Dark Sector Thermal History

Written for Apiary – where the buzz of bees meets the hum of self‑governing AI agents.

Introduction

In the grand tapestry of the cosmos, the visible matter that makes up stars, planets, and the honey‑sweet nectar we cherish is just the tip of an iceberg. Roughly 85 % of the Universe’s matter is dark, interacting with us only through gravity. Yet the story of how this dark matter (DM) came to fill the cosmos is intimately tied to the thermal history of the early Universe—a narrative written in the language of temperature, interaction rates, and expansion.

Two leading chapters of that story are known as freeze‑out and freeze‑in. In the freeze‑out picture, a dark particle once roamed the hot plasma in thermal equilibrium, only to “decouple” when the Universe cooled and its annihilation rate fell below the Hubble expansion rate. In the freeze‑in scenario, the dark particle never attains equilibrium; instead, a feeble portal leaks a tiny number of particles from the hot bath into the hidden sector, gradually “filling” the dark side.

Why do these mechanisms matter beyond the realm of particle physics? The same principles that govern the birth of dark matter—population dynamics, resource scarcity, and temperature‑dependent interactions— echo in ecosystems of honeybees and in the emergent behavior of autonomous AI agents. Understanding the thermal evolution of a hidden sector can therefore illuminate how delicate balances are struck in nature and technology alike, offering fresh perspectives for conservation and for designing self‑governing systems that respect the limits of their environment.

In this pillar article we will explore the physics of freeze‑out and freeze‑in, the role of hidden‑sector temperatures, and the observable fingerprints they leave on the cosmic microwave background (CMB), big‑bang nucleosynthesis (BBN), and large‑scale structure. Along the way we’ll draw honest bridges to bee colonies and AI agents, illustrating how the same concepts of thermal coupling, feedback, and resource allocation appear across scales.

1. The Standard Cosmological Thermal Narrative

The early Universe, from 10⁻⁴⁰ s after the Big Bang to a few seconds later, was a hot, dense plasma of photons, leptons, quarks, and gluons. Its temperature \( T \) fell as the scale factor \( a \) expanded, roughly following \( T \propto a^{-1} \) during the radiation‑dominated era.

The Hubble rate in this epoch is given by

\[ H(T) \simeq 1.66 \, \sqrt{g_\* (T)} \, \frac{T^{2}}{M_{\rm Pl}}, \]

where \( g_\* (T) \) counts the relativistic degrees of freedom (DOF) and \( M_{\rm Pl}=1.22\times10^{19}\,{\rm GeV} \) is the Planck mass. At \( T \sim 1\,{\rm MeV} \), the Universe expands at \( H \sim 10^{-22}\,{\rm GeV} \) (about 0.1 s⁻¹).

As the temperature drops, particle species either remain in equilibrium (if reaction rates \( \Gamma \gg H \)) or decouple (if \( \Gamma \lesssim H \)). Decoupling freezes the comoving number density, which later manifests as the relic abundance we observe today. For the Standard Model (SM), the most famous decoupling events are:

• Neutrino decoupling at \( T \approx 2\,{\rm MeV} \), leading to a temperature ratio \( T_\nu/T_\gamma = (4/11)^{1/3} \approx 0.714 \).
• Electron‑positron annihilation around \( T \approx 0.5\,{\rm MeV} \), heating photons relative to neutrinos.

These events set the stage for big‑bang nucleosynthesis (big-bang-nucleosynthesis) and the cosmic microwave background (cosmic-microwave-background). Any hidden sector that interacts with the SM will experience a similar thermal history, but the details—especially the temperature of the hidden bath—can differ dramatically.

