted line shows the adopted Planck 2018+BAO 95% bound; the vertical dashed line marks the resulting lower bound on mS1m_{S_1}.

ΔNeff,i=(TνiRTνL)4=[10.75g∗s(Tdec,i)]4/3.(36)\Delta N_{\mathrm{eff},i}=\left(\frac{T_{\nu_{iR}}}{T_{\nu_L}}\right)^4=\left[\frac{10.75}{g_{*s}(T_{\mathrm{dec},i})}\right]^{4/3}. \tag*{(36)}

Since the three generations can decouple at different temperatures, their contributions add independently:

ΔNeff=∑i=13ΔNeff,i.(37)\Delta N_{\mathrm{eff}}=\sum_{i=1}^{3}\Delta N_{\mathrm{eff},i}. \tag*{(37)}

Figure 9 shows the individual and total contributions for a representative one-dimensional mass scan, together with the adopted cosmological bound.

Figure 10 translates the same bound into the scalar-mass plane for three mixing choices. At maximal mixing, the equal scalar weights produce an exclusion contour symmetric under mS1↔mS2m_{S_1}\leftrightarrow m_{S_2}. The endpoint curves at sin⁡θ=0\sin\theta=0 and 11 are thermal limits; since sin⁡(2θ)=0\sin(2\theta)=0, they do not generate radiative neutrino masses.

For equal scalar masses at maximal mixing, Figure 10 gives mS1=mS2≳5.4 TeVm_{S_1}=m_{S_2}\gtrsim5.4\,\mathrm{TeV} for YR=I3Y^R=I_3. A direct calculation with the same thermal treatment and YR=4πI3Y^R=\sqrt{4\pi}I_3, near the perturbative limit, raises this lower bound to approximately 19.3 TeV. For this particular configuration, the cosmological bound lies more than an order of magnitude above the characteristic O(1 TeV)O(1\,\mathrm{TeV}) limits from current direct searches, illustrating the potential strength of NeffN_{\mathrm{eff}}. This comparison is indicative, since a quantitative LHC limit requires a dedicated recast of the mixed-leptoquark spectrum [45, 46].

Figure 10. Planck 2018+BAO 95% exclusion regions in the (mS1,mS2)(m_{S_1},m_{S_2}) plane for YR=diag⁡(1,1,1)Y^R=\operatorname{diag}(1,1,1) and sin⁡θ=0\sin\theta=0, 1/21/\sqrt{2}, 11. The curves mark the adopted ΔNeff\Delta N_{\mathrm{eff}} bound and shaded regions are excluded. At maximal mixing, the contour is symmetric under mS1↔mS2m_{S_1}\leftrightarrow m_{S_2}.

Numerical analysis and discussion

We now combine the neutrino-mass, meson-decay, charged-lepton flavor, and cosmological results to identify viable Yukawa textures. Figure 11 shows how the constraints accumulate at fixed scalar spectra, while Figure 12 displays the resulting reach in the scalar mass plane.

Figure 12. Search for textures satisfying all four conditions on a 31×3131 \times31 logarithmic mass grid from 1 to 100 TeV. At each mass pair, white indicates that no perturbative neutrino-mass texture can be reconstructed, light gray that a perturbative texture can be reconstructed but no texture satisfying all constraints was found, and green that at least one texture satisfying all constraints was found.

Physical inputs and constraints

We adopt the restricted flavor texture and neutrino-mass reconstruction defined in Eqs. (15) and (16). Thus, at each scalar spectrum and mixing angle, only the positive parameters y2=Y22Ry_2 = Y^{R}_{22} and y3=Y33Ry_3 = Y^{R}_{33} are varied, while ζ\zeta and YL=V∗ζY^{L} = V^*\zeta are reconstructed using the exact loop function in Eq. (8). Neutrino masses and mixing are therefore imposed algebraically. The elementwise perturbativity condition in Eq. (17) is applied to YRY^{R}, ζ\zeta, and YLY^{L}, retaining the CKM-induced correlations in Eq. (16).

For reproducibility, we use the NuFIT 6.0 inputs collected in Eq. (13). The PMNS and CKM phases are retained in Eq. (16); oscillation and form-factor uncertainties are not scanned.

The objective function is based on the inclusive B+→K+ννˉB^+ \to K^+\nu\bar{\nu} branching fraction defined in Eq. (21), with the differential rate given in Eq. (38). We adopt the combined Belle II measurement and its 1σ1\sigma interval [37], [1.6596876,3.0071068]×10−5[1.6596876, 3.0071068] \times10^{-5}, obtained by combining statistical and systematic errors in quadrature on each side. The objective vanishes inside this interval and penalizes the squared departure outside it. Two additional requirements are imposed: B(μ→eγ)≤1.5×10−13\mathcal{B}(\mu\to e\gamma) \leq1.5 \times10^{-13}, from § 0.5, Eqs. (23) and (24) [58], and ΔNeff≤0.286\Delta N_{\mathrm{eff}} \leq0.286, evaluated according to Sec. VI, Eq. (37) [61]. These criteria use their respective experimental confidence levels and do not define a joint confidence region.

Rare top decays can also be induced in this model but are not included among the numerical constraints because their rates are negligible for the Y~R=0\widetilde{Y}^{R} = 0 benchmark adopted here; see Appendix § 0.7.2. A nonzero Y~R\widetilde{Y}^{R} would modify both the CLFV and rare-top phenomenology and require a separate multidimensional analysis.

Freeze-in production is not included. In particular, ν1R\nu_{1R} has no thermal contribution for the adopted Y11R=0Y^{R}_{11} = 0 texture.

Combined constraints and scalar mass reach

Figure 11 shows the cumulative effect of perturbativity, Belle II, MEG II, and ΔNeff\Delta N_{\mathrm{eff}} for three scalar spectra and mixing choices. For each fixed scalar spectrum and mixing angle, we evaluate a 201×201201 \times201 logarithmic grid in (y2,y3)(y_2,y_3), supplemented by 401 points resolving the Belle II contours. The four conditions are then applied cumulatively at each sampled point.

The Belle II bands contain two regimes. For larger y2y_2 and y3y_3, the reconstructed ζ\zeta and YLY^L entries are suppressed, and the RR vector contribution dominates. A strong hierarchy between y2y_2 and y3y_3 instead enhances one row of ζ\zeta and can produce a mixed-chirality scalar– tensor solution.7 MEG II preferentially removes the latter regime, whereas ΔNeff\Delta N_{\rm eff} constrains larger right-handed Yukawas. Both mechanisms survive for the (1,1.5)(1, 1.5) TeV spectrum at maximal mixing. As the mixing approaches its endpoints, however, the radiative mass relation requires larger Yukawa products, while the scalar weights increasingly select different mass eigenstates. For (mS1,mS2)=(10,15)(m_{S_1},m_{S_2}) = (10, 15) TeV, no sampled point survives MEG II at sin⁡θ=10−3\sin\theta= 10^{-3}, while RR-vector solutions remain at maximal mixing and for cos⁡θ=10−3\cos\theta= 10^{-3}.

