betaFeedback

WorksPhysics and Astronomy

Interacting scalar-fermion two-component dark matter in a gauged U(1)B−L\boldsymbol{U(1)_{B-L}} model

Abstract

We propose an interacting two-component dark matter model based on a gauged U(1)B−LU(1)_{B-L} extension of the Standard Model, which also accommodates neutrino masses through the type-I seesaw mechanism. The dark sector consists of a Dirac fermion χχ and a complex scalar SS, which can both be stable without imposing an additional discrete symmetry and are directly coupled by a gauge-invariant Yukawa interaction. The resulting conversion, semi-conversion, and semi-annihilation processes lead to a genuinely coupled freeze-out. After imposing collider and dark matter constraints, we identify viable solutions away from the Z′Z' resonance for both mass hierarchies, mS>mχm_S > m_χ and mS<mχm_S < m_χ. Despite the strong direct-detection constraints, part of the surviving parameter space remains within the reach of next-generation experiments.

Introduction

The existence of dark matter (DM) is firmly established by astrophysical and cosmological observations, including galaxy rotation curves, gravitational lensing, large-scale structure formation, and measurements of the cosmic microwave background. Nevertheless, its particle nature remains unknown. Among the most extensively studied candidates are weakly interacting massive particles (WIMPs), whose abundance may be generated through thermal freeze-out in the early Universe. The apparent connection between a weak-scale annihilation cross section and the observed DM abundance—the so-called WIMP miracle—has traditionally motivated models in which a single particle accounts for the entire DM density. The absence of convincing non-gravitational signals, together with increasingly stringent limits from collider, direct-detection, and indirect-detection searches, has significantly restricted the simplest realizations of this paradigm [1, 2, 3, 4].

An equally plausible possibility is that the dark sector contains more than one stable particle. In multi-component DM scenarios, the observed density is shared among two or more species, each contributing only a fraction of the total abundance [5, 6, 7, 8, 9, 10, 11]. More importantly, the thermal histories of the different components need not be independent [12]. Interactions among the dark particles can induce conversion, semi-annihilation, and semi-conversion processes, thereby coupling their Boltzmann equations and qualitatively modifying the freeze-out dynamics [13, 14, 15]. These processes can weaken the usual connection between the relic abundance and the couplings probed by direct-detection experiments, opening viable regions that would not be available in the corresponding single-component models.

A central model-building question in such scenarios is the origin of the stability of the different DM particles. In many constructions, stability is imposed through additional discrete symmetries, such as ZNZ_N or products thereof [8, 14, 16, 17]. Although perfectly consistent at the phenomenological level, these symmetries are often introduced specifically for this purpose. It is therefore worthwhile to investigate frameworks in which the stability of multiple dark particles follows instead from an underlying gauge symmetry. Such a framework becomes especially attractive when the same gauge extension also addresses another open problem of the Standard Model (SM), such as the origin of neutrino masses.

The gauged U(1)B−LU(1)_{B-L} extension provides a simple realization of this idea [18, 19, 20, 21, 22, 23]. In its minimal anomaly-free implementation, it contains three right-handed neutrinos and a complex scalar singlet ΦBL\Phi_{BL}. The vacuum expectation value of ΦBL\Phi_{BL} spontaneously breaks the U(1)B−LU(1)_{B-L} symmetry, generates the mass of the new gauge boson Z′Z^\prime, and gives Majorana masses to the right-handed neutrinos. Together with their Dirac Yukawa interactions, this leads to light neutrino masses through the type-I seesaw mechanism [24, 25]. The scalar excitation associated with ΦBL\Phi_{BL} also mixes with the SM Higgs boson, giving rise to a second physical scalar, denoted by h2h_2.

In this work, we extend this framework by introducing only two additional fields: a Dirac fermion χ\chi and a complex scalar SS, both singlets under the SM gauge group and charged under U(1)B−LU(1)_{B-L}. Their charges are chosen so that gauge invariance allows a Yukawa interaction that directly couples the two DM particles. This interaction is a defining feature of the model, as it enables conversion, semi-annihilation, and semi-conversion processes that couple the thermal evolution of the two components. For suitable B−LB-L charge assignments, gauge invariance forbids additional renormalizable operators that would destabilize the dark sector, while the kinematic condition mS<2mχm_S < 2m_\chi prevents the decay of the scalar into a dark fermion pair. Both particles can therefore be stable without imposing an additional discrete symmetry, and their relic abundances are determined by a coupled network of dark-sector processes rather than by two independent freeze-out histories.

We perform a numerical exploration of the multidimensional parameter space, solving the coupled Boltzmann equations and imposing theoretical consistency conditions together with collider and dark matter constraints. We analyze separately the two mass hierarchies, mS>mχm_S > m_\chi and mS<mχm_S < m_\chi, and identify viable regions in both regimes. In particular, viable solutions occur over broad regions away from the Z′Z^\prime resonance, showing that the observed relic abundance does not require resonant enhancement. The coupled dark-sector processes can instead play a central role in determining the relic composition. Direct detection provides a strong constraint on the model, while part of the surviving parameter space remains within the projected reach of next-generation experiments.

The remainder of this paper is organized as follows. In § 2, we introduce the model and discuss its scalar and dark matter sectors. § 3 describes the coupled freeze-out dynamics, while the relevant experimental constraints are presented in § 4. The numerical results are discussed in § 5, followed by a broader discussion in § 6. We summarize our conclusions in § 7.

The U(1)B−LU(1)_{B-L} two-component dark matter model

Field content and Lagrangian

We consider a gauge extension of the SM based on

SU(3)C⊗SU(2)L⊗U(1)Y⊗U(1)B−L,(1)SU(3)_C \otimes SU(2)_L \otimes U(1)_Y \otimes U(1)_{B-L}, \tag*{(1)}

where BB and LL denote baryon and lepton number, respectively. The cancellation of the gauge and mixed gravitational anomalies requires the introduction of three right-handed neutrinos, NiRN_{iR}, with i=1,2,3i = 1, 2, 3. The U(1)B−LU(1)_{B-L} symmetry is spontaneously broken by a complex scalar singlet ΦBL\Phi_{BL} with B−LB-L charge equal to 2. The mass of the new gauge boson is mZ′=2gBLvΦm_{Z^\prime} = 2g_{BL}v_\Phi, where gBLg_{BL} is the U(1)B−LU(1)_{B-L} gauge coupling and ⟨ΦBL⟩=vΦ/2\langle\Phi_{BL}\rangle= v_\Phi/\sqrt{2} defines the ΦBL\Phi_{BL}’s vacuum expectation value.

The dark sector contains a Dirac fermion χ\chi and a complex scalar SS, both singlets under the SM gauge group. The left- and right-handed components of χ\chi carry the same B−LB-L charge nχn_\chi, so that χ\chi is vector-like and does not introduce additional gauge anomalies. To obtain a genuinely interacting two-component dark sector, rather than a simple superposition of two independent single-component DM models, we require a renormalizable Yukawa interaction directly coupling χ\chi and SS. Gauge invariance of this interaction fixes their charges to satisfy QB−L(S)=−2QB−L(χ)=−2nχQ_{B-L}(S) = -2Q_{B-L}(\chi) = -2n_\chi. The complete field content and charge assignments are summarized in Table 1.

FieldSU(3)CSU(3)_CSU(2)LSU(2)_LU(1)YU(1)_YU(1)B−LU(1)_{B-L}
qiLq_{iL}321/61/61/31/3
uiRu_{iR}312/32/31/31/3
diRd_{iR}31−1/3-1/31/31/3
ℓiL\ell_{iL}12−1/2-1/2−1-1
eiRe_{iR}11−1-1−1-1
NiRN_{iR}110−1-1
HH121/21/20
ΦBL\Phi_{BL}1102
χ\chi110nχn_\chi
SS110−2nχ-2n_\chi

Table 1. Field content and gauge-charge assignments of the two-component DM model.

The B−LB-L charge nχn_\chi plays a crucial role in determining the structure of the dark sector. While most values lead to the minimal interacting two-component model considered here, certain special assignments either prevent the existence of two stable DM particles or allow additional interactions that modify the model. The relevant cases are summarized in Table 2. Any value of nχn_\chi not listed in Table 2 defines a consistent realization of the minimal model at the renormalizable level. For the numerical analysis, we adopt the benchmark charge assignment nχ=3/4n_\chi= 3/4, which is an O(1)O(1) value compatible with the minimal two-component dark matter scenario defined above.

Classificationnχn_\chiConsequence
Nonviable00, ±13\pm\frac{1}{3}, ±1\pm1, ±2\pm2, ±3\pm3The two-component dark matter scenario considered here cannot be realized.
Viable but nonminimal±12\pm\frac{1}{2}Additional renormalizable interactions are allowed, leading to an extended two-component dark matter model.
Conditionally viable±4\pm4The renormalizable model remains viable, but higher-dimensional operators can induce the decay of the scalar dark matter particle.

Table 2. Special values of the dark B−LB-L charge nχn_\chi and their implications for the two-component dark matter scenario.

The most general renormalizable gauge Lagrangian also contains kinetic mixing between U(1)YU(1)_Y and U(1)B−LU(1)_{B-L}. For simplicity, we neglect gauge kinetic mixing throughout the phenomenological analysis. The terms relevant for neutrino masses and DM phenomenology can then be written as

L⊃LSMB−L−14Fμν′F′μν+∑i=13iN‾iRγμDμNiR+iχ‾γμDμχ−mχχ‾χ+(DμΦBL)†(DμΦBL)+(DμS)†(DμS)−[yijDLi‾H~NjR+12yijMNiRc‾NjRΦBL+12Sχc‾(ys+ypγ5)χ+H.c.]−V(H,ΦBL,S),(2)\begin{aligned} \mathcal{L} \supset{} & \mathcal{L}_{\mathrm{SM}}^{B-L} - \frac{1}{4}F'_{\mu\nu}F'^{\mu\nu} + \sum_{i=1}^{3} i\overline{N}_{iR}\gamma^\mu D_\mu N_{iR} + i\overline{\chi}\gamma^\mu D_\mu\chi- m_\chi\overline{\chi}\chi\\ & + (D_\mu\Phi_{BL})^\dagger(D^\mu\Phi_{BL}) + (D_\mu S)^\dagger(D^\mu S) \\ & - \left[y^D_{ij}\overline{L_i}\widetilde{H}N_{jR} + \frac{1}{2}y^M_{ij}\overline{N^c_{iR}}N_{jR}\Phi_{BL} + \frac{1}{2}S\overline{\chi^c}(y_s+y_p\gamma_5)\chi+ \mathrm{H.c.}\right] \\ & - V(H,\Phi_{BL},S), \tag*{(2)} \end{aligned}

where LSMB−L\mathcal{L}_{\mathrm{SM}}^{B-L} denotes the SM Lagrangian with the covariant derivatives extended to include the U(1)B−LU(1)_{B-L} gauge interaction, and H~=iσ2H∗\widetilde{H}=i\sigma_2H^*. For a field carrying B−LB-L charge njn_j, the corresponding contribution to the covariant derivative is Dμ⊃∂μ+igBLnjZμ′D_\mu\supset\partial_\mu+ig_{BL}n_jZ'_\mu. After electroweak and U(1)B−LU(1)_{B-L} symmetry breaking, the Yukawa couplings yijDy^D_{ij} and yijMy^M_{ij} generate the Dirac and Majorana neutrino mass matrices, respectively, leading to light neutrino masses through the type-I seesaw mechanism. These parameters can be chosen to reproduce the observed neutrino masses and mixings while keeping the right-handed neutrinos sufficiently heavy that their production and decay channels are kinematically inaccessible throughout the parameter space considered. The neutrino sector therefore does not affect the dark matter phenomenology studied below.