2. Hidden Sectors: What They Are and How They Couple

A hidden sector (also called a dark sector) comprises particles that are singlets under the SM gauge groups. They may carry their own gauge symmetries (e.g., a dark U(1)\_D), scalar fields, or fermions. The sector is “hidden” because SM particles do not feel its forces directly; instead, portal interactions provide the only bridge. Common portals include:

The temperature ratio \( \xi \equiv T_{\rm dark}/T_{\rm SM} \) is a crucial parameter. If the portal is strong enough to keep the two baths in equilibrium until a temperature \( T_{\rm dec} \), then after decoupling the hidden sector cools faster (or slower) depending on its own relativistic DOF \( g_{\rm dark} \). A simple estimate for the ratio after decoupling is

\[ \xi = \left( \frac{g_{\}^{\rm SM}(T_{\rm dec})}{g_{\}^{\rm SM}(T_{\rm now})} \right)^{1/3} \left( \frac{g_{\}^{\rm dark}(T_{\rm now})}{g_{\}^{\rm dark}(T_{\rm dec})} \right)^{1/3}. \]

Typical scenarios give \( \xi \) anywhere from \(10^{-3}\) (a “cold” hidden sector) to \(0.9\) (almost thermalized). The value of \( \xi \) directly influences the relic density, as we will see.

3. Freeze‑Out: The Classic WIMP Paradigm

3.1 The Boltzmann Equation

For a dark particle \( \chi \) that annihilates to SM states (or to other dark particles) with cross section \( \sigma v \), the comoving number density \( Y \equiv n_\chi / s \) (where \( s \) is the entropy density) obeys

\[ \frac{dY}{dx} = -\frac{s \langle \sigma v \rangle}{Hx}\left( Y^{2} - Y_{\rm eq}^{2} \right), \qquad x \equiv \frac{m_\chi}{T}. \]

When \( \Gamma_{\rm ann} = n_\chi \langle \sigma v \rangle \) drops below \( H \), the term in parentheses forces \( Y \) to freeze at a constant value \( Y_\infty \).

3.2 The “Thermal Relic” Benchmark

A remarkable coincidence is that a weak‑scale cross section

\[ \langle \sigma v \rangle_{\rm WIMP} \approx 3 \times 10^{-26}\,{\rm cm^{3}\,s^{-1}} \]

produces the observed dark matter density

\[ \Omega_{\rm DM} h^{2} \approx 0.12, \]

provided the particle mass lies between 10 GeV and 10 TeV. This “WIMP miracle” is the cornerstone of the freeze‑out narrative.

3.3 Freeze‑Out Temperature and Relic Density

The freeze‑out temperature is roughly

\[ x_{\rm fo} \equiv \frac{m_\chi}{T_{\rm fo}} \simeq \ln\!\left[ 0.038\,\frac{g\,M_{\rm Pl}\,m_\chi \langle \sigma v \rangle}{\sqrt{g_\* x_{\rm fo}}} \right], \]

where \( g \) counts the internal DOF of \( \chi \). For a 100 GeV WIMP with \( \langle \sigma v \rangle = 3\times10^{-26}\,{\rm cm^{3}\,s^{-1}} \) and \( g=2 \), one finds \( x_{\rm fo} \approx 20–25 \), i.e. \( T_{\rm fo} \approx 4–5\) GeV.

The final relic abundance is

\[ Y_\infty \approx \frac{3.79 \times 10^{-11}\,{\rm GeV^{-1}}}{g_{\*}^{1/2} \, m_\chi \, \langle \sigma v \rangle}, \]

which translates to a present‑day number density \( n_\chi = Y_\infty s_0 \) with entropy today \( s_0 = 2891\,{\rm cm^{-3}} \).

3.4 Hidden‑Sector Freeze‑Out

If the dark particle annihilates within the hidden sector (e.g., \( \chi\chi \to \phi\phi \) where \( \phi \) is a dark mediator), the calculation is identical, but the temperature entering the Boltzmann equation is \( T_{\rm dark} = \xi T_{\rm SM} \). Consequently, the freeze‑out condition becomes

\[ \Gamma_{\rm ann}(T_{\rm dark}) \simeq H(T_{\rm SM}), \]

which for a given \( \langle \sigma v \rangle \) pushes freeze‑out to a higher SM temperature if \( \xi < 1 \). The relic density scales as

\[ \Omega_{\chi} h^{2} \propto \frac{\xi}{\langle \sigma v \rangle}, \]

so a colder hidden sector (small \( \xi \)) requires a larger cross section to achieve the same abundance. This is a key lever for model builders.