Figure 12 extends the search to independently varied scalar masses. On a 31×3131 \times31 logarithmic mass grid, we scan the positive Yukawa parameters y2,y3>0y_2, y_3 > 0 for each mass pair and mixing angle, refining the search near Belle II-compatible regions. As both scalar masses increase, propagator suppression drives the branching fraction back toward its SM value, making it too small to reproduce the Belle II measurement with perturbative Yukawa couplings. Consequently, no complete solution is found once both masses exceed approximately 12–14 TeV, although one mass can reach 100 TeV if the other remains light. The endpoint panels are approximately transposed because the scalar weights interchange. The lower explored mass of 1 TeV is motivated by current collider bounds on this scale, although a dedicated recast is required for the present two-leptoquark framework [45, 46].

Table II makes the interplay among the constraints explicit through four maximal-mixing textures selected from Figure 11. Benchmark A is RR-vector dominated and has negligible flavor violation, whereas B realizes the mixed-chirality regime, with larger left-handed couplings and a μ→eγ\mu\to e\gamma rate near the MEG II bound. Benchmark C is again RR-vector dominated but instead lies close to the cosmological limit. Finally, D accommodates the Belle II enhancement at (mS1,mS2)=(10,15)(m_{S_1},m_{S_2}) = (10,15) TeV at the cost of right-handed Yukawa couplings of order three. All four reproduce the same neutrino spectrum and mixing parameters, while their meson, flavor, and cosmological predictions highlight the complementarity of the constraints. We emphasize that, for each benchmark, YRY^{R} is diagonal, with the two positive nonzero entries listed in Table II; the matrices ζ\zeta and YL=V∗ζY^{L} = V^{*}\zeta then follow uniquely from the reconstruction in Eq. (16).

Conclusions

We have studied a simple leptoquark model connecting radiative Dirac neutrino masses with the enhancement of B+→K+ννˉB^{+} \to K^{+}\nu\bar{\nu}. The scalar mixing needed for the one-loop mass mechanism also links the left- and right-handed interactions entering the meson decay. The renormalizable model conserves baryon and lepton numbers, preserving proton stability and the Dirac nature of neutrinos after soft breaking of the discrete symmetries.

Within the adopted flavor ansatz, the neutrino-mass reconstruction leaves two right-handed Yukawa entries free. Their allowed values are restricted jointly by perturbativity, the Belle II branching fraction interval, μ→eγ\mu\to e\gamma, and ΔNeff\Delta N_{\mathrm{eff}}. Figure 11 and Figure 12 show that compatible textures occur in both the RR-vector-dominated and mixed-chirality scalar–tensor regimes. The latter can approach the charged-lepton flavor bound through larger reconstructed left-handed couplings, while RR-vector solutions can approach the cosmological bound with negligible flavor violation. We provide four benchmarks that make this complementarity explicit. Under the baseline assumption Y~R=0\widetilde{Y}^{R} = 0, rare top decays remain suppressed for the flavor structure considered and do not supply an additional restriction in the numerical study.

Within the adopted texture and scan, no complete solution is found once both scalar masses exceed approximately 12–14 TeV, although one scalar can be much heavier if the other remains sufficiently light. Direct collider searches provide the complementary restriction at low masses, but existing LHC limits cannot be applied quantitatively without a dedicated recast. The two mixed scalar leptoquarks originate from fields with identical gauge quantum numbers but distinct Yukawa structures, leading to flavor-dependent decay patterns [45, 46, 69]. Consequently, the lower mass reach is not universal. Nevertheless, ΔNeff\Delta N_{\mathrm{eff}} provides a powerful probe of leptoquark scenarios with radiative Dirac neutrino masses and, for some benchmark configurations, can surpass the nominal reach of current direct searches.

Parameter or observableABCD
mS1m_{S_1} [TeV]11110
mS2m_{S_2} [TeV]1.51.51.515
sin⁡θ\sin\theta1/21/\sqrt{2}1/21/\sqrt{2}1/21/\sqrt{2}1/21/\sqrt{2}
∣Y22R∣\lvert Y^{R}_{22}\rvert0.09812.102×10−72.102 \times10^{-7}0.27472.7119
∣Y33R∣\lvert Y^{R}_{33}\rvert1.03001.95580.42963.2445
max⁡ij∣YijL∣\max_{ij}\lvert Y^{L}_{ij}\rvert7.812×10−87.812 \times10^{-8}0.03652.799×10−82.799 \times10^{-8}2.839×10−92.839 \times10^{-9}
105B(B+→K+ννˉ)10^{5}\mathcal{B}(B^{+}\to K^{+}\nu\bar{\nu})2.33341.65973.00711.8843
B(μ→eγ)\mathcal{B}(\mu\to e\gamma)3.038×10−363.038 \times10^{-36}1.444×10−131.444 \times10^{-13}4.930×10−384.930 \times10^{-38}5.180×10−465.180 \times10^{-46}
ΔNeff\Delta N_{\mathrm{eff}}0.17080.08700.27850.2606
RR vector share [%]79.43<10−8<10^{-8}84.0474.53
Scalar + tensor share [%]<10−8<10^{-8}71.09<10−8<10^{-8}<10−8<10^{-8}

Table 2. Four illustrative textures from the maximal-mixing panels of Figure 11, with Y~R=0\widetilde{Y}^{R}=0. All satisfy the adopted Belle II, MEG II, and cosmological conditions and the perturbativity criterion in Eq. (17). The neutrino masses are the same for all four: m1=0m_{1}=0, m2=8.654×10−3 eVm_{2}=8.654\times10^{-3}\,\mathrm{eV}, and m3=5.013×10−2 eVm_{3}=5.013\times10^{-2}\,\mathrm{eV}, with the oscillation inputs specified in the text. Operator shares refer to the total meson branching fraction in Eq. (38), including the SM contribution. Values are rounded; acceptance is checked with the stored full-precision Yukawa matrices. Collider limits are not imposed in this table.

