Singlet Scalar Dark Matter Model

• The singlet scalar dark matter model introduces a neutral scalar S, stabilized by a Z₂ symmetry and coupled via the Higgs portal, which drives its relic density and detection signals.
• It supports distinct thermal histories, with WIMP freeze-out near the Higgs resonance and FIMP freeze-in scenarios emerging from extremely low portal couplings.
• Extensions of the model include self-interactions, semi-annihilation, and mixing with additional fields, broadening its phenomenology beyond standard direct detection limits.

Searching arXiv for relevant singlet scalar dark matter papers to ground the article. Singlet scalar dark matter denotes a class of dark-sector models in which the Standard Model is extended by at least one scalar field that is neutral under the Standard Model gauge group and whose stability follows from a symmetry or charge assignment. In its minimal and most studied realization, one adds a single real scalar $S$ odd under a $Z_2$ symmetry, so that $S\to -S$ forbids odd powers of $S$, keeps $\langle S\rangle=0$, and makes the field stable. The resulting Higgs-portal framework is among the simplest renormalizable dark-matter models, yet it supports multiple cosmological realizations—notably thermal freeze-out and thermal freeze-in—and has motivated a wide family of extensions involving nonminimal Higgs sectors, semi-annihilation, gauge portals, composite dynamics, and radiative neutrino-mass constructions (Guo et al., 2010, Yaguna, 2011, Cline et al., 2013, Collaboration et al., 2017).

1. Minimal Higgs-portal construction

The minimal real-singlet model augments the Standard Model by a real scalar $S$, neutral under $SU(3)_c\times SU(2)_L\times U(1)_Y$, and odd under a stabilizing $Z_2$ symmetry. In the convention used in the classic real-singlet analyses, the Lagrangian is

$$\mathcal{L} = \mathcal{L}_{\rm SM} + \frac{1}{2}\partial_\mu S \partial^\mu S - \frac{m_0^2}{2} S^2 - \frac{\lambda_S}{4} S^4 - \lambda S^2 H^\dagger H \, ,$$

with $H$ the Standard Model Higgs doublet, $Z_2$0 the singlet bare mass parameter, $Z_2$1 the singlet quartic self-coupling, and $Z_2$2 the Higgs-portal coupling (Guo et al., 2010, Yaguna, 2011). After electroweak symmetry breaking, the Higgs vev contributes to the physical singlet mass, giving

$Z_2$3

or, in the alternative notation frequently used in later Higgs-portal studies,

$Z_2$4

The phenomenology is then controlled primarily by the singlet mass and the Higgs-portal coupling (Cline et al., 2013, Collaboration et al., 2017).

Within this minimal setup, the portal interaction generates the $Z_2$5 and $Z_2$6 vertices that govern annihilation into Standard Model states, elastic scattering on nuclei through $Z_2$7-channel Higgs exchange, and invisible Higgs decay when $Z_2$8 (Guo et al., 2010). The singlet quartic $Z_2$9 is present in the full renormalizable theory, but standard relic-density, direct-detection, and invisible-width phenomenology are usually dominated by the Higgs portal; $S\to -S$0 becomes central mainly for self-interactions, vacuum structure, and some electroweak phase-transition applications (Guo et al., 2010, Cline et al., 2013).

A common source of confusion is notation rather than physics. Some papers write the portal as $S\to -S$1, others as $S\to -S$2; both describe the same Higgs-portal mechanism. What makes the model distinctive is not the notation but the fact that all observable couplings to Standard Model matter proceed through the Higgs sector in the minimal realization (Cline et al., 2013).

2. Thermal histories: freeze-out and freeze-in

The historically standard realization is the WIMP one. In this regime the singlet is initially in thermal equilibrium with the plasma, annihilates through an $S\to -S$3-channel Higgs into Standard Model fermions, gauge bosons, or Higgs pairs, and later freezes out as the annihilation rate drops below the Hubble expansion rate (Guo et al., 2010, Collaboration et al., 2017). The relic density is computed from the Boltzmann equation

$S\to -S$4

with a full thermal average required near the Higgs pole. In this regime, increasing the portal coupling lowers the relic density by making annihilation more efficient (Guo et al., 2010). A central feature is the Higgs resonance at

$S\to -S$5

where thermal averaging over the Higgs propagator becomes essential and the required portal coupling can become very small (Cline et al., 2013, Collaboration et al., 2017).

The same Lagrangian also supports a qualitatively different FIMP realization. Yaguna showed that if the portal coupling is reduced to the range

$S\to -S$6

the singlet never thermalizes and is instead slowly produced from the Standard Model bath by freeze-in (Yaguna, 2011). In that regime the initial abundance is taken to be negligible,

$S\to -S$7

and the relevant Boltzmann equation for pair production becomes

$S\to -S$8

with the crucial approximation $S\to -S$9 at all times (Yaguna, 2011). Because $S$0, the freeze-in abundance grows with the interaction strength rather than decreasing with it: $S$1 The viable FIMP band identified in that analysis spans singlet masses from the GeV scale to the TeV scale, with essentially no direct- or indirect-detection signal expected (Yaguna, 2011).

The coexistence of freeze-out and freeze-in within the same minimal Higgs-portal theory is conceptually important. The model is therefore not solely a WIMP benchmark: it is a minimal singlet framework with two distinct thermal production mechanisms, both realized without enlarging the field content beyond the real scalar singlet (Yaguna, 2011).

3. Minimal-model parameter space and experimental constraints

The minimal Higgs-portal singlet is highly predictive because the same portal that fixes the relic density also controls spin-independent scattering on nuclei and invisible Higgs decay. In the 2013 update using the measured Higgs mass, the surviving low-mass region was narrowed to a resonance window near

$S$2

with masses below about $S$3 already robustly excluded in that analysis by the combination of relic density, invisible Higgs width, and direct detection (Cline et al., 2013). The same study found a heavy-mass tail extending to multi-TeV masses, but requiring increasingly large portal couplings, with projected XENON1T sensitivity reaching to about $S$4 (Cline et al., 2013).

Those conclusions were sharpened by the 2017 global fit using GAMBIT, which varied the singlet mass and portal coupling together with 13 nuisance parameters and combined Planck relic density, LUX, PandaX, SuperCDMS, XENON100, invisible Higgs-decay limits, IceCube, and Fermi-LAT dwarf-galaxy searches (Collaboration et al., 2017). That fit found viable solutions at couplings of order unity for singlet masses between the Higgs mass and about $S$5, and at masses above $S$6; only in the latter case can the singlet constitute all of dark matter in the nonresonant region. The same analysis also showed that the low-mass resonance region can account for all of dark matter in a frequentist treatment, but contributes less than $S$7 of the total posterior mass in the Bayesian analysis and is correspondingly identified as fine-tuned (Collaboration et al., 2017).

A second major refinement was the consistent treatment of subdominant dark matter. Rather than forcing the singlet to saturate the observed relic density everywhere, the 2013 and 2017 analyses allowed it to make up only a fraction of the total abundance, with direct-detection signals rescaled by

$S$8

and indirect signals by $S$9 (Cline et al., 2013, Collaboration et al., 2017). This changes the interpretation of exclusions substantially: a point with efficient annihilation and small relic fraction can survive direct and indirect searches even if it would be excluded under the full-DM assumption.

The minimal model is therefore constrained, but not uniformly excluded. Exclusions derived for one regime, especially the conventional WIMP region, do not automatically remove the entire singlet-scalar framework once the resonance region, subdominant-density solutions, or the FIMP realization are taken into account (Yaguna, 2011, Collaboration et al., 2017).

4. Self-interactions, semi-annihilation, and light-mass variants

The singlet scalar framework has also been studied as a self-interacting dark-matter model. Interpreting the Abell 3827 offset as evidence for

$\langle S\rangle=0$0

Aarssen, Bringmann, and Pfrommer showed that the minimal perturbative Higgs-portal singlet model is driven either to very light masses,

$\langle S\rangle=0$1

or to an extremely narrow resonant regime near $\langle S\rangle=0$2 (Campbell et al., 2015). They further argued that ordinary thermal freeze-out does not work in the minimal model under this self-interaction requirement: in the light-mass regime annihilation through the Higgs portal is too weak once collider limits are imposed, while in the resonant regime annihilation is too strong. Their preferred minimal-model solution is again freeze-in, with

$\langle S\rangle=0$3

for very light dark matter (Campbell et al., 2015).

A distinct modification arises when the stabilizing symmetry is $\langle S\rangle=0$4 rather than $\langle S\rangle=0$5. In that case the singlet must be complex, the scalar potential includes the cubic term

$\langle S\rangle=0$6

and semi-annihilation processes such as

$\langle S\rangle=0$7

contribute to the relic density (Bélanger et al., 2012). This weakens the usual minimal-model correlation between relic abundance and spin-independent direct detection because the relic density is no longer controlled solely by the same Higgs-portal coupling that sets nuclear scattering. At the same time, too large a cubic term breaks the $\langle S\rangle=0$8 symmetry spontaneously, so vacuum structure imposes an upper bound on $\langle S\rangle=0$9 and therefore a lower bound on the direct-detection cross section. In that analysis the dark-matter mass could not lie below $S$0 because of Higgs-decay constraints (Bélanger et al., 2012).

A more astrophysics-oriented variant treated the singlet as a very light, weakly self-interacting scalar capable of Bose-Einstein condensation and cored halo formation, emphasizing representative scales such as

$S$1

rather than a standard Higgs-portal WIMP analysis (Matos et al., 2014). That work explicitly did not present a full relic-density computation and framed its conclusions in terms of condensation, halo structure, and weak collider bounds rather than a precision freeze-out study. A plausible implication is that “singlet scalar dark matter model” encompasses not only the canonical heavy Higgs-portal WIMP, but also light scalar-field and self-interaction-oriented realizations whose cosmological logic is different (Matos et al., 2014).

5. Extended Higgs sectors and nonminimal portal structures

Once the Higgs sector is enlarged, the tight minimal-model relation between relic density and direct detection can be relaxed. In the leptophilic Type-X two-Higgs-doublet model with an added real singlet $S$2, the two portal couplings $S$3 and $S$4 play different roles: in the alignment and large-$S$5 limit, $S$6 controls the coupling to the SM-like Higgs and therefore direct detection, while $S$7 controls couplings to the nonstandard Higgs sector and therefore annihilation (Bandyopadhyay et al., 2017). The dominant relic-density channel becomes

$S$8

and indirect detection excludes approximately $S$9, while substantial viable parameter space remains above that scale (Bandyopadhyay et al., 2017).

An analogous decoupling appears in the Georgi–Machacek model extended by a real singlet scalar. There, the singlet retains the $SU(3)_c\times SU(2)_L\times U(1)_Y$0-stabilized Higgs-portal identity, but the enlarged custodial Higgs sector provides additional annihilation channels into new scalar states. As a result, the combination of relic density and direct detection excludes singlet masses below about $SU(3)_c\times SU(2)_L\times U(1)_Y$1, but above the Higgs pole substantial parameter space can evade near-future direct-detection experiments for masses as low as $SU(3)_c\times SU(2)_L\times U(1)_Y$2, with cross sections falling below the neutrino floor for $SU(3)_c\times SU(2)_L\times U(1)_Y$3 (Campbell et al., 2016). This is explicitly contrasted with the ordinary Standard Model plus singlet case, where future experiments were expected to exclude masses above the Higgs pole up to the multi-TeV range (Campbell et al., 2016).

Composite-Higgs embeddings provide a structurally different generalization. In the $SU(3)_c\times SU(2)_L\times U(1)_Y$4 model, the singlet dark-matter candidate is a CP-odd pNGB $SU(3)_c\times SU(2)_L\times U(1)_Y$5, stabilized by an exact $SU(3)_c\times SU(2)_L\times U(1)_Y$6 remnant of the coset structure (Cai et al., 2020). Its couplings are non-linear rather than purely renormalizable Higgs-portal ones, which introduces direct $SU(3)_c\times SU(2)_L\times U(1)_Y$7 contact terms and non-linear $SU(3)_c\times SU(2)_L\times U(1)_Y$8 interactions. Those non-linearities allow viable thermal relic density for

$SU(3)_c\times SU(2)_L\times U(1)_Y$9

with $Z_2$0 dominating below about $Z_2$1 and diboson or $Z_2$2 final states dominating above that; the surviving parameter space lies close to the direct-detection boundary and was argued to be within reach of XENONnT and LZ (Cai et al., 2020).

Gauge extensions change the stabilization mechanism itself. In the $Z_2$3 scalar model, the dark matter field is a complex scalar singlet under the Standard Model but charged under $Z_2$4, so stability follows from the $Z_2$5 charge assignment rather than an imposed $Z_2$6 (Rodejohann et al., 2015). In that framework the observed relic density can be obtained through gauge or scalar interactions, and semi-annihilation can be important for special charge assignments. If the relic density is controlled by the $Z_2$7 gauge interaction, the dark matter mass should lie below $Z_2$8 and the direct-detection cross section can be easily probed by XENON1T; if it is controlled by scalar interactions, the mass can be much larger and the detection prospects are less certain (Rodejohann et al., 2015).

6. Embedded singlets, mixed states, and conceptual scope

The singlet scalar idea also survives in models where dark matter is not purely singlet in the mass basis. In the ScotoSinglet model, a real scalar singlet $Z_2$9 is added to the scotogenic model and mixes with the inert-doublet CP-even scalar through the trilinear term $$\mathcal{L} = \mathcal{L}_{\rm SM} + \frac{1}{2}\partial_\mu S \partial^\mu S - \frac{m_0^2}{2} S^2 - \frac{\lambda_S}{4} S^4 - \lambda S^2 H^\dagger H \, ,$$0 (Beniwal et al., 2020). The dark-matter state $$\mathcal{L} = \mathcal{L}_{\rm SM} + \frac{1}{2}\partial_\mu S \partial^\mu S - \frac{m_0^2}{2} S^2 - \frac{\lambda_S}{4} S^4 - \lambda S^2 H^\dagger H \, ,$$1 interpolates continuously between a pure singlet ($$\mathcal{L} = \mathcal{L}_{\rm SM} + \frac{1}{2}\partial_\mu S \partial^\mu S - \frac{m_0^2}{2} S^2 - \frac{\lambda_S}{4} S^4 - \lambda S^2 H^\dagger H \, ,$$2), a pure inert-doublet scalar ($$\mathcal{L} = \mathcal{L}_{\rm SM} + \frac{1}{2}\partial_\mu S \partial^\mu S - \frac{m_0^2}{2} S^2 - \frac{\lambda_S}{4} S^4 - \lambda S^2 H^\dagger H \, ,$$3), and a mixed singlet-doublet state. This mixing softens direct-detection constraints, introduces a new neutrino-mass contribution proportional to $$\mathcal{L} = \mathcal{L}_{\rm SM} + \frac{1}{2}\partial_\mu S \partial^\mu S - \frac{m_0^2}{2} S^2 - \frac{\lambda_S}{4} S^4 - \lambda S^2 H^\dagger H \, ,$$4, and opens a large portion of the parameter space of the original scotogenic model (Beniwal et al., 2020).

A related lesson comes from the scalar singlet–triplet model, where a real scalar singlet and a real electroweak triplet are both $$\mathcal{L} = \mathcal{L}_{\rm SM} + \frac{1}{2}\partial_\mu S \partial^\mu S - \frac{m_0^2}{2} S^2 - \frac{\lambda_S}{4} S^4 - \lambda S^2 H^\dagger H \, ,$$5-odd and mix after electroweak symmetry breaking (Fischer et al., 2013). In the pure-singlet limit the model reproduces the standard Higgs-portal singlet behavior, but triplet admixture introduces electroweak gauge interactions and coannihilations. The “relaxed singlet scenario” shows explicitly how electroweak partners can lower the direct-detection tension of the minimal singlet model and reopen masses down to about $$\mathcal{L} = \mathcal{L}_{\rm SM} + \frac{1}{2}\partial_\mu S \partial^\mu S - \frac{m_0^2}{2} S^2 - \frac{\lambda_S}{4} S^4 - \lambda S^2 H^\dagger H \, ,$$6 (Fischer et al., 2013).

These nonminimal constructions clarify several persistent misconceptions. Singlet scalar dark matter is not equivalent to a real $$\mathcal{L} = \mathcal{L}_{\rm SM} + \frac{1}{2}\partial_\mu S \partial^\mu S - \frac{m_0^2}{2} S^2 - \frac{\lambda_S}{4} S^4 - \lambda S^2 H^\dagger H \, ,$$7-odd WIMP interacting only through a single portal coupling. Stability may come from $$\mathcal{L} = \mathcal{L}_{\rm SM} + \frac{1}{2}\partial_\mu S \partial^\mu S - \frac{m_0^2}{2} S^2 - \frac{\lambda_S}{4} S^4 - \lambda S^2 H^\dagger H \, ,$$8, $$\mathcal{L} = \mathcal{L}_{\rm SM} + \frac{1}{2}\partial_\mu S \partial^\mu S - \frac{m_0^2}{2} S^2 - \frac{\lambda_S}{4} S^4 - \lambda S^2 H^\dagger H \, ,$$9, an exact remnant symmetry, or a gauge charge; the relic density may be set by freeze-out, freeze-in, semi-annihilation, coannihilation, or non-linear interactions; and the dark matter state itself may be a pure singlet, a mixed singlet-doublet state, or a composite pNGB with singlet quantum numbers under the Standard Model (Yaguna, 2011, Bélanger et al., 2012, Rodejohann et al., 2015, Beniwal et al., 2020, Cai et al., 2020).

In that broader sense, the singlet scalar dark matter model is best understood as a family of theories organized around a simple core principle: a scalar neutral under the Standard Model gauge group can be stable and cosmologically relevant with minimal new field content, but the physical consequences depend decisively on how stability is implemented, how the scalar communicates with the visible sector, and which term in the Boltzmann dynamics controls the abundance.