4. Freeze‑In: The Feeble Interaction Regime

4.1 The Basic Idea

In the freeze‑in scenario, the portal coupling is so tiny that the hidden sector never reaches equilibrium with the SM bath. Instead, a steady, low‑rate production of dark particles occurs, typically from decays or scatterings of SM particles. The integral form of the Boltzmann equation for a decay \( A \to B + \chi \) is

\[ Y_\chi (T) = \int_{T}^{\infty} \frac{dT'}{H(T')\,s(T')} \, \gamma_{A\to B\chi}(T'), \]

where \( \gamma \) is the reaction density. Because the integrand falls rapidly as the temperature drops below the mass of the parent particle, most of the yield is generated when \( T \sim m_A \).

4.2 A Simple Example: Higgs Portal Freeze‑In

Consider a scalar DM candidate \( S \) coupled via the Higgs portal with coupling \( \lambda_{HS} \). The dominant production channel at temperatures above the electroweak scale is Higgs boson decay \( h \to SS \). The decay width

\[ \Gamma_{h\to SS} = \frac{\lambda_{HS}^{2} v^{2}}{32\pi m_h} \sqrt{1-\frac{4m_S^{2}}{m_h^{2}}} \]

(with \( v = 246\,{\rm GeV} \) the Higgs vacuum expectation value) yields a relic density

\[ \Omega_{S} h^{2} \approx 0.12 \left( \frac{\lambda_{HS}}{10^{-11}} \right)^{2} \left( \frac{m_S}{100\,{\rm keV}} \right). \]

Thus a tiny coupling \( \lambda_{HS} \sim 10^{-11} \) suffices to generate the full DM abundance, provided the scalar mass is in the sub‑MeV range.

4.3 Freeze‑In Temperature Dependence

If the hidden sector has its own temperature \( T_{\rm dark} \), the production rate can be altered. For a cold hidden sector (\( \xi \ll 1 \)), the final DM number density is reduced because the phase‑space factor \( (1 + f_{\rm dark}) \) is negligible, and the hidden bath cannot re‑absorb the produced particles. Conversely, a warm hidden sector (\( \xi \sim 1 \)) can lead to dark sector thermalization if enough particles are produced—a phenomenon called “dark freeze‑in” that blends the two mechanisms.

4.4 The “Infrared Freeze‑In”

A useful classification distinguishes UV‑dominated freeze‑in, where production is controlled by high‑temperature processes (e.g., scatterings at \( T \gg m_{\rm SM} \)), from IR‑dominated freeze‑in, where decays near the mass threshold dominate. The former is sensitive to the reheating temperature \( T_{\rm RH} \); for a vector portal with kinetic mixing \( \varepsilon \), the relic density scales as

\[ \Omega_{\rm DM} h^{2} \propto \varepsilon^{2} \left( \frac{T_{\rm RH}}{10^{9}\,{\rm GeV}} \right). \]

Thus, the thermal history of the Universe—including inflationary reheating—can directly shape the dark matter abundance.