The last Yukawa interaction in Eq. (2) is allowed because the dark-sector charges satisfy QB−L(S)=−2QB−L(χ)Q_{B-L}(S)=-2Q_{B-L}(\chi). It directly couples the two dark particles and induces the conversion, semi-annihilation, and semi-conversion processes that distinguish this model from two independently freezing-out DM components. Finally, the scalar potential is

V(H,ΦBL,S)=μH2H†H+μΦ2ΦBL†ΦBL+μS2S†S+λH(H†H)2+λΦ(ΦBL†ΦBL)2+λS(S†S)2+λHΦ(H†H)(ΦBL†ΦBL)+λSH(H†H)(S†S)+λSΦ(S†S)(ΦBL†ΦBL).(3)\begin{aligned} V(H,\Phi_{BL},S) ={}& \mu_H^2H^\dagger H+\mu_\Phi^2\Phi_{BL}^\dagger\Phi_{BL}+\mu_S^2S^\dagger S \\ &+\lambda_H(H^\dagger H)^2+\lambda_\Phi(\Phi_{BL}^\dagger\Phi_{BL})^2+\lambda_S(S^\dagger S)^2 \\ &+\lambda_{H\Phi}(H^\dagger H)(\Phi_{BL}^\dagger\Phi_{BL})+\lambda_{SH}(H^\dagger H)(S^\dagger S) \\ &+\lambda_{S\Phi}(S^\dagger S)(\Phi_{BL}^\dagger\Phi_{BL}). \tag*{(3)} \end{aligned}

The mass-squared parameters μH2\mu_H^2, μΦ2\mu_\Phi^2, and μS2\mu_S^2 are not assumed to be positive. Only HH and ΦBL\Phi_{BL} acquire vacuum expectation values (VEVs), whereas the vacuum is required to satisfy ⟨S⟩=0\langle S\rangle= 0. The conditions under which this configuration is a consistent minimum are discussed in § 2.3.

Symmetry breaking and scalar mixing

After electroweak and U(1)B−LU(1)_{B-L} symmetry breaking, the Higgs doublet and the B−LB-L scalar acquire VEVs. In the unitary gauge, they can be parametrized as

H=12(0vH+h),ΦBL=12(vΦ+ϕ),(4)H = \frac{1}{\sqrt{2}}\begin{pmatrix} 0 \\ v_H+h \end{pmatrix}, \qquad\Phi_{BL} = \frac{1}{\sqrt{2}}(v_\Phi+\phi), \tag*{(4)}

where vH=246 GeVv_H=246\ \mathrm{GeV}. We require the dark scalar not to acquire a VEV, ⟨S⟩=0\langle S\rangle=0, because its CP-even excitation would generically mix with those of HH and ΦBL\Phi_{BL}. Through this mixing, the predominantly-SS state would acquire couplings to visible particles and would no longer be stable, thereby spoiling the two-component DM scenario. In § 2.3, we discuss the conditions under which the vacuum with ⟨S⟩=0\langle S\rangle=0 is realized.

Assuming vH,vΦ≠0v_H,v_\Phi\ne0, the minimal conditions of the scalar potential are

μH2+λHvH2+12λHΦvΦ2=0,μΦ2+λΦvΦ2+12λHΦvH2=0.(5)\mu_H^2+\lambda_Hv_H^2+\frac{1}{2}\lambda_{H\Phi}v_\Phi^2=0,\qquad\mu_\Phi^2+\lambda_\Phi v_\Phi^2+\frac{1}{2}\lambda_{H\Phi}v_H^2=0. \tag*{(5)}

Since SS does not mix with the other neutral scalars, it remains a physical DM state with mass

mS2=μS2+12λSHvH2+12λSΦvΦ2.(6)m_S^2=\mu_S^2+\frac{1}{2}\lambda_{SH}v_H^2+\frac{1}{2}\lambda_{S\Phi}v_\Phi^2. \tag*{(6)}

The remaining CP-even scalar fields, hh and ϕ\phi, mix through the portal coupling λHΦ\lambda_{H\Phi}. Their mass-squared matrix in the (h,ϕ)(h,\phi) basis is

M2=(2λHvH2λHΦvHvΦλHΦvHvΦ2λΦvΦ2).(7)\mathcal{M}^2=\begin{pmatrix} 2\lambda_Hv_H^2 & \lambda_{H\Phi}v_Hv_\Phi\\ \lambda_{H\Phi}v_Hv_\Phi& 2\lambda_\Phi v_\Phi^2 \end{pmatrix}. \tag*{(7)}

The mass eigenstates are obtained through the orthogonal transformation

h1=cos⁡θ h−sin⁡θ ϕ,h2=sin⁡θ h+cos⁡θ ϕ.(8)h_1=\cos\theta\,h-\sin\theta\,\phi,\qquad h_2=\sin\theta\,h+\cos\theta\,\phi. \tag*{(8)}

where h1h_1 is identified with the observed Higgs boson, with mh1=125 GeVm_{h_1}=125\ \mathrm{GeV}, while h2h_2 is the predominantly B−LB-L scalar. With this rotation convention, the mixing angle satisfies

tan⁡2θ=λHΦvHvΦλΦvΦ2−λHvH2.(9)\tan2\theta=\frac{\lambda_{H\Phi}v_Hv_\Phi}{\lambda_\Phi v_\Phi^2-\lambda_Hv_H^2}. \tag*{(9)}

For the numerical analysis, it is convenient to express the quartic couplings λH\lambda_H, λΦ\lambda_\Phi, and λHΦ\lambda_{H\Phi} in terms of the physical scalar masses, the mixing angle, and the vacuum expectation values. The corresponding relations are

λH=mh12+mh22+(mh12−mh22)cos⁡2θ4vH2,(10)\lambda_H=\frac{m_{h_1}^2+m_{h_2}^2+(m_{h_1}^2-m_{h_2}^2)\cos2\theta}{4v_H^2}, \tag*{(10)}
λΦ=mh12+mh22−(mh12−mh22)cos⁡2θ4vΦ2,(11)\lambda_\Phi=\frac{m_{h_1}^2+m_{h_2}^2-(m_{h_1}^2-m_{h_2}^2)\cos2\theta}{4v_\Phi^2}, \tag*{(11)}
λHΦ=(mh22−mh12)sin⁡2θ2vHvΦ.(12)\lambda_{H\Phi}=\frac{(m_{h_2}^2-m_{h_1}^2)\sin2\theta}{2v_Hv_\Phi}. \tag*{(12)}

The mass parameters μH2\mu_H^2 and μΦ2\mu_\Phi^2 are then fixed by the minimal conditions. Thus, once mh1m_{h_1} and vHv_H are fixed to their measured values, the mixed hh–ϕ\phi sector can be parametrized by {mh2,mZ′,gBL,θ}\{m_{h_2},m_{Z'},g_{BL},\theta\}, with vΦ=mZ′/(2gBL)v_\Phi=m_{Z'}/(2g_{BL}).

Boundedness and local vacuum stability

A consistent scalar sector requires the potential to be bounded from below and the stationary point

⟨H⟩=vH2,⟨ΦBL⟩=vΦ2,⟨S⟩=0(13)\langle H\rangle= \frac{v_H}{\sqrt{2}}, \qquad\langle\Phi_{BL}\rangle= \frac{v_\Phi}{\sqrt{2}}, \qquad\langle S\rangle= 0 \tag*{(13)}

to be locally stable. These two requirements constrain different properties of the potential: boundedness from below controls its large-field behavior, whereas local stability is determined by the scalar mass matrix evaluated at the stationary point.

At large field values, the quartic part of the scalar potential dominates. Applying the copositivity criteria to the three non-negative field bilinears H†HH^\dagger H, ΦBL†ΦBL\Phi^\dagger_{BL}\Phi_{BL}, and S†SS^\dagger S, the necessary and sufficient bounded-from-below conditions are [26] λH≥0\lambda_H \ge0, λΦ≥0\lambda_\Phi\ge0, λS≥0\lambda_S \ge0, together with

λˉHΦ≡λHΦ+2λHλΦ≥0,λˉSH≡λSH+2λSλH≥0,λˉSΦ≡λSΦ+2λSλΦ≥0,\bar{\lambda}_{H\Phi} \equiv\lambda_{H\Phi} + 2\sqrt{\lambda_H\lambda_\Phi} \ge0,\qquad\bar{\lambda}_{SH} \equiv\lambda_{SH} + 2\sqrt{\lambda_S\lambda_H} \ge0,\qquad\bar{\lambda}_{S\Phi} \equiv\lambda_{S\Phi} + 2\sqrt{\lambda_S\lambda_\Phi} \ge0,

and

λHλΦλS+12(λHΦλS+λSHλΦ+λSΦλH+λˉHΦλˉSHλˉSΦ)≥0.(14)\sqrt{\lambda_H\lambda_\Phi\lambda_S} + \frac{1}{2}\left(\lambda_{H\Phi}\sqrt{\lambda_S} + \lambda_{SH}\sqrt{\lambda_\Phi} + \lambda_{S\Phi}\sqrt{\lambda_H} + \sqrt{\bar{\lambda}_{H\Phi}\bar{\lambda}_{SH}\bar{\lambda}_{S\Phi}}\right) \ge0. \tag*{(14)}

All these inequalities must be imposed simultaneously; the last condition is additional to the three pairwise conditions. Local stability in the hh–ϕ\phi sector follows from the positivity of the physical scalar masses, mh12>0m_{h_1}^2 > 0, mh22>0m_{h_2}^2 > 0. Since SS does not mix with the other scalars at this stationary point, local stability along both real components of the complex field SS requires mS2>0m_S^2 > 0.

In the numerical analysis, we impose the bounded-from-below conditions together with positive physical scalar masses.

Dark matter stability

Assuming the vacuum configuration ⟨S⟩=0\langle S\rangle= 0, the stability of the dark particles can be inferred directly from the renormalizable Lagrangian in Eq. (2). We first consider the dark fermion χ\chi. Its gauge interactions are bilinear in χ\chi, and the dark Yukawa interaction also contains two fermionic fields. Moreover, the Lagrangian contains no operator involving a single χ\chi and only lighter particles. Consequently, any process mediated by the dark Yukawa interaction with one χ\chi in the initial state would require another particle of mass mχm_\chi in the final state and is therefore kinematically forbidden. The fermion χ\chi is thus always stable within the renormalizable model.

The situation is different for the complex scalar SS. Its gauge and portal interactions contain SS only through the bilinear S†SS^\dagger S and hence cannot mediate the decay of a single SS particle. However, the Yukawa interaction

12Sχˉc(ys+ypγ5)χ+H.c.(15)\frac{1}{2}S\bar{\chi}^{c}(y_s+y_p\gamma_5)\chi+\mathrm{H.c.} \tag*{(15)}

allows SS and S†S^\dagger to decay into pairs of dark antifermions and fermions, respectively, whenever these channels are kinematically open. Therefore, two qualitatively different regimes are possible:

mS>2mχ:S is unstable and χ is the only DM particle,mS<2mχ:S and χ are both stable.(16)\begin{aligned} m_S > 2m_\chi&: \qquad S\text{ is unstable and }\chi\text{ is the only DM particle},\\ m_S < 2m_\chi&: \qquad S\text{ and }\chi\text{ are both stable}. \tag*{(16)} \end{aligned}

Throughout this work, we focus exclusively on the two-component DM regime and therefore impose mS<2mχm_S < 2m_\chi. No additional discrete symmetry is then required to stabilize the two dark matter particles. Hence, our work differs from previous studies in the literature, where more symmetries are often invoked.

