---
title: Two-Component Scalar Dark Matter
url: https://www.emergentmind.com/topics/two-component-scalar-dark-matter
type: topic
---

# Two-Component Scalar Dark Matter

Two-component scalar dark matter denotes dark-sector frameworks in which the observed dark matter abundance is shared by two stable species and at least one of them is a scalar; in many constructions both components are scalars. Across the literature, the defining ingredients are an exact discrete or residual gauge symmetry that stabilizes more than one state, coupled abundance evolution rather than independent one-component freeze-out, and scalar interactions that are usually organized by Higgs-portal, dark-scalar, or electroweak gauge couplings. Representative realizations range from singlet models with one scalar and one fermion [1406.0617, 2112.07020], purely scalar singlet sectors based on \(Z_5\) or \(Z_{2n}\) symmetries [2006.14922, 2106.11889], inert doublet or doublet–triplet constructions [1904.04837, 2102.06032], and scale-invariant or gauge-extended models in which the scalar sector is tightly linked to radiative symmetry breaking or residual discrete gauge symmetries [1511.08432, 2512.18568].

## 1. Symmetry structure and model-building patterns

The common model-building problem is to stabilize two dark states without introducing fast decays between them. A minimal singlet example adds a fermion \(\chi\), a real singlet scalar \(S\), and an additional singlet scalar \(\phi\), with a single \(Z_2\) under which \(\chi\) and \(S\) are odd while the Standard Model and \(\phi\) are even; the heavier odd particle is also stable because the Lagrangian contains no allowed operators involving both odd fields, so an accidental symmetry prevents \(\chi\) from decaying into \(S\) or vice versa [1406.0617]. Closely related \(Z_4\) models take a real singlet scalar \(S\) and a singlet Dirac fermion \(\psi\) with
\[
S\to -S,\qquad \psi\to i\,\psi,
\]
and impose the kinematic condition \(M_S<2M_\psi\) so that both particles remain stable [2112.07020]. In the type-I 2HDM realization, a real scalar \(s\) and a Dirac fermion \(\chi\) are stabilized by \(Z_4^{\rm DM}\) together with \(m_s<2m_\chi\) [2603.18158].

Purely scalar two-component sectors are often organized by a single discrete symmetry rather than a product symmetry. The \(Z_5\) model contains two complex singlet scalars \(\phi_1\) and \(\phi_2\) with different \(Z_5\) charges, \(\phi_1\sim \omega_5\) and \(\phi_2\sim \omega_5^2\), and stability requires \(\frac{M_1}{2}<M_2<2M_1\) together with vanishing singlet vacuum expectation values [2006.14922]. The \(Z_{2n}\) construction instead stabilizes one complex scalar \(\phi_A\) and one real scalar \(\phi_B\) with \(\phi_B\to -\phi_B\); the explicitly analyzed prototypes are the \(Z_4\), \(Z_6(13)\), and \(Z_6(23)\) scenarios [2106.11889]. A different two-singlet scalar framework uses an unbroken \(Z_2\times Z_2'\) and two real gauge-singlet scalars \(S_2\) and \(S_3\), neither of which acquires a vacuum expectation value, so both remain stable and never mix with the Higgs [1709.05955].

Electroweak multiplet realizations exploit inert sectors. One minimal scotogenic construction adds two inert scalar doublets \(\eta_1,\eta_2\) and two right-handed neutrinos \(N_1,N_2\), with unbroken \(\mathbb{Z}_2\times\mathbb{Z}'_2\); the two neutral CP-even scalars \(H_1\) and \(H_2\) serve as the two dark matter candidates [1904.04837]. Another model combines an inert scalar doublet \(\Phi\) and a real scalar triplet \(\mathcal T\), again protected by \(Z_2\times Z_2'\), so that the stable neutral states are \(H_0\) and \(T_0\) [2102.06032]. A collider-oriented realization in the I(2+1)HDM uses one active and two inert doublets with a \(Z_2\times Z_2'\) symmetry that stabilizes the lightest neutral scalar in each inert sector, yielding two scalar dark matter candidates \(H_1\) and \(H_2\) [2607.00243].

