---
title: 'Models in the Dark: Dark-Sector Constructions'
url: https://www.emergentmind.com/topics/models-in-the-dark
type: topic
---

# Models in the Dark: Dark-Sector Constructions

Searching arXiv for papers relevant to “Models in the Dark” and dark-sector model building.
arXivSearch:
{"query":"all:\"Models in the Dark\" OR ti:\"Models in the Dark\" OR abs:\"hidden sector dark matter models\" OR ti:\"Simplified Dark Matter Models\"","max_results":10,"sort_by":"relevance"}
I found several relevant arXiv papers on dark-sector model building, including "Simplified Dark Matter Models" [1804.01245], "Perturbative benchmark models for a dark shower search program" [2103.01238], and multiple hidden-sector, interacting-dark-energy, and inert-doublet constructions relevant to the theme.
“Models in the Dark” (Editor's term) denotes a family of constructions in which dark matter, dark energy, or both are given explicit symmetry structure, field content, mediator sectors, and cosmological dynamics, rather than being represented only by effective contact operators. In this literature, the dark sector may be stabilized by an exact \(Z_2\), a residual local discrete gauge symmetry, an unbroken local gauge symmetry, topology, or accidental symmetries; it may be connected to the Standard Model through scalar, vector, Higgs, dark-photon, graviton, or flavor portals; and it may participate in late-time cosmic acceleration, matter asymmetry generation, or hidden-sector astrophysics. A recurrent theme is that the phenomenology of darkness is controlled not by one mechanism but by the interplay of symmetry, mediation, cosmological background, and observational inference [1503.05412, 1804.01245, 2204.11676].

## 1. Programmatic scope

A central divide in dark-sector model building is between effective descriptions and explicit mediator models. Simplified dark matter models were introduced as an intermediate framework between very generic effective field theories and fully specified ultraviolet-complete theories: one starts from the Standard Model, adds a dark matter particle \(\chi\), adds one mediator field, couples the mediator renormalizably to dark matter and to quarks, gluons, leptons, or Higgs sectors, and keeps the parameter set minimal. This makes mediator propagators, resonance structure, widths, and complementary search channels explicit, which is precisely where effective field theory breaks down at LHC momentum transfers [1804.01245].

The same logic appears outside collider missing-energy searches. Hidden-sector dark matter models with local dark gauge symmetries are organized by the principle that dark matter should be stabilized and structured by its own local gauge symmetry, just as ordinary matter is shaped by the Standard Model gauge group. Interacting dark energy–dark matter theories replace ad hoc continuity-equation couplings by conformal or disformal metric relations, while perturbative hidden-valley benchmarks replace generic “exotic long-lived” signatures by a small set of reproducible shower models. The phrase therefore captures a methodological stance as much as a particle-physics subject: dark sectors are treated as structured dynamical systems, not merely as missing energy [1503.05412, 2103.01238].

## 2. Symmetry, stability, and dark-sector architecture

The most economical “dark” constructions stabilize new states through discrete or gauge symmetries.

| Construction | Organizing principle | Representative field content |
|---|---|---|
| Dark 2HDM / IDM [0911.2457] | exact \(Z_2\), inert vacuum | \(h\), \(H\), \(A\), \(H^\pm\) |
| KNT-like three-loop models [1404.6033] | \(Z_2\) forbids tree-level neutrino Yukawa couplings | \(Z_2\)-odd fermion or inert scalar dark matter |
| DMFV up-type model [1709.01930] | \(\mathrm{U}(3)_\chi\) with residual \(\mathbb{Z}_3\) symmetry | triplet \(\chi_i\) and scalar mediator \(\phi\) |
| \(Z_{N\ge 3}\) companion model [2405.05694] | semi-annihilation allowed by \(Z_N\) | Dirac \(\Psi\), companion \(S\), scalar doublet \(\eta\) |
| Dark little Higgs / NLH [1304.7835] | duplicated sigma sector with inert \(\Delta\)-doublet | \(\Sigma\)-sector \(h\), \(\Delta\)-sector \(\xi\) |

In the Dark 2HDM, also called the Inert Doublet Model, the exact symmetry is
\[
\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2,
\]
with \(\phi_1\) and all Standard Model fields \(Z_2\)-even, \(\phi_2\) \(Z_2\)-odd, and \(\langle \phi_2\rangle=0\). The lightest \(Z_2\)-odd scalar is then stable and can be dark matter; in practice the candidate is typically \(H\), though \(A\) is also possible. Because the inert doublet does not couple directly to fermions, the dark scalars communicate through electroweak gauge interactions and the Higgs portal [0911.2457].