Interaction structure and model parameters

The phenomenology of the model is governed by three types of interactions. The U(1)B−LU(1)_{B-L} gauge coupling gBLg_{BL} connects both DM particles to the Z′Z^\prime and mediates annihilation into SM states as well as conversion between the two dark components. The scalar portal couplings λSH\lambda_{SH} and λSΦ\lambda_{S\Phi} connect SS to the physical scalars h1h_1 and h2h_2, providing an additional link between the scalar DM particle and the visible sector. The fermion χ\chi has no direct coupling to h1h_1 or h2h_2.

Finally, the dark Yukawa couplings ysy_s and ypy_p directly connect χ\chi and SS and generate conversion, semi-conversion, and semi-annihilation processes. These interactions couple the thermal evolution of the two dark matter components and constitute a defining feature of the model.

After fixing nχ=34n_\chi= \frac{3}{4}, mh1=125m_{h_1} = 125 GeV, vH=246v_H = 246 GeV, the independent parameters relevant to the DM phenomenology are

{mχ,mS,mZ′,mh2,gBL,ys,yp,λSH,λSΦ,sin⁡θ}.(17)\{m_\chi, m_S, m_{Z^\prime}, m_{h_2}, g_{BL}, y_s, y_p, \lambda_{SH}, \lambda_{S\Phi}, \sin\theta\}. \tag*{(17)}

The B−LB-L symmetry-breaking scale is fixed by vΦ=mZ′/(2gBL)v_\Phi= m_{Z^\prime}/(2g_{BL}), while λH\lambda_H, λΦ\lambda_\Phi, λHΦ\lambda_{H\Phi}, μH2\mu_H^2, and μΦ2\mu_\Phi^2 are determined by the physical scalar masses, mixing angle, and VEVs through the relations in § 2.2. The parameter μS2\mu_S^2 is similarly fixed by the physical mass mSm_S. The scalar self-coupling λS\lambda_S does not enter directly into the relic-density or detection observables considered below, but it enters the bounded-from-below conditions of § 2.3; throughout the numerical analysis we fix λS=1\lambda_S = 1.

For the phenomenological analysis, the model was implemented in FeynRules [27] and cross-checked against an independent implementation in LanHEP [28].

Relic abundance and dark-sector dynamics

We assume that both DM particles were initially in thermal and chemical equilibrium with the primordial plasma. As the temperature decreases below their masses, their abundances depart from equilibrium and eventually freeze out through a network of annihilation, conversion, semi-conversion, and semi-annihilation processes. Because the scalar and fermionic DM particles interact directly, their thermal histories are coupled and must be described by a system of Boltzmann equations.

We assume vanishing particle-antiparticle asymmetries and define the total comoving abundances of the two dark sectors as

Yχ≡nχ+nχˉs,YS≡nS+nS†s,(18)Y_\chi\equiv\frac{n_\chi+n_{\bar{\chi}}}{s}, \qquad Y_S \equiv\frac{n_S+n_{S^\dagger}}{s}, \tag*{(18)}

where ss is the entropy density. The corresponding equilibrium abundances are denoted by YχeqY_\chi^{\rm eq} and YSeqY_S^{\rm eq}. We use the dimensionless inverse-temperature variable

z≡μT,μ≡mχmSmχ+mS.(19)z \equiv\frac{\mu}{T}, \qquad\mu\equiv\frac{m_\chi m_S}{m_\chi+m_S}. \tag*{(19)}

where TT is the temperature of the thermal bath and μ\mu is the reduced mass of the two DM particles.

Following the sector notation adopted in micrOMEGAs, we label χ\chi and χˉ\bar{\chi} as sector 1, SS and S†S^\dagger as sector 2, and all remaining particles as sector 0 [29, 30]. In this notation, the coupled Boltzmann equations are

dYχdz=−λ1100[Yχ2−(Yχeq)2]−λ1120[Yχ2−YS(YχeqYSeq)2]−λ1122[Yχ2−YS2(YχeqYSeq)2](20)\frac{dY_\chi}{dz} = -\lambda_{1100}\left[Y_\chi^2-\left(Y_\chi^{\rm eq}\right)^2\right]-\lambda_{1120}\left[Y_\chi^2-Y_S\left(\frac{Y_\chi^{\rm eq}}{Y_S^{\rm eq}}\right)^2\right]-\lambda_{1122}\left[Y_\chi^2-Y_S^2\left(\frac{Y_\chi^{\rm eq}}{Y_S^{\rm eq}}\right)^2\right] \tag*{(20)}
dYSdz=−λ2200[YS2−(YSeq)2]−λ2211[YS2−Yχ2(YSeqYχeq)2]−12λ1210Yχ(YS−YSeq)+12λ1120[Yχ2−YS(YχeqYSeq)2](21)\frac{dY_S}{dz} = -\lambda_{2200}\left[Y_S^2-\left(Y_S^{\rm eq}\right)^2\right]-\lambda_{2211}\left[Y_S^2-Y_\chi^2\left(\frac{Y_S^{\rm eq}}{Y_\chi^{\rm eq}}\right)^2\right]-\frac{1}{2}\lambda_{1210}Y_\chi\left(Y_S-Y_S^{\rm eq}\right)+\frac{1}{2}\lambda_{1120}\left[Y_\chi^2-Y_S\left(\frac{Y_\chi^{\rm eq}}{Y_S^{\rm eq}}\right)^2\right] \tag*{(21)}

The temperature-dependent coefficients appearing in Eqs. (20) and (21) are defined as

λijkl(z)≡s(z)⟨σv⟩ijkl(z)zHˉ(z),(22)\lambda_{ijkl}(z) \equiv\frac{s(z)\langle\sigma v\rangle_{ijkl}(z)}{z\bar{H}(z)}, \tag*{(22)}

where ⟨σv⟩ijkl\langle\sigma v\rangle_{ijkl} is the thermally averaged cross section for a reaction involving particles from sectors i,j→k,li,j \to k,l [31]. The quantity

Hˉ(T)≡H(T)1+13dln⁡g∗s(T)dln⁡T(23)\bar{H}(T) \equiv\frac{H(T)}{1+\frac{1}{3}\frac{d\ln g_{*s}(T)}{d\ln T}} \tag*{(23)}

accounts for the temperature dependence of the effective number of entropy degrees of freedom, g∗sg_{*s}. In the limit in which g∗sg_{*s} is constant, one has Hˉ=H\bar{H}=H. The reaction classes contributing to the coupled thermal evolution are summarized in Table 3. Here X0X_0 denotes any kinematically accessible particle belonging to sector 0. Charge-conjugate reactions are included implicitly.

ProcessLabelClassification
χχˉ→X0X0\chi\bar{\chi} \to X_0X_01100Annihilation
χχ→S†X0\chi\chi\to S^\dagger X_0 and c.c.1120Semi-conversion
χχˉ→SS†\chi\bar{\chi} \to SS^\dagger1122Conversion
SS†→X0X0SS^\dagger\to X_0X_02200Annihilation
SS†→χχˉSS^\dagger\to\chi\bar{\chi}2211Conversion
χS→χˉX0\chi S \to\bar{\chi}X_0 and c.c.1210Semi-annihilation

Table 3. Reaction classes contributing to the thermal evolution of the two DM components. Sector 1 contains χ\chi and χˉ\bar{\chi}, sector 2 contains SS and S†S^\dagger, and sector 0 contains all remaining particles.

The annihilation processes 1100 and 2200 deplete the corresponding DM component into particles in the thermal bath. The conversion processes 1122 and 2211 redistribute the abundance between the two dark sectors without changing the total number of dark particles.

By contrast, the semi-conversion process 1120 transforms two fermionic DM particles into one scalar DM particle and one thermal plasma state. It therefore removes two particles from the fermionic sector while producing only one particle in the scalar sector, accounting for the relative factor of 1/21/2 in Eq. (21). Finally, the 1210 process depletes the scalar sector while leaving the total fermionic abundance unchanged and therefore contributes only to the evolution equation for YSY_S.

We solve the coupled system in Eqs. (20) and (21) numerically using micrOMEGAs 7.1.1 [29, 30, 32, 33]. The numerical calculation automatically includes all kinematically accessible channels belonging to the reaction classes listed in Table 3, including channels containing a Z′Z^\prime boson in the final state whenever they are open [34].

The present-day DM abundance is the sum of the scalar and fermionic contributions,

ΩDMh2=Ωχh2+ΩSh2.(24)\Omega_{\mathrm{DM}}h^2 = \Omega_{\chi}h^2 + \Omega_{S}h^2. \tag*{(24)}

In terms of the asymptotic abundances obtained by solving the coupled Boltzmann equations,

ΩDMh2≃2.74×108(mχYχ0+mSYS0) GeV−1,(25)\Omega_{\mathrm{DM}}h^2 \simeq2.74 \times10^8 \left(m_{\chi}Y_{\chi}^{0} + m_{S}Y_{S}^{0}\right)\ \mathrm{GeV}^{-1}, \tag*{(25)}

where Yχ0Y_{\chi}^{0} and YS0Y_{S}^{0} denote the present-day values of the corresponding comoving abundances. The observed cold DM abundance is [35]

ΩDMh2=0.120±0.001.\Omega_{\mathrm{DM}}h^2 = 0.120 \pm0.001.

We retain parameter points for which the sum of the two contributions is compatible with this value. It is also convenient to define the fractional contribution of each component to the total DM density,

ξχ≡ΩχΩDM,ξS≡ΩSΩDM,ξχ+ξS=1.(26)\xi_{\chi} \equiv\frac{\Omega_{\chi}}{\Omega_{\mathrm{DM}}}, \qquad\xi_{S} \equiv\frac{\Omega_{S}}{\Omega_{\mathrm{DM}}}, \qquad\xi_{\chi} + \xi_{S} = 1. \tag*{(26)}

These fractions will be used below to rescale the predicted direct- and indirect-detection signals of the two DM components.

In the next section, we confront the model with the relevant collider, direct-detection, and indirect-detection constraints.

Experimental constraints

Collider constraints

Collider searches constrain the mediators connecting the dark and visible sectors, namely the scalar states h1h_1 and h2h_2 and the U(1)B−LU(1)_{B-L} gauge boson Z′Z^\prime. Although the scalar and gauge sectors are related through the spontaneous breaking of U(1)B−LU(1)_{B-L}, it is convenient to discuss their collider constraints separately.