Acknowledgments

P. E. acknowledges support from CNPq under grant No. 151612/2024-2 and from the Fundação de Amparo e Promoção da Ciência, Tecnologia e Inovação do Rio Grande do Norte (FAPERN) under Call No. 06/2026–FAPERN/IIP, process No. 10910022.000493/2024-04. G. G. acknowledges support from the National Council for Science, Technology and Technological Innovation (CONCYTEC) and the National Program for Scientific Research and Advanced Studies (PROCIENCIA) under the call “E041-2026-01 Proyectos de Investigación Básica” (contract No. PE5010103523-2026), and thanks the International Institute of Physics (IIP-UFRN) in Natal for its kind hospitality, financial support, and stimulating research environment during the development of this work. V. P. acknowledges financial support from the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq). This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) – Finance Code 001. C. A. S. P. thanks CNPq for financial support under grant No. 301734/2026-7. F. S. Q. is supported by the Humboldt Foundation (Friedrich Wilhelm Bessel Research Award), ANID – Millennium Science Initiative Program – ICN2019_044, IIF-FINEP grant 213/2024, FAPESP grants 2021/14335-0 and 2021/00449-4, and CNPq grants 307130/2021-5, 403521/2024-6, 309465/2026-5, and 446711/2025-0.

Meson decay details

Differential rate and tree-level matching

In this appendix, we detail the calculation of the inclusive branching fraction shown in Eq. (21) of § 0.4. Neglecting neutrino masses in the decay kinematics, the differential rate for the dimension-six vector, scalar, and tensor operators is given by [70]

dΓdq2=GF2∣VtbVts∗∣2αem23072π5mB3qλK1/2(q2)∑i,j=13{(∣CLLSMδij+[CLLV]ij+[CRLV]ij∣2+∣[CLRV]ij+[CRRV]ij∣2)HV2+32(∣[CRLS]ij+[CLLS]ij∣2+∣[CRRS]ij+[CLRS]ij∣2)HS2+8(∣[CLLT]ij∣2+∣[CRRT]ij∣2)HT2},(38)\begin{aligned} \frac{d\Gamma}{dq^2}={}&\frac{G_F^2\lvert V_{tb}V_{ts}^{*}\rvert^2\alpha_{\mathrm{em}}^2}{3072\pi^5m_B^3}q\lambda_K^{1/2}(q^2)\sum_{i,j=1}^{3}\Bigg\{\left(\left\lvert C_{LL}^{\mathrm{SM}}\delta_{ij}+[C_{LL}^{V}]_{ij}+[C_{RL}^{V}]_{ij}\right\rvert^2+\left\lvert[C_{LR}^{V}]_{ij}+[C_{RR}^{V}]_{ij}\right\rvert^2\right)H_V^2\\ &+\frac{3}{2}\left(\left\lvert[C_{RL}^{S}]_{ij}+[C_{LL}^{S}]_{ij}\right\rvert^2+\left\lvert[C_{RR}^{S}]_{ij}+[C_{LR}^{S}]_{ij}\right\rvert^2\right)H_S^2\\ &+8\left(\left\lvert[C_{LL}^{T}]_{ij}\right\rvert^2+\left\lvert[C_{RR}^{T}]_{ij}\right\rvert^2\right)H_T^2\Bigg\}, \tag*{(38)} \end{aligned}

where GFG_F is the Fermi constant, αem\alpha_{\mathrm{em}} is the electromagnetic fine-structure constant, mBm_B is the charged-BB-meson mass, and Vqq′V_{qq'} are CKM matrix elements. The dimensionless SM Wilson coefficient is CLLSM=−2Xt/sin⁡2θWC^{\mathrm{SM}}_{LL}=-2X_t/\sin^2\theta_W, for which we use Xt=1.469X_t=1.469, sin⁡2θW=0.23121\sin^2\theta_W=0.23121, and αem−1(MZ)=127.952\alpha_{\mathrm{em}}^{-1}(M_Z)=127.952 [53]. The CKM elements are evaluated using the standard-parametrization inputs specified in § 0.2.

At tree level, the nonzero leptoquark coefficients are CLL,aV,S,TC^{V,S,T}_{LL,a} and CRR,aV,S,TC^{V,S,T}_{RR,a} for each physical scalar SaS_a (a=1,2a=1,2). At the matching scale μ=mSa\mu=m_{S_a} they read

[CLL,aV(mSa)]ij=KB(YL(a))bj(YL(a))si∗mSa2,[CRR,aV(mSa)]ij=KB(YR(a))bj(YR(a))si∗mSa2,(39)\left[C^V_{LL,a}(m_{S_a})\right]_{ij}=\mathcal{K}_B\frac{\left(\mathcal{Y}^{(a)}_L\right)_{bj}\left(\mathcal{Y}^{(a)}_L\right)^*_{si}}{m_{S_a}^2}, \qquad\left[C^V_{RR,a}(m_{S_a})\right]_{ij}=\mathcal{K}_B\frac{\left(\mathcal{Y}^{(a)}_R\right)_{bj}\left(\mathcal{Y}^{(a)}_R\right)^*_{si}}{m_{S_a}^2}, \tag*{(39)}
[CLL,aS(mSa)]ij=KB(YL(a))bj(YR(a))si∗mSa2,[CRR,aS(mSa)]ij=KB(YR(a))bj(YL(a))si∗mSa2,(40)\left[C^S_{LL,a}(m_{S_a})\right]_{ij}=\mathcal{K}_B\frac{\left(\mathcal{Y}^{(a)}_L\right)_{bj}\left(\mathcal{Y}^{(a)}_R\right)^*_{si}}{m_{S_a}^2}, \qquad\left[C^S_{RR,a}(m_{S_a})\right]_{ij}=\mathcal{K}_B\frac{\left(\mathcal{Y}^{(a)}_R\right)_{bj}\left(\mathcal{Y}^{(a)}_L\right)^*_{si}}{m_{S_a}^2}, \tag*{(40)}
[CLL,aT(mSa)]ij=−14[CLL,aS(mSa)]ij,[CRR,aT(mSa)]ij=−14[CRR,aS(mSa)]ij,(41)\left[C^T_{LL,a}(m_{S_a})\right]_{ij}=-\frac{1}{4}\left[C^S_{LL,a}(m_{S_a})\right]_{ij}, \qquad\left[C^T_{RR,a}(m_{S_a})\right]_{ij}=-\frac{1}{4}\left[C^S_{RR,a}(m_{S_a})\right]_{ij}, \tag*{(41)}

where KB≡πv2/(αemVtbVts∗)\mathcal{K}_B\equiv\pi v^2/(\alpha_{\mathrm{em}}V_{tb}V^*_{ts}) and the physical Yukawa couplings YL,R(a)\mathcal{Y}^{(a)}_{L,R} are defined in Eq. (22).

