Fuzzy Dark Sector (Interacting FDM)

• Fuzzy Dark Sector (FDS) is an interacting extension of FDM that models dark matter with coupled ultra-light fields exhibiting wave-like behavior.
• In its Abelian–Higgs realization, FDS produces a scalar Higgs excitation and a massive dark photon, leading to unique predictions in halo core dynamics and structure formation.
• FDS phenomenology influences galaxy cores, soliton stability, and gravitational-wave interactions, motivating new observational tests across cosmic scales.

Fuzzy Dark Sector (FDS) denotes an interacting extension of the single-component, non-interacting Fuzzy Dark Matter (FDM) paradigm in which dark matter is modeled not as one ultra-light bosonic field but as a coupled system of ultra-light degrees of freedom with wave-like behavior on astrophysical scales. In the concrete realization introduced as an ultra-light Abelian–Higgs model, the dark sector contains a complex scalar $\phi$ and an Abelian gauge field $A_\mu$; after symmetry breaking this yields a real scalar Higgs excitation $h$ and three physical polarizations of a massive dark photon, with linear structure formation controlled by a single characteristic scale $k_*$ and galactic halos exhibiting a diversity not present in canonical single-field FDM (Capanelli et al., 18 Sep 2025). In the broader literature, closely related fuzzy-dark-sector phenomenology includes the Schrödinger–Poisson and Gross–Pitaevskii–Poisson descriptions of ultralight bosons, solitonic core formation, multi-field halo diversity, environmental deformation of solitons, and observational tests ranging from dwarf galaxies and the Milky Way center to large-scale structure and gravitational-wave sources (Sreenath, 2018).

1. From canonical FDM to an interacting fuzzy dark sector

Canonical FDM models dark matter as one ultra-light bosonic field whose astrophysical phenomenology follows from the competition between gravity and quantum pressure. The FDS scenario generalizes this to a genuinely interacting, multi-component system. In the Abelian–Higgs realization, the relativistic action is

$$S=\int d^4x \sqrt{-g}\left[\frac{R}{16\pi G}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}+\frac{1}{2}|D_\mu\phi|^2-\frac{\lambda}{4!}(|\phi|^2-v^2)^2\right], \qquad D_\mu=\partial_\mu+i g A_\mu .$$

After spontaneous symmetry breaking,

$\phi=(h+v)e^{i\theta},$

and in unitary gauge $\theta=0$, the spectrum contains a scalar Higgs mode and a massive dark photon with

$m_A^2=g^2 v^2, \qquad m_h^2=2\lambda v^2 .$

Because the masses are ultra-light, $m_A,m_h\ll \mathrm{eV}$, the couplings satisfy $\lambda,g\ll 1$, placing the model in the feebly interacting massive particle regime (Capanelli et al., 18 Sep 2025).

In the nonrelativistic limit, the scalar and vector modes become coupled Schrödinger–Poisson fields. Their dynamics are governed by gravitational attraction, Higgs self-interaction, and Higgs–vector cross-interaction. The resulting equations show a mixed-sign interaction structure: the Higgs has a repulsive self-interaction from $A_\mu$0, the Higgs–vector interaction is attractive in the nonrelativistic equations, and the dark photon has no self-interaction. Each species separately conserves particle number in the EFT. A convenient dimensionless control parameter is

$A_\mu$1

which remains after rescaling the equations (Capanelli et al., 18 Sep 2025).

This explicit construction is presented as a framework rather than a single unique model. The same paper also sketches another realization involving a complex scalar with radial and angular modes, opposite-sign self-interactions, and a nontrivial kinetic coupling. A related but distinct route to fuzzy-dark-sector phenomenology is the two-field fuzzy dark matter model, in which two ultralight wavefunctions $A_\mu$2 and $A_\mu$3 share a common gravitational potential and form a minimal multi-axion realization motivated by the axiverse (Luu et al., 2024).