Higgs and scalar-sector constraints

The scalar singlet ΦBL\Phi_{BL} mixes with the neutral CP-even component of the SM Higgs doublet, giving rise to the physical states h1h_1 and h2h_2. We identify h1h_1 with the observed SM-like Higgs boson and fix mh1=125m_{h_1} = 125 GeV, whereas h2h_2 denotes the additional, predominantly B−LB-L scalar. The couplings of h1h_1 to SM particles are universally rescaled by cos⁡θ\cos\theta relative to their SM values [36]. We restrict the analysis to the region in which decays of h1h_1 into dark-sector particles are kinematically forbidden. Its production cross sections and visible partial decay widths are therefore rescaled by cos⁡2θ\cos^{2}\theta, while its total width is modified by the same factor. Consequently, the signal-strength modifier for any visible SM final state is

μh1≡σ(pp→h1)σSM(pp→h)BR⁡(h1→XSM)BR⁡SM(h→XSM)=cos⁡2θ.(27)\mu_{h_1} \equiv\frac{\sigma(pp \to h_1)}{\sigma_{\mathrm{SM}}(pp \to h)} \frac{\operatorname{BR}(h_1 \to X_{\mathrm{SM}})}{\operatorname{BR}_{\mathrm{SM}}(h \to X_{\mathrm{SM}})} = \cos^{2}\theta. \tag*{(27)}

The ATLAS and CMS Higgs signal-strength measurements adopted in this analysis are [37, 38] μATLAS=1.05±0.06\mu_{\mathrm{ATLAS}} = 1.05 \pm0.06, μCMS=1.00±0.06\mu_{\mathrm{CMS}} = 1.00 \pm0.06, which yield the approximate combined value μATLAS+CMS≃1.03±0.06\mu_{\mathrm{ATLAS+CMS}} \simeq1.03 \pm0.06. Requiring consistency at the 95% confidence level gives the bound

∣sin⁡θ∣≲0.27.(28)\lvert\sin\theta\rvert\lesssim0.27. \tag*{(28)}

Direct searches for the additional scalar h2h_2 provide complementary constraints that depend on its mass, mixing angle, and branching fractions. Since h2h_2 couples to SM particles through its Higgs-doublet component, its production cross sections through SM mechanisms are suppressed by sin⁡2θ\sin^{2}\theta: σ(pp→h2)=sin⁡2θ σSM(mh2)\sigma(pp \to h_2) = \sin^{2}\theta\,\sigma_{\mathrm{SM}}(m_{h_2}), where σSM(mh2)\sigma_{\mathrm{SM}}(m_{h_2}) is the production cross section of an SM Higgs boson evaluated at the mass mh2m_{h_2}. Gluon fusion provides the dominant production mechanism over most of the mass range, while vector-boson fusion and associated production offer additional sensitivity.

The most relevant searches involve the decay channels pp→h2→WWpp \to h_2 \to WW, pp→h2→ZZpp \to h_2 \to ZZ, and pp→h2→h1h1pp \to h_2 \to h_1h_1. We impose the 95% confidence-level limits from the ATLAS and CMS heavy-scalar searches [39, 40, 41, 42, 43, 44, 45], using the recast described in Ref. [46]. For every parameter point, the predicted production rate multiplied by the corresponding branching fraction is compared with the experimental upper limit:

σ(pp→h2)BR⁡(h2→X)≤[σ(pp→h2)BR⁡(h2→X)]95% CL,(29)\sigma(pp \to h_2)\operatorname{BR}(h_2 \to X) \leq\left[\sigma(pp \to h_2)\operatorname{BR}(h_2 \to X)\right]_{95\%\,\mathrm{CL}}, \tag*{(29)}

with X=WW,ZZ,h1h1X = WW, ZZ, h_1h_1. The branching fractions of h2h_2 are evaluated within the full model. In addition to its decays into SM particles, the channels h2→h1h1h_2 \to h_1h_1 and h2→SS†h_2 \to SS^{\dagger} may modify its total width whenever they are kinematically open. Possible decays into other non-SM states must likewise be included when allowed.

Searches for the Z′Z^{\prime} gauge boson

The Z′Z^{\prime} boson couples to SM quarks and leptons according to their B−LB-L charges and can therefore be resonantly produced at hadron colliders through the Drell–Yan process, pp→Z′→ffˉpp \to Z^{\prime} \to f\bar{f}. Searches in dilepton final states provide particularly strong constraints because of their clean experimental signature, pp→Z′→ℓ+ℓ−pp \to Z^{\prime} \to\ell^{+}\ell^{-}, ℓ=e,μ\ell= e,\mu, with dedicated searches performed by both ATLAS and CMS [47, 48]. Dijet and top-quark final states provide complementary, though typically weaker, sensitivity [49, 50].

The predicted dilepton rate, σ(pp→Z′)BR⁡(Z′→ℓ+ℓ−)\sigma(pp \to Z^{\prime})\operatorname{BR}(Z^{\prime} \to\ell^{+}\ell^{-}), depends on mZ′m_{Z^{\prime}}, gBLg_{\mathrm{BL}}, and the total Z′Z^{\prime} width. In particular, the branching fraction into charged leptons can be reduced when decays into dark matter particles or other non-SM states are kinematically accessible. Collider constraints on mZ′m_{Z^{\prime}} must therefore be applied as a function of gBLg_{\mathrm{BL}} and of the complete particle spectrum, rather than as a universal lower bound on the Z′Z^{\prime} mass [51, 52]. For every parameter point, we compare the predicted dilepton signal with the 95% confidence-level upper limits reported by ATLAS and CMS [47, 48],

σ(pp→Z′)BR⁡(Z′→ℓ+ℓ−)≤[σ(pp→Z′)BR⁡(Z′→ℓ+ℓ−)]95% CL.(30)\sigma(pp \to Z^{\prime})\operatorname{BR}(Z^{\prime} \to\ell^{+}\ell^{-}) \leq\left[\sigma(pp \to Z^{\prime})\operatorname{BR}(Z^{\prime} \to\ell^{+}\ell^{-})\right]_{95\%\,\mathrm{CL}}. \tag*{(30)}

Motivated by the strong sensitivity of current dilepton searches to B−LB-L gauge bosons in the multi-TeV mass range [51, 52], we restrict the numerical scan to mZ′≥4m_{Z^\prime} \ge4 TeV. This lower limit defines the mass range explored in the present analysis; it should not be interpreted as a coupling-independent experimental exclusion, since the collider sensitivity decreases rapidly for small values of gBLg_{BL} and depends on the available Z′Z^\prime decay modes.

Direct detection bounds

Direct-detection experiments search for nuclear recoils induced by dark matter scattering off nucleons [53]. In the present model, both dark matter components give rise to spin-independent (SI) interactions. For the fermion χ\chi, the relevant contributions arise from tree-level Z′Z^\prime exchange and from one-loop scalar exchange induced by the dark Yukawa interaction. For the scalar SS, both Z′Z^\prime and h1,2h_{1,2} exchange contribute already at tree level [54]. These processes are illustrated in Figure 1.

Figure 1. Relevant contributions to the SI dark matter–nucleon scattering amplitudes. For the fermionic component χ\chi, the leading contributions arise from tree-level Z′Z^\prime exchange and one-loop scalar exchange. For the scalar component SS, scattering proceeds at tree level through Z′Z^\prime and h1,2h_{1,2} exchange.

We define the trilinear scalar interactions through

L⊃−∑i=1,2gSShihiS†S,(31)\mathcal{L} \supset- \sum_{i=1,2} g_{SSh_i} h_i S^\dagger S, \tag*{(31)}

with

gSSh1=λSHvHcos⁡θ−λSΦvΦsin⁡θ,(32)g_{SSh_1} = \lambda_{SH}v_H \cos\theta- \lambda_{S\Phi}v_\Phi\sin\theta, \tag*{(32)}
gSSh2=λSHvHsin⁡θ+λSΦvΦcos⁡θ.(33)g_{SSh_2} = \lambda_{SH}v_H \sin\theta+ \lambda_{S\Phi}v_\Phi\cos\theta. \tag*{(33)}

The couplings of the physical scalars to SM particles are controlled by κ1=cos⁡θ\kappa_1 = \cos\theta and κ2=sin⁡θ\kappa_2 = \sin\theta.

For the fermionic component, the effective coefficient associated with tree-level Z′Z^\prime exchange is

CχZ′=gBL2nχnNmZ′2,(34)C_\chi^{Z^\prime} = \frac{g_{BL}^2 n_\chi n_N}{m_{Z^\prime}^2}, \tag*{(34)}

where nN=1n_N = 1 is the effective B−LB-L charge of either a proton or a neutron [51]. The scalar contribution is generated at one loop and can be written as

Cχh=fNmNvH∑i=1,2κigSShimhi2∣ys∣2f(r)+∣yp∣2g(r)16π2mχ,(35)C_\chi^h = \frac{f_N m_N}{v_H} \sum_{i=1,2} \frac{\kappa_i g_{SSh_i}}{m_{h_i}^2} \frac{|y_s|^2 f(r) + |y_p|^2 g(r)}{16\pi^2 m_\chi}, \tag*{(35)}

where mNm_N is the nucleon mass, fN≃0.3f_N \simeq0.3 is the scalar nucleon form factor [55], and r=mS2mχ2r = \frac{m_S^2}{m_\chi^2}.

The loop functions are [14, 56, 57]

f(r)=r2−5r+4(4−r)rarctan⁡(4−rr)+12[2−(r−3)ln⁡r],(36)f(r) = \frac{r^2 - 5r + 4}{\sqrt{(4-r)r}} \arctan\left(\sqrt{\frac{4-r}{r}}\right) + \frac{1}{2}\left[2 - (r-3)\ln r\right], \tag*{(36)}
g(r)=(r−3)r4−rarctan⁡(4−rr)+12[2−(r−1)ln⁡r].(37)g(r) = (r-3)\sqrt{\frac{r}{4-r}} \arctan\left(\sqrt{\frac{4-r}{r}}\right) + \frac{1}{2}\left[2 - (r-1)\ln r\right]. \tag*{(37)}

These expressions are real throughout the two-component dark matter regime, mS<2mχm_S < 2m_\chi. The contribution proportional to ∣yp∣2|y_p|^2 is not velocity suppressed, whereas the mixed term proportional to ysypy_s y_p is suppressed in the non-relativistic limit.

For the scalar component, the corresponding tree-level coefficients are

CSZ′=gBL2nSnNmZ′2,CSh=mNfN2mSvH∑i=1,2κigSShimhi2,(38)C_S^{Z'} = \frac{g_{BL}^2 n_S n_N}{m_{Z'}^2}, \qquad C_S^h = \frac{m_N f_N}{2m_S v_H}\sum_{i=1,2}\frac{\kappa_i g_{SSh_i}}{m_{h_i}^2}, \tag*{(38)}

with nS=−2nχn_S = -2n_\chi.

Assuming vanishing particle–antiparticle asymmetries, the observable SI cross sections are obtained by averaging over χ\chi, χˉ\bar{\chi} and SS, S†S^\dagger. Since the vector contribution changes sign under charge conjugation whereas the scalar contribution does not, the interference between them cancels in the averaged rates. The resulting cross sections are

σ‾χNSI=μχN2π(∣CχZ′∣2+∣Cχh∣2),(39)\overline{\sigma}^{\mathrm{SI}}_{\chi N} = \frac{\mu_{\chi N}^2}{\pi}\left(|C_\chi^{Z'}|^2 + |C_\chi^h|^2\right), \tag*{(39)}
σ‾SNSI=μSN2π(∣CSZ′∣2+∣CSh∣2),(40)\overline{\sigma}^{\mathrm{SI}}_{S N} = \frac{\mu_{S N}^2}{\pi}\left(|C_S^{Z'}|^2 + |C_S^h|^2\right), \tag*{(40)}

where

μXN=mXmNmX+mN,X=χ,S.(41)\mu_{X N} = \frac{m_X m_N}{m_X + m_N}, \qquad X = \chi, S. \tag*{(41)}

The interference between h1h_1 and h2h_2 remains encoded in the coherent scalar amplitudes CχhC_\chi^h and CShC_S^h.

