Two-Component Scalar Dark Matter

• Two-component scalar dark matter is defined by frameworks where the dark matter abundance is shared by two stable species, with at least one being a scalar stabilized by discrete or residual gauge symmetries.
• Coupled Boltzmann equations govern the relic density evolution through annihilations, conversions, and semi-annihilations, allowing parameter regions inaccessible to single-component models.
• Scalar interactions via Higgs portals and mediator effects lead to distinctive experimental signatures in direct detection and collider searches despite competing depletion channels.

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 (Esch et al., 2014, Yaguna et al., 2021), purely scalar singlet sectors based on $Z_5$ or $Z_{2n}$ symmetries (Bélanger et al., 2020, Yaguna et al., 2021), inert doublet or doublet–triplet constructions (Borah et al., 2019, Chakrabarty et al., 2021), and scale-invariant or gauge-extended models in which the scalar sector is tightly linked to radiative symmetry breaking or residual discrete gauge symmetries (Ghorbani et al., 2015, Chen et al., 21 Dec 2025).

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 $Z_{2n}$0 or vice versa (Esch et al., 2014). Closely related $Z_{2n}$1 models take a real singlet scalar $Z_{2n}$2 and a singlet Dirac fermion $Z_{2n}$3 with

$Z_{2n}$4

and impose the kinematic condition $Z_{2n}$5 so that both particles remain stable (Yaguna et al., 2021). In the type-I 2HDM realization, a real scalar $Z_{2n}$6 and a Dirac fermion $Z_{2n}$7 are stabilized by $Z_{2n}$8 together with $Z_{2n}$9 (Escalona et al., 18 Mar 2026).

Purely scalar two-component sectors are often organized by a single discrete symmetry rather than a product symmetry. The $\chi$0 model contains two complex singlet scalars $\chi$1 and $\chi$2 with different $\chi$3 charges, $\chi$4 and $\chi$5, and stability requires $\chi$6 together with vanishing singlet vacuum expectation values (Bélanger et al., 2020). The $\chi$7 construction instead stabilizes one complex scalar $\chi$8 and one real scalar $\chi$9 with $S$0; the explicitly analyzed prototypes are the $S$1, $S$2, and $S$3 scenarios (Yaguna et al., 2021). A different two-singlet scalar framework uses an unbroken $S$4 and two real gauge-singlet scalars $S$5 and $S$6, neither of which acquires a vacuum expectation value, so both remain stable and never mix with the Higgs (Pandey et al., 2017).

Electroweak multiplet realizations exploit inert sectors. One minimal scotogenic construction adds two inert scalar doublets $S$7 and two right-handed neutrinos $S$8, with unbroken $S$9; the two neutral CP-even scalars $\phi$0 and $\phi$1 serve as the two dark matter candidates (Borah et al., 2019). Another model combines an inert scalar doublet $\phi$2 and a real scalar triplet $\phi$3, again protected by $\phi$4, so that the stable neutral states are $\phi$5 and $\phi$6 (Chakrabarty et al., 2021). A collider-oriented realization in the I(2+1)HDM uses one active and two inert doublets with a $\phi$7 symmetry that stabilizes the lightest neutral scalar in each inert sector, yielding two scalar dark matter candidates $\phi$8 and $\phi$9 (Belyaev et al., 30 Jun 2026).

Residual gauge symmetries provide a more structural origin of stability. In the $Z_2$0 model, two complex scalar dark matter fields $Z_2$1 are stabilized by the remnant $Z_2$2 gauge symmetry after a dark Higgs $Z_2$3 acquires a vacuum expectation value (Justino et al., 8 Sep 2025). In the hidden $Z_2$4 model, symmetry breaking $Z_2$5 leaves two stable dark relics, one scalar $Z_2$6 and one vector $Z_2$7, both carrying nontrivial residual charge (Chen et al., 21 Dec 2025). Gauge extensions with $Z_2$8 or flavor-dependent $Z_2$9 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 (Das et al., 2022, Duy et al., 25 Mar 2026).

Realization	Stable dark states	Stabilizing structure
Singlet scalar–fermion models	$\chi$0 or $\chi$1	$\chi$2 or $\chi$3 (Esch et al., 2014, Yaguna et al., 2021)
Purely scalar singlet sectors	$\chi$4, $\chi$5, $\chi$6	$\chi$7, $\chi$8, $\chi$9 (Bélanger et al., 2020, Yaguna et al., 2021, Pandey et al., 2017)
Inert electroweak sectors	$S$0, $S$1, $S$2	$S$3 (Borah et al., 2019, Chakrabarty et al., 2021, Belyaev et al., 30 Jun 2026)
Residual gauge-symmetry models	$S$4, $S$5	remnant $S$6 or $S$7 (Justino et al., 8 Sep 2025, Chen et al., 21 Dec 2025)

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 $S$8 and $S$9 satisfy the standard two-component system with annihilation into $\phi$0-even states and conversion terms,

$\phi$1

and the total abundance is

$\phi$2

with $\phi$3 (Esch et al., 2014). The same structural point recurs in inert-doublet and doublet–triplet models, where $\phi$4 or $\phi$5 conversion terms appear explicitly in the coupled equations and shift the final relic fractions (Borah et al., 2019, Chakrabarty et al., 2021).