5. Temperature Ratios and Dark Sector Thermodynamics

5.1 Entropy Conservation and the Ratio \( \xi \)

When the two sectors decouple, each conserves its own entropy \( s = \frac{2\pi^{2}}{45} g_{\*S} T^{3} \). The ratio \( \xi \) can be tracked via

\[ \xi(T) = \left( \frac{g_{\S}^{\rm SM}(T)}{g_{\S}^{\rm SM}(T_{\rm dec})} \right)^{1/3} \left( \frac{g_{\S}^{\rm dark}(T_{\rm dec})}{g_{\S}^{\rm dark}(T)} \right)^{1/3}. \]

If the hidden sector contains only a dark photon (two polarizations) plus a light fermion, then \( g_{\S}^{\rm dark}=2+ \frac{7}{8}\times 2 = 3.5 \). If decoupling occurs before the QCD phase transition (when SM \( g_{\S} \approx 61.75 \)), the ratio today becomes

\[ \xi_{0} \approx \left( \frac{3.5}{61.75} \right)^{1/3} \approx 0.36. \]

A colder hidden sector reduces the effective number of relativistic species \( N_{\rm eff} \) that appear in the CMB.

5.2 Impact on \( N_{\rm eff} \)

The contribution of a hidden relativistic component to the radiation energy density is

\[ \Delta N_{\rm eff} = \frac{8}{7}\left( \frac{11}{4} \right)^{4/3} \sum_i g_i \xi^{4}, \]

where the sum runs over all hidden relativistic DOF. For the dark photon example with \( \xi = 0.36 \),

\[ \Delta N_{\rm eff} \approx 0.027, \]

well below the current Planck 2018 limit \( N_{\rm eff} = 2.99 \pm 0.17 \). Future CMB Stage‑4 experiments aim for a sensitivity of \( \sigma(N_{\rm eff}) \sim 0.02 \), meaning that such a hidden sector could be detectable if \( \xi \gtrsim 0.5 \).

5.3 Dark Acoustic Oscillations

If the hidden sector possesses self‑interactions (e.g., dark photons coupling to dark baryons), the resulting dark acoustic oscillations (DAO) can suppress matter power on small scales. The characteristic DAO scale \( k_{\rm DAO} \) depends on the sound speed \( c_s \) in the dark plasma and the time of dark recombination. For a temperature ratio \( \xi = 0.5 \) and a dark fine‑structure constant \( \alpha_D = 0.01 \), calculations (e.g., Cyr‑Racine & Sigurdson 2013) yield

\[ k_{\rm DAO} \sim 10\,h\,{\rm Mpc^{-1}}, \]

corresponding to a suppression of the halo mass function below \( M \sim 10^{9}\,M_\odot \). This can be probed by Lyman‑α forest data and by the satellite counts of Milky Way–like galaxies, offering a window onto hidden‑sector temperatures.

6. Cosmological Imprints: CMB, BBN, and Structure Formation

6.1 Big‑Bang Nucleosynthesis (BBN)

During BBN ( \( T \sim 0.1–1\,{\rm MeV} \) ), the expansion rate is sensitive to the total radiation density. A hidden sector with \( \Delta N_{\rm eff} = 0.2 \) would increase the Hubble rate by roughly 10 %, leading to a higher helium‑4 mass fraction \( Y_p \) by \( \Delta Y_p \approx 0.013 \). Current observations of primordial deuterium and helium constrain \( \Delta N_{\rm eff} \lesssim 0.3 \) at the 95 % confidence level, already limiting many dark‑photon models.

6.2 Cosmic Microwave Background

Beyond \( N_{\rm eff} \), the phase shift of acoustic peaks in the CMB is altered by any extra relativistic component that free‑streams. A cold hidden sector (small \( \xi \)) that is tightly coupled internally but not to photons produces a different phase shift than a free‑streaming neutrino‑like component. Detailed analyses (e.g., Baumann et al. 2016) show that the polarization spectra are especially sensitive, allowing future experiments to discriminate between dark photons, dark neutrinos, and standard neutrinos.