QCD evolution

The Wilson coefficients entering Eq. (38) must be evaluated at a common low-energy scale. We choose μb=4.8 GeV\mu_b=4.8\ \mathrm{GeV}, of the order of the bottom-quark mass and equal to the scale at which the tensor form factor fT(q2,μ)f_T(q^2,\mu) is provided in Ref. [52]. This choice avoids an additional evolution of fTf_T. Accordingly, the scalar and tensor Wilson coefficients and the running quark masses entering the scalar matrix element are also evaluated at μb=4.8 GeV\mu_b=4.8\ \mathrm{GeV}. Since the two leptoquarks can have different masses, each contribution must first be evolved from its own matching scale and only then added:

[CAAX(μb)]ij=∑a=12ΩX(μb,mSa)[CAA,aX(mSa)]ij,X=V,S,T,A=L,R.(42)\left[C^X_{AA}(\mu_b)\right]_{ij}=\sum_{a=1}^{2}\Omega_X(\mu_b,m_{S_a})\left[C^X_{AA,a}(m_{S_a})\right]_{ij}, \qquad X=V,S,T,\quad A=L,R. \tag*{(42)}

At leading-logarithmic order, for mSa>mtm_{S_a}>m_t, the QCD evolution factor is [70]

ΩX(μb,mSa)=[αs(μb)αs(mt)]−γX/β1(5)[αs(mt)αs(mSa)]−γX/β1(6),β1(nf)=2nf−336,(43)\Omega_X(\mu_b,m_{S_a})=\left[\frac{\alpha_s(\mu_b)}{\alpha_s(m_t)}\right]^{-\gamma_X/\beta^{(5)}_1}\left[\frac{\alpha_s(m_t)}{\alpha_s(m_{S_a})}\right]^{-\gamma_X/\beta^{(6)}_1}, \qquad\beta^{(n_f)}_1=\frac{2n_f-33}{6}, \tag*{(43)}

where nfn_f denotes the number of active quark flavors at the corresponding renormalization scale, and

γV=0,γS=2,γT=−23.(44)\gamma_V = 0,\qquad\gamma_S = 2,\qquad\gamma_T = -\frac{2}{3}. \tag*{(44)}

Consequently, the vector coefficients do not run, ΩV=1\Omega_V = 1, whereas the scalar and tensor coefficients evolve as

ΩS(μb,M)=[αs(μb)αs(mt)]12/23[αs(mt)αs(M)]4/7,ΩT(μb,M)=[αs(μb)αs(mt)]−4/23[αs(mt)αs(M)]−4/21.(45)\begin{aligned} \Omega_S(\mu_b,M) &= \left[\frac{\alpha_s(\mu_b)}{\alpha_s(m_t)}\right]^{12/23}\left[\frac{\alpha_s(m_t)}{\alpha_s(M)}\right]^{4/7},\\ \Omega_T(\mu_b,M) &= \left[\frac{\alpha_s(\mu_b)}{\alpha_s(m_t)}\right]^{-4/23}\left[\frac{\alpha_s(m_t)}{\alpha_s(M)}\right]^{-4/21}. \tag*{(45)} \end{aligned}

The matching relation CAA,aT(mSa)=−CAA,aS(mSa)/4C_{AA,a}^{T}(m_{S_a})=-C_{AA,a}^{S}(m_{S_a})/4 holds at each leptoquark matching scale. It does not hold at μb\mu_b, because the scalar and tensor operators have different QCD anomalous dimensions.

Helicity amplitudes and form factors

Neutrino masses are retained in the reconstruction of ζ\zeta but neglected in the decay kinematics. The Källén function is

λK(q2)=mB4+mK4+(q2)2−2mB2mK2−2mK2q2−2mB2q2,(46)\lambda_K(q^2) = m_B^4 + m_K^4 + (q^2)^2 - 2m_B^2m_K^2 - 2m_K^2q^2 - 2m_B^2q^2, \tag*{(46)}

and

HV(q2)=λK(q2)q2 f+(q2),HS(q2)=mB2−mK2mb(μb)−ms(μb)f0(q2),HT(q2)=−λK(q2)mB+mKfT(q2,μb).(47)\begin{aligned} H_V(q^2) &= \sqrt{\frac{\lambda_K(q^2)}{q^2}}\,f_+(q^2),\\ H_S(q^2) &= \frac{m_B^2-m_K^2}{m_b(\mu_b)-m_s(\mu_b)}f_0(q^2),\\ H_T(q^2) &= -\frac{\sqrt{\lambda_K(q^2)}}{m_B+m_K}f_T(q^2,\mu_b). \tag*{(47)} \end{aligned}

The lattice-QCD results for the form factors in Ref. [52] are fitted using the Bourrely–Caprini–Lellouch (BCL) parametrization [71]

f0(q2)=LB1−q2/M02∑n=0N−1an0z(q2)n,f_0(q^2) = \frac{\mathcal{L}_B}{1-q^2/M_0^2}\sum_{n=0}^{N-1}a_n^0 z(q^2)^n,
fi(q2)=LB1−q2/M∗2∑n=0N−1ani[z(q2)n−(−1)n−NnNz(q2)N],i=+,T.(48)f_i(q^2) = \frac{\mathcal{L}_B}{1-q^2/M_*^2}\sum_{n=0}^{N-1}a_n^i\left[z(q^2)^n-(-1)^{n-N}\frac{n}{N}z(q^2)^N\right],\qquad i=+,T. \tag*{(48)}

For N=3N = 3, t0=0t_0 = 0, and t+=(mB+mK)2t_+ = (m_B + m_K)^2, the conformal variable is

z(q2)=t+−q2−t+−t0t+−q2+t+−t0.(49)z(q^2) = \frac{\sqrt{t_+ - q^2} - \sqrt{t_+ - t_0}}{\sqrt{t_+ - q^2} + \sqrt{t_+ - t_0}}. \tag*{(49)}

The coefficients used in the calculation are listed in Table III, while the remaining BCL parameters are given in Eq. (50).

nnan0a_n^0an+a_n^+anTa_n^T
00.2546162730.2546162730.254907841
10.211016714-0.708471001-0.654938491
20.0269077670.3096901520.362659041

Table III. Central coefficients of the N=3N = 3 BCL expansion used for the B→KB \to K form factors.

LB=1.303655672,M0=5.729495108 GeV,M∗=5.415766152 GeV.(50)\mathcal{L}_B = 1.303655672,\qquad M_0 = 5.729495108\ \mathrm{GeV},\qquad M_* = 5.415766152\ \mathrm{GeV}. \tag*{(50)}

The tensor coefficients correspond to fT(q2,μb)f_T(q^2,\mu_b) at μb=4.8 GeV\mu_b = 4.8\ \mathrm{GeV}. The phenomenological use of these form factors in B→Kℓ+ℓ−B \to K\ell^+\ell^- and B→KννˉB \to K\nu\bar{\nu} is discussed in Ref. [53].

Rare top decays

The charged-lepton interactions of the model can induce rare top decays. Here we estimate their rates and show why they are negligible for the Y~R=0\widetilde{Y}_R = 0 benchmark adopted in the numerical analysis. The tree-level topology for the three-body decay and the representative one-loop topologies for the two-body channels are shown in Figure 13.