We compare the predicted SI cross sections with the XENONnT and LZ results and use the projected DARWIN/XLZD sensitivity to illustrate the future experimental reach [58, 59, 60]. Since each species constitutes only a fraction of the local dark matter density, the quantities relevant for comparison with the single-species experimental limits are

σX,Nth=ξXσ‾XNSI,ξX=ΩXΩDM,X=χ,S.(42)\sigma^{\mathrm{th}}_{X,N} = \xi_X \overline{\sigma}^{\mathrm{SI}}_{X N}, \qquad\xi_X = \frac{\Omega_X}{\Omega_{\mathrm{DM}}}, \qquad X = \chi, S. \tag*{(42)}

The rescaling is linear in ξX\xi_X because the direct-detection event rate is proportional to the local density of the incident dark matter component.

Indirect detection limits

Dark matter annihilation in regions of large density can produce gamma-ray signals observable by present and future telescopes. We use continuum gamma-ray limits from Fermi-LAT [61] and H.E.S.S. [62], together with the projected sensitivity of CTAO [63, 64], to estimate the indirect-detection reach of the model. Since our aim is not to perform a dedicated gamma-ray likelihood analysis, these bounds are applied as approximate constraints on the present-day annihilation and semi-conversion rates.

For conventional pair annihilation of either dark matter component, we define

⟨σv⟩ffˉtot≡∑f⟨σv⟩XXˉ→ffˉ,X=χ,S,(43)\langle\sigma v\rangle^{\mathrm{tot}}_{f\bar{f}} \equiv\sum_f \langle\sigma v\rangle_{X\bar{X}\to f\bar{f}}, \qquad X=\chi,S, \tag*{(43)}

where the sum runs over kinematically accessible charged SM fermions. Because each species constitutes only a fraction of the total dark matter density, the corresponding gamma-ray flux is suppressed by the square of its relic fraction. The effective rate is therefore [13, 14, 65]

⟨σv⟩XXˉ→ffˉeff=ξX2⟨σv⟩ffˉtot.(44)\langle\sigma v\rangle^{\mathrm{eff}}_{X\bar{X}\to f\bar{f}}=\xi_X^2\langle\sigma v\rangle^{\mathrm{tot}}_{f\bar{f}}. \tag*{(44)}

The model also contains semi-conversion channels of the form χχ→S†hi\chi\chi\to S^\dagger h_i, i=1,2i=1,2, together with their charge-conjugate processes. These reactions can provide an important contribution to the gamma-ray signal through the subsequent decays of hih_i. Their comparison with standard annihilation limits requires accounting for the modified kinematics and decay spectra of the visible state in the final state [13, 66, 67]. We therefore use the corresponding semi-annihilation recasting prescription as an approximate estimate of the constraint, with the overall normalization rescaled by the fermionic relic fraction,

⟨σv⟩χχ→S†hieff=ξχ2⟨σv⟩χχ→S†hi.(45)\langle\sigma v\rangle^{\mathrm{eff}}_{\chi\chi\to S^\dagger h_i}=\xi_\chi^2\langle\sigma v\rangle_{\chi\chi\to S^\dagger h_i}. \tag*{(45)}

The resulting indirect-detection limits should consequently be interpreted as indicative rather than as a dedicated spectral recast.

Numerical Results

We explore the phenomenology of the model through a global scan of its multidimensional parameter space using the differential-evolution algorithm implemented in SciPy. The scan ranges are chosen to satisfy perturbativity requirements and the kinematic condition mS<2mχm_S < 2m_\chi, which guarantees the stability of the scalar DM component. We consider

mχ∈[100,3000] GeV,mSmχ∈[0.05,1.95],mZ′∈[4000,10000] GeV,mh2∈[200,4000] GeV,gBL∈[10−3,1],∣ys∣,∣yp∣∈[10−2,3],∣λSH∣,∣λSΦ∣∈[10−3,3],sin⁡θ∈[10−6,1].(46)\begin{aligned} m_\chi&\in[100,3000]~\mathrm{GeV}, & \frac{m_S}{m_\chi} &\in[0.05,1.95],\\ m_{Z'} &\in[4000,10000]~\mathrm{GeV}, & m_{h_2} &\in[200,4000]~\mathrm{GeV},\\ g_{BL} &\in[10^{-3},1], & |y_s|,|y_p| &\in[10^{-2},3],\\ |\lambda_{SH}|,|\lambda_{S\Phi}| &\in[10^{-3},3], & \sin\theta&\in[10^{-6},1]. \tag*{(46)} \end{aligned}

For each trial point, we first impose the theoretical consistency conditions discussed in § 2 together with the collider constraints of § 4.1. We then pass points that satisfy these requirements to micrOMEGAs to solve the coupled Boltzmann equations and compute the individual relic abundances Ωχh2\Omega_\chi h^2 and ΩSh2\Omega_S h^2. The scan is guided by a χ2\chi^2-like objective function that favors points satisfying Planck observation [35]

Ωχh2+ΩSh2=0.120±0.001.(47)\Omega_\chi h^2+\Omega_S h^2=0.120\pm0.001. \tag*{(47)}

Several independent differential-evolution runs are combined in order to improve the coverage of the allowed multidimensional parameter space.

Once the relic abundances are known, the corresponding fractions ξχ\xi_\chi and ξS\xi_S are used to rescale the direct- and indirect-detection rates. Current direct-detection constraints are then imposed on both DM components following the prescription described in § 4.2.

Throughout this section, we refer to a parameter point as viable if it satisfies the theoretical consistency conditions, the collider constraints, the observed total DM abundance, and the current direct-detection bounds. Indirect-detection constraints are not included in this definition; instead, they are applied subsequently to the viable sample.

The scatter plots shown below represent projections of this viable multidimensional parameter space. Since differential evolution is an optimization algorithm rather than a statistical sampling procedure, a statistical interpretation should not be assigned to the density of points in a particular region.

To quantify the effect of the different reaction classes on the relic abundance, we also recompute the freeze-out evolution after switching off one process at a time. For a DM component X=χ,SX = \chi, S, we define

ζXijkl≡ΩX(−ijkl)ΩXfull,(48)\zeta_X^{ijkl} \equiv\frac{\Omega_X^{(-ijkl)}}{\Omega_X^{\mathrm{full}}}, \tag*{(48)}

where ΩX(−ijkl)\Omega_X^{(-ijkl)} is the relic abundance obtained after removing the reaction class ijkl∈{1100,2200,1120,1210,1122,2211}ijkl \in\{1100, 2200, 1120, 1210, 1122, 2211\}, and ΩXfull\Omega_X^{\mathrm{full}} denotes the abundance obtained with all processes included. Thus, ζXijkl=1\zeta_X^{ijkl} = 1 indicates that removing the corresponding process does not affect the final abundance, whereas deviations from unity quantify its impact.

The numerical results are presented separately for the two possible mass hierarchies, mS>mχm_S > m_\chi and mS<mχm_S < m_\chi.

The mS>mχm_S > m_\chi regime

We first consider the hierarchy mS>mχm_S > m_\chi, for which the fermion is the lighter DM particle. Figure 2 shows the relic-density fractions ξχ\xi_\chi and ξS\xi_S for the viable points projected onto selected model parameters.

Relic-density fractions $\xi_\chi$ and $\xi_S$ as functions of selected model parameters

Figure 2. Relic-density fractions ξχ\xi_\chi and ξS\xi_S as functions of selected model parameters for mS>mχm_S > m_\chi. Blue triangles and red circles denote the fermionic and scalar components, respectively. The upper panels use linear scales and show the dependence on mχm_\chi and mS/mχm_S/m_\chi, while the lower panels use logarithmic scales and show the dependence on the dark Yukawa couplings. All displayed points satisfy the theoretical constraints, collider bounds, the observed total relic abundance, and current direct-detection limits. Indirect-detection constraints are not imposed.

A clear feature of this regime is that the fermionic component dominates the relic density throughout most of the viable parameter space. Nevertheless, viable points with a sizable scalar contribution are also present, and the two components can contribute comparable fractions in restricted regions. Large mass splittings tend to favor a strongly fermion-dominated relic density, while the dependence on the dark Yukawa couplings is more pronounced for ∣yp∣|y_p| than for ∣ys∣|y_s|. Moreover, the scalar fraction can become highly suppressed, spanning several orders of magnitude.

The processes responsible for this pattern can be identified more directly through the off-channel analysis shown in Figure 3. For the fermionic component, the contrast between the 1100 annihilation and the 1120 semi-conversion processes is particularly clear. Removing the 1100 channel leaves Ωχ\Omega_\chi close to its full value for a large fraction of the viable points. By contrast, switching off 1120 can increase Ωχ\Omega_\chi by several orders of magnitude. Thus, over a substantial part of the viable parameter space, the fermionic abundance is controlled primarily by the semi-conversion reaction χχ→S†X0\chi\chi\to S^\dagger X_0 and its charge-conjugate process, rather than by conventional χχˉ\chi\bar{\chi} annihilation.

Impact of individual reaction classes on the relic abundances for $m_S > m_\chi$

Figure 3. Impact of individual reaction classes on the relic abundances for mS>mχm_S > m_\chi. The left and right panels show ζχijkl=Ωχ(−ijkl)/Ωχfull\zeta_{\chi}^{ijkl} = \Omega_{\chi}^{(-ijkl)}/\Omega_{\chi}^{\mathrm{full}} and ζSijkl=ΩS(−ijkl)/ΩSfull\zeta_{S}^{ijkl} = \Omega_{S}^{(-ijkl)}/\Omega_{S}^{\mathrm{full}}, respectively, as functions of the lighter DM mass mχm_\chi. The superscript (−ijkl)(-ijkl) denotes a calculation in which reaction class ijklijkl has been removed. The reference value ζXijkl=1\zeta_{X}^{ijkl} = 1 corresponds to no change relative to the full calculation. Values above unity indicate that the removed process normally depletes component XX, whereas values below unity indicate that it contributes, directly or indirectly, to its production. The 1120 contribution is not displayed in the SS panel because removing this process leads to an extremely strong suppression of the scalar relic abundance.

The same reaction has the opposite effect on the scalar component: the 1120 process produces one scalar DM particle and therefore acts as a source for SS. Indeed, removing this channel leads to an extremely strong suppression of ΩS\Omega_S and is not shown in the right panel of Figure 3 because of the resulting dynamic range. On the other hand, removing the

2200 annihilation, 2211 conversion, or 1210 semi-annihilation processes generally increases ΩS\Omega_S, showing that these reactions contribute to the depletion of the scalar population. The 1120 semi-conversion therefore plays a dual role in this hierarchy, simultaneously reducing the fermionic abundance and feeding the scalar component.

Figure 4 shows complementary projections of the viable parameter space. Viable solutions occur both near and well away from the Z′Z' resonance conditions, demonstrating that the observed relic abundance does not require resonant annihilation. The dark Yukawa couplings also span broad ranges, with no narrow correlation between ∣ys∣|y_s| and ∣yp∣|y_p| selected by the combined constraints.

Selected projections of the viable parameter space for $m_S > m_\chi$

Figure 4. Selected projections of the viable parameter space for mS>mχm_S > m_\chi. The upper-left panel shows the gauge sector in the (mZ′,gBL)(m_{Z^\prime}, g_{BL}) plane. The upper-right and lower-left panels show mZ′m_{Z^\prime} as a function of mχm_\chi and mSm_S, respectively, with reference lines indicating the resonance conditions mZ′=2mχm_{Z^\prime} = 2m_\chi and mZ′=2mSm_{Z^\prime} = 2m_S. The lower-right panel shows the correlation between the dark Yukawa couplings ∣ys∣|y_s| and ∣yp∣|y_p|. All displayed points satisfy the theoretical constraints, collider bounds, the observed total relic abundance, and current direct-detection limits.