Conversion processes are not a small correction. In the singlet scalar–fermion model, conversion can change $\phi$6 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 (Esch et al., 2014). In the inert-doublet plus radiative neutrino-mass construction, the entire intermediate range $\phi$7, 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 (Borah et al., 2019). In the doublet–triplet model, the IDM desert region $\phi$8 and triplet masses below about $\phi$9, both under-abundant in the standalone limits, can reproduce the Planck value once $\chi$0 drives efficient $\chi$1 conversion (Chakrabarty et al., 2021).

Semi-annihilation provides an additional depletion channel that has no analogue in the minimal one-component singlet-scalar model. In the $\chi$2 scalar–fermion setup, the relevant dark processes include $\chi$3, $\chi$4, and $\chi$5; semi-annihilation is especially important in the $\chi$6 ordering because it opens viable scalar-mass regions below the $\chi$7 GeV threshold excluded in the ordinary scalar singlet model (Yaguna et al., 2021). In $\chi$8 and $\chi$9 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 (Bélanger et al., 2020, Yaguna et al., 2021). In the type-I 2HDM with $Z_{2n}$00, semi-annihilation is singled out as a distinctive feature of $Z_{2n}$01-stabilized multi-component dark matter, although annihilation and conversion are often more important in setting the final relic abundances (Escalona et al., 18 Mar 2026).

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

$Z_{2n}$02

within the PLANCK range $Z_{2n}$03 (Pandey et al., 2017). A central difference from WIMP freeze-out is that the relic density increases with the portal couplings in the FIMP case, roughly as $Z_{2n}$04 (Pandey et al., 2017).

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 $Z_{2n}$05, $Z_{2n}$06, and $Z_{2n}$07, the scalar $Z_{2n}$08 interacts through the Higgs portal $Z_{2n}$09 and the $Z_{2n}$10-mediated terms $Z_{2n}$11 and $Z_{2n}$12; after electroweak symmetry breaking, the Higgs and singlet mix into

$Z_{2n}$13

with $Z_{2n}$14 identified as the observed 125 GeV Higgs boson (Esch et al., 2014). The scalar relic density then exhibits the familiar Higgs resonance near $Z_{2n}$15, a new resonance near $Z_{2n}$16, and the new final state

$Z_{2n}$17

which becomes important when $Z_{2n}$18 (Esch et al., 2014).

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 $Z_{2n}$19- and $Z_{2n}$20-exchange contributions,

$Z_{2n}$21

with $Z_{2n}$22, so constructive or destructive interference can enhance or suppress the rate (Esch et al., 2014). The same qualitative feature appears in the generic $Z_{2n}$23 model, where the scalar candidate $Z_{2n}$24 scatters through both $Z_{2n}$25 and $Z_{2n}$26, and destructive interference between the two Higgs exchanges is explicitly noted (Das et al., 2022).

Scale invariance leads to a distinctive scalar mediator, the scalon. In the classically scale-invariant Standard Model with two real scalar dark matter fields $Z_{2n}$27, the tree-level scalon is massless and acquires its mass only through one-loop corrections,

$Z_{2n}$28

which is phenomenologically essential because a massless scalar mediator would produce enormous dark matter–nucleon scattering and exclude the model (Ghorbani et al., 2015). 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 $Z_{2n}$29, whereas the single-component scalar case is viable mainly for $Z_{2n}$30 (Ghorbani et al., 2015). A related scale-invariant scalar–fermion model also uses a loop-generated scalon $Z_{2n}$31 and finds that the scalar mass must satisfy $Z_{2n}$32 while direct-detection constraints exclude large regions unless mediator effects suppress the nucleon cross section (Ayazi et al., 2018).

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 $Z_{2n}$33, $Z_{2n}$34, and related states are too efficient in the interval $Z_{2n}$35; in the two-component theory this under-abundance is compensated by the second component rather than treated as a failure of the model (Frank et al., 5 May 2025). In the 3HDM collider study, the key kinematic control variables are the mass splittings

$Z_{2n}$36

which fix the visible dilepton kinematics through the cascade $Z_{2n}$37 and generate a preselection double-bump structure in the dimuon invariant-mass spectrum because the two dark sectors have different $Z_{2n}$38 (Belyaev et al., 30 Jun 2026).

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

$Z_{2n}$39

(Gutiérrez-Luna et al., 2021). The BBN-era constraint is implemented through $Z_{2n}$40 and $Z_{2n}$41 at $Z_{2n}$42. Within this framework, the classical-plus-Higgs-like model survives only if the classical component dominates sufficiently, requiring $Z_{2n}$43, equivalently a Higgs-like fraction $Z_{2n}$44, whereas the axion-plus-Higgs-like model is discarded because there is no value of $Z_{2n}$45 for which $Z_{2n}$46 stays within the BBN bounds throughout nucleosynthesis (Gutiérrez-Luna et al., 2021).

