Dark Photon Solitons

• Dark photon solitons are localized, self-gravitating configurations of massive vector fields that behave as coherent dark matter condensates.
• Their dynamics are governed by a vector Schrödinger–Poisson system, leading to distinct polarization effects and spin-dependent self-interactions.
• Formation in the early universe, stability criteria, and radiative signatures offer observable probes for ultralight vector dark matter substructures.

Dark photon solitons are localized, self-gravitating, nonrelativistic configurations of a massive spin-1 field, usually described as a dark photon or Proca boson. In the dark-matter context they appear as coherent ground-state clumps of wave-like vector dark matter, and the literature refers to closely related objects as “dark photon stars,” “Proca stars,” and “vector solitons” (Gorghetto et al., 2022, Amin et al., 2023, Zhang, 2024). Across these formulations, the common structure is a bound condensate whose leading dynamics are governed by a Schrödinger–Poisson system, with polarization, spin density, self-interactions, and nonminimal gravitational interactions controlling the detailed profile, stability, and phenomenology (Zhang, 2024).

1. Terminology and scope

The term dark photon soliton denotes an astrophysical or cosmological soliton built from a massive vector field in the dark sector. In the minimal setup the field is a Proca boson with mass $m$, and in the weak-field, nonrelativistic regime its bound states are described by three complex mode functions corresponding to the vector components (Gorghetto et al., 2022). When the emphasis is on compact-object phenomenology, the same solutions are often called “dark photon stars” or “Proca stars”; when the emphasis is on wave mechanics and EFT, the name “vector soliton” is common (Gorghetto et al., 2022, Amin et al., 2023).

A useful terminological distinction is that this subject is not the same as the nonlinear-optics literature on “dark” photonic solitons. “Dark topological valley Hall edge solitons,” for example, are intensity-dip edge excitations in a photonic topological insulator formed at domain walls between honeycomb lattices with broken inversion symmetry, described by a $(2+1)$-dimensional nonlinear Schrödinger equation in waveguide arrays (Ren et al., 2022). This is a distinct usage of dark: there it refers to a dark soliton on a finite optical background, not to a dark-sector vector field.

Within dark-matter theory, the broader unifying point is that solitons are a generic prediction for ultralight dark matter. A “unified view” has shown that real or complex, scalar or vector dark matter can exhibit universal nonrelativistic soliton properties such as conserved charges, mass–radius relations, stability criteria, and approximate profiles, while vector-specific effects arise from macroscopic spin density and polarization-dependent self-interactions (Zhang, 2024).

2. Field-theoretic description and nonrelativistic limit

In a curved FRW spacetime, the minimal dark-photon setup starts from the Proca action

$$S = \int d^4x \sqrt{-g}\left[-\frac14 F_{\mu\nu}F^{\mu\nu} + \frac12 m^2 A_\mu A^\mu \right],$$

with $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$. Variation yields the Proca equation $\nabla_\nu F^{\nu\mu}+m^2A^\mu=0$ together with the constraint $\nabla_\mu A^\mu=0$ (Gorghetto et al., 2022). In the nonrelativistic, weak-field limit one decomposes the spatial components as

$$A_i(x,t)=\mathrm{Re}\!\left[\frac{\psi_i(x,t)}{\sqrt{2m^2a^3}}e^{-imt}\right],$$

and introduces a Newtonian potential $\Phi$ satisfying $\nabla^2\Phi=4\pi G(\rho-\langle\rho\rangle)$ (Gorghetto et al., 2022).

The resulting dynamics are a vector Schrödinger–Poisson system,

$$i\partial_t\psi_i=\left[-\frac{\nabla^2}{2m}+m\Phi\right]\psi_i,\qquad \nabla^2\Phi=4\pi G\,\frac{\sum_i|\psi_i|^2-\langle|\psi_i|^2\rangle}{a},$$

which is formally identical to three copies of a scalar Schrödinger–Poisson problem after eliminating $(2+1)$0, but with an intrinsic spin

$(2+1)$1

that can take any value $(2+1)$2, where $(2+1)$3 is the particle number (Gorghetto et al., 2022). The ground-state soliton has zero orbital angular momentum but can carry macroscopic spin.