2. Wave mechanics, quantum pressure, and collapse

The dynamical foundation of FDS is inherited from the wave mechanics of FDM. In the standard ultralight-boson picture, dark matter is treated as a Bose–Einstein condensate or classical coherent field. In the nonrelativistic limit, the field obeys the Schrödinger–Poisson system; with self-interactions included, the appropriate formulation is the Gross–Pitaevskii–Poisson system,

$A_\mu$4

Under the Madelung transform, this becomes a fluid-like system with continuity, Euler, and Poisson equations. The Euler equation contains the ordinary gravitational force, a self-interaction pressure term, and the quantum potential

$A_\mu$5

which resists collapse and is the physical origin of wave-supported cores and the suppression of structure below the quantum Jeans scale (Sreenath, 2018).

The cosmological Schrödinger equation used in high-resolution FDM simulations makes the same mechanism explicit: $A_\mu$6 In the fluid form, the linear perturbation equation contains a scale-dependent term $A_\mu$7, and the Jeans wavenumber scales as $A_\mu$8, giving a Jeans mass $A_\mu$9 (Elgamal et al., 2023). This wave-supported scale is the basis for both small-scale cutoff phenomenology and core formation.

A tractable analytic realization appears in the spherical-collapse treatment of FDM. For a power-law overdensity profile, the shell radius $h$0 obeys

$h$1

The three terms correspond respectively to self-interaction, gravity, and quantum pressure. In the non-interacting case, the shell does not collapse to zero radius but oscillates between a nonzero $h$2 and $h$3, so quantum pressure prevents complete collapse in this toy model. Repulsive self-interaction pushes the turning points outward, whereas sufficiently attractive self-interaction can overcome quantum support and allow collapse (Sreenath, 2018). This suggests that FDS retains the characteristic wave-supported resistance to singular collapse even when the dark sector is generalized beyond a single field.

3. Linear cosmology and the suppression of small-scale structure

The central cosmological result of the Abelian–Higgs FDS construction is that, despite multiple coupled ultra-light fields, linear perturbations are governed by a single characteristic collapse scale $h$4. The matter power spectrum is written as

$h$5

with a transfer function of the standard fuzzy form,

$h$6

where $h$7 is built from the effective FDS scale $h$8 rather than the single-field FDM Jeans scale. In the limits $h$9 and $k_*$0, $k_*$1 reduces to the expected single-species Jeans scales. On large scales, $k_*$2, so the model resembles CDM; on smaller scales, power is suppressed. The cutoff can shift to higher or lower $k_*$3 depending on the FDS parameters, so cosmological data can mimic either a heavier or a lighter FDM particle if interpreted in a single-field framework (Capanelli et al., 18 Sep 2025).

Large-scale numerical work on canonical FDM quantifies the corresponding suppression of halo and subhalo populations. Using AX-GADGET to evolve the quantum potential throughout cosmic time, one set of simulations finds a fitting formula for the cumulative FDM-to-CDM subhalo abundance ratio in Milky-Way-like hosts,

$k_*$4

with best-fit parameters $k_*$5, $k_*$6, $k_*$7, and $k_*$8. In those simulations, an extended FDM particle-mass interval can reproduce observed substructure counts and at the same time create substantial cores, $k_*$9 kpc, in dwarf-galaxy haloes of mass $$S=\int d^4x \sqrt{-g}\left[\frac{R}{16\pi G}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}+\frac{1}{2}|D_\mu\phi|^2-\frac{\lambda}{4!}(|\phi|^2-v^2)^2\right], \qquad D_\mu=\partial_\mu+i g A_\mu .$$0–$$S=\int d^4x \sqrt{-g}\left[\frac{R}{16\pi G}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}+\frac{1}{2}|D_\mu\phi|^2-\frac{\lambda}{4!}(|\phi|^2-v^2)^2\right], \qquad D_\mu=\partial_\mu+i g A_\mu .$$1; the authors identify a viable window roughly $$S=\int d^4x \sqrt{-g}\left[\frac{R}{16\pi G}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}+\frac{1}{2}|D_\mu\phi|^2-\frac{\lambda}{4!}(|\phi|^2-v^2)^2\right], \qquad D_\mu=\partial_\mu+i g A_\mu .$$2 (Elgamal et al., 2023).

