Dark Stars: Theory and Implications

Updated 11 November 2025

Dark Stars are theoretical stellar objects powered by dark matter annihilation rather than nuclear fusion, forming in metal-free early Universe minihalos.
Their formation relies on DM heating outweighing baryonic cooling, with simulations confirming stable, accretion-driven growth under precise physical conditions.
Dark Stars influence cosmic reionization, serve as seeds for supermassive black holes, and exhibit unique spectral features observable by instruments like JWST.

Dark Stars (DSs) are a theoretical class of stellar objects formed in the early Universe whose power source during an initial evolutionary phase is the annihilation of dark matter (DM) particles, typically Weakly Interacting Massive Particles (WIMPs), rather than conventional nuclear fusion. DSs are predicted to form in metal-free (Population III) minihalos at high redshift (z ≈ 10–50), where the combination of baryonic collapse and DM physics enables unique stellar properties unparalleled in classical stellar evolution. While their existence remains observationally unconfirmed, recent advances have dramatically clarified their astrophysical, particle-physics, and cosmological implications, as well as established direct connections to observed phenomena such as the high-redshift luminous candidates seen by JWST.

1. Formation Mechanisms and Physical Conditions

The critical requirement for DS formation is a protostellar environment where DM annihilation heating overtakes all relevant baryonic cooling processes before the onset of fusion. In standard cosmology, primordial gas collapses inside DM minihalos (M_h ≈ 10^{6–10^8 M_⊙)} to central densities n_H ≳ 10^{13–10^{14 cm⁻³.}} Simultaneously, adiabatic contraction (AC) of the DM halo due to baryon infall increases the central WIMP density, ρχ, according to a scaling ρχ ≈ 5 GeV cm⁻³ (n_H/cm⁻³)^0.81 (Freese et al., 2015, 0903.3070). If DM is composed of self-annihilating WIMPs (benchmark m_χ ≈ 100 GeV, ⟨σv⟩ ≈ 3 × 10^{−26 cm³ s⁻¹),} the volumetric heating rate is

$Q_\mathrm{DM}(r) = f_Q\,⟨σv⟩\,ρ_χ(r)^2 / m_χ$

with f_Q ≈ 2/3 the energy deposition fraction; the remaining annihilation power is lost primarily to neutrinos. The DS phase is triggered if at every radius:

$Q_\mathrm{DM}(r) > Λ_\mathrm{cool}(n,T)$

where Λ_\mathrm{cool} is the dominant baryonic cooling rate (e.g. H₂ line cooling at early stages) (Gondolo et al., 2010). In most thermal WIMP scenarios, this condition is robustly met; only special particle-physics configurations, such as resonant s-channel annihilation at freezeout but not inside the protostar, coannihilations, or low-mass neutrinophilic DM, fail to ignite DSs (Gondolo et al., 2010).

Recent analyses demonstrate that self-interacting dark matter (SIDM) scenarios accelerate the gravothermal evolution of the DM halo in the presence of baryon collapse, facilitating the required central densities for DS formation even more efficiently (Wu et al., 2022). In all viable mechanisms, the nascent DS reaches hydrostatic equilibrium supported by DM annihilation luminosity rather than contraction or fusion.

2. Stellar Structure, Growth, and Evolution

Once WIMP heating dominates, the DS achieves radiative and hydrostatic equilibrium. DSs emerge as large (radii ~1–10 AU), cool (T_eff ≲ 10⁴ K), and extremely luminous (L ~ 10⁶–10⁷ L_⊙ for M ≲ 10³ M_⊙; up to L ~ 10⁹–10¹⁰ L_⊙ for supermassive cases) objects (Freese et al., 2015, Rindler-Daller et al., 2014, 0903.3070). The interior can be described by a polytropic equation of state (P = K ρ^1+1/n), with n≈1.5 (fully convective) at low mass and n≈3 (radiative diffusion) for supermassive DSs; self-consistent MESA calculations confirm this structure across the mass range (Rindler-Daller et al., 2014, Rindler-Daller et al., 2020).