A more general complex-vector theory augments the Proca action by a quartic self-interaction and nonminimal couplings to curvature,

$(2+1)$4

In the nonrelativistic limit this produces a vector Gross–Pitaevskii–Poisson system with a spin density

$(2+1)$5

and a gradient-dependent term proportional to $(2+1)$6, where $(2+1)$7 (Zhang, 2024). The associated conserved quantities include particle number $(2+1)$8, total spin $(2+1)$9, gravitational mass $$S = \int d^4x \sqrt{-g}\left[-\frac14 F_{\mu\nu}F^{\mu\nu} + \frac12 m^2 A_\mu A^\mu \right],$$0, and an energy functional containing a spin-spin contribution proportional to $$S = \int d^4x \sqrt{-g}\left[-\frac14 F_{\mu\nu}F^{\mu\nu} + \frac12 m^2 A_\mu A^\mu \right],$$1 (Zhang, 2024). This is the precise sense in which dark photon solitons generalize scalar Schrödinger–Poisson solitons.

3. Formation in the early universe and dark-matter substructure

One concrete formation channel begins during inflation. If a massive dark photon is present while $$S = \int d^4x \sqrt{-g}\left[-\frac14 F_{\mu\nu}F^{\mu\nu} + \frac12 m^2 A_\mu A^\mu \right],$$2, the longitudinal mode behaves like a light scalar and acquires a nearly scale-invariant spectrum at horizon exit,

$$S = \int d^4x \sqrt{-g}\left[-\frac14 F_{\mu\nu}F^{\mu\nu} + \frac12 m^2 A_\mu A^\mu \right],$$3

After reheating and radiation domination, the perturbations redshift differently above and below the scale $$S = \int d^4x \sqrt{-g}\left[-\frac14 F_{\mu\nu}F^{\mu\nu} + \frac12 m^2 A_\mu A^\mu \right],$$4 defined by $$S = \int d^4x \sqrt{-g}\left[-\frac14 F_{\mu\nu}F^{\mu\nu} + \frac12 m^2 A_\mu A^\mu \right],$$5 (Gorghetto et al., 2022). In parallel, quantum pressure enters the Euler equation through the Bohm potential and defines a comoving quantum Jeans scale

$$S = \int d^4x \sqrt{-g}\left[-\frac14 F_{\mu\nu}F^{\mu\nu} + \frac12 m^2 A_\mu A^\mu \right],$$6

A key result is the parametric coincidence $$S = \int d^4x \sqrt{-g}\left[-\frac14 F_{\mu\nu}F^{\mu\nu} + \frac12 m^2 A_\mu A^\mu \right],$$7 at matter–radiation equality, so perturbations at $$S = \int d^4x \sqrt{-g}\left[-\frac14 F_{\mu\nu}F^{\mu\nu} + \frac12 m^2 A_\mu A^\mu \right],$$8 first become nonlinear just as quantum pressure becomes subdominant (Gorghetto et al., 2022).

This mechanism yields a rich hierarchy of substructure rather than a smooth dark-matter fluid. A substantial fraction of the dark matter collapses into gravitationally bound solitons, which are fully quantum coherent objects (Gorghetto et al., 2022). The characteristic masses are

$$S = \int d^4x \sqrt{-g}\left[-\frac14 F_{\mu\nu}F^{\mu\nu} + \frac12 m^2 A_\mu A^\mu \right],$$9

and the central densities are typically a factor $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$0 larger than the local background dark matter density, with the detailed study reporting a range of $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$1–$F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$2 times the ambient density depending on the population considered (Gorghetto et al., 2022). Numerical simulations find formation with $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$3, $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$4, and today’s most abundant solitons near $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$5 (Gorghetto et al., 2022).

The same simulations indicate that a comparable fraction of the energy density is initially stored in, and subsequently radiated from, long-lived quasi-normal modes, and that solitons are surrounded by characteristic “fuzzy” halos in which wave effects are enhanced relative to virialized dark matter expectations (Gorghetto et al., 2022). The soliton mass function $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$6 peaks at $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$7 and integrates to $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$8 of the total dark matter, while lower-density compact halos of mass $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$9 are also produced at larger scales (Gorghetto et al., 2022). A plausible implication is that dark photon solitons should be treated not as rare exotica but as a potentially generic component of vector dark matter substructure whenever the production history populates the relevant wave modes.