6.3 Large‑Scale Structure (LSS)

The matter power spectrum \( P(k) \) is suppressed on scales below the free‑streaming length of the dark particle. For a freeze‑out WIMP with mass \( m_\chi = 100\,{\rm GeV} \), the corresponding comoving free‑streaming horizon is tiny (\( k_{\rm fs} \sim 10^{6}\,h\,{\rm Mpc^{-1}} \)), making the effect invisible to current surveys. In contrast, a freeze‑in keV‑scale sterile neutrino yields \( k_{\rm fs} \sim 5\,h\,{\rm Mpc^{-1}} \), suppressing dwarf‑galaxy formation—a scenario that can be tested with satellite counts and strong lensing flux‑ratio anomalies.

7. Model Building: Portals and Mediators

7.1 Higgs Portal Freeze‑Out

A simple model: a real scalar \( S \) with a quartic coupling to the Higgs. The annihilation cross section into SM particles is

\[ \langle \sigma v \rangle_{SS\to \rm SM} \simeq \frac{\lambda_{HS}^{2} v^{2}}{4\pi m_S^{2}} \frac{1}{(4m_S^{2}-m_h^{2})^{2}+m_h^{2}\Gamma_h^{2}}. \]

Choosing \( m_S = 60\,{\rm GeV} \) (just below the Higgs resonance) and \( \lambda_{HS}=0.02 \) yields \( \langle \sigma v \rangle \approx 3\times10^{-26}\,{\rm cm^{3}\,s^{-1}} \), giving the right relic density via freeze‑out. The hidden temperature matters: if the dark sector is colder by a factor \( \xi=0.5 \), the freeze‑out temperature doubles, and the relic density would be overproduced unless \( \lambda_{HS} \) is increased to \( \sim 0.03 \).

7.2 Vector Portal Dark Photon Freeze‑In

A dark photon \( A' \) kinetically mixes with the SM photon via \( \varepsilon \). Production occurs mainly through electron‑positron annihilation \( e^{+}e^{-} \to \gamma A' \) when \( T \gtrsim m_{e} \). The reaction density

\[ \gamma_{e^{+}e^{-}\to \gamma A'} \approx \frac{\varepsilon^{2}\alpha^{2} T^{6}}{2\pi^{3}}, \]

integrated over temperature, gives a relic density

\[ \Omega_{A'} h^{2} \approx 0.12 \left( \frac{\varepsilon}{10^{-10}} \right)^{2} \left( \frac{m_{A'}}{10\,{\rm MeV}} \right). \]

Thus, a tiny kinetic mixing \( \varepsilon \sim 10^{-10} \) can produce the full DM abundance without ever equilibrating.

7.3 Dark QCD‑Like Sectors

If the hidden sector has its own non‑abelian gauge group that confines at a scale \( \Lambda_D \sim 100\,{\rm MeV} \), the lightest dark baryon can be a DM candidate. The freeze‑out temperature is typically \( T_{\rm fo} \sim \Lambda_D \), and the relic abundance depends on the dark pion mass and the dark confinement temperature. Lattice simulations suggest that the annihilation cross section can be as large as \( \sigma v \sim 10^{-24}\,{\rm cm^{3}\,s^{-1}} \), easily achieving the observed density even for sub‑GeV masses, provided the hidden sector temperature is modest (\( \xi \sim 0.3 \)).

8. Implications for Direct and Indirect Detection

8.1 Direct Detection Sensitivity to Temperature

Traditional direct‑detection experiments (e.g., XENONnT, LZ) rely on elastic scattering of DM off nuclei. The recoil spectrum depends on the DM velocity distribution, usually assumed to be a Maxwell‑Boltzmann distribution with a dispersion of 220 km s⁻¹. If the hidden sector is colder, the velocity dispersion scales as \( \xi^{1/2} \), leading to narrower recoil spectra. For a 50 GeV WIMP with \( \xi = 0.5 \), the typical speed drops to ≈ 155 km s⁻¹, reducing the expected rate by roughly 30 %. Experiments must therefore consider non‑standard velocity distributions when interpreting null results.