Figure 5 summarizes the detection prospects in this hierarchy. Current direct-detection bounds are already imposed on the displayed sample, but part of the surviving parameter space remains within the projected sensitivity of DARWIN/XLZD. Current gamma-ray limits provide only an additional restriction on part of this sample, while CTAO offers complementary future sensitivity.

Figure 5. Direct- and indirect-detection predictions for the viable points with mS>mχm_S > m_\chi. The upper panels show the rescaled SI scattering cross sections for χ\chi (left) and SS (right), together with the current LZ limit and the projected DARWIN/XLZD sensitivity. The lower panel shows the effective gamma-ray annihilation rate together with the current Fermi-LAT and H.E.S.S. limits and the projected CTAO sensitivity. Current direct-detection constraints are already imposed on the displayed sample, whereas indirect-detection constraints are shown but not imposed.

The mS<mχm_S < m_\chi regime

We now consider the opposite hierarchy, mS<mχm_S < m_\chi, in which the scalar is the lighter DM particle. Figure 6 shows the relic-density fractions ξχ\xi_\chi and ξS\xi_S for the viable points projected onto the same set of representative model parameters used in the previous subsection.

Relic-density fractions for the fermionic and scalar components as functions of selected model parameters

Figure 6. Relic-density fractions ξχ\xi_\chi and ξS\xi_S as functions of selected model parameters for mS<mχm_S<m_\chi. Blue triangles and red circles denote the fermionic and scalar components, respectively. The upper panels use linear scales and show the dependence on mχm_\chi and mS/mχm_S/m_\chi, while the lower panels use logarithmic scales and show the dependence on the dark Yukawa couplings. All displayed points satisfy the theoretical constraints, collider bounds, the observed total relic abundance, and current direct-detection limits. Indirect-detection constraints are not imposed.

In contrast to the mS>mχm_S > m_\chi regime, viable solutions now span a much broader range of relic compositions. Scalar-dominated, fermion-dominated, and mixed configurations are all possible, even though SS is the lighter state. Thus, the mass ordering alone does not determine which particle dominates the present-day abundance. Either component can become highly subdominant, and the relic composition shows a more pronounced dependence on ∣yp∣\lvert y_p\rvert than on ∣ys∣\lvert y_s\rvert.

The off-channel analysis in Figure 7 identifies the processes responsible for these relic compositions. For the fermionic component, both the 1120 semi-conversion and 1122 conversion processes can substantially affect the final abundance. Removing either process can significantly increase Ωχ\Omega_\chi, showing that reactions transferring fermionic dark matter into the scalar sector provide important depletion mechanisms for χ\chi. Conventional 1100 annihilation generally has a smaller impact.

Impact of individual reaction classes on the relic abundances for $m_S < m_\chi$

Figure 7. Impact of individual reaction classes on the relic abundances for mS<mχm_S < m_\chi. The left and right panels show ζχijkl=Ωχ(−ijkl)/Ωχfull\zeta_\chi^{ijkl} = \Omega_\chi^{(-ijkl)}/\Omega_\chi^{\mathrm{full}} and ζSijkl=ΩS(−ijkl)/ΩSfull\zeta_S^{ijkl} = \Omega_S^{(-ijkl)}/\Omega_S^{\mathrm{full}}, respectively, as functions of the lighter DM mass mSm_S. The superscript (−ijkl)(-ijkl) denotes a calculation in which reaction class ijklijkl has been removed. The reference value ζXijkl=1\zeta_X^{ijkl} = 1 corresponds to no change relative to the full calculation. Values above unity indicate that the removed process normally contributes to the depletion of component XX, whereas values below unity indicate that it contributes, directly or indirectly, to its production.

The scalar abundance is instead particularly sensitive to the 2200 annihilation channel, whose removal can strongly increase ΩS\Omega_S. The 1120 semi-conversion plays the opposite role by feeding the scalar population, while 2211 conversion and 1210 semi-annihilation provide additional, point-dependent modifications. The final composition reflects the competition between scalar depletion and fermion-to-scalar transfer.

Figure 8 shows complementary projections of the viable parameter space for the mS<mχm_S < m_\chi hierarchy. As in the opposite hierarchy, viable solutions occur both near and well away from the Z′Z' resonance conditions, confirming that resonant annihilation is not required. The dark Yukawa couplings also span broad ranges, with no preferred relation between ∣ys∣\lvert y_s\rvert and ∣yp∣|y_p| selected by the combined constraints. Figure 9 summarizes the detection prospects in this hierarchy. Current direct-detection bounds are already imposed on the displayed sample, while part of the surviving parameter space remains within the projected sensitivity of DARWIN/XLZD. Current gamma-ray limits exclude only part of this sample, and CTAO provides complementary future sensitivity.

Selected projections of the viable parameter space for $m_S < m_\chi$

Figure 8. Selected projections of the viable parameter space for mS<mχm_S < m_\chi. The upper-left panel shows the gauge sector in the (mZ′,gBL)(m_{Z'},g_{BL}) plane. The upper-right and lower-left panels show mZ′m_{Z'} as a function of mχm_\chi and mSm_S, respectively, with reference lines indicating the resonance conditions mZ′=2mχm_{Z'} = 2m_\chi and mZ′=2mSm_{Z'} = 2m_S. The lower-right panel shows the correlation between the dark Yukawa couplings ∣ys∣|y_s| and ∣yp∣|y_p|. All displayed points satisfy the theoretical constraints, collider bounds, the observed total relic abundance, and current direct-detection limits.

Figure 9. Direct- and indirect-detection predictions for the viable points with mS<mχm_S < m_\chi. The upper panels show the rescaled SI scattering cross sections for χ\chi (left) and SS (right), together with the current LZ limit and the projected DARWIN/XLZD sensitivity. The lower panel shows the effective gamma-ray annihilation rate together with the current Fermi-LAT and H.E.S.S. limits and the projected CTAO sensitivity. Current direct-detection constraints are already imposed on the displayed sample, whereas indirect-detection constraints are shown but not imposed.

Discussion

The numerical results reveal a common physical picture across the two mass hierarchies. The relic abundances of χ\chi and SS cannot, in general, be understood as the outcome of two approximately independent freeze-out processes, since reactions connecting the two dark sectors can both deplete one component and feed the other. Consequently, the mass ordering alone does not determine the relic composition. Figure 10 illustrates this result directly by combining the viable points from both hierarchies: scalar- and fermion-dominated solutions occur on either side of mS=mχm_S = m_\chi, and the lighter dark matter particle need not be the dominant component.

Ratio of the scalar and fermionic relic abundances

Figure 10. Ratio of the scalar and fermionic relic abundances, ΩS/Ωχ\Omega_S/\Omega_\chi, as a function of the dark matter mass ratio mS/mχm_S/m_\chi, combining the viable points from both mass hierarchies. The line mS/mχ=1m_S/m_\chi= 1 separates the mS<mχm_S < m_\chi and mS>mχm_S > m_\chi regimes, while ΩS/Ωχ=1\Omega_S/\Omega_\chi= 1 corresponds to equal relic abundances. Points above (below) this value are scalar- (fermion-) dominated. All displayed points satisfy the theoretical constraints, collider bounds, the observed total relic abundance, and current direct-detection limits.

This behavior is particularly clear for mS>mχm_S > m_\chi, where the [11, 20] semi-conversion process can strongly deplete the fermionic abundance while simultaneously sourcing the scalar population. For mS<mχm_S < m_\chi, the final composition instead results from a broader competition among annihilation, conversion, semi-annihilation, and semi-conversion processes.

A second important feature is that the correct relic abundance does not require resonant enhancement. Viable solutions populate broad regions away from the Z′Z^\prime resonance, where the freeze-out dynamics is controlled by the interplay among the gauge portal, the scalar portal, and the dark-Yukawa interaction. The gauge coupling gBLg_{BL} connects both dark matter particles to the Z′Z^\prime, the scalar portal couplings connect SS to the visible sector through h1h_1 and h2h_2, and the couplings ysy_s and ypy_p directly connect χ\chi and SS. The relic composition is therefore determined by the combined action of these interactions rather than by any single coupling or by the mass hierarchy alone.

The numerical analysis presented above was performed for the benchmark charge assignment nχ=3/4n_\chi= 3/4. It is therefore important to determine which of these features are specific to this choice and which persist for other B−LB-L charges. For generic values of nχn_\chi outside the special cases listed in Table 2, the renormalizable interaction structure of the model remains unchanged: the relation QB−L(S)=−2nχQ_{B-L}(S) = -2n_\chi continues to allow the dark Yukawa interaction, and the same classes of annihilation, conversion, semi-annihilation, and semi-conversion processes remain present. Changing nχn_\chi, however, modifies the strength of the gauge interactions of both dark matter particles and can therefore alter the balance between the Z′Z^\prime portal. and the Yukawa-driven dark-sector dynamics. To assess this dependence, we compare the benchmark with one smaller and one larger non-exceptional charge assignment,

nχ=25,34,75.(49)n_\chi=\frac{2}{5},\quad\frac{3}{4},\quad\frac{7}{5}. \tag*{(49)}

Since nS=−2nχn_S=-2n_\chi, varying nχn_\chi directly affects the interactions mediated by the Z′Z^\prime gauge boson, whereas the Yukawa and scalar interactions do not depend explicitly on nχn_\chi.

The relative importance of the Z′Z^\prime interaction is not the same for all reaction classes. The fermionic annihilation process 11→0011\to00 is particularly simple. Since χ\chi has no tree-level coupling to h1h_1 or h2h_2, its annihilation into sector-0 states proceeds predominantly through the Z′Z^\prime portal. Moreover, throughout the scan mχ<mZ′m_\chi<m_{Z^\prime}, so annihilation into a pair of on-shell Z′Z^\prime bosons is kinematically forbidden. Away from the Z′Z^\prime resonance and kinematic thresholds, the corresponding rate therefore scales as ⟨σv⟩11→00∝nχ2gBL4\langle\sigma v\rangle_{11\to00}\propto n_\chi^2g_{BL}^4.

To isolate this charge dependence numerically, we take each viable point from the benchmark scan with nχ=3/4n_\chi=3/4 and recompute the reaction rates after changing only nχn_\chi, while keeping all masses and couplings fixed. We define

Rijkl(nχ)=log⁡10[⟨σv⟩ijkl(nχ)⟨σv⟩ijkl(3/4)].(50)R_{ijkl}^{(n_\chi)} = \log_{10}\left[\frac{\langle\sigma v\rangle_{ijkl}^{(n_\chi)}}{\langle\sigma v\rangle_{ijkl}^{(3/4)}}\right]. \tag*{(50)}

The ratios Rijkl(nχ)R_{ijkl}^{(n_\chi)} therefore isolate the explicit dependence of each reaction rate on the dark B−LB-L charge. For the two alternative charge assignments, this scaling predicts

R1100(2/5)=2log⁡10(2/53/4)≃−0.55,R1100(7/5)=2log⁡10(7/53/4)≃+0.54.(51)R_{1100}^{(2/5)} = 2\log_{10}\left(\frac{2/5}{3/4}\right) \simeq-0.55,\qquad R_{1100}^{(7/5)} = 2\log_{10}\left(\frac{7/5}{3/4}\right) \simeq+0.54. \tag*{(51)}

Thus, if the 1100 process is indeed controlled by the Z′Z^\prime portal, the numerical results should form approximately horizontal bands at these two values. The corresponding rate ratios are shown in Figure 11.