Residual gauge symmetries provide a more structural origin of stability. In the \(U(1)'\to Z_4\) model, two complex scalar dark matter fields \(\phi_1,\phi_2\) are stabilized by the remnant \(Z_4\) gauge symmetry after a dark Higgs \(\chi\) acquires a vacuum expectation value [2509.06728]. In the hidden \(SU(2)_D\) model, symmetry breaking \(SU(2)_D\to Z_3\) leaves two stable dark relics, one scalar \(\rho_1^{(*)}\) and one vector \(X^\pm\), both carrying nontrivial residual charge [2512.18568]. Gauge extensions with \(U(1)_X\) or flavor-dependent \(U(1)_X\) also admit mixed scalar–fermion two-component sectors in which the scalar state is stabilized by exact discrete remnants or by inert-sector parity assignments [2202.01443, 2603.24072].

| Realization | Stable dark states | Stabilizing structure |
|---|---|---|
| Singlet scalar–fermion models | \(S+\chi\) or \(S+\psi\) | \(Z_2\) or \(Z_4\) [1406.0617, 2112.07020] |
| Purely scalar singlet sectors | \(\phi_1+\phi_2\), \(\phi_A+\phi_B\), \(S_2+S_3\) | \(Z_5\), \(Z_{2n}\), \(Z_2\times Z_2'\) [2006.14922, 2106.11889, 1709.05955] |
| Inert electroweak sectors | \(H_1+H_2\), \(H_0+T_0\), \(H_1+H_2\) | \(\mathbb Z_2\times\mathbb Z_2'\) [1904.04837, 2102.06032, 2607.00243] |
| Residual gauge-symmetry models | \(\phi_1+\phi_2\), \(\rho_1^{(*)}+X^\pm\) | remnant \(Z_4\) or \(Z_3\) [2509.06728, 2512.18568] |

## 2. Coupled relic-density evolution

The phenomenological hallmark of two-component scalar dark matter is that the abundances must be evolved with coupled Boltzmann equations. In the minimal singlet scalar–fermion model, the yields \(Y_\chi\) and \(Y_S\) satisfy the standard two-component system with annihilation into \(Z_2\)-even states and conversion terms,
\[
\chi\chi \leftrightarrow SS,
\]
and the total abundance is
\[
\Omega_{\rm DM}h^2=\Omega_Sh^2+\Omega_\chi h^2=0.1199\pm0.0027
\]
with \(\xi_\chi+\xi_S=1\) [1406.0617]. The same structural point recurs in inert-doublet and doublet–triplet models, where \(H_1H_1\leftrightarrow H_2H_2\) or \(T_0T_0\leftrightarrow H_0H_0\) conversion terms appear explicitly in the coupled equations and shift the final relic fractions [1904.04837, 2102.06032].

Conversion processes are not a small correction. In the singlet scalar–fermion model, conversion can change \(\Omega_\chi\) by more than two orders of magnitude in the examples shown, because the heavier species can annihilate into the lighter one and the lighter species can receive a residual contribution from the heavier partner after freeze-out [1406.0617]. In the inert-doublet plus radiative neutrino-mass construction, the entire intermediate range \(M_W\le M_{\rm DM}\lesssim 500~{\rm GeV}\), where one scalar doublet dark matter candidate cannot satisfy the correct relic density on its own, becomes allowed in the two-component theory because the two neutral CP-even scalars can inter-convert in the presence of neutrino Yukawa couplings with the dark sector [1904.04837]. In the doublet–triplet model, the IDM desert region \(100~{\rm GeV}\lesssim m_{H_0}\lesssim 500~{\rm GeV}\) and triplet masses below about \(1.8~{\rm TeV}\), both under-abundant in the standalone limits, can reproduce the Planck value once \(\lambda_{\Phi T}\) drives efficient \(T_0T_0\leftrightarrow H_0H_0\) conversion [2102.06032].

Semi-annihilation provides an additional depletion channel that has no analogue in the minimal one-component singlet-scalar model. In the \(Z_4\) scalar–fermion setup, the relevant dark processes include \(\psi\psi\to Sh\), \(S\psi\to \bar\psi h\), and \(\bar\psi\psi\to SS\); semi-annihilation is especially important in the \(M_S<M_\psi\) ordering because it opens viable scalar-mass regions below the \(\sim 950\) GeV threshold excluded in the ordinary scalar singlet model [2112.07020]. In \(Z_5\) and \(Z_{2n}\) models the trilinear dark couplings induce both conversion and semi-annihilation, and the presence or absence of such trilinear terms largely determines whether the viable region extends broadly below the TeV scale or remains confined to quasi-degenerate spectra [2006.14922, 2106.11889]. In the type-I 2HDM with \(Z_4^{\rm DM}\), semi-annihilation is singled out as a distinctive feature of \(Z_4\)-stabilized multi-component dark matter, although annihilation and conversion are often more important in setting the final relic abundances [2603.18158].