Another direction emphasizes self-interactions and late-time structure formation. In the self-resonant $Z_{2n}$47 model, the resonance condition

$Z_{2n}$48

enhances elastic co-scattering through a Yukawa potential with effective mediator mass

$Z_{2n}$49

which becomes small near resonance and yields velocity-dependent self-interactions suitable for addressing small-scale structure problems at galaxies (Justino et al., 8 Sep 2025). The same setup also predicts Sommerfeld-enhanced semi-annihilation and boosted dark matter from the Galactic Center, with benchmark Sommerfeld factors $Z_{2n}$50 (Justino et al., 8 Sep 2025).

Cosmological consistency can also impose unexpectedly strong mass bounds. In the two-singlet scalar FIMP model, self-interaction constraints based on $Z_{2n}$51 push the viable masses of both components to roughly $Z_{2n}$52 for maximal allowed couplings, even though freeze-in relic-density considerations alone permit masses from GeV down to keV (Pandey et al., 2017). In the hidden $Z_{2n}$53 model, the massless dark gauge boson $Z_{2n}$54 contributes to dark radiation and induces ellipticity constraints, leading to $Z_{2n}$55 and upper bounds on $Z_{2n}$56 from halo-shape considerations (Chen et al., 21 Dec 2025).

5. Detection phenomenology

Direct detection in multicomponent theories is governed by abundance-weighted rates rather than bare cross sections. The standard rescaling is

$Z_{2n}$57

or equivalent notation such as $Z_{2n}$58 (Esch et al., 2014, Escalona et al., 18 Mar 2026). 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 $Z_{2n}$59 contributing only $Z_{2n}$60 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 (Esch et al., 2014). Comparable conclusions recur in the $Z_{2n}$61 scalar–fermion model, where both dark matter particles may be observed in future direct-detection experiments (Yaguna et al., 2021), in the $Z_{2n}$62 scalar model, where current and future direct-detection experiments may be sensitive to signals from both dark matter particles (Bélanger et al., 2020), and in the $Z_{2n}$63 prototypes, which may lead to observable signals in direct detection experiments across wider mass intervals than the ordinary Higgs-portal singlet model (Yaguna et al., 2021).

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 (Yaguna et al., 2021, Escalona et al., 18 Mar 2026). This makes scalar-involved scenarios more constrained: in the flavor-dependent $Z_{2n}$64 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 $Z_{2n}$65 (Duy et al., 25 Mar 2026). In the generic $Z_{2n}$66 model, scalar dark matter below roughly $Z_{2n}$67 GeV is strongly constrained by XENON1T unless resonance or interference effects suppress the rate (Das et al., 2022). In the hidden $Z_{2n}$68 model, by contrast, the benchmark scalar cross sections are around $Z_{2n}$69, below current XENON1T, XENONnT, and LZ limits (Chen et al., 21 Dec 2025).

Collider probes are increasingly model-specific but already nontrivial. The most explicit LHC study considers the I(2+1)HDM signal

$Z_{2n}$70

from $Z_{2n}$71 followed by $Z_{2n}$72, with $Z_{2n}$73 (Belyaev et al., 30 Jun 2026). For the representative benchmark BP1, the detector-level analysis yields $Z_{2n}$74 and $Z_{2n}$75 at Run 3 with $Z_{2n}$76, rising to $Z_{2n}$77 under a statistical-only extrapolation to $Z_{2n}$78 (Belyaev et al., 30 Jun 2026). 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 (Belyaev et al., 30 Jun 2026). Other electroweak-multiplet models predict the familiar inert-doublet channels $Z_{2n}$79 or disappearing charged tracks from nearly degenerate triplet states (Frank et al., 5 May 2025, Chakrabarty et al., 2021). 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 (Escalona et al., 18 Mar 2026).

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_{2n}$80 model admits viable sub-TeV masses that are excluded in the simple singlet scalar model (Bélanger et al., 2020). The scale-invariant two-component scalar model allows masses down to roughly $Z_{2n}$81 GeV per component, whereas the one-component case is driven to $Z_{2n}$82 (Ghorbani et al., 2015). The doublet–triplet model revives the inert-doublet desert and the sub-TeV triplet region (Chakrabarty et al., 2021), and the scotogenic two-doublet model opens the entire intermediate inert-doublet range (Borah et al., 2019).

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 (Esch et al., 2014, Das et al., 2022, Duy et al., 25 Mar 2026).

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 (Esch et al., 2014, Yaguna et al., 2021), inert multiplet models (Borah et al., 2019, Chakrabarty et al., 2021), and discrete-symmetry scalar sectors (Bélanger et al., 2020, Yaguna et al., 2021). A plausible implication is that phenomenological exclusions derived under the assumption that one candidate constitutes $Z_{2n}$83 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 (Gutiérrez-Luna et al., 2021).

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 (Esch et al., 2014, Pandey et al., 2017, Belyaev et al., 30 Jun 2026).