Dependence of representative annihilation and conversion rates on the dark $B-L$ charge

Figure 11. Dependence of representative annihilation and conversion rates on the dark B−LB-L charge. We show the logarithmic ratios Rijkl(nχ)R_{ijkl}^{(n_\chi)} defined in Eq. (50), using nχ=3/4n_\chi= 3/4 as the reference value and comparing it with nχ=2/5n_\chi= 2/5 and 7/57/5. The left panel shows the 1100 annihilation and 1122 conversion processes as functions of mχm_\chi, while the right panel shows the 2200 annihilation and 2211 conversion processes as functions of mSm_S.

The situation is different for the scalar annihilation and conversion processes. The 2200 rate receives both gauge and scalar-portal contributions, while χχˉ↔SS†\chi\bar{\chi} \leftrightarrow SS^\dagger receives, in addition to ss-channel Z′Z^\prime exchange, contributions involving two dark-Yukawa vertices. These latter contributions carry no explicit dependence on nχn_\chi. Since the Z′Z^\prime is heavy and gBLg_{BL} is restricted by collider and direct-detection constraints, whereas the scalar-portal and dark-Yukawa couplings can remain sizable, there is no a priori reason for the 2200, 1122, or 2211 rates to be dominated by the gauge contribution.

The behavior of the 1100 process provides a particularly clear verification of the expected charge dependence. The points form two approximately horizontal bands at R1100(2/5)≃−0.55R_{1100}^{(2/5)} \simeq-0.55 and R1100(7/5)≃+0.54R_{1100}^{(7/5)} \simeq+0.54, in excellent agreement with Eq. (51). This confirms that the dependence of the fermionic annihilation rate on nχn_\chi is controlled primarily by the Z′Z^\prime portal.

The remaining processes behave quite differently. The 2200 scalar annihilation rate shows essentially no systematic dependence on nχn_\chi, while the 1122 and 2211 conversion rates remain clustered around Rijkl(nχ)=0R_{ijkl}^{(n_\chi)} = 0, apart from isolated points with larger deviations. Thus, although their gauge-mediated contributions depend explicitly on the dark B−LB-L charge, their total rates are largely insensitive to nχn_\chi. This behavior reflects the competing scalar and dark-Yukawa contributions, which carry no explicit charge dependence and dilute the dependence associated with the Z′Z' portal.

The fixed-parameter comparison above isolates the microscopic dependence of the reaction rates on nχn_\chi. We now ask whether these changes propagate to the full phenomenology once the remaining model parameters are allowed to vary. To this end, we repeat the complete numerical analysis for nχ=2/5n_\chi= 2/5 and 7/57/5 and compare the resulting viable samples with the benchmark nχ=3/4n_\chi= 3/4. The resulting comparison is shown in Figure 12.

Comparison of the viable parameter space: (a) Relic composition; (b) Gauge-sector parameter space

Figure 12. (a) Relic composition. (b) Gauge-sector parameter space. Comparison of the viable parameter space for nχ=2/5n_{\chi}=2/5, 3/43/4, and 7/57/5. The left panel shows the ratio of the scalar and fermionic relic abundances as a function of the dark matter mass ratio. The right panel shows the corresponding viable points in the (mZ′,gBL)(m_{Z'},g_{BL}) plane. All samples satisfy the same theoretical, collider, relic-density, and direct-detection requirements used in the main numerical analysis.

The three samples largely overlap in both projections. The relic composition retains the same qualitative structure found for the benchmark: both scalar- and fermion-dominated solutions occur for mS<mχm_S < m_\chi, whereas the mS>mχm_S > m_\chi branch is predominantly fermion dominated. Likewise, no clear systematic displacement appears in the (mZ′,gBL)(m_{Z'},g_{BL}) plane as the charge is varied.

A similarly weak dependence is found in direct detection. Although the Z′Z'-mediated amplitudes depend explicitly on nχn_\chi, the relic-fraction-rescaled spin-independent cross sections show no clear systematic shift among the three scans. This is not in conflict with the charge dependence of the gauge contribution, since the observable rates also depend on gBLg_{BL}, mZ′m_{Z'}, the scalar interactions, and the relic fractions ξχ\xi_\chi and ξS\xi_S.

Taken together, these results show that the explicit charge dependence of the gauge interaction does not translate into an equally strong dependence of the full phenomenology. The clearest effect appears in the gauge-dominated χχˉ\chi\bar{\chi} annihilation rate, where the expected nχ2n_\chi^2 scaling is reproduced quantitatively. By contrast, the scalar annihilation and conversion rates are much less sensitive to the charge, and the resulting relic composition and viable parameter space remain essentially unchanged. Within the range explored here, the benchmark choice nχ=3/4n_\chi= 3/4 is therefore representative of the broader phenomenology of the model.

Conclusions

We have proposed an interacting scalar–fermion two-component dark matter model within a gauged U(1)B−LU(1)_{B-L} extension of the SM. The dark sector consists of a Dirac fermion χ\chi and a complex scalar SS, whose charges allow a renormalizable Yukawa interaction directly connecting the two particles. For suitable charge assignments and mS<2mχm_S < 2m_\chi, both particles are stable without the need for an additional discrete symmetry. The same interaction generates conversion, semi-conversion, and semi-annihilation processes, leading to a genuinely coupled two-component freeze-out.

Our numerical analysis shows that viable solutions exist for both mS>mχm_S > m_\chi and mS<mχm_S < m_\chi. In particular, semi-conversion can play a central role in depleting the fermionic component while simultaneously feeding the scalar population. The model remains viable over broad non-resonant regions of parameter space, showing that the observed DM abundance can be reproduced without relying on Z′Z' resonance enhancement.

Current constraints still leave sizable viable regions of parameter space. Direct detection provides one of the strongest restrictions on the model, but part of the surviving parameter space remains within the projected reach of next-generation experiments. Finally, varying the dark B−LB-L charge over representative non-exceptional values leaves the overall phenomenology largely unchanged, indicating that the benchmark choice nχ=3/4n_{\chi}=3/4 captures the main features of the model over the range of charges explored.

Acknowledgments

The authors thank Giorgio Arcadi, Juri Smirnov, and Josef Pradler for insightful discussions. JPN thanks Patricio Escalona for his immense help with the multidimensional parameter space analysis and for providing valuable code advice, and Guillermo Gambini for helpful discussions about the model-building aspects of this work. JPN thanks the support from the Austrian Federal Ministry of Women, Science and Research (BMFWF) via the Ernst Mach Grant, weltweit (Award number: MPC-2025-01008), the Fundação de Amparo e Promoção da Ciência, Tecnologia e Inovação do Rio Grande do Norte (FAPERN) via the call number 06/2026 - FAPERN/IIP (Process number: 10910022.000493/2024-04.), the National Council for Scientific and Technological Development (CNPq) (Grant number: 307130/2021-5), and the Brazilian Federal Agency for Support and Evaluation of Graduate Education (CAPES) (Grant number: 88887.712383/2022-00). JPN thanks the Marietta Blau Institute for Particle Physics (MBI/OAW) for the hospitality during the partial development of this project. This work was supported by the Simons Foundation (Award Number: 1023171-RC). FSQ 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.

References

  1. [1]D. Abercrombie et. al., Dark Matter benchmark models for early LHC Run-2 Searches: Report of the ATLAS/CMS Dark Matter Forum, Phys. Dark Univ. 27 (2020) 100371, [1507.00966].DOI
  2. [2]J. Abdallah et. al., Simplified Models for Dark Matter Searches at the LHC, Phys. Dark Univ. 9-10 (2015) 8–23, [1506.03116].DOI
  3. [3]G. Arcadi, M. Dutra, P. Ghosh, M. Lindner, Y. Mambrini, M. Pierre, S. Profumo, and F. S. Queiroz, The waning of the WIMP? A review of models, searches, and constraints, Eur. Phys. J. C 78 (2018), no. 3 203, [1703.07364].DOI
  4. [4]G. Arcadi, D. Cabo-Almeida, M. Dutra, P. Ghosh, M. Lindner, Y. Mambrini, J. P. Neto, M. Pierre, S. Profumo, and F. S. Queiroz, The Waning of the WIMP: Endgame?, Eur. Phys. J. C 85 (2025), no. 2 152, [2403.15860].DOI
  5. [5]D. Tucker-Smith and N. Weiner, Inelastic dark matter, Phys. Rev. D 64 (2001) 043502, [hep-ph/0101138].DOI
  6. [6]K. M. Zurek, Multi-Component Dark Matter, Phys. Rev. D 79 (2009) 115002, [0811.4429].DOI
  7. [7]S. Esch, M. Klasen, and C. E. Yaguna, A minimal model for two-component dark matter, JHEP 09 (2014) 108, [1406.0617].DOI
  8. [8]C. E. Yaguna and O. Zapata, Multi-component scalar dark matter from a Z_N symmetry: a systematic analysis, JHEP 03 (2020) 109, [1911.05515].DOI
  9. [9]G. Bélanger, A. Pukhov, C. E. Yaguna, and Ó. Zapata, The Z_5 model of two-component dark matter, JHEP 09 (2020) 030, [2006.14922].arxiv.org/abs/2006.14922
  10. [10]G. Bélanger, A. Pukhov, C. E. Yaguna, and Ó. Zapata, The Z_7 model of three-component scalar dark matter, JHEP 03 (2023) 100, [2212.07488].DOI
  11. [11]F. Costa, S. Khan, and J. Kim, A two-component dark matter model and its associated gravitational waves, JHEP 06 (2022) 026, [2202.13126].DOI
  12. [12]G. Belanger and J.-C. Park, Assisted freeze-out, JCAP 03 (2012) 038, [1112.4491].DOI
  13. [13]B. Díaz Sáez, P. Escalona, S. Norero, and A. R. Zerwekh, Fermion singlet dark matter in a pseudoscalar dark matter portal, JHEP 10 (2021) 233, [2105.04255].arxiv.org/abs/2105.04255
  14. [14]C. E. Yaguna and O. Zapata, Fermion and scalar two-component dark matter from a Z_4 symmetry, Phys. Rev. D 105 (2022), no. 9 095026, [2112.07020].DOI
  15. [15]P. Escalona, J. P. Neto, M. J. Neves, C. Ramos, and D. Suarez, Two-Component Dark Matter in the Type-I 2HDM, 2603.18158.
  16. [16]X. Qi and H. Sun, A two-component dark matter model with Z_2 × Z_4 symmetry, JCAP 08 (2025) 002, [2504.12876].arxiv.org/abs/2504.12876
  17. [17]M. Frank, P. Ghosh, C. Majumdar, and S. Senapati, Exploring two component doublet dark matter, Phys. Rev. D 112 (2025), no. 5 055014, [2505.02816].DOI
  18. [18]L. Basso, S. Moretti, and G. M. Pruna, A Renormalisation Group Equation Study of the Scalar Sector of the Minimal B-L Extension of the Standard Model, Phys. Rev. D 82 (2010) 055018, [1004.3039].DOI
  19. [19]W. Rodejohann and C. E. Yaguna, Scalar dark matter in the B−L model, JCAP 12 (2015) 032, [1509.04036].DOI
  20. [20]M. Klasen, F. Lyonnet, and F. S. Queiroz, NLO+NLL collider bounds, Dirac fermion and scalar dark matter in the B−L model, Eur. Phys. J. C 77 (2017), no. 5 348, [1607.06468].DOI
  21. [21]N. Bernal, D. Restrepo, C. Yaguna, and Ó. Zapata, Two-component dark matter and a massless neutrino in a new B−L model, Phys. Rev. D 99 (2019), no. 1 015038, [1808.03352].DOI
  22. [22]S. Okada, Z' Portal Dark Matter in the Minimal B−L Model, Adv. High Energy Phys. 2018 (2018) 5340935, [1803.06793].arxiv.org/abs/1803.06793
  23. [23]A. Das, S. Gola, S. Mandal, and N. Sinha, Two-component scalar and fermionic dark matter candidates in a generic U(1)X model, Phys. Lett. B 829 (2022) 137117, [2202.01443].DOI
  24. [24]R. N. Mohapatra and G. Senjanovic, Neutrino Mass and Spontaneous Parity Nonconservation, Phys. Rev. Lett. 44 (1980) 912.DOI
  25. [25]P. Van Dong, C. H. Nam, and D. Van Loi, Canonical seesaw implication for two-component dark matter, Phys. Rev. D 103 (2021) 095016, [2007.08957].DOI
  26. [26]K. Kannike, Vacuum Stability Conditions From Copositivity Criteria, Eur. Phys. J. C 72 (2012) 2093, [1205.3781].DOI
  27. [27]A. Alloul, N. D. Christensen, C. Degrande, C. Duhr, and B. Fuks, FeynRules 2.0 - A complete toolbox for tree-level phenomenology, Comput. Phys. Commun. 185 (2014) 2250–2300, [1310.1921].DOI
  28. [28]A. Semenov, LanHEP — A package for automatic generation of Feynman rules from the Lagrangian, Version 3.2, Comput. Phys. Commun. 201 (2016) 167–170, [1412.5016].
  29. [29]G. Bélanger, F. Boudjema, A. Pukhov, and A. Semenov, micrOMEGAs4.1: two dark matter candidates, Comput. Phys. Commun. 192 (2015) 322–329, [1407.6129].DOI
  30. [30]G. Alguero, G. Belanger, F. Boudjema, S. Chakraborti, A. Goudelis, S. Kraml, A. Mjallal, and A. Pukhov, micrOMEGAs 6.0: N-component dark matter, Comput. Phys. Commun. 299 (2024) 109133, [2312.14894].DOI
  31. [31]P. Gondolo and G. Gelmini, Cosmic abundances of stable particles: Improved analysis, Nucl. Phys. B 360 (1991) 145–179.DOI
  32. [32]G. Alguero, G. Belanger, F. Boudjema, S. Chakraborti, A. Goudelis, S. Kraml, A. Mjallal, and A. Pukhov, micromegas 6.0, Mar., 2024.DOI
  33. [33]G. Belanger, A. Belyaev, N. Bernal, F. Boudjema, S. Chakraborti, A. Goudelis, and A. Pukhov, micrOMEGAs 7: Beyond standard cosmology, 2606.06645.arxiv.org/abs/2606.06645
  34. [34]J. L. Feng and J. Smolinsky, Impact of a resonance on thermal targets for invisible dark photon searches, Phys. Rev. D 96 (2017), no. 9 095022, [1707.03835].DOI
  35. [35]Planck Collaboration, N. Aghanim et. al., Planck 2018 results. VI. Cosmological parameters, Astron. Astrophys. 641 (2020) A6, [1807.06209]. [Erratum: Astron. Astrophys. 652, C4 (2021)].
  36. [36]P. N. Bhattiprolu and J. D. Wells, Depleted Higgs boson: Searches for universal coupling suppression, invisible decays, and mixed-in scalars, Phys. Rev. D 107 (2023), no. 5 055022, [2204.03435].DOI
  37. [37]ATLAS Collaboration, G. Aad et. al., A detailed map of Higgs boson interactions by the ATLAS experiment ten years after the discovery, Nature 607 (2022), no. 7917 52–59, [2207.00092]. [Erratum: Nature 612, E24 (2022)].DOI
  38. [38]CMS Collaboration, A. Tumasyan et. al., A portrait of the Higgs boson by the CMS experiment ten years after the discovery., Nature 607 (2022), no. 7917 60–68, [2207.00043]. [Erratum: Nature 623, (2023)].DOI
  39. [39]CMS Collaboration, V. Khachatryan et. al., Search for a Higgs boson in the mass range from 145 to 1000 GeV decaying to a pair of W or Z bosons, JHEP 10 (2015) 144, [1504.00936].DOI
  40. [40]CMS Collaboration, A. M. Sirunyan et. al., Search for a new scalar resonance decaying to a pair of Z bosons in proton-proton collisions at √s = 13 TeV, JHEP 06 (2018) 127, [1804.01939]. [Erratum: JHEP 03, 128 (2019)].DOI
  41. [41]ATLAS Collaboration, M. Aaboud et. al., Combination of searches for heavy resonances decaying into bosonic and leptonic final states using 36 fb−1 of proton-proton collision data at √s = 13 TeV with the ATLAS detector, Phys. Rev. D 98 (2018), no. 5 052008, [1808.02380].DOI
  42. [42]ATLAS Collaboration, G. Aad et. al., Search for heavy diboson resonances in semileptonic final states in pp collisions at √s = 13 TeV with the ATLAS detector, Eur. Phys. J. C 80 (2020), no. 12 1165, [2004.14636].DOI
  43. [43]ATLAS Collaboration, G. Aad et. al., Search for heavy resonances decaying into a pair of Z bosons in the ℓ+ℓ−ℓ+ℓ− and ℓ+ℓ−νν̄ final states using 139 fb−1 of proton-proton collisions at √s = 13 TeV with the ATLAS detector, Eur. Phys. J. C 81 (2021), no. 4 332, [2009.14791].
  44. [44]CMS Collaboration, A. Tumasyan et. al., Search for a heavy Higgs boson decaying into two lighter Higgs bosons in the ττbb final state at 13 TeV, JHEP 11 (2021) 057, [2106.10361].DOI
  45. [45]ATLAS Collaboration, G. Aad et. al., Search for resonant and non-resonant Higgs boson pair production in the bb̄τ+τ− decay channel using 13 TeV pp collision data from the ATLAS detector, JHEP 07 (2023) 040, [2209.10910].
  46. [46]T. Ferber, A. Grohsjean, and F. Kahlhoefer, Dark Higgs bosons at colliders, Prog. Part. Nucl. Phys. 136 (2024) 104105, [2305.16169].DOI
  47. [47]ATLAS Collaboration, G. Aad et. al., Search for high-mass dilepton resonances using 139 fb−1 of pp collision data collected at √s = 13 TeV with the ATLAS detector, Phys. Lett. B 796 (2019) 68–87, [1903.06248].DOI
  48. [48]CMS Collaboration, A. M. Sirunyan et. al., Search for resonant and nonresonant new phenomena in high-mass dilepton final states at √s = 13 TeV, JHEP 07 (2021) 208, [2103.02708].DOI
  49. [49]ATLAS Collaboration, G. Aad et. al., Search for new resonances in mass distributions of jet pairs using 139 fb−1 of pp collisions at √s = 13 TeV with the ATLAS detector, JHEP 03 (2020) 145, [1910.08447].DOI
  50. [50]ATLAS Collaboration, M. Aaboud et. al., Search for heavy particles decaying into a top-quark pair in the fully hadronic final state in pp collisions at √s = 13 TeV with the ATLAS detector, Phys. Rev. D 99 (2019), no. 9 092004, [1902.10077].DOI
  51. [51]G. Arcadi, G. C. Dorsch, J. P. Neto, F. S. Queiroz, and Y. M. Oviedo-Torres, Probing a dark sector with collider physics, direct detection, and gravitational waves, Phys. Lett. B 848 (2024) 138382, [2307.06376].DOI
  52. [52]F. S. Queiroz, J. Zamora-Saa, R. C. Silva, and Y. M. Oviedo-Torres, LHC and HL-LHC bounds on visible and invisible Z′ decays in the B-L model, Phys. Rev. D 111 (2025), no. 9 095021, [2501.00610].arxiv.org/abs/2501.00610
  53. [53]S. Bhattacharya and D. Pradhan, Direct Search signal of two-component Dark Matter, 2501.17862.arxiv.org/abs/2501.17862
  54. [54]C.-F. Chang, X.-G. He, and J. Tandean, Two-Higgs-Doublet-Portal Dark-Matter Models in Light of Direct Search and LHC Data, JHEP 04 (2017) 107, [1702.02924].DOI
  55. [55]M. Hoferichter, P. Klos, J. Menéndez, and A. Schwenk, Improved limits for Higgs-portal dark matter from LHC searches, Phys. Rev. Lett. 119 (2017), no. 18 181803, [1708.02245].DOI
  56. [56]U. Haisch and F. Kahlhoefer, On the importance of loop-induced spin-independent interactions for dark matter direct detection, JCAP 04 (2013) 050, [1302.4454].DOI
  57. [57]A. Amiri, B. Diaz Saez, and K. Mö hling, Freeze-in at Low Reheating and Direct Detection of Fermion Dark Matter, 2511.21520.arxiv.org/abs/2511.21520
  58. [58]XENON Collaboration, E. Aprile et. al., First Dark Matter Search with Nuclear Recoils from the XENONnT Experiment, Phys. Rev. Lett. 131 (2023), no. 4 041003, [2303.14729].DOI
  59. [59]LZ Collaboration, J. Aalbers et. al., Dark Matter Search Results from 4.2 Tonne-Years of Exposure of the LUX-ZEPLIN (LZ) Experiment, Phys. Rev. Lett. 135 (2025), no. 1 011802, [2410.17036].DOI
  60. [60]L. Baudis, DARWIN/XLZD: A future xenon observatory for dark matter and other rare interactions, Nucl. Phys. B 1003 (2024) 116473, [2404.19524].arxiv.org/abs/2404.19524
  61. [61]Fermi-LAT Collaboration, M. Ackermann et. al., Searching for Dark Matter Annihilation from Milky Way Dwarf Spheroidal Galaxies with Six Years of Fermi Large Area Telescope Data, Phys. Rev. Lett. 115 (2015), no. 23 231301, [1503.02641].DOI
  62. [62]H.E.S.S. Collaboration, H. Abdalla et. al., Search for Dark Matter Annihilation Signals in the H.E.S.S. Inner Galaxy Survey, Phys. Rev. Lett. 129 (2022), no. 11 111101, [2207.10471].DOI
  63. [63]CTA Collaboration, A. Acharyya et. al., Sensitivity of the Cherenkov Telescope Array to a dark matter signal from the Galactic centre, JCAP 01 (2021) 057, [2007.16129].
  64. [64]H. Silverwood, C. Weniger, P. Scott, and G. Bertone, A realistic assessment of the CTA sensitivity to dark matter annihilation, JCAP 03 (2015) 055, [1408.4131].DOI
  65. [65]F. S. Queiroz, C. Siqueira, and C. E. Yaguna, Multi-Component Dark Matter as a Solution to the Galactic Center GeV Excess, 2603.17095.arxiv.org/abs/2603.17095
  66. [66]G. Elor, N. L. Rodd, T. R. Slatyer, and W. Xue, Model-Independent Indirect Detection Constraints on Hidden Sector Dark Matter, JCAP 06 (2016) 024, [1511.08787].DOI
  67. [67]F. S. Queiroz and C. Siqueira, Search for Semi-Annihilating Dark Matter with Fermi-LAT, H.E.S.S., Planck, and the Cherenkov Telescope Array, JCAP 04 (2019) 048, [1901.10494].DOI

Paper details

vlnc:W_q5wn95fb
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