Cosmological Solutions in dRGT

• Cosmological solutions in dRGT are derived from a ghost-free massive gravity framework that imposes strict FLRW constraints, leading to selective self-accelerating and static branches.
• Methodologies such as singular reference metrics and inhomogeneous Stückelberg profiles are employed to evade no-go theorems and enable viable de Sitter expansions.
• Extended models including BD-dRGT and DBI-dRGT stabilize perturbations and generate effective fluid components that mimic dark energy and dark matter effects for observational concordance.

The de Rham-Gabadadze-Tolley (dRGT) theory of massive gravity supplies a ghost-free nonlinear completion of Fierz-Pauli massive gravity, with the graviton acquiring a strictly nonzero mass $m$. A central avenue of investigation in this framework is its cosmological sector: can the theory accommodate spatially homogeneous and isotropic Friedmann-Lemaître-Robertson-Walker (FLRW) solutions that fit the universe’s observed expansion and cosmic acceleration? Addressing this requires detailed scrutiny of background solutions, their stability, phenomenology, and data concordance, both in the pure dRGT theory and in its extended incarnations. The following sections systematically survey the technical developments and key results in cosmological solutions within dRGT massive gravity and its extensions.

1. Foundation: dRGT Action and FLRW No-Go Results

The core dRGT action in four dimensions is

$$S = \frac{M_{\mathrm{Pl}}^2}{2} \int d^4x\, \sqrt{-g}\left[ R[g] - m^2\sum_{n=0}^4\beta_n\,e_n(\sqrt{g^{-1}f}) \right] + S_{\text{matter}}[g, \Psi],$$

where $g_{\mu\nu}$ is the physical metric, $f_{\mu\nu}$ the fixed fiducial metric, $\beta_n$ are dimensionless parameters, and $e_n$ the elementary symmetric polynomials constructed from the square root matrix $\sqrt{g^{-1}f}$ (Hinterbichler, 2017, Zhang et al., 2019, Hinterbichler et al., 2013). The theory’s ghost freedom is preserved by the specific structure of the interaction terms.

For standard FLRW ansätze—$g_{\mu\nu}$ and $f_{\mu\nu}$ both homogeneous and isotropic—the pure dRGT theory does not admit nontrivial flat or closed FLRW solutions (i.e., with $k=0$ or $$S = \frac{M_{\mathrm{Pl}}^2}{2} \int d^4x\, \sqrt{-g}\left[ R[g] - m^2\sum_{n=0}^4\beta_n\,e_n(\sqrt{g^{-1}f}) \right] + S_{\text{matter}}[g, \Psi],$$0): the system imposes strong constraints that force the physical scale factor $$S = \frac{M_{\mathrm{Pl}}^2}{2} \int d^4x\, \sqrt{-g}\left[ R[g] - m^2\sum_{n=0}^4\beta_n\,e_n(\sqrt{g^{-1}f}) \right] + S_{\text{matter}}[g, \Psi],$$1 to be static, except for Minkowski space (Hinterbichler, 2017, Hinterbichler et al., 2013, Heisenberg et al., 2024). Open ($$S = \frac{M_{\mathrm{Pl}}^2}{2} \int d^4x\, \sqrt{-g}\left[ R[g] - m^2\sum_{n=0}^4\beta_n\,e_n(\sqrt{g^{-1}f}) \right] + S_{\text{matter}}[g, \Psi],$$2) FLRW models do exist, with two solution branches:

• Normal branch: Corresponds to generalized Milne universes (hyperbolic slicing of Minkowski), with no genuine cosmic acceleration.
• Self-accelerating branch: An algebraic constraint fixes the ratio $$S = \frac{M_{\mathrm{Pl}}^2}{2} \int d^4x\, \sqrt{-g}\left[ R[g] - m^2\sum_{n=0}^4\beta_n\,e_n(\sqrt{g^{-1}f}) \right] + S_{\text{matter}}[g, \Psi],$$3 (with $$S = \frac{M_{\mathrm{Pl}}^2}{2} \int d^4x\, \sqrt{-g}\left[ R[g] - m^2\sum_{n=0}^4\beta_n\,e_n(\sqrt{g^{-1}f}) \right] + S_{\text{matter}}[g, \Psi],$$4 the Stüeckelberg temporal field), producing a genuine de Sitter expansion $$S = \frac{M_{\mathrm{Pl}}^2}{2} \int d^4x\, \sqrt{-g}\left[ R[g] - m^2\sum_{n=0}^4\beta_n\,e_n(\sqrt{g^{-1}f}) \right] + S_{\text{matter}}[g, \Psi],$$5, even in the absence of a bare cosmological constant (Hinterbichler, 2017, Hinterbichler et al., 2013).