Beyond two-point statistics, nearest-neighbour analysis has been proposed as a higher-order probe of the wave nature of ultralight dark matter. On halo catalogues at $$S=\int d^4x \sqrt{-g}\left[\frac{R}{16\pi G}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}+\frac{1}{2}|D_\mu\phi|^2-\frac{\lambda}{4!}(|\phi|^2-v^2)^2\right], \qquad D_\mu=\partial_\mu+i g A_\mu .$$3, with additional tests at $$S=\int d^4x \sqrt{-g}\left[\frac{R}{16\pi G}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}+\frac{1}{2}|D_\mu\phi|^2-\frac{\lambda}{4!}(|\phi|^2-v^2)^2\right], \qquad D_\mu=\partial_\mu+i g A_\mu .$$4 and $$S=\int d^4x \sqrt{-g}\left[\frac{R}{16\pi G}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}+\frac{1}{2}|D_\mu\phi|^2-\frac{\lambda}{4!}(|\phi|^2-v^2)^2\right], \qquad D_\mu=\partial_\mu+i g A_\mu .$$5, the spherical-contact function $$S=\int d^4x \sqrt{-g}\left[\frac{R}{16\pi G}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}+\frac{1}{2}|D_\mu\phi|^2-\frac{\lambda}{4!}(|\phi|^2-v^2)^2\right], \qquad D_\mu=\partial_\mu+i g A_\mu .$$6, nearest-neighbour function $$S=\int d^4x \sqrt{-g}\left[\frac{R}{16\pi G}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}+\frac{1}{2}|D_\mu\phi|^2-\frac{\lambda}{4!}(|\phi|^2-v^2)^2\right], \qquad D_\mu=\partial_\mu+i g A_\mu .$$7, and especially

$$S=\int d^4x \sqrt{-g}\left[\frac{R}{16\pi G}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}+\frac{1}{2}|D_\mu\phi|^2-\frac{\lambda}{4!}(|\phi|^2-v^2)^2\right], \qquad D_\mu=\partial_\mu+i g A_\mu .$$8

separate FDM from CDM and from the baryonic TNG50 simulations. The most discriminating observables are $$S=\int d^4x \sqrt{-g}\left[\frac{R}{16\pi G}-\frac{1}{4}F_{\mu\nu}F^{\mu\nu}+\frac{1}{2}|D_\mu\phi|^2-\frac{\lambda}{4!}(|\phi|^2-v^2)^2\right], \qquad D_\mu=\partial_\mu+i g A_\mu .$$9, the logarithmic skewness $\phi=(h+v)e^{i\theta},$0 of the nearest-neighbour distance PDF, and the angle-fit coefficient $\phi=(h+v)e^{i\theta},$1. At $\phi=(h+v)e^{i\theta},$2, the sign of $\phi=(h+v)e^{i\theta},$3 changes between CDM/TNG and FDM, and lower FDM mass strengthens the deviation (Kousha et al., 2023). A plausible implication is that an interacting FDS with an FDM-like transfer function but richer internal dynamics could preserve the small-scale cutoff in the cosmic power spectrum while generating more complex halo phenomenology than single-field FDM.

4. Solitons, halo diversity, and the breakdown of universality

The characteristic equilibrium configuration of canonical FDM is the soliton. In the idealized Schrödinger–Poisson system, the soliton profile is fixed by the particle mass up to scaling. A standard empirical fit is

$\phi=(h+v)e^{i\theta},$4

with

$\phi=(h+v)e^{i\theta},$5

In that idealized picture, once $\phi=(h+v)e^{i\theta},$6 is fixed, the soliton profile is essentially fixed up to scaling (Tan et al., 2024).

The central theme of more recent work is that this universality breaks down in realistic or multi-field settings. In two-field fuzzy dark matter, the shared gravitational potential couples a light field $\phi=(h+v)e^{i\theta},$7 and a heavy field $\phi=(h+v)e^{i\theta},$8, each with its own de Broglie scale and collapse history. A high-resolution cosmological simulation with $\phi=(h+v)e^{i\theta},$9, $\theta=0$0, and $\theta=0$1 finds three broad halo populations at $\theta=0$2: nested-core haloes in which both species contribute to the center, $\theta=0$3-only haloes, and rarer $\theta=0$4-only haloes created by asymmetric tidal stripping. Halo type depends on formation time, abundance ratio, mass hierarchy, and tidal interaction, and the diversity is not generic for all parameters: for $\theta=0$5 it appears near $\theta=0$6–$\theta=0$7, whereas other choices can eliminate the diversity (Luu et al., 2024).