4. Internal structure, polarization, and stability

For isolated bound states, a general polarized vector-soliton ansatz may be written as

$\nabla_\nu F^{\nu\mu}+m^2A^\mu=0$0

with $\nabla_\nu F^{\nu\mu}+m^2A^\mu=0$1, $\nabla_\nu F^{\nu\mu}+m^2A^\mu=0$2 a real spherically symmetric profile, and $\nabla_\nu F^{\nu\mu}+m^2A^\mu=0$3 specifying the polarization or spin state (Amin et al., 2023). Empirically, the profile is fitted by

$\nabla_\nu F^{\nu\mu}+m^2A^\mu=0$4

and the mass, binding energy, and radius scale as

$\nabla_\nu F^{\nu\mu}+m^2A^\mu=0$5

(Amin et al., 2023). The choice of $\nabla_\nu F^{\nu\mu}+m^2A^\mu=0$6 distinguishes, for example, linear polarization from a purely circular soliton with $\nabla_\nu F^{\nu\mu}+m^2A^\mu=0$7.

The cosmological analysis of dark photon stars gives a complementary structural description. Stationary, minimal-energy solitons obey

$\nabla_\nu F^{\nu\mu}+m^2A^\mu=0$8

so $\nabla_\nu F^{\nu\mu}+m^2A^\mu=0$9 and the central density scales as $\nabla_\mu A^\mu=0$0 (Gorghetto et al., 2022). The physical radius is

$\nabla_\mu A^\mu=0$1

and the density profile is well matched by

$\nabla_\mu A^\mu=0$2

Outside the core, the surrounding fuzzy halo is described by an NFW form with $\nabla_\mu A^\mu=0$3 and $\nabla_\mu A^\mu=0$4 (Gorghetto et al., 2022).

Self-interactions and nonminimal gravitational interactions change these relations in polarization-dependent ways. For a stationary polarized ansatz

$\nabla_\mu A^\mu=0$5

with $\nabla_\mu A^\mu=0$6, the radial profile equation contains

$\nabla_\mu A^\mu=0$7

so the effective quartic self-interaction depends on polarization (Zhang, 2024). In the minimal case one obtains the approximate analytic relations

$\nabla_\mu A^\mu=0$8

while in the strong repulsive Thomas–Fermi limit with $\nabla_\mu A^\mu=0$9,

$$A_i(x,t)=\mathrm{Re}\!\left[\frac{\psi_i(x,t)}{\sqrt{2m^2a^3}}e^{-imt}\right],$$0

(Zhang, 2024).

Stability is controlled by the mass–radius curve. The linear criterion is

$$A_i(x,t)=\mathrm{Re}\!\left[\frac{\psi_i(x,t)}{\sqrt{2m^2a^3}}e^{-imt}\right],$$1

so the negative-slope branch is stable and the turnover point $$A_i(x,t)=\mathrm{Re}\!\left[\frac{\psi_i(x,t)}{\sqrt{2m^2a^3}}e^{-imt}\right],$$2 marks the onset of instability (Zhang, 2024). Attractive self-interactions $$A_i(x,t)=\mathrm{Re}\!\left[\frac{\psi_i(x,t)}{\sqrt{2m^2a^3}}e^{-imt}\right],$$3 yield a maximum mass and minimum radius, while $$A_i(x,t)=\mathrm{Re}\!\left[\frac{\psi_i(x,t)}{\sqrt{2m^2a^3}}e^{-imt}\right],$$4 produces an analogous turnover from nonminimal gravity. For $$A_i(x,t)=\mathrm{Re}\!\left[\frac{\psi_i(x,t)}{\sqrt{2m^2a^3}}e^{-imt}\right],$$5, regularity imposes an absolute amplitude bound,

$$A_i(x,t)=\mathrm{Re}\!\left[\frac{\psi_i(x,t)}{\sqrt{2m^2a^3}}e^{-imt}\right],$$6

because the effective inertia denominator $$A_i(x,t)=\mathrm{Re}\!\left[\frac{\psi_i(x,t)}{\sqrt{2m^2a^3}}e^{-imt}\right],$$7 must remain positive (Zhang, 2024). This bound caps the central amplitude and therefore constrains the high-mass tail of the soliton distribution.

5. Couplings to photons and radiative processes

A central question is whether dark photon solitons emit observable electromagnetic radiation. In vacuo, gauge–kinetic mixing

$$A_i(x,t)=\mathrm{Re}\!\left[\frac{\psi_i(x,t)}{\sqrt{2m^2a^3}}e^{-imt}\right],$$8

does not source photon emission from a coherent dark photon condensate (Amin et al., 2023). The leading interactions that do so in the isolated-soliton problem are instead dimension-6 operators

$$A_i(x,t)=\mathrm{Re}\!\left[\frac{\psi_i(x,t)}{\sqrt{2m^2a^3}}e^{-imt}\right],$$9

with five independent gauge- and Lorentz-invariant structures built from $\Phi$0, $\Phi$1, and $\Phi$2 (Amin et al., 2023). In the time-periodic background of a vector soliton, the photon mode equations acquire periodic coefficients, and Floquet analysis shows exponential growth for modes near the resonance band $\Phi$3,

$\Phi$4

For $\Phi$5, the maximal growth rate in the linear-polarization case is

$\Phi$6

whereas for circular polarization only $\Phi$7 and $\Phi$8 resonate and $\Phi$9 is suppressed at $\nabla^2\Phi=4\pi G(\rho-\langle\rho\rangle)$0 (Amin et al., 2023).