8.2 Indirect Detection and Dark Radiation

Freeze‑out models often predict annihilation today in high‑density regions (e.g., dwarf spheroidal galaxies). The thermal-averaged cross section needed for relic abundance (\(3\times10^{-26}\,{\rm cm^{3}\,s^{-1}}\)) is within reach of gamma‑ray observations. However, if the hidden sector temperature is low, the present‑day annihilation rate can be suppressed because the DM number density is unchanged but the relative velocity is lower, leading to a \(p\)-wave or Sommerfeld‑enhanced cross section that scales with \( v^{2} \) or \( 1/v \).

For freeze‑in models, the annihilation cross section is typically far below the thermal benchmark, rendering indirect searches ineffective. Instead, decay signatures (e.g., a sterile neutrino decaying to an X‑ray photon) become the primary probe. The X‑ray line at 3.5 keV observed in several galaxy clusters remains a tantalizing hint that could be interpreted as a 7 keV sterile neutrino produced via freeze‑in.

9. Lessons from Bees and Self‑Governing AI

9.1 Thermal Coupling as Resource Sharing

A bee colony regulates its internal temperature through collective behavior: foragers bring in nectar (energy), while workers fan their wings to dissipate heat. The temperature ratio between the hive interior and the external environment determines the development rate of brood. Similarly, a hidden sector’s temperature ratio \( \xi \) dictates how quickly dark particles “grow” or “freeze”. In both cases, a feedback loop—the colony’s response to temperature changes—is essential for stability.

9.2 Freeze‑Out as “Resource Exhaustion”

When a bee population exhausts its nectar stores, foraging slows, and the colony enters a maintenance mode. This mirrors a dark particle’s freeze‑out: the annihilation “resource” (other dark particles) becomes scarce relative to the expansion “demand”. The colony’s self‑governing protocols (e.g., task allocation) are analogous to the Boltzmann equation, which balances production and loss terms to achieve a steady state.

9.3 Freeze‑In and “Gentle Seeding”

In some beekeeping practices, queen replacement involves introducing a new queen at low density, allowing her to gradually assume reproductive dominance. This gentle seeding mirrors freeze‑in, where a tiny portal coupling slowly seeds the hidden sector without overwhelming it. The success of both processes hinges on maintaining a low interaction rate that avoids destabilizing the existing system.

9.4 AI Agents as Dark‑Sector Simulators

Self‑governing AI agents tasked with managing a shared resource (e.g., a power grid) can be programmed to emulate thermal decoupling: each agent’s “temperature” reflects its workload, and the communication bandwidth acts as the portal. By adjusting the interaction strength (analogous to \( \lambda_{HS} \) or \( \varepsilon \)), designers can explore regimes where agents synchronize (thermal equilibrium) versus operate independently (decoupled). Such simulations can test the robustness of policies against resource shocks, much like cosmologists test dark‑matter models against sudden changes in the expansion rate.

10. Future Directions and Experiments

The next decade promises precision cosmology that can differentiate between a cold hidden sector (\( \xi \ll 1 \)) and a warm one (\( \xi \sim 1 \)). Complementary laboratory experiments will probe the tiny portal couplings that drive freeze‑in. Together, these efforts will either pin down the thermal history of the dark side or push us toward more exotic frameworks (e.g., asymmetric dark matter or dark‑matter‑baryon interactions).

Why It Matters

Understanding freeze‑out and freeze‑in is more than an academic exercise; it tells us how the Universe allocated its mass after the Big Bang, and it informs us about the limits of interaction that nature tolerates. In bee colonies, the same principles of resource limitation, feedback regulation, and gradual seeding keep ecosystems resilient. For self‑governing AI agents, modeling analogous thermal decoupling can guide the design of robust, low‑interference coordination protocols.

By charting the hidden sector’s temperature evolution, we gain a universal framework that links the microscopic world of particle interactions with the macroscopic realms of ecology and technology. In doing so, we not only edge closer to uncovering the nature of dark matter, but we also learn how to steward the delicate balances—whether in a hive, a data center, or the very cosmos itself.