Environmental effects can also diversify solitons even in a nominally single-field setting. One study identifies four mechanisms: a gravitoelectric field from a central black hole, a gravitomagnetic field, an extra denser and more compact FDM soliton, and an ellipsoidal baryon background. The gravitoelectric field is a considerable effect: a larger central black hole makes the soliton denser and more compact. Gravitomagnetic effects are very weak and negligible in practice compared with ordinary gravitational effects. In extreme-density-ratio soliton encounters, the smaller dense soliton is almost unaffected whereas the larger dilute soliton shrinks dramatically. In an ellipsoidal baryonic background motivated by the Milky Way, the fine structure of the baryon profile does not matter much, but the baryonic environment can noticeably deform the soliton, especially for lighter FDM particles (Tan et al., 2024).

The interacting Abelian–Higgs FDS produces an even stronger departure from single-field universality. Using a variational Gaussian ansatz, the halo core radii depend on the mass fraction carried by the Higgs and gauge sectors, the mass ratio $\theta=0$8, and the interaction strengths. The Higgs–vector attraction introduces a critical soliton mass above which the variational solution becomes complex, signaling instability. This instability is absent in the decoupled case (Capanelli et al., 18 Sep 2025). The broader phenomenological lesson is that the universal soliton of canonical FDM is a zeroth-order baseline rather than a generic prediction once realistic fuzzy-dark-sector structure is allowed.

5. Galaxy-scale phenomenology, empirical successes, and empirical tensions

The Milky Way center provides a direct astrophysical test of a key FDM prediction: the replacement of a cuspy inner halo by a quantum-pressure-supported soliton core. High-resolution hydrodynamical simulations of gas flow in a rigidly rotating barred Milky Way potential, performed with Athena++, use the Central Molecular Zone (CMZ) as the principal tracer of the inner mass distribution. The CMZ requires a dense central component to support $\theta=0$9 orbits and produce the observed ring-like structure. A compact nuclear bulge alone can reproduce the CMZ if its mass-to-light ratio varies with radius, but a less massive bulge plus a soliton also fits. The preferred soliton range is

$m_A^2=g^2 v^2, \qquad m_h^2=2\lambda v^2 .$0

with a representative model $m_A^2=g^2 v^2, \qquad m_h^2=2\lambda v^2 .$1, $m_A^2=g^2 v^2, \qquad m_h^2=2\lambda v^2 .$2 pc. Interpreted as FDM, this corresponds to $m_A^2=g^2 v^2, \qquad m_h^2=2\lambda v^2 .$3. The lower bound $m_A^2=g^2 v^2, \qquad m_h^2=2\lambda v^2 .$4 is relatively stronger, while the upper bound is weaker because a more compact nuclear bulge could reduce the need for a soliton (Li et al., 2020).

Galaxy rotation curves provide a broader but more contentious test. A Bayesian analysis of SPARC galaxies compares FDM against NFW, Burkert, and coreNFW halo profiles. In purely statistical terms, 72 out of 143 galaxies substantially prefer FDM over all other tested halo models, and no galaxy strongly prefers a cuspy NFW profile. However, the inferred boson mass is not universal: the 95% credible intervals for $m_A^2=g^2 v^2, \qquad m_h^2=2\lambda v^2 .$5 are often mutually inconsistent across galaxies, and best-fit core densities and radii show large scatter relative to the expected FDM scaling relations. The work therefore supports the soliton-like halo shape while challenging the simplest single-flavor, single-mass, universal-scaling version of FDM. It explicitly leaves room for modified fuzzy dark sector models, mixed FDM+CDM, and multi-state or multi-flavor scenarios (Khelashvili et al., 2022).