The emitted radiation carries operator- and polarization-specific signatures. The central frequency is

$\nabla^2\Phi=4\pi G(\rho-\langle\rho\rangle)$1

with bandwidth

$\nabla^2\Phi=4\pi G(\rho-\langle\rho\rangle)$2

For $\nabla^2\Phi=4\pi G(\rho-\langle\rho\rangle)$3 and $\nabla^2\Phi=4\pi G(\rho-\langle\rho\rangle)$4, the emission is isotropic and unpolarized; for $\nabla^2\Phi=4\pi G(\rho-\langle\rho\rangle)$5 and $\nabla^2\Phi=4\pi G(\rho-\langle\rho\rangle)$6 with a linearly polarized soliton along $\nabla^2\Phi=4\pi G(\rho-\langle\rho\rangle)$7, it peaks in the equatorial plane $\nabla^2\Phi=4\pi G(\rho-\langle\rho\rangle)$8, with $\nabla^2\Phi=4\pi G(\rho-\langle\rho\rangle)$9 linearly polarized parallel to $$i\partial_t\psi_i=\left[-\frac{\nabla^2}{2m}+m\Phi\right]\psi_i,\qquad \nabla^2\Phi=4\pi G\,\frac{\sum_i|\psi_i|^2-\langle|\psi_i|^2\rangle}{a},$$0 and $$i\partial_t\psi_i=\left[-\frac{\nabla^2}{2m}+m\Phi\right]\psi_i,\qquad \nabla^2\Phi=4\pi G\,\frac{\sum_i|\psi_i|^2-\langle|\psi_i|^2\rangle}{a},$$1 polarized in the azimuthal $$i\partial_t\psi_i=\left[-\frac{\nabla^2}{2m}+m\Phi\right]\psi_i,\qquad \nabla^2\Phi=4\pi G\,\frac{\sum_i|\psi_i|^2-\langle|\psi_i|^2\rangle}{a},$$2 direction; for a circular soliton in the $$i\partial_t\psi_i=\left[-\frac{\nabla^2}{2m}+m\Phi\right]\psi_i,\qquad \nabla^2\Phi=4\pi G\,\frac{\sum_i|\psi_i|^2-\langle|\psi_i|^2\rangle}{a},$$3 plane, only $$i\partial_t\psi_i=\left[-\frac{\nabla^2}{2m}+m\Phi\right]\psi_i,\qquad \nabla^2\Phi=4\pi G\,\frac{\sum_i|\psi_i|^2-\langle|\psi_i|^2\rangle}{a},$$4 and $$i\partial_t\psi_i=\left[-\frac{\nabla^2}{2m}+m\Phi\right]\psi_i,\qquad \nabla^2\Phi=4\pi G\,\frac{\sum_i|\psi_i|^2-\langle|\psi_i|^2\rangle}{a},$$5 resonate, producing circularly polarized radiation along $$i\partial_t\psi_i=\left[-\frac{\nabla^2}{2m}+m\Phi\right]\psi_i,\qquad \nabla^2\Phi=4\pi G\,\frac{\sum_i|\psi_i|^2-\langle|\psi_i|^2\rangle}{a},$$6 with the same handedness as the soliton (Amin et al., 2023). The resonance condition is $$i\partial_t\psi_i=\left[-\frac{\nabla^2}{2m}+m\Phi\right]\psi_i,\qquad \nabla^2\Phi=4\pi G\,\frac{\sum_i|\psi_i|^2-\langle|\psi_i|^2\rangle}{a},$$7, equivalently $$i\partial_t\psi_i=\left[-\frac{\nabla^2}{2m}+m\Phi\right]\psi_i,\qquad \nabla^2\Phi=4\pi G\,\frac{\sum_i|\psi_i|^2-\langle|\psi_i|^2\rangle}{a},$$8 (Amin et al., 2023).

A distinct radiative channel appears when external electromagnetic fields or charge densities are present. In that case, a dimension-6 dipole operator

$$i\partial_t\psi_i=\left[-\frac{\nabla^2}{2m}+m\Phi\right]\psi_i,\qquad \nabla^2\Phi=4\pi G\,\frac{\sum_i|\psi_i|^2-\langle|\psi_i|^2\rangle}{a},$$9

allows linearly polarized solitons to induce an oscillating charge/current at $(2+1)$00, while gauge kinetic mixing implies $(2+1)$01, so electrons in the soliton background experience an oscillating Lorentz force and produce a current at $(2+1)$02 (Schiappacasse et al., 8 Jan 2026). For both mechanisms, plasma effects are crucial: in the resonant limits $(2+1)$03 for the dipole channel and $(2+1)$04 for the mixing channel, the exponential suppression from finite soliton size is removed and replaced by a milder power-law behavior (Schiappacasse et al., 8 Jan 2026). This suggests that compact magnetospheres and dense plasma environments can be more important observationally than empty space.

6. Observational signatures, survival, and open distinctions

The survival of dark photon solitons depends both on gravitational disruption and on radiative decay. From the gravitational side, a clump of mean density $(2+1)$05 survives tidal forces from a host of mean density $(2+1)$06 if $(2+1)$07. In the Milky Way, where $(2+1)$08, the soliton cores with $(2+1)$09–$(2+1)$10 are robust, and fuzzy halos survive down to local densities $(2+1)$11 (Gorghetto et al., 2022). The same study argues that, at minimum, the solitons are likely to survive to the present day without being tidally disrupted (Gorghetto et al., 2022).

Radiative stability is more model-dependent. If the parametric-resonance channel is open, an isolated soliton decays on a timescale

$(2+1)$12

so long-lived present-day objects require $(2+1)$13 and therefore lie below the critical mass

$(2+1)$14

(Amin et al., 2023). Heavier solitons would have evaporated through photon emission. This resolves a potential misconception: kinetic mixing or higher-dimensional couplings do not automatically imply persistent emission from every extant soliton; for isolated objects in vacuo, the kinematics and coupling regime matter sharply (Amin et al., 2023).

Detection strategies therefore span several regimes. Passing solitons may be probed by pulsar timing arrays, with $(2+1)$15 and encounter rate $(2+1)$16 for $(2+1)$17; microlensing is sensitive to compact halos of mass $(2+1)$18 and to solitons if $(2+1)$19; direct detection with resonant cavities or atomic clocks can see transient density boosts $(2+1)$20–$(2+1)$21 over minutes to hours (Gorghetto et al., 2022). If dimension-6 photon couplings are present, mergers of two subcritical solitons can temporarily exceed $(2+1)$22, triggering monochromatic radio bursts at $(2+1)$23, with $(2+1)$24–$(2+1)$25, duration $(2+1)$26, and brightness above $(2+1)$27 even at cosmological distances; the polarization then encodes both the operator type and the soliton spin state (Amin et al., 2023).

Environmental conversion channels motivate radio searches in magnetospheres and plasmas. Radio telescopes such as SKA and FAST can probe line fluxes of order $(2+1)$28 at $(2+1)$29–$(2+1)$30, and a resonant neutron-star magnetosphere with $(2+1)$31 and $(2+1)$32–$(2+1)$33 yields

$(2+1)$34

for $(2+1)$35 and $(2+1)$36–$(2+1)$37 (Schiappacasse et al., 8 Jan 2026). Space-based arrays such as OLFAR would extend coverage to $(2+1)$38–$(2+1)$39, corresponding to $(2+1)$40–$(2+1)$41, a band inaccessible from the ground because of the ionosphere (Schiappacasse et al., 8 Jan 2026). The same analysis identifies open parameter windows at $(2+1)$42–$(2+1)$43, $(2+1)$44–$(2+1)$45, and $(2+1)$46–$(2+1)$47, especially when resonant enhancement in magnetized or plasma-filled environments is available (Schiappacasse et al., 8 Jan 2026).

The main conceptual distinction that remains active in the literature is not whether vector solitons exist in principle, but how much of their phenomenology is universal and how much is genuinely vector-specific. The unified treatment indicates that mass–radius relations, profile equations, and stability criteria are largely shared with scalar solitons in the nonrelativistic regime, whereas spin density, polarization dependence, and gradient-induced amplitude bounds are distinctive to vector dark matter (Zhang, 2024). This suggests that observational discrimination will likely rely on precisely those vector-specific handles: polarization-dependent radiation, spin-sensitive self-interaction effects, and merger signatures of spinning solitons.