Figure 13. Leptoquark contributions to rare top decays. The left panel shows the tree-level exchange inducing t→ceα+eβ−t \to c e^{+}_{\alpha}e^{-}_{\beta}. The center and right panels show representative one-loop diagrams for t→cXt \to cX, with the gauge boson coupled to the leptoquark and the charged lepton, respectively. Here SaS_a denotes either physical leptoquark (a=1,2a = 1,2); X=γ,g,ZX = \gamma,g,Z in the center panel, while X=γ,ZX = \gamma,Z in the right panel.

The charged-lepton interaction in Eq. (3) induces t→ceα+eβ−t \to ce^+_\alpha e^-_\beta at tree level. For the numerical estimates below, we set Y~R=0\widetilde{Y}_R = 0, consistently with the main analysis. For mS1,mS2≫mtm_{S_1},m_{S_2} \gg m_t and neglecting the final-state fermion masses, its branching fraction is approximately

Bαβ≡B(t→ceα+eβ−)≃mt56144π3Γt∣Y3αLY2βL∗∣2(cos⁡2θmS12+sin⁡2θmS22)2,(51)\mathcal{B}_{\alpha\beta} \equiv\mathcal{B}(t \to ce^+_\alpha e^-_\beta) \simeq\frac{m_t^5}{6144\pi^3\Gamma_t}\left|Y^L_{3\alpha}Y^{L*}_{2\beta}\right|^2\left(\frac{\cos^2\theta}{m_{S_1}^2}+\frac{\sin^2\theta}{m_{S_2}^2}\right)^2, \tag*{(51)}

where mt≃173 GeVm_t \simeq173\ \mathrm{GeV} and Γt≃1.4 GeV\Gamma_t \simeq1.4\ \mathrm{GeV} [6]. The suppression is particularly strong when YLY^L is reconstructed from neutrino masses using Eq. (12). For example, mS1=1 TeVm_{S_1} = 1\ \mathrm{TeV},

mS2=1.5 TeVm_{S_2} = 1.5\ \mathrm{TeV}, sin⁡θ=1/2\sin\theta= 1/\sqrt{2} and YR=diag⁡(0,1,1)Y^{R} = \operatorname{diag}(0,1,1) give max⁡α∣Y2αL∣=7.6×10−9\max_{\alpha}|Y^{L}_{2\alpha}| = 7.6 \times10^{-9} and max⁡α∣Y3αL∣=1.4×10−9\max_{\alpha}|Y^{L}_{3\alpha}| = 1.4 \times10^{-9} for normal ordering with m1=0m_1 = 0. The largest individual three-body branching fraction is then 3.3×10−413.3 \times10^{-41}, far below experimental sensitivities of order 10−710^{-7}–10−610^{-6} [72].

The two-body channels t→cXt \to cX, with X=γ,g,ZX = \gamma,g,Z, arise at one loop and are further suppressed, as illustrated in the center and right panels of Figure 13. For Y~R=0\widetilde{Y}^{R} = 0, their leading dipole amplitudes can be written as

MX=uˉciσμνqν(ALXPL+ARXPR)utϵμ∗,ARX≃gXmt16π2∑αY3αLY2αL∗[cθ2mS12fX(meα2mS12)+sθ2mS22fX(meα2mS22)],ALX=O(mcmt)ARX,(52)\begin{aligned} \mathcal{M}_{X} &= \bar{u}_{c}i\sigma^{\mu\nu}q_{\nu}\left(A^{X}_{L}P_{L}+A^{X}_{R}P_{R}\right)u_{t}\epsilon^{*}_{\mu},\\ A^{X}_{R} &\simeq\frac{g_{X}m_{t}}{16\pi^{2}}\sum_{\alpha}Y^{L}_{3\alpha}Y^{L*}_{2\alpha}\left[\frac{c_{\theta}^{2}}{m_{S_{1}}^{2}}f_{X}\left(\frac{m_{e_{\alpha}}^{2}}{m_{S_{1}}^{2}}\right)+\frac{s_{\theta}^{2}}{m_{S_{2}}^{2}}f_{X}\left(\frac{m_{e_{\alpha}}^{2}}{m_{S_{2}}^{2}}\right)\right],\\ A^{X}_{L} &= \mathcal{O}\left(\frac{m_{c}}{m_{t}}\right)A^{X}_{R}, \tag*{(52)} \end{aligned}

where gγ=eg_{\gamma} = e, gg=gsg_{g} = g_{s}, gZ=g/cWg_{Z} = g/c_{W}, and the dimensionless function fXf_{X} includes the diagrams with XX attached to the scalar and, for X=γ,ZX = \gamma,Z, to the charged lepton. They are of order unity in the heavy-leptoquark limit. The corresponding dipole contribution is

Γ(t→cX)=κXmt316π(1−mX2mt2)3(∣ALX∣2+∣ARX∣2),κγ=κZ=1,κg=CF=43.(53)\Gamma(t \to cX) = \frac{\kappa_{X}m_{t}^{3}}{16\pi}\left(1-\frac{m_{X}^{2}}{m_{t}^{2}}\right)^{3}\left(|A^{X}_{L}|^{2}+|A^{X}_{R}|^{2}\right),\qquad\kappa_{\gamma}=\kappa_{Z}=1,\qquad\kappa_{g}=C_{F}=\frac{4}{3}. \tag*{(53)}

With the reconstruction in § 0.2, ζζ†\zeta\zeta^{\dagger} is diagonal for the restricted texture, but

YLYL†=V∗(ζζ†)VT,(54)Y^{L}Y^{L\dagger}=V^{*}(\zeta\zeta^{\dagger})V^{T}, \tag*{(54)}

is not diagonal in general. Therefore, the mass-independent flavor factor ∑αY3αLY2αL∗=(YLYL†)32\sum_\alpha Y^L_{3\alpha}Y^{L*}_{2\alpha} = (Y^L Y^{L\dagger})_{32} does not vanish exactly. For the benchmark above, its modulus is only 5.8×10−185.8 \times10^{-18}, leading to two-body branching fractions below approximately 10−4110^{-41} for order-one loop functions. Both the tree-level and one-loop rates are therefore negligible, and rare top decays are not included as constraints in the numerical analysis with Y~R=0\widetilde{Y}^R = 0.