Freeze-in constitutes a nonthermal alternative. In the two-real-singlet scalar FIMP model, the dark matter particles are produced from nearly vanishing initial abundances by Higgs decays and annihilations, remain out of equilibrium because the portal couplings are extremely feeble, and satisfy
\[
\Omega_{\rm tot}h^2=\Omega_{s_2}h^2+\Omega_{s_3}h^2
\]
within the PLANCK range \(0.1172\le \Omega_{\rm DM}h^2\le 0.1226\) [1709.05955]. A central difference from WIMP freeze-out is that the relic density increases with the portal couplings in the FIMP case, roughly as \(\Omega h^2\sim \lambda_{HS_i}^2\) [1709.05955].

## 3. Scalar interactions, portals, and mediator effects

Scalar components are typically tied to the visible sector through Higgs-portal-like interactions, but extended scalar sectors alter both annihilation and scattering in ways that are absent in single-component models. In the minimal singlet model with \(\chi\), \(S\), and \(\phi\), the scalar \(S\) interacts through the Higgs portal \(\lambda (H)S^2/4\) and the \(\phi\)-mediated terms \(\mu_{\phi SS}\phi S^2/2\) and \(\lambda_{\phi\phi SS}\phi^2S^2/2\); after electroweak symmetry breaking, the Higgs and singlet mix into
\[
H_1=h\cos\alpha+\phi\sin\alpha,\qquad H_2=\phi\cos\alpha-h\sin\alpha,
\]
with \(H_1\) identified as the observed 125 GeV Higgs boson [1406.0617]. The scalar relic density then exhibits the familiar Higgs resonance near \(M_S\simeq M_{H_1}/2\), a new resonance near \(M_S\simeq M_{H_2}/2\), and the new final state
\[
SS\to H_2H_2,
\]
which becomes important when \(M_S\gtrsim M_{H_2}\) [1406.0617].

The direct-detection amplitude is likewise modified by multiple mediators. In the same model, the scalar spin-independent cross section contains an interference term between \(H_1\)- and \(H_2\)-exchange contributions,
\[
\sigma_{SI,S}=\frac{m_r^2}{4\pi M_S^2}\left[\lambda\frac{v}{2}\left(\frac{1}{M_{H_1}^2}-\frac{1}{M_{H_2}^2}\right)+\mu_{\phi SS}\left(\frac{1}{M_{H_2}^2}-\frac{1}{M_{H_1}^2}\right)\right]^2 g_{Hp}^2,
\]
with \(g_{Hp}\simeq 1.4\times10^{-3}\), so constructive or destructive interference can enhance or suppress the rate [1406.0617]. The same qualitative feature appears in the generic \(U(1)_X\) model, where the scalar candidate \(\chi_R\) scatters through both \(h_1\) and \(h_2\), and destructive interference between the two Higgs exchanges is explicitly noted [2202.01443].

Scale invariance leads to a distinctive scalar mediator, the scalon. In the classically scale-invariant Standard Model with two real scalar dark matter fields \(\varphi_1,\varphi_2\), the tree-level scalon is massless and acquires its mass only through one-loop corrections,
\[
\delta m_s^2 = 2Bv_\phi^2,
\]
which is phenomenologically essential because a massless scalar mediator would produce enormous dark matter–nucleon scattering and exclude the model [1511.08432]. This radiative mass makes it possible for the two-component scalar theory to satisfy relic-density and Xenon100/LUX constraints, and the lower bound on each dark matter mass drops to roughly \(m_{\rm DM}\gtrsim 300~{\rm GeV}\), whereas the single-component scalar case is viable mainly for \(m_{\rm DM}\gtrsim 2~{\rm TeV}\) [1511.08432]. A related scale-invariant scalar–fermion model also uses a loop-generated scalon \(H_2\) and finds that the scalar mass must satisfy \(M_s\gtrsim 310~{\rm GeV}\) while direct-detection constraints exclude large regions unless mediator effects suppress the nucleon cross section [1808.08706].

Electroweak multiplet models replace singlet portals by gauge-strength annihilation. In the inert-doublet setup, the scalar sector is often under-abundant because annihilation and co-annihilation into \(W^+W^-\), \(ZZ\), and related states are too efficient in the interval \(80~{\rm GeV}\lesssim m_{H^0}\lesssim 525~{\rm GeV}\); in the two-component theory this under-abundance is compensated by the second component rather than treated as a failure of the model [2505.02816]. In the 3HDM collider study, the key kinematic control variables are the mass splittings
\[
\Delta M_i=M_{A_i}-M_{H_i},
\]
which fix the visible dilepton kinematics through the cascade \(A_i\to H_i Z^\ast\to H_i\mu^+\mu^-\) and generate a preselection double-bump structure in the dimuon invariant-mass spectrum because the two dark sectors have different \(\Delta M_i\) [2607.00243].