Nevertheless, these FLRW backgrounds in pure dRGT confront severe obstacles in their linear perturbation theory. Specifically, vector and scalar modes can acquire vanishing kinetic terms (helicity-1 and helicity-0 become infinitely strongly coupled), rendering the theory unstable or non-predictive at cosmological scales (Hinterbichler, 2017).

2. Evading the No-Go: Singular Reference Metrics and Stückelberg Inhomogeneity

A crucial development for constructing isotropic and homogeneous cosmologies invoked either singular reference metrics or inhomogeneous, anisotropic Stüeckelberg profiles. These prescriptions relax the rigid structure in $$S = \frac{M_{\mathrm{Pl}}^2}{2} \int d^4x\, \sqrt{-g}\left[ R[g] - m^2\sum_{n=0}^4\beta_n\,e_n(\sqrt{g^{-1}f}) \right] + S_{\text{matter}}[g, \Psi],$$6 and the Stückelberg sector, opening nontrivial dynamical FLRW solutions (Heisenberg et al., 2024, Zhang et al., 2019).

• Singular Reference Metrics: Making $$S = \frac{M_{\mathrm{Pl}}^2}{2} \int d^4x\, \sqrt{-g}\left[ R[g] - m^2\sum_{n=0}^4\beta_n\,e_n(\sqrt{g^{-1}f}) \right] + S_{\text{matter}}[g, \Psi],$$7 (or more generally, dropping rank in the reference metric) allows the lapse term in the massive potential to vanish, evading the constraint that would otherwise freeze $$S = \frac{M_{\mathrm{Pl}}^2}{2} \int d^4x\, \sqrt{-g}\left[ R[g] - m^2\sum_{n=0}^4\beta_n\,e_n(\sqrt{g^{-1}f}) \right] + S_{\text{matter}}[g, \Psi],$$8 (Zhang et al., 2019). For example, $$S = \frac{M_{\mathrm{Pl}}^2}{2} \int d^4x\, \sqrt{-g}\left[ R[g] - m^2\sum_{n=0}^4\beta_n\,e_n(\sqrt{g^{-1}f}) \right] + S_{\text{matter}}[g, \Psi],$$9 yields a consistent Hamiltonian constraint and viable cosmological evolution.
• Inhomogeneous Stückelbergs: Enforcing FLRW symmetry only at the level of the square-root tensor $g_{\mu\nu}$0 (not pointwise in $g_{\mu\nu}$1) leads to two branches of flat FLRW solutions (Heisenberg et al., 2024):
  • The $g_{\mu\nu}$2-branch (“self-accelerating”) has $g_{\mu\nu}$3, manifesting as a cosmological constant-like graviton stress-energy.
  • The mixed branch (“bi-fluid”): The graviton sector acts as a mixture of several perfect fluids, producing more general behavior.

In both cases, the dynamical solutions acquire dark energy and dark matter-like effective fluids, with energy densities scaling as $g_{\mu\nu}$4, $g_{\mu\nu}$5, and $g_{\mu\nu}$6 respectively (Zhang et al., 2019).

3. Friedmann Equations and Effective Cosmic Components

Upon varying the dRGT action in mini-superspace ansatz, the modified Einstein equations yield generalized Friedmann and Raychaudhuri equations. For singular reference metrics as in (Zhang et al., 2019): $g_{\mu\nu}$7 where $g_{\mu\nu}$8 are the coefficients of the ghost-free interaction terms. The graviton “fluid” energy and pressure are

$g_{\mu\nu}$9

The emergent terms provide:

• $f_{\mu\nu}$0 scaling (“domain wall-like”): acts as dynamical dark energy.
• $f_{\mu\nu}$1 scaling: mimics dark matter.
• $f_{\mu\nu}$2 scaling: shifts effective curvature.

At late times, the $f_{\mu\nu}$3 term (with $f_{\mu\nu}$4) dominates and can drive cosmic acceleration without an explicit cosmological constant. For certain parameter choices, the effective equation of state $f_{\mu\nu}$5 can cross $f_{\mu\nu}$6 (Zhang et al., 2019).

4. Extensions with Additional Scalar Fields and Higher Dimensions

Several model extensions introduce new scalar DoFs or higher-dimensional origins to ameliorate pure dRGT’s pathologies:

• Brans-Dicke–dRGT (BD-dRGT): Augments the theory with a Brans-Dicke–type scalar $f_{\mu\nu}$7 (with coupling function $f_{\mu\nu}$8) and a rescaling in the graviton mass term. The full action in the Jordan and Einstein frames admits flat FLRW, self-accelerating de Sitter vacua, with all perturbations (tensor, vector, scalar) free of ghosts or gradient instabilities (Kazempour et al., 2022). Stability of vector and scalar modes is set by explicit no-ghost and sound-speed conditions on the kinetic matrices (e.g., $f_{\mu\nu}$9 for scalars).
• DBI-dRGT: Inclusion of a Dirac-Born-Infeld scalar with noncanonical kinetic structure, coupled to the graviton mass potential, yields self-accelerating de Sitter branches. Stabilizing vector and scalar perturbations requires constraints both on the graviton mass and on the DBI tension $\beta_n$0; phenomenological viability is determined by satisfying $\beta_n$1 and sound-speed bounds (Kazempour et al., 2022).
• Dimensional Reduction: Reducing higher-dimensional dRGT yields 4D effective actions combining quasi-dilaton or mass-varying massive gravity. Late-time de Sitter attractors arise with stabilization of the radion, and the onset of acceleration is controlled by fixed points in the autonomous dynamical system, characterized by the parameters $\beta_n$2, curvature of the extra dimensions, and scalar potential (Nakarachinda et al., 2017).

5. Observational Concordance and Parameter Constraints

Phenomenological viability necessitates consistency with supernovae (SNe Ia), cosmic microwave background (CMB), and baryon acoustic oscillation (BAO) data sets. For theories with singular reference metrics (Zhang et al., 2019), the Friedmann equation is rewritten in terms of observable energy densities: $\beta_n$3 with $\beta_n$4 and other scale-dependent terms determined by $\beta_n$5.

Markov chain Monte Carlo fits (e.g., via CosmoMC) yield allowed regions for the graviton mass and interaction coefficients, such as

$\beta_n$6

with $\beta_n$7 and $\beta_n$8 contours leaving significant parameter space for dRGT contributions (Zhang et al., 2019). In the BD-dRGT and DBI-dRGT extensions, additional constraints from gravitational wave propagation (via $\beta_n$9 eV from GW170817) and linear-perturbation stability further restrict the viable parameter regions (Kazempour et al., 2022, Kazempour et al., 2022).

6. Pathologies: Lapse Function, Big-Brake Singularities, and Strong Coupling

Certain parameter regions in the pure dRGT FLRW sector exhibit unphysical lapse function behavior—specifically, $e_n$0, rendering the metric signature ill-defined. Addition of a cosmological background energy density with negative pressure ($e_n$1) rectifies this, restoring a positive-definite lapse and guaranteeing acceleration at late times (Alatas et al., 2019).

“Big-brake” singularities—a divergence in $e_n$2 at finite $e_n$3—can arise generically in these cosmologies due to the structure of the modified Friedmann equation. Avoiding such future singularities demands parameter tuning such that the denominator of $e_n$4 does not vanish for finite $e_n$5 (Hinterbichler et al., 2013, Nakarachinda et al., 2017).

Strong coupling in scalar/vector perturbations is a generic feature in self-accelerating branches of both pure dRGT and certain mixed branches with inhomogeneous Stüeckelbergs: kinetic terms vanish (for example, $e_n$6 on $e_n$7-branch), and only tensor modes propagate at quadratic order (Heisenberg et al., 2024, Hinterbichler, 2017). Extensions with extra scalar degrees of freedom are engineered to alleviate or eliminate this pathology (Kazempour et al., 2022, Kazempour et al., 2022).

7. Comparative Overview of Solution Branches and Extensions

Model/Branch	FLRW Solutions	Stability	Late-Time Acceleration	Observational Fit
Pure dRGT (Minkowski $e_n$8, homogeneous Stückelbergs)	Flat/Closed: Only static; Open: two branches (Hinterbichler, 2017, Hinterbichler et al., 2013)	Self-acc. branch: strong coupling in scalar/vector (Hinterbichler, 2017, Heisenberg et al., 2024)	Self-acceleration in open branch	Open parameter space allowed (Zhang et al., 2019) (extensions preferred)
Singular $e_n$9 or inhomogeneous Stückelbergs	Flat FLRW now viable (Heisenberg et al., 2024, Zhang et al., 2019)	$\sqrt{g^{-1}f}$0-branch: tensor only; mixed: partial restoration, but with strong-coupling issues (Heisenberg et al., 2024)	Yes, via effective $\sqrt{g^{-1}f}$1	Yes, fits SNe, CMB, BAO (Zhang et al., 2019)
BD-dRGT, DBI-dRGT, higher-dimensional, quasi-dilaton, etc.	Flat FLRW with self-acceleration (Kazempour et al., 2022, Kazempour et al., 2022, Nakarachinda et al., 2017)	Ghost-/gradient-free for suitable parameters	Robust, dynamical in scalar sector	Consistent with late-time expansion, GW bounds

These results establish that, while pure dRGT admits only limited cosmological solutions compatible with observation and often suffers from severe strong-coupling pathologies, allowing singular reference metrics or leveraging scalar-tensor and higher-dimensional generalizations yields technically natural models admitting robust, healthy cosmic acceleration sourced by the graviton mass term, with phenomenology that can be tuned to current data.