If the couplings in Y~R\widetilde{Y}^R are allowed to be of order unity, rare top decays can become moderately relevant for a light leptoquark spectrum. For example, with the diagonal ansatz Y~R=diag⁡(y~u,y~c,y~t)\widetilde{Y}^R = \operatorname{diag}(\widetilde{y}_u,\widetilde{y}_c,\widetilde{y}_t), maximal mixing, and (mS1,mS2)=(1,1.5) TeV(m_{S_1},m_{S_2}) = (1,1.5)\ \mathrm{TeV}, the dominant channel gives

B(t→cτ+μ−)≃2.97×10−7∣y~ty~c∣2,(55)\mathcal{B}(t \to c\tau^{+}\mu^{-}) \simeq2.97 \times10^{-7} \lvert\widetilde{y}_t\widetilde{y}_c\rvert^2, \tag*{(55)}

which is below, but within the same order of magnitude as, the ATLAS bound B(t→μτq)<8.7×10−7\mathcal{B}(t \to\mu\tau q) < 8.7 \times10^{-7} at 95% C.L. [72]. Since the rate scales approximately as mS−4m_S^{-4}, the same prediction falls to 2.97×10−11∣y~ty~c∣22.97 \times10^{-11}\lvert\widetilde{y}_t\widetilde{y}_c\rvert^2 for (mS1,mS2)=(10,15) TeV(m_{S_1},m_{S_2}) = (10,15)\ \mathrm{TeV}.

The loop factor additionally suppresses the two-body channels. Moreover, for the diagonal ansatz the pure right-handed flavor factor vanishes, while the remaining mixed terms involve the tiny reconstructed YLY^L couplings and are therefore negligible. Thus, order-one Y~R\widetilde{Y}^R couplings can bring rare top decays close to current sensitivity only for leptoquark masses near the TeV scale; the rates rapidly become negligible for heavier spectra.