The DS further grows by sustained accretion, enabled by the reduced ionizing flux (as cool T_eff depresses photoionization feedback) (0903.3070). Baryonic accretion rates, dictated by minihalo environment (SMH: Ṁ ≈ 10⁻³ M_⊙ yr⁻¹, LMH: Ṁ ≈ 10⁻¹ M_⊙ yr⁻¹), allow growth from ~1 M_⊙ to M_DS ≳ 10⁶ M_⊙ over 10⁵–10⁷ yr—while the DM fuel persists (Freese et al., 2015, Rindler-Daller et al., 2014). The total DM heating power at peak is generally

$L_\mathrm{DM} ≈ (⟨σv⟩/m_χ) ∫ ρ_χ^2\,dV$

with luminosity scaling almost linearly with stellar mass once a fully radiative (n=3) structure is established. The central temperatures remain orders of magnitude below those needed for fusion-burning stars of comparable mass (T_c ≲ 10⁶ K vs T_c ≳ 10⁸ K), which slows pre-main-sequence contraction and maintains large radii and low effective temperatures (Freese et al., 2015, Rindler-Daller et al., 2014).

Additional late-time DM supply may occur via WIMP capture through elastic scattering off baryonic nuclei, a process that can further prolong the DS phase if the local DM density and scattering cross-section are sufficiently high (0903.3070).

3. Astrophysical and Cosmological Role

DSs directly impact several key epochs and processes in cosmic history.

Reionization: DSs are exceptionally inefficient producers of ionizing UV radiation relative to fusion-powered Population III stars due to their low T_eff. In semi-analytical reionization models, introducing a DS phase with efficiency parameters (mass fraction f_DS ≳ 0.7–1, lifetimes t_DS ~ 10^8–10⁹ yr) shifts the completion of reionization to later epochs, reducing the integrated Thomson optical depth τ to the values measured by Planck (τ ≈ 0.054 ± 0.007) even for high Pop III star formation efficiencies (f_*,m ≳ 1%) (Gondolo et al., 2021). DSs thus alleviate tension between theoretical Pop III SFR predictions and CMB constraints.
Supermassive Black Hole (SMBH) Seeds: Once the DM fuel is exhausted or no longer replenished, DSs with M ≳ 10^{5–10^6 M_⊙} can collapse directly into black holes of similar mass without passing through a strong fusion phase (Freese et al., 2015, Schwemberger et al., 24 Dec 2024). Such massive remnants provide a natural explanation for the rapid assembly of ≳10^9 M_⊙ quasars observed at z ≳ 7. The abundance of such seeds is determined by the fraction of minihalos hosting DSs (f_SMDS ≈ 10^{−3–10⁻²⁾} (Schwemberger et al., 24 Dec 2024).
Diffuse Extragalactic Backgrounds: The cumulative emission from a DS population produces an imprint in the near- to mid-infrared extragalactic background light (EBL), peaking at λ ≈ 2–10 μm. Calculations show that models with high luminosity-to-mass ratios (LMR≳10⁴ L_⊙/M_⊙), high SFRs (SFR_norm≳10^{−4 M_⊙ yr⁻¹ Mpc⁻³),} or long-lived phases (Δt_DS≳10^8 yr) are excluded by EBL upper limits (~5–25 nW m⁻² sr⁻¹ at 2–10 μm), while moderate scenarios remain viable (Maurer et al., 2012).
Diffuse Neutrino Flux: The fraction of DM-annihilation energy lost to neutrinos in DSs can lead to a diffuse, high-z neutrino background at energies up to E_ν ≲ m_χ/(1+z_lim) (Schwemberger et al., 24 Dec 2024). Current atmospheric neutrino data from Super-Kamiokande and IceCube provide bounds on DS and annihilation parameter space that are independent and complementary to traditional indirect searches.

4. Spectral Signatures and Observational Diagnostics

DSs possess unique spectral features that distinguish them from other high-redshift stellar populations:

Continuum Slope: Cooler (T_eff ≲ 10⁴ K) DSs display redder near-infrared continua, with peak emission shifted to JWST NIRCam wavelengths at z ≳ 10 (Freese et al., 2015, Zackrisson, 2011). Supermassive DSs (M ≳ 10^6 M_⊙, L ≳ 10^10 L_⊙) at z ≈10–15 yield AB magnitudes m_AB ≈ 26–27, well within JWST's reach (Rindler-Daller et al., 2014, Zackrisson, 2011).
He II λ1640 Absorption: DS atmospheres (particularly SMDSs) are predicted to show He II λ1640 Å absorption in contrast to the nebular emission from star-forming Population III galaxies (Ilie et al., 9 May 2025, Ilie et al., 2023). The appearance of this feature in the highest-redshift JADES candidate, JADES-GS-z14-0, is a potential "smoking-gun" signature (Ilie et al., 9 May 2025).
Absence of Metals: In classical DSs, the absence of heavy elements eliminates metal lines and nebular forbidden transitions, although emerging observations suggest some DSs may co-exist with metal-rich environments, requiring extensions of standard formation models (Ilie et al., 9 May 2025).
Photometric Colors: DSs can be identified as dropouts in J or H bands due to the Lyman break at high redshift, with very red F200W–F356W colors and steep SEDs distinguishing them from both Pop III clusters and low-z impostors (Ilie et al., 2023, Zackrisson, 2011).
Pulsational Variability: Modeling of adiabatic and non-adiabatic pulsations predicts periods ranging from days to years (observer-frame, redshifted to months to years at z ~ 15) (Rindler-Daller et al., 2014, Rindler-Daller et al., 2020). These oscillations, if detected, could serve as unique temporal signatures and potentially as standard candles.

5. Numerical Simulations and Theoretical Stability

Three-dimensional hydrodynamic simulations incorporating AC, baryonic physics, and WIMP annihilation confirm the formation of central hydrostatic DSs at densities n_H ≳ 5 × 10^14 cm⁻³ (Gondolo et al., 2013). DM heating suppresses protostellar disk fragmentation, stabilizing accretion and promoting the growth of isolated DSs rather than multiple competing protostars, provided the minihalo is not strongly perturbed post-formation. Sink-particle approaches recover the macroscopic DS properties predicted by analytic models once the resolution limit is considered.

Self-consistent stellar evolution codes (MESA) further demonstrate that DSs are stable against dynamical and radiative instabilities up to at least M ≈ 10^3 M_⊙, without strong super-Eddington winds or mass loss, and only exhibit non-adiabatic acoustic-mode pulsational driving for M ≲ 200 M_⊙ (Rindler-Daller et al., 2020). The macroscopic properties (R, L, T_eff) are robust to more sophisticated treatments of the annihilation energy deposition.

6. Observational Searches, Machine Learning, and Current Candidates

The recent deployment and deep imaging surveys of JWST have ushered in an era of direct DS candidate identification. Advanced pipelines now employ feed-forward neural networks (FFNNs) trained on synthetic DS photometric and spectroscopic templates (using TLUSTY and CLOUDY for H/He atmospheres) (Mahmud et al., 6 Nov 2025, Siddiqa et al., 6 Nov 2025), enabling:

Robust and rapid classification of DS candidates among thousands of high-z JWST sources;
Fast parameter inference (DS mass, redshift, DM channel) surpassing traditional χ² minimization by ~10⁴× in speed;
Re-identification and subsequent spectroscopic confirmation of multiple JADES DS candidates (e.g., JADES-GS-z11-0, z13-0, z14-0, z14-1) (Ilie et al., 9 May 2025, Ilie et al., 2023, Siddiqa et al., 6 Nov 2025).

Current photometric and spectroscopic fits reinforce the viability of the DS hypothesis for several extreme-luminosity, high-redshift, point-source JWST objects—although the presence of metal lines in some cases suggests formation scenarios more complex than classical isolated DSs (Ilie et al., 9 May 2025).

Alternative "dark star" approaches in the context of general relativity introduce compact, horizonless objects (RDM stars, Planck-core dark stars) that do not collapse to black holes but are supported by quantum-gravity effects and produce extended r⁻² dark-matter–like halos (Nikitin, 2021). These objects can explain observed flat galactic rotation curves, with the emission of extremely low-energy (λ ∼ 10¹⁴ m) quanta, and accommodate fast radio bursts (FRBs) through accretion-induced events. These horizonless DSs challenge conventional interpretations and propose a unified explanation for galactic dark matter and certain high-energy phenomena.

Dark Stars thus comprise a multi-faceted and predictive theoretical framework connecting fundamental particle physics, the microphysics of the first star-forming environments, and a broad array of cosmological observations. Their existence and properties can be tested with ongoing and future data across the electromagnetic and neutrino spectra, with specific spectral and temporal signatures critical for their identification amid normal population III stars and galaxies.