## 4. Cosmological extensions and nonstandard scalar regimes

Two-component scalar dark matter is not restricted to thermal WIMP freeze-out. One direction treats the dark sector as a superposition of cosmological scalar fields. In the two-scalar-field cosmology built from a classical complex scalar plus either an axion-like field or a Higgs-like inert scalar, the fields are assumed to be matter-like today and their present-day fractions are parametrized by
\[
\rho_1(a=1)=\eta\,\Omega_{\rm DM}\rho_{\rm crit},\qquad \rho_2(a=1)=(1-\eta)\,\Omega_{\rm DM}\rho_{\rm crit}
\]
[2110.10258]. The BBN-era constraint is implemented through \(N_{\mathrm{eff}}=3.56\pm0.23\) and \(w(z_{\rm eq})<0.001\) at \(z_{\rm eq}\approx 3365\). Within this framework, the classical-plus-Higgs-like model survives only if the classical component dominates sufficiently, requiring \(\eta\gtrsim 0.423\), equivalently a Higgs-like fraction \(\lesssim 58\%\), whereas the axion-plus-Higgs-like model is discarded because there is no value of \(f_a\) for which \(N_{\mathrm{eff}}\) stays within the BBN bounds throughout nucleosynthesis [2110.10258].

Another direction emphasizes self-interactions and late-time structure formation. In the self-resonant \(U(1)'\to Z_4\) model, the resonance condition
\[
m_2\simeq 2m_1,\qquad \Delta\equiv 1-\frac{m_2}{2m_1}\ge 0
\]
enhances elastic co-scattering through a Yukawa potential with effective mediator mass
\[
M=m_2\sqrt{2-\frac{m_2}{m_1}},
\]
which becomes small near resonance and yields velocity-dependent self-interactions suitable for addressing small-scale structure problems at galaxies [2509.06728]. The same setup also predicts Sommerfeld-enhanced semi-annihilation and boosted dark matter from the Galactic Center, with benchmark Sommerfeld factors \(S_0\sim 10^3-10^6\) [2509.06728].

Cosmological consistency can also impose unexpectedly strong mass bounds. In the two-singlet scalar FIMP model, self-interaction constraints based on \(\sigma/m<0.47~\mathrm{cm^2/g}\) push the viable masses of both components to roughly \(m_{s_2},m_{s_3}\lesssim 0.2~\mathrm{GeV}\) for maximal allowed couplings, even though freeze-in relic-density considerations alone permit masses from GeV down to keV [1709.05955]. In the hidden \(SU(2)_D\) model, the massless dark gauge boson \(X_3\) contributes to dark radiation and induces ellipticity constraints, leading to \(T^{\rm dec}\gtrsim 375~\mathrm{MeV}\) and upper bounds on \(\alpha_D=g_D^2/(4\pi)\) from halo-shape considerations [2512.18568].

## 5. Detection phenomenology

Direct detection in multicomponent theories is governed by abundance-weighted rates rather than bare cross sections. The standard rescaling is
\[
\xi_i\,\sigma_i^{\rm SI},\qquad \xi_i=\frac{\Omega_i}{\Omega_{\rm DM}},
\]
or equivalent notation such as \(\hat\sigma_{\rm SI}^i=\frac{\Omega_i}{\Omega_s+\Omega_\chi}\sigma_{\rm SI}^i\) [1406.0617, 2603.18158]. This rescaling does not imply automatic invisibility for a subdominant component. In the minimal singlet scalar–fermion model, the authors explicitly emphasize that reduced relic fraction does not necessarily imply a reduced direct-detection signal because the couplings generally increase as the abundance decreases; one benchmark has \(S\) contributing only \(\sim 3\%\) of the dark matter, yet its direct-detection signal remains projected to be observable in a 1-ton experiment, and future experiments like XENON1T can probe a subdominant scalar component at the percent level [1406.0617]. Comparable conclusions recur in the \(Z_4\) scalar–fermion model, where both dark matter particles may be observed in future direct-detection experiments [2112.07020], in the \(Z_5\) scalar model, where current and future direct-detection experiments may be sensitive to signals from both dark matter particles [2006.14922], and in the \(Z_{2n}\) prototypes, which may lead to observable signals in direct detection experiments across wider mass intervals than the ordinary Higgs-portal singlet model [2106.11889].

The scalar component is usually the more visible direct-detection target. In mixed scalar–fermion models the scalar scatters at tree level through Higgs exchange, while the fermion often scatters only at one loop or through suppressed heavy-mediator exchange [2112.07020, 2603.18158]. This makes scalar-involved scenarios more constrained: in the flavor-dependent \(U(1)_X\) model, the mixed fermion–scalar case is significantly more restricted than the purely fermionic case, and the viable region tends to lie near future experimental sensitivity around \(\sigma^{\rm SI}\sim 10^{-47}-10^{-48}~\mathrm{cm}^2\) [2603.24072]. In the generic \(U(1)_X\) model, scalar dark matter below roughly \(500\) GeV is strongly constrained by XENON1T unless resonance or interference effects suppress the rate [2202.01443]. In the hidden \(SU(2)_D\) model, by contrast, the benchmark scalar cross sections are around \(10^{-48}~\mathrm{cm}^2\), below current XENON1T, XENONnT, and LZ limits [2512.18568].

Collider probes are increasingly model-specific but already nontrivial. The most explicit LHC study considers the I(2+1)HDM signal
\[
\mu^+\mu^- + E_T^{\rm miss} + j
\]
from \(pp\to H_iA_i\) followed by \(A_i\to H_iZ^\ast\to H_i\mu^+\mu^-\), with \(m_{\mu^+\mu^-}^{\max}=m_{A_i}-m_{H_i}\) [2607.00243]. For the representative benchmark BP1, the detector-level analysis yields \(S/B\simeq 9.8\%\) and \(S/\sqrt{B}=1.35\) at Run 3 with \({\cal L}=300~{\rm fb}^{-1}\), rising to \(S/\sqrt{B}=4.93\) under a statistical-only extrapolation to \({\cal L}=4~{\rm ab}^{-1}\) [2607.00243]. Before the full selection, the two dark sectors generate a double-bump structure in the dimuon invariant-mass distribution, but after cuts optimized for inclusive sensitivity this feature is not statistically robust enough to establish the two-component origin of the signal [2607.00243]. Other electroweak-multiplet models predict the familiar inert-doublet channels \(\ell^+\ell^-+\slashed E_T\) or disappearing charged tracks from nearly degenerate triplet states [2505.02816, 2102.06032]. In the type-I 2HDM realization, collider bounds strongly constrain the scalar sector and create tension with dark-matter-favored regions, particularly in the sub-TeV regime [2603.18158].

## 6. Recurring phenomenological lessons

Several general lessons recur across otherwise very different constructions. First, two-component scalar dark matter is often viable precisely where the corresponding one-component scalar model is not. The \(Z_5\) model admits viable sub-TeV masses that are excluded in the simple singlet scalar model [2006.14922]. The scale-invariant two-component scalar model allows masses down to roughly \(300\) GeV per component, whereas the one-component case is driven to \(m_{\rm DM}\gtrsim 2~\text{TeV}\) [1511.08432]. The doublet–triplet model revives the inert-doublet desert and the sub-TeV triplet region [2102.06032], and the scotogenic two-doublet model opens the entire intermediate inert-doublet range [1904.04837].

Second, the viability of a scalar component is usually controlled by a competition between efficient depletion channels and direct-detection pressure. Increasing portal or conversion couplings can lower the relic abundance, but the same couplings typically raise Higgs-mediated nucleon scattering; the most successful parameter regions therefore often lie near resonances, near destructive-interference conditions, or in spectra where the scalar is only one part of the relic abundance [1406.0617, 2202.01443, 2603.24072].

Third, multicomponent cosmology is not equivalent to adding two one-component models. The coupled Boltzmann systems, semi-annihilation terms, and threshold-sensitive conversion channels introduce qualitatively new dynamics. This is explicit in singlet scalar–fermion systems [1406.0617, 2112.07020], inert multiplet models [1904.04837, 2102.06032], and discrete-symmetry scalar sectors [2006.14922, 2106.11889]. A plausible implication is that phenomenological exclusions derived under the assumption that one candidate constitutes \(100\%\) of the halo can mischaracterize viable parameter space in multi-component theories; this point is emphasized directly in the cosmological scalar-field analysis, which argues that direct-detection strategies should not assume a single 100% dark-matter species [2110.10258].

Finally, two-component scalar dark matter has become a framework rather than a single model. Minimal singlet sectors, inert electroweak multiplets, scale-invariant theories, residual discrete gauge symmetries, freeze-in sectors, and scalar-field cosmologies all realize the same organizing idea: the dark matter density may be partitioned among multiple stable states, and the scalar component can remain both cosmologically relevant and experimentally accessible even when it is not dominant [1406.0617, 1709.05955, 2607.00243].

Source: https://www.emergentmind.com/topics/two-component-scalar-dark-matter