References

References

  1. [1]R. N. Mohapatra and G. Senjanovic, “Neutrino Mass and Spontaneous Parity Nonconservation,” Phys. Rev. Lett. 44, 912 (1980).DOI
  2. [2]E. Ma and U. Sarkar, “Neutrino masses and leptogenesis with heavy Higgs triplets,” Phys. Rev. Lett. 80, 5716 (1998), arXiv:hep-ph/9802445.DOI
  3. [3]A. Zee, “A Theory of Lepton Number Violation, Neutrino Majorana Mass, and Oscillation,” Phys. Lett. B 93, 389 (1980), [Erratum: Phys.Lett. B 95, 461 (1980)].DOI
  4. [4]K. S. Babu, “Model of ’Calculable’ Majorana Neutrino Masses,” Phys. Lett. B 203, 132 (1988).DOI
  5. [5]J. Schechter and J. W. F. Valle, “Neutrinoless Double beta Decay in SU(2) x U(1) Theories,” Phys. Rev. D 25, 2951 (1982).DOI
  6. [6]F. Takahashi et al. (Particle Data Group), “Review of Particle Physics,” Int. J. Mod. Phys. A 41, 2630011 (2026).
  7. [7]S. Abe et al. (KamLAND-Zen), “Search for Majorana Neutrinos with the Complete KamLAND-Zen Dataset,” Phys. Rev. Lett. 135, 262501 (2025), arXiv:2406.11438 [hep-ex].DOI
  8. [8]M. Agostini et al. (GERDA), “Final Results of GERDA on the Search for Neutrinoless Double-β Decay,” Phys. Rev. Lett. 125, 252502 (2020), arXiv:2009.06079 [nucl-ex].DOI
  9. [9]D. Q. Adams et al. (CUORE), “Search for Majorana neutrinos exploiting millikelvin cryogenics with CUORE,” Nature 604, 53 (2022), arXiv:2104.06906 [nucl-ex].DOI
  10. [10]G. Anton et al. (EXO-200), “Search for Neutrinoless Double-β Decay with the Complete EXO-200 Dataset,” Phys. Rev. Lett. 123, 161802 (2019), arXiv:1906.02723 [hep-ex].DOI
  11. [11]D. Chang and R. N. Mohapatra, “Small and Calculable Dirac Neutrino Mass,” Phys. Rev. Lett. 58, 1600 (1987).pubmed.ncbi.nlm.nih.gov/10034486
  12. [12]R. N. Mohapatra, “Left-right Symmetry and Finite One Loop Dirac Neutrino Mass,” Phys. Lett. B 201, 517 (1988).DOI
  13. [13]K. S. Babu and X. G. He, “DIRAC NEUTRINO MASSES AS TWO LOOP RADIATIVE CORRECTIONS,” Mod. Phys. Lett. A 4, 61 (1989).DOI
  14. [14]M. Lindner, T. Ohlsson, and G. Seidl, “Seesaw mechanisms for Dirac and Majorana neutrino masses,” Phys. Rev. D 65, 053014 (2002), arXiv:hep-ph/0109264.DOI
  15. [15]S. Centelles Chuliá, R. Srivastava, and J. W. F. Valle, “Seesaw roadmap to neutrino mass and dark matter,” Phys. Lett. B 781, 122 (2018), arXiv:1802.05722 [hep-ph].DOI
  16. [16]S. Centelles Chuliá, R. Srivastava, and J. W. F. Valle, “Seesaw Dirac neutrino mass through dimension-six operators,” Phys. Rev. D 98, 035009 (2018), arXiv:1804.03181 [hep-ph].
  17. [17]V. Oliveira, P. Escalona, L. Angel, C. A. de S. Pires, and F. S. Queiroz, “Type-II seesaw mechanism for Dirac neutrinos and its implications on N_eff and lepton flavor violation in a 3-3-1 model,” JHEP 07, 197 (2025), arXiv:2502.01760 [hep-ph].arxiv.org/abs/2502.01760
  18. [18]J. Calle, D. Restrepo, C. E. Yaguna, and Ó. Zapata, “Minimal radiative Dirac neutrino mass models,” Phys. Rev. D 99, 075008 (2019), arXiv:1812.05523 [hep-ph].DOI
  19. [19]J. C. Pati and A. Salam, “Unified Lepton-Hadron Symmetry and a Gauge Theory of the Basic Interactions,” Phys. Rev. D 8, 1240 (1973).
  20. [20]I. Doršner, S. Fajfer, A. Greljo, J. F. Kamenik, and N. Košnik, “Physics of leptoquarks in precision experiments and at particle colliders,” Phys. Rept. 641, 1 (2016), arXiv:1603.04993 [hep-ph].DOI
  21. [21]U. Mahanta, “Neutrino masses and mixing angles from leptoquark interactions,” Phys. Rev. D 62, 073009 (2000), arXiv:hep-ph/9909518.
  22. [22]K. S. Babu and J. Julio, “Two-Loop Neutrino Mass Generation through Leptoquarks,” Nucl. Phys. B 841, 130 (2010), arXiv:1006.1092 [hep-ph].DOI
  23. [23]D. Zhang, “Radiative neutrino masses, lepton flavor mixing and muon g − 2 in a leptoquark model,” JHEP 07, 069 (2021), arXiv:2105.08670 [hep-ph].DOI
  24. [24]K. S. Babu, P. S. B. Dev, S. Jana, and A. Thapa, “Unified framework for B-anomalies, muon g − 2 and neutrino masses,” JHEP 03, 179 (2021), arXiv:2009.01771 [hep-ph].DOI
  25. [25]A. Doff, J. P. Pinheiro, and C. A. de S. Pires, “Leptoquark-induced radiative masses for active and sterile neutrinos within the framework of the 3-3-1 model,” Phys. Lett. B 863, 139375 (2025), arXiv:2412.15055 [hep-ph].arxiv.org/abs/2412.15055
  26. [26]T. Enkhbat, “Scalar leptoquarks and Higgs pair production at the LHC,” JHEP 01, 158 (2014), arXiv:1311.4445 [hep-ph].DOI
  27. [27]S.-L. Chen, W.-w. Jiang, and Z.-K. Liu, “Combined explanations of B-physics anomalies, (g − 2)e,μ and neutrino masses by scalar leptoquarks,” Eur. Phys. J. C 82, 959 (2022), arXiv:2205.15794 [hep-ph].DOI
  28. [28]M. Bauer and M. Neubert, “Minimal Leptoquark Explanation for the RD(*), RK, and (g−2)μ Anomalies,” Phys. Rev. Lett. 116, 141802 (2016), arXiv:1511.01900 [hep-ph].
  29. [29]D. Bečirević, S. Fajfer, N. Košnik, and O. Sumensari, “Leptoquark model to explain the B-physics anomalies, RK and RD,” Phys. Rev. D 94, 115021 (2016), arXiv:1608.08501 [hep-ph].DOI
  30. [30]I. Bigaran, J. Gargalionis, and R. R. Volkas, “A near-minimal leptoquark model for reconciling flavour anomalies and generating radiative neutrino masses,” JHEP 10, 106 (2019), arXiv:1906.01870 [hep-ph].DOI
  31. [31]S. Sahoo and R. Mohanta, “Leptoquark effects on b → sνν̄ and B → Kl+l− decay processes,” New J. Phys. 18, 013032 (2016), arXiv:1509.06248 [hep-ph].
  32. [32]L. Calibbi, A. Crivellin, and T. Li, “Model of vector leptoquarks in view of the B-physics anomalies,” Phys. Rev. D 98, 115002 (2018), arXiv:1709.00692 [hep-ph].DOI
  33. [33]C.-H. Chen and C.-W. Chiang, “Flavor anomalies in leptoquark model with gauged U(1)Lμ−Lτ,” Phys. Rev. D 109, 075004 (2024), arXiv:2309.12904 [hep-ph].
  34. [34]W. Altmannshofer, A. Crivellin, H. Haigh, G. Inгуglia, and J. Martin Camalich, “Light new physics in B→K(*)νν̄,” Phys. Rev. D 109, 075008 (2024), arXiv:2311.14629 [hep-ph].arxiv.org/abs/2311.14629
  35. [35]A. Crivellin, S. Iguro, and T. Kitahara, “Discriminating tauphilic leptoquark explana- tions of the B anomalies via K→πνν̄ and B→Kνν̄,” Phys. Rev. D 112, 095016 (2025), arXiv:2505.05552 [hep-ph].
  36. [36]C.-H. Chen, C.-W. Chiang, and L. M. G. de la Vega, “Leptoquark-mediated Dirac neutrino mass and its impact on \(B \to K\nu\bar\nu\) and \(K \to \pi\nu\bar\nu\) decays,” JHEP 09, 055 (2025), arXiv:2503.22431 [hep-ph].arxiv.org/abs/2503.22431
  37. [37]I. Adachi et al. (Belle-II), “Evidence for B+→K+νν̄ decays,” Phys. Rev. D 109, 112006 (2024), arXiv:2311.14647 [hep-ex].
  38. [38]A. J. Buras and P. Stangl, “On the interplay of constraints from \(B_s\), \(D\), and \(K\) meson mixing in \(Z'\) models with implications for \(b \to s\nu\bar\nu\) transitions,” Eur. Phys. J. C 85, 519 (2025), arXiv:2412.14254 [hep-ph].arxiv.org/abs/2412.14254