At the dwarf-galaxy scale, FDM has also been invoked to explain extreme stellar diffuseness through dynamical heating. In the almost dark dwarf galaxy Nube, numerical simulations using PyUltraLight model a solitonic core embedded in an NFW-like envelope and evolve $m_A^2=g^2 v^2, \qquad m_h^2=2\lambda v^2 .$6 test-particle stars for $m_A^2=g^2 v^2, \qquad m_h^2=2\lambda v^2 .$7 Gyr. The fluctuating FDM density field, granular halo structure, and random walk of the soliton center heat the stars and spread their projected distribution outward. Among three representative models, the one with $m_A^2=g^2 v^2, \qquad m_h^2=2\lambda v^2 .$8 and $m_A^2=g^2 v^2, \qquad m_h^2=2\lambda v^2 .$9 matches the observed 2D stellar density profile most closely. The paper notes that a future observation of a faint stellar population at $m_A,m_h\ll \mathrm{eV}$0 kpc would be a natural prediction of the FDM explanation, while also acknowledging tension with Ly$m_A,m_h\ll \mathrm{eV}$1 forest analyses, subhalo mass functions, and other dynamical-heating arguments (Yang et al., 2024).

Taken together, these studies do not establish a single canonical mass or halo architecture. They instead delineate a mixed empirical picture: soliton-like cores and wave-induced heating can fit specific systems well, yet the assumption of universal single-field scaling is under pressure.

6. Gravitational-wave channels, portals, and open problems

Fuzzy-dark-sector phenomenology is not restricted to static halo structure. A distinct interaction channel appears near strong gravitational-wave sources. For an ultralight axion-like field with $m_A,m_h\ll \mathrm{eV}$2, gravitational waves can induce parametric resonance in the curved-space Klein–Gordon equation, exciting the field in a spherical shell around the source rather than everywhere in the near zone. The resonance condition is

$m_A,m_h\ll \mathrm{eV}$3

and for $m_A,m_h\ll \mathrm{eV}$4,

$m_A,m_h\ll \mathrm{eV}$5

A finite-size cutoff requires roughly $m_A,m_h\ll \mathrm{eV}$6, and the dominant-mode threshold is estimated as $m_A,m_h\ll \mathrm{eV}$7 nHz, far below compact-binary frequencies. For a GW150914-like binary-black-hole waveform, field excitations arise toward the end of merger and in some cases persist slightly into ringdown. On a representative hypersurface, average potential-energy growth reaches about $m_A,m_h\ll \mathrm{eV}$8 or $m_A,m_h\ll \mathrm{eV}$9 for two waveform parameter sets, increasing to $\lambda,g\ll 1$0 or $\lambda,g\ll 1$1 if the strain is doubled. The observable implication is late-time waveform distortion through energy transfer from the GW to the fuzzy field (Dave et al., 2021).

The explicit FDS framework also raises questions of primordial production and coupling to the Standard Model. Production mechanisms mentioned for the Abelian–Higgs FDS include gravitational particle production, freeze-in through portals, parametric resonance, and an internal mechanism in which the Higgs-field phase $\lambda,g\ll 1$2 couples through $\lambda,g\ll 1$3, producing dark photons through a Chern–Simons or tachyonic instability. Proposed portals include a millicharged Higgs, quadratic Higgs–photon coupling, dark-photon current coupling $\lambda,g\ll 1$4, kinetic mixing, Higgs-portal terms such as $\lambda,g\ll 1$5, and $\lambda,g\ll 1$6. The same work stresses that a fully natural and unified production mechanism for both species remains open (Capanelli et al., 18 Sep 2025).

Several unresolved issues recur across the literature. Improved stellar mass-to-light measurements in the Galactic center are needed to break the degeneracy between a compact nuclear bulge and a less compact bulge plus a soliton (Li et al., 2020). Dark-matter-only simulations that support a viable FDM mass window still require fully hydrodynamical follow-up for final quantitative constraints (Elgamal et al., 2023). Nearest-neighbour probes of the cosmic web require larger FDM simulations and better observational reconstruction of the 3D web (Kousha et al., 2023). The cumulative implication is that FDS should be regarded not as a single prediction but as a structured research program: cosmological linear structure can remain effectively one-scale and FDM-like, while halo cores, soliton stability, and astrophysical observables become highly non-universal once interactions, multiple fields, or realistic environments are included.