The same symmetry logic appears in radiative neutrino-mass models. In the Krauss–Nasri–Trodden class and its variants, the dark matter particle propagates in the inner loop of a three-loop neutrino-mass diagram, and the same \(Z_2\) symmetry that stabilizes it also forbids the tree-level neutrino Yukawa couplings that would otherwise generate ordinary seesaw masses. The simplest realizations use Majorana dark matter; related Dirac-mediator constructions replace fermionic dark matter by inert singlet, doublet, or triplet scalar dark matter [1404.6033].

Other models use flavor or replicated symmetry structure. In Dark Minimal Flavour Violation, the dark sector carries its own \(\mathrm{U}(3)_\chi\), the dark matter consists of a triplet of Dirac fermions \(\chi_i\), and the lightest \(\chi\) is stabilized by a residual \(\mathbb{Z}_3\) symmetry. In the dark little Higgs construction, a duplicated \(SU(5)_\Sigma/SO(5)_\Sigma\) and \(SU(5)_\Delta/SO(5)_\Delta\) structure produces an inert \(\Delta\)-sector doublet because fermions are charged only under the original \(\Sigma\) sector; after integrating out the heavy states, the low-energy scalar theory becomes precisely an inert doublet model [1709.01930, 1304.7835].

The broadest symmetry-based framework is the hidden-sector gauge approach. There, dark matter may be stabilized by unbroken local gauge symmetry, residual discrete gauge symmetry after symmetry breaking, topology, or accidental symmetries, and the particle contents and their dynamics are fixed by local gauge symmetries. This yields a spectrum in which dark gauge bosons and dark Higgs bosons are not optional embellishments but inevitable mediators [1503.05412].

## 3. Mediators, portals, and simplified phenomenology

Mediator taxonomy is the backbone of modern dark-sector phenomenology. For fermionic dark matter \(\chi\), the standard first-generation simplified models include vector, axial-vector, scalar, and pseudoscalar \(s\)-channel mediators, as well as \(t\)-channel colored mediators. The virtue of these constructions is not merely notational economy: explicit widths,
\[
\Gamma = \Gamma_\chi + \sum_f \Gamma_f + \Gamma_{gg},
\]
and Breit–Wigner production,
\[
\sigma \propto \frac{g_\chi^2 g_q^2}{(s-m_{\rm med}^2)^2+m_{\rm med}^2\Gamma^2},
\]
make resonance regions, off-shell suppression, and search complementarity manifest. At the same time, these models are not automatically self-consistent, a point returned to below [1804.01245].

A more collider-specific realization is the perturbative hidden-valley program. Five benchmark dark-shower models span a broad range of experimentally relevant topologies by choosing one visible dark particle and one decay portal: gluon, photon, vector, Higgs, or dark photon. The initiating state is taken to be a heavy \(s\)-channel mediator \(H\to\bar\psi\psi\), identified with the Standard Model Higgs when \(m_H=125\) GeV, and the dark shower is modeled in PYTHIA 8 with \(N_c=3\) and \(N_\psi=1\). Once the portal is fixed, branching ratios are determined by the visible dark particle mass and, in some cases, by its CP or spin structure, while the ultraviolet completion implies lower bounds on the lifetime \(c\tau\) [2103.01238].

Direct detection revisits the same mediator structures from a different angle. In five simplified spin-0 mediator models labeled D2, D3, D4, C, and V, the tree-level WIMP–nucleon amplitudes are momentum suppressed, yet one-loop box diagrams induce effective operators of the form
\[
\sum_q \frac{m_q^2}{v_0^2}\,\mathcal{C}(m_{\rm DM},m_{\rm Med},m_q)\,\overline{\rm DM}\,{\rm DM}\,\bar q q,
\]
which generate non-suppressed spin-independent scattering. After matching onto nucleons,
\[
\sigma_{\rm DM\,N}^{\rm SI} = \frac{\mu_{\rm DM\,N}^2}{\pi}\,|\mathcal{I}_i(m_{\rm DM},m_{\rm Med})|^2,
\]
and Xenon1T excludes mediator masses up to about \(30\)–\(50\) GeV in substantial regions of parameter space, showing that tree-level momentum suppression does not guarantee weak direct-detection constraints [1804.02120].

The mediator concept can even be generalized to curved spacetime. In gravity-mediated dark matter models on de Sitter space, the interaction is written universally as
\[
S_{\text{int}}=\alpha \int d^4x\, \sqrt{-g}\, T^{\mu\nu} h_{\mu\nu},
\]
with dark matter entering only through its energy-momentum tensor. The Euclidean generating functional on \(S^4\) yields free and interacting Green’s functions for the symmetric traceless divergence-less graviton sector, making the dark matter–graviton interaction computable by standard quantum field theory techniques in curved spacetime [1804.00943].

## 4. Coupled dark sectors, dark energy, and asymmetry transfer

Dark-sector interaction is not limited to portals into the Standard Model. In interacting dark energy models, the background conservation equations are
\[
\dot\rho_m + 3H\rho_m = Q,\qquad
\dot\rho_w + 3H(1+w)\rho_w = -Q,
\]
with the paper studying the ansatz
\[
Q(z)=\alpha H(1+z)^{-\beta}\rho_w.
\]
Because energy transfer changes the matter content of a fixed comoving volume between recombination and the present, the matter density inferred from the CMB under the assumption of no interaction is generically shifted with respect to the true present-day value. The reconstructed effective equation of state becomes
\[
w_{\mathrm{eff}} =\frac{w}{1+\frac{\Omega_{m0}}{1-\Omega_{m0}}\frac{\Delta g(z)}{f(z)}(1+z)^{-3w}},
\]
and even perfect knowledge of \(H(z)\) is insufficient to recover the true \(w(z)\) if the interaction is ignored. A non-phantom interacting model can therefore mimic phantom background evolution and primary CMB anisotropies [1201.0550].

A geometrically sharper version replaces phenomenological \(Q\) by a distinct dark-matter metric,
\[
\tilde g_{\mu\nu}=C(\phi)g_{\mu\nu}+D(\phi)\partial_\mu\phi\,\partial_\nu\phi.
\]
Standard matter, radiation, and the scalar field live on \(g_{\mu\nu}\), while dark matter follows geodesics of \(\tilde g_{\mu\nu}\). The resulting interaction induces a fifth force on dark matter, and singularity structure becomes frame dependent: with suitable conformal coupling, a Big Bang at finite \(t\) can map to \(\tilde t\to -\infty\), and a Big Rip at finite \(t_s\) can map to \(\tilde t\to +\infty\). The same paper finds that simple phenomenological models of this type fit Union 2.1 SNe Ia, BAO, and \(H(z)\) data about as well as \(\Lambda\)CDM, with \(\chi^2_{\rm red}\approx 0.96\) [2204.11676].

Interacting holographic dark energy realizes the same theme at the Lagrangian level. The dark-energy density is fixed by the holographic bound
\[
\rho_\Lambda = 3c^2 M_{Pl}^2 L^{-2},
\]
with \(L\) taken to be the future event horizon, and dark energy is coupled to a fermionic dark matter field through a Yukawa term \((M-\beta\varphi)\bar\Psi\Psi\). For both interacting quintessence and interacting tachyon models, the scalar potential is reconstructed rather than assumed, and combined fits to CMB distance information, BAO, lookback time, and the Constitution supernova sample favor a negative coupling \(\delta<0\), meaning that dark energy decays into dark matter and the coincidence problem is alleviated [1009.6198].

Interaction can also generate asymmetry. In \(Z_{N\ge 3}\) dark matter–companion models, semi-annihilation,
\[
N_1N_1 \to S_1^* h,
\]
violates dark number, complex couplings provide CP violation, and freeze-out supplies the departure from equilibrium. The companion subsequently decays via
\[
S_1 \to N_1 + \nu_L,
\]
transferring asymmetry to the leptonic sector before sphalerons partially convert it into baryon asymmetry. The paper emphasizes that thermal motion enhances the CP violation parameter for the first time in this context, and preliminary Boltzmann-equation analysis shows that both correct relic density of dark matter and baryon asymmetry can be accommodated [2405.05694].

## 5. Inference, lensing, stars, and other diagnostics

The diversity of dark-sector models has motivated equally diverse discrimination strategies. One example is model selection in dark energy with a hybrid variational autoencoder / generative adversarial network trained on Union2.1 type Ia supernova distance moduli. The network learns an analytical variational approximation to the true posterior of latent cosmological parameters, replacing repeated Bayesian-evidence integration,
\[
p(\boldsymbol{x}\mid M_i) = \int p(\boldsymbol{x}\mid \boldsymbol{\theta}_i, M_i)\,p(\boldsymbol{\theta}_i\mid M_i)\,d\boldsymbol{\theta}_i,
\]
by amortized inference. For Union2.1 it assigns
\[
\mathcal{D}(\Lambda CDM \mid \boldsymbol{\mu}_{obs}) = 56.2\%, \quad
\mathcal{D}(\omega CDM \mid \boldsymbol{\mu}_{obs}) = 28.6\%, \quad
\mathcal{D}(CPL \mid \boldsymbol{\mu}_{obs}) = 15.1\%,
\]
consistent with the Bayesian ordering \(\Lambda\)CDM \(>\omega\)CDM \(>\) CPL, while also reconstructing and interpolating the distance-modulus curve. The quoted classification accuracy of about \(47.9\%\) with observational noise is interpreted as reflecting intrinsic overlap of the model distributions rather than a failure of the method [1907.00568].

Cosmological background geometry can itself be diagnostic. In a singular isothermal sphere treatment of gravitational lensing, the optical depth for multiple imaging is
\[
\tau = \frac{\mathcal F}{30} r_s^3,\qquad
{\cal F} \equiv 16\pi^3 n_0 a_0^3 \left(\frac{\sigma_v}{c}\right)^4.
\]
Because \(r_s(z)\) depends on the Hubble history \(h(z)\), the optical depth varies across CDM, \(\Lambda\)CDM, Bose–Einstein condensate dark matter, Chaplygin gas, viscous fluid, and holographic dark energy cosmologies. In the models studied, increasing \(\Omega_m\) lowers \(\tau\) in CDM, realistic changes in \(\Omega_\Lambda\) produce only modest effects in \(\Lambda\)CDM, and some holographic and generalized Chaplygin models are nearly degenerate in lensing statistics [1210.1868].

Dark-matter phenomenology also enters stellar structure. A dark star is defined broadly as any stellar object whose structure or evolution has been affected by dark matter annihilation. About \(90\%\) of the WIMP rest mass is deposited locally as heat; because stars have negative specific heat, this extra support tends to make them larger, cooler, less dense, and more diffuse than ordinary stars of the same mass. The public code \(\textsf{DarkStars}\) computes grids of such stellar evolutionary models including capture, internal WIMP distribution, annihilation heating, and conductive transport, and the resulting Pop III populations can either delay reionisation, if they remain very cool and dark-matter dominated, or shift it to higher redshift, if they become more massive and more luminous [1101.1029].

## 6. Consistency, ultraviolet completion, and conceptual limits

Minimal benchmark models are useful precisely because they are incomplete. The review of simplified dark matter models stresses that naive scalar and pseudoscalar couplings such as \(S\bar q q\) and \(P\bar q\gamma_5 q\) are not fully invariant under the Standard Model electroweak gauge group, that vector and axial-vector constructions are often only invariant under unbroken \(\mathrm{SU}(3)_c\times \mathrm{U}(1)_{\rm em}\), and that axial-vector mediators can violate perturbative unitarity. Once the mediator is promoted to a genuine gauge boson, anomaly cancellation, Higgs charge assignments, and additional scalar states become unavoidable. A plausible implication is that the most realistic “simplified” models tend to grow into richer sectors with dark Higgs fields, extra fermions, or extended gauge structure [1804.01245].

Hidden-sector gauge theories provide such structure explicitly. In local dark gauge-symmetry models, dark gauge bosons and dark Higgs bosons are the two natural force mediators, and ultraviolet completion can qualitatively alter low-energy expectations. For example, in a singlet-scalar completion of a Higgs-portal fermion dark matter model, the renormalized spin-independent cross section satisfies
\[
\sigma_{\rm SI}^{\rm ren} = \sigma_{\rm SI}^{\rm EFT} \left(1-\frac{m_{125}^2}{m_1^2}\right)^2 \cos^4\alpha,
\]
showing destructive interference and mediator mixing effects absent in the effective theory. The same framework supports local \(Z_3\) dark matter, hidden-sector monopoles, dark radiation from an unbroken \(U(1)_X\), GeV gamma-ray excess explanations through dark Higgs cascades, and Higgs inflation assisted by dark Higgs [1503.05412].

At the ultraviolet extreme, the nonsupersymmetric \(SO(16)\times SO(16)'\) heterotic string provides a tachyon-free, modular invariant, anomaly-free setting with a hidden nonabelian gauge sector and a bi-fundamental portal fermion
\[
\chi \sim (16,16').
\]
The hidden \(SO(16)'\) sector can confine into dark glueballs or other composite dark states, while compactification on \(S^2\times S^2\times S^2\) with flux yields an effective radion potential
\[
V_{\text{eff}}(b)=(4\pi)^3 b^{-6} \left(-3b^{-2}+\lambda_{10}+\frac{1}{2}f^2 b^{-4}\right),
\]
with a local minimum at \(b_m \approx 630.93\) and a small positive four-dimensional cosmological constant \(\lambda_4 \approx 1.16061\times 10^{-22}\). In this setting, dark matter, dark energy, Higgs phenomenology, and gauge–Higgs unification are not separate topics but coupled consequences of a single visible-plus-hidden construction [1907.01944].

Taken together, these models show that darkness in contemporary theory is not a synonym for agnosticism. It is a domain in which symmetry protection, mediator structure, cosmological interaction, and data-driven discrimination are developed with the same formal specificity as in visible-sector model building, even though the observational handles range from Xenon1T and the LHC to supernovae, reionisation, gravitational lensing, and the geometry of de Sitter space.

Source: https://www.emergentmind.com/topics/models-in-the-dark