  39. [39]P. Athron, R. Martinez, and C. Sierra, “B meson anomalies and large \(B^+ \to K^+\nu\bar\nu\) in non-universal U(1)′ models,” JHEP 02, 121 (2024), arXiv:2308.13426 [hep-ph].DOI
  40. [40]K. Fridell, M. Ghosh, T. Okui, and K. Tobioka, “Decoding the B→Kνν̄ excess at Belle II: Kinematics, operators, and masses,” Phys. Rev. D 109, 115006 (2024), arXiv:2312.12507 [hep-ph].DOI
  41. [41]E. Gabrielli, L. Marzola, K. Müürsepp, and M. Raidal, “Explaining the \(B^+ \to K^+\nu\bar\nu\) excess via a massless dark photon,” Eur. Phys. J. C 84, 460 (2024), arXiv:2402.05901 [hep-ph].arxiv.org/abs/2402.05901
  42. [42]B.-F. Hou, X.-Q. Li, M. Shen, Y.-D. Yang, and X.-B. Yuan, “Deciphering the Belle II data on \(B \to K\nu\bar\nu\) decay in the (dark) SMEFT with minimal flavour violation,” JHEP 06, 172 (2024), arXiv:2402.19208 [hep-ph].arxiv.org/abs/2402.19208
  43. [43]F.-Z. Chen, Q. Wen, and F. Xu, “Correlating \(B \to K^{(*)}\nu\bar\nu\) and flavor anomalies in SMEFT,” Eur. Phys. J. C 84, 1012 (2024), arXiv:2401.11552 [hep-ph].DOI
  44. [44]A. J. Buras, J. Harz, and M. A. Mojahed, “Disentangling new physics in \(K \to \pi\nu\bar\nu\) and \(B \to K(K^*)\nu\bar\nu\) observables,” JHEP 10, 087 (2024), arXiv:2405.06742 [hep-ph].
  45. [45]ATLAS Collaboration, Leptoquark summary plots for scalar or vector models, Tech. Rep. ATL-PHYS-PUB-2025-013 (CERN, 2025).DOI
  46. [46]G. Aad et al. (ATLAS), “Search for resonant leptoquark production via lepton-jet signatures in \(pp\) collisions at \(\sqrt{s}=13\) TeV and \(\sqrt{s}=13.6\) TeV with the ATLAS detector,” JHEP 12, 180 (2025), arXiv:2507.03650 [hep-ex].
  47. [47]Y. Farzan and E. Ma, “Dirac neutrino mass generation from dark matter,” Phys. Rev. D 86, 033007 (2012), arXiv:1204.4890 [hep-ph].DOI
  48. [48]P. O. Ludl and W. Grimus, “A complete survey of texture zeros in the lepton mass matrices,” JHEP 07, 090 (2014), [Erratum: JHEP 10, 126 (2014)], arXiv:1406.3546 [hep-ph].DOI
  49. [49]I. Esteban, M. C. Gonzalez-Garcia, M. Maltoni, I. Martinez-Soler, J. P. Pinheiro, and T. Schwetz, “Nufit-6.0: Updated global analysis of three-flavor neutrino oscillations,” JHEP 12, 216 (2024), arXiv:2410.05380 [hep-ph].DOI
  50. [50]U. Khasianevich, D. Stöckinger, H. Stöckinger-Kim, and J. Wünsche, “Constraint on scalar leptoquark from low-energy leptonic observables,” Phys. Rev. D 108, 095027 (2023), arXiv:2305.05016 [hep-ph].arxiv.org/abs/2305.05016
  51. [51]L. Allwicher, P. Arnan, D. Barducci, and M. Nardecchia, “Perturbative unitarity constraints on generic Yukawa interactions,” JHEP 10, 129 (2021), arXiv:2108.00013 [hep-ph].DOI
  52. [52]W. G. Parrott, C. Bouchard, and C. T. H. Davies ((HPQCD collaboration)§, HPQCD), “B→K and D→K form factors from fully relativistic lattice QCD,” Phys. Rev. D 107, 014510 (2023), arXiv:2207.12468 [hep-lat].DOI
  53. [53]W. G. Parrott, C. Bouchard, and C. T. H. Davies (HPQCD), “Standard Model predictions for B → Kℓ+ℓ−, B → Kℓ1+ℓ2− and B → Kνν̄ using form factors from Nf = 2 + 1 + 1 lattice QCD,” Phys. Rev. D 107, 014511 (2023), arXiv:2207.13371 [hep-ph].
  54. [54]B. W. Lee and R. E. Shrock, “Natural Suppression of Symmetry Violation in Gauge Theories: Muon - Lepton and Electron Lepton Number Nonconservation,” Phys. Rev. D 16, 1444 (1977).
  55. [55]D. Palo, “Status of the MEG II Experiment and Performance Results from the First Year’s Data Taking,” Phys. Sci. Forum 8, 6 (2023).DOI
  56. [56]L. Calibbi and G. Signorelli, “Charged Lepton Flavour Violation: An Experimental and Theoretical Introduction,” Riv. Nuovo Cim. 41, 71 (2018), arXiv:1709.00294 [hep-ph].DOI
  57. [57]L. Lavoura, “General formulae for f(1) —> f(2) gamma,” Eur. Phys. J. C 29, 191 (2003), arXiv:hep-ph/0302221.DOI
  58. [58]K. Afanaciev et al. (MEG II), “New limit on the μ+ → e+γ decay with the MEG II experiment,” Eur. Phys. J. C 85, 1177 (2025), arXiv:2504.15711 [hep-ex].arxiv.org/abs/2504.15711
  59. [59]K. N. Abazajian and J. Heeck, “Observing Dirac neutrinos in the cosmic microwave background,” Phys. Rev. D 100, 075027 (2019), arXiv:1908.03286 [hep-ph].DOI
  60. [60]X. Luo, W. Rodejohann, and X.-J. Xu, “Dirac neutrinos and Neff,” JCAP 06, 058 (2020), arXiv:2005.01629 [hep-ph].
  61. [61]N. Aghanim et al. (Planck), “Planck 2018 results. vi. cosmological parameters,” Astron. As- trophys. 641, A6 (2020), arXiv:1807.06209 [astro-ph.CO].DOI
  62. [62]J. J. Bennett, G. Buldgen, P. F. de Salas, M. Drewes, S. Gariazzo, S. Pastor, and Y. Y. Y. Wong, “Towards a precision calculation of N_eff in the Standard Model II: Neutrino decoupling in the presence of flavour oscillations and finite-temperature QED,” JCAP 04, 073 (2021), arXiv:2012.02726 [hep-ph].DOI
  63. [63]G. Mangano, G. Miele, S. Pastor, T. Pinto, O. Pisanti, and P. D. Serpico, “Relic neutrino decoupling including flavor oscillations,” Nucl. Phys. B 729, 221 (2005), arXiv:hep-ph/0506164.DOI
  64. [64]E. Calabrese et al. (ACT), “The Atacama Cosmology Telescope: DR6 Constraints on Extended Cosmological Models,” JCAP 11, 063 (2025), arXiv:2503.14454 [astro-ph.CO].arxiv.org/abs/2503.14454
  65. [65]M. Cannoni, “Relativistic < σv_rel > in the calculation of relics abundances: a closer look,” Phys. Rev. D 89, 103533 (2014), arXiv:1311.4494 [astro-ph.CO].
  66. [66]P. Gondolo and G. Gelmini, “Cosmic abundances of stable particles: Improved analysis,” Nucl. Phys. B 360, 145 (1991).
  67. [67]L. Husdal, “On Effective Degrees of Freedom in the Early Universe,” Galaxies 4, 78 (2016), arXiv:1609.04979 [astro-ph.CO].DOI
  68. [68]L. A. Anchordoqui, H. Goldberg, and G. Steigman, “Right-Handed Neutrinos as the Dark Radiation: Status and Forecasts for the LHC,” Phys. Lett. B 718, 1162 (2013), arXiv:1211.0186 [hep-ph].DOI
  69. [69]G. Aad et al. (ATLAS), “Search for pair-produced scalar and vector leptoquarks decaying into third-generation quarks and first- or second-generation leptons in pp collisions with the ATLAS detector,” JHEP 2306, 188 (2023), arXiv:2210.04517 [hep-ex].
  70. [70]T. E. Browder, N. G. Deshpande, R. Mandal, and R. Sinha, “Impact of B→Kνν̄ measurements on beyond the Standard Model theories,” Phys. Rev. D 104, 053007 (2021), arXiv:2107.01080 [hep-ph].DOI
  71. [71]C. Bourrely, L. Lellouch, and I. Caprini, “Model-independent description of b → πlν decays and a determination of |V_ub|,” Phys. Rev. D 79, 013008 (2009), arXiv:0807.2722 [hep-ph].arxiv.org/abs/0807.2722
  72. [72]G. Aad et al. (ATLAS), “Search for charged-lepton-flavour violating μτqt interactions in top-quark production and decay in pp collisions at √s = 13 TeV with the ATLAS detector at the LHC,” Phys. Rev. D 110, 012014 (2024), arXiv:2403.06742 [hep-ex].

Paper details

vlnc:W_dj